On 9 October 2026, Bloomberg reported that Ramp, the corporate card and spend software company, was raising about $1.85 billion at roughly $60 billion. Four months earlier the number had been $44 billion. Eleven months earlier, $32 billion. In the same quarter, CB Insights counted only 726 fintech funding deals worldwide, the fewest in more than four years. Those two facts describe 2026 better than any trend report: money is plentiful for a few dozen companies that can show revenue, licences and growth at once, and scarce for almost everyone else. This list is about the first group, and about the smaller firms most likely to join it.

How we chose these fintech startups

We selected 14 private or recently listed fintech companies using measurable criteria as of October 2026: a dated funding round or valuation from 2025 or 2026, reported revenue or volume where available, regulatory licences, and product breadth across payments, stablecoin infrastructure, business banking and consumer finance. We grouped them by category rather than ranking them, and we mixed geographies deliberately: the US, Europe, India, Saudi Arabia, Nigeria, Mexico and Asia-Pacific.

We excluded companies where we could not verify a 2025 or 2026 number, and left out prediction markets and pure AI software firms. Every figure below is as reported by the linked source. For what investors test before writing these cheques, see how investors evaluate fintech startups.

Which payments infrastructure companies matter most in 2026?

In 2026 the payments infrastructure companies attracting the largest valuations are those that process trillions of dollars for other businesses and are adding stablecoin rails on top. Stripe remains the reference point for the whole sector, while Airwallex and Column show that cross-border accounts and bank-as-infrastructure models can grow quickly when paired with licences.

Stripe

Stripe is no longer a startup by any normal definition, but it sets the price of everything else in payments. Its February 2026 annual letter said businesses on Stripe processed $1.9 trillion in 2025, up 34%, and a tender offer for employee shares valued the company at $159 billion, according to The Block. The more interesting part for this list is its stablecoin unit: Bridge, bought for $1.1 billion in late 2024, received conditional OCC approval in February 2026 to become a federally chartered national bank.

Airwallex

Airwallex, founded in Australia, sells multi-currency business accounts, cards and payouts. It raised a $330 million Series G at $8 billion in December 2025, then closed a $320 million Series H led by Addition at an $11 billion valuation in June 2026, according to FinTech Futures. Annualised revenue was $1.3 billion as of March 2026, up 74% on a year earlier, and the company holds more than 85 licences, with the new money earmarked for more.

Column

Column is an FDIC-insured bank built to serve fintechs and developers, led by Plaid co-founder William Hockey. Forbes reported in February 2026 that it more than doubled revenue in 2025 to over $200 million. Sponsor-bank risk is one of the main constraints on US fintech growth, and a bank built for API customers answers it directly.

Which stablecoin infrastructure startups should you watch?

The stablecoin infrastructure startups to watch in 2026 are the ones that let ordinary businesses move, hold and spend dollar stablecoins without running crypto operations themselves. Rain issues stablecoin-linked cards, BVNK runs enterprise stablecoin accounts and is being acquired by Mastercard, and Stripe's Bridge is pursuing a US bank charter. Valuations here rose faster than anywhere else in fintech.

The backdrop is a stablecoin supply of about $306 billion as of October 2026, according to DefiLlama, with genuine stablecoin payments estimated at around $390 billion in 2025 by McKinsey and Artemis. Our wider list of stablecoin companies to watch goes deeper on issuers.

Rain

Rain lets companies launch Visa-linked cards that spend from stablecoin balances. It raised a $250 million Series C led by ICONIQ at a $1.95 billion valuation on 9 January 2026, according to The Block. It reported over $3 billion in annualised transactions, more than 200 partners including Western Union, and card issuance in over 150 countries. Forbes noted its valuation tripled in five months.

BVNK

BVNK, based in London, provides tools for businesses to send, receive, store and convert stablecoins and fiat across more than 130 countries. On 17 March 2026 Mastercard agreed to acquire it for up to $1.8 billion, including $300 million in contingent payments, with closing expected by the end of 2026, as reported by The Block. It is on this list because the deal is the clearest price signal yet for stablecoin payments infrastructure.

Which business finance and embedded finance startups are growing fastest?

Business finance is the fastest-repricing corner of fintech in 2026. Ramp's valuation rose from $32 billion to about $60 billion in under a year, and Mercury reported $650 million of 2025 revenue with a third straight profitable year. Both sell software-led accounts and cards to companies, and both earn largely from interchange and deposits.

Ramp

Ramp sells corporate cards, bill pay and spend management to US companies. Its June 2026 announcement said annualised revenue had passed $1 billion, the company was free-cash-flow positive and it served more than 70,000 customers; the October 2026 round, co-led by Dragoneer and Thrive Capital, valued it at about $60 billion before the new money, per FinTech Futures, citing Bloomberg. Bloomberg separately reported that annualised revenue had passed $1.5 billion by June; Ramp has not confirmed the October round.

Mercury

Mercury offers business checking and working capital loans, mainly to startups. Forbes reported 2025 revenue of $650 million and a third consecutive profitable year, and CB Insights listed a $200 million Series D among the largest digital banking rounds of the second quarter of 2026.

Which consumer fintech and digital banks are worth watching?

The consumer fintechs worth watching in 2026 combine a licence, a large customer base and profits. Revolut is now valued at $75 billion; Plata in Mexico doubled its valuation within seven months; and Moniepoint in Nigeria processes more than $22 billion a month. Each is moving from one product into a full bank.

Revolut

Revolut, headquartered in London, was valued at $75 billion in a November 2025 share sale led by Coatue, Greenoaks, Dragoneer and Fidelity, up from $48 billion in August 2025, according to TechCrunch. Its 2024 revenue rose 72% to $4 billion with $1 billion of net profit, and it has said it wants 100 million customers by mid-2027.

Plata

Plata is a Mexican digital lender that began with credit cards. It raised a $250 million Series B at $3.1 billion on 21 October 2025, more than double its $1.5 billion valuation in March 2025, and holds approval for a full banking licence from Mexico's CNBV granted in December 2024, per FinTech Futures.

Moniepoint

Moniepoint started with point-of-sale terminals for Nigerian merchants and now runs accounts, loans and payments. In October 2025 it completed a $200 million round with Visa, Development Partners International, Leapfrog and Google's Africa Investment Fund, at a valuation above $1 billion, and reported more than one billion transactions and over $22 billion of payment volume a month, according to Launch Base Africa. Acquisitions in the UK and Kenya point to its next markets; our Africa fintech list has the regional context.

Which fintech startups in India and the Gulf should you watch?

India and Saudi Arabia produced several of 2026's largest fintech financings. CRED raised about $900 million led by Meta in June 2026; Groww and Pine Labs listed in November 2025; and Tabby, the Gulf's largest buy-now-pay-later company, reached a $6.5 billion valuation in September 2026. Each shows a different route from startup to scale.

CRED

CRED, the Indian fintech, began as a credit card bill payment app and now offers lending, payments and wealth products. On 22 June 2026 it announced a Series H of about Rs 8,550 crore (roughly $900 million) led by Meta as a minority investor, at a post-money valuation of about $4.5 billion, with around 1.7 crore members using its products monthly, per The Daily Guardian.

Groww and Pine Labs

Both listed in India in November 2025, so neither is a startup any longer, but they set public-market benchmarks for every Indian fintech raising today. Groww's IPO raised about $748 million, the largest Indian fintech listing of 2025, at a valuation near $9 billion, and its shares closed 29% above issue on debut; it reported over 14 million active users. Pine Labs, the merchant payments and point-of-sale company, raised $440 million and closed its first day 14% higher, according to This Week in Fintech. Our India fintech companies guide covers the wider field.

Tabby

Tabby, headquartered in Saudi Arabia, offers instalment payments, cards and a shopping app across the Gulf. It raised $233 million at a $6.5 billion valuation in September 2026, led by Blue Pool Capital, according to FinTech Futures. The company reports more than 25 million registered users, between 65,000 and 70,000 merchant partners depending on the report, and more than $18 billion in annualised transaction volume. In February 2025 its Series E had valued it at $3.3 billion. See our UAE fintech companies piece for the neighbouring market.

How do these fintech companies compare?

Worked example: what are investors paying for each dollar of volume?

Valuation headlines are hard to compare, so it helps to divide them by something the company actually moves. Using only the figures above:

1. Stripe: $159 billion divided by $1.9 trillion of 2025 volume is about 8.4 cents of valuation per dollar processed in a year.
2. Tabby: $6.5 billion divided by $18 billion of annualised volume is about 36 cents per dollar. Instalment lenders earn far more per transaction than processors, and carry credit risk to match.
3. Rain: $1.95 billion divided by $3 billion of annualised card volume is about 65 cents per dollar, a premium for 38x volume growth in a year.
4. Ramp: about $60 billion against $1 billion of annualised revenue is roughly 60 times revenue; on the $1.5 billion run-rate Bloomberg reported, it is about 40 times.

None of these ratios says a company is cheap or expensive. They do show what investors are betting on: for Stripe, scale; for Tabby, margin per transaction; for Rain and Ramp, growth that has not yet slowed.

In 2026 the market pays for licences plus revenue plus growth; any two of the three gets a much smaller cheque.

What should you watch next in fintech?

The next six to nine months will test whether 2026's private valuations hold. These are the dated events that will tell us:

• Mastercard and BVNK, by end-2026. Closing is expected by the end of 2026, subject to regulatory approval. A completed deal would set the template for card networks buying stablecoin infrastructure, a theme in our stablecoin startup funding tracker.
• Bridge's national bank charter. The OCC's February 2026 approval was conditional. Final approval would give Stripe a federally chartered issuer and custodian.
• Tabby's listing decision. Its chief executive said in September 2026 that the profitable company does not need to force the timing of an IPO and will pick whichever market makes most sense; no date has been set.
• Airwallex's licence expansion. The June 2026 Series H was earmarked for licences in new markets, on top of the more than 85 it already holds.
• Revolut's customer target. It aims for 100 million customers by mid-2027.
• CB Insights Q3 2026 data. Due later this autumn, it will show whether the second quarter's four-year low in deal count was a trough or a trend.

This list is editorial research, not investment advice or an endorsement of any company. Company figures are as reported by the linked sources and may since have changed.

At about nine on a Friday night, the queue outside an exchange house in Dubai's Al Karama runs past the pharmacy next door. Most people in it already have a money transfer app on their phone. They are queuing anyway, some for cash, some for a rate they trust, some because the cashier knows their name. That scene is a useful corrective to the idea that the remittance business has already gone digital. The biggest companies in 2026 are a mix of 170-year-old agent networks, listed apps growing at 30% a year, and Gulf exchange houses with hundreds of branches. Size depends on which yardstick you pick.

How we chose and ranked these companies

We ranked companies by measurable scale as of October 2026: revenue from money transfer in the latest reported full year, send volume, and active customers, using annual reports, quarterly filings and company releases. Where a company is private, we use the figures it has disclosed and say so. We do not rank on price or service quality, and nothing here is a recommendation.

One caution before the numbers. These businesses do not report the same things. Western Union and Euronet count transactions. Wise reports cross-border volume and a financial year that ends in March. Remitly reports send volume. Gulf exchange houses report operating income in dirhams. We have kept each figure in the form the company gives it, with the period attached, rather than forcing them into a single column that would hide the differences.

For context, the World Bank's Remittance Prices Worldwide database put the global average cost of sending $200 at 6.36% in Q3 2025. Every company below is competing for a share of that margin, and the fastest-growing ones are the ones pushing it down.

Who are the largest remittance companies in 2026?

By 2025 money transfer revenue, the largest remittance company in the world is still Western Union, at about $4.1 billion of total revenue. Wise is second on net revenue for its year to March 2026, followed by Euronet's Ria segment and Remitly. By send volume the order changes: Wise moved $243.5 billion across borders in its latest year, more than three times Remitly's $74.9 billion.

Which companies lead on revenue?

The revenue leaders in 2025 were Western Union, Wise, Euronet's money transfer arm and Remitly. Western Union's revenue is shrinking slightly while the digital players grow quickly, so the gap is closing. Remitly grew revenue 29% in 2025; Western Union's consumer money transfer revenue fell 8%.

Western Union

Western Union reported 2025 revenue of $4.05 billion, down 4%, with consumer money transfer revenue of $3.51 billion across 285.9 million transactions, according to its full-year results. Its branded digital business made up 30% of consumer transfer revenue and 39% of transactions in the fourth quarter. It operates in more than 200 countries and territories. On 4 May 2026 it launched its own dollar stablecoin, USDPT, issued by Anchorage Digital Bank on Solana and used first for settlement with agents in the Philippines and Bolivia, The Block reported. A consumer feature, Stable by Western Union, is due in more than 40 countries during 2026.

Wise

Wise is the largest by volume. In its year to 31 March 2026 it moved $243.5 billion across borders, up 31%, for 18.9 million active customers, on net revenue of $2.50 billion and pre-tax income of $660.4 million, per its FY26 results. Its cross-border take rate fell to 0.52%. Wise is now listed on both Nasdaq and the London Stock Exchange, and sells its infrastructure to banks through Wise Platform.

Ria Money Transfer (Euronet)

Euronet's money transfer segment, built around Ria, earned $1.78 billion of revenue in 2025, up 6%, on 183.4 million transactions, according to Euronet's results. Its network reached about 639,000 locations at the end of 2025. Direct-to-consumer digital transactions in internationally originated transfers grew 31%, a sign that even agent-heavy players are moving customers to apps.

Remitly

Remitly is the fastest-growing of the large listed players. Revenue rose 29% to $1.64 billion in 2025, send volume rose 37% to $74.9 billion, and it reported net income of $67.9 million after a loss the year before, per its fourth-quarter release. It had 9.3 million active customers in Q4 2025 and guided to revenue of $1.94 billion to $1.96 billion for 2026.

Who are the big private and retail-network players?

MoneyGram, Intermex and Zepz are the largest remitters outside the Western Union, Wise and Remitly trio. MoneyGram is privately held and discloses little. Intermex is being bought by Western Union, a deal held up in California as of October 2026. Zepz has cut costs hard and is growing again.

MoneyGram

MoneyGram says it serves more than 50 million customers across 200-plus countries and territories through nearly 500,000 retail locations, according to a 2026 announcement with the Stellar Development Foundation. It has launched a dollar stablecoin balance in its app, built with Circle's USDC on Stellar, first in Colombia and then in El Salvador. Customers can cash out at its agents, which is the part crypto-native apps cannot easily copy.

Intermex

Intermex is a US-to-Latin America specialist that reported 2024 revenue of $658.6 million and $24.4 billion of principal sent. Western Union agreed in August 2025 to buy it for $16 a share in cash, about $500 million. New York approved the deal, but on 13 August 2026 California's regulator suspended its approval extension pending further review, according to the companies' joint update. As of mid-October 2026 we found no announcement that the deal had closed. Intermex's first-half 2026 revenue fell 17% to $253 million.

Zepz (WorldRemit and Sendwave)

Zepz has more than 9 million customers, about 8 million transactions a month, 50 sending markets and 130 receiving countries, according to an FXC Intelligence interview in January 2026. Revenue fell to $420.1 million in 2024 on $14.7 billion of volume, but the company says it ended 2025 growing about 25%, after cutting annual operating costs nearly in half since 2022. Sendwave launched a USDC-based dollar wallet in October 2025.

Which remittance companies are biggest in the UAE and the Gulf?

In the UAE, the largest remittance companies are exchange houses rather than apps. Lulu Exchange remitted $19.7 billion in 2024 and claims more than 14% of Gulf remittance volume. Al Ansari, listed in Dubai, earned AED1.29 billion of operating income in 2025 from 444 branches. Both now push digital channels, but branches still carry most of the money.

Lulu Exchange

Lulu Exchange, part of Lulu Financial Holdings, handled about $19.7 billion of remittances in 2024 and projected about $22 billion for 2025, roughly 2% of global flows, a senior Lulu executive told Semafor in December 2025. These are company-stated figures; Lulu Financial Holdings is private and does not publish audited volumes. It says it has more than 14% of Gulf remittance volume and is aiming for a double-digit global share by 2030, an ambitious target for an operator of its size.

Al Ansari Exchange

Al Ansari Financial Services reported 2025 operating income of AED1.29 billion, up 12%, and net profit of AED401 million, with 444 branches after buying Bahrain's BFC Group, per Zawya. In the first half of 2025 it processed 28 million transactions, and digital channels carried 23% of outward remittances.

The Gulf matters because of where the money goes. India received $135.46 billion of remittances in 2024-25, the most of any country. RBI's survey for 2023-24 found the US supplied 27.7% of that and the UAE 19.2%, with the UK at 10.8%, as Business Standard reported. The UAE-to-India corridor is covered in detail in our UAE to India money transfer guide, and the wider corridor in Gulf to South Asia remittances.

Where do Revolut, PayPal Xoom and the newer apps fit?

Revolut and PayPal are large in money movement but are not remittance companies first. Revolut reported $6.0 billion of revenue in 2025 across all its banking products. PayPal's Xoom serves about 160 countries but PayPal does not break out its figures. Newer diaspora apps such as Félix and Taptap Send are smaller but growing on narrow corridors.

Revolut

Revolut reported 2025 revenue of $6.0 billion, pre-tax profit of $2.3 billion and 68.3 million retail customers, with total transaction volume of $1.7 trillion, per its annual results. It does not report a cross-border transfer figure, so it cannot be ranked on remittance volume. It is a bank in more than 30 of its 40 markets.

PayPal Xoom

Xoom sends money to about 160 countries and in April 2024 let US users fund transfers with PayPal's PYUSD stablecoin, according to IBS Intelligence. Its scale is hidden inside PayPal's consolidated results.

Félix and other diaspora apps

Félix, which serves US-to-Latin America senders in 11 markets including Mexico and Colombia, raised $200 million in September 2026 in equity and debt, led by Andreessen Horowitz, and says it has more than 6 million users and $8 billion processed since launch, per FinTech Futures. Taptap Send, which focuses on underserved corridors into Africa and Asia, raised a $65 million Series B led by Spark Capital in December 2021, TechCrunch reported; it has not disclosed a newer round that we could find. Both are ones to watch because they show how a single-corridor focus can still produce scale.

Worked example: what each company earns on a transfer

Dividing revenue by transactions or volume gives a rough sense of what each business earns. These are averages across all transfer sizes and corridors, not prices any customer will be quoted.

• Western Union: $3,507.4 million of consumer transfer revenue over 285.9 million transactions is about $12.27 per transfer.
• Ria: $1,782.4 million over 183.4 million transactions is about $9.72 per transfer.
• Remitly: $1,635 million on $74.9 billion of send volume is about 2.2% of the amount sent.
• Wise: its reported cross-border take rate is 0.52%.

Now apply those to a $200 transfer. The World Bank's 6.36% average implies about $12.72 of cost. Western Union's average revenue per transfer, $12.27, sits close to that, although its average transfer is larger than $200. Remitly's 2.2% would be about $4.40. Wise's 0.52% would be about $1.04, but Wise's average transfer is much larger than $200, so a small transfer can cost more than the percentage suggests. The point is not which is cheapest on a given day. It is that the business models differ by an order of magnitude, and that difference is why digital players keep taking share.

The biggest remittance company by revenue and the biggest by volume are not the same firm, and the gap between them explains the market.

For how fees and exchange-rate margins combine in practice, see the cost of sending money home in 2026 and mid-market rates and zero-fee transfers.

What to watch next

The order of this list could change within a year. Four things will decide it.

• The Intermex deal: Western Union and Intermex said in August 2026 they would close promptly once California reinstates approval. Until then, Western Union's 2026 guidance of 5% to 8% revenue growth leans heavily on the acquisition.
• Stablecoin cash-out: MoneyGram's stablecoin wallet is expanding across Central and South America during 2026, and Western Union, which launched USDPT in May 2026, plans to roll out its Stable consumer feature in more than 40 countries this year. Both bet that agent networks become the cash ramp for digital dollars. Our stablecoin payments trends piece tracks the wider shift.
• Remitly's 2026 target: revenue of $1.94 billion to $1.96 billion would put it within reach of Euronet's money transfer segment on revenue by its full-year 2026 results, due early in 2027.
• Gulf digital share: Al Ansari's digital channels carried 23% of outward remittances in early 2025. Whether that passes a third by 2027 will show how fast Gulf exchange houses can defend their base against apps.

This list is editorial research, not investment advice or an endorsement. Company figures are as reported by the linked sources and periods differ between companies.

It is nine on a Saturday night in Seattle and already Sunday morning in Coimbatore, which is when a software architect logs into his father's electricity account and finds the bill went overdue on Thursday. He sends money to his father's savings account, phones to explain which counter to visit, and hopes the payment reaches the utility before the disconnection notice does. Millions of non-resident Indians run some version of this ritual every month. In 2026 there are better ways to do it, and the right one depends on where you live, which Indian accounts you hold and how much control you want to keep.

What NRIs actually pay for back home

India received a record $135.46 billion in private transfers in 2024-25, according to RBI data reported by All India Radio. A large share of that is not savings or investment. It is recurring household spending: electricity and piped gas for parents, a rented flat for a sibling, school or college fees for a niece, health and life insurance premiums, municipal property tax on the family home, and the postpaid mobile or broadband plan that someone in India uses but you fund.

These bills share three traits. They are small relative to the cost of an international transfer, they recur on fixed dates, and missing them has consequences, from late fees to a lapsed insurance policy. That profile is exactly why sending cash and asking someone else to pay is so inefficient, a point we make at length in why bills are the real remittance.

Option one: pay the biller directly through Bharat Connect

Bharat Connect, the rebranded Bharat Bill Payment System run by NPCI Bharat BillPay (NBBL), lists more than 22,400 billers and merchants. Its category list runs to about 30 entries, including electricity, water, piped gas, LPG, municipal taxes, housing society maintenance, rent, insurance, education fees, loan repayments and mobile postpaid. For an NRI, that covers almost the entire household budget.

The cross-border piece arrived in stages. In August 2022 the RBI said the system could accept inbound bill payments from abroad, with the elderly parents of overseas Indians explicitly in mind. Early corridors ran through Gulf exchange houses in the UAE, Oman, Kuwait and Bahrain. In August 2024 an NBBL circular widened the scope to personal non-trade bills of a current-account nature, subject to FEMA and KYC rules. In February 2025 a Canadian fintech, working with Bharat Connect and an Indian sponsor bank, began letting users pay more than 21,000 Indian billers in Canadian dollars.

The practical upside is that you enter the consumer number once, the network fetches the amount due, and the biller is credited directly. Nobody has to queue or remember a password. The catch is coverage: it only works if a provider in your country of residence has connected to the network, so check before assuming. Our Bharat Connect explainer covers the plumbing.

Option two: UPI with an international mobile number

Since January 2023, NPCI has allowed NRE and NRO account holders to use UPI with mobile numbers from ten countries: the United States, Canada, the United Kingdom, the UAE, Singapore, Australia, Hong Kong, Oman, Qatar and Saudi Arabia. Once your Indian bank enables it, you can pay most Indian billers from a UPI app as a resident would, including through the bill payment section many apps carry.

This route has no currency conversion at the moment of payment, because the money is already rupees sitting in your NRE or NRO account. The conversion cost was paid earlier, when you funded the account. For many NRIs that is the cleanest setup: move money in larger chunks when the rate and fees suit you, then pay bills in rupees on the due date.

Option three: let family pay, within limits

Some households prefer a parent to keep paying locally. NPCI's UPI Circle, launched in August 2024, lets a primary user delegate payments to up to five trusted people. Under full delegation, a secondary user can spend up to Rs 15,000 a month and Rs 5,000 per transaction; under partial delegation, they raise a request that the primary user approves with a PIN. Confirm that your bank and app support delegation from an NRI account, but the model suits the common case of a parent paying for the gas cylinder and groceries while you keep the account.

The old fallback, a monthly transfer into a parent's savings account, still works. It just brings back the delay and leakage that direct payment removes.

NRE or NRO: which account should fund your bills

Under FEMA the two accounts behave very differently. Money in an NRE account must come from income earned abroad; principal and interest are fully repatriable, and the interest is tax-free in India. An NRO account holds Indian income such as rent, dividends or pension. Its interest attracts 30% TDS plus surcharge and cess unless a tax treaty lowers the rate, and repatriation is capped at $1 million per financial year, as DBS Bank India summarises.

For bills, the logic is straightforward. If rent from an Indian flat lands in an NRO account, paying that flat's property tax, society maintenance and insurance from the same account keeps Indian money in India and avoids a round trip through foreign currency. If you are funding bills from a foreign salary, the NRE account is the natural source.

What it really costs

Fees on the bill itself are usually small. The real cost sits in turning dollars, pounds or dirhams into rupees. The World Bank's Remittance Prices Worldwide survey for the third quarter of 2025 put the average cost of sending about $200 to India at 1.91% from the UK, 2.92% from the US, 3.72% from the UAE and 4.28% from Canada. The best options in each corridor cost well under 1%, while some bank branch transfers cost more than 20%. Always compare what you receive with the mid-market rate rather than trusting a zero-fee label.

Risks worth taking seriously

• Fake disconnection messages. Texts warning that power will be cut tonight unless you pay through a link are among India's most common frauds. Pay only through your bank, a UPI app or a Bharat Connect flow. Our guide to money transfer scams lists the usual patterns.
• Wrong consumer numbers. Direct payment is only as good as the reference you enter. Save each biller once and check the name the system fetches.
• Account status. Once you become non-resident, resident savings accounts should be redesignated as NRO. Paying bills from an account that should have been converted creates paperwork later.
• Tax records. Rental income you receive in India is taxable there and may be taxable where you live. Keep receipts for property tax and maintenance you pay against it.

The cheapest bill payment is the one that never passes through a third pair of hands.

A sensible setup for 2026

For most NRIs a layered approach works. Fund an NRE account in larger, less frequent transfers when costs are low, link it to UPI with your overseas number, and pay utilities, insurance and fees in rupees on the due date. Where a provider in your country offers Bharat Connect cross-border payments, use it for bills you want settled straight from a foreign account. Give a parent delegated UPI access for small daily spending. If you live in the Gulf, our UAE to India guide covers the local options. We think the biggest gain is not the saving on fees. It is the end of the midnight phone call about a disconnection notice.

This article is general information, not tax, legal or financial advice. Check current rules with your bank and a qualified adviser.

Strip away the token charts and the stablecoin business looks like a very old one: take in dollars, pay the holder nothing, buy Treasury bills, keep the difference. That single fact explains most of the investment case for stablecoins in 2026, and most of what can go wrong with it.

How big the market is, as of October 2026

Total stablecoin supply stood at about $306 billion in mid-October 2026, according to DefiLlama's stablecoin dashboard. Tether's USDT accounts for roughly $184 billion of that and Circle's USDC for about $73 billion. Together the two hold around 84% of all supply. The long tail is real but small: Sky's USDS near $7 billion, Ethena's USDe and Dai under $5 billion each, PayPal USD close to $2.9 billion and Ripple USD around $2.4 billion.

Forecasts are much larger. Citi lifted its 2030 base case to $1.9 trillion in late 2025, with a bull case of $4 trillion, and a16z has sketched a path past $3 trillion. Treat those as scenarios rather than destinations. The more useful number is the gap between how much stablecoins move and how much of that is commerce. a16z counted $46 trillion of transactions over the year to its 2025 report, or $9 trillion after stripping out bots and churn. McKinsey, looking only at genuine payments, put the figure near $390 billion a year based on December 2025 activity, about 0.02% of global payment flows.

That gap is the opportunity in one line. Supply is large; payment use is still a rounding error against card networks and bank wires. Investors are betting on the gap closing, and the question is who gets paid when it does.

Revenue pool one: the float

Issuers earn interest on reserves. Circle's second-quarter 2026 results show the shape of it. Total revenue and reserve income was $701 million, up 7% on the year, while USDC in circulation grew 19% to $73.3 billion. Revenue lagged supply because the reserve return rate fell to 3.5%, down 66 basis points. Distribution and transaction costs came to about $410 million, which means well over half of the top line flowed to the exchanges and platforms that hold and promote USDC. Adjusted EBITDA was $143 million and net income $48 million.

Tether runs the same model with far less sharing. It reported roughly $1.5 billion of net operating profit for the quarter, per Decrypt, with an excess reserve buffer of about $4.1 billion. Our explainer on how stablecoin issuers make money goes through the mechanics. The lesson for investors is that the float is enormous, but it is mostly a rates trade, and whoever controls distribution takes a large cut.

Revenue pool two: the pipes that move the money

The second pool is fee income from moving stablecoins between people and businesses. It includes payment orchestration, on- and off-ramps into local currency, card programs, and business payouts. McKinsey estimated business-to-business stablecoin payments at around $226 billion a year, about 60% of real payment volume, after growth of more than 700% in 2025. Payroll and remittances added roughly $90 billion, and card-linked spending about $4.5 billion.

Incumbents are building here too. Circle's own payments network was running at $14.7 billion of annualized volume at the end of June, with 175 financial institutions enrolled, and its non-interest revenue rose 41% to $34 million. Visa says its stablecoin settlement has passed a $20 billion annual run rate. These businesses earn a fee per transaction, so their economics do not collapse when the Fed cuts rates. That is why they attract a different kind of investor than issuers do.

What the M&A tape says

Prices paid for private companies are the clearest signal of what strategic buyers value. Stripe paid $1.1 billion for Bridge in February 2025. In March 2026 Mastercard agreed to buy London-based BVNK for up to $1.8 billion, including $300 million of contingent payments, according to Fortune. BVNK had been valued at $750 million in its December 2024 Series B. Coinbase had earlier walked away from talks near $2 billion, and Mastercard's reported pursuit of Zerohash at $1.5 billion to $2 billion fell through.

What are buyers actually paying for? Mostly licences, banking partners, compliance teams and live corridors, the things that take years to assemble. Technology alone rarely commands these prices. A card network executive described the goal as reaching new addressable markets, and the target in Mastercard's case was cross-border business flows that cards never captured well.

Public market signals

Circle's June 2025 listing gave public investors their first pure stablecoin stock, and our piece on Circle's IPO and stablecoins in public markets covers that debut. Since then the share price has behaved like a sensitivity gauge for two variables: USDC growth and short-term interest rates. The second quarter showed both forces at once. On-chain USDC volume reached $14.8 trillion, up 151%, yet revenue grew only 7%. Circle is also launching its own blockchain, Arc, whose public mainnet was scheduled for 16 September 2026 with validators including BlackRock, DTCC and Visa. Investors are watching whether network businesses like this can diversify income away from the float.

The float pays the bills today, but the businesses that earn a fee on every payment are the ones built for a world of lower rates.

A framework for what investors look at

Whether the target is a listed issuer, a private payouts company or an infrastructure provider, the same handful of questions keeps coming up in diligence:

• Rate sensitivity. How much revenue disappears with each cut? On $73 billion of reserves, a 25 basis point drop removes roughly $180 million of gross annual income before any sharing.
• Distribution economics. Who owns the customer, and what share of income must be paid to keep them?
• Licences and jurisdictions. Is the company authorised under the GENIUS Act, MiCA, or Gulf and Asian regimes? See our guide to GENIUS Act and MiCA rules.
• Real payment volume. Raw on-chain volume flatters everyone. Investors want merchant, payroll and supplier flows with repeat customers.
• Take rate and corridor mix. What does the company earn per $1,000 moved, and how concentrated is volume in one or two corridors?

The risks that matter

Rates. Issuer income is a direct function of Treasury yields. Circle's quarter already showed revenue growth trailing supply growth as yields eased.

Regulation. The GENIUS Act in the United States and Canada's new Stablecoin Act both bar issuers from paying yield to holders, which protects banks but caps one route to growth. Rules on reserves, redemption and licensing are still being written in several countries, and a change in any of them can reshape margins.

Concentration. Two issuers hold about 84% of supply, and USDT alone holds about 60%. Any company whose volume rides on one token or one banking partner carries that dependency. A loss of confidence in a major issuer would hit the whole value chain.

Volume quality. Headline transaction numbers mix trading, internal transfers and automated activity. The $9 trillion versus $390 billion contrast is a reminder that diligence starts by asking which figure a company is quoting.

For a wider view of supply trends, read our stablecoin market report for 2026.

This article presents data and frameworks for general information. It is not investment advice and does not recommend any security, token or company.

On 21 August 2026, Indian banks had collected $65.4 billion in fresh foreign currency deposits from non-residents under a Reserve Bank of India swap scheme. Ten days later the figure was $127.2 billion. Nearly half of the biggest NRI deposit drive in the country's history arrived in a final sprint, and almost all of it went into one kind of account: the FCNR(B). For most of the 2026 rupee slide, the question for families abroad was never whether to keep money in India. It was which of three accounts to keep it in, and the answer is more mechanical than the bank brochures suggest.

The three accounts are the Non-Resident External (NRE) account, the Non-Resident Ordinary (NRO) account and the Foreign Currency Non-Resident (Bank) deposit, usually written FCNR(B). All three are governed by the Foreign Exchange Management (Deposit) Regulations, 2016 and the RBI's Master Direction on deposits and accounts. They differ on three things that matter: where the money came from, what currency it sits in, and how freely it can leave India again. Get the first one right and the other two mostly follow.

Three accounts, three different jobs

The cleanest way to think about it is by source of funds. Money you earned abroad belongs in an NRE account or an FCNR(B) deposit. Money you earn or hold in India, such as rent from a flat in Coimbatore, a pension, dividends on shares bought before you left, or the proceeds of selling inherited land, belongs in an NRO account. The RBI describes NRO as an account for bona fide rupee transactions, which is a polite way of saying it is the bucket for everything Indian.

NRE and FCNR(B) share the same rules on repatriation, joint holding and loans. The difference is currency. An NRE account converts your dollars or dirhams into rupees the moment they land, so you carry the exchange rate risk. An FCNR(B) deposit keeps the money in the foreign currency itself. SBI, for example, offers FCNR(B) in US dollars, pounds, euros, Canadian and Australian dollars and yen, with terms from one to five years. FCNR(B) is a term deposit only. There is no FCNR savings account, which is why most NRIs run an NRE savings account alongside it as the day-to-day wallet.

What can go in and what can come out

The table below condenses the rules as the RBI sets them out. It is the part most people get wrong, usually by depositing Indian income into an NRE account because the interest is tax-free.

Two credits trip people up. First, current income such as rent can only go into an NRE account if the bank is satisfied that tax has been deducted or paid on it, so the default home for rent is still NRO. Second, NRO money can move to NRE, but that transfer counts against the USD 1 million annual cap and needs the tax paperwork described below. You cannot top up an NRE account from Indian income just by moving it between your own accounts.

The USD 1 million NRO rule, read properly

The NRO cap is widely misquoted. The rule allows repatriation of current income, meaning rent, interest, dividends and pension, without counting against any ceiling once tax is settled. On top of that, an NRI or person of Indian origin can send out up to USD 1 million per financial year from NRO balances, which covers capital such as sale proceeds or inheritance. Transfers into an NRE account count inside that million. Going above it needs RBI approval.

The paperwork changed this year. Since 1 April 2026, remittances fall under the Income-tax Act, 2025, and the old Forms 15CA and 15CB have been replaced by Forms 145 and 146 under the Income-tax Rules, 2026. Form 145 is the sender's declaration of what the money is and what tax applies. Form 146 is a chartered accountant's certificate, required for taxable remittances above Rs 5 lakh in a year, and it now carries a UDIN so the bank can check that it is genuine. If you claim a treaty rate, you quote your tax residency certificate number. Banks will not release a large NRO remittance without these forms, and they will not release it at all while tax on the underlying income is unpaid.

The NRO cap is a ceiling on capital, not on income. Most families never get near it.

How the interest is taxed

NRE interest is exempt from Indian income tax. The new law moved the exemption into a schedule but kept its logic, and the Income Tax Department's FAQs, as reported in July 2026, confirm it still turns on your status under FEMA rather than on the income-tax day count. That distinction matters for anyone who spends long stretches in India: you can be a non-resident for FEMA and still trip the tax residency tests, or the other way round. Check both.

FCNR(B) interest is also exempt while you are a non-resident, and it stays exempt during the transitional "resident but not ordinarily resident" (RNOR) period after you return, which is why returning NRIs often leave FCNR deposits to run to maturity.

NRO interest is fully taxable and the bank deducts tax at source. The base rate is 30 percent plus surcharge and a 4 percent cess, which works out to 31.2 percent for interest up to Rs 50 lakh and rises with the amount. A double taxation agreement can lower that. Under the India-UAE treaty the withholding rate on interest is 12.5 percent, provided you give the bank a tax residency certificate, Form 10F and a PAN. Note too that your country of residence may tax the interest as well, with credit for Indian tax. For how gifts and transfers to family are treated, see our guide to tax on money sent to India, and check your own position with a qualified adviser.

Joint holders, mandates and the resident relative

Joint holding is where the three accounts diverge in ways that matter for ageing parents. An NRE or FCNR(B) account can be held jointly with other NRIs, or with a resident relative, but only on a "former or survivor" basis. That means the resident can operate the account only after the NRI's death, or under a mandate, and the money is still treated as the NRI's. An NRO account can be held with any resident, again former or survivor.

Power of attorney holders on an NRO account are tightly fenced. They can make local rupee payments and remit current income, but they cannot repatriate capital to anyone except the account holder, cannot make gifts to residents from it and cannot move money to another NRO account. If the plan is for a sibling in India to manage a parent's bills, an NRO account with a mandate usually fits better than trying to stretch an NRE account. Our piece on caring for parents from abroad covers the practical setup.

The 2026 FCNR(B) rush and what it tells you

The summer's numbers show how sensitive NRI money is to price. With the rupee at a record 96.25 to the dollar in May 2026, as oil climbed above $110 a barrel, the RBI announced a concessional swap facility on 5 June and opened it on 8 June. Banks that raised fresh three to five year FCNR(B) deposits could swap the dollars with the central bank, which took on the hedging cost, and deposits raised this way were exempted from reserve requirements. On 17 June the RBI also lifted, until 30 September, the ceiling on fresh three to five year FCNR(B) rates and the rule tying long NRE rates to domestic ones.

The RBI's own tallies show $17.4 billion by 17 July and $36.7 billion by 31 July. The window, due to run to the end of September, was closed early on 31 August at $127.2 billion, with most money in five-year deposits.

The flip side showed up in NRE accounts. In April to July 2026, net NRE inflows were just $207 million against $2.42 billion a year earlier, while FCNR(B) took $34.53 billion, according to RBI data reported by IANS. NRE outstanding fell from $102.03 billion to $98.02 billion over the year. When the rupee is falling, people prefer the account that does not convert. The lesson for anyone deciding now is that the window has closed, the special rates applied only to deposits opened in it, and new FCNR(B) deposits are back under normal pricing.

A worked example: one family, three accounts

Take an engineer working in Abu Dhabi whose father in Thrissur has passed away, leaving a flat. The numbers below use real tax rates and limits; the interest rates are illustrative assumptions, not quotes.

• Salary savings. She sends AED 4,000 a month to India. Kept in an NRE fixed deposit of Rs 10 lakh at an assumed 6.5 percent, it earns Rs 65,000 a year with no Indian tax. The UAE does not levy personal income tax, so that is the full return, but the rupee value of her savings moves with the exchange rate.
• Dollar reserve. A separate $20,000 sits in a five-year FCNR(B) deposit. If the rupee weakens, the deposit's value in her home currency does not change, and the interest is tax-free in India.
• Inherited flat. The flat sells for Rs 1.2 crore. The proceeds must go to an NRO account. At an illustrative Rs 95 to the dollar, USD 1 million is about Rs 9.5 crore, so the whole sale fits inside one year's cap. Before the bank sends it, capital gains tax must be settled, she files Form 145, and a chartered accountant issues Form 146.
• Parked proceeds. If Rs 10 lakh of the sale money stays in an NRO fixed deposit at an assumed 7 percent, interest is Rs 70,000. Default TDS at 31.2 percent is Rs 21,840. With a UAE tax residency certificate and Form 10F, the treaty rate of 12.5 percent cuts that to Rs 8,750, a saving of Rs 13,090 a year.

The order of operations matters more than the rates. Paperwork for the treaty rate should be with the bank before the first interest credit, not after, because recovering over-deducted tax means filing a return and waiting for a refund.

Mistakes that cost real money

• Keeping a resident savings account after leaving. FEMA requires resident accounts to be redesignated as NRO once you become a non-resident.
• Putting Indian rent into NRE. Unless tax has been dealt with, rent belongs in NRO. Mixing it into NRE muddies the tax-free status of the whole balance.
• Assuming NRE is tax-free everywhere. It is exempt in India. The US, UK and Canada generally tax worldwide interest of their residents.
• Ignoring the FEMA and tax mismatch. A long visit home can change your tax residency even while your accounts are correctly non-resident. Read our FEMA rules for NRIs explainer alongside the tax rules.

For moving money into these accounts in the first place, the cost of the transfer often outweighs a quarter-point of interest. Our corridor guides on UAE to India and the cost of sending money home show where the spread goes.

What to watch next

• 16 October 2026: last date for banks to use the RBI swap on FCNR(B) deposits raised in the window. After that, the facility's deposit leg is done; the five-year deposits it created mature in 2031, which is a refinancing question for banks to watch.
• 31 December 2026: the parallel swap window for external commercial borrowings and overseas foreign currency borrowings closes.
• RBI deposit regulations: whether the temporary removal of NRE and FCNR(B) rate ceilings, which lapsed on 30 September, returns if the rupee comes under pressure again.
• Tax year 2026-27 returns: the first returns under the Income-tax Act, 2025 will be filed in 2027. Watch for final section mappings and any change in how the NRE exemption is drafted.

This article explains rules as published by the RBI and the Income Tax Department. It is general information, not tax or legal advice. Rules change and individual circumstances differ, so check your position with a qualified adviser before moving money. If you are planning to return, our moving back to India money checklist covers what happens to each account.

A treasury analyst at a freight forwarder in Rotterdam opens her dashboard on a Monday in October 2026 and sees three dollar balances that are not held at any bank: USDT from a supplier in Shenzhen, USDC from a client in Austin, and a small float of PYUSD that a marketplace paid out last week. Each token is a promise from a different company, held to a different standard, redeemable through a different chain of partners. Picking which stablecoin companies matter is no longer an exercise for crypto traders. It is supplier due diligence.

How we chose. We looked at companies with measurable scale as of October 2026: circulating supply for issuers, and volume, licences, funding or acquisition price for infrastructure firms. We also weighed regulatory standing and product breadth. We group companies by role rather than numbering them, and within the issuer group we order by reported supply. Every figure below is dated and linked to the source we checked.

Who are the largest stablecoin companies in 2026?

As of October 2026, the largest stablecoin company by supply is Tether, issuer of USDT at about $184 billion, followed by Circle, issuer of USDC at about $73 billion, according to DefiLlama data. Together they account for roughly 84% of a market worth about $306 billion. Behind them sits a long tail of issuers, each under $10 billion, and a fast-consolidating layer of infrastructure companies.

The concentration is the headline, but the more interesting story of 2026 is the money flowing into the plumbing. Card networks, payment processors and prime brokers have spent several billion dollars buying stablecoin infrastructure in the past twenty months, and those deals tell you where incumbents think the profits will be. For the history of how the market reached this size, see our stablecoin market report.

Which companies issue the biggest stablecoins?

In 2026 the biggest dollar stablecoins are issued by Tether, Circle, Sky, Ethena, World Liberty Financial, Paxos (including for PayPal) and Ripple. Only Tether and Circle exceed $10 billion. As of mid-October 2026 the others range from about $2 billion to $7 billion on DefiLlama, and their supply has been volatile, with some losing a third or more within months as exchange incentives changed.

Tether (USDT)

Tether is the largest issuer by a wide margin. Its first-quarter 2026 reserve report showed net profit of $1.04 billion, total assets of $191.7 billion against liabilities of $183.5 billion, and direct and indirect US Treasury exposure of $141 billion, which Crowdfund Insider reported would make it the 17th-largest holder of Treasuries in the world. Reserves also included about $20 billion of physical gold and $7 billion of bitcoin. The attestation was prepared by BDO; Tether says a full audit has begun. In January 2026 it launched USAT, a US-regulated dollar token issued through Anchorage Digital.

Circle (USDC)

Circle, now a listed company, had $73.3 billion of USDC in circulation at the end of June 2026, up 19% on a year earlier. Its second-quarter results showed $701 million of revenue and reserve income, $48 million of net income from continuing operations and $14.8 trillion of onchain USDC volume, up 151%. Its Circle Payments Network had 175 financial institutions enrolled and an annualised run rate of $14.7 billion, and its Arc blockchain was due for public mainnet on September 16, 2026.

Sky (USDS and DAI)

Sky, the rebranded MakerDAO, runs the largest decentralised dollar. Its newer USDS token had about $6.9 billion in circulation in mid-October 2026, ahead of legacy DAI at about $4.7 billion, according to DefiLlama's stablecoin dashboard. Both are backed by onchain collateral and real-world assets rather than a single bank account, which makes Sky the main counterweight to the custodial model.

World Liberty Financial (USD1)

USD1, issued by World Liberty Financial with reserves custodied by BitGo, had about $4.15 billion in circulation in mid-October 2026, per DefiLlama, making it the sixth-largest dollar stablecoin. The project's political links draw scrutiny, and readers should note that independent reserve reporting is thinner than for the larger issuers.

Ethena (USDe)

Ethena's USDe is a synthetic dollar backed by crypto collateral and offsetting derivatives positions rather than cash. Supply stood at about $4.8 billion in mid-October 2026, per DefiLlama, slightly ahead of DAI. Its returns come from funding rates on derivatives, which rise and fall with crypto markets. It is a yield product with market risk, not a payment token in the conventional sense.

Paxos and PayPal (PYUSD, USDG)

Paxos is the white-label issuer behind PayPal's PYUSD and the Global Dollar (USDG). In December 2025 it received approval to convert its New York trust charter into a federal one supervised by the OCC, as Skadden, its counsel, announced. PYUSD reached about $4.1 billion in March 2026, when PayPal expanded it from two markets to 70, COIN360 reported, then fell to about $2.9 billion by mid-October, per DefiLlama. Paxos's own USDG was slightly larger, at about $3.0 billion.

Ripple (RLUSD)

Ripple's RLUSD, issued under a New York trust charter, grew from about $130 million in mid-2025 to about $2.4 billion in mid-October 2026, per DefiLlama. Ripple uses it as a settlement asset inside its payments product, which we explain in how Ripple Payments works.

First Digital (FDUSD)

Hong Kong's First Digital shows how fragile second-tier supply can be. FDUSD was once a leading Binance trading pair, but by mid-October 2026 it was no longer among the 23 stablecoins above about $490 million on DefiLlama. In April 2025 it briefly lost its dollar peg after Tron founder Justin Sun alleged that First Digital Trust could not meet redemptions, The Defiant reported. It remains one to watch because of its base in Hong Kong, where the new stablecoin licensing regime is taking shape.

Which companies build stablecoin infrastructure?

Stablecoin infrastructure companies issue tokens for others, connect bank accounts to blockchains, hold keys and run payout networks. In 2025 and 2026 the leaders were Bridge (owned by Stripe), BVNK (being bought by Mastercard), Zerohash, Fireblocks and Anchorage Digital. Their valuations and acquisition prices now run from about $1 billion to over $4 billion.

Bridge (Stripe)

Stripe bought Bridge for $1.1 billion in February 2025, its largest acquisition, and in February 2026 Bridge won conditional OCC approval to organise a national trust bank, according to FinTech Futures. Bridge provides APIs that let companies accept, hold and issue stablecoins, and its charter would let it hold reserves under federal supervision rather than relying on partner banks.

BVNK

London-founded BVNK runs stablecoin payment accounts and payouts for businesses in more than 130 countries. Mastercard agreed in March 2026 to acquire it for up to $1.8 billion, including $300 million of contingent payments, with completion expected by the end of 2026, COIN360 reported. BVNK had been valued at about $750 million in 2024, and Coinbase had earlier held talks near $2 billion.

Zerohash

Chicago-based Zerohash supplies crypto and stablecoin rails to brokers and fintechs including Interactive Brokers, Stripe and Shift4. It was valued at $1 billion in September 2025 after a $104 million round led by Interactive Brokers, and in May 2026 was reported to be seeking a valuation above $1.5 billion, per crypto.news, citing CoinDesk. Mastercard had considered buying it before choosing BVNK.

Fireblocks

Fireblocks provides wallet and key-management technology to banks and exchanges. In August 2026 its Network for Payments was handling more than $100 billion a month in stablecoin transactions across more than 40 live payment providers, banks and fintechs, with over $200 billion a month across its wider platform, according to Crypto Briefing.

Anchorage Digital

Anchorage operates a federally regulated digital asset bank in the US and issues Tether's USAT. In February 2026 Tether invested $100 million at a $4.2 billion valuation, as The Block reported, alongside an employee tender at the same price. It is the clearest example of an offshore issuer buying a regulated US channel.

How do the leading stablecoin companies compare?

The table below sets out each company's base, role and latest dated metric as of October 2026. Supply figures for issuers come from DefiLlama and company results; infrastructure figures come from company results and press reports. The metrics are not directly comparable across rows, because issuers are measured by tokens outstanding and infrastructure firms by price, valuation or volume.

How much does each dollar of stablecoin earn? A worked example

Stablecoin issuers earn mainly from interest on reserves, so revenue per dollar outstanding is a useful yardstick. Using 2026 results, Circle earned about 3.8 cents of gross revenue per USDC dollar on an annualised basis, while Tether's annualised profit was about 2.3 cents per dollar of liabilities. The gap between gross revenue and profit is where distribution deals bite.

Here is the arithmetic, using the companies' own reported figures:

• Circle, Q2 2026. $701 million of quarterly revenue and reserve income, multiplied by four, is about $2.8 billion a year. Divided by $73.3 billion of USDC, that is roughly 3.8%. But net income from continuing operations was only $48 million in the quarter, or about 0.26% annualised. Most of the gross income is paid away to distribution partners and spent on operations.
• Tether, Q1 2026. $1.04 billion of quarterly profit, multiplied by four, is about $4.2 billion. Divided by $183.5 billion of liabilities, that is about 2.3% as profit, not revenue, and Tether's figure includes gains on gold and bitcoin as well as interest.
• Concentration. USDT at about $184 billion plus USDC at about $73 billion is $257 billion, or 84% of the $306 billion market. The next six tokens in this article (USDS, DAI, USDe, USD1, PYUSD and RLUSD) add up to roughly $26 billion, about 8%, on mid-October DefiLlama data.

The lesson is that size alone does not decide profitability. Circle's business model rewards the exchanges and wallets that hold USDC, while Tether keeps most of its spread. Every smaller issuer faces the same choice: pay heavily for distribution, as PYUSD and FDUSD did through incentives, or stay small. Our explainer on how stablecoin issuers make money goes deeper.

The stablecoin market has two giants and a crowd of companies renting their way to relevance.

Why are incumbents buying stablecoin companies?

Between February 2025 and March 2026, Stripe, Mastercard, Ripple and others agreed deals worth more than $5 billion for stablecoin infrastructure. The buyers want licences, banking relationships and engineering teams that would take years to build, and they want them before US rules under the GENIUS Act fully take effect.

The pattern is consistent. Stripe paid $1.1 billion for Bridge's issuance and orchestration stack. Mastercard agreed up to $1.8 billion for BVNK's payout network. Ripple paid $1.25 billion for prime broker Hidden Road, $200 million for Rail and $1 billion for treasury software firm GTreasury. Tether bought into Anchorage at $4.2 billion. Each deal joins a company with distribution (merchants, cardholders, institutions) to a company with stablecoin rails.

For businesses choosing providers, consolidation cuts both ways. A Mastercard-owned BVNK or Stripe-owned Bridge brings a balance sheet and compliance resources a start-up cannot match. It also brings product roadmaps set by a parent whose interests may not align with a small corridor. Our pieces on stablecoin payments for business and stablecoin start-up funding track the trade-offs.

What should you watch next?

Over the next six months, the events most likely to reshape this list are the closing of Mastercard's BVNK deal, US federal trust charters moving from conditional to final, Tether's first full audit, and the implementation of GENIUS Act rules for issuers. Each could shift market share or open new distribution.

• BVNK closing, by end of 2026. Mastercard expects completion by year end, subject to regulatory approvals. Watch whether BVNK's corridors are folded into Mastercard's cross-border products.
• Federal trust charters. Paxos, Bridge and Ripple all hold OCC approvals of varying finality. Final charters would let them hold reserves under one federal supervisor.
• Tether's audit. Tether says an audit has begun. A full audit from a major firm would remove the largest standing objection to USDT among US institutions.
• GENIUS Act implementation. Rules from US regulators will determine how offshore issuers such as Tether can serve American users, and why USAT exists. Our GENIUS Act tracker follows each rulemaking.
• Zerohash's raise. A round above $1.5 billion would confirm that independent infrastructure can still command premium valuations after the card networks' buying spree.

This list is editorial research, not investment advice or an endorsement of any company or token. Company figures are as reported by the linked sources and may have changed since publication.

On the Ring Road textile markets in Surat, a saree wholesaler can close a day's trading without touching cash. Retail buyers scan a QR code taped to the counter, a speaker box reads out each credit in Gujarati, bulk orders go through a card terminal on the desk, and by evening the wholesaler is checking a mutual fund SIP on a phone app. Each of those steps belongs to a different Indian fintech company, and in 2026 several of them are public, profitable and arguing over who gets to keep the customer.

How we chose these companies

We picked the companies below using measurable criteria as of October 2026: scale (UPI share, users, customer assets), reported revenue and profit, listing status, funding and valuation, and breadth of regulated products. They are grouped by business type rather than ranked, because a payments app, a stockbroker and an insurance marketplace cannot sensibly be ranked against one another on a single number. Every figure is dated and linked to the source that reported it.

The backdrop is the payments rail itself. UPI processed 24.07 billion transactions in September 2026, or about 802 million a day on average, the first month above 800 million, according to NPCI figures reported by StartupTalky. Because the rail is largely free to use, most Indian fintechs earn elsewhere: lending, broking, insurance, devices or software. Our UPI explainer covers how it works.

Who leads UPI payments in India in 2026?

PhonePe is the largest UPI app in India in 2026, with 45.64% of transaction volume in August 2026, followed by Google Pay at 32.26% and Paytm at 8.03%, per NPCI data cited by StartupTalky. Together the top three handled about 86% of volume. Their grip is loosening slowly: PhonePe and Google Pay combined fell below 80% for the first time in May 2026.

The concentration story matters because NPCI has a standing rule that no single third-party app should exceed 30% of UPI volume, and the deadline for compliance was pushed to 31 December 2026, as Business Today reported when the extension was announced at the end of 2024. That date is now less than three months away, and as of mid-October 2026 we found no NPCI announcement of a further extension or of how the cap will be enforced.

PhonePe

The Walmart-backed app is India's biggest payments brand by volume and has been preparing a listing for over a year. It filed confidentially in September 2025 and filed updated papers in March 2026, but the plan stalled; Inc42 reported in September 2026 that it is now targeting a February to March 2027 listing at roughly $10 billion. In FY26 its operating revenue rose 11.5% to Rs 7,920.5 crore, while net loss widened 62% to Rs 2,792 crore.

Google Pay and the challengers

Google Pay is not Indian, but no honest map of UPI can leave it out. Behind the two leaders, a group of newer apps is chipping away. Outlook Business reported in June 2026 that Navi and super.money together had reached about 5.5% of volume roughly two years after launch, and that the government's own BHIM app had grown fivefold to about 1% share. These are distribution plays, buying users on the rail to sell them credit later.

Which Indian fintech companies are listed, and how are they doing?

As of October 2026, India's listed fintech group includes Paytm (One97 Communications), Groww (Billionbrains Garage Ventures), Pine Labs and PB Fintech, the parent of PolicyBazaar. All four reported a profit for the quarter ended June 2026. Groww and Pine Labs listed in November 2025, during a run of five fintech IPOs in the second half of 2025, according to Tracxn data reported in July 2026.

Paytm

Paytm has completed a turnaround that few investors priced in two years ago. In Q1 FY27 it reported operating revenue of Rs 2,448 crore, up 28%, and profit after tax of Rs 220 crore, up 79%, with a record quarterly EBITDA of Rs 203 crore, according to the company's results release. It had 8 crore monthly transacting users and 1.57 crore merchants paying for devices such as soundboxes. Financial services revenue, mostly loan distribution, grew 45% to Rs 814 crore, which tells you where the margin really lives.

Groww

Groww's parent raised about $752 million in its IPO and closed its first trading day, 12 November 2025, up about 31% with a market value of roughly $9.1 billion, Bloomberg reported. In Q1 FY27 it earned Rs 735 crore of net profit on Rs 1,501 crore of revenue, per YourStory, with 2.2 crore transacting users and Rs 3.6 lakh crore of customer assets.

Pine Labs

Pine Labs is the merchant-side counterweight to the UPI apps: card terminals, checkout software and issuing technology for banks and brands. It listed on 14 November 2025 after a Rs 3,900 crore IPO and closed its debut at a market value of about Rs 28,937 crore, per Business Standard. In Q1 FY27 operating revenue rose 20% to Rs 736.9 crore and net profit reached Rs 19.6 crore on gross transaction value of about Rs 4.22 lakh crore, Inc42 reported.

PB Fintech (PolicyBazaar)

The insurance and credit marketplace has quietly become one of the most consistent earners in Indian fintech. Q1 FY27 operating revenue rose 40% to Rs 1,888.3 crore and net profit nearly doubled to Rs 162.9 crore, on Rs 8,372 crore of insurance premium, according to Inc42.

Which private fintechs are the largest in India?

The largest private Indian fintechs in 2026 are Zerodha, the discount broker that remains the most profitable fintech in the country, payment gateway Razorpay, which plans to list within FY27, credit card rewards app CRED, valued at about $4.5 billion in June 2026, and payment orchestration company Juspay, which became a unicorn in 2025 and was valued at about $1.2 billion in January 2026.

Zerodha

Zerodha has never raised outside capital and still earns more than any listed peer. Its net profit was Rs 4,283 crore in FY26, up about 1% from Rs 4,231 crore, on revenue roughly flat with FY25's Rs 8,847 crore, per YourStory and Inc42. Zerodha shares indicative figures rather than audited filings at that point, so treat them as approximate. SEBI's tighter derivatives rules have flattened the top line, and its share of NSE active clients slipped to about 14.9% in July 2026. It is a cash machine that has stopped growing.

Razorpay

Razorpay is a payment gateway for online businesses that has added SME accounts, payroll and lending. Its FY25 operating revenue rose 65% to Rs 3,783 crore, but one-time ESOP and tax costs tied to moving its domicile back to India pushed it into a loss, according to Inc42's October 2025 report. Its CFO said a listing within FY27 is the goal, which would put it in the same IPO window as PhonePe.

CRED

CRED raised $900 million in a Series H led by Meta in June 2026 at a post-money valuation of about $4.5 billion, Inc42 reported. In FY25 it posted Rs 2,735 crore of operating revenue, up 16%, and cut its operating loss by 51% to Rs 298 crore. Its affluent card-holder base suits lending and bill payment products.

Juspay

Juspay builds payment orchestration and checkout infrastructure used by more than 500 enterprises and banks, and says it processes over 300 million transactions a day with annualised volume above $1 trillion, per Entrackr's January 2026 report of a $50 million round at about $1.2 billion. FY25 revenue was Rs 540 crore with a Rs 62 crore profit, a rare profitable infrastructure business at this size.

Which Indian fintechs are turnaround or challenger stories?

The clearest turnaround stories in Indian fintech in 2026 are BharatPe, which turned EBITDA positive in FY25 after years of heavy losses, and Jupiter, a digital-banking app that raised fresh money in late 2025. CoinDCX is the largest name in the regulated crypto segment, with Coinbase investing at a $2.45 billion valuation in October 2025.

BharatPe

BharatPe built a merchant QR, card terminal and lending business for small shops. In FY25 operating revenue rose 16.9% to Rs 1,667 crore, EBITDA turned positive at Rs 47 crore and the net loss fell 82% to Rs 88 crore, per Entrackr. Management has talked about an IPO after FY26.

Jupiter

Jupiter raised $15 million from existing investors in October 2025, Entrackr reported, after last being valued at about $710 million in 2021. It says it has over 3 million customers, about 60% of them active, and more than a million active users of its Account Aggregator feature.

CoinDCX

CoinDCX is one of India's largest crypto exchanges, with more than 20 million users as of mid-2025. Coinbase invested in it in October 2025 at a post-money valuation of about $2.45 billion, Outlook Business reported, months after CoinDCX lost roughly $44 million in a July 2025 breach that it said did not touch customer funds. The sector's direction depends heavily on policy; our India crypto regulation guide covers the current rules.

How do India's leading fintechs compare?

The table sets out each company's core business and latest dated metric. Metrics differ by company, so treat it as a map, not a league table.

How much funding did Indian fintech raise in 2025 and 2026?

Indian fintech companies raised about $2.4 billion in 2025, up 2% on 2024, keeping India third globally behind the US and UK, according to a Tracxn report covered by IANS in January 2026. In the first half of 2026 the sector raised about $2 billion, up 42% year on year, across far fewer rounds.

One data caution. In July 2025 Tracxn reported H1 2025 fintech funding of $889 million, as Business Standard covered. Its July 2026 report compares H1 2026 against a revised H1 2025 figure of $1.4 billion, which is where the 42% growth comes from. Databases restate past periods as late-disclosed rounds come in; we use the revised $1.4 billion so the comparison is like for like.

The shape of the money changed more than its size. In 2025, early-stage funding jumped 78% to $1.2 billion while late-stage money fell to $1 billion. In H1 2026 that flipped: late-stage funding rose 331% to $1.6 billion, and the number of rounds dropped to 106 from 186 a year earlier, IANS reported in July 2026. CRED's $900 million round alone accounts for a large share of that late-stage total. Bengaluru took 70% of H1 2026 fintech funding.

Indian fintech money is no longer spread across a hundred bets. It is concentrated on the handful of companies that can show a profit.

IPOs and acquisitions are the other half. Tracxn counted five fintech IPOs in H2 2025, none in H1 2025, and two in H1 2026, Turtlemint and Kissht. Our fintech startups to watch list covers the earlier-stage names globally.

Worked example: what the 30% UPI cap would mean

Take September 2026's 24.07 billion UPI transactions as a monthly baseline. A 30% ceiling would allow any single app about 7.22 billion transactions a month. PhonePe, at its August share of 45.64%, would be running at about 10.99 billion on that baseline. To comply, it would need to shed roughly 3.77 billion transactions a month, or about 125 million a day.

Google Pay, at 32.26%, would be about 7.77 billion, roughly 0.54 billion over the line, or 18 million a day. Paytm and everyone else are already well under it. Put another way, enforcing the cap on 1 January 2027 would require moving more than 4.3 billion payments a month to smaller apps, which is more than double Paytm's entire monthly volume. That arithmetic is why the deadline has already been extended once.

Illustrative calculation using NPCI September 2026 volume and August 2026 app shares; actual monthly volumes vary.

What to watch next in Indian fintech

The next six months contain several dated decisions that will shape the ranking above. Anyone weighing a listed fintech stock should consult a qualified adviser.

• 31 December 2026: the NPCI deadline for the 30% UPI market share cap. Extension, phased enforcement or a hard cap would each reshape PhonePe's IPO story.
• End of 2026: PhonePe's expected updated IPO filing, ahead of a targeted February to March 2027 listing.
• By March 2027: the end of FY27, Razorpay's stated window for a listing.
• Quarterly results: Paytm, Groww, Pine Labs and PB Fintech report Q2 FY27 in late October and November 2026; watch whether lending distribution keeps growing faster than payments.
• UPI pricing: any change to merchant discount rules for large merchants, which Inc42 notes leaves about 96% of UPI transactions free.

India's fintech winners increasingly look like financial services firms with a payments front door. The rail stays public, as our pieces on India Stack and Bharat Connect explain; the profit sits in what is sold over it.

This list is editorial research, not investment advice or an endorsement of any company. Company figures are as reported by the linked sources and may be revised.

My aunt asked me last Diwali what a stablecoin was. She had heard the word on a WhatsApp forward, sandwiched between a cricket score and a recipe. I gave her the textbook answer and watched her eyes glaze over. So this is the answer I should have given.

The short version

A stablecoin is a digital dollar. Not a dollar in your bank, and not a coin in your pocket, but a token on the internet that is designed to always be worth one US dollar. You can send it to anyone with a phone in a few seconds, on a Sunday, at 2 a.m., without asking a bank's permission.

That's it. Everything else is detail.

How does a digital dollar stay worth a dollar?

The big stablecoins are backed by real money held somewhere else. For every token in circulation, the company that issued it keeps roughly a dollar of cash or very short-term US government debt in reserve. If you hand the token back, you're meant to get your dollar back. That promise is what keeps the price pinned.

It is a bit like a cloakroom ticket. The ticket itself is just paper, but everyone trusts it because the coat is hanging behind the counter. The quality of a stablecoin comes down to one question: is the coat really there?

That's exactly what new laws now check. In the United States, the GENIUS Act requires issuers to hold reserves at least equal to every token outstanding, restricts what those reserves can be invested in, and demands monthly attestations from outside accountants. Europe's MiCA rules do something similar. We wrote a separate piece on what those rules change.

How big is this, really?

Those numbers matter less than the direction. Stablecoins started as a tool for crypto traders who needed somewhere to park money between trades. In the last two years they quietly became a payments rail. Businesses paying overseas suppliers moved the most money, but freelancers, exporters and families are close behind.

Is it crypto? Isn't crypto risky?

Stablecoins live on the same blockchains as Bitcoin, which is why they get lumped together. But they behave very differently. Bitcoin's price swings because people are betting on it. A good stablecoin's price doesn't swing because nobody is betting on it; it's meant to be boring.

The real risks are different ones:

• Issuer risk. If the company behind the coin mismanages its reserves, the dollar promise weakens. This is why regulated, audited issuers matter.
• Access risk. Getting from stablecoin back into rupees, pesos or shillings still runs through licensed exchanges and banks. That last step is where most of the friction, and most of the fees, still live.
• User error. Send a token to the wrong address and there's no customer helpline to reverse it. Good apps hide the address entirely.

So why should a normal family care?

Because the most expensive part of sending money abroad has never been the money. It's the pipes. A typical international transfer still hops through several banks, each taking a cut and a day. Globally, sending money home costs an average of about 6.36% of the amount sent, according to the World Bank's latest data for the third quarter of 2025. On a $200 transfer that's nearly thirteen dollars, every single month.

Stablecoins don't magically remove every cost. But they replace the slowest, most expensive middle part of the journey with something that settles in seconds for a fraction of a cent. When that saving is passed on, it shows up as a better exchange rate, a bill that's paid on time, or simply money that arrives when your mother needs it rather than three days later.

The best stablecoin experience is one where you never have to hear the word "stablecoin."

What careful beginners tend to do

1. Many stay with large coins from regulated issuers. USDC and USDT are the two most widely used; we compare them here.
2. They use licensed services that convert into local currency, rather than juggling exchanges themselves.
3. They treat "yield" offers on money meant for spending with suspicion. Boring is the point.
4. They check the rate actually received, not only the fee shown. A "zero fee" transfer at a bad exchange rate isn't free.

That's the guide I wish I had given my aunt. She would have liked the cloakroom ticket.

Judged by its headline number alone, 2026 was a dull year for stablecoins. Total supply opened the year at roughly $308 billion, touched a record of about $321 billion in April and May and slipped back to around $306 billion by mid-October, according to DefiLlama. Look underneath that flat line, though, and almost every measure of real use went up. This report pulls the main numbers into one place and explains which ones actually matter.

The float: from $27 billion to $300 billion in five years

The long view is still remarkable. Year-end figures from DefiLlama, compiled by Transak, show the market growing roughly elevenfold since 2020, with two years of retreat in the middle.

The 2026 pattern deserves a closer look. Supply peaked near $321 billion in May, then June brought the steepest monthly decline since the Terra crash of 2022. Analysts tracking the quarter put the drop down to cooling crypto trading and lower demand for collateral on exchanges, not to any loss of faith in stablecoins as payment instruments. That distinction matters: the float shrank, but the activity riding on it did not.

A methodology note before going further. Trackers disagree at the margin because they include or exclude different tokens. On 14 October 2026, for instance, rwa.xyz listed total stablecoin value at about $293 billion, several billion below other counts. Treat any single figure as a range of a few percent.

Who issues it: two companies, about 84 percent

Concentration is the defining feature of this market. Tether's USDT stood at about $184 billion in early October 2026, roughly 60% of total supply. Circle's USDC ended June at $73.3 billion. Together they hold around 84% of everything in circulation, and the top five issuers held about 89% in the first quarter.

Below them sits a crowded tail: Sky's USDS and legacy DAI, World Liberty Financial's USD1, Ethena's USDe, Paxos's USDG, PayPal's PYUSD and Ripple's RLUSD, each between roughly half a percent and three percent of the market as of August 2026. Some of that tail is fragile. Ethena's USDe shrank from about $14.7 billion before the October 2025 liquidation cascade to about $6.3 billion two months later. We compare the two leaders in detail in USDT vs USDC.

Where it lives: supply by chain

DefiLlama data cited on 25 September 2026 ranked the main networks this way: Ethereum about $145.9 billion, Tron about $94.1 billion, Solana a record $16.6 billion, BNB Chain about $13.3 billion and Base about $5.1 billion. Ethereum's share has fallen from above 90% in 2020 to roughly half today, though its dollar balance has grown enormously. Tron is effectively a USDT network, used heavily for retail and cross-border transfers where low fees matter more than anything else.

The chain map is starting to move again because of purpose-built payment networks such as Stripe's Tempo and Circle's Arc, both live in 2026. We compare fees and settlement times in our guide to the best blockchains for stablecoin payments.

How much it moves: three numbers, three meanings

Volume is where stablecoin statistics get abused. There are at least three honest ways to measure it, and they differ by orders of magnitude.

• Raw onchain transfers. a16z crypto counted about $46 trillion in stablecoin transactions over the year to late 2025. In the first quarter of 2026 alone, transfers topped $28 trillion, but around three quarters of that was bot activity.
• Adjusted volume. Stripping out bots and repeated internal hops, a16z put the figure nearer $9 trillion, up 87% on the year before.
• Actual payments. A joint analysis by McKinsey and Artemis found about $390 billion of genuine payments in 2025, around 0.02% of global payment flows. Business-to-business transfers made up $226 billion of it.

Our read: the third number is the one that tells you whether stablecoins are becoming money people spend, and it is growing quickly. Identified payments reached at least $401 billion in the first eight months of 2026, up 42% on the same period of 2025. Standard Chartered also noted in March 2026 that the average stablecoin now changes hands about six times a month, double the pace of two years earlier. More use from a flat float is exactly what you would expect if the asset is shifting from parking spot to payment rail. The details are in our 2026 payments trends piece.

The float went sideways in 2026. The payments built on top of it did not.

Forecasts: what the banks expect, and what that implies

The big institutional forecasts cluster around $2 trillion within a few years.

• US Treasury Borrowing Advisory Committee (April 2025): about $2 trillion by 2028, up from roughly $240 billion at the time, according to BeInCrypto's summary.
• Standard Chartered: $2 trillion by 2028, reaffirmed in March 2026, with about $1 trillion of extra demand for US Treasury bills along the way, as reported by The Block.
• Citi: a revised base case of $1.9 trillion by 2030 and a bull case of $4 trillion, published in September 2025.

Some simple arithmetic puts these in perspective. Getting from about $306 billion in October 2026 to $2 trillion by the end of 2028 would require supply to multiply roughly 6.5 times in just over two years, or about 130% a year. This year's growth so far is close to zero. The forecasts may still prove right, since a single large shift, such as banks or card networks moving settlement balances onchain, could add hundreds of billions quickly. But they are bets on a step change, not extrapolations of the current trend.

What to watch into 2027

1. GENIUS Act implementation. The US law takes effect no later than 18 January 2027. Final rules will decide how much room issuers and their partners have to share reserve income. See our explainer on GENIUS and MiCA.
2. Open USD. The OUSD token launched on 30 September 2026 backed by Visa, Mastercard, Stripe, Shopify and Coinbase. Its early volume will show whether distribution partners can dent the USDT and USDC duopoly.
3. Bank tokens. A consortium of 21 banks plans a dollar stablecoin for the first half of 2027.
4. USDC versus USDT. USDC drives most of the rise in velocity; USDT dominates low-velocity savings use in emerging markets. Which one grows faster tells you which use case is winning.

Figures are as of the dates stated and come from the linked sources. Different trackers count supply differently, so small discrepancies between sources are normal.

A Filipino seafarer on a container ship in the Strait of Malacca gets a few hours of satellite connection a week. In that window he has to make sure the family's Meralco electricity bill, his son's tuition and the monthly housing loan in Batangas all get paid. For years the answer was to send money home and trust that someone would get to each counter before the due date. A growing set of networks now lets him pay those companies directly, and that small change says a lot about where cross-border payments are heading.

What a cross-border bill payment actually is

A cross-border bill payment starts in one country and lands in a specific biller's account in another, tagged with the reference the biller needs: a consumer number, a policy number, a student ID. The sender pays in their own currency. The utility, school or insurer receives local currency and marks the bill as settled. Nobody in between has to collect, convert or carry anything.

That sounds like a small tweak to a remittance. The scale says otherwise. The World Bank estimates $685 billion flowed to low- and middle-income countries in 2024, with India at about $129 billion, Mexico at $68 billion and the Philippines at $40 billion. IFAD estimates that around 75% of remittance money goes to immediate needs such as food, health, education and housing. A large share of that ends up, sooner or later, at a utility or a school's accounts office.

Why paying the biller beats sending cash

Three things go wrong when money travels as a cash transfer first and becomes a bill payment later.

• Leakage. Money sent for school fees competes with every other need in the receiving household. Nobody has to be dishonest for it to be spent elsewhere first.
• Double friction. The transfer carries its own fee and exchange margin, and the recipient may pay again to settle the bill at an agent. The World Bank's latest global survey put the average cost of sending $200 at 6.36% in the third quarter of 2025, and cash-based services cost more than that.
• Timing. A transfer that arrives on day one and gets paid on day nine misses due dates and triggers late fees. A direct payment can be scheduled against the due date itself.

There is a quieter benefit too. A paid bill leaves a record in the sender's name, which is useful when a migrant needs to show that they support a parent or child back home.

How biller networks and aggregators work

No overseas provider wants to integrate with thousands of utilities one by one. Instead they plug into a domestic hub, and three layers do the work:

1. The domestic biller network. A national switch or aggregator that already connects billers, validates customer numbers and fetches the amount due.
2. The cross-border gateway. A licensed bank or exchange house in the destination country that accepts foreign funds, runs compliance checks, converts the currency and pushes the payment into the domestic network.
3. The sending front end. The bank, app or exchange house the customer uses abroad, which collects the money and passes on the biller details.

Mastercard's 2021 purchase of Arcus in Mexico shows why the hub matters. Arcus had direct links to many of Mexico's largest billers and retailers and approval to process through SPEI, the central bank's real-time system. Buying the hub bought the reach.

India: a national network opens to the diaspora

India has gone furthest in building this into public infrastructure. Bharat Connect connects more than 22,400 billers across about 30 categories and handled more than 260 million transactions in October 2025, according to its own overview. The RBI cleared it for inbound cross-border bill payments in 2022, and a 2024 circular widened the scope to personal non-trade bills. Gulf exchange houses were the first partners, and a Canadian fintech went live in 2025. The design choice that matters is that foreign providers join one network rather than signing thousands of bilateral deals. Our Bharat Connect explainer goes into the detail.

The Philippines, Mexico and Africa

The Philippines received a record $35.6 billion in cash remittances in 2025, according to central bank figures reported by The Philippine Star. Philippine banks with branches abroad have long offered overseas bills payment, letting customers in the US, Hong Kong and Japan pay Meralco, PLDT, Globe, water utilities, SSS and Pag-IBIG contributions, insurers and schools. The model works, but much of it still runs through branch counters and phone hotlines rather than open interfaces.

Mexico shows the pressure on the old cash model. FXC Intelligence, citing Banxico, says remittances fell 5% in 2025 to roughly $62 billion, the steepest drop since 2009, after a record $64.7 billion in 2024. When volumes shrink, providers compete harder on services such as direct bill pay. Our US-Mexico corridor analysis looks at what is driving the decline.

Sub-Saharan Africa is where the gap is widest. It remains the most expensive region to send money to, at 8.78% for $200 in the first quarter of 2025 according to the World Bank, even though mobile money, at 3.63%, is the cheapest way to fund a transfer worldwide. Diaspora apps have let Kenyans abroad pay electricity, water and school fees into mobile money systems since at least 2022, but coverage across the continent is patchy and split country by country.

A remittance tells a family that money is coming. A bill payment tells them the lights will stay on.

Where stablecoin settlement fits

Bill payments still need a local biller network at the destination. What stablecoins can change is the middle leg: how the sending provider funds the receiving gateway. Instead of pre-funding bank accounts in dozens of currencies or waiting on correspondent banks, a provider can settle in a dollar stablecoin at any hour and convert to local currency at the edge. Card networks are already testing this at scale. Visa said in April 2026 that its stablecoin settlement had reached a $7 billion annualised run rate across nine blockchains, The Block reported. In a bill payment the customer would never see a token. They would just see the bill marked paid sooner. We track the wider shift in stablecoin payment trends for 2026.

What still needs fixing

The G20's roadmap wants 75% of cross-border payments to arrive within an hour by 2027. An FXC Intelligence analysis of 2025 data put the share at about 35%. Biller coverage is uneven, protections when a bill payment fails differ from country to country, and very few national bill networks talk to each other. We think the next real gain will come from linking bill networks the way instant payment systems are now being linked, so a sender in Dubai, London or Houston sees the same bill a resident sees. That is part of the agenda behind the G20's cross-border payment targets.

A founder pitching a stablecoin payments company in autumn 2026 walks into an odd market. There are fewer cheques overall, fewer active funds, and yet more of the money that is moving is aimed squarely at what she is building.

The numbers, read carefully

Start with the long view. Galaxy Research counted about $20 billion of venture investment into crypto and blockchain companies in 2025 across 1,660 deals, the biggest annual total since 2022. The fourth quarter alone brought $8.5 billion, though eleven deals above $100 million made up 85% of it, led by a $3 billion round for Revolut. Later-stage companies took 57% of the year's capital, the highest share Galaxy has recorded.

2026 has been choppier. Galaxy put second-quarter crypto venture funding at $5.68 billion across 384 deals, up 31% from the first quarter, with roughly 77% going to late-stage companies and only about 7% to seed and pre-seed. Only five new crypto funds closed that quarter, raising about $3.9 billion, the weakest fund formation since late 2019.

The third quarter looked thinner still. CryptoRank's Q3 2026 recap found 127 venture rounds with $2.62 billion of disclosed capital, down about 42% from the prior quarter. Galaxy and CryptoRank count differently, so their figures should not be stacked side by side, but the direction is consistent: fewer deals, bigger average cheques, and a median round that rose to $8 million.

Valuations tell a similar story of selectivity. Galaxy put the median deal size in the fourth quarter of 2025 at $4 million, matching prior highs, and the median pre-money valuation at $70 million, though it cautioned that valuation data covered only about a tenth of deals and leaned toward later-stage companies. In other words, the headline medians say more about the companies that could command a price than about a typical seed round. Founders should read them as a ceiling, not a benchmark.

Payments took the top spot

Inside that shrinking pie, payments led. CryptoRank's data showed payments companies raising $791 million across 24 rounds in the third quarter, more rounds and more disclosed capital than any other category. The larger payments deals included Augustus with a $180 million Series B led by Tiger Global, Felix Pago with an $87 million Series C led by Andreessen Horowitz, Fasset at $68 million, Yellow Card at $40 million, and Series A rounds of $38 million for Velocity, $37 million for HIFI and $35 million for Latitude.

Earlier in the year, card issuer Rain raised a $250 million Series C at a $1.95 billion valuation, telling Crunchbase News it was processing more than $3 billion in annualized transactions. The momentum is not new. DL News, using DefiLlama data, reported that stablecoin-focused startups raised $537 million in 2025 by September, against $84 million in all of 2024.

Acquisitions are the other half of the story. Stripe paid $1.1 billion for Bridge in early 2025, and Mastercard agreed in March 2026 to buy BVNK for up to $1.8 billion. Those sales matter to early investors because they prove there are buyers for licensed payment infrastructure, which is not true of every crypto niche.

Why payments-led stablecoin companies attract capital

Three reasons come up repeatedly in conversations with investors. First, revenue is legible. A business that moves supplier payments or payroll earns a fee per transaction, and that looks like fintech, a sector generalist funds understand. Second, regulation has given larger investors cover. Galaxy expects the GENIUS Act to strengthen the US share of deals, and EU firms now have MiCA as a reference point. Third, the demand is visible in the data: McKinsey estimates business-to-business stablecoin payments grew more than 700% in 2025. Our stablecoin payments trends for 2026 piece covers that growth.

The wider fintech market helps too. Crunchbase counted $51.9 billion of global venture funding into fintech in 2025, up about 27% on 2024, even as deal counts fell. That pattern of fewer, larger rounds now applies to stablecoin startups as well.

Investors are no longer paying for a token or a chain. They are paying for corridors that already carry real money.

What each stage expects

There is no rulebook, but founders and investors describe fairly consistent expectations for payments companies.

Seed and pre-seed have become harder. CryptoRank counted 40 seed and pre-seed rounds worth $255 million in the third quarter, down from 75 rounds and $369 million a year earlier. Outside crypto, Carta's Q2 2026 pre-seed data shows US instruments averaging a record $276,000, up 27% on the year, with AI companies taking about half of all pre-seed dollars. Payments founders are competing for the same angels and funds.

How founders are actually raising

• SAFEs at the earliest stage. Pre-seed money in the US mostly arrives through SAFEs and convertible notes, which Carta tracks at more than 11,500 instruments in the quarter.
• Strategic investors. Corporate venture arms are active. Coinbase Ventures joined 13 rounds in the third quarter, more than any other fund in CryptoRank's data, though it led none.
• Licence first, volume second. Many teams spend early capital on registrations and banking partners before marketing, because that is what later investors and acquirers price.
• One corridor, done well. A narrow focus, such as Gulf-to-South Asia payouts or Latin American contractor payroll, is easier to diligence than a promise of global coverage. Our coverage of the stablecoin investment opportunity explains why corridor depth matters.

What makes investors walk away

The concerns are predictable. Volume that comes from trading or internal transfers rather than commerce. Dependence on a single banking partner that could pull out. A business model that only works if one token keeps its current share. And unclear licensing, especially for companies that touch customer funds across borders. Fund concentration adds pressure: CryptoRank counted 333 active institutions in the third quarter, down from 767 a year earlier, so each investor can afford to be choosier.

For founders, the practical read is that the window is open but narrow. Capital exists for payments companies with real flows, licences in progress and a clear first market. Teams still at the idea stage face a smaller pool and longer processes. For background on the rules shaping these companies, see our guide to the GENIUS Act and MiCA.

Funding figures come from different trackers with different methods and are not directly comparable across sources. Nothing here is investment advice.

Anyone in Washington whose job is to read stablecoin rules has had a heavy autumn. The OCC's main proposal runs to 376 pages. The Federal Reserve added two companion proposals on 29 September 2026. Treasury published an interim final rule the next day, and its proposal on what counts as "offering" a stablecoin in the United States closes for comments on 19 October. Fourteen months after President Trump signed the GENIUS Act, the law is being turned into actual rulebooks, and it is running late.

This piece maps where each agency stands as of mid-October 2026, what the draft rules actually require, and which dates matter between now and July 2028. For the statute itself and how it compares with Europe, see our earlier explainer on GENIUS Act and MiCA stablecoin rules.

The statute in five numbers

The Guiding and Establishing National Innovation for U.S. Stablecoins Act became Public Law 119-27 on 18 July 2025, after passing the Senate 68-30 on 17 June and the House 308-122 on 17 July, according to the Congressional Research Service. Five numbers do most of the work:

• 1:1 reserves. Every payment stablecoin must be backed at least dollar for dollar by cash, insured deposits, Treasury bills, Treasury-backed repos, government money market funds and similar assets.
• 93 days. Treasury securities in the reserve must mature in 93 days or less, as Fenwick's summary of the Act sets out.
• $10 billion. Non-bank issuers below this outstanding amount can choose a certified state regime. Above it, they move under federal supervision.
• $50 billion. Issuers above this level must publish annual audited financial statements.
• 18 months. The Act takes effect on the earlier of 18 months after enactment (18 January 2027) or 120 days after regulators finish their rules.

The law also told regulators to write their rules within a year. That deadline was 18 July 2026, and seven federal agencies missed it, as coverage of Treasury's first binding rule noted. Missing it has no legal penalty. What it changes is the order of events: the effective date is now almost certainly the fixed 18 January 2027 backstop, not a date set by finished rules.

Who is writing which rule

GENIUS spreads the work across the agencies that already supervise banks, credit unions and money flows. The table below is the scorecard as of 15 October 2026, drawn from Morgan Lewis's tracker, Arnold & Porter's review of the Fed proposals and the Federal Register.

The pattern is clear. The OCC carries the heaviest load because it supervises every non-bank federal issuer, and it is the only agency that has publicly committed to a final rule date. The Fed was last to publish, which matters because its proposals cover large state-chartered issuers once they cross $10 billion.

What the OCC wants: reserves, redemption and capital

The OCC's proposal was approved on 25 February 2026 and published on 2 March. It is the template others are following, and its detail is where the business models get decided.

On reserves, the OCC offered two options. Under Option A, a safe harbour checked every business day, an issuer would keep at least 10% of reserves in demand deposits or money at a Federal Reserve Bank, at least 30% in that form or in assets that come due within five business days, no more than 40% at any one bank, and a weighted average maturity of 20 days or less. Option B would make those benchmarks mandatory. Issuers with at least $25 billion outstanding would also hold at least 0.5% of reserves, capped at $500 million, as insured deposits, according to Gibson Dunn's analysis.

Redemption must generally happen within two business days. If requests exceed 10% of outstanding coins within 24 hours, the issuer can stretch that to seven calendar days, provided it tells the OCC within a day. Capital starts at the greater of $5 million or whatever the approval order says, and the OCC itself signals that a viable business probably needs $6.05 million to $25 million.

The yield ban is where the real fight sits, because it decides who gets paid for the float.

The OCC also writes the statutory ban on paying "any form of interest or yield" for simply holding a stablecoin into rule text, with a rebuttable presumption that affiliate or third-party reward schemes that mimic interest are caught too. Exchanges that share reserve income with users through partner programmes are the obvious target. For how that income is earned in the first place, see how stablecoin issuers make money.

The Fed's late entry and its capital formula

The Fed's two proposals, published in the Federal Register on 29 September 2026 under Docket R-1899, cover subsidiaries of insured state member banks and uninsured state-chartered institutions with $10 billion or more outstanding. They mirror the OCC on 1:1 segregated reserves and two-day redemption, and add a tiered operational-risk capital charge that is easy to model.

Worked example: capital for a $60 billion issuer

Take a hypothetical Fed-supervised issuer with $60 billion in circulation and average non-reserve revenue (fees, services) of $200 million a year over three years. Under the proposed formula:

• 2.0% on the first $20 billion: $400 million
• 1.5% on the next $30 billion: $450 million
• 1.0% on the remaining $10 billion: $100 million
• 25% of $200 million non-reserve revenue: $50 million

That totals about $1.0 billion of operational-risk capital, roughly 1.7% of outstanding coins, before any separate 2% charge on uninsured deposit exposure in the reserve. If the same issuer earned 4% on its $60 billion reserve, gross income would be around $2.4 billion a year, so the capital ask is meaningful but not crippling. A smaller issuer at $5 billion would face about $100 million plus its revenue add-on, which is a far larger bite relative to income. This is a stylised illustration of a proposal, not a forecast; issuers should take advice on the final rule.

The Fed also proposes a 12-month exam cycle, weekly confidential reports and quarterly financial reports. Where a bank owns the issuer, the bank would deduct the subsidiary's minimum capital from its own CET1, dollar for dollar. That deduction may quietly steer large banks toward tokenised deposits instead of a separate stablecoin subsidiary.

State versus federal: the $10 billion line

The state route is where the politics live. Treasury's 3 April principles say state regimes can differ in form and procedure but must be "substantially similar" in substance. Its 30 September interim final rule then set up the procedure for the Stablecoin Certification Review Committee, chaired by Treasury with the heads of the Fed and FDIC, to certify states. States must win initial certification by 18 January 2028 and recertify every year, though the committee cannot accept filings until the Office of Management and Budget clears the paperwork.

Once a state-licensed issuer crosses $10 billion, the OCC's draft gives it five business days to notify, 270 days to complete a capital analysis and 360 days to move to federal oversight, win a waiver or stop new issuance. In practice the state lane is for smaller and regional issuers. Circle, with USDC at roughly $73 billion as of October 2026, is already on the federal track: it received conditional OCC approval for a national trust bank on 12 December 2025 alongside Ripple, BitGo, Fidelity Digital Assets and Paxos, and final approval in July 2026. Notably, USDC issuance itself is slated to sit in a New York limited-purpose trust company, with reserve management at the national bank a later step.

Foreign issuers: the door is drawn but not built

For readers outside the US, the most consequential text is Treasury's 18 August proposal on issuance, offer and sale. It treats a stablecoin as issued in the US if either the issuer or the recipient is located there, and it counts answering an unsolicited US inquiry as an offer. That removes the reverse-solicitation comfort that many offshore platforms rely on in other US regimes, as Freshfields points out.

A foreign payment stablecoin issuer can reach US users through a platform only if Treasury first finds its home regime comparable, and the issuer registers with the OCC, accepts US enforcement and holds enough reserves at a US institution to cover US redemptions. Treasury has not yet proposed how comparability will be judged. So in January 2027 there may be a legal door for offshore issuers with no process to walk through it. Platforms can lean on an issuer's representation that it can comply with lawful freeze orders, if they do reasonable diligence.

This matters most for USDT, at about $184 billion the largest stablecoin. Tether is not organised under US law, so the state route is closed to it. Its path to US platforms after July 2028 runs either through comparability for its home jurisdiction or through a US-law vehicle. Our USDT vs USDC comparison covers the reserve differences that make that harder or easier.

Penalties and AML: the teeth

The Act carries criminal penalties of up to $1 million per violation and five years' imprisonment for knowing participation in unlawful issuance, civil penalties of up to $100,000 per day, and fines of up to $500,000 for marketing something as a payment stablecoin that was not issued under the Act. Issuers must be able to block, freeze and reject transfers when lawfully ordered, and certify their AML and sanctions programmes within 180 days of approval and annually after that, per the CRS. A joint customer identification rule from FinCEN, the OCC, Fed, FDIC and NCUA appeared on 22 June 2026. One open question the statute leaves hanging: whether issuer AML duties stop at the first sale or follow coins through secondary on-chain transfers. The answer will shape how AML and sanctions compliance works for every wallet that touches a US-regulated coin.

What to watch next

• 19 October 2026: comments close on Treasury's issuance, offer and sale proposal. Watch whether Treasury moves toward a Regulation S-style conduct test, which it floated.
• November 2026: the OCC's target for its final rule. It is a target, not a promise, but it would be the first complete federal rulebook.
• 30 November 2026: comments close on the Fed proposals and the certification interim final rule.
• 18 January 2027: the Act takes effect. The OCC expects to process issuer applications from 2027, and existing applicants may get up to 12 months of regulatory forbearance.
• 18 January 2028: deadline for initial state certifications.
• 18 July 2028: US platforms may generally offer only coins from permitted or qualifying foreign issuers.

Our read: the federal track is now the default for anything large, and the real competitive variable is how strictly the yield ban is enforced against platform rewards. The market structure bill still in Congress will decide the rest; see our piece on the US crypto market structure law. Nothing here is legal advice, and firms affected by these proposals should work with qualified US counsel.

A finance manager at a textile exporter in Tiruppur waits on a $48,000 payment from a buyer in Dubai. The buyer's bank sends it on a Thursday afternoon. It passes through a correspondent in New York, sits over a US weekend, and lands on Monday with a lifting fee deducted that nobody quoted in advance. Ripple has spent more than a decade arguing that this journey should take minutes rather than days. In 2026, with a national trust bank approval, a stablecoin of its own and several billion dollars of acquisitions behind it, the company is closer to testing that claim at scale than it has ever been.

What is Ripple Payments in 2026?

Ripple Payments is Ripple's cross-border payments service for banks, payment companies and businesses. It combines what used to be RippleNet messaging and On-Demand Liquidity, letting customers send value using XRP, the RLUSD stablecoin or fiat through a network of payout partners. As of December 2025, Ripple said the network had processed more than $95 billion in total volume.

The product has changed shape. Earlier, RippleNet was mainly a messaging network that banks used alongside traditional nostro accounts, while On-Demand Liquidity used XRP to avoid pre-funding. Today Ripple sells an end-to-end service: it can collect funds, convert them, move them across a blockchain and pay out locally, holding the licences itself. In a May 2025 release on its UAE customers, Ripple described more than 60 licences and registrations and over 90 payout markets, covering more than 90% of daily FX markets. For the broader context of the company's year, see our overview of Ripple in 2026.

Note what the volume figure is: cumulative, self-reported and across all years. It is not an annual number, and Ripple does not break it down by corridor or settlement asset.

How does a Ripple payment work, step by step?

In a typical Ripple Payments transfer in 2026, the sending institution submits a payment, Ripple or a partner converts the local currency into a digital asset such as XRP or RLUSD, the asset moves across a blockchain in seconds, and a payout partner converts it into the destination currency and credits the recipient. The sending institution never needs a pre-funded account abroad.

1. Instruction. A bank or payment firm, say a UAE remittance company, sends a payment request through Ripple's API with amount, currency and beneficiary details.
2. Quote and compliance. Ripple returns an FX quote and runs screening. Because Ripple holds licences in the relevant markets, it can take on part of the compliance work.
3. On-ramp. Dirhams are converted into a digital asset. For XRP routes this happens through exchange liquidity; for RLUSD routes, through issuance or purchase of the stablecoin.
4. Transfer. The asset moves on the XRP Ledger or another supported chain. Settlement on the ledger takes seconds and runs around the clock.
5. Off-ramp and payout. A partner in the destination market sells the asset for local currency and pays the beneficiary through a domestic rail such as a bank transfer or mobile wallet.

The economic point of steps three to five is pre-funding. In correspondent banking, a sender must keep money parked in accounts abroad, or pay a bank that does, so that a payment can be credited on arrival. Ripple's model replaces those idle balances with assets bought and sold on demand.

What is the difference between XRP and RLUSD in Ripple settlement?

XRP is a floating-price digital asset used as a bridge between two currencies, so the customer is exposed to its price for the seconds the payment is in flight. RLUSD is a dollar stablecoin issued by Ripple under a New York trust charter, so there is no price movement in transit. In 2026 Ripple Payments supports both, and corridor liquidity decides which is cheaper.

RLUSD launched on 17 December 2024 on the XRP Ledger and Ethereum, backed by dollar deposits, US government bonds and cash equivalents, with monthly attestations promised, according to Ripple's launch release. Its market value passed $1 billion within a year and stood at about $2.4 billion in mid-October 2026, per DefiLlama. That makes it a mid-sized stablecoin, far behind USDT and USDC, as our stablecoin companies list shows.

For banks, the stablecoin route answers the most common objection to XRP: treasury teams do not want price risk on their balance sheet, even for minutes. For XRP-based routes, the argument is depth: in corridors where XRP trades actively against both currencies, using it can avoid two stablecoin conversions. Neither choice is a statement about the asset's future value, and readers should not treat it as one.

Is Ripple the same as XRP?

No. Ripple is a private US company that sells payments, custody, treasury and brokerage services. XRP is a digital asset native to the XRP Ledger, an open-source blockchain that Ripple helped create but does not own. Ripple holds a large amount of XRP and releases one billion XRP from escrow each month, which links the two closely without making them the same thing.

The distinction has legal weight. The US Securities and Exchange Commission's case against Ripple ended on 7 August 2025, when both sides dismissed their appeals, leaving in place a final judgment with a civil penalty of $125,035,150 and an injunction against violating securities registration rules, according to the SEC's litigation release. The case concerned how Ripple sold XRP, not whether the XRP Ledger could be used for payments.

It also matters commercially. A bank can use Ripple Payments without ever touching XRP, by settling in RLUSD or fiat. And anyone can build on the XRP Ledger without Ripple, as several stablecoin issuers do. Our comparison of Stellar's payments blockchain, which has a similar split between a foundation and an open network, is a useful parallel.

What licences and acquisitions does Ripple have?

By late 2025 Ripple said it held more than 60 licences and registrations, including a New York trust charter for RLUSD and a Dubai Financial Services Authority licence. It received conditional OCC approval for a national trust bank in December 2025 and a full MiCA licence in the EU in July 2026. It has spent billions buying firms in brokerage, treasury and stablecoin payments.

The figures come from The Block's report on the $40 billion round, which also noted more than $4 billion of strategic investments in 2025; Ripple's Rail announcement, which said Rail was expected to handle over 10% of a $36 billion B2B stablecoin market in 2025 (the deal completed on 11 August 2025); Ripple's October 2025 GTreasury announcement; and Ripple's December 2025 release on the trust bank. In 2026 Ripple added customers including Convera in March and South Korea's Jeonbuk Bank in August, and CEO Brad Garlinghouse has set a target of a $1 billion annual operating revenue run rate by the end of 2026, excluding XRP sales.

Ripple's bet is no longer a token. It is a licensed, vertically integrated payments company that happens to own one.

How does Ripple compare with SWIFT correspondent banking?

SWIFT is a messaging network, while correspondent banks move the money. Ripple Payments bundles messaging, settlement and payout into one service. SWIFT says 90% of payments on its network reach the destination bank within an hour, but only 43% reach the end customer that fast, which is where Ripple claims an edge.

Those SWIFT figures, announced in October 2024 and reported by PYMNTS, already beat the G20 target of 75% of payments processed within an hour by 2027 at the bank-to-bank level. The weak point is the last mile, where domestic processing, compliance checks and cut-off times slow delivery. SWIFT's ISO 20022 migration is designed to reduce that friction with richer data.

Ripple's advantage is strongest where correspondent banking is thinnest: smaller currencies, weekend payments and corridors where pre-funding is expensive. Its disadvantage is reach. SWIFT connects nearly every bank in the world; Ripple's 90-plus payout markets depend on local partners, and in some countries regulation limits what digital assets can be used. In India, for instance, Ripple's publicly announced deployments have been bank-led, such as LuLu Exchange and Federal Bank connecting through RippleNet Cloud in 2021 for transfers from LuLu's 76 UAE branches.

A worked example: what the cost gap means

Ripple does not publish a public price list, so a direct fee comparison is not possible. What can be measured is the cost Ripple says it can undercut. The World Bank's global average for sending $200 was 6.36% in the third quarter of 2025. On a $200 transfer, that is $12.72. On $1,000 sent as five $200 transfers over a year, the same average adds up to $63.60.

Now take the Tiruppur exporter's $48,000 invoice. Suppose, purely as an illustration, that the correspondent route charges a $40 combined fee plus 0.5% hidden in the FX rate. That is $40 plus $240, or $280, and three days of float. A provider that settled in minutes for, say, 0.3% all-in would cost $144 and release the cash the same day. Those percentages are hypothetical; the point is that savings come mainly from the FX spread and pre-funding, not the headline fee. Our cost of sending money home report shows how to read a quote.

Can Ripple Payments be used for India and the UAE?

Ripple Payments has a licensed base in the UAE through its DFSA licence, granted in March 2025, and announced Zand Bank and Mamo as UAE customers in May 2025. For India, deployments have run through banks and licensed money transfer firms, since the user receives rupees through the domestic banking system rather than holding any digital asset.

The Gulf-to-India corridor is one of the largest in the world, and India received $135.46 billion in remittances in 2024-25, according to the RBI. That size is why every network wants a share. For senders, the practical question is not whether a provider uses Ripple behind the scenes, but the total cost and delivery time quoted. Our guide to UAE to India money transfer compares providers on those terms.

What should you watch next?

The developments most likely to change Ripple Payments over the coming year are its national trust bank moving to a final charter, its push to hit a $1 billion revenue run rate by end-2026, EU expansion under its July 2026 MiCA licence, and US market structure legislation that would define XRP's status.

• Revenue target, December 2026. Ripple's year-end run-rate goal excludes XRP sales, so it will be a test of whether payments, custody, treasury and brokerage can stand alone.
• Trust bank charter. A final OCC charter would put RLUSD under federal supervision alongside NYDFS oversight.
• EU rollout. MiCA passporting from Luxembourg lets Ripple serve customers across the EU. Watch for named European bank customers.
• US crypto market structure law. Senate progress on the CLARITY Act, after a procedural vote scheduled for 15 September 2026, would affect how XRP and other tokens are classified. Our Ripple vs Stellar comparison will track the impact.

This article explains how a payments product works. It is not investment advice and makes no prediction about the price of XRP or any other asset. Businesses considering any cross-border provider should take professional advice.

In a tailoring shop in Karama, the owner keeps three ways to get paid within arm's reach: a card terminal from one of the big acquirers, a phone that accepts tap-to-pay through a licensed wallet, and a QR code for the central bank's instant payment app. Half his customers split their festive orders into instalments. Two streets away, his cousin sends dirhams home from an exchange house counter he has used for a decade, while his nephew does the same thing from an app. That is the UAE fintech market in 2026 in a single block: incumbents and challengers selling to the same people, each under a different licence.

How we chose these companies

We selected UAE fintech companies using measurable criteria as of October 2026: reported revenue or profit, users and merchants, funding and valuation, and the licences each firm holds from the Central Bank of the UAE (CBUAE), Dubai's Virtual Assets Regulatory Authority (VARA) or Abu Dhabi Global Market (ADGM). Companies are grouped by category, not ranked, and every figure is dated and linked.

A note on borders: several of the region's biggest fintechs operate across the Gulf. Tabby moved its headquarters to Saudi Arabia, and Tamara is a Saudi company with UAE customers. We include firms with a material UAE business, and cover the Saudi side in our Saudi Arabia fintech guide.

Who regulates fintech in the UAE in 2026?

In 2026 UAE fintech is regulated by several authorities depending on activity and location. The CBUAE licenses banks, payment firms and stored value facilities onshore; VARA licenses virtual asset firms in Dubai outside the financial free zone; and ADGM's regulator covers firms in Abu Dhabi's financial centre. Dubai's DIFC has its own regulator, the DFSA. Choosing the right licence is often a company's first strategic decision.

The DIFC route deserves its own line. The DFSA, the independent regulator of financial services in the centre, has offered an Innovation Testing Licence since May 2017, a restricted permission that lets a fintech test a product for six to 12 months before seeking full authorisation. The centre is now large: DIFC had 10,018 active registered companies at the end of June 2026, including 1,933 AI, fintech and innovation firms, up 39% in a year, and its Innovation Hub added 361 companies in the first half, according to the Dubai Media Office.

VARA's framework covers eight regulated virtual asset activities, from exchange and broker-dealer services to custody, lending and token issuance, and requires every licensee to follow four compulsory rulebooks on company structure, compliance, technology and market conduct, according to VARA's licensing page. The CBUAE, for its part, runs the payment plumbing directly: its subsidiary Al Etihad Payments launched Aani, an instant payment platform that lets users pay with a phone number, on 16 October 2023, per the CBUAE announcement. See our UAE crypto regulation guide.

Which are the largest consumer fintechs in the UAE?

The largest consumer fintechs active in the UAE in 2026 are Tabby, the buy now, pay later company valued at $6.5 billion in September 2026, and Wio Bank, the Abu Dhabi digital bank that earned AED 622 million in 2025. Tamara is the other major instalment provider, and Ziina has built a wallet and payments business under a central bank licence.

Tabby

Founded in the UAE in 2019, Tabby announced in 2023 that it would move its headquarters to Saudi Arabia after getting a Saudi Central Bank licence, Finance Middle East reported. In September 2026 it raised $233 million in a Series F at a $6.5 billion valuation, up from $4.5 billion at an October 2025 share sale, with 25 million registered users, 70,000 business partners and more than $18 billion of annualised transaction volume, according to EnterpriseAM. In the UAE it holds a stored value facility licence for TabbyCash, a spending account. The company says it has been profitable since 2023 and is in no rush to list.

Wio Bank

Wio is the clearest proof that a digital bank can make real money in the Gulf. It reported 2025 revenue of AED 1.24 billion, up 55%, and net profit of AED 622 million, up 57%, with deposits above AED 57 billion at year end, per Entrepreneur Middle East. It serves more than 270,000 individuals and 120,000 businesses.

Ziina

Ziina became the first venture-backed startup to get a CBUAE stored value facility licence, as Mubasher reported in August 2024, and said at the time it supported over 557,000 businesses. It lets individuals send money and lets small merchants accept cards, Apple Pay and Google Pay, including tap-to-pay on a phone. Ziina's own site cites a $22 million Series A in September 2024.

Tamara

Tamara is Saudi-licensed but sells to UAE shoppers too, with instalment plans of up to 24 months and no late fees, according to its website, which lists Saudi Arabia, the UAE, Kuwait and Bahrain as markets. It became a unicorn with a $340 million Series C in December 2023, and in September 2025 secured an asset-backed facility of up to $2.4 billion from Goldman Sachs, Citi and Apollo-managed funds, citing more than 20 million customers and 87,000 merchants, Arab News reported. It has not published a newer valuation.

Which companies run the UAE's payments infrastructure?

The UAE's merchant payments in 2026 are dominated by Network International, which merged with First Abu Dhabi Bank's former payments arm Magnati in October 2025 under a Brookfield-led consortium, alongside the central bank's own Aani instant payment rail. Bank-linked players such as Mashreq's Neo Pay and data firms like Lean Technologies fill the gaps.

Network International (with Magnati)

Network International was taken private by a Brookfield-led consortium that included First Abu Dhabi Bank, Mubadala, ADQ and Olayan, after a recommended cash offer of 400 pence a share that valued it at about £2.2 billion; the deal closed in September 2024, per Linklaters, which advised the buyers. Network and Magnati then completed their merger at the start of October 2025, creating a group operating in 56 markets across the Middle East and Africa, Gulf News reported. Before the merger Network alone served more than 240,000 merchants and over 250 financial institutions. It is also an early stablecoin mover: Gulf News reported in January 2026 that Network planned to accept AE Coin, a CBUAE-licensed dirham stablecoin, at point of sale and online.

Lean Technologies

Lean sells APIs that let apps verify bank accounts, pull financial data and take pay-by-bank payments. Its website says it has processed more than AED 16 billion, connected over 2 million accounts in MENA and serves 350+ clients, including Tabby and Careem. It holds ADGM authorisation and a CBUAE in-principle approval for open finance services, which makes it a bet on the UAE's open finance rules being switched on in full.

Mashreq Neo and the banks

Mashreq's digital arm Neo is one of the bank-owned challengers, and Neo Pay was among the first acquirers rolling out Aani merchant QR payments, according to the CBUAE launch announcement. Banks are also moving into tokenised money: First Abu Dhabi Bank, ADQ and IHC said in April 2025 they would launch a dirham-backed stablecoin subject to CBUAE approval, The National reported, and RAKBANK has in-principle approval for its own.

Which UAE fintechs serve remittances and property?

Al Ansari Financial Services, the listed exchange house, is the largest dedicated remittance brand in the UAE covered here, with 441 branches across four countries in March 2026. In property finance, Dubai-based Huspy says its mortgage unit processes more than a quarter of Dubai's residential home financing.

The size of the remittance pool explains the competition. Outward remittances through UAE exchange houses rose 27.8% to Dh188.8 billion in 2025, of which Dh115.7 billion were personal remittances and Dh63.3 billion trade-related, according to the CBUAE's Financial Stability Report 2025 as reported by Khaleej Times.

Al Ansari Financial Services

Al Ansari reported H1 2026 operating income of AED 635 million, down 0.5%, and net profit of AED 151 million, down 28.7%, per its results release on Zawya. In Q1 the company blamed regional conflict hitting tourism, margin pressure from competitors including fintechs, and higher financing costs, while digital transactions rose to 29% of the total from 24% a year earlier, AGBI reported. For background on the corridor, see our UAE to India transfer guide.

Huspy

Huspy raised a $59 million Series B led by Balderton Capital in July 2025, taking total funding past $100 million, MENAbytes reported. It runs a mortgage broker portal linked to more than 25 banks and says it facilitates over $7 billion of property transactions a year across Dubai, Abu Dhabi and Spain.

Which crypto companies are licensed in the UAE?

In 2026 the UAE's licensed crypto firms include BitOasis, which holds a full VARA licence, Binance, whose Nest entities received full ADGM authorisation in December 2025, and Rain, which won VARA in-principle approval in May 2026.

BitOasis

BitOasis has moved from a provisional permit to a full VASP licence from VARA, allowing it to serve retail, institutional and qualified investors, per The Fintech Times.

Binance

Binance FZE received an operational minimum viable product licence from VARA in July 2023, initially limited to institutional and qualified retail investors, Blockworks reported. In December 2025 Binance received full ADGM authorisation for three entities, an exchange, a clearing and custody house and a broker-dealer, starting services on 5 January 2026, per AGBI.

Rain

Rain, a crypto brokerage and custodian with Bahraini roots, received VARA in-principle approval for exchange, broker-dealer and margin trading services in May 2026, a regional crypto news site reported. It says it has over two million users and has processed more than $11 billion since 2017, and it already holds an ADGM licence.

How do the UAE's leading fintechs compare?

The table lists each company's main business and its latest dated metric. Metrics differ by company, so treat this as a map rather than a ranking.

Where are UAE fintech profits and funding moving?

In 2026 UAE fintech profits are shifting from transfer fees towards account balances, credit and merchant relationships, while funding is concentrating in a few large rounds such as Tabby's $233 million Series F. The figures below, taken from company results and MAGNiTT data, show both trends in numbers.

Worked example: margins at Wio and Al Ansari

Two sets of numbers show where the money is going. Start with Al Ansari. In H1 2026 it earned AED 151 million of net profit on AED 635 million of operating income, a net margin of about 23.8%. A 28.7% fall in profit implies H1 2025 net profit of about AED 212 million, on operating income of about AED 638 million, a margin of roughly 33.2%. On essentially flat income, the exchange house kept about 9.4 fewer fils of profit from every dirham it earned.

Now Wio. AED 622 million of 2025 profit on AED 1.24 billion of revenue is a net margin of about 50%. Spread across its roughly 390,000 personal and business customers, that is about AED 1,595 of profit per customer for the year. Most of that is earned on deposits and business accounts, not on transfer fees.

The UAE's fintech profits are drifting away from the transfer counter and towards whoever holds the account.

The lesson is not that exchange houses are finished. Al Ansari's digital volumes grew 69% year on year in Q1 2026. It is that pricing power on the transfer itself is eroding, while balances, credit and merchant relationships keep their margins. That is also why Tabby is pushing from instalments into accounts and SME finance.

Margins calculated from the reported figures; implied H1 2025 values are derived from the stated year-on-year changes and are approximate.

How much venture money is going into UAE fintech?

UAE startups raised a record $1.4 billion in the first nine months of 2025, up 188% year on year across 164 deals, according to MAGNiTT data reported by EnterpriseAM. MENA fintech as a whole took $965 million over the same period, up 97%.

Those totals have since been dwarfed by single rounds. Tabby's $233 million Series F in September 2026 alone equals about a sixth of the UAE's nine-month 2025 total. KPMG's Pulse of Fintech puts EMEA fintech investment at $11.3 billion in H1 2026, down from $18 billion in the second half of 2025, a reminder that Gulf money is not immune to the global cycle.

What to watch next in UAE fintech

• Q4 2026: whether Rain and other in-principle VARA approvals convert to full licences, and how many more exchanges follow BitOasis to full status.
• Dirham stablecoins: the CBUAE's decisions on the FAB-backed token and RAKBANK's in-principle approval, and how quickly merchants accept AE Coin through Network International.
• Tabby's listing choice: its CEO said in September 2026 it will pick the market that makes the most sense; a Tadawul or UAE listing would set the region's first big fintech IPO benchmark.
• Open finance: full licensing of providers such as Lean would let pay-by-bank compete directly with cards on fees.
• Remittance pricing: watch Al Ansari's Q3 2026 results for whether digital growth offsets margin pressure. Our Gulf to South Asia corridor analysis tracks prices.

Anyone considering a UAE fintech product or licence should take advice from a qualified local adviser, as permissions differ sharply between onshore, Dubai and Abu Dhabi regimes. Our explainer on Gulf payroll cards covers another part of the market where fintechs are competing with banks.

This list is editorial research, not investment advice or an endorsement of any company. Company figures are as reported by the linked sources; company-stated metrics have not been independently audited.

Ask a vegetable seller in Pune whether she accepts the digital rupee and she will probably point at her UPI QR code. In a sense she is right, because an e₹ wallet can scan that same code. But the money moving through it is different, and nearly four years into the Reserve Bank of India's pilot, that difference is still the whole story.

What the e₹ actually is

The e₹ is India's central bank digital currency. According to the RBI's own FAQ, it is legal tender, it sits at par with paper cash, and it is a direct liability of the RBI. That last point is the one people miss. The balance in your bank account is a promise from your bank. A rupee note in your wallet, and an e₹ in your phone, is a promise from the central bank itself.

The RBI creates the e₹ and hands it to banks and approved non-banks, which then give customers a wallet app. You load the wallet from your savings account, any time of day, with no fee and no minimum balance. The wallet pays no interest, which is deliberate: the RBI wants it to behave like cash, not like a savings product that pulls deposits away from banks.

Two pilots, two very different jobs

India runs two separate experiments. The wholesale pilot started in November 2022 and is aimed at institutions. The RBI lists 16 participants and four use cases: settling government bonds in the secondary market, call money, and tokenised issuance and settlement of certificates of deposit and corporate bonds. Ordinary people never touch it.

The retail pilot began on 1 December 2022. This is the version with phone wallets, merchant QR codes and person-to-person transfers. When two e₹ wallets pay each other, the transfer settles instantly and never passes through either bank account. If you scan a UPI QR from an e₹ app instead, the payment works but settles on UPI's timelines.

The numbers, as of late 2025 and 2026

User growth has been steady rather than explosive. RBI data cited by Business Standard shows the retail pilot going from 1.3 million customers in June 2023 to 5 million a year later, while merchants rose from 0.3 million to 0.42 million. The RBI's 2024-25 annual report put users at about six million across 17 banks. By December 2025, the deputy governor said the figure was a little above eight million.

Circulation tells a more awkward story. The RBI's statistical tables show total e₹ outstanding at ₹16.39 crore in March 2023, ₹234.12 crore in March 2024 and ₹1,016.50 crore in March 2025. Then it fell, to ₹771.66 crore by March 2026, according to the RBI handbook table published in July 2026. A pilot can shrink for dull reasons, such as banks ending incentive campaigns or institutions rebalancing wholesale holdings, and the table alone does not say which. But it is a reminder that headline growth in earlier years was partly pushed rather than pulled.

Non-banks join the party

In January 2025, MobiKwik and Cred became the first non-banks to offer e₹ wallets, with Yes Bank as the sponsor bank. Users complete video KYC, load the wallet through UPI, and can send up to ₹10,000 per transfer and ₹50,000 a day. This matters more than it looks. Most Indians already live inside a handful of payment apps, and asking them to download a separate bank wallet for a pilot was always a hard sell.

The features that make it more than a copy of UPI

Programmable money

This is the e₹'s real selling point. A government department or an employer can attach conditions to the money it sends: an expiry date, a location, a merchant category, even a specific merchant. The wallet shows the user what those conditions are. The RBI's annual report pointed to pilots such as direct benefit transfers under Odisha's Subhadra Yojana and company-issued fuel and meal allowances.

Picture a state scheme paying a farmer in Koraput a fertiliser subsidy that can only be spent at registered agri-input shops before the sowing season ends. That is hard to do cleanly with bank transfers. With programmable e₹ it is a setting.

Offline payments

The RBI said in its 2024-25 annual report that an offline capability had been introduced so payments can work with weak or no connectivity, though it has not published usage numbers. For a country where a trip into a hill district can mean losing signal for hours, this is the one feature that cash still beats every app on.

Cross-border ambitions

The RBI has said it is exploring cross-border CBDC pilots, without giving a timeline. If you want to understand how central banks are thinking about that, our pieces on Project mBridge and Project Nexus cover the two main models.

The e₹ will not win by being a faster UPI. It wins only where money needs rules attached.

So why is everyday usage still small?

Put the e₹ next to UPI and the scale gap is stark. UPI handled an average of about 802 million transactions a day in September 2026. The e₹ took roughly three years to reach 120 million in total. Most people do not have a problem that the e₹ solves better than what they already use, which is why the RBI keeps leaning on programmability and government flows. We look at that trade-off in detail in e₹ vs UPI.

There is also no reward for holding it. No interest, plus a per-transaction cap on some wallets, makes the e₹ a pass-through rather than a place to keep money. That is by design, and we think it is the right call for financial stability, but it limits organic growth.

What to watch next

• Government payouts. If more state schemes move subsidies onto programmable e₹, user numbers will rise for reasons that have little to do with consumer choice.
• Circulation in 2026-27. Whether the March 2026 dip reverses will say a lot about genuine demand.
• Wholesale tokenisation. The bond and certificate-of-deposit trials could turn out to be the quieter success story.
• Cross-border pilots. Any formal announcement would change the conversation for the large Indian diaspora.

Figures are from RBI publications and statements reported by Indian media. Circulation figures are end-March totals for retail and wholesale e₹ combined.

On the last day of April 2026, a line cook in Abu Dhabi from Thrissur checked his phone between lunch and dinner service and saw a dirham buying almost 26 rupees, a level Gulf News called one of the strongest windows yet for remitters. Two years earlier the same dirham averaged closer to 22.8. He sent three months of savings that night. Whether that was clever or lucky is a question every Indian in the Emirates is asking this year. It sits on top of a duller question that decides more of what actually arrives: what each transfer costs.

How big the UAE-India corridor is

The UAE is India's second-largest source of remittances. The RBI's sixth remittances survey, published in its monthly bulletin, found it accounted for 19.2% of inflows in 2023-24, behind only the United States at 27.7%. Applied to that year's total of about $118.7 billion, that implies roughly $23 billion a year moving from the Emirates to India. Indian officials put the number of Indians living in the UAE at more than three million.

The Gulf's share is slowly shrinking as more Indians work in advanced economies; the six GCC states together fell from 46.7% of India's remittances in 2016-17 to 37.9% in 2023-24. In volume terms, though, UAE to India remains one of the busiest corridors anywhere. We look at the wider region in the Gulf to South Asia corridor.

What it costs: the World Bank numbers

The World Bank's corridor data for the third quarter of 2025 put the average total cost of sending $200 (735 dirhams) from the UAE to India at 3.72%. That breaks down into an average fee of about 21.8 dirhams plus an exchange-rate margin of just 0.75%. It is below the South Asia regional average of 4.80% but well above the UK-India corridor, where the same transfer averaged 1.91%.

The average hides a wide spread. Across 24 listings from 12 firms, the cheapest bank option cost 0.77% and the cheapest money transfer operator 1.04%. At the other extreme, some bank transfers in the sample cost more than 21%. On a $500 transfer the best options fell below 1%. The pattern is clear: in this corridor the flat fee matters more than the exchange margin, which is why so many people send larger amounts less often.

Exchange houses, apps and banks compared

In the World Bank sample, money transfer operators outnumbered banks 20 listings to four, and their costs sat in a tight band of roughly 1% to 3.4%. Bank pricing ranged from the cheapest in the corridor to the most expensive. The lesson is to compare specific products, not types of institution. Globally, digital transfers averaged 4.85% in early 2025 against 7.16% for non-digital ones, so moving from the counter to an app is usually the single easiest saving.

Counters still have a role. Workers paid in cash, people without a card that works online, and anyone sending a large sum for a property purchase often value a human who can check the paperwork. The point is to use the branch deliberately, for the transfers that need it, rather than out of habit for every monthly payment home.

UPI and Aani: what is live and what is still a plan

The UAE's Al Etihad Payments, a subsidiary of the central bank, launched Aani, its instant payments platform, on 16 October 2023. It lets users send money to a phone number, split bills and pay by QR code. India and the UAE agreed to link UPI with Aani during Prime Minister Narendra Modi's visit in February 2024, and officials have spoken of deeper integration since.

What has arrived first is merchant acceptance for Indian visitors. NPCI International said in July 2025 that UPI QR payments were accepted at outlets including Dubai Duty Free and Lulu Hypermarket through local acquirers. A full person-to-person link, where a UAE resident pays a phone number in India from a local bank app in seconds, is the piece remittance users are waiting for. Until it is broadly available, check which leg of an advertised instant transfer is actually instant. Our UPI global expansion piece tracks the links country by country.

Timing the rupee: why chasing the rate is overrated

The rupee's slide in 2026 has turned many Indians in the UAE into part-time currency traders. Because the dirham is pegged to the US dollar, the dirham-rupee rate simply mirrors the rupee against the dollar. Khaleej Times reported in early June 2026 that the rupee was trading around 95.8 to the dollar after nearly touching 97 in May, putting a dirham at about 26 rupees. Someone sending 5,000 dirhams a month would have received about Rs 114,000 on average in 2024 and about Rs 130,000 at June 2026 rates.

That is a real windfall, but picking the peak is a bet on oil prices, conflict and capital flows. A calmer approach is to send fixed obligations such as loan EMIs and school fees on schedule, and move only discretionary savings when the rate looks unusually good. A rate alert is useful. A week spent waiting for one more paisa usually is not. And a strong headline rate with a high fee can deliver fewer rupees than a slightly weaker rate with none.

In this corridor the fee, not the exchange rate, usually decides who gets the better deal.

Paying bills in India directly from the UAE

The UAE was among the first countries where India's Bharat Connect network went live for cross-border bill payments, alongside Oman, Kuwait and Bahrain. Some exchange houses let you pay a relative's electricity bill, insurance premium or school fees against the biller's own consumer number, so the money never sits in someone else's account. If you hold an NRE account, UPI with a UAE mobile number is another way to pay Indian bills in rupees. Our guide to bill payments for NRIs compares both.

Before you send: a short checklist

• Compare the total rupees delivered, not the fee or the rate on its own.
• Use digital channels where you can; they are cheaper on average than counter services.
• Confirm the provider is licensed by the Central Bank of the UAE.
• Ignore social media agents offering rates far above the market; they are a classic front for fraud, as our money transfer scams guide explains.
• Keep receipts for large transfers, particularly property payments and deposits into NRE accounts.

A furniture importer in Dubai owes a workshop in Jakarta $48,000, and the wire she sent on Monday is still somewhere between two correspondent banks on Thursday. That gap, measured in days and fees, is why business use of USDC and USDT has grown faster than any consumer use case.

Business is where the stablecoin volume is

Most stablecoin activity is still trading. When McKinsey tried to isolate genuine payments, it estimated about $390 billion a year based on December 2025 activity, and roughly 60% of that, about $226 billion, was business-to-business. That segment grew by more than 700% in 2025. Payroll and remittances added about $90 billion. The figures are small against global B2B flows, but the growth rate is why finance teams are asking questions.

The two tokens they ask about are the two that dominate supply. As of October 2026, USDT had about $184 billion in circulation and USDC about $73 billion, according to DefiLlama. Our comparison of USDT vs USDC covers reserves and issuers in detail. For a business, the practical difference is often about counterparties: USDT is widely used by traders and in many emerging markets, while USDC is the token most often supported by large US and European payment processors.

Four ways businesses use stablecoins

For supplier and payroll flows, the stablecoin leg is usually the easy part. The hard parts are the edges: converting into USDC from a bank account, and converting out of it into rupees, pesos or naira on the other side. Many firms use a payout provider that handles both ends, so the supplier never sees a wallet at all. Our article on correspondent banking explains what the stablecoin leg replaces.

Take a design agency in London paying twelve freelancers across Lagos, Manila and Mexico City each month. With wires, it pays a fee per transfer, a foreign exchange margin, and often a receiving-bank charge, and some workers wait days. With a stablecoin payout provider, the agency funds one USDC balance and the provider sends each worker either tokens or local currency. The savings depend almost entirely on how good the local cash-out is in each country, which is why the same tool can be excellent in one corridor and pointless in another.

Network choice matters as well. The same token exists on several blockchains with different fees and confirmation times, and both sides must agree on which one to use before the first payment. Finance teams that write this into supplier onboarding avoid most of the early mistakes.

How acceptance works in practice

Large processors have added stablecoins as a payment method, and the details matter.

• Stripe. According to Stripe's documentation, customers can pay in USDC on Ethereum, Solana, Polygon, Base and Tempo, plus USDP and USDG for US merchants. Payments settle into the merchant's Stripe balance in local currency. Acceptance is available to US businesses, with a private preview in the EU, Hong Kong, Mexico and Switzerland. There is a $10,000 limit per customer transaction, refunds go back in stablecoins, and there are no chargebacks. Note that USDT is not on Stripe's accepted list.
• Shopify. In June 2025 Shopify announced USDC checkout on Base, built with Coinbase and Stripe. By default merchants receive local currency, with an option to take USDC into their own wallet.
• Card networks. Visa says stablecoin settlement with its partners has passed a $20 billion annual run rate. Here the merchant usually sees nothing different: a card is swiped, and the issuer settles with Visa in stablecoins behind the scenes.

Stripe has also offered stablecoin financial accounts in 101 countries, letting businesses hold balances in stablecoins and send them out. That is aimed at the treasury use case, particularly for firms outside the US that want dollar access.

The stablecoin leg is the fast part. The work lives at the two ends, where tokens meet bank accounts.

Accounting and tax basics

Accounting is catching up. In the US, the FASB proposed guidance in August 2026 with examples of when certain stablecoins can be treated as cash equivalents under existing definitions, with comments due by 19 November 2026. Until a final standard arrives, treatment can vary, so finance teams should agree an approach with auditors before holding meaningful balances.

Tax reporting is also tightening. US brokers report digital asset sales on the new Form 1099-DA, with gross proceeds reported for sales from 1 January 2025. Final IRS rules include a $10,000 annual de minimis threshold for qualifying stablecoin sales, above which brokers may report in aggregate, and a stablecoin that loses its peg during the year loses that status. Other countries have their own rules, and a business paying contractors in tokens still has payroll and withholding duties. A local adviser is worth the fee.

Risks finance teams should weigh

• Wrong network, lost funds. USDC on Solana and USDC on Ethereum are not interchangeable at the wallet level. A transfer sent on the wrong chain can be unrecoverable.
• Issuer and peg risk. Both major issuers can freeze addresses, and any stablecoin can trade below a dollar in stress.
• Off-ramp dependence. A supplier who cannot convert locally has been paid in an asset they cannot use. Check liquidity in the receiving country first.
• Compliance. Sanctions screening, know-your-customer checks and record keeping apply as they do to wires. In the EU, transfers of e-money tokens can count as payment services that need authorisation.
• Regulatory change. Rules under the GENIUS Act, MiCA and Gulf frameworks are still being implemented. Our guide to stablecoin rules under GENIUS and MiCA tracks the main ones.

A sensible way to start

Most businesses that adopt stablecoins begin small. A common pattern is to pick one painful corridor, often a supplier or contractor group in a country where wires take three days, and run a few months of payments through a licensed provider alongside the existing bank route. Compare total cost including conversion spreads, time to funds, and how many payments needed manual follow-up. If the numbers hold, expand. If the off-ramp on the receiving side is thin, the savings on the stablecoin leg will not matter. For where this is heading, see our look at stablecoin payments trends in 2026.

This article is general information about payment methods and does not endorse any provider. It is not legal, tax or financial advice.

At 2 a.m. in Singapore, a transaction monitoring analyst at a regional exchange watches an alert queue refill. The US Treasury has just designated four Iranian exchanges, and the blockchain analytics feed has re-scored thousands of addresses overnight. One customer's deposit of 25,000 USDT now traces back, three hops away, to a wallet tied to a newly sanctioned platform. Does she freeze it, ask questions, or let it through? That decision, multiplied across every regulated crypto business, is what anti-money laundering and sanctions compliance looks like in 2026.

Crypto compliance used to be a story about bitcoin and darknet markets. In 2026 it is a story about stablecoins, state actors and the businesses that sit between them. This piece pulls together the official data, the year's sanctions actions and what compliance teams actually check.

The threat picture in official numbers

The FATF's annual reviews give the clearest regulator's-eye view. Its June 2025 update recorded the theft of USD 1.46 billion from the exchange Bybit by North Korean actors, the largest single virtual asset theft on record, with only 3.8% recovered at the time. It cited one industry estimate of about USD 51 billion in illicit on-chain activity tied to fraud and scams in 2024, and concluded that most on-chain illicit activity now involves stablecoins.

The July 2026 update went further on how the money moves. Competent authorities found that one Cambodia-based conglomerate laundered at least USD 4 billion between August 2021 and January 2025, including at least USD 37 million from North Korean cyber-heists. A stablecoin issuer froze more than USD 29 million in a wallet linked to the same group, which had reportedly marketed its own stablecoin as immune to freezing. In Spain, Operation Borrelli uncovered about EUR 460 million laundered from more than 5,000 victims. The FATF also found that terrorist groups increasingly prefer stablecoins over bitcoin, using rotating wallets and multi-hop transfers.

The common thread is speed. Stablecoins settle across borders in seconds and can be fragmented through decentralised exchanges, bridges and multiple platforms faster than any investigator can follow. That is why compliance has shifted from reviewing transactions after the fact to screening them before they settle. For the consumer side of this, see our guides to money transfer scams and AI fraud in payments.

2026's defining sanctions campaign: Iran's crypto exchanges

The US Treasury's "Economic Fury" campaign against Iran has produced the most intensive run of crypto sanctions to date. On 2 June 2026, OFAC designated Nobitex, Iran's largest exchange, alongside Wallex, Bitpin and Ramzinex, under counterterrorism and Iran financial-sector executive orders. The Treasury press release said Nobitex handled over half of all Iranian digital asset inflows in 2025, Wallex 12% and Bitpin 10%, and that Ramzinex processed more than USD 2.45 billion in transactions. It also said Nobitex helped Iran's central bank access hundreds of millions of dollars in stablecoins. Nobitex claims more than 11 million users, according to Elliptic, whose analysts had earlier traced its wallet links to addresses tied to sanctioned and hacking groups.

More followed. In July, Treasury designated Iran-linked wallets holding more than USD 130 million, which an on-chain analyst identified as four Tron addresses holding about USD 131 million in USDT. August brought designations of two more exchanges, Shelbit and Aban Tether, and on 17 September Treasury sanctioned BitBank, alleging it moved hundreds of millions of dollars in bitcoin to the IRGC, as The Block reported. On 9 October, Treasury Secretary Scott Bessent said the US would probably seize about USD 1 billion more in Iran-linked crypto, while earlier totals remained unclear, per The Block.

Stablecoin issuers as the new chokepoint

Unlike bitcoin, centrally issued stablecoins can be frozen by the issuer at the token level. That has turned issuers into a front line of sanctions enforcement. Tether says it froze more than USD 344 million in USDT across two Tron addresses on 23 April 2026 in coordination with OFAC and US law enforcement. In a September 2026 statement, it put Iran-linked freezes over the past year at about USD 550 million, cumulative freezes at more than USD 5 billion, and said it works with over 340 agencies in 70 countries. It has also aligned its wallet-freezing policy with OFAC's SDN list, including secondary-market wallets. These are the company's own figures.

Congress is not convinced that is enough. A Senate Permanent Subcommittee on Investigations report released on 28 September 2026 analysed 846 sanctioned or blocked Iran-linked wallets and found that 87% of 757 wallets implicated in Iranian terrorism financing had transacted mainly in USDT, and argued the issuer had failed to act on some wallets with strong illicit-finance indicators, The Block reported. Tether rejected the claim. The episode shows where the policy fight is heading: issuers will be judged not only on whether they freeze when asked, but on whether they spot problems first. Our comparison of USDT and USDC covers the two largest issuers' approaches.

A freeze button is only a control if someone presses it before the money moves on.

What compliance teams actually screen

A regulated crypto business runs several overlapping checks on every customer and transfer. The table below maps each to its legal hook and a 2026 example.

Two technical points trip up even experienced teams. First, OFAC's 50% rule concerns ownership: any entity 50% or more owned by blocked persons is itself blocked, even if it is not named. It does not mean funds become sanctioned once half of them pass through a listed address. Second, sanctions liability is strict. Treasury's release reminds firms that OFAC can impose civil penalties without proving intent, and that non-US firms can be exposed if they cause US persons to violate sanctions. Our guide to the FATF Travel Rule explains the data layer that makes counterparty screening possible.

Worked example: one deposit, three decisions

Return to the Singapore analyst and her 25,000 USDT deposit. Suppose the analytics tool breaks it down like this.

1. Direct exposure: zero. The sending address itself is not on any sanctions list. No mandatory block yet.
2. Indirect exposure: 7,500 USDT, or 30%. That portion arrived three hops back from a wallet attributed to an exchange designated by OFAC in June. The other 17,500 USDT came from two regulated exchanges.
3. Customer profile. The account holder is a freelancer who says she was paid by a client through a P2P trader. P2P is exactly the channel 58 of 66 assessing jurisdictions told the FATF they rate as high risk.

A typical risk-based response: hold the deposit, request evidence of the underlying work and payment, and escalate to the money laundering reporting officer. If the explanation does not hold, file a suspicious transaction report with the local financial intelligence unit and consider ending the relationship. If the exchange has US nexus or uses a US-regulated stablecoin, legal will also check whether any portion constitutes blocked property. None of this is automatic. Analytics tools produce a risk score; people still make the call, and regulators examine whether the call was reasonable.

The numbers matter. A tool tuned to flag any indirect exposure above 5% will catch this deposit and thousands of innocent ones. Tuned to 50%, it would have missed it. Every compliance team is choosing a threshold, and the right one depends on its customers, products and regulators.

Where the system is still weak

The FATF's own 2026 data shows the gaps. Only 23% of jurisdictions (31 of 133) collect P2P market data. Just 26 of 142 have assessed DeFi risks, and only one reported enforcement against a DeFi arrangement. Of 91 jurisdictions with Travel Rule laws, 55 had taken no supervisory or enforcement action on them. Criminals route through whichever link is weakest: an unlicensed OTC desk, a lightly supervised offshore platform, a bridge. Our explainer on OTC crypto desks covers one of those channels.

Our view: sanctions enforcement has become far more effective at the stablecoin layer, because issuers and large exchanges can be compelled to act. The weakness is everything that does not touch a regulated chokepoint. Expect regulators to keep pushing obligations outward, to issuers, to analytics on unhosted wallets and to the licensing of offshore platforms, rather than inventing new rules for the core.

What to watch next

• October 2026: whether Treasury follows through on the roughly USD 1 billion Iran-linked seizure Secretary Bessent flagged on 9 October, and how much is held in stablecoins.
• Late 2026: the response of Justice and Treasury to Senator Blumenthal's request for investigations following the Senate subcommittee's 28 September report.
• Mid-2027: the FATF's next targeted update on virtual assets, which will show whether supervisors have started enforcing Travel Rule and licensing obligations.
• 25 October 2027: the UK's FSMA crypto regime takes effect, bringing conduct supervision to firms that today are registered only for AML; see our guide to crypto licences.
• End of 2030: the deadline for the FATF's revised Recommendation 16, which will apply similar originator and beneficiary data rules to all cross-border payments.

This article explains AML and sanctions frameworks for general information and is not legal advice. Firms should take advice from qualified sanctions and AML counsel on specific obligations.

At a public sector bank's treasury desk in Mumbai, a dealer lends to another bank overnight in the call money market. Since October 2023, for a group of nine banks, the cash leg of some of those deals has moved not over the RTGS system but as wholesale digital rupees, settled the same day. Nobody outside the dealing room noticed. That is the point of wholesale CBDC: it changes the plumbing under markets without asking the public to download anything.

What a wholesale CBDC is, and what it is not

Central banks already provide digital money to banks. Reserve balances at the RBI, the ECB or the Swiss National Bank are entries in a central bank ledger, moved through real-time gross settlement systems. So what is new? A wholesale CBDC puts central bank money into a form that can live on a shared or programmable ledger, often alongside tokenised securities. That lets the cash leg and the asset leg of a trade settle together in one step, a pattern known as atomic delivery versus payment, and lets institutions write rules into the settlement itself.

The RBI's October 2022 concept note drew the line clearly: retail CBDC for everyone, wholesale CBDC restricted to select financial institutions for interbank transfers and wholesale settlement. A wholesale CBDC is not a stablecoin and not a tokenised bank deposit; it is a claim on the central bank itself. For the full taxonomy, see stablecoins vs CBDC vs tokenised deposits.

India's e-rupee-W: bonds, then call money, then tokenised CDs

The wholesale pilot began on 1 November 2022, a month before the retail pilot. Nine banks took part: State Bank of India, Bank of Baroda, Union Bank of India, HDFC Bank, ICICI Bank, Kotak Mahindra Bank, Yes Bank, IDFC First Bank and HSBC, settling secondary-market trades in government securities, as FinTech Futures reported at launch. The RBI's stated aim was a more efficient interbank market with lower costs, by reducing the need for settlement guarantee infrastructure or collateral.

The second use case followed in October 2023. According to Business Standard, nine banks, four public and five private, began settling call money trades in e-rupee instead of RTGS, on a same-day basis, with the rest of the operating process unchanged. The roadmap described by people familiar with it ran from bonds to call money to repo and then asset tokenisation.

The tokenisation step arrived in 2025-26. The RBI said in its annual report that it had begun a pilot to issue and trade certificates of deposit in tokenised form on its Unified Markets Interface, with settlement in wholesale CBDC, as OrissaPOST reported. For 2026-27 the bank has said it will explore more tokenisation pilots, widen participants and set up a CBDC and Asset Tokenisation sandbox. We cover the asset side in real-world asset tokenisation.

Worked example: what changes in a call money trade

Take a single overnight loan. Bank A lends ₹200 crore to Bank B at 6.72%, the weighted average call rate on the day Business Standard reported the pilot. Interest for one night is ₹200 crore × 6.72% ÷ 365, or about ₹3.68 lakh. Bank B repays ₹200.0368 crore the next day.

• Through RTGS. The trade is agreed on the dealing system, cleared, and the cash moves between the two banks' reserve accounts at the RBI.
• Through e-rupee-W. The trade is agreed the same way. The cash leg moves as wholesale e-rupee tokens from Bank A's wallet to Bank B's, settling the same day. On maturity, ₹200.0368 crore in tokens moves back.

The interest is identical. The credit risk is identical. What changes is the settlement asset and the ledger it moves on. For a plain overnight loan, the benefit is modest. The value shows up when the cash leg can be tied to something else: a tokenised CD that transfers only if the payment does, a repo where collateral and cash swap in one step, or a cross-border trade where two currencies move together. That is why the RBI's sequence ends with tokenisation rather than call money.

The ₹200 crore trade is illustrative; volumes in the e-rupee-W call money pilot have not been published.

Switzerland's Helvetia: wholesale CBDC in production

The Swiss National Bank has gone furthest toward making wholesale CBDC routine. It has provided wholesale CBDC on SIX Digital Exchange since the end of 2023, when the first tokenised transaction was settled, so that digital bonds traded there can settle against central bank money on a delivery-versus-payment basis. The pilot, known as Helvetia III, was extended by two years in June 2024 and, after a review, by another year to at least mid-2027, according to Finadium.

The 2025 extension also widened the approach. BX Digital, which plans a trading venue for tokenised assets, gets a production connection to the Swiss Interbank Clearing system, so tokenised trades can settle in ordinary central bank money through a bridge rather than as a token. The SNB stresses that none of this is a promise to offer wholesale CBDC permanently. That caution is telling: even the most advanced central bank in this area keeps both routes open.

Cross-border: Mariana, Agora and mBridge

Domestic settlement is the easy half. The bigger prize is cross-border, where a payment today passes through a chain of correspondent banks, as we explain in correspondent banking explained.

Project Mariana, run by the BIS with the Banque de France, the Monetary Authority of Singapore and the Swiss National Bank, tested trading and settling hypothetical euro, Singapore dollar and Swiss franc wholesale CBDCs on a public blockchain using an automated market maker. Its final report in September 2023 found the approach feasible, while stressing that the work was experimental.

Project Agorá is the more consequential effort. Launched in April 2024 with eight central banks, including those of five major reserve currencies, and more than 40 financial institutions convened by the Institute of International Finance, it built a shared programmable platform where tokenised reserves and tokenised commercial bank deposits sit side by side. Its main report came out on 27 May 2026. In July 2026, 28 institutions and central banks ran real-value tests in several currencies, about CHF 800,000 in total across 17 scenarios, with individual payments from CHF 9,000 to CHF 125,000.

Project mBridge, a multi-CBDC platform built by a group of central banks, shows what scale looks like. By the end of November 2025 it had processed 4,047 cross-border payments worth a cumulative 387.2 billion yuan, with the digital yuan accounting for about 95.3% of transaction value, according to BeInCrypto figures republished by Yahoo Finance. That concentration is the story: a multi-currency platform that runs mainly on one currency. See our mBridge explainer.

Europe's Pontes: the bridge went live in September 2026

The Eurosystem took a different route. Instead of issuing a wholesale CBDC token, it connected DLT platforms to its existing settlement system. On 21 September 2026 it launched Pontes, which lets wholesale transactions in tokenised assets settle in central bank money, according to the ECB press release. Four market DLT operators were named: Axiology, Cashlink, Clearstream and SWIAT. Thirteen market participants plus the Bundesbank had onboarded at launch.

The ECB's own timeline sets out what comes next: settlement finality, support for all T2 currencies and T2-aligned operating hours in 2027, then overnight liquidity for tokenised central bank money and a stated aim of 24/7 opening in the second half of 2028. Alongside sits Appia, a longer-term roadmap for a tokenised European financial system run with Danmarks Nationalbank, which aims for a blueprint by 2028. This followed exploratory trials from May to November 2024 involving 64 market participants and more than 50 experiments, according to Payment Expert.

The projects compared

Why central banks start with wholesale

Look at where the real money is moving and a pattern stands out. Wholesale projects are live or near-live in India, Switzerland and the euro area. Retail CBDC is still in pilot in India and still in legislation in Europe. That ordering is deliberate.

First, wholesale CBDC avoids the hardest political question in retail CBDC: what happens to bank deposits. In a non-binding technical analysis reported by MNI in October 2025, ECB staff estimated that in a severe stress case where every user held the maximum €3,000 in a retail digital euro, outflows could reach €699 billion, about 8.2% of retail sight deposits. A wholesale CBDC sits among institutions that already hold central bank money, so there is no such question.

Second, tokenised markets are coming whether or not central banks join. If tokenised bonds settle in stablecoins or tokenised deposits only, the safest settlement asset, central bank money, is absent from the newest markets. Pontes and Helvetia are, at heart, defensive moves to keep central bank money as the anchor.

Third, wholesale is where programmability is uncontroversial. Nobody worries about civil liberties when a bond and its payment swap atomically. The same feature in a citizen's wallet raises questions we explore in the CBDC privacy debate.

Wholesale CBDC is the part of the CBDC story where central banks are protecting their role rather than reinventing the citizen's wallet.

What to watch next

• 2026-27, India: the RBI's plans for more tokenisation pilots and the CAT sandbox; the first published volumes for e-rupee-W would be a useful signal.
• 2027, euro area: Pontes adds settlement finality, multi-currency T2 support and 22.5-hours-a-day, five-days-a-week operation.
• Mid-2027, Switzerland: the current end date for Helvetia III; watch whether the SNB extends again or commits to a permanent service.
• Second half of 2028: Pontes enhanced products and Appia's blueprint for an integrated DLT-based financial system.
• Agorá's next phase: the BIS says testing continues with a larger role for the private sector.

For the retail side of the same story, see our CBDC tracker.

A knitwear exporter in Tiruppur ships 4,000 sweaters to a buyer in Rotterdam. The invoice is in euros, the mill's costs are in rupees, and the cash will pass through at least two companies the exporter has never heard of before it lands. One collects the euros, another converts them, a third pays out on a local rail in India. That invisible chain is the cross-border payments industry for businesses. In 2026 it is crowded, profitable and changing quickly, as card networks, fintechs and stablecoin firms all claim the same pipes.

How we chose these companies

We selected companies by scale and reach as of October 2026: reported revenue or payment volume, the number of countries and currencies they pay into, licences held, and recent funding. We grouped them by what they do rather than numbering them, because a payout network and an SME account provider are not substitutes. Figures are dated and linked. Nothing here is a recommendation.

The market they share is enormous. FXC Intelligence estimated the total cross-border payments market at $208 trillion in 2025, with B2B flows from small and mid-sized firms generating a revenue pool of about $191 billion, the largest of any retail segment. The same research found take rates fell in every retail segment between 2022 and 2025. Growth is real; margins are thinning.

Who are the leading cross-border payment companies in 2026?

The leading cross-border payment companies for businesses in 2026 fall into four groups: SME account and FX platforms (Airwallex, Payoneer, Ebury, Corpay), payout networks sold to other firms (Thunes, Nium, Wise Platform, Rapyd), emerging-market specialists (dLocal), and network giants (Visa Direct, Mastercard Move, Stripe with Bridge, Ripple). Airwallex, dLocal and Corpay grew fastest by revenue.

Which platforms serve small and mid-sized businesses?

For small and mid-sized businesses, the largest cross-border platforms in 2026 are Airwallex, Payoneer, Ebury and Corpay. They offer multi-currency accounts, FX conversion and supplier payments. Airwallex is the fastest-growing, Payoneer is the most established with online sellers and freelancers, and Corpay is the largest by revenue after buying Alpha Group.

Airwallex

Airwallex raised $320 million in a Series H in June 2026 at an $11 billion valuation, up from $8 billion six months earlier. It reported annualised revenue of $1.3 billion as of March 2026, up 74%, more than 676,000 business customers and over 85 licences, per FinTech Futures. It is co-headquartered in San Francisco and Singapore.

Payoneer

Payoneer, listed on Nasdaq, passed $1 billion of revenue in 2025, reaching $1.05 billion on $87.5 billion of volume, according to its full-year results. Its B2B small business revenue grew 28% to $237 million. Interest on customer balances is still a large part of the total: revenue excluding interest was $821 million.

Ebury

Ebury, the international payments firm backed by Santander, secured about £550 million of funding in April 2026 led by Centerbridge Partners. It serves more than 27,000 businesses from 30 regulated markets, paying into 160 countries in more than 140 currencies, according to Santander, which says revenue has grown more than 30% a year since 2020.

Corpay

Corpay's corporate payments segment, which includes its cross-border FX business, earned $1.64 billion of revenue in 2025, up 34%, out of group revenue of $4.53 billion, per its annual results. It closed the purchase of Alpha Group in the fourth quarter, its largest cross-border deal so far. Since April 2025 it has also been Mastercard's exclusive provider of large-ticket cross-border payments to Mastercard's bank clients.

Which companies run the payout networks other firms use?

The payout networks behind many apps in 2026 are Thunes, Nium, Wise Platform and Rapyd. They connect to local bank systems, mobile wallets and cards in well over 100 countries and sell that reach to banks, remitters and platforms. Thunes and Nium have added stablecoin settlement in the past year.

Thunes

Thunes, which split from TransferTo in 2019, says its network reaches more than 140 countries and 90 currencies, with direct connections to over 12 billion bank accounts and wallets. In March 2026 it let the 11,500 institutions on Swift send payouts to stablecoin wallets using their existing connections, following a pay-to-stablecoin-wallet service launched in October 2025, according to its announcement. These are company figures.

Nium

Nium pays out in more than 190 countries and 100 currencies, with real-time delivery in over 100 of them, and holds licences in more than 40 countries, according to a May 2026 Circle announcement connecting USDC settlement to Nium's payouts. Its last disclosed valuation was $1.4 billion, after a $50 million round from Brunei's sovereign fund in June 2024, a drop from its peak, DealStreetAsia reported; we found no newer valuation as of October 2026.

Wise Platform

Wise sells its transfer infrastructure to banks and businesses. Morgan Stanley uses it for corporate clients' international settlements and Standard Chartered uses it for its SC Remit service, according to Finance Magnates. Wise as a whole moved $243.5 billion across borders in its year to March 2026 at a 0.52% take rate.

Rapyd

Rapyd sells a single API for local payment methods, cards and bank transfers. In March 2025 it completed its purchase of PayU's European-based payments business from Prosus for about $610 million, funded largely by a $500 million raise, and said it would then operate in more than 100 countries with licences in 41, The Paypers reported. A month earlier, Bloomberg reported that Rapyd was in talks to raise money at a valuation of about $3.5 billion, down from $9 billion in 2021; Rapyd declined to comment. It is private and does not publish revenue. It is one to watch because it is trying to combine acceptance and payouts in one contract, which most rivals still sell separately.

Who specialises in emerging markets?

In 2026 dLocal is the largest listed specialist in emerging-market payments, processing $40.8 billion in 2025 for global merchants that need to collect and pay out in Latin America, Africa, the Middle East and Asia. It earns more per dollar than developed-market rivals because local methods and FX are harder to handle.

dLocal

dLocal's total payment volume rose 60% to $40.8 billion in 2025 and revenue grew 47% to $1.09 billion, according to its fourth-quarter filing. Its business is helping global merchants collect and pay out locally without building that capability country by country.

How are Visa, Mastercard, Stripe and Ripple competing?

Visa, Mastercard, Stripe and Ripple compete on network reach rather than service to individual businesses. Visa Direct handled 12.6 billion transactions in fiscal 2025, Mastercard Move reaches over 95% of the world's banked population, Stripe processed $1.9 trillion in 2025, and Ripple has processed more than $95 billion. All four are building stablecoin settlement.

Visa Direct

Visa Direct processed 12.6 billion transactions in fiscal 2025, up 27%, according to Visa's fourth-quarter earnings call. Visa says its wider network reaches about 12 billion endpoints across cards, accounts and wallets. In September 2025 it announced a pilot letting banks and remitters prefund Visa Direct with stablecoins.

Mastercard Move

Mastercard Move pays out in more than 200 countries and territories and 150-plus currencies, reaching over 95% of the world's banked population through accounts, wallets, cards and cash, according to Mastercard. Its partnership with Corpay added near real-time payments to 22 more markets in September 2025.

Stripe and Bridge

Stripe processed $1.9 trillion of total payment volume in 2025, up 34%, and was valued at $159 billion in a February 2026 tender offer, per its annual letter. It estimated stablecoin payments roughly doubled to about $400 billion in 2025, around 60% of it B2B, and said volume through Bridge, its stablecoin unit, more than quadrupled. That estimate sits close to the roughly $390 billion of genuine stablecoin payments McKinsey and Artemis counted for 2025.

Ripple

Ripple raised $500 million at a $40 billion valuation in November 2025 and said it had processed more than $95 billion in payments and held 75 licences, per FinTech Futures. Its dollar stablecoin RLUSD passed a $1 billion market cap within its first year, and it bought treasury software firm GTreasury and prime broker Hidden Road. More detail is in Ripple in 2026.

Worked example: what a $50,000 receivable costs

Return to the Tiruppur exporter, now collecting a $50,000 payment. Published price lists vary by plan and corridor, so a fairer comparison uses what each company actually earns per dollar, from its own figures.

• Wise: reported take rate of 0.52% would be about $260.
• Payoneer: $1,052.8 million of revenue on $87.5 billion of volume is about 1.20%, or $602. Excluding interest income, it is 0.94%, or about $469.
• dLocal: $1,093.6 million on $40.8 billion is about 2.68%, or $1,340, reflecting harder emerging-market rails and pay-in work.

These are blended averages across very different clients, not quotes. They show the shape of the market: providers serving simpler, high-volume corridors earn fractions of a percent, while those handling local collections in difficult markets earn several times more. An Indian exporter should also factor in the conversion margin on rupee payout and the documentation rules covered in our guide to Indian exporters' payment rules. For the business case, see cross-border B2B payments in 2026.

Reach is now cheap to claim. The real difference is who holds the licence, the liquidity and the local bank link in the last mile.

What about cross-border payments between India and the UAE?

India and the UAE signed a local currency settlement agreement in July 2023 so trade can be invoiced and settled in rupees and dirhams without converting through dollars. Uptake has been modest, according to a July 2025 Khaleej Times analysis, so most business flows still pass through banks and the platforms above. The UAE is also India's second-largest source of personal remittances, at 19.2% in RBI's 2023-24 survey. The rails used for the corridor are compared in what are payment rails.

What to watch next

• Stablecoin prefunding: Visa's Direct pilot, Thunes's Swift link and Nium's USDC settlement all went live between September 2025 and May 2026. Whether banks use them for real volume by mid-2027 will show if stablecoins reduce trapped liquidity in prefunded accounts. Our B2B stablecoin settlement case studies follow early users.
• Airwallex's run rate: its $1.3 billion annualised revenue in March 2026 already tops Payoneer's 2025 total. Its next disclosed figure will show whether 74% growth holds at that size.
• Corpay's integration of Alpha Group through 2026, which will test whether a card-heavy company can run a large FX book.
• Ripple's acquisitions: folding GTreasury into its payments product would put a stablecoin firm directly inside corporate treasury software.

This list is editorial research, not investment advice or an endorsement. Company figures are as reported by the linked sources and periods differ between companies.

The controller of a 40-person HVAC contractor in Phoenix closes her books on the third business day of the month. Her corporate cards come from Ramp, her customers pay invoices through a Stripe checkout, her bank feeds arrive through a Plaid connection, and half her technicians are paid into Chime accounts. She has never thought of herself as a fintech customer. That is the clearest sign of how far American fintech has travelled: in 2026 the leading firms are no longer apps people try, they are the plumbing of ordinary businesses and households, and the numbers they now report look like those of banks.

How did we choose the leading US fintech companies for 2026?

We selected US fintech companies for 2026 on measurable criteria as of October 2026: scale (revenue, payment volume or customers in the latest reported period), growth rate, licences held or pending, funding or market value, and product breadth. We grouped them by category rather than ranking them one to thirteen, because a payments processor and a consumer bank app cannot be compared on one number.

Every figure below is dated and linked. Where a private company does not publish revenue, we use the latest valuation or customer count that was reported. Rippling, the payroll and HR platform that has pushed into corporate cards and bill pay, is a credible name for this list, but we could not verify a 2026 figure, so it is mentioned here rather than profiled.

The money backdrop matters. US fintech attracted $80.8 billion across 933 deals in the first half of 2026, most of the $86.9 billion recorded for the Americas, according to KPMG's Pulse of Fintech as reported by InvestmentNews. Mergers drove the total, led by FIS buying Worldpay for $24.3 billion. Venture money is concentrating in fewer, larger winners, which is why the same dozen names keep appearing.

Which US payments and infrastructure companies lead in 2026?

By volume and valuation, Stripe is the largest private US fintech in 2026, processing $1.9 trillion in 2025 and valued at $159 billion in February 2026. Plaid leads bank-account connectivity with an $8 billion valuation set the same month, and Circle, issuer of USDC, is the largest US-listed stablecoin company, with $73.3 billion of USDC outstanding at the end of June 2026.

Stripe

Stripe's annual letter put business volume on its platform at $1.9 trillion in 2025, up 34%, and a tender offer for staff shares valued the company at $159 billion, against roughly $100 billion in late 2025, as its February 2026 annual letter and The Block reported. Stripe does not publish revenue; the letter said it remained robustly profitable and that its Revenue suite of billing tools was on track for a $1 billion annual run rate in 2026. Stablecoin payment volume across the market roughly doubled in 2025 to about $400 billion, by Stripe's estimate. Stripe has also bought its way into stablecoins: it acquired Bridge for $1.1 billion, and Bridge won conditional OCC approval for a national bank charter in February 2026.

Plaid

Plaid connects apps to bank accounts for payments, identity and income checks. An employee share sale in February 2026 valued it at $8 billion, up 31% from $6.1 billion in April 2025 but still well below its $13.4 billion peak in 2021, according to TechCrunch. Its chief executive has called an IPO part of the path for the coming years without naming a date.

Circle

Circle listed on the NYSE in 2025 and now reports like a rate-sensitive financial company. In the second quarter of 2026 it booked $701 million of revenue and reserve income, up 7%, with net income of $48 million and USDC circulation of $73.3 billion, up 19%, per Circle's results release. Our analysis of Circle as a public company covers why falling rates squeeze its margin.

Which US consumer fintechs are largest by revenue and customers?

By quarterly revenue in the three months to June 2026, Robinhood ($1.31 billion) and SoFi ($1.22 billion) are the largest US consumer fintechs, followed by Chime ($670 million). By users, Block's Cash App leads with 59 million monthly transacting actives in June 2026. All four were profitable in that quarter, which marks a clear break from their loss-making years.

Robinhood

Robinhood reported record second-quarter 2026 revenue of $1.31 billion, up 32%, and net income of $573 million, which included $129 million of one-off gains. It had 28.4 million funded customers, $369 billion in platform assets and 4.8 million Gold subscribers, according to its July 29, 2026 results. The subscriber line, up 39%, is the quiet engine.

SoFi

SoFi is a chartered bank with a fintech's growth rate. Its second-quarter 2026 GAAP net revenue reached $1.22 billion, up 43%, with net income of $156.6 million, 15.8 million members (up 35%) and $45.5 billion of deposits, per its earnings release. The bank charter lets it fund loans with cheap deposits rather than wholesale money.

Chime

Chime listed on Nasdaq in June 2025 at $27 a share, valuing it at $18.4 billion at debut, according to American Banker. In the second quarter of 2026 revenue rose 27% to $670 million, active members reached 10.4 million and the company posted a second straight quarterly profit of $28 million, as Reuters reported. It lifted full-year growth guidance to 25-26%.

Block (Cash App and Square)

Block's second-quarter 2026 gross profit was $3.17 billion, up 25%. Cash App contributed $1.97 billion, up 31%, from 59 million monthly transacting actives, while Square processed $72.8 billion of gross payment volume, up 13%, according to its shareholder letter filed with the SEC. It raised full-year gross profit guidance to $12.51 billion.

Which buy now, pay later firms lead the US market in 2026?

Affirm is the largest US-headquartered buy now, pay later lender by volume in 2026, with $50.2 billion of gross merchandise volume in the fiscal year to June 2026. Klarna, Swedish-founded but listed on the NYSE since September 2025, is larger globally at $36.6 billion of GMV in one quarter, and its US volume is growing faster than the rest of its business.

Affirm

Affirm's fiscal fourth quarter of 2026 delivered GMV of $14.1 billion, up 36%, revenue of about $1.17 billion and 27.8 million active consumers, according to Crowdfund Insider's summary of the August 27 results. Management guided to more than $64 billion of GMV in fiscal 2027.

Klarna

Klarna priced its US IPO at $40 on September 10, 2025, at a $15 billion valuation. In the second quarter of 2026 it reported revenue of $1,042 million, up 27%, GMV of $36.6 billion, US GMV of $7.9 billion (up 27%), 120 million active consumers and a $9 million profit, per its filing with the SEC.

Which business finance fintechs are growing fastest?

Ramp is the fastest-growing US business finance fintech by valuation in 2026, reportedly valued at about $60 billion in October 2026, with Bloomberg putting annualized revenue above $1.5 billion by June. Brex left the independent field when Capital One completed its $5.15 billion purchase in April 2026, and Mercury, the startup bank, is moving toward its own national charter.

Ramp

Bloomberg reported on October 9, 2026 that Ramp was raising about $1.85 billion at a roughly $60 billion pre-money valuation, co-led by Thrive Capital and Dragoneer, as FinTech Futures summarised; Ramp had not confirmed it. In June the company closed a $750 million Series F at $44 billion, saying annualized revenue had passed $1 billion and it served more than 70,000 businesses. Bloomberg has since reported the run rate passed $1.5 billion by June. For context on the product category, see our explainer on corporate expense management cards.

Brex

Brex agreed in January 2026 to sell to Capital One for $5.15 billion in cash and stock, and the deal closed on April 7, 2026, according to Capital One. It is now a bank-owned platform, which tells you how incumbents prefer to buy rather than build in this segment.

Mercury

Mercury raised a $200 million Series D at a $5.2 billion valuation on May 20, 2026, and says it serves more than 300,000 businesses. It applied for an OCC national bank charter in December 2025 and received conditional approval in late April 2026, according to PYMNTS.

How do the leading US fintech companies compare?

Compared on their latest dated metric as of October 2026, the US leaders split into three types: public companies reporting quarterly revenue (Robinhood, SoFi, Block, Affirm, Klarna, Chime, Circle), private companies valued through tender offers or rounds (Stripe, Ramp, Plaid, Mercury), and one acquired platform (Brex). The table below lines them up.

Metrics use different bases (revenue, gross profit, GMV, valuation) and are not directly comparable across rows.

What are investors paying for US fintech revenue? A worked example

In 2026 investors are paying about 40 times annualized revenue for Ramp, and about 8 cents of valuation for every dollar Stripe processed in 2025, far richer terms than listed peers get. Stripe does not disclose revenue, so volume is the only clean yardstick. Those prices assume years of growth, not current earnings.

Here is the arithmetic. Stripe: $159 billion divided by $1.9 trillion of 2025 volume is about 8.4 cents of valuation per dollar processed in a year. Ramp: $60 billion divided by the $1.5 billion run rate Bloomberg reported gives 40 times; on the $1 billion Ramp itself cited in June, it would be 60 times. Either way, the figure excludes the new $1.85 billion of cash.

Take rates explain why a dollar of volume is worth such different amounts. Affirm's fourth-quarter revenue of $1.17 billion on $14.1 billion of GMV works out to about 8.3%, because it earns interest on the loans it holds. Klarna's $1,042 million on $36.6 billion is about 2.8%, because most of its volume is short, interest-free pay-in-four. Applied to guidance, Affirm's outlook of $64 billion of fiscal 2027 GMV at an 8.5% take rate implies roughly $5.4 billion of revenue.

A dollar of payment volume is worth eight cents to one lender and under three to another. Valuation follows the cents, not the dollars.

None of this is a view on whether these prices are right. It shows what has to be true: Ramp has to keep compounding at startup speed for years to grow into a 40 times multiple, while the listed companies are already being judged on profit.

What should you watch in US fintech next?

The next six months will test whether US fintech's 2026 profitability holds as interest rates fall and stablecoin rules take effect. Key dated events are SoFi's third-quarter results on October 27, 2026, Block's on November 3, 2026, and rulemaking under the GENIUS Act that will decide how Bridge, Circle and bank issuers compete.

• October 27, 2026: SoFi reports third-quarter results, a read on whether deposit growth keeps funding its lending.
• November 3, 2026: Block reports, with guidance of $3.13 billion in third-quarter gross profit.
• Charters: Bridge and Mercury both hold conditional OCC approvals. Final approval would put fintech-owned national banks in direct competition with regional banks. Our guide to GENIUS Act implementation tracks the stablecoin side.
• IPO queue: Stripe, Ramp and Plaid all have the scale to list, and Plaid has said an IPO is on its path. Any filing would reset private valuations across the sector.
• Instant rails: As FedNow and RTP spread, account-to-account payments will pressure card-based business models over time.

For how these names stack up against other markets, see our companion lists on leading UK fintech companies and Canadian fintech companies.

This list is editorial research, not investment advice or an endorsement. Company figures are as reported by the linked sources and may be restated.

In September 2026 India's Unified Payments Interface processed 24.07 billion transactions in thirty days. Abroad, the same system handled roughly 755,000 cross-border payments in the whole of the 2024-25 financial year. That gap is the real story of UPI's global ambitions, and it says more about how payment systems travel than any list of countries ever could.

The numbers at home

Start with scale, because nothing else in retail payments looks like it. NPCI data shows UPI handled 24.07 billion transactions worth Rs 29.37 lakh crore in September 2026, according to a monthly breakdown published by StartupTalky. The headline volume dipped 1.8% from August, but only because September is a day shorter. Daily volume actually climbed to a record 802 million, the first month above 800 million, and volume was up 23% on a year earlier.

For the full 2025-26 financial year, the government told Parliament that UPI processed about 24,161 crore transactions (roughly 241.6 billion) worth Rs 314.23 lakh crore, with about 55.49 crore users onboarded as of June 2026. An IMF fintech note from June 2025 called UPI the largest retail fast-payment system in the world by volume, and an ACI Worldwide report put its share of global real-time transactions at about 49%.

How a UPI payment actually moves

UPI is a switch, not a bank. A customer installs an app (PhonePe, Google Pay, Paytm, a bank's own app) and links a bank account to a virtual payment address such as name@bank. When a shopkeeper in Indore shows a QR code, the app sends an instruction through NPCI's central switch to the customer's bank, which debits the account and credits the merchant's bank in seconds. Settlement between banks happens later in batches, but the money is usable at once.

Two design choices did most of the work. First, the app layer is open, so fintechs could compete on experience while banks held the accounts. Second, person-to-person and small merchant payments were kept free to the user, which made UPI cheaper than cash for a chai stall. The side effect is concentration: in August 2026 PhonePe and Google Pay together carried about 78% of volume, a share regulators have been trying to dilute for years.

Where UPI works abroad

According to a government reply in the Lok Sabha summarised in July 2026, UPI is live in 10 countries through 12 separate links. They fall into two buckets.

• Merchant acceptance (P2M): Bhutan (2021), Singapore (2021), the UAE (2022), France, Mauritius and Sri Lanka (2024), Nepal (2024), Qatar (September 2025) and Cambodia (June 2026). An Indian tourist scans a local QR code and pays from an Indian bank account.
• Person-to-person transfers (P2P): Singapore's PayNow link (February 2023), Nepal (June 2026) and Greece (May 2026).

The Singapore-PayNow linkage was the first fast-payment-to-fast-payment link of its kind. At launch, the Reserve Bank of India capped outward transfers at Rs 60,000 a day, with a handful of banks on each side. In France, NIPL partnered with payments firm Lyra in February 2024, and by June 2026 the Eiffel Tower was accepting UPI for tickets, with Paris and Nice airports next on the list, MediaNama reported.

UPI's export is less about Indians paying abroad and more about other countries borrowing the blueprint.

Why the overseas numbers stay small

MediaNama's figures show cross-border UPI transactions rising from 180 in 2021-22 to more than 755,000 in 2024-25, worth about Rs 258 crore. That is impressive growth from nothing, and still a rounding error next to a single hour of domestic traffic.

The reasons are structural. Most of the merchant links serve Indian tourists, which caps the market at outbound travel. An engineer in Calgary or a hotel cook in Muscat who wants to pay a local bill in local currency gets no help from a QR code that debits a rupee account. And the P2P corridors that would matter for remittances are the hardest to build, because they need two central banks to agree on limits, foreign exchange handling, compliance checks and who carries settlement risk. The PayNow link took years of negotiation for a single corridor.

There is also a fee problem in reverse. UPI is free at home because the government subsidises it and merchants pay little or nothing. Abroad, someone has to pay the acquirer, the currency conversion and the network. Those costs reappear, and UPI loses the price edge that made it famous.

The quieter strategy: exporting the rails

The more interesting work at NPCI International Payments Limited is not acceptance at shops. It is selling the design. In May 2024 NIPL signed with the Bank of Namibia to build a UPI-style system, and in June 2024 it signed a similar pact with Peru's central bank, Business Standard reported. Trinidad and Tobago followed later that year, and an agreement with Uzbekistan's payments operator was signed in August 2026.

This is a smarter long game. A country that runs a UPI-like switch is far easier to link with later, because message formats, addressing and dispute rules already rhyme. It mirrors what Brazil did with Pix domestically and what the BIS-backed Project Nexus is trying to do multilaterally: connect fast-payment systems once through a common hub rather than negotiating corridor by corridor.

What would make UPI matter globally

In our view, three things would move the overseas numbers from thousands to millions.

1. Remittance corridors with the Gulf. The UAE, Saudi Arabia and Qatar send enormous sums to India. A P2P link that lets a worker in Doha push money straight into a family's UPI account, priced near the mid-market rate, would be the first overseas use case with genuine volume.
2. Non-resident access. Letting overseas Indians use international mobile numbers on NRE and NRO accounts was a start. Extending that to more countries and more banks widens the base.
3. Joining multilateral hubs. Bilateral links do not scale. Plugging into a shared hub would put UPI one connection away from many systems at once.

The bottom line

UPI is a domestic phenomenon with a growing passport. The tourist QR codes in Paris and Phnom Penh make for good headlines, but the bigger contribution may be the countries that build their own UPI-shaped systems. For anyone comparing it to India's central bank money experiments, our explainer on the e-rupee versus UPI covers why the two are not rivals.

Domestic figures are NPCI data as reported for September 2026. International counts reflect a government statement from July 2026 and may have changed since.

A junior doctor in Leeds sends £400 on the first of every month to cover her brother's education loan instalment in Pune. For two years she used her high-street bank because it felt safe. Then she compared the rupees that actually arrived with the rate quoted on a search engine that morning, and realised the gap between the two was her real fee. Her story is common in a corridor that is now among the cheapest in the world, but only for people who know where to look.

A corridor that has grown in weight

The UK matters far more to India's inflows than it did five years ago. The RBI's remittances survey shows the UK's share rising from 6.8% in 2020-21 to 10.8% in 2023-24, making it India's third-largest source after the US and the UAE. On that year's total of about $118.7 billion, that implies roughly $12.8 billion. The 2021 census counted about 920,000 India-born residents in England and Wales, the largest foreign-born group.

The UK's own numbers are fuzzier. The Migration Observatory at Oxford notes that World Bank estimates put outward remittances at about £9.3 billion in 2023, a figure it considers a significant undercount, and that India is the top destination.

What it costs in 2026

The World Bank's corridor data for the third quarter of 2025 put the average cost of sending £120 (about $200) from the UK to India at 1.91%, made of an average fee of about £1.58 and an exchange margin of 0.59%. That is far below the 5.20% average for UK senders across all corridors in early 2025, and already under the UN's 3% target. The cheapest bank option in the sample cost 0.15% and the cheapest money transfer operator 0.49%. The most expensive listings cost more than 3%.

Low averages do not make every transfer cheap. When the cost is a percentage margin, it scales with the amount, so the gap between a good and a bad provider on a house deposit or a wedding is much bigger than on a monthly £120. See how the UK compares with other sending countries in the cost of sending money home in 2026.

The mid-market rate and where the margin hides

The mid-market rate is the midpoint between the wholesale buy and sell prices for a currency, the number search engines and data sites show. Providers either convert near that rate and charge a visible fee, or quote a weaker rate and advertise low or zero fees. Both are legal. The second simply buries the cost in the rupees delivered. The test is easy: multiply your pounds by the mid-market rate at the moment you send and compare it with what the provider promises to deliver. We explain the mechanics in zero-fee transfers and the mid-market rate.

The rate has been unusually volatile in 2026. The rupee weakened sharply against the dollar in the spring, trading around 95 to 96 per dollar by June according to Khaleej Times. That makes each pound worth more at home, which recipients welcome. It also makes it easier for a fat margin to go unnoticed while everyone is pleased with the headline number.

Who you are dealing with: FCA-regulated provider types

Three kinds of firm move most money from Britain to India:

• Banks and building societies, whose deposits carry FSCS protection, though international transfers are often priced mainly through the exchange rate.
• Authorised payment institutions, the licensed money transfer operators, which move funds on your behalf without taking deposits.
• E-money institutions, which hold balances in accounts or wallets and often offer multi-currency features.

For the second and third groups the key protection is safeguarding, not deposit insurance: customer money must be kept apart from the firm's own funds so it can be returned if the firm fails. The FCA found that payment firms which went insolvent between 2018 and 2023 had average shortfalls of 65% of customer funds. Its tighter safeguarding rules, in force since 7 May 2026, add daily checks, monthly reporting and annual audits. Check the firm on the FCA register, and do not leave large balances sitting in a wallet longer than you need to.

A cheap corridor rewards the people who check, and quietly charges the people who assume.

Faster Payments: the domestic leg

Much of the speed in this corridor starts on the UK side. Faster Payments, launched in 2008, moves money between UK accounts in seconds around the clock, with a scheme limit of £1 million per payment, though individual banks set lower limits. When you fund a transfer by bank payment, the provider usually receives it almost immediately, and many transfers to Indian bank accounts now land within minutes once the Indian side credits them. Delays, when they happen, are usually fraud or compliance checks on a new payee rather than the rails themselves.

HMRC basics for NRIs and British Indians

HMRC's guidance on foreign income ties UK tax to your residence status and the type of income you have, not to the act of moving money abroad. If you are UK resident you are normally taxed on worldwide income, including rent from a flat in Bengaluru or interest on an Indian NRO account, usually reported through Self Assessment.

The rules changed on 6 April 2025, when the old non-dom regime ended. It was replaced by a residence-based system that gives 100% relief on eligible foreign income and gains for the first four years of UK tax residence, for people who were not UK resident in the previous ten years. Where India has already taxed the income, for instance through 30% TDS on NRO interest, double taxation relief may reduce the UK bill. India's Liberalised Remittance Scheme governs money leaving India, so it plays no part in transfers from the UK into India. If you own Indian property or investments, an adviser who understands both systems is usually worth the fee.

What British Indians send money for

Typical uses are steady from household to household: parents' running costs, education loan instalments, insurance premiums, property purchases and upkeep, and weddings. Many of these are bills with a fixed payee in India. Paying them directly avoids sending money twice. NRIs with UK mobile numbers can link UPI to NRE or NRO accounts, and NPCI said in 2023 that the UK would be next for Bharat Connect's cross-border bill service. Our guide to paying bills in India as an NRI sets out the routes. Whichever you choose, treat any social media agent promising rates well above the mid-market as a red flag.

This article is general information, not tax or financial advice.

The plastic looks the same everywhere, and the merchant in Dubai, Manchester or Austin is paid in local currency either way. What changes from country to country is everything behind the swipe: who may hold your crypto, who may convert it, and which regulator answers when something breaks.

How much people actually spend

Crypto cards have grown from a curiosity into a measurable channel. Data firm Artemis estimated on-chain card spending at about $100 million a month in January 2023 and around $1.5 billion a month by August 2025, an annualized pace above $18 billion by late 2025, as reported by Crowdfund Insider. Visa carries more than 90% of that on-chain card volume. Visa itself says more than 160 stablecoin-linked card programs were live as of its fiscal second quarter, with spend on them up nearly 200% from a year earlier.

Estimates differ by method. McKinsey counted about $4.5 billion of card-linked stablecoin payments in 2025, a much narrower definition. Either way, cards have become one of the main places stablecoins touch everyday commerce. Our explainer on how crypto cards work covers the mechanics.

United Arab Emirates

The UAE splits oversight. The Central Bank (CBUAE) regulates payment tokens under its Payment Token Services Regulation, published in June 2024. Under that framework, licensed dirham-backed tokens are the ones intended for paying for goods and services onshore. Foreign stablecoins must be registered, and the registration regime is designed so they are not used directly for local retail payments, though holding them for investment is not restricted, according to a regional summary of the regulation. Dubai's Virtual Assets Regulatory Authority (VARA) licenses most crypto service providers in the emirate, and the free zones have their own regulators.

For a nurse in Dubai using a crypto card, the practical effect is that the card converts her balance into dirhams before the merchant is paid. The questions to ask are whether the provider holding her balance is licensed by VARA or another UAE regulator, and what conversion spread it charges.

United Kingdom

The UK is between regimes. Today the Financial Conduct Authority mainly supervises crypto firms through anti-money-laundering registration and rules on financial promotions. A full authorisation regime covering stablecoin issuance, custody, trading platforms and dealing has now been finalised, with the FCA's application gateway opening on 30 September 2026 and the regime expected to apply from 25 October 2027, according to Lewis Silkin. Until then, UK users of crypto cards rely heavily on the card issuer's own e-money or payment licence and the provider's AML registration.

United States

US crypto cards are usually issued by a bank under a Visa or Mastercard program, while the company holding customers' crypto typically needs state money transmitter licences. The GENIUS Act, signed in July 2025, creates a federal framework for payment stablecoin issuers. It takes effect on the earlier of 18 January 2027 or 120 days after final regulations, and federal agencies were still publishing proposals through mid-2026.

Tax is the issue most US users miss. Spending crypto is generally a disposal. Brokers now report sales on Form 1099-DA, and final IRS rules allow aggregate reporting for qualifying stablecoin sales above a $10,000 annual threshold. Spending a stablecoin that stays at a dollar usually creates little gain, but spending bitcoin or ether through a card can.

European Union

MiCA gives the EU the most complete rulebook. Stablecoins pegged to a single currency are e-money tokens, and providers need MiCA authorisation. The European Banking Authority has also said that transferring e-money tokens can be a payment service requiring authorisation under payment services law, with a no-action transition that ended on 1 March 2026. MiCA's own transition for existing crypto firms ended on 1 July 2026. For cardholders, that means EU providers should now be able to point to both kinds of authorisation where they apply. Our guide to GENIUS Act and MiCA rules has more detail.

Canada and India

Canada. The Stablecoin Act arrived through Bill C-15, which received Royal Assent on 26 March 2026, according to DLA Piper. The Bank of Canada supervises issuers, who must hold full reserves, redeem at par and may not pay yield. The framework is expected to come into force in 2027 once regulations are written. A freelancer in Vancouver using a crypto card today is still dealing mostly with provincial securities rules for the platform and the issuing bank's rules for the card.

India. India taxes income from virtual digital assets at a flat 30% with 1% tax deducted at source on transfers, and Budget 2026 left both unchanged while adding penalties for failing to file crypto transaction statements, effective from 1 April. A resident using any crypto-linked card should treat each spend of crypto as a taxable transfer and keep records. We take no view on India's broader crypto policy.

Ask who holds your money before you ask what the card costs. The answer decides which regulator you can turn to.

A short checklist before applying

• Licence. Find the name of the regulator on the provider's website and check it on the regulator's public register.
• Custody. Confirm whether the card spends from a custodial account held by the provider or from a wallet the user controls.
• Conversion point. Ask when crypto becomes local currency: at top-up, or at the moment of each purchase.
• Residency. Many programs are limited to residents of specific countries, and moving abroad can mean the card is closed.
• Records. Make sure transaction history can be exported for tax filing.

These questions are dull, but they are the ones that matter when a provider pauses withdrawals or a regulator changes the rules.

Fees to check in any country

Cashback paid in a volatile token can look generous and lose value before it is used. A card with no monthly fee but a wide conversion spread can easily cost more than one with a small subscription, so compare the total cost of a typical month of spending rather than any single headline fee. For travel and freelance use cases, see crypto cards for travellers and freelancers, and for Gulf and Asian rules, stablecoin regulation in Asia and the Gulf.

Regulatory summaries are simplified and current as of October 2026. This article does not recommend any card, exchange or provider, and is not legal, tax or financial advice.

On Friday 9 October 2026 a leather goods exporter in Kanpur with a $200,000 receivable due in December watched the rupee close at 96.73 to the dollar. By the time the market opened on Monday it was 95.75. Nothing about his order, his buyer or his margins had changed, yet the rupee value of that one invoice had fallen by roughly Rs 1.96 lakh over a weekend, because the Reserve Bank of India had rewritten the rules for currency derivatives on Saturday.

That weekend is the best argument we know for treating foreign exchange as an operating risk, not a treasury hobby. Small firms usually have a relationship manager, a habit of converting on the day money arrives, and a hope that the rate will be kind. This guide sets out what the tools actually are, what they cost, and how the rules changed in October 2026.

Why a weekend can wipe out a margin

The arithmetic is unforgiving. A garment or leather exporter working on an 8 to 10 percent net margin can lose most of it on a 5 percent currency swing in the wrong direction, and the rupee has moved far more than that. According to a DSIJ report on the October measures, the rupee was down about 6.6 percent in 2026 and had come close to its all-time low of 96.96 the week before the RBI acted. Brent crude above $100 a barrel kept oil importers buying dollars, and on 7 October the RBI raised the repo rate by 25 basis points to 5.50 percent, its first hike since February 2023.

For an exporter, a weaker rupee looks like good news, and often it is. The problem is the round trip. A rupee that falls 6 percent can bounce 1 percent in a morning, as it did on 12 October when it opened 98 paise stronger at 95.75. Anyone who priced a contract assuming 97 and left the receivable open learned that volatility cuts both ways. For an importer paying for machinery, components or crude derivatives in dollars, the year has simply been expensive.

The market small firms are trading in

Every three years central banks count the world's currency trading. The BIS Triennial Survey published in September 2025 found average daily turnover of $9.6 trillion in April 2025, up 28 percent from 2022. The dollar sat on one side of 89 percent of all trades. The euro accounted for 28.9 percent and the yen 16.8 percent.

Two details in the BIS statistical release matter for smaller businesses. First, outright forwards, the plain-vanilla hedge that exporters use, grew 60 percent to about $1.8 trillion a day, or 19 percent of turnover. Second, non-financial customers, meaning real businesses rather than banks and funds, made up only 5 percent of turnover, down from 7 percent in 2019. The market is enormous and mostly professional. A small firm's pricing from its bank can be far worse than the screen rate.

What the RBI changed on 10 October 2026

India regulates who may hedge and how. Under the January 2024 directions on hedging of foreign exchange risk, effective 5 April 2024, a user could hold up to $100 million of notional hedges across all banks without having to prove the underlying exposure upfront, provided a genuine exposure existed. Retail users had to be shown the mid-market price before booking a derivative, a useful protection that still applies.

The RBI press release of 10 October 2026 tightened this sharply:

• Lower threshold. The no-documentation limit for OTC hedges of contracted exposures fell from $100 million to $5 million, across all banks combined. The same $5 million cap now applies to exchange-traded rupee currency derivatives.
• No rebooking. Once a rupee derivative is cancelled, the bank cannot let the customer rebook it. Rolling over at maturity is still allowed.
• Single-bank undertaking. Users must sign an undertaking that the same exposure is not hedged with another bank.
• A 20 percent reserve. For contracts above $2 million where the customer buys foreign currency against rupees to hedge current account flows, banks must park 20 percent of the rupee notional with the RBI as a Foreign Exchange Risk Reserve.

The asymmetry is the important part. The reserve targets dollar buyers, which means importers. An exporter selling dollars forward is not caught by it. Bankers quoted by DSIJ estimated that hedging costs for affected contracts could rise by 80 to 150 basis points, depending on how much banks pass through. An importer who used to cancel and rebook forwards to chase a better rate has lost that option entirely.

Most small firms do not lose money on currencies because they hedged badly. They lose it because they never decided anything at all.

The toolkit, compared

There are only a handful of real choices. The table sets them out with honest trade-offs.

The research on smaller firms is remarkably consistent: they hedge a portion, not all, and the share falls as the horizon lengthens. A Duke FinReg blog review of central bank data found New Zealand exporters hedged about 39 percent of foreign-currency export transactions on average, with small firms at 33 percent and large firms at 49 percent. Cover was around 50 percent for the next quarter and below 10 percent two years out. That pattern is sensible: you hedge what you know, and you know the near term best.

A worked example with rupee numbers

Take our Kanpur exporter. He has a $200,000 invoice due in 90 days. The figures below use the actual rates from October 2026 for spot and an illustrative forward quote; your bank's quote will differ.

1. Unhedged on Friday. At the 9 October close of 96.73, the receivable is worth Rs 1,93,46,000.
2. Unhedged on Monday. At 95.75 it is worth Rs 1,91,50,000. The weekend cost him Rs 1,96,000, about 1 percent of the invoice.
3. With a forward. Suppose he books a 90-day forward at an illustrative 96.40. He will receive Rs 1,92,80,000 whatever happens. If the rupee strengthens to 93, he is Rs 6.8 lakh better off than unhedged. If it weakens to 99, he gives up Rs 5.2 lakh of upside.
4. With an option. Suppose instead he buys a dollar put with a 95.50 strike for an illustrative premium of 1 percent, Rs 1.93 lakh. If the rupee goes to 93, he exercises and receives Rs 1,91,00,000 less the premium. If it goes to 99, he lets the option lapse, converts at spot for Rs 1,98,00,000, and nets about Rs 1,96,07,000 after the premium.
5. Layered. Forward half, leave half open. His worst case at 93 is about Rs 1,89,40,000 and his best at 99 is about Rs 1,95,40,000. He has narrowed the range of outcomes to roughly Rs 6 lakh instead of Rs 12 lakh.

None of these is right in the abstract. What matters is whether the worst-case number still pays wages and suppliers. If the answer is no, the exposure is too large to leave open. Because he is a dollar seller, the new 20 percent reserve does not apply to his forward, and at $200,000 he is well below the $5 million threshold, though his bank will now ask for the signed undertaking.

Invoicing currency is a strategic decision

The cheapest hedge is never taking the risk. India's RBI software exports survey for 2025-26, released in September 2026, shows how rare that is. Of $221.4 billion in software services exports, 72.6 percent was invoiced in dollars, 9.6 percent in euros, 6.0 percent in sterling and just 6.3 percent in rupees. Goods exporters face the same pull: buyers in dollar markets want to pay in dollars, and economists including the authors of a CEPR study on invoicing patterns have shown that dollar invoicing has risen even as the US share of world trade has fallen.

Rupee invoicing is now better rewarded by regulation. Under India's new export and import regulations that took effect on 1 October 2026, proceeds must be realised within nine months, or 12 months for exports invoiced in rupees, after a September 2026 amendment reversed the earlier 15-month relaxation. We cover those rules in detail in our guide to getting paid as an Indian exporter. Still, a regulatory nudge does not change a buyer's treasury policy. For most small exporters, the practical question is whether some customers, typically in the Gulf or Southeast Asia, will accept a rupee or local-currency price.

Natural hedges and the importer's problem

An exporter who also imports inputs can net the two. A Tiruppur knitwear unit buying synthetic yarn in dollars and selling finished garments in dollars has already offset part of its exposure; a foreign-currency account lets it pay suppliers from export receipts without converting twice. The Duke review notes that German firms are unusually good at this operational matching, and it is the cheapest hedge available.

Pure importers have it hardest in late 2026. They are dollar buyers into a weakening rupee, they now face the 20 percent reserve on hedges over $2 million, and they can no longer cancel and rebook to manage a bad entry point. For them the priorities are simpler: hedge known payables early and in tranches, negotiate longer supplier terms where possible, and ask whether a supplier will accept rupees or a third currency. Freelancers and small service exporters face a smaller version of the same issue, which we explore in our piece on paying and getting paid as a contractor.

Writing a one-page FX policy

A small business does not need a treasury manual. It needs one page, agreed by the owner and the accountant, that answers five questions:

• What is our exposure? List foreign-currency receivables and payables by month for the next six months.
• What is our budget rate? The rate at which each order was priced. Any hedge at or better than this rate protects the margin you already promised yourself.
• How much do we hedge, and when? For example, 50 to 75 percent of confirmed invoices within 90 days, 25 percent of forecast flows beyond that.
• Which tools are allowed? Forwards and bought options only; no sold options or structured products that the business does not fully understand.
• Who decides, and how is it recorded? One named person books hedges; every booking is logged against an invoice.

The last point now has regulatory weight in India, because banks must collect undertakings that the same exposure is not hedged elsewhere. A clean record of invoice-to-hedge matching saves time with your bank and with your auditor. For context on how cross-border flows move once they are hedged, see our explainers on correspondent banking and the mid-market rate. This is general information, not advice; check any hedging programme with a qualified adviser and your bank's treasury team.

What to watch next

• 16 October 2026: the RBI's higher daily minimum cash reserve requirement, 99 percent of the prescribed amount instead of 90 percent, starts with the new fortnight, tightening rupee liquidity and possibly forward premia.
• October to December 2026: how far banks pass on the cost of the 20 percent reserve to importers. Watch the quotes on your own forwards against the mid-market rate the RBI requires banks to disclose.
• The special dollar window: from 12 October the RBI is meeting state oil companies' dollar needs directly. If it is withdrawn, importer demand returns to the market.
• The RBI's next monetary policy review: after the October hike to 5.50 percent, any further move will feed straight into rupee forward pricing.
• 2028: the next BIS Triennial Survey, which will show whether the 2026 measures reduced the rupee's share of offshore derivatives trading.

The rupee may recover or slide further; nobody can tell you which. What a small business can control is how much of its margin depends on the answer. For the wider context of business payments, read our guide to cross-border B2B payments in 2026.

In July 2020, as the monsoon floods in Bangladesh built toward one of their longest peaks in decades, more than 23,000 families received 4,500 taka each on their mobile phones. The money arrived about five days before the water crested. When researchers followed up, those families were 52% less likely to have gone a full day without food during the flood, according to a case study of the WFP pilot. They had used the cash to move livestock, store grain and buy food before prices jumped.

That pilot is the cleanest illustration of an argument humanitarians have been making for fifteen years: in a disaster, money that moves digitally and early does more good than goods that move physically and late. The argument has largely been won on evidence. What is new in 2026 is that the money itself is drying up, and the remaining budget has to work harder. This piece looks at what the data says about speed and leakage, how India's benefit-transfer plumbing fits in, and where digital relief still breaks down.

A shrinking pot, a settled argument

Cash and voucher assistance, the sector's term for aid given as money rather than goods, totalled US$7.3 billion in 2025, down 11% on the year and 31% below its 2022 peak, according to the ALNAP State of the Humanitarian System 2026. It was the third straight annual decline. Cash held steady at about 21% of all international humanitarian assistance, which means the fall reflects the overall squeeze on aid rather than a loss of faith in the method.

Part of the drop is mechanical. The very large, highly digitised cash programmes in Ukraine and after the Turkiye-Syria earthquakes wound down. Part is political: some governments in West and Central Africa restricted cash programming, and ALNAP notes that cash is also easier to cut quickly than a warehouse full of food. That last point deserves more attention than it gets. Flexibility cuts both ways. A modality that can be scaled up in days can be scaled down in days too.

The World Food Programme remains the largest single deliverer. It transferred about US$2.2 billion in cash-based assistance in 2024 and reached 124.4 million people with all forms of aid, according to its annual performance report on ReliefWeb.

The efficiency case, in rupees

The strongest evidence on leakage now comes from comparing what share of a budget actually reaches people. A CALP Network study cited by ALNAP looked at US$11.4 billion of spending between 2014 and 2025. Cash and voucher programmes delivered 71% to 77% of their budgets directly to recipients. In-kind food aid delivered 54% to 56%. Overall, cash came out 37.5% more efficient, and larger cash programmes got better with scale while in-kind did not.

Worked example: a Rs 10 crore relief budget

Apply those ratios to a district relief allocation of Rs 10 crore. Delivered as cash, somewhere between Rs 7.1 crore and Rs 7.7 crore lands in households' hands. Delivered as food kits, between Rs 5.4 crore and Rs 5.6 crore does. The gap is Rs 1.5 crore to Rs 2.3 crore, or roughly the price of an extra 20,000 to 30,000 families receiving a Rs 7,500 transfer. ALNAP puts the system-wide prize at US$1.1 billion to US$3.3 billion a year if the efficiency levers in the study were applied everywhere.

The caveats matter. The cash advantage shrank in places with heavy security vetting, such as Syria, in high-inflation economies where market checks are costly, and in sudden disasters where no cash system existed beforehand. Cash beats goods when markets work and when a delivery channel is already in place. Where neither is true, it is not automatically better.

The rail is rarely the bottleneck in disaster relief. The list of who is eligible almost always is.

India's version: DBT meets disaster norms

India approaches the problem from the other end. Instead of building a humanitarian payment channel and attaching it to a crisis, it has a vast standing benefit-transfer system and plugs disaster payments into it. A report published in April 2025 by a Viksit Bharat senior fellow, summarised by Business Today, estimated that direct benefit transfer saved Rs 3.48 lakh crore between 2009 and 2024 by curbing leakages, with food subsidy reforms accounting for about 53% of that. Treat the precise figure with some care: it is an estimate built on assumptions about what would have leaked, not an audited number.

Speed is where the standing system pays off. In January 2022, Odisha credited Rs 3,000 each to the bank accounts of 32 lakh families for repairs to houses built under state housing schemes, within 15 days of the cabinet decision, as reported by Orissa Post. That is a scale few humanitarian agencies could match in two weeks, and it worked because the beneficiary list and the bank accounts already existed. The Jan Dhan account base is, in effect, disaster infrastructure.

What a family can actually receive

The amounts are set by the State Disaster Response Fund norms. The 2022-23 to 2025-26 schedule, published on an Assam district portal, lists Rs 4 lakh ex gratia for a death, Rs 1.2 lakh for a fully damaged house in the plains and Rs 1.3 lakh in hilly areas, and Rs 2,500 each for lost clothing and utensils.

Take a farming family in a plains district whose house is destroyed and whose 1.5 hectares of irrigated crop is lost. The norms add up to Rs 1,20,000 for the house, Rs 5,000 for clothing and utensils, and Rs 25,500 in crop input subsidy at Rs 17,000 per hectare. That is Rs 1,50,500 in total. For a household of that kind, the timing of the money matters almost as much as the amount. A crop subsidy that arrives after the next sowing window has passed is a compensation payment, not a recovery tool. Families should check the current norms and eligibility with their district administration, as states can top up these amounts.

Where it breaks: verification, not payment

Sri Lanka's response to Cyclone Ditwah at the end of 2025 shows the pattern clearly. By 30 December, about 87.4% of 450,225 eligible families had received a flat Rs 25,000 relief allowance, according to the Ministry of Defence figures reported by Ada Derana. But only 8.63% of 153,593 families eligible for the larger Rs 50,000 payment had been paid, and under 15% of schoolchildren had received their education grant.

The flat payment needed one check: is this family on the list? The larger tier needed damage assessment. That is the universal shape of disaster finance. Small, flat, early transfers move at the speed of the payment rail. Large, assessed transfers move at the speed of a field officer with a clipboard. Digital tools such as satellite imagery and geotagged photos can speed up assessment, but they introduce their own exclusion risks for informal housing and unregistered land.

The other failure point is the financial system itself. ALNAP records that diaspora money raised for Afghanistan was held up by the banking system and missed the winter it was meant for. In sanctioned or de-banked economies, the fastest rail on paper can be the slowest in practice. This is where some agencies have experimented with purpose-bound and digital-voucher designs, which narrow what money can be spent on in exchange for working in fragile settings.

Mobile money is the crisis rail in most of the world

Outside India and a handful of other countries with deep bank-account coverage, the default disaster channel is mobile money. East Africa's mobile money networks and South Asia's wallet providers already reach households that no bank branch serves. Their advantages in a crisis are practical: a phone number is easier to verify remotely than a bank account, agents can cash out in the village, and payments can be sent in batches to tens of thousands of numbers in an afternoon.

The limits are just as practical. Floods knock out cell towers and power. Agent networks run out of physical cash when everyone withdraws on the same day. And women in many households have less access to a phone in their own name, which affects who controls the transfer. Programmes that work well tend to pre-register recipients, pre-position agent liquidity and build in a fallback to physical distribution if networks fail.

Our view on what actually matters

We think the debate about cash versus in-kind is over, and the debate about digital versus physical cash is mostly over too. The open questions are about lists and timing. Who is on the list before the disaster? Can the trigger fire before the damage, as in the Bangladesh pilot? Can a national safety net, like government-to-person payment systems elsewhere, absorb emergency top-ups without new paperwork?

India is unusually well placed on the first and third questions because of its benefit-transfer base. It is less advanced on the second. Anticipatory payments tied to forecasts, such as a cyclone track or a river gauge reading, are still rare in Indian state practice, even though the payment plumbing to deliver them already exists. That is the gap worth closing.

What to watch next

• October to December 2026: the northeast monsoon and the Bay of Bengal cyclone season, the next real-world test for state relief payments in Odisha, Andhra Pradesh, Tamil Nadu and Sri Lanka.
• 2026-27: the first full year of disaster-fund allocations under the Sixteenth Finance Commission period, after the 2022-26 SDRF norms cycle ended in March 2026. Watch whether the revised schedule raises the Rs 1.2 lakh house assistance figure.
• Early December 2026: the UN's annual Global Humanitarian Overview for 2027, which will show whether cash funding falls for a fourth year from US$7.3 billion.
• 30 November 2026: the end of the Atlantic hurricane season, after which agencies typically publish speed-of-payment reviews for Caribbean and Central American responses.

In January 2026 an accounting firm in a routine attestation confirmed that a single stablecoin issuer earned more than $10 billion of net profit in 2025. That company was Tether, issuer of the world's largest stablecoin, and the figure put it ahead of Coinbase, Robinhood and Kraken combined. The biggest crypto companies in 2026 are not the ones with the loudest marketing or the highest trading volume. They are the ones sitting on reserves, custody balances and licences, and the numbers show a sector whose profits are moving from trading fees toward interest income and infrastructure.

How we chose the top crypto companies

We ranked companies only by stated, measurable criteria as of October 2026: 2025 revenue, 2025 net profit, registered or funded users, and assets on platform or in circulation, as reported by company filings, attestations or established crypto and financial media. Where a company does not publish full financials, we describe it with the figures it does disclose. We grouped companies by type rather than giving a single 1-to-10 order, because an exchange, a stablecoin issuer and a broker earn money in very different ways.

Coverage spans the US, Europe, India and the UAE. Our list of stablecoin companies to watch goes deeper on issuers.

Which is the most profitable crypto company in 2026?

Tether is the most profitable crypto company by reported net profit. It earned more than $10 billion in 2025, according to an attestation by BDO, down from more than $13 billion in 2024. That compares with $1.26 billion for Coinbase and $1.9 billion for Robinhood as a whole, making the stablecoin issuer by far the sector's largest earner.

Tether

Tether issues USDT, the largest stablecoin. At the end of 2025 it had more than $186 billion of USDT in circulation backed by $193 billion of assets, including about $122 billion of US Treasuries and roughly 140 tonnes of gold, with $6.3 billion of excess reserves, according to The Block. As of October 2026 USDT stands at about $184 billion, roughly 60% of a $306 billion stablecoin market, per DefiLlama. Our comparison of USDT and USDC explains the difference in structure.

Circle

Circle, which issues USDC and went public in 2025, reported second-quarter 2026 revenue and reserve income of $701 million, up 7%, and net income from continuing operations of $48 million, according to its results release. USDC in circulation was $73.3 billion at the end of June, up 19%, with $14.8 trillion of onchain transaction volume in the quarter. For the full year 2025, Circle reported $2.7 billion of revenue and reserve income, up 64%, and a $70 million net loss from continuing operations driven by IPO-related stock compensation, per its fourth-quarter release. Its margins are thinner than Tether's because it pays a large share of reserve income to distribution partners. Our piece on Circle's IPO covers that arrangement.

Which is the largest crypto company by revenue?

Coinbase is the largest crypto company by reported revenue, with $7.18 billion in 2025 and net income of $1.26 billion. Kraken reported more than $2.2 billion of 2025 revenue. Robinhood earned $4.5 billion across its whole business, though crypto is only one part of it. Binance does not publish audited group revenue, so it cannot be ranked on this measure.

Coinbase

Coinbase is the largest US-listed crypto exchange. It reported 2025 revenue of $7.18 billion, including $2.83 billion of subscription and services revenue, and net income of $1.26 billion, despite a fourth-quarter net loss of $667 million tied to falling crypto prices, according to Traders Union's summary of its results. The downturn continued into 2026: Quartz reported second-quarter revenue of about $1.2 billion and a net loss of $359.5 million. Coinbase also took a minority stake in India's CoinDCX in October 2025.

Kraken

Kraken, owned by Payward, reported 2025 revenue up 33% to more than $2.2 billion, adjusted EBITDA of $530.6 million, 5.7 million funded accounts and $48.2 billion of assets on platform, according to Finance Magnates. An $800 million raise valued it at $20 billion. In the second quarter of 2026 adjusted revenue was $508 million, up 17%, but Payward is not expected to list before the second quarter of 2027, per Crowdfund Insider. CB Insights recorded Payward buying derivatives exchange Bitnomial for $550 million and card issuer Reap for up to $600 million during the quarter.

Robinhood

Robinhood is a broker rather than a pure crypto company, but it is one of the largest retail crypto venues in the US and, through Bitstamp, in Europe. It reported record 2025 net revenue of $4.5 billion and net income of $1.9 billion, with 27 million funded customers. Crypto revenue in the fourth quarter of 2025 was $221 million, down 38%, on $82 billion of notional volume, of which $48 billion ran through Bitstamp, according to its investor release.

Which crypto exchange has the most users?

Binance has the most users of any crypto exchange, with more than 300 million registered users and over $125 trillion of cumulative trading volume as of December 2025. It is followed at a distance by regional platforms such as India's CoinDCX, with 20.4 million users, and Europe's Bitpanda, with about 30 million customers. Registered users overstate active ones everywhere.

Binance

Binance is the world's largest crypto exchange by users. In December 2025 it received full authorisation from Abu Dhabi Global Market for three entities covering exchange, clearing and custody, and broker-dealer activity, starting operations on 5 January 2026, as AGBI reported. It cited more than 300 million registered users. The ADGM licence gives the group a single primary regulator for its global platform for the first time.

Bybit

Bybit, which holds UAE licences, describes itself as the world's second-largest exchange by trading volume. In October 2025 it became the first crypto exchange to receive a full Virtual Asset Platform Operator licence from the UAE's federal Securities and Commodities Authority, eight months after in-principle approval. It does not publish audited financials, so we list it on licences and scale claims only. Our UAE crypto regulation guide explains how SCA, VARA and ADGM licences differ.

Hyperliquid

Hyperliquid is a decentralised exchange for perpetual futures. Forbes named it to its 2026 Fintech 50 in February 2026, one of five crypto companies on the list, and it is the largest venue for onchain perpetual futures by trading volume. It does not publish audited accounts, so we do not rank it on revenue. It is the clearest sign that a meaningful share of crypto trading revenue is moving onchain.

Which crypto companies lead in India and Europe?

In India, CoinDCX is the largest crypto platform with disclosed figures in 2026: 20.4 million users and $141 million of annualised revenue at a $2.45 billion valuation in October 2025. WazirX is rebuilding after a $234 million hack. In Europe, Vienna-based Bitpanda serves about 30 million customers and put a planned Frankfurt listing on hold in September 2026.

CoinDCX

CoinDCX, founded in 2018, received investment from Coinbase in October 2025 at a $2.45 billion post-money valuation, according to The Block. As of July 2025 it reported over 20.4 million users, $165 billion of annualised transaction volume, over $1.2 billion of assets under custody and $141 million of annualised revenue. The same month it lost $44 million in an exploit that began with malware planted on an engineer's laptop. India's competition regulator cleared the Coinbase stake in December 2025. Our India crypto regulation explainer sets out the tax and compliance rules these platforms operate under.

WazirX

WazirX was once India's best-known exchange. A July 2024 breach drained about $234 million and affected roughly 4.3 million users. In October 2025 Singapore's High Court sanctioned its restructuring scheme after 95.7% of creditors voted in favour, and the company said it would restart operations within ten business days of the scheme taking effect, working with custodian BitGo, as Inc42 reported.

Bitpanda

Bitpanda, founded in Vienna in 2014, operates across Austria, Germany, Switzerland, Italy and France and recently entered the UK. In January 2026 Bloomberg reported it was preparing a Frankfurt listing at a valuation of €4 billion to €5 billion with about 30 million customers, per The Block. In September 2026 it paused the listing, citing market conditions, and said its chief executive would step down in the first quarter of 2027, as Finance Magnates reported.

Which institutional crypto firms matter?

Galaxy is a US-listed institutional crypto firm with $12 billion of assets on platform at the end of 2025. It posted a 2025 net loss of $241 million as crypto prices fell, but its trading and asset management arm earned record adjusted gross profit.

Galaxy

Galaxy runs trading, lending, asset management and a fast-growing data centre business. For 2025 it reported a net loss of $241 million, record digital-asset adjusted gross profit of $505 million, total equity of $3.0 billion and $2.6 billion of cash and stablecoins, according to its results release. Its Helios campus in Texas has over 1.6 gigawatts of approved power, with CoreWeave contracted for 526 megawatts of critical IT load, a pivot that makes it partly an AI infrastructure company.

How do the biggest crypto companies compare?

Worked example: why a stablecoin issuer out-earns an exchange

The profit gap between issuers and exchanges comes down to arithmetic. Circle's second-quarter 2026 figures make it visible.

1. Reserve income: $668 million in the quarter, while USDC in circulation moved between about $77 billion at the end of March and $73.3 billion at the end of June. On a rough $75 billion average, each $100 of USDC earned Circle about $3.50 annualised.
2. After distribution: revenue less distribution costs was $289 million, a 41% margin. So of that $3.50, roughly $1.44 stayed with Circle and the rest went to partners such as exchanges.
3. Tether by comparison: it reported more than $10 billion of 2025 profit on roughly $186 billion of USDT at year end. That is above $5 of profit per $100 of tokens, helped by gold and other investment gains on top of interest income.
4. Coinbase by comparison: $1.26 billion of net income on $7.18 billion of revenue in 2025 is a 17.5% net margin, and trading revenue falls sharply when prices do, as the second quarter of 2026 showed.

Exchanges earn when people trade; issuers earn when people simply hold, and holding has proved the steadier business.

The flip side is rate sensitivity. If short-term US rates fall by one percentage point, an issuer's income on each $100 of reserves falls by about $1, which is why issuers are racing to add payments and other fee income. Our explainer on how stablecoin issuers make money covers that shift.

What should you watch next among crypto companies?

• Q3 2026 earnings season. Coinbase, Robinhood, Circle and Galaxy report third-quarter results over the coming weeks, the first full read on whether the 2026 trading slump has bottomed.
• Kraken's IPO, not before Q2 2027. Payward has filed confidentially with the SEC; timing depends on market conditions.
• Bitpanda's listing. A Frankfurt IPO at €4 billion to €5 billion was planned for the first half of 2026 but paused in September; the company says it remains on the agenda, with a CEO handover due in the first quarter of 2027.
• Mastercard's BVNK deal, by end-2026. A completed acquisition would mark a card network directly owning stablecoin infrastructure.
• WazirX distributions. Creditor token distributions follow the scheme taking legal effect and will be watched closely in India.

This list is editorial research, not investment advice or an endorsement of any company or token. Company figures are as reported by the linked sources; private and offshore firms are not audited to public-company standards. Readers should check with a qualified adviser before making any investment decision.

A pharmacist in Abidjan's Treichville market pays her wholesaler from her phone at the end of each week. The transfer costs her a flat 1%, and taking cash out at an agent costs nothing. Older telecom wallets charged several times that. That gap, between what African money used to cost and what a handful of companies now charge, is where most of the continent's fintech value has been built. It is also why the list of companies that matter in 2026 looks very different from the list of companies that raised the most in 2021.

How we chose. This is a list of 13 African fintech companies grouped by what they do rather than ranked one to thirteen. We looked at scale (users, transaction value, revenue), growth, licences held, funding and valuation, and product breadth, using figures each company or its investors disclosed, as of October 2026. Where no recent number is verifiable, we say so. Global card networks and banks are left out.

How much money went into African fintech in 2025?

African tech companies raised about $4.1 billion in 2025, up 25% on 2024, according to Partech Africa, with fintech taking $769 million of the $2.4 billion in equity. Briter's separate count put total disclosed funding at $3.8 billion, up 32%, with fintech again the most funded sector. The headline story of 2025 was debt, not equity: lenders, not venture funds, financed the biggest growth.

The 2025 Partech Africa Tech VC report counted 570 transactions, with debt rising 63% to a record $1.6 billion and making up 41% of all capital deployed, up from 31% a year earlier. Fintech's slice of equity shrank as cleantech, health and enterprise software grew; investors are no longer writing payments cheques on growth alone.

Briter's Africa Investment Report 2025 split the money by country: South Africa took 32% of funding, Kenya 29%, Egypt 15% and Nigeria just 8%. Nigeria had its lowest share since 2019 but the largest number of deals. Nigerian fintech is still the busiest market on the continent, but its rounds have got smaller in dollar terms. Kenya, by Partech's count, led the continent with $1.04 billion and four of the year's nine megadeals.

Which mobile money companies lead African fintech in 2026?

Mobile money remains the largest part of African fintech in 2026. M-Pesa, run by Safaricom in Kenya and Ethiopia and by Vodacom elsewhere, is the biggest by users and value moved, and Wave is the fastest-growing challenger in francophone West Africa. For the history, see our guide to M-Pesa and mobile money.

M-Pesa (Safaricom, Kenya)

Safaricom's results for the year to 31 March 2026 put active M-Pesa users in Kenya at 43.69 million, up 18%, and M-Pesa revenue at Ksh 182.73 billion, up 13.4%, according to HapaKenya's breakdown of the results. M-Pesa now earns 45.6% of Safaricom's Kenya service revenue, which makes the telco as much a payments company as a phone company. In Ethiopia, M-Pesa reached 5.2 million 90-day customers, more than double a year earlier.

M-Pesa (Vodacom Group)

Vodacom, a major shareholder in Safaricom, reported 103.0 million financial services customers for the year to 31 March 2026, a figure that counts Safaricom on a 100% basis, and said the transaction value it processed reached $525.6 billion, up 16.6%, according to its FY2026 results announcement. Because Safaricom is included, these group numbers should not be added to the Kenyan figures above. Its other markets include Tanzania, Mozambique, Lesotho and Egypt.

Wave (Senegal)

Wave built its business on a flat 1% transfer fee and free cash-in and cash-out, undercutting telecom wallets that charged 5% to 10%. In June 2025 it raised about $137 million (EUR 117 million) in debt led by Rand Merchant Bank, with British International Investment, Norfund and Finnfund, as Launch Base Africa reported. At the time it had more than 20 million monthly active users, about 150,000 agents and operations in eight countries. Its last disclosed valuation was $1.7 billion, set in 2021.

Who are the biggest fintech companies in Nigeria in 2026?

Nigeria's largest fintech companies in 2026 are OPay and Moniepoint by users and transaction value, with Paystack and Flutterwave leading merchant payments. OPay processed $358 billion in 2025 and is preparing a New York listing; Moniepoint says it handles more than $250 billion a year. Nigeria has the most fintech deals in Africa even though its share of funding has fallen.

OPay

OPay is the largest consumer fintech in Nigeria by the numbers it has disclosed. Its gross transaction value more than doubled to $358 billion in 2025, monthly active users rose 57% to 39.3 million, and revenue grew 161% to $536.3 million, according to Nairametrics. It turned an operating profit of $107.1 million. Nigeria supplied 88.1% of revenue, with Indonesia and Egypt making up most of the rest. Citigroup, Deutsche Bank and JPMorgan are managing a planned NYSE listing at a targeted valuation of about $4 billion.

Moniepoint

Moniepoint's products range from handheld point-of-sale devices for cash deposits, withdrawals and card payments to a personal banking app. In October 2025 it raised $90 million to complete a $200 million Series C at a valuation of about $1 billion, Semafor reported. It says it serves more than 10 million businesses and individuals and processes more than $250 billion a year. In April 2025 it launched UK-to-Nigeria remittances, its first step outside domestic payments.

Paystack

Paystack, bought by Stripe in 2020 for $200 million, serves about 300,000 businesses, mostly in Nigeria. In early 2026 it took over Ladder Microfinance Bank, giving a sister company, Paystack MFB, its own banking licence, according to Launch Base Africa. The move followed a reported N250 million Central Bank of Nigeria fine over its Zap transfer app, which regulators said had gone beyond its deposit-taking permissions. Bank transfers made up 66% of its Nigerian volume in 2024.

Flutterwave

Flutterwave is one of the most widely used pan-African payment processors for merchants selling across borders. Founder Olugbenga Agboola wrote in June 2026 that it had passed 1 billion transactions and more than $40 billion in payments since inception, and had won new licences in Senegal, Zambia and Cameroon during the year, per Nairametrics. In January 2026 it bought the Nigerian open-banking start-up Mono in an all-stock deal, adding bank-data access and account-to-account payments.

Which digital banks and lenders are scaling outside Nigeria and Kenya?

Outside Nigeria and Kenya, the fastest-scaling African fintechs in 2026 are digital banks and lenders in South Africa and Egypt. Tyme Group, which runs GoTyme Bank in South Africa and the Philippines, reached a $1.5 billion valuation after Nubank led its Series D, and Egypt's MNT-Halan was valued at $1.4 billion in a June 2026 round led by Al Ahly Capital.

GoTyme Bank (Tyme Group, South Africa)

Tyme Group raised a $250 million Series D in December 2024 at a $1.5 billion valuation, with Brazil's Nubank putting in $150 million, according to FinTech Futures. That remains its latest disclosed valuation. In April 2026 the South African Reserve Bank published the renaming of TymeBank as GoTyme Bank, matching its Philippine sister bank, and the group then had more than 15 million customers, over 10 million of them in South Africa, Financial Afrik reported. It is one of the few African digital bank models running at scale in Asia.

MNT-Halan (Egypt)

MNT-Halan combines consumer and small-business lending with payments. Its latest round, led by Al Ahly Capital, the investment arm of the National Bank of Egypt, valued it at $1.4 billion at first close in June 2026, Wamda reported. It also operates in Turkiye and entered the UAE in 2024.

Who builds Africa's cross-border and stablecoin payment rails?

Africa's cross-border rails in 2026 are built by a small group: Onafriq connects mobile wallets and bank accounts across borders, LemFi serves the diaspora sending money home, and Yellow Card runs a licensed stablecoin on- and off-ramp. Chipper Cash spans app-based transfers and newer stablecoin integrations.

Onafriq

Onafriq runs the plumbing that lets a wallet in one country pay a wallet or bank account in another. When it marked 15 years in June 2025 it said its network could reach 961 million registered mobile wallets and 464 million bank accounts across more than 2,000 corridors, per HapaKenya.

LemFi

LemFi, based in London, serves African and Asian migrants sending money home. Its $53 million Series B in January 2025 was led by Highland Europe, and it reported about $1 billion in monthly payment volume and 1 million active users, according to The Condia. It operated in 27 send-from markets and 20 receiving countries and was buying a licensed Irish firm to passport across the EU.

Yellow Card

Yellow Card is one of the longest-running stablecoin on- and off-ramps on the continent. It received a crypto asset service provider licence in South Africa in late 2024, had raised $33 million, and said it had processed more than $3 billion since 2016 across 20 African countries, per TechCabal.

Chipper Cash

Chipper Cash, a cross-border transfer app that raised a $100 million Series C during the last funding boom, said in December 2025 that it had seven million customers across nine African countries, when it announced a partnership to run stablecoin payments over Stable, a USDT-focused blockchain, per its own announcement. It was also among the wallets that integrated PayPal's PYUSD stablecoin on the Stellar network in September 2025, per the Stellar Development Foundation.

How do Africa's leading fintech companies compare?

Compared side by side in 2026, OPay and M-Pesa lead on users and transaction value, Moniepoint and OPay lead Nigeria, and Tyme and MNT-Halan carry the highest disclosed valuations outside Nigeria. Metrics differ between companies, so read across rows with care.

A worked example shows why scale alone is not the full story. Divide Safaricom's M-Pesa revenue of Ksh 182.73 billion by 43.69 million active users and each user generated about Ksh 4,182 in the year. Do the same for OPay, $536.3 million over 39.3 million monthly users, and the figure is about $13.65 per user. Wave's model sits at the other end: on a transfer equal to $100, its 1% fee is $1, against the $5 to $10 that the same transfer cost on older telecom wallets. Cheap transfers win users; lending, merchant fees and float are what pay the bills. Hence OPay's 285% jump in new loans in 2025.

In African fintech the transfer is the hook. The licence and the loan book are the business.

What does this mean for Nigeria and Kenya?

For Nigeria and Kenya in 2026, the lesson is that the leaders are profitable incumbents rather than fresh start-ups. Both markets are led by companies that moved from payments into lending and banking.

In Kenya, a fintech either plugs into M-Pesa or targets a niche it ignores, and the sharpest growth is now in Ethiopia. For how these systems connect, see our report on payments in Africa in 2026.

In Nigeria, the regulator is the swing factor. Paystack's fine over Zap, the push by fintechs to buy microfinance banks and Yellow Card's wait for an SEC decision all point the same way: the Central Bank of Nigeria and the SEC want fintechs inside the regulatory perimeter, holding licences that match what they actually do. Diaspora flows matter too: Moniepoint's UK corridor and LemFi target the same money that our analysis of the cost of sending money to Africa shows is still among the most expensive in the world to move.

What to watch next in African fintech

The next six months will test whether African fintech can reach public markets. OPay's planned NYSE listing, MNT-Halan's second close and Partech's 2026 report, due in early 2027, are the dates that matter. Stablecoin licensing in Nigeria and South Africa will decide how quickly dollar balances spread inside African apps.

• OPay's listing, expected later in 2026. A successful float near the reported $4 billion target would be one of the largest African fintech listings to date and could reset valuations for Moniepoint and others.
• MNT-Halan's second closing. Still pending as of June 2026; its size will show how much Gulf and Egyptian bank capital is willing to back fintech lending.
• Partech's 2026 report, likely January 2027. The key number is whether debt keeps growing as a share of capital. If it does, expect more fintechs to look like lenders.
• Stablecoin rules. Yellow Card's Nigerian application and further South African CASP licences will decide whether stablecoin balances stay a niche or become a standard feature. Our stablecoin companies to watch and fintech start-ups to watch lists track the wider field.

This list is editorial research, not investment advice or an endorsement of any company. Company figures are as reported by the linked sources and may use different definitions of users and volume.

If stablecoins were airlines, USDT and USDC would be the two carriers that fly almost every route. Together they make up a little more than four-fifths of every stablecoin in existence. Most people pick one by accident, usually whichever their exchange or a friend happened to use. It's worth five minutes to understand the difference.

The quick facts

USDT: the people's dollar

Tether is the older and much bigger of the two, and it won on reach. Walk into a mobile-money kiosk in Lagos, a phone shop in Istanbul or a trading group in Manila and "USDT" is simply what a digital dollar is called. A huge share of it lives on Tron, a blockchain that's cheap and fast enough for everyday transfers; over 97% of the stablecoins on Tron are USDT.

Its weak spot has always been transparency and regulation. Tether publishes reserve attestations, but it has historically kept its distance from the strictest regulators. That choice caught up with it in Europe: when MiCA's transition period ended on 1 July 2026, the big EU exchanges restricted USDT trading for European users.

Tether's answer for the US market was a separate coin, USAT, issued through a federally chartered bank partner and built for the GENIUS Act. That's a clever move. It also means there are now effectively two Tethers: a global one and an American one.

USDC: the regulator's dollar

Circle took the opposite route. It chased licences early, publishes detailed reserve reports and positioned USDC as the stablecoin that banks, payment companies and auditors can sign off on. When MiCA arrived, USDC was ready, authorised through Circle's e-money licence in France.

That makes USDC the default for regulated businesses, especially in the US and Europe, and for newer, low-cost blockchains like Base and Solana. On Base, about 90% of stablecoins are USDC.

The trade-off is reach. In many corridors where people actually send money home, USDT is still the coin that local exchanges quote and local agents accept.

Which one is "safer"?

Honestly, for most people spending or sending modest amounts, the difference in day-to-day risk is small. Both have kept their dollar peg through some rough markets. The difference is more about where you can use it and who will accept it without questions.

• For people who live or bank in the EU or US, or who deal mainly with regulated partners, USDC tends to be the more widely accepted coin.
• In markets across Asia, Africa and Latin America, USDT often has deeper local liquidity, which can mean a better rate at the other end.
• For businesses, it's increasingly normal to hold both and route each payment through whichever gives the better final rate.

The coin matters less than the last mile. What counts is how many rupees, pesos or naira land at the other end.

The bigger picture

For a while it looked like this would be a winner-takes-all fight. It isn't. USDT and USDC are settling into two roles that fit together: one is the street dollar of emerging markets, the other is the compliance-ready dollar of regulated finance. For the person sending money, the choice of coin increasingly happens out of sight: the provider picks whichever route delivers more at the other end and shows the final amount before the transfer is confirmed.

That is how it should be. Which digital dollar moved underneath should be the provider's problem, not the customer's.

A cardiologist in Houston sends $3,000 to his parents in Hyderabad every quarter from his checking account, and the new US remittance tax does not touch him. A construction worker in Queens who sends $300 to Ludhiana in cash at a neighbourhood money transfer counter pays an extra $3 every time. Same corridor, same destination country, different tax, and the difference comes down to how the money was paid in. That split sums up the North America to India corridor in 2026: overwhelmingly digital, increasingly cheap, and with a shrinking cash minority that is now paying more.

The biggest single source of India's remittances

India received a record $135.46 billion in private transfers in 2024-25, up 14% on the year, according to RBI data. The United States is the largest contributor. The RBI's sixth remittances survey found the US accounted for 27.7% of inflows in 2023-24, up from 23.4% in 2020-21. On that year's total of about $118.7 billion, that implies roughly $33 billion. Advanced economies led by the US, UK, Canada, Australia and Singapore now supply more than half of India's remittances.

That shift towards skilled migrants in rich countries changes the shape of the money. Transfers are larger and more digital: 73.5% of remittance transactions to India were digital in 2023-24, the RBI survey found.

The 1% US remittance tax: what it covers

Since 1 January 2026, a 1% federal excise tax has applied to some remittance transfers from the US under section 4475 of the tax code, added by the One Big Beautiful Bill Act signed in July 2025. The scope is narrow. It applies only when a transfer is funded with cash, a money order, a cashier's check or a similar physical instrument. Proposed Treasury regulations published in April 2026 indicate that transfers funded from a US bank account or with a US-issued debit or credit card are exempt, as are ordinary wires and ACH payments, according to a summary by Greenback Tax Services. The test is where the sender is located, not citizenship. Providers collect the tax at the point of sale and report it to the IRS.

For India, the direct hit is small. The Niskanen Center estimates inflows could fall by about $466 million, a fraction of a percent of the corridor. In Mexico, where FXC Intelligence says cash-funded transfers were only about 1% of inflows, the effect has been limited too. The bigger effect is behavioural: the tax gives cash users one more reason to switch to a bank account or card. Our US-Mexico corridor analysis looks at that shift.

What sending costs from the US and Canada

World Bank data for the third quarter of 2025 shows a clear gap between the two North American corridors. Sending $200 from the US to India cost 2.92% on average, with a fee of about $4.20 and an exchange margin of about 0.82%. Sending CAD 200 from Canada to India cost 4.28%, with a fee of about CAD 6.09 and a margin of 1.23%. The UK, by comparison, averaged 1.91%, as we explain in our UK to India guide.

The ranges are wider than the averages suggest. In the US sample, one bank-linked option was priced slightly better than the reference rate, while the dearest money transfer operator cost 6.35%. In Canada the cheapest bank transfer cost 0.86%, but a branch transfer at a large domestic bank cost 28.33% of the amount sent. In both countries, branch counters are where costs balloon.

The new US tax is less a toll on remittances than a push away from cash.

Canada: a smaller corridor reshaped by student caps

Canada supplied 3.8% of India's remittances in 2023-24, roughly $4.5 billion on that year's total. Its profile is unusual. Many Indians in Canada arrived as students and moved into work through post-graduation permits, so money flows both ways: families in India pay tuition and living costs, and graduates who find jobs start sending money home.

That pipeline has narrowed. Ottawa has capped new international study permits since 2024, and reports based on government data indicate that permits for Indian students fell by about half in 2025, according to The Wire. Fewer students today means fewer young workers remitting later. It also explains why cost matters so much here: at 4.28%, Canada is the most expensive of India's big Western corridors, and many senders are on entry-level wages. For the outbound side, see our guide to paying international student fees.

Digital versus cash

For most Indian Americans and Indo-Canadians the question is no longer whether to go digital but which digital route to use. The World Bank's global data shows digital transfers averaging 4.85% in early 2025 against 7.16% for non-digital ones, and in the US the cash side now carries an extra 1%. Anyone still sending cash can usually save simply by opening a basic bank account or paying the same provider by debit card. The other habit worth changing is the reflexive use of a big bank's branch wire, which in both countries produced some of the worst prices in the World Bank sample.

Speed has stopped being a reason to pay more. Many digital transfers to Indian bank accounts now credit within minutes once the provider releases the funds, so the main trade-off is between price and convenience features such as rate alerts, scheduled transfers and saved recipients. Those are worth having, but not at the cost of a percentage point on every transfer.

Where the money goes, and paying it directly

A lot of North America's money to India covers recurring obligations: parents' medical and household costs, home loan instalments, insurance premiums, property tax and younger siblings' school fees. A Canadian fintech began offering direct payment to more than 21,000 Indian billers through Bharat Connect in Canadian dollars in February 2025, and NRIs with US or Canadian mobile numbers can link UPI to NRE or NRO accounts and pay bills in rupees. Our NRI bill payments guide compares the routes.

What to watch next

Three things. First, the final Treasury rules on the remittance tax, which should settle edge cases the proposed version left open. Second, Canada's student intake, which will shape that corridor's remittances for years. Third, the rupee: after sliding past 95 to the dollar in 2026, each dollar sent buys more at home, and a weak rupee has historically pulled transfers forward. None of that changes the basic advice, which is to fund transfers electronically, compare the rupees delivered and stay away from branch wires.

This article is general information, not tax or financial advice.

In 2020 the G20 did something unusual for a club of finance ministers: it put numbers on a promise. By the end of 2027, cross-border payments were meant to be cheaper, faster, more widely available and more transparent, with specific percentages attached. With a little over a year to go, the people tracking those numbers have said plainly that the world will miss most of them.

What the targets actually say

The targets differ by segment, as summarised in a February 2026 Reserve Bank of Australia bulletin:

• Cost. Retail payments should average below 1%, with no corridor above 3%. Remittances should average below 3%, with no corridor above 5%, by 2030.
• Speed. For retail, remittances and wholesale alike, 75% of payments should reach the recipient within one hour and the rest within one business day.
• Access. Every user should have at least one electronic option for sending and receiving, and more than 90% of individuals should have a means of receiving remittances electronically.
• Transparency. Every provider should show the total cost, expected delivery time, tracking and terms before a payment is made.

Most of these fall due at the end of 2027. The remittance cost goal runs to 2030, matching the UN development target.

Where the world stands

The Financial Stability Board's 2025 progress report found only slight improvement since indicators were first measured in 2023. The board said it is unlikely that satisfactory global improvement will arrive in line with the 2027 timetable.

The detail is sobering. Only 35.4% of retail cross-border payments were credited within an hour in 2025, up 1.9 percentage points. For remittances the figure was 54.4%. Both are well short of 75%. Wholesale bank-to-bank payments did better, with more than half arriving within an hour. On cost, person-to-person retail payments averaged about 2.5%, stable since 2023, against a 1% goal. Sending $200 as a remittance cost about 6.4% on average, more than double the 3% target.

The World Bank's own price survey tells the same story from the ground. In the first quarter of 2025, the global average cost of sending $200 was 6.49%, and the World Bank's Q3 2025 reading was still 6.36%. Banks averaged 14.55%, while digital services averaged 4.85%. Sub-Saharan Africa remained the most expensive region to send to at 8.78%, and South Asia the cheapest at 4.80%.

Averages hide wide regional gaps. On FXC Intelligence's reading of the FSB data, Sub-Saharan Africa is the most expensive region to send from, at just over 4% for a $1,000 person-to-person payment, while Europe and Central Asia is the cheapest at 1.93%. About 18% of corridors still average above 3%. Speed varies by how money goes in: cash-funded transfers often arrive fast, while about 42% of transfers funded from bank accounts take longer than a full business day. A warehouse supervisor in Riyadh who sends money home from a bank app may be waiting longer than her colleague who pays cash at a counter.

There is good news in access. FXC Intelligence's reading of the FSB data shows 78.7% of adults now have a transaction account, up nearly five points since 2022, and transparency is improving, with about 63% of services showing both cost and speed up front.

Why the gap is so stubborn

The FSB's own diagnosis is that most of the policy work is done, but it has not reached the end user. Domestic payment systems are the first and last mile of every cross-border payment, and if a receiving bank in Lagos or Dhaka only credits accounts in batches, a fast international leg does not help. Long correspondent chains add cost and compliance checks at every hop, which we cover in correspondent banking explained. And many corridors are small, so providers have little reason to compete on price.

The international leg of a payment is rarely the slow part anymore. The delays live at the edges, inside domestic systems and compliance queues.

The rails that could close the gap

ISO 20022 messaging

On 22 November 2025, Swift ended the coexistence period for cross-border payment messages, making ISO 20022 mandatory. Swift said about 97% of instructions were already in the new format just before the deadline. Richer structured data should mean fewer payments stopped for manual checks, which is where a lot of delay comes from. Our piece on Swift and ISO 20022 in 2026 goes deeper.

Linking instant payment systems

The most direct fix for speed is connecting domestic instant systems to each other. Nexus, first designed at the BIS, is the most ambitious attempt. Its operating company appointed a joint venture of Malaysia's PayNet and Singapore's NETS as technical operator in February 2026, with India, Malaysia, the Philippines, Singapore and Thailand as founding partners and go-live targeted for 2027, according to a Nexus Global Payments announcement. The design aims for transfers in under a minute. See Project Nexus explained.

Stablecoins

Stablecoins settle in minutes at any hour, which speaks directly to the speed target. Business payments in stablecoins reached an estimated $226 billion a year by late 2025, according to McKinsey. But they do not by themselves solve the last mile: the recipient still needs a way into local currency, and the FSB has been cautious about their risks. They are likely to help most in corridors where banks are slow and costly, rather than replace bank rails generally. Bill payments and supplier invoices in high-cost corridors are where that is already happening.

What happens after 2027

Missing a target is not the same as failing. The roadmap gave the industry common measurements, and the FSB is now pushing for more regional and national action rather than more global plans. Expect the next phase to focus on the hardest corridors, especially those into Sub-Saharan Africa, on getting more domestic systems to run instant payments, and on linking them. For a parent in Leeds paying school fees in Kochi, or a shop in Nairobi paying a supplier in Mumbai, the change will show up as fewer days of waiting and a clear price before pressing send. Those are the two things the targets were always really about.

The bigger question is whether the official sector or the market moves faster. Central banks are building links like Nexus on multi-year timelines, while private payment companies are wiring stablecoin corridors in months. The likely outcome is a mix: regulated bank rails for large and routine flows, and newer rails filling gaps where those rails are slowest. Corridors such as the ones covered in our UAE to India money transfer guide will be an early test of which approach users prefer.

Figures are from the FSB's 2025 consolidated progress report and the World Bank's Remittance Prices Worldwide survey, the latest available as of October 2026.

A small-batch hot sauce maker in Albuquerque gets her first order from Sao Paulo on a Tuesday and her first from Munich on the Friday. Both customers found her through the same video. The Brazilian abandons checkout because there is no way to pay by Pix. The German pays, but by card, and when the processor's statement arrives she sees that the order cost her almost twice as much to accept as a sale to New Mexico. Cross-border e-commerce is mostly won or lost in those two moments: whether the customer can pay the way they normally pay, and what it costs the merchant when they do.

This guide is for founders, e-commerce managers and finance leads selling abroad. It uses the 2026 edition of the Worldpay Global Payments Report for market data, a large processor's published US price list for cost maths, and Baymard Institute's checkout research for the conversion side. The conclusion is blunt: in most large markets outside the United States, a card-only checkout is a niche checkout.

The world does not pay online by card

The 2026 Global Payments Report, based on more than 63,000 consumers in 42 markets, found that digital wallets accounted for 56 percent of global e-commerce transaction value in 2025 and 33 percent of in-store spending. That headline needs unpacking, because "wallet" covers very different things: a card stored in a phone wallet, a stored-value app in China, or an app sitting on top of a real-time bank rail such as India's UPI.

The market-by-market view from a Payment Expert summary of the report is where the practical decisions live. Alipay and WeChat Pay take 89 percent of Chinese e-commerce value. UPI-driven wallets take 68 percent of online spending in India. Brazil's Pix takes 42 percent of e-commerce. In Germany wallets reach 52 percent of online value and buy now, pay later another 18 percent. The UK is still card-heavy online at 46 percent. And the United States, where many sellers design their checkout, is the outlier: cards were 49 percent of US online value in 2025, with wallets at 40 percent.

The pattern is that where a country built a fast, cheap domestic rail, consumers moved to it online. For background on those rails, see our explainers on UPI and its global expansion and on Brazil's Pix.

What checkout friction actually costs

Baymard Institute's long-running research, summarised by Contentsquare, puts the average documented cart abandonment rate at about 70 percent. Among shoppers who abandoned for reasons other than browsing, 39 percent cited unexpected extra costs such as shipping, taxes or fees, 19 percent said they did not trust the site with payment details, and 10 percent said there were not enough payment methods. Those reasons overlap, but every one of them is sharper across a border: duties appear late, the brand is unfamiliar, and the familiar method is missing.

Localisation matters on top. A seven-country Pitney Bowes and Morning Consult survey of 7,000 shoppers ranked localisation as the second most common driver of cross-border abandonment after price and convenience, and identified prices not shown in the shopper's own currency as a leading problem. Older vendor data points the same way: FastSpring reported that clients adding local currencies typically saw conversion rise 5 to 20 percent, though that data dates from 2014 and comes from a seller of the service.

A foreign shopper forgives slow shipping more readily than an unfamiliar way to pay.

A worked example: the real cost of an $80 order

Our Albuquerque merchant uses a processor whose published US price list charges 2.9 percent plus 30 cents for a domestic card, adds 1.5 percent for international cards and another 1 percent when currency conversion is needed. We use it because the numbers are public; other providers price differently, and large merchants negotiate.

1. Domestic card, $80: 2.9 percent is $2.32, plus $0.30, for $2.62, or 3.3 percent of the order.
2. UK customer, card, priced in dollars: add 1.5 percent ($1.20) for the international card. Total $3.82, or 4.8 percent.
3. UK customer, card, priced in sterling: add another 1 percent ($0.80) for conversion. Total $4.62, or 5.8 percent. Localised pricing has cost her a point of margin, but it probably saved the sale.
4. Brazilian customer paying by Pix through the processor's wallet product: listed at 1.8 percent plus 30 cents, plus the same 1.5 percent international and 1 percent conversion surcharges. Total $3.74, or 4.7 percent, cheaper than the UK card sale and, crucially, the sale happens.
5. Stablecoin payment: listed at 1.5 percent, or $1.20. Few consumers pay this way today, but for some business buyers it is the lowest-cost option on the sheet.

Now scale it. At 1,000 UK card orders a month, the gap between domestic and localised international acceptance is about $2,000 a month. That is real money, but it is small next to the revenue lost when shoppers leave. Of 10,000 carts started, roughly 7,000 are typically abandoned; if a tenth of those shoppers cite missing payment options, several hundred potential orders are at stake each month. Pricing a market correctly means comparing the fee premium with the conversion you buy for it, not just minimising fees.

FX at checkout: who takes the spread?

Every cross-border sale involves a currency conversion somewhere. The only question is where, and who pays for it.

• Price in your currency, let the customer's bank convert. Cheapest for the merchant, worst for conversion. The shopper sees an unfamiliar number, then a foreign transaction fee on their statement.
• Price in the customer's currency, convert at settlement. The merchant pays a conversion fee, typically around 1 percent on processor price lists, and carries rate risk between pricing and settlement. This is usually the best compromise.
• Hold balances in foreign currency. Merchants with costs abroad, such as a European warehouse or ad spend in euros, can settle into multi-currency balances and convert only the surplus. That turns a fee into a natural hedge.
• Dynamic currency conversion offered to the shopper. The shopper chooses to pay in their own currency at a rate set by the acquirer. It looks helpful, but the rate often carries a markup well above the mid-market rate, and regulators in several markets require that markup to be disclosed.

One structural point is easy to miss. The dollar sits on one side of 89 percent of all FX trades, according to the BIS 2025 Triennial Survey. A seller in Lisbon pricing in Brazilian reais may find that the conversion goes through dollars twice behind the scenes. Ask your provider how many conversions happen and at what rate.

Wallets, rails and the cross-border shortcut

The newest development is that domestic rails are starting to work across borders. Worldpay's report notes that Pix is accepted in Argentina, Chile, Portugal, Spain and the United States, UPI in Malaysia, Singapore, Thailand, the UAE and elsewhere, and that Alipay+ connects 1.8 billion users to 100 million merchants across 14 markets. For a merchant, that means a tourist or diaspora shopper may already expect to pay with a home-country app.

Stablecoins sit at the edge of this picture. Card networks and large processors have added stablecoin acceptance and payouts, and the fee schedule above shows why some merchants care: 1.5 percent with no international surcharge is competitive with any cross-border card. Consumer usage is still small; the more meaningful use is on the payout side, where merchants pay suppliers, creators or marketplace sellers abroad. Our guide to stablecoin payments for business covers that side.

A practical checklist for selling abroad

• Rank markets by revenue, then fix the top three. Use the table above to see which local methods those markets expect.
• Show prices in local currency, including duties. Unexpected extra costs are the top abandonment reason; landed-cost pricing removes the late surprise.
• Add the one dominant local method in each priority market before adding a long tail of options.
• Watch authorisation rates for foreign cards separately. Issuers decline cross-border card-not-present transactions more often; local acquiring or network tokens can help.
• Decide where FX happens and record it in your pricing model, not just your processor statement.
• Reconcile by market monthly: fees, FX cost, chargebacks and refunds, so you see the true margin per country.

For the business-to-business side of the same problem, see our guide to cross-border B2B payments in 2026. Nothing here is a recommendation of any provider; compare current pricing and contract terms yourself.

What to watch next

• 1 January 2027: the stablecoin rate on the processor price list we used rises from 0.8 percent to 1.0 percent for its wallet product, an early sign of how stablecoin acceptance will be priced once promotions end.
• Spring 2027: the next Global Payments Report. The 2026 edition was published on 31 March; watch whether India's UPI share and Brazil's Pix share keep climbing.
• 2025 to 2030: Worldpay forecasts wallet spending at the point of sale to grow about 11 percent a year globally, US wallet spending to reach $4.1 trillion by 2030, and UK wallet spending to rise from GBP 269 billion to GBP 453 billion.
• Cross-border rail links: each new UPI or Pix acceptance corridor changes what travelling and diaspora shoppers expect at your checkout.

The direction is clear even if the pace is not. Checkout is becoming local everywhere, and the merchants who treat payment methods as part of market entry, alongside language, shipping and pricing, will keep the customers that a single global card form loses.

In a district treasury office in Odisha, an officer reviewing a welfare scheme can now ask a question that cash never allowed: could this subsidy be issued as digital rupees that only work at approved fertiliser dealers? The Reserve Bank of India has run pilots doing exactly that. But the code that makes such a rupee programmable, the ledger that records it and the wallet that holds it were not all written inside the central bank. Behind almost every central bank digital currency sits a small group of technology vendors, and choosing among them is one of the least discussed decisions in modern monetary policy.

How many countries are building a CBDC in 2026?

As of May 2026, 146 countries and currency unions representing more than 98% of global GDP are exploring a central bank digital currency, according to the Atlantic Council's tracker. Of those, a record 77 are in development, pilot or launch, the tracker counts 41 pilot projects, and only three countries have fully launched: the Bahamas, Jamaica and Nigeria.

The gap between 146 and 3 is the market that vendors are chasing. Most central banks do not have large in-house engineering teams, and almost none want to run a consumer payments business. They need someone to supply the core ledger, the issuance and redemption tools, the interfaces for commercial banks, and often the wallets and offline hardware too. The Atlantic Council tracker shows how uneven progress is, and our own CBDC tracker follows country-level changes.

What does a CBDC technology provider actually supply?

A CBDC technology provider supplies some or all of four layers: a core ledger where the central bank mints and destroys currency, a distribution layer for commercial banks and payment firms, end-user wallets and merchant tools, and specialised features such as offline payments or programmability. In 2026 few central banks buy all four from one vendor.

Ripple's description of its own platform is a useful map of the layers: an issuer module for minting, redemption and destruction; an operator module for banks distributing the currency; end-user wallets; and the ledger itself, with the controls central banks require. Bitt's work with the Eastern Caribbean Central Bank followed the same pattern, supplying a central bank system, a financial institution system, a merchant app and a free wallet.

The architectural choices behind those layers matter more than the brand on the contract:

• Token or account. In a token model the digital note itself carries value, like cash. India's retail e-rupee is described as a digital token representing legal tender, issued in the same denominations as notes. In an account model, balances sit in ledger accounts tied to identities.
• One tier or two. In a two-tier model the central bank issues and commercial banks distribute and handle customers. India's pilot distributes through banks; almost every retail CBDC under development uses this structure.
• Retail or wholesale. Retail CBDCs reach the public. Wholesale CBDCs settle between banks and are the focus of BIS projects such as Agora. Our wholesale CBDC explainer covers the second.
• Online or offline. Offline payment capability, where two devices transfer value without a network, is a major differentiator for inclusion-focused countries.

Who are the main CBDC technology providers?

The CBDC technology providers with verified central bank work as of 2026 include Giesecke+Devrient, Bitt, Soramitsu, R3, Consensys and Ripple for retail and pilot projects, and Fnality and BIS-led consortia for wholesale settlement. Each has a distinct technical approach and a different track record of moving from proof of concept to live use.

Giesecke+Devrient (Filia)

The German banknote and security printer sells a CBDC product called Filia, designed around offline payments that can be made consecutively without a network. The Bank of Ghana signed with G+D to pilot the e-Cedi, a general-purpose retail CBDC, in three phases covering design, implementation and field testing through mobile apps and smart cards, as The Fintech Times reported. The aim includes payments without a bank account or smartphone.

Bitt

Barbados-based Bitt is the only vendor on this list behind two currencies the public actually used. It built DCash for the Eastern Caribbean Central Bank, launched on 31 March 2021 as the first retail CBDC issued within a currency union, per the ECCB's announcement. In August 2021 the Central Bank of Nigeria chose Bitt as technical partner for the eNaira, citing competence, security, interoperability and implementation experience, TheCable reported. Nigeria is one of the Atlantic Council's three launched CBDCs, though public use has stayed thin: in June 2026 the central bank's Payments System Vision 2028 acknowledged slow adoption and set out to reposition the eNaira around government-to-person and cross-border payments, Central Banking reported. DCash had a harder ending. Its pilot concluded in January 2024, and in February 2026 the ECCB suspended work on DCash 2.0 to prioritise a regional fast payment system, MNI reported.

Soramitsu

The Japanese firm built Cambodia's Bakong system on Hyperledger Iroha, live since 2020. Soramitsu reports more than 30 million wallets in a country of about 17 million people, and 608 million transactions worth over $150 billion in 2024, according to its September 2025 release. Its own material classes Bakong as closer to tokenised deposits than a pure CBDC. It has also run proofs of concept in Laos, the Solomon Islands and Papua New Guinea.

R3 (Corda)

R3's permissioned Corda platform has been used in several central bank experiments, including the Bank of Thailand's Inthanon project from 2018. In April 2023 the Central Bank of the UAE named R3 as technology provider and G42 Cloud as infrastructure provider for its CBDC strategy, with Clifford Chance on legal oversight, according to Crowdfund Insider. The plan included a CBDC bridge proof of concept with India.

Consensys

Consensys brings enterprise Ethereum technology to central bank pilots. In 2020 it was selected by the Hong Kong Monetary Authority and the Bank of Thailand for phase two of their joint Inthanon-LionRock cross-border project, working with PwC.

Ripple

Ripple markets a CBDC platform based on XRP Ledger technology. In 2023 it said it was in talks with more than 20 countries and named work with Palau on a national digital currency and with the Central Bank of Montenegro on a pilot and strategy. Neither has become a launched retail CBDC.

Fnality (wholesale)

Fnality, owned by a consortium of global banks, launched its sterling Fnality Payment System in December 2023, letting banks settle in digital tokens backed one for one by funds at the Bank of England. It raised a $136 million Series C in September 2025 led by investors including WisdomTree, Bank of America and Citi, per The Block.

BIS Innovation Hub projects

mBridge connected the central banks of China, Hong Kong, Thailand and the UAE for cross-border wholesale CBDC payments; the BIS withdrew in late 2024, saying the project had graduated. Project Agora, also BIS-led, brings together the New York Fed, central banks from Europe, Korea, Mexico and Japan, and more than 40 commercial banks. User testing began around January 2026 and was expected to run about six months, Reuters reported via Malay Mail. Our mBridge explainer has the detail.

How do CBDC providers compare?

The table summarises each provider's verified central bank work as of October 2026. It is limited to partnerships confirmed by a central bank, the company or a reliable news report. Many vendors claim further confidential engagements; we have left those out.

How does a central bank choose a CBDC platform?

Central banks typically choose a CBDC platform through a staged process: research, a request for information, proofs of concept with several vendors, a competitive selection for a pilot, then a decision on issuance. In 2021 Nigeria's central bank said it chose Bitt for technological competence, efficiency, security, interoperability and implementation experience, a list most central banks would recognise.

In practice, five questions dominate:

1. Sovereignty and control. Can the central bank run, audit and if necessary replace the system? The ECCB's governor framed it bluntly: "we will not outsource our development." Vendors that hand over source code and operational control score higher.
2. Resilience and offline capability. A national payment system must keep working in outages and in areas with weak connectivity, which is why G+D leads with offline payments.
3. Interoperability. The CBDC must connect to existing real-time systems such as India's UPI, which handles about 802 million payments a day as of September 2026, and to other central banks for cross-border use.
4. Privacy and compliance. Designs must satisfy anti-money-laundering rules while protecting user data, a debate covered in our CBDC privacy explainer.
5. Proven delivery. Very few vendors have moved a CBDC from pilot to public use. Bitt and Soramitsu are the main exceptions on retail.

Building a CBDC ledger is the easy part; getting a population to want it is not.

Which technology runs India's digital rupee?

India's digital rupee is issued by the Reserve Bank of India and distributed through commercial banks in a two-tier, token-based model. The retail pilot began on 1 December 2022 with SBI, ICICI Bank, YES Bank and IDFC First Bank, followed by Bank of Baroda, Union Bank of India, HDFC Bank and Kotak Mahindra Bank, according to Business Standard.

The RBI has not published a single named vendor for the core platform, and we have not found a verified one. What it has published is direction. Its 2025-26 annual report, summarised by Outlook Money, describes programmable pilots under central and state direct benefit transfer schemes, a memorandum of understanding with the Monetary Authority of Singapore, and talks with MAS and the Central Bank of the UAE on a cross-border CBDC pilot. Our guide to the digital rupee explains how this works for users.

A worked example: what adoption looks like in numbers

Compare two systems built for different purposes. India's retail e-rupee in circulation fell from Rs 1,016.46 crore on 31 March 2025 to Rs 771.66 crore on 31 March 2026. The change is Rs 244.80 crore, and 244.80 divided by 1,016.46 is a decline of about 24% in a year.

Cambodia's Bakong, by Soramitsu's figures, processed 608 million transactions worth over $150 billion in 2024. Dividing $150 billion by 608 million gives an average transaction of roughly $247, and 30 million wallets for 17 million people works out to about 1.8 wallets per person. The comparison is imperfect, because Bakong functions as a fast payment system with tokenised deposits and the e-rupee competes with UPI. That is the point: a technically sound CBDC that duplicates an excellent existing rail struggles for users, as our e-rupee vs UPI analysis argues.

What should governments and vendors watch next?

The next twelve months will test whether CBDC vendors can move from pilots to production. Key dates include the RBI's financial year 2026-27 plans for wider pilots and cross-border tests, Agora's report to decision makers after its 2026 testing phase, and decisions in large economies that could create the biggest contracts yet.

• RBI pilots through March 2027. The RBI plans to expand direct benefit transfer and retail pilots, add tokenisation pilots, run a CBDC and Asset Tokenisation sandbox, and explore bilateral and multilateral cross-border pilots in financial year 2026-27.
• Agora results, late 2026. After roughly six months of testing from January 2026, participants were due to report to decision makers on what a launch would require, possibly adding currencies already settled through CLS.
• UAE to India. Any operational CBUAE and RBI pilot would be one of the first live retail-adjacent corridor tests involving India.
• Wholesale tokenised settlement. Fnality's expansion beyond sterling, funded by its 2025 round, will show whether privately run, central-bank-backed settlement can compete with direct wholesale CBDCs.

Partnerships listed are those confirmed by a central bank, the company or a reliable report as of October 2026. This article is research, not procurement or legal advice.

A drywall subcontractor in Calgary finishes a job on a Friday afternoon and sends his two helpers their pay by Interac e-Transfer from his phone. The money lands in minutes, but behind it sits a national batch system that still settles the banks' positions later, and a payroll file he cannot automate because Canada has no live open banking regime. Canadian fintech in 2026 is built around that gap: a handful of well-funded consumer and business apps, one dominant domestic network, and two pieces of national infrastructure that have been promised for years and are finally close.

How did we choose the leading Canadian fintech companies for 2026?

We chose Canadian fintech companies for 2026 using measurable criteria as of October 2026: customers or clients, assets or volume, latest funding round or valuation, licences held or sought, and product breadth. We grouped them by category rather than numbering them, and we include the two national systems, Interac and Payments Canada's Real-Time Rail, because they shape what every fintech here can build.

Capital is scarcer than it was. Canadian fintech raised US$996.7 million across 47 deals in the first half of 2026, down 40% by value from US$1.7 billion a year earlier, according to KPMG's Pulse of Fintech as reported by Wealth Professional. Fewer deals, bigger cheques: the average round grew even as the count fell. That favours companies that already have scale, which is most of this list.

Which Canadian consumer fintechs are largest in 2026?

By valuation and assets, Wealthsimple is Canada's largest consumer fintech in 2026, valued at C$10 billion in October 2025 with C$100 billion under administration for 3 million clients. KOHO is next by members, with more than 2.5 million and a C$1.33 billion valuation in June 2026, while Neo Financial reports more than one million customers and Nesto leads digital mortgages.

Wealthsimple

Wealthsimple began as a robo-adviser and now offers trading, crypto, tax filing and cash accounts. In October 2025 it announced an equity round of up to C$750 million at a C$10 billion post-money valuation, led by Dragoneer with CPP Investments joining, according to Torys, which advised on the deal. At the time it served 3 million Canadians and held C$100 billion in assets under administration.

KOHO

Toronto-based KOHO offers spending and savings accounts, credit-building and international transfers through partner institutions. It raised C$130 million at a C$1.33 billion valuation in June 2026, with Mubadala and Shopify founder Tobi Lutke among new backers, and says it serves more than 2.5 million Canadians, per its June 11 announcement. It is pursuing a federal banking licence, subject to ministerial approval.

Neo Financial

Calgary's Neo issues credit cards, savings accounts and mortgages, often to people with thin credit files. It closed C$68.5 million of equity in January 2026 to fund a securitization program and has raised more than C$650 million in debt and equity since launch, according to BetaKit. It reports more than one million customers. Its 2024 round reportedly cut its valuation to US$510 million, a reminder that not every Canadian fintech has re-rated upward.

Nesto

Montreal's Nesto is a mortgage platform that also licenses its software to lenders. It raised a C$302 million Series E of primary and secondary capital at a C$1.47 billion valuation in June 2026, with La Caisse among new investors, and says it is profitable with more than C$80 billion of mortgages under administration, per La Caisse's release. KPMG counted the round as the largest Canadian fintech deal of the first half.

Which Canadian payments and commerce companies lead?

Shopify is by far the largest Canadian fintech business by payment volume in 2026, processing US$78.1 billion through Shopify Payments in the second quarter alone. Nuvei, taken private by Advent for about US$6.3 billion, and Lightspeed, with US$1.23 billion of fiscal 2026 revenue, are the other Canadian-founded payments heavyweights, both serving merchants worldwide.

Shopify (Shopify Payments)

Shopify is rarely called a fintech, but payments are now most of its economics. In the second quarter of 2026 merchants processed US$115.6 billion of GMV, and US$78.1 billion of that ran through Shopify Payments, a 68% penetration rate, up from 64% a year earlier, according to EcomCrew's breakdown of the results. Revenue rose 34% to US$3.58 billion.

Nuvei

Montreal-founded Nuvei processes card and alternative payments for merchants in many markets. Advent International agreed in 2024 to take it private at US$34 a share, valuing it at about US$6.3 billion, when quarterly payment volume was running at US$61.3 billion, as Finance Magnates reported. As a private company it no longer publishes quarterly figures.

Lightspeed

Lightspeed sells point-of-sale software and embedded payments to retailers and restaurants. For the fiscal year to March 31, 2026 it reported revenue of US$1.227 billion, up 14%, and a net loss of US$144.4 million, much narrower than the prior year, per its results filing. Fourth-quarter payment volume was US$9.6 billion, 42% of transaction volume.

Which Canadian business finance fintechs are growing fastest?

Float is the fastest-growing Canadian business finance fintech by funding momentum in 2026, raising C$85 million in June at a valuation up 70% and serving more than 7,500 businesses. Clearco, the e-commerce financier that shrank sharply after 2022, is rebuilding with a US$100 million credit facility from Macquarie announced in August 2026.

Float

Toronto's Float offers corporate cards, bill pay and business accounts in Canadian and US dollars. Its C$85 million Series C, led by Inovia Capital on June 25, 2026, came with more than 7,500 active business customers and revenue up more than 120% since its previous round, according to Dealroom. Float, KOHO and Wise joined Payments Canada as members earlier in 2026.

Clearco

Clearco funds inventory and marketing for online brands in exchange for fixed repayments. After a 2021 valuation above US$2 billion, it cut staff and recapitalised in 2023. In August 2026 it secured US$100 million of asset-backed financing from Macquarie for about 400 customers and expects to break even by the fourth quarter of 2026, according to BetaKit.

What infrastructure do Canadian fintechs depend on?

Canadian fintechs in 2026 depend on Interac, which handled about 1.6 billion e-Transfers and 7 billion debit transactions in 2025, and on two systems still being built: Payments Canada's Real-Time Rail, expected to launch in the fourth quarter of 2026, and the consumer-driven banking framework, whose payments phase targets mid-2027.

Interac

Interac is a network, not a startup, but it is the rail most Canadian fintechs ride. In 2025 it processed about 7 billion debit transactions and 1.6 billion e-Transfers, with a record 149 million e-Transfers in October, and losses to fraud of about 4 cents per $100 transacted, according to its corporate year in review.

Payments Canada and the Real-Time Rail

The Real-Time Rail was first targeted for 2019. In mid-2026 the finance minister approved its by-law and rules, which came into force on August 24, 2026, and launch is expected in the fourth quarter of 2026, as BetaKit reported. Our overview of instant payment systems around the world shows how late Canada is.

Consumer-driven banking

Canada's open banking law is being built in two phases under Bank of Canada oversight. Read-only data sharing has no committed start date, while write access for payments targets mid-2027 and depends on the Real-Time Rail being in wide use, according to a timeline compiled by Banq. India's approach, covered in our account aggregator explainer, offers a useful contrast.

How do the leading Canadian fintech companies compare?

As of October 2026, Canada's fintech leaders range from Shopify, processing tens of billions of US dollars a quarter, to private companies valued between roughly US$500 million and C$10 billion. The table lines up each company's latest dated metric.

What is a Canadian fintech customer worth? A worked example

Dividing valuation by customers shows investors in 2026 value a Wealthsimple client at roughly C$3,300 and a KOHO member at roughly C$530. The gap reflects what each customer brings: Wealthsimple clients hold an average of about C$33,000 in assets, while KOHO earns mainly on spending and subscriptions.

The arithmetic is simple. Wealthsimple: C$10 billion divided by 3 million clients is about C$3,333 per client, and C$100 billion of assets over 3 million clients is about C$33,333 each. KOHO: C$1.33 billion divided by 2.5 million members is about C$532 each. Neo: its reported US$510 million valuation over one million customers is roughly US$510 each, close to KOHO once currency is considered.

Now take the subcontractor from the opening, sending C$1,000 to his helpers each week by e-Transfer. At Interac's 2025 fraud rate of about 4 cents per $100, the expected system-wide fraud loss on that C$52,000 a year is around C$21. Low fraud is a big reason the incumbent network is hard to dislodge, and it sets the bar the Real-Time Rail and future open banking payments must clear.

Canada's fintechs have the customers. What they have lacked is the plumbing, and that arrives on someone else's timetable.

What should you watch in Canadian fintech next?

Three dated tests will shape Canadian fintech through 2027: whether the Real-Time Rail goes live in the fourth quarter of 2026 as planned, whether KOHO wins its federal banking licence, and whether open banking write access arrives by mid-2027. Each would widen what non-bank firms can offer against the Big Six banks.

• Q4 2026: Real-Time Rail launch. A further slip into 2027 would also push back open banking payments.
• Q4 2026: Clearco's target to reach break-even.
• Licences: KOHO's bank licence application, subject to ministerial approval, is the one to watch. Neo has not said whether it will seek one.
• Mid-2027: Target for the payments phase of consumer-driven banking under Bank of Canada oversight.
• Cross-border: Canadians sending money abroad will feel faster domestic rails first. Our guide to USA and Canada to India remittances covers that corridor.

For comparison with larger markets, see our lists of leading US fintech companies and UK fintech companies.

This list is editorial research, not investment advice or an endorsement. Company figures are as reported by the linked sources and may be restated.

A barista in Lisbon taps a card, the terminal beeps, and a few cents of euro land in the cafe's account the next day. Nothing about that moment tells you the customer was spending dollar stablecoins held in a crypto wallet. That invisibility is the whole design of the crypto card, and it explains why the product has quietly become one of the busiest corners of the digital asset economy.

The card is ordinary. The account behind it is not

Strip away the branding and a crypto card is a regular debit or prepaid card running on Visa or Mastercard. The merchant sees a normal card transaction in local currency. What differs is the funding source. Instead of a checking account at a bank, the balance sits in stablecoins or other crypto assets, either with a custodial provider or in the customer's own wallet.

Several companies sit in the chain. A card programme usually involves an issuing bank or licensed principal member of the card network, a programme manager that runs the app and the customer relationship, a processor that handles authorisations, and a crypto custodian or wallet provider holding the assets. Some infrastructure firms bundle several of those roles. Rain, for example, is a Visa principal member, and The Block reported in September 2026 that it had financed roughly $2 billion of card settlement through an onchain credit facility since 2023.

What happens in the two seconds at the till

When the card is tapped, the network sends an authorisation request to the issuer's processor, just as it would for any debit card. The processor checks whether the linked crypto balance can cover the amount at the current rate. If it can, the matching amount of stablecoin or crypto is reserved or sold, the authorisation is approved and the merchant gets paid in local currency through normal card settlement.

There are two common models:

• Custodial cards. The provider holds your assets and converts them at the moment of purchase. This is fast and simple, but you are trusting a company with your money.
• Self-custodial cards. Funds stay in your own wallet, and you approve a smart contract that lets the card programme pull a set amount when you spend. You keep control, but you take on the job of protecting your keys. We cover the trade-off in more detail in our guide to self-custody versus custodial wallets.

Behind the scenes, the card programme still owes the network money every day. Historically that meant converting crypto into dollars or euros to settle. Increasingly, networks accept settlement in stablecoins directly. Visa said its stablecoin settlement had passed a $20 billion annualised run rate, more than 15 times the year before, according to The Block in September 2026.

How big has this become?

Card spending is one of the few stablecoin use cases where ordinary consumer behaviour shows up clearly in the data. Analytics firm Artemis estimated in January 2026 that annualised crypto card spending had passed $18 billion, putting it within reach of peer-to-peer stablecoin transfers, as reported by The Defiant.

The smaller, fully onchain cards show the same curve. Dune Analytics data cited by Cointelegraph showed monthly net spend across six self-custodial Visa cards rising from $14.6 million in January 2025 to $91.3 million in December, with the EtherFi card alone accounting for $55.4 million over the year. Visa told investors it had more than 160 stablecoin-linked card programmes live in its fiscal second quarter of 2026.

Mastercard has been moving too. In March 2026 it launched a Crypto Partner Program with more than 85 companies, including exchanges, wallet providers and stablecoin issuers, aimed at settlement, payouts and consumer payments.

Where the fees hide

Crypto cards rarely advertise a single headline charge, because the cost is spread across several layers. When comparing programmes, look at each of these separately:

• Conversion spread. If you hold bitcoin or ether, the card sells it at a rate that may sit above or below the market price. With dollar stablecoins this spread is usually thin for dollar purchases.
• Foreign exchange. Paying in euros from a dollar stablecoin balance involves a currency conversion, and the rate used matters as much as any stated fee. Our explainer on the mid-market rate shows how to spot the markup.
• Network and onchain costs. Self-custodial cards may pass on blockchain transaction fees for top-ups or settlement.
• Card fees. Issuance, monthly, inactivity and ATM charges vary widely.

Rewards can offset some of this, but cashback paid in a volatile token is not the same as cashback paid in dollars.

A crypto card is only as cheap as its worst conversion, so price the whole journey, not the advertised fee.

The rules now apply to the card, and the coin

Because the card leg runs on Visa and Mastercard, it is regulated like any card product: the issuer needs a bank charter or e-money licence, and the programme must run identity checks and anti-money-laundering controls. What changed in 2025 and 2026 is that the stablecoin leg now has its own rulebook in the largest markets.

In the United States, the GENIUS Act became law on 18 July 2025, requiring payment stablecoin issuers to hold at least one dollar of permitted reserves for every dollar issued and barring them from paying interest to holders, according to the Congressional Research Service. In the European Union, MiCA treats euro and dollar stablecoins as e-money tokens; by July 2026, 21 authorised issuers across 12 member states were issuing 35 such tokens. Our overview of GENIUS and MiCA goes deeper.

Tax is the piece most users underestimate. In many countries, swapping crypto for goods is treated as a disposal, so a card purchase funded by bitcoin can create a taxable gain or loss. Spending a dollar stablecoin usually produces little or no gain, but the record-keeping obligation can still apply.

Our view: the card is a bridge, not the destination

Crypto cards succeed because they ask nothing of the merchant. That is also their limit. Every purchase still passes through interchange, network fees and an issuing bank, so the cost structure looks a lot like the card system it rides on. The more interesting shift is happening underneath, as networks start settling in stablecoins and programme managers fund cards from onchain credit. If that plumbing keeps maturing, the card may end up as the familiar front end for money that moves very differently behind the counter.

This article is for information only and is not financial or tax advice. Card features and fees change often; check current terms with any provider.

A software developer in Cordoba invoices a client in Austin every month. Five years ago she would have waited for a wire, then watched the official exchange rate take a slice before the pesos reached her account. Today she is paid in dollar stablecoins, keeps most of the balance in that form, and converts only what she needs for rent and groceries. Multiply her by a few million people across Argentina, Brazil, Venezuela, Colombia and Mexico, and you get the most stablecoin-intensive region on earth. Chainalysis found that stablecoins accounted for more than half of local-currency exchange purchases in both Argentina and Brazil between July 2024 and June 2025.

This piece looks at what the data says about each major market, why the drivers differ from country to country, and how stablecoins are now competing with, and connecting to, Brazil's Pix and the region's remittance channels.

The regional numbers

Chainalysis's 2025 Latin America report recorded nearly $1.5 trillion in crypto transaction volume across the region from July 2022 to June 2025. Monthly volume climbed from $20.8 billion in July 2022 to a record $87.7 billion in December 2024, before easing to $47.9 billion in June 2025. That peak lined up with the global crypto rally at the end of 2024, a reminder that some regional activity is speculation rather than payments.

The country breakdown is striking. Brazil received $318.8 billion, nearly a third of regional activity, after growth of almost 110 percent. Argentina came second with $93.9 billion, followed by Mexico at $71.2 billion, Venezuela at $44.6 billion and Colombia at $44.2 billion.

Argentina: dollarisation by app

Argentines have saved in dollars for generations, in cash under the mattress or in safe deposit boxes. Stablecoins are the digital continuation of that habit. They offer a dollar balance without a US bank account, accessible from a phone, and usable for payments at a growing number of local merchants through apps that convert at the point of sale.

Inflation has fallen sharply but remains high by any international standard. Argentina's statistics agency INDEC reported consumer prices up 31.8 percent in the year to September 2025, with a monthly rise of 2.1 percent. That is a big improvement on recent triple-digit years, but it still means peso savings that earn nothing lose a quarter of their purchasing power in a year. Chainalysis cites persistent inflation, currency volatility and capital controls as the forces behind regional stablecoin demand, and Argentina is the clearest case of all three.

In Argentina, a stablecoin is not a crypto bet. It is a savings account in the only currency people trust.

Brazil: the business and investment market

Brazil is a different story. The real is volatile but not in crisis, and Brazilians have one of the best domestic payment systems in the world in Pix. Yet Brazil is by far the largest crypto market in Latin America, and officials have said more than 90 percent of the flows are stablecoin-related.

Why? Mostly cross-border and business use. Stablecoins give Brazilian companies and investors a faster, cheaper way to move dollars for imports, offshore investments and international services than traditional FX contracts. That is also why the Banco Central do Brasil has paid close attention. Under the Brazilian Virtual Assets Law, the central bank became the AML authority for crypto firms, and its consultations 109, 110 and 111 of 2024 set out to bring crypto service providers, including stablecoin transactions with an FX character, into a licensed framework. Chainalysis expected final rules by the end of 2025. The direction is clear: Brazil wants stablecoin flows inside its regulated FX system, not beside it.

Pix and stablecoins are complements more than rivals. Pix handles domestic real payments, nearly 280 million on a record day in June 2025, with more than 175 million users. Stablecoins handle the dollar leg. Most exchanges in Brazil use Pix as the on-ramp and off-ramp, so a user can move from reais to a dollar token and back within minutes. See our instant payments comparison for how Pix stacks up globally.

Mexico: remittances meet stablecoins

Mexico's crypto story is tied to the United States. It is one of the world's largest remittance recipients, receiving $61.8 billion in 2025 according to Banco de Mexico data reported by El Imparcial, down 4.6 percent from 2024. Local exchanges have built remittance services that take dollars in the US, move them as stablecoins, and pay out pesos in Mexico through the domestic SPEI system.

Chainalysis describes Mexican activity as relatively stable rather than booming. That makes sense. The peso has been comparatively strong, so the savings motive that drives Argentina is weaker. Stablecoins in Mexico are mainly a payments tool, competing with money transfer operators on the US to Mexico corridor, which we cover in detail in the US-Mexico remittance corridor.

Venezuela and Colombia: two kinds of need

Venezuela shows the extreme version of stablecoin adoption. After years of extreme inflation and restricted access to the international banking system, USDT functions as everyday money for many businesses, used to price goods, pay suppliers and hold savings. The $44.6 billion received is large relative to the size of the economy.

Colombia, at $44.2 billion, is closer to the regional middle: a stable but depreciating currency, large inflows from family members abroad and growing retail interest in holding digital dollars. Together they show that stablecoin adoption is not one phenomenon. It is a response to different failures in local money and banking.

That matters for anyone building or regulating in the region. A product designed for Argentine savers, who want to hold dollars for months, looks nothing like one designed for Mexican remittance recipients, who want pesos within minutes, or for Brazilian importers, who need a dollar leg for a single invoice. The data from Chainalysis mixes all three into a single headline figure. Read country by country, the region is really three separate stablecoin markets sharing a currency peg. For a primer on how the tokens themselves work, see USDT vs USDC.

A worked example: one year of savings in Argentina

Here is a simple illustration using INDEC's inflation figure. It is not investment advice, and the outcome for any real saver depends on exchange rate moves and fees.

1. A saver holds ARS 1,000,000 in a non-interest current account for the 12 months to September 2025.
2. With consumer prices up 31.8 percent, the money's purchasing power at the end is 1,000,000 divided by 1.318, about ARS 759,000 in starting-year prices. She has lost roughly 24 percent in real terms.
3. If instead she had converted to a dollar stablecoin at the start, her balance would hold its dollar value. Whether that beats the peso in local purchasing power depends on how much the peso depreciated against the dollar over the same year, and on US inflation.
4. She also faces costs: the exchange spread when buying, any fee for converting back, and the risks of the platform and the token itself, which we discuss in a history of stablecoin depegs.

The example shows why demand is so persistent. Even in a year of falling inflation, idle peso balances lost about a quarter of their value. When the alternative is that certain loss, a few percentage points of conversion cost and some platform risk look acceptable to many people.

Risks and the policy response

Mass stablecoin use creates real problems for central banks. Moody's has warned about cryptoization, where dollar tokens weaken a central bank's control over rates and the exchange rate and pull deposits out of local banks. Latin America is among the regions where Moody's sees the fastest adoption. We examine the deposit question in the stablecoins and bank deposits debate.

There is also concentration risk. USDT, with around $184 billion in circulation as of October 2026 and about 60 percent of the market, is the token most associated with everyday use in Argentina and Venezuela. USDC follows at about $73 billion. Total stablecoin supply was about $306 billion as of October 2026, so two issuers account for the large majority of the digital dollars Latin Americans hold. A problem at either issuer would hit Latin American savers harder than almost anyone else, because for many of them the token is not a trade but their savings.

What to watch next

• Brazil, 2026-2027: how the central bank's crypto and FX rules are applied in practice, especially reporting for stablecoin transfers to self-custody wallets.
• Monthly INDEC releases: Argentina's inflation path is the main variable for peso-to-stablecoin demand.
• Monthly Banxico data: whether the US-Mexico corridor keeps shrinking, and how much of it shifts to digital routes after the US 1 percent tax on cash-funded remittances took effect in January 2026.
• Late 2026: Chainalysis's next geography report, which will show whether Brazil's growth continued after the 2024 peak.
• Ongoing: regional rules for exchanges and issuers. Expect more licensing regimes for exchanges and issuers across the region.

This article explains market data and regulation. It is not financial or tax advice; rules on holding and converting digital assets vary by country, so check with a qualified adviser.

A grandmother in a village outside Sylhet gets a text from her operator: her balance has been topped up. Her grandson sent it from a bus depot in Doha, between shifts, and it took him less time than buying tea. No bank was involved, no exchange house queue, no account number. International mobile recharge is the smallest and most frequent form of cross-border money, and for millions of families it is how support arrives between the bigger monthly transfers. It is also a business with a wholesale layer that almost no customer ever sees.

What is international mobile recharge?

International mobile recharge, also called airtime top-up, is the purchase of prepaid credit or a data bundle for a phone number in another country. The sender pays in their own currency; the recipient's operator credits the line in local currency, usually within seconds. In 2026 it is sold by money transfer apps, telecom retailers and specialist top-up apps, almost all of which rely on a few wholesale aggregators.

The product exists because most phones in developing countries run on prepaid plans. India alone had 1,282.38 million mobile subscriptions at the end of June 2026, according to TRAI's quarterly performance indicators, and the regulator tracks prepaid as its own segment, with prepaid users averaging 1,055 minutes of calls a month. We looked for a current official prepaid share of connections for India, Sub-Saharan Africa and the Philippines; the TRAI and GSMA publications we checked report prepaid revenue and usage but not a subscriber split, so we do not quote a percentage. Prepaid users buy credit in small amounts, often weekly, which is exactly the type of payment that migrant relatives are well placed to cover.

The top-up itself is not a money transfer in the normal sense. The recipient cannot withdraw it as cash. What arrives is a telecom service, which is why top-ups often sit outside the paperwork and identity checks that apply to remittances. That simplicity is the appeal, and also the reason for some of the risks below.

How does an international top-up actually work?

An international top-up passes through three layers: the retail app or website where the sender pays, a wholesale aggregator that holds connections to hundreds of operators, and the destination operator that credits the line. The aggregator buys airtime at a discount, the retailer adds its margin, and the operator receives settlement in local currency.

1. Sender: picks a country and enters the number. The app usually detects the operator and shows a list of fixed amounts or data plans.
2. Retailer: charges a card, bank transfer or wallet in the sender's currency, then calls an aggregator's API.
3. Aggregator: routes the order to the right operator, either through a direct agreement or a local distributor, and settles with it.
4. Operator: credits the prepaid line and sends the recipient a text. Some products deliver a PIN that has to be redeemed instead.

Delivery is quick. Western Union's top-up page, operated with Ding, says credit usually arrives a few seconds after payment, in a description on its UK site. Postpaid bills are a different product: they need a biller connection, not an airtime API, and India has its own bill-payment network, which we cover in Bharat Connect explained.

Who are the wholesale players behind mobile top-ups?

The main wholesale aggregators for international mobile recharge in 2026 are DT One, Ding and Reloadly, with smaller players such as Prepay Nation. DT One, the former TransferTo, says it connects to more than 900 operators and billers in 180 countries. Ding's network is described as 700-plus operators on Western Union's site. Reloadly says it reaches over 800 operators. All of these are company claims.

DT One and the TransferTo story

The history explains how the industry is structured. TransferTo began as a mobile top-up business in the mid-2000s and later added cross-border payments. In February 2019 it split into two companies: DT One kept airtime and rewards, with more than 1,000 partners in 160 countries at the time, and Thunes became a payments network. Today DT One's network page lists more than 900 operators and billers in 180 countries and over 23,000 products, from airtime to eSIMs and prepaid utility vouchers. Private equity firm Apis Partners invested in May 2022 and names MoneyGram, Western Union and WorldRemit among the money transfer firms it works with.

Ding

Ding, founded in Dublin in 2006, runs its own consumer app and supplies other brands. When Pollen Street Capital bought a majority stake in September 2021, Ding said it had sent almost half a billion top-ups, according to the deal announcement. In 2023 it began powering Western Union's top-ups in Canada, and in December 2024 Intermex launched top-ups to more than 130 countries through Ding, whose network then included over 600 operators, per Intermex's release. Western Union's 2026 page describes a network of more than 700 operators in 150-plus countries.

Reloadly

Reloadly, founded in Barcelona in 2018, sells a developer API rather than a consumer app. Its product pages in 2026 claim more than 800 operators, with country counts of 140-plus in one place and 170-plus in another, and it lets developers send local-currency amounts in some African markets. Prepay Nation, based in Berwyn, Pennsylvania, says it operates in more than 150 countries; in June 2026 it added Airtel Malawi top-ups and data bundles to its partner network, according to a company release.

What does an international top-up really cost?

An international top-up costs the face value of the credit plus the retailer's margin and any exchange-rate markup, which are usually bundled into one price. Reloadly's own guidance to resellers in 2020 described typical margins of 5% to 10%, with bulk discounts of 6% to 15% from operators. Fees are rarely shown separately, so the true cost is the gap between what you pay and the local value received.

Worked example: a $10 top-up paid in euros

Reloadly's developer documentation shows a sample order from a euro-paying sender to a dollar-denominated operator at an exchange rate of 1.10278, with fixed local amounts such as $5, $10 and $20. Use those figures:

• A $10 top-up at 1.10278 is worth about €9.07 at that rate.
• If a retailer charges €10.00 for it, the sender pays €0.93 more than the credit is worth, a cost of about 9.3%.
• At a 5% margin the price would be about €9.52; at 10%, about €9.98.

Compare that with the World Bank's global average of 6.36% to send $200 in Q3 2025. A small top-up is often more expensive in percentage terms than a cash transfer, which is the price of convenience on a very small amount. The easiest check is to compare the local credit a recipient gets with what the operator sells the same credit for at home.

Taxes can shave the amount further. Reloadly gives a Nigerian example in which a tax is deducted from a ₦20 top-up; at 1%, the line would receive ₦19.80. Local-currency products that are treated as domestic sales may avoid that deduction, depending on the country.

How do people recharge an Indian mobile from the UAE, USA or UK?

To recharge an Indian prepaid mobile from abroad in 2026, a sender uses a remittance or top-up app that supports Indian operators, enters the +91 number, picks a plan and pays by card or bank transfer. The plan is credited in rupees within seconds. Many apps show Indian operators' bundled data plans rather than plain talk-time.

The corridors that matter most mirror remittances. India received $135.46 billion of remittances in 2024-25, and the US and UAE were the two largest sources in RBI's 2023-24 survey, at 27.7% and 19.2%. Top-ups are a rounding error against those sums, but they are frequent, which is why money transfer apps bundle them: a customer who recharges a parent's phone every week opens the app far more often than one who sends money monthly. For bigger household costs, our guide to bill payments for NRIs covers electricity, insurance and school fees.

The same pattern holds for the Philippines, where overseas workers top up family lines alongside regular transfers (see Philippines remittances), and across Africa, where airtime and mobile money sit side by side (see M-Pesa and mobile money).

What top-up scams should senders watch for?

The most common mobile recharge scams in 2026 are fake websites offering heavily discounted credit, messages asking a relative to top up an unfamiliar number in an emergency, and requests from strangers to pay in airtime because it is fast and hard to reverse. A top-up sent to the wrong number generally cannot be pulled back.

• Too-good discounts: a site selling credit well below face value is often taking card details, not delivering airtime.
• "New number" messages: a text claiming a family member has changed phones and needs credit urgently. Call the old number first.
• Airtime as payment: fraudsters ask for top-ups because they can resell credit quickly and there is no chargeback route through the operator.
• Wrong digits: operators can deliver credit in seconds, and reversals depend on the retailer's policy, which may offer none.

Our explainer on money transfer scams covers the wider patterns.

How is international mobile top-up regulated?

In most countries international mobile top-up is regulated as the sale of a telecom service, not as a money transfer, so sellers usually do not need a remittance licence. Operators set the terms, consumer and tax rules apply, and the card or bank payment the sender makes is regulated as normal. Rules tighten when airtime starts to behave like money.

That last point matters. Where credit can be transferred between users or cashed out, it begins to look like stored value, which central banks oversee. Mobile money began in exactly that space, and regulators separated it from airtime by licensing e-money. Two concrete examples show where the lines sit. In the EU, Article 3(l) of the second Payment Services Directive, PSD2, takes payments made through a telecom provider for digital content and voice-based services out of payments regulation only while each transaction stays at or below EUR 50 and a subscriber's monthly total at or below EUR 300; the UK's payment rules were built on the same directive. In the US, FinCEN's 2011 prepaid access rule lists mobile phones among prepaid devices and puts some providers and sellers under anti-money-laundering duties, while exempting products of $1,000 or less and closed-loop products of $2,000 or less. Whether a given top-up product falls inside either regime depends on its design. Taxes are also local: a country can levy duty on airtime sold from abroad differently from a domestic sale, which is why some aggregators push local-currency products. Senders should check the terms of the service they use; this is general information, not legal advice.

A top-up is the only cross-border payment most families receive without filling in a single form.

What to watch next

• Data over talk-time: DT One now sells eSIMs and data alongside airtime, and recipients increasingly want data bundles. Expect more apps to lead with data plans through 2027.
• Stablecoin wallets in the same apps: Sendwave launched a USDC-based dollar wallet in October 2025, and Thunes enabled payouts to stablecoin wallets the same month. Top-ups could become one more thing paid from a digital dollar balance.
• Bundled bills: DT One already lists prepaid utility vouchers for electricity and water. The line between top-up apps and cross-border bill payment is blurring, as we explore in paying utility bills across borders.
• Money transfer app bundling: as Remitly, Western Union and others fight for daily engagement in 2026, top-ups are one of the cheapest ways to keep customers opening the app.

This article explains how international mobile recharge works and does not recommend any specific app or provider. Figures are as reported by the linked sources; vendor claims are labelled as such.

Hand a stablecoin issuer one dollar and it hands you back one token. It then buys a US Treasury bill with your dollar and keeps the interest. You get nothing extra, by design and, in the United States, by law. Multiply that by a few hundred billion dollars and you have one of the most profitable business models in finance. Yet the two biggest issuers, running nearly identical models, end up with very different results.

The basic model: borrow at zero, lend to the government

A fully reserved stablecoin is economically a deposit that pays no interest. The issuer holds cash, short-dated Treasury bills and overnight repurchase agreements against every token. When short-term rates sit around 3.5%, each billion dollars of tokens in circulation earns tens of millions a year in interest.

Holders cannot claim a share. America's GENIUS Act bars licensed issuers from paying interest or yield to holders simply for holding the coin, a point we cover in our piece on yield-bearing stablecoins. That leaves the issuer with two levers: grow the float, and decide how much of the interest to share with whoever helps it grow.

That second lever is the whole story of how Tether and Circle diverged.

Tether: the money machine that keeps most of it

Tether reported more than $10 billion of net profit for 2025, according to its year-end attestation prepared by BDO. That was down about 23% from roughly $13 billion in 2024 as interest rates fell, but it remains an extraordinary sum for a company of its size. At the end of 2025 it held more than $122 billion of US Treasuries directly and over $141 billion including indirect exposure such as reverse repos, against about $186 billion of USDT in circulation.

The quarterly figures for 2026 show the same engine running a little cooler. According to its second-quarter attestation, released on 31 July 2026, Tether earned about $1.5 billion in the quarter and held about $115 billion of Treasury bills. Its reserves also include more than 146 tonnes of gold, valued near $18.8 billion, and roughly $5.8 billion of bitcoin. Falling prices for those two assets helped cut its excess reserve buffer from $8.2 billion in March to about $4.1 billion in June, as reported by Stablecoin Insider.

Why does Tether keep so much? Because it does not pay for distribution. USDT became the default dollar on Tron and on many offshore exchanges years ago, and that position renews itself. Nobody needs to be paid to list it.

The trade-off is risk appetite. Gold and bitcoin in the reserve mix add upside but also volatility, which is part of why S&P Global cut its assessment of USDT's ability to hold its peg to "weak" in November 2025. Tether disputed the rating. Investors have also been cautious about pricing the business: reports in early 2026 said Tether had scaled back fundraising talks from as much as $20 billion at a $500 billion valuation to about $5 billion after pushback, according to The Block.

Circle: a great business that rents its shelf space

Circle's model starts the same way. In 2024 it booked about $1.68 billion of revenue, more than 99% of it from interest on USDC reserves. The difference is what happened next: it paid Coinbase about $908 million that year.

The arrangement, disclosed in Circle's IPO filing, gives Coinbase all the reserve income on USDC held on its own platform plus half of the residual income on USDC held elsewhere. Circle confirmed in August 2026 that the agreement had renewed on existing terms through 2029. Payments to distributors took about 40% of Circle's revenue in 2022, about half in 2023 and more than 60% in 2024.

Scale has not changed the ratio much. For the second quarter of 2026, Circle reported $701 million of total revenue and reserve income, of which $668 million came from reserves. Distribution, transaction and other costs were $412 million, or about 59% of revenue, and net income from continuing operations was $48 million. For the full year 2025, revenue reached $2.7 billion according to Circle's results release.

A stablecoin issuer is a bank with no interest bill, until it has to pay someone to put its coin on the shelf.

The rate problem

Both businesses are leveraged to one variable they do not control: the short-term interest rate. Circle's reserve return rate fell to 3.48% in the second quarter of 2026, down 66 basis points from a year earlier.

Some rough arithmetic makes the exposure concrete. On about $73 billion of reserves, every quarter-point cut in rates removes roughly $180 million of annual interest income before any distribution payments. On Tether's larger Treasury book the hit is bigger in dollars, although its lower payout ratio means it keeps a larger share of what remains.

That is why both companies talk constantly about other revenue. Circle reported $34 million of other revenue in the second quarter of 2026 from subscriptions, services and transaction fees, and it launched its own Arc blockchain in September. Tether holds more than $20 billion of proprietary investments outside the USDT reserves. Neither has yet built a second engine anywhere near the size of the first.

The new entrants are copying Coinbase, not Tether

The most telling development of 2026 is that newcomers are designing around Circle's problem by giving distributors equity instead of a revenue cut. Open USD, which launched on 30 September 2026, is owned in equal stakes by Visa, Mastercard, Stripe, Shopify and Coinbase, with partners sharing most of the reserve income and the issuer keeping a small management fee. Bank consortia planning their own tokens are following similar logic.

The lesson is that reserve income is a commodity. Any regulated issuer can buy Treasury bills. What is scarce is the place where people actually hold and spend the dollar token, whether that is an exchange, a wallet, a checkout page or a payroll platform. Whoever controls that spot will capture most of the interest, whatever name is on the token. We look at how public investors are valuing this in our analysis of Circle's first 16 months as a listed company, and at the two leading coins in USDT vs USDC.

What it means for everyday users

For someone sending money from Kuwait City to Kollam, none of this shows up as a line item. But it explains three things. First, why issuers fight so hard for listings and partnerships. Second, why some apps offer "rewards" on stablecoin balances: they are often passing on part of a distribution payment, which regulators are now scrutinising. Third, why fees on the user side can fall toward zero: when the float itself earns 3% to 4%, the transfer can be priced as a loss leader.

Company figures are taken from issuer attestations, results releases and press reports cited above. This article is not investment advice.

On 15 September 2026, the US Senate voted 49 to 50 on whether to move ahead with the CLARITY Act. The bill needed 60. Within days, the chairs of the Securities and Exchange Commission and the Commodity Futures Trading Commission were promising to write the rules themselves. That vote, and what followed it, tells you most of what you need to know about American crypto regulation in 2026: Congress has written the outline, and the agencies are now drawing inside it without waiting for a signature.

This piece explains what the market structure bill would do, how far it got, the fault lines that stopped it, and what actually changes for exchanges, token issuers and users while the law sits unfinished.

What "market structure" means and why it matters

The US already has a stablecoin law. The GENIUS Act cleared the House 308 to 122 on 17 July 2025 and was signed the next day, as The Block reported. That law governs who may issue a payment stablecoin and what reserves must back it. Our guide to GENIUS Act and MiCA rules covers it in detail.

Market structure is the bigger, messier question: for every other token, who is the regulator? If a token is a security, the SEC governs its sale and the venues that trade it. If it is a commodity, the CFTC polices fraud and manipulation, but until recently had no direct authority over spot trading at all. For a decade that ambiguity was settled by lawsuits rather than rules. The CLARITY Act tries to replace case-by-case enforcement with a statute that sorts tokens into categories and assigns each one a regulator.

What the House passed in July 2025

The House approved the CLARITY Act on 17 July 2025 by 294 to 134, with every Republican and dozens of Democrats in favour. According to The Block's explainer, the bill divides digital assets into categories including digital commodities, stablecoins and tokenised securities, with investment contracts still handled under SEC rules. It sets benchmarks for when a network is decentralised enough that its token stops being treated as a security, and treats DeFi protocols as infrastructure rather than intermediaries that must register.

For centralised platforms the House text required federal registration, retail disclosures and segregation of customer funds from company money. One concern raised at the time: decentralised exchanges would not get the same federal pre-emption over state law, leaving them exposed to 50 state regimes.

The Senate's version and the three fights

The Senate split the work between two committees. Agriculture, which oversees the CFTC, advanced its draft in January 2026 on party lines. Banking took longer, stalled for months over stablecoin rewards, then voted 15 to 9 on 14 May 2026 to advance its bill, with Democratic senators Ruben Gallego and Angela Alsobrooks supplying the decisive votes, per The Block. More than 100 amendments had been filed.

Three disputes then dominated the summer.

• Stablecoin rewards. Banks argue that exchanges paying "rewards" on stablecoin balances are offering deposit interest without being banks, and that this will drain community bank deposits. The compromise text bans passive yield paid by issuers and platforms while allowing activity-based incentives, and the final September draft added an 18-month "circuit breaker" the Treasury Secretary could trigger if stablecoins cause substantial deposit outflows. Bank groups said that kicks in too late. Our explainer on the stablecoins and bank deposits debate sets out both sides.
• Ethics. Democrats wanted limits on senior officials, including the President, profiting from crypto ventures. A deal announced in July barred officials and their spouses from issuing or sponsoring digital assets, but left enforcement with the Justice Department. A later draft gave state attorneys general a role, which critics still called unenforceable.
• Developers and DeFi. The Blockchain Regulatory Certainty Act, folded into the bill, says non-custodial developers are not money transmitters. Changes to how it protects developers from criminal liability split even the bill's supporters.

The bill did not die over what a token is. It died over who gets to profit from one.

Worked example: the 60-vote arithmetic

Senate procedure means a bill like this needs 60 votes to end debate and proceed. Here is how the numbers looked on 15 September 2026, based on The Block's account of the vote.

1. The tally: 49 yes, 50 no. Senator Thom Tillis voted yes, then switched to no, a procedural move that allows him to file a motion to reconsider. Counting him, the bill had 50 supporters.
2. The gap: 60 minus 50 leaves 10 votes short, almost all of which would have to come from Democrats.
3. The swing bloc: the two Democrats who carried the bill through Banking, Gallego and Alsobrooks, had said they would not vote for it without adequate ethics language. Senator Mark Warner said the provision fell well short.
4. The calendar: even had the Senate passed it, the text differs from the House bill, so it must go back to the House, which was not expected to act until after the 3 November 2026 midterm elections. The day before the vote, TD Cowen put the odds of enactment in 2026 at 25%.

Compare that with the House margin of 294 to 134, or about 69% in favour. The Senate result was roughly 49% of those voting. The policy content was broadly bipartisan; the politics of the moment were not.

Plan B: the agencies move without Congress

The more consequential story of 2026 may be what the regulators did while Congress argued. The SEC, under Chairman Paul Atkins, and the CFTC, under Chairman Michael Selig, who took office in December 2025, announced a joint effort called Project Crypto in January 2026, with Selig directing staff to work on joint codification of an SEC "token taxonomy" for deciding which tokens are securities, The Block reported. In March 2026 the two agencies issued guidance saying most cryptocurrencies are not securities.

Then came concrete rules:

• Listed spot crypto. Following September 2025 joint SEC-CFTC guidance, Bitnomial became the first CFTC-registered exchange to list leveraged spot crypto products in December 2025, per The Block.
• SEC innovation exemption. On 17 September 2026, two days after the Senate vote, the SEC released a five-year exemption that lets "tokenised securities venues" offer onchain trading of certain tokenised stocks without being defined as exchanges, and lets some automated liquidity providers avoid dealer registration. Synthetic instruments are excluded and issuers can block their shares, as The Block detailed.
• CFTC Regulation CTX and CAM. On 5 October 2026 the CFTC launched rulemaking for a federal framework for leveraged and margined retail crypto trading, including a possible "crypto asset market" category of registered exchange, as an alternative to state licensing. It cannot force all spot crypto onto CFTC venues without Congress, according to The Block.

The SEC's July 2026 regulatory agenda also flagged amendments to broker-dealer net capital, customer protection and recordkeeping rules to address crypto assets, plus possible safe harbours for token offerings.

Law versus rulemaking: what each route delivers

That last row is the whole argument for legislation. JPMorgan analysts noted after the vote that agency rules are less durable than statute. A different SEC chair in 2029 could withdraw an exemption or reverse a taxonomy. For anyone building a business on these rules, that is a five-year horizon at best.

What changes for exchanges, tokens and users

Exchanges. US platforms still operate mostly under state money transmitter licences, which our guide to crypto licences explains. The CFTC's proposal offers a federal alternative for leveraged retail products. Without the CLARITY Act, there is no single federal licence for spot trading of non-security tokens.

Token issuers. The March 2026 guidance and the coming SEC taxonomy lower the risk that a widely used, decentralised token is treated as a security. Projects at an early, centralised stage still face securities rules, softened by any safe harbour the SEC finalises.

Tokenisation. The innovation exemption opens a regulated route for trading tokenised shares onchain, a development covered in our piece on real-world asset tokenisation.

Stablecoin holders. The fight over rewards is unresolved in law. Whether a platform can pay you for holding a dollar stablecoin depends on GENIUS Act implementing rules, explained in our guide to GENIUS Act implementation.

Our view: the agencies are delivering most of the practical clarity the industry asked for, faster than Congress could have. What they cannot deliver is permanence, state pre-emption or developer protection. Those three are the real stakes of the bill now.

What to watch next

• 3 November 2026: US midterm elections. The House cannot take up a revised bill before then, and a change in control would reset the negotiation.
• Late 2026: whether Senator Tillis's motion to reconsider produces a second floor vote in the lame-duck session.
• 12 November 2026 onward: EU direct supervision of USDC's reserves begins, a reminder that other jurisdictions are pressing ahead with stablecoin oversight.
• Through 2026-27: final text of CFTC Regulation CTX and CAM, SEC broker-dealer crypto amendments, and any SEC token offering safe harbour.
• 2028: analysts at TD Cowen have warned the CFTC framework may need to be completed before the 2028 election to survive a change of administration.

This article explains legislation and regulation in general terms and is not legal advice.

A family office partner in Abu Dhabi has two fintech decks open side by side. Both show $50 million a month of payment volume. One startup keeps 0.2% of it after paying banks, card networks and FX providers; the other keeps 1.1% but has written off 3% of the loans it extended last year. The headline number is identical and almost meaningless. What separates an investable fintech from a crowded one in 2026 is what happens to each dollar after it arrives, how long the customer stays, and whether a regulator would let a competitor copy the business next month.

How much money is going into fintech startups in 2026?

Fintech funding in 2026 is fewer, larger cheques. CB Insights counted $52.7 billion of fintech venture funding in 2025, up 35.5% on 2024, while deal count fell almost 19% to 3,631. In the second quarter of 2026, funding fell to $11.7 billion across 726 deals, the fewest quarterly deals in more than four years.

The 2025 totals come from CB Insights data reported by Let's Data Science, which put the average deal at about $20 million. The second-quarter figures are from CB Insights' State of Fintech Q2 2026, which also logged only four fintech IPOs in the quarter against 25 in the fourth quarter of 2025.

KPMG counts more broadly, including M&A and private equity. Its Pulse of Fintech H1 2026 recorded $103.1 billion across 2,100 deals in the first half, up from $72.2 billion in the second half of 2025, with one $24.3 billion payments acquisition doing much of the lifting. Venture capital alone was $31.5 billion across 1,641 deals. The US took $80.8 billion of the total; KPMG estimated India at around $2 billion. By segment, payments drew $44.2 billion in the half, more than double its $20.2 billion for all of 2025, while AI-related fintech drew $21.4 billion and digital assets $11.1 billion.

The lesson for founders is uncomfortable. Capital has not disappeared, but it is pooling in companies that already look like category leaders, such as those in our list of fintech startups to watch in 2026.

What do investors look at first in a fintech startup?

Investors in 2026 look first at revenue quality: how much of each transaction a fintech keeps after paying banks, networks, FX providers and credit losses. Gross volume and gross revenue impress in a pitch, but net revenue, sometimes called revenue less transaction costs, is the number that decides valuation multiples.

Three ratios do most of the work:

• Take rate. Revenue divided by volume processed. A card processor might earn a fraction of a percent; an instalment lender earns several percent because it also carries credit risk.
• Net take rate. Revenue after interchange, network fees, sponsor-bank fees, FX spread costs and funding costs, divided by volume. This shows who really holds pricing power.
• Contribution margin. Net revenue minus losses, fraud, customer support and servicing, before marketing and overheads. If this is negative at scale, more growth makes the business worse.

Lending and buy-now-pay-later models add a fourth: loss rates by vintage. Investors want charge-offs tracked by the month loans were made, not blended across a growing book, because a fast-growing lender can hide deteriorating credit behind new, unseasoned loans. Payments and stablecoin businesses face an equivalent in fraud and chargeback losses. Our explainer on how stablecoin issuers make money shows how interest-driven revenue can flatter margins when rates are high.

Worked example: reading a public fintech's numbers the way a VC would

Public filings let us test these ratios on real figures. Affirm, the US instalment lender, reported fiscal 2026 (the year to 30 June 2026) revenue of $4.26 billion on gross merchandise volume of $50.2 billion, according to Crowdfund Insider's results summary.

1. Gross take rate: $4.26 billion divided by $50.2 billion is about 8.5%. Affirm's own fiscal 2027 guidance assumes a revenue take rate near 8.5% of volume.
2. Net revenue share: In the fourth quarter, revenue was about $1.17 billion and revenue less transaction costs was $589 million. So roughly half of every revenue dollar went to funding, processing and loss provisions.
3. Net take rate: $589 million against $14.1 billion of quarterly volume is about 4.2%.
4. Engagement: 27.8 million active consumers made about 7.0 purchases each over the trailing year, up 20%, a signal investors use as a proxy for retention.

Now apply the same lens to a seed-stage cross-border payments startup. Suppose it processes $50 million a month at a 0.8% gross take rate, earning $400,000. If bank, FX and network costs consume 0.55% of volume, net revenue is $125,000, or 0.25%. Subtract $20,000 of fraud losses and $30,000 of support and compliance cost and contribution is $75,000 a month. An investor would then ask what it cost to acquire the customers producing it. If the startup spent $1.2 million on sales and marketing to win them, payback on contribution is 16 months. That is the number the term sheet is really about.

The seed-stage figures above are illustrative. The Affirm figures are as reported for fiscal 2026.

What unit economics benchmarks do VCs use for fintech?

For fintech startups in 2026, VCs commonly test customer acquisition cost payback, net revenue retention and contribution margin. Andreessen Horowitz's widely used guide puts typical CAC payback at 12 to 18 months, 6 to 12 months for small-business products and 18 to 24 months for enterprise sales. Fintechs with credit risk face stricter thresholds.

The a16z framework, first published in 2020 and still cited in diligence memos, calculates payback as sales and marketing spend divided by net new recurring revenue multiplied by gross margin. Fintech investors adapt it by using net revenue in place of subscription revenue, since a payments company's top line includes pass-through costs.

A fintech's gross volume tells you how busy it is; its net take rate tells you whether it is a business.

Why do licences count as a moat for fintech startups?

Licences count as a moat because they take years and real capital to obtain, and they decide which products a fintech can legally offer. In 2026 investors value fintechs holding their own payment, e-money, banking or virtual-asset licences more highly than those renting them from partners, because a partner can withdraw.

India illustrates the cost. The Reserve Bank of India's Payment Aggregator Directions of 15 September 2025 require a non-bank applicant to have a net worth of Rs 15 crore when it applies and Rs 25 crore by the end of its third financial year after authorisation, maintained on an ongoing basis. Merchant funds must sit in escrow at a scheduled commercial bank and be settled within T+1, and cross-border aggregators face a Rs 25 lakh per-transaction cap. A startup holding that authorisation has cleared a hurdle a newcomer cannot buy quickly.

The same logic shows up at scale. Airwallex holds more than 85 licences worldwide, and Stripe's stablecoin unit Bridge won conditional approval from the US Office of the Comptroller of the Currency in February 2026 to become a national bank. In the Gulf, crypto firms have spent years securing approvals from Abu Dhabi's ADGM, Dubai's VARA and the federal Securities and Commodities Authority. Our guide to crypto licences explains the main categories.

What are pre-seed and seed benchmarks for fintech in 2026?

Pre-seed rounds in 2026 are larger and priced higher than a year earlier. Carta data shows US startups raised $3.19 billion through more than 11,500 pre-seed SAFEs and notes in Q2 2026, with an average instrument of $276,000, up 27%. The median valuation cap on SAFEs above $2.5 million reached $35 million, up 40% year on year.

The figures are from Carta's State of Pre-Seed Q2 2026 and its analysis of SAFE valuation caps. In the second quarter, 91% of SAFEs on Carta were post-money, and 94% of post-money SAFEs in the first half carried a valuation cap. One caution for fintech founders: AI companies took 49% of pre-seed dollars in the first half of 2026, so the medians are pulled upward by a sector that is not theirs.

The bar for the next round has also risen. Research summarised in January 2026 put the median revenue of fintechs raising a Series A at about $4 million, four times the 2021 level, according to Crowdfund Insider. A seed-stage fintech raising today is being underwritten on whether it can reach that.

How do fintech investors in India and the UAE differ?

India and the UAE are both growing fintech markets, but their funding looks different in 2025-2026. Indian fintech raised about $2.4 billion in 2025, with seed funding down 40%, while the UAE raised a record $1.4 billion across all sectors in the first nine months of 2025, mostly from foreign investors.

In India, Tracxn data reported by IANS showed early-stage fintech funding up 78% to $1.2 billion in 2025, while seed funding fell 40% to $177 million and late-stage funding fell 26% to $1 billion. India ranked third globally behind the US and UK, with Bengaluru taking 42% of funding and Mumbai 29%. Indian investors tend to weigh RBI licensing and distribution through UPI heavily, and they price in the possibility that regulation can reshape a product category quickly.

In the UAE, MAGNiTT data reported by Enterprise showed venture funding up 188% to $1.4 billion in the first nine months of 2025 across 164 deals, with about 82% of capital coming from non-UAE investors. Fintech across the wider Middle East and North Africa raised $965 million over the same period, up 97%. Gulf investors often look for regional licences in several jurisdictions and a credible path into Saudi Arabia. Our UAE fintech companies and India fintech companies guides profile the leaders in each market.

What should fintech founders and investors watch next?

Several dated data points over the coming months will show whether 2026's concentration continues:

• CB Insights State of Fintech Q3 2026, expected later this autumn, will show whether Q2's four-year low in deal count was the bottom.
• Carta's Q3 2026 pre-seed and seed data will show whether rising SAFE caps hold outside AI.
• RBI payment aggregator authorisations. Non-banks had to apply by 31 December 2025 and wind down by 28 February 2026 if they did not comply; the list of authorised firms will shape Indian payments for years.
• KPMG Pulse of Fintech H2 2026, typically published early in the following year, will give the full-year total including M&A.
• Mastercard's acquisition of BVNK, expected to close by the end of 2026, will set a reference price for stablecoin infrastructure.

This article explains how professional investors commonly assess fintech companies. It is not investment, tax or legal advice. Anyone considering investing in a private company should take advice from a qualified adviser.

A treasury analyst at a mid-sized money transfer company in Houston has two vendor decks open. One, from Ripple, promises a licensed payout network, a regulated dollar stablecoin and a bank charter in progress. The other, from a Stellar-based partner, offers an open network where MoneyGram, PayPal's stablecoin and a few hundred local cash-out firms already sit. Both say "settles in seconds". Both cost a sliver of a cent per transaction on-chain. The choice she actually faces is not between two blockchains. It is between buying a service from a company and building on a public network run by a foundation.

Is Ripple the same as Stellar?

No. In 2026 Ripple is a private US company that sells payment, custody and treasury services and issues the RLUSD stablecoin; it is closely linked to the XRP Ledger but does not own it. Stellar is a separate public blockchain, supported by the nonprofit Stellar Development Foundation (SDF), whose native asset is XLM. The two share early history and a payments focus, but they are different networks.

The confusion is understandable. Both ledgers were designed more than a decade ago for moving money rather than running general-purpose applications, both use a form of voting consensus rather than mining, and both have native assets (XRP and XLM) used mainly to pay network fees and meet account reserves. The real difference is commercial. Ripple's business is selling a product to banks and payment firms. The SDF's job is to grow an open network that companies such as MoneyGram, Circle and Paxos build on directly.

That shapes everything below. When people compare "Ripple vs Stellar", they often mean Ripple Payments vs the Stellar network, which is a comparison between a vendor and an ecosystem. We cover Ripple's corporate strategy in more depth in Ripple in 2026 and Stellar's year in our Stellar explainer.

How do Ripple and Stellar compare for payments in 2026?

For payments in 2026, Ripple offers a managed service for licensed institutions, with its own stablecoin and more than $95 billion processed through Ripple Payments. Stellar offers an open network where regulated firms issue and redeem dollars directly, now carrying PayPal's PYUSD and MoneyGram's MGUSD. Fees on both ledgers are tiny; the cost sits in on- and off-ramps.

Two caveats. The "disclosed scale" row compares different things: Ripple publishes cumulative value through its product, while Stellar's figures come from on-chain activity. And on-chain fees are only a small part of what a customer pays; foreign exchange spreads, compliance and cash-out fees dominate.

What does Ripple Payments actually offer banks and fintechs?

Ripple Payments is a licensed cross-border service in 2026: Ripple takes funds in one currency and pays out in another, using RLUSD or XRP for settlement where useful and local partners for the last mile. Ripple said in December 2025 that it had processed more than $95 billion and that its payout network covered over 90% of the global FX market.

Those figures come from Ripple's announcement of conditional OCC approval for Ripple National Trust Bank, a federally supervised trust bank, on 12 December 2025. That approval matters more to a bank's procurement team than any throughput benchmark. It means RLUSD would sit under federal oversight from the Office of the Comptroller of the Currency as well as state oversight from the New York Department of Financial Services.

The customer experience is closer to using a correspondent bank than using a blockchain. A payment firm sends dollars to Ripple, Ripple handles conversion and routing, and a licensed partner credits the recipient. The firm does not need to hold XRP or run XRPL infrastructure itself. That is Ripple's pitch, and its limitation: you are buying a service with Ripple in the middle. For a step-by-step view, see our planned guide to how Ripple Payments works.

How do Stellar anchors, USDC and MoneyGram work?

On Stellar in 2026, regulated companies called anchors issue digital dollars or other assets and let users move between them and local cash or bank money. MoneyGram used USDC on Stellar for more than five years before launching its own MGUSD there in June 2026, and PayPal's PYUSD went live on Stellar in September 2025.

The anchor model means no single company sits in the middle. Circle issues USDC directly on Stellar. Paxos issues PYUSD, which the SDF announced as live on 18 September 2025, with early wallet integrations including Chipper Cash, Bitcoin.com and Lobstr. At the time, Stellar had nearly 10 million accounts and more than 20 billion operations to date.

MoneyGram is the clearest example of the model working at consumer scale. On 2 June 2026 it launched MGUSD, a dollar stablecoin issued by Stripe's Bridge under the GENIUS Act framework and running natively on Stellar, according to The Block. MoneyGram cited more than 60 million active customers and nearly 500,000 retail locations. The launch started in the US, with wider rollout planned. A family in Lagos or Manila receiving MGUSD can, in principle, collect cash at a MoneyGram counter, which is the off-ramp most blockchains lack.

Network activity has grown alongside. Nansen's Q2 2026 Stellar report counted about 4.9 million transactions a day on average, about 63,000 daily active addresses and more than $3 billion of tokenised real-world assets. Protocol 26, a network upgrade, went live on mainnet on 6 May 2026.

Which stablecoins run on the XRP Ledger and Stellar?

As of mid-October 2026, Ripple's RLUSD had about $2.4 billion in circulation, and in mid-September roughly 57% of it sat on Ethereum and 43% on the XRP Ledger. Stellar carries USDC, PayPal's PYUSD and MoneyGram's MGUSD. PYUSD, across all its chains, was about $2.9 billion, somewhat larger than RLUSD.

The chain split comes from Bitcoin.com News, citing DefiLlama on 14 September 2026: about $1.37 billion on Ethereum and $1.05 billion on the XRP Ledger, after most new supply since July went to Ethereum. Current totals come from DefiLlama. The split is revealing. Even Ripple's own stablecoin lives mostly on another chain, because that is where institutional liquidity and DeFi collateral demand sit. Both totals are small next to the overall stablecoin market of about $306 billion, of which USDT holds roughly 60% and USDC about $73 billion.

The practical point for a payments firm is availability, not market share. If a corridor partner only settles in USDC, Stellar and many other chains will do; if it settles in RLUSD through Ripple, the XRP Ledger or Ethereum are the options. Our comparison of the best blockchains for stablecoin payments covers how Ethereum, Solana, Tron and others fit in.

What does a payment actually cost on each network?

On-chain costs on both networks are far below a cent per payment in 2026. The XRP Ledger's minimum fee is 10 drops (0.00001 XRP), which is destroyed rather than paid to anyone, and Stellar's minimum is 100 stroops (0.00001 XLM) per operation, with ledgers closing in about five seconds. What users actually pay depends on exchange spreads and cash-out fees.

A worked example. A payout company sends 100,000 stablecoin payments a month on the XRP Ledger. At the reference transaction cost of 10 drops, the network fee for the whole month is 1,000,000 drops, exactly 1 XRP, though it can rise when the network is busy. Each recipient wallet must also hold a reserve: under the current XRPL reserve rules, 1 XRP base plus 0.2 XRP for the trust line needed to hold RLUSD, so 1.2 XRP locked per wallet. For a company onboarding 10,000 new recipients, that is 12,000 XRP of locked reserves, a balance-sheet item that often matters more than fees. Stellar has an equivalent rule: under its published reserve rules, an account must hold a 1 XLM minimum plus 0.5 XLM for each trustline, so 1.5 XLM per wallet holding USDC.

Now compare the whole journey. The World Bank's global average cost of sending $200 was 6.36% in Q3 2025, or $12.72. Even if the on-chain leg costs nothing, a sender still pays for converting dollars in, converting out to naira or pesos, and compliance checks. A blockchain can remove correspondent fees and delays; it cannot remove the cost of the cash counter at the end.

The ledger fee is the cheapest part of any blockchain payment. The off-ramp is where the money goes.

Have Ripple or Stellar been used for CBDCs?

Both have been used in central bank digital currency pilots, but neither runs a launched national CBDC as of 2026. Ripple said in 2023 it was working with Palau, Bhutan's Royal Monetary Authority and Montenegro's central bank on CBDC and stablecoin pilots. Stellar hosted Ukraine's electronic hryvnia pilot with TASCOMBANK and Bitt in December 2021.

Ripple's offering, reported by Finextra, is a private-ledger platform with tools for minting, distributing and redeeming tokens, plus a trial with Fubon Bank inside Hong Kong's e-HKD pilot programme. Stellar's approach used the public network: the Ukrainian pilot tested programmable payroll for public employees, peer-to-peer transfers and merchant payments under the supervision of the National Bank of Ukraine.

Context helps. The Atlantic Council tracker counted 146 countries exploring a CBDC, 41 pilot projects and just 3 launches as of May 2026. Most central banks building retail CBDCs have chosen their own or permissioned infrastructure rather than a public chain. Vendor pilots like these are best read as proofs of concept. Our CBDC tracker follows which projects have moved past that stage.

What to watch next for Ripple and Stellar

Over the next year, the questions are regulatory and commercial. Ripple needs final approval for its national trust bank and continued RLUSD growth; Stellar needs MGUSD and PYUSD to move beyond US launch markets. GENIUS Act implementing rules will apply to both stablecoin models.

• Ripple National Trust Bank. Conditional OCC approval came in December 2025. Final approval and opening would put RLUSD under direct federal supervision.
• RLUSD's chain split. Watch whether the XRP Ledger's 44% share rises. If Ripple's own coin stays mostly on Ethereum, the XRPL's role in its payment stack is narrower than its marketing suggests.
• MGUSD outside the US. MoneyGram said global scaling would follow the June 2026 US launch. The first non-US corridors will show whether stablecoin cash-out works at retail scale.
• GENIUS Act rules. The US stablecoin law, signed in July 2025, is still being turned into detailed regulations. They will shape RLUSD, PYUSD and MGUSD alike; see our GENIUS Act implementation guide.

For the analyst in Houston, the answer depends less on technology than on her company's appetite to run its own infrastructure. A firm that wants one contract and one counterparty will lean towards Ripple. A firm that wants to pick its own issuers, anchors and wallets will lean towards Stellar. Both are legitimate choices, and plenty of firms will use both.

This article explains how two payment networks work. It is not investment advice and says nothing about the price of XRP, XLM or any token. Check with a qualified adviser before making financial decisions.

Two years ago the stablecoin market was a two-horse race with a few ponies at the back. In 2026 the field looks more like a marathon start line. Payment companies, a remittance giant, banks and even blockchains themselves now issue their own digital dollars. Most won't overtake the big two. Several will matter anyway, because of who they're built for.

PYUSD: the checkout dollar

PayPal's stablecoin is the one most likely to reach people who have never opened a crypto app. Its advantage isn't technology; it's the hundreds of millions of accounts already sitting inside PayPal and Venmo. By late September 2026 it was a little ahead of Ripple's coin in size, roughly in the $2.7 billion range. Watch it for merchant payments and person-to-person transfers inside the PayPal world.

RLUSD: Ripple's institutional dollar

Ripple's stablecoin launched at the end of 2024 and has grown quickly with a very specific audience in mind: banks, payment providers and treasuries. Its supply passed about $2.49 billion at the end of September 2026, up more than 86% since January, split between Ethereum and Ripple's own XRP Ledger. Its reserves sit with a New York trust company, and it gained approval in Japan this year. Our Ripple deep dive covers where it fits in Ripple's bigger plan.

USAT: Tether goes American

Rather than reshape USDT for US rules, Tether launched a separate dollar in January 2026, issued through Anchorage Digital and designed around the GENIUS Act. It's a sensible way to keep USDT's global reach while giving US institutions a version they can legally hold. It also shows where the industry is heading: one brand, several regulated wrappers, each tuned to a jurisdiction.

EURC: the euro finally shows up

Almost every stablecoin is a dollar. Circle's EURC is the most visible exception, authorised under MiCA through the same French licence as USDC. It's still small, around €474 million in October 2026, but Europe's rules are deliberately nudging businesses towards euro-denominated tokens. For anyone paying salaries, suppliers or rent in euros, it's worth knowing it exists.

MGUSD: remittance gets its own coin

This one matters most to anyone who has ever queued to send money home. MoneyGram, a name generations of migrants know from the counter at their local shop, launched its own regulated digital dollar on the Stellar network in 2026. That's a remittance company admitting, in effect, that the future of sending money home runs on stablecoin rails, with cash pickup as one of the payout options rather than the whole business.

Bank coins and everything else

Banks in the US, Europe and Asia are testing their own tokenised deposits and stablecoins, and some newer coins pay a yield. We'd treat the yield-bearing ones carefully: under the GENIUS Act, ordinary payment stablecoins aren't allowed to pay you interest just for holding them. A coin built for savings and a coin built for spending are different products, and it's fine to want both, as long as you know which is which.

What this means for people who just want to pay

• More choice, less loyalty. No single coin will win every corridor, so providers will increasingly route between them.
• Local currencies are coming. Euro stablecoins are the start; expect more local-currency tokens as regulators get comfortable.
• Brands you already trust are moving in. PayPal and MoneyGram issuing stablecoins does more for mainstream trust than any crypto ad campaign.

The question is no longer whether stablecoins will be used for everyday money. It's which one moves your payment, and who decides.

We think the answer should be whichever coin gets the most money to the other side, with the final amount shown clearly before anyone pays. Nobody should have to read a guide like this just to pay their mother's electricity bill. (But we're glad you did.)

Somewhere in a district treasury office in Karnataka, a drawing and disbursing officer submits one of roughly 36.75 lakh bills the state processed in 2024-25. That bill, and the 36 lakh others like it, moved Rs 4.43 lakh crore out of a single state's accounts in a year. Multiply that machinery across 28 states and you get the least glamorous, most consequential payment system in India.

Most writing on Indian government payments starts with UPI or direct benefit transfer. Those are the parts citizens see. Underneath them sits the state treasury, the system that decides when money is released, which account it sits in while waiting, and whether the accounts reconcile at the end of the year. This explainer follows the money through that system using the most recent CAG, RBI and state data, and argues that the next round of reform is about float and reconciliation, not speed.

The scale: Rs 42.4 lakh crore and rising

The Comptroller and Auditor General's first decade-long compendium of state finances, released in September 2025, put combined state spending at Rs 42.4 lakh crore in 2022-23, up from Rs 15.9 lakh crore in 2013-14, as The Tribune reported. Uttar Pradesh spent about Rs 4.8 lakh crore and Maharashtra about Rs 4.7 lakh crore. Revenue spending, meaning salaries, pensions, interest, subsidies and grants rather than building things, made up 84.73% of the total, or Rs 35.96 lakh crore.

Within that, committed expenditure on salaries, pensions and interest reached about Rs 15.64 lakh crore, 2.49 times its 2013-14 level, according to Business Standard's reading of the CAG data. Subsidies were Rs 3.10 lakh crore and grants-in-aid Rs 11.26 lakh crore. Capital spending was just 15.27% of the total in 2022-23, against a decade peak of 19.68% in 2016-17.

The RBI's State Finances study published in January 2026 uses a narrower definition and so gets smaller ratios. For all states and union territories, it shows committed expenditure at 31.6% of revenue spending in 2023-24, 29.6% in the 2024-25 revised estimates and 30.6% in the 2025-26 budgets. Pensions alone took 11.8% of revenue spending in 2025-26 budgets and interest 12.2% of revenue receipts. States covered 57.3% of their revenue spending from their own revenue in the 2025-26 budgets. The rest arrives as devolution and grants from the Centre, which is exactly where the treasury plumbing gets complicated.

Three layers of plumbing

It helps to think of state money moving through three layers.

The state IFMS. Each state runs an integrated financial management system that controls budget releases, accepts bills from drawing and disbursing officers (DDOs), passes them, and pushes payments. Karnataka's version is Khajane-II, Tamil Nadu's is IFHRMS, and most other states have their own. Bills are increasingly electronic vouchers with digital signatures. The IFMS also usually hosts the receipt side: citizens and businesses paying taxes, fees and deposits through netbanking, cards, UPI or NEFT, generating an electronic challan.

The Centre's PFMS. The Public Financial Management System tracks central money as it flows to states and implementing agencies. For centrally sponsored schemes it now sits alongside the single nodal agency (SNA) framework, under which each scheme in each state has one designated account from which implementing agencies draw, with balances visible to the Centre.

The banker. The RBI acts as banker to state governments, and its core banking system, e-Kuber, holds state balances, handles the government's cash position and runs the auction calendar through which states borrow. When a treasury runs short, the gap is met through ways and means advances or market borrowing, both of which carry a cost.

Each layer was digitised separately. The friction appears at the seams between them.

What one state's ledger looks like

Karnataka published unusually detailed numbers in a presentation to the CAG's conference of state finance secretaries in September 2025. They show both the payment and receipt sides of Khajane-II over four years.

The system also connects to HRMS for about 5.03 lakh government employees and supports DBT, account-based transfers and e-RUPI vouchers as payment modes. From 1 June 2025 the state stopped sending printed copies of around 7.5 lakh salary-bill vouchers a year to the Accountant General, relying on the electronic record instead. That sounds administrative. It is actually the point at which the digital record becomes the legal record.

Worked example: reading the Karnataka numbers

A few divisions turn that table into something more intuitive.

• Average bill size in 2024-25: Rs 4,42,613.89 crore across 36,75,112 bills is about Rs 12.04 lakh per bill. Most bills are far smaller; a few very large transfers to local bodies and corporations pull the average up.
• Workload per DDO: 36,75,112 bills across 31,556 DDOs is about 116 bills per officer per year, roughly one every two working days, each handling about Rs 14 crore annually.
• Payment growth: payments rose about 44.5% between 2021-22 and 2024-25, while the number of bills rose only about 15%. Treasury throughput is growing in value, not in transaction count.
• Receipts: Rs 1,14,546 crore across 1,90,71,021 challans in 2024-25 is an average of about Rs 60,000 per challan. Challan volume grew about 78% over the four years, while value grew about 41%, consistent with more small payments moving online.

The receipt side is where citizens touch the treasury most directly, through property registrations, fees, and departmental deposits. The rising count of smaller challans suggests the digital front door is being used by individuals, not only by large corporate filers.

The next frontier in treasury reform is not paying faster. It is knowing, on any given morning, where every rupee is sitting and why.

The float problem: SNA balances and unreconciled money

The single nodal agency system was meant to end the old practice of central scheme money being parked across thousands of implementing-agency bank accounts. It has helped, but audit reports show the reconciliation is still messy. The CAG's 2023-24 State Finances audit of Punjab found about Rs 3,372.86 crore lying unspent in SNA accounts at 31 March 2024, according to PFMS data. PFMS recorded Rs 6,487.22 crore transferred to SNAs that year, while the state's own figures put it at Rs 6,677.24 crore, a mismatch of roughly Rs 190 crore between two official ledgers.

The same audit found Rs 406.98 crore of cess and levies collected by DDOs but not deposited into the government account, 2,037 abstract contingent bills worth Rs 2,804.85 crore still awaiting detailed vouchers, and committed expenditure at 65% of the state's revenue spending. None of these are payment-speed problems. They are visibility problems: money has moved, but the ledger cannot fully say where it went or whether it was used.

Why does this matter to anyone outside a finance department? Because unspent balances in scheme accounts are money that the state could have used elsewhere, or that the Centre could have released to a state that needed it. Each crore idling in a nodal account while another treasury borrows through ways and means advances is a real interest cost.

Payroll, the biggest single draw

Salaries were the largest component of committed spending in 19 states in 2022-23, according to the CAG data. That is why so many states are consolidating HR and payroll into their treasury systems. Tamil Nadu sanctioned Rs 288.91 crore for IFHRMS to link about 29,000 DDOs and digitise around 9 lakh service registers, according to a 2018 state government release. Telangana set a 15 October 2026 deadline to finish digitising records for about 3.48 lakh employees across 38,710 offices and directed that DBT sanctions use Aadhaar authentication, according to The Hans India.

Integrating HR with the treasury closes an old gap: the employee master and the payment record used to be separate systems, which is where ghost employees and duplicate payments hide. Our companion piece on government payroll and wage payments covers that side in more detail, including the shift to 60:40 funding for rural wages that lands directly on state treasuries from July 2026.

Where programmable money might fit

There is an obvious argument that tagged or programmable money could solve the float problem: if a scheme rupee could only be spent for its intended purpose, and only released when a milestone is met, the SNA balance would effectively become a set of conditional payments rather than a pooled account. Karnataka already lists e-RUPI as a supported payment mode, and our explainers on e-RUPI vouchers and purpose-bound money look at how that works in practice.

We are cautious about over-reading this. The Punjab audit's problems were missing vouchers and undeposited collections, not a lack of programmability. A treasury that cannot reconcile two ledgers today will not fix that by adding a third, more complex one. The sequence that makes sense is integration first (one employee master, one scheme account view, one receipt ledger), and only then conditional money for specific flows such as contractor milestone payments. India's broader digital public infrastructure gives states the identity and payment rails; the ledger discipline has to come from inside each finance department. For background on how money actually moves between banks and the RBI, see what are payment rails.

What to watch next

• 15 October 2026: Telangana's e-HRMS completion deadline. Whether it holds is a useful test of how quickly a large state can consolidate its employee master.
• October 2026 to March 2027: the first two full quarters in which states fund 40% of rural employment wages under the VB-G RAM G Act. Watch ways and means advance usage at the RBI and any reports of states delaying their share.
• 1 February 2027: the Union Budget for 2027-28, which will set transfers to states and show whether the Centre tightens SNA release conditions further.
• Early 2027: the RBI's next State Finances study, which will carry 2024-25 accounts and 2026-27 budgets, and the first audited numbers on whether committed spending ratios moved after the new wage-sharing rule.

State finance data are revised as accounts are audited. Ratios from the RBI and the CAG use different definitions of committed expenditure and are not directly comparable.

On the evening wages land, the queue outside a money transfer counter in Jubail's industrial city runs past the pharmacy next door. Inside, a site engineer from Kerala sends part of his salary home while checking, on the same phone, how much of a new refrigerator he has left to pay off in instalments. Both transactions run on Saudi rails that the central bank either owns or licenses. Together they describe the two halves of Saudi fintech in 2026: one of the world's largest flows of money leaving a country, and a domestic consumer finance boom built on apps.

How we chose these companies

We selected the companies below using measurable criteria as of October 2026: licences from the Saudi Central Bank (SAMA), funding and valuation, users and merchants, and product breadth. They are grouped by category rather than ranked. Where we could not verify a recent number for a company, we describe it qualitatively and say so. That applies more often in Saudi Arabia than in India or the UAE, because several leading firms are private and publish few figures.

This guide is narrower than a full directory. It covers the firms with the clearest public footprint and the infrastructure they run on, and pairs them with SAMA's own remittance data, which is the best public measure of how money actually moves out of the kingdom.

How much money is sent from Saudi Arabia in 2026?

Expatriate workers in Saudi Arabia sent about SAR 165.5 billion abroad in 2025, roughly $44 billion, according to SAMA's monthly figures as compiled by Argaam (our sum of the twelve months). In March 2026 expat remittances hit SAR 16.92 billion, up 9% year on year and a record for the month. Saudi nationals' own transfers abroad fell 20% in March.

The monthly series, published by Argaam from SAMA data in May 2026, shows how steady the flow is. Expat remittances ran between SAR 12.6 billion and SAR 15.5 billion every month of 2025. The first quarter of 2026 totalled SAR 42.79 billion, up about 1.9% on the same quarter of 2025, after soft January and February months were offset by a 35% jump in March, which fell around the end of Ramadan.

The other line in the data is more surprising. Remittances by Saudi nationals fell to SAR 15.2 billion in Q1 2026 from SAR 18.82 billion a year earlier, a drop of about 19%. Across 2025 they came to roughly SAR 70.4 billion. Nationals' transfers are a mix of investment, tuition and family spending abroad, and the decline is worth watching as a signal of domestic confidence or tighter household budgets.

For fintech, the expat number is the prize. A market of over SAR 165 billion a year, sent mostly in small, regular amounts, is precisely the business that digital wallets and app-based remitters want to take from counters. Our Gulf to South Asia corridor analysis covers where the money goes.

Who runs Saudi Arabia's payment rails?

Saudi Arabia's core payment rails in 2026 are owned by the state. Saudi Payments, a wholly owned SAMA subsidiary set up in 2019, runs the mada card and point-of-sale network, the SADAD bill payment system, Esal and the sarie instant payment system. Fintechs build on top of these rails rather than replacing them, which keeps the central bank at the centre of almost every transaction in the kingdom.

Mada's roots go back to 1990, when SAMA set up the Saudi Payments Network, according to Saudipedia. The newer rail is sarie, which SAMA launched on 20 February 2021: round-the-clock transfers between banks for a fee of no more than SAR 1, sent to a mobile number instead of an IBAN, with amounts under SAR 20,000 credited instantly. It handled more than 74 million operations worth SAR 121 billion in its first half-year, Asharq Al-Awsat reported. The cumulative effect shows in SAMA's own numbers: electronic payments made up 85% of retail payments in 2025, up from 79% in 2024, across 14.6 billion transactions, per a SAMA statement in April 2026. Mada e-commerce alone reached about SAR 325.2 billion in 2025, up from SAR 197.4 billion, Zawya reported.

SAMA has also been selective about cross-border experiments. It joined the multi-central-bank mBridge platform as a full participant in June 2024, then ended its participation after completing a proof of concept on 13 May 2025, Business Today reported in September 2026, citing the Financial Times. Our mBridge explainer covers the project.

Which are the largest fintech companies in Saudi Arabia?

The largest Saudi fintech companies in 2026 are the two instalment and consumer finance providers, Tabby and Tamara. Tabby, headquartered in Saudi Arabia since its 2023 move from the UAE, was valued at $6.5 billion in September 2026. Tamara holds a SAMA finance company licence and operates across four Gulf markets. Both are expanding from buy now, pay later into broader financial services.

They sit in a fast-growing field. SAMA's governor said the kingdom had 281 fintech companies by August 2025, up from 82 in 2022, with cumulative investment above SAR 8.9 billion, in a September 2025 speech. The count reached 301 by the end of 2025, according to the Vision 2030 annual report as covered by Mubasher, and the governor put it at 371 by the end of August 2026 at Money20/20 in Riyadh, Arab News reported, against a National Fintech Strategy target of 525 active companies by 2030.

Tabby

Tabby raised $233 million in a Series F at a $6.5 billion valuation in September 2026, up from $4.5 billion in October 2025, with more than $18 billion of annualised transaction volume and 25 million registered users, according to EnterpriseAM. Over the past year it secured SAMA consumer and SME finance licences and bought Tweeq, a SAMA-licensed wallet, adding accounts, cards and transfers. The round remains subject to SAMA approval, per FinTech Futures and other coverage.

Tamara

Tamara Finance Company is a Saudi joint-stock company supervised by SAMA under licence number 95/A Sh/202502, with share capital of SAR 515 million, according to its website. It offers Sharia-compliant instalments of up to 24 months with no late fees and lists Saudi Arabia, the UAE, Kuwait and Bahrain as markets. Tamara became Saudi Arabia's first fintech unicorn with a $340 million Series C in December 2023. In September 2025 it secured a Sharia-compliant asset-backed facility of up to $2.4 billion from a consortium of Goldman Sachs, Citi and Apollo-managed funds, $1.4 billion of it committed upfront, and said it had more than 20 million customers and 87,000 merchants, per Arab News. It has not disclosed a valuation since the Series C.

Which Saudi fintechs serve merchants and small businesses?

For merchants and SMEs, the leading Saudi fintechs in 2026 include Geidea, a Riyadh-based payments and point-of-sale provider that says it serves over 530,000 businesses, HALA, which raised a $157 million Series B in September 2025 and serves more than 140,000 small businesses, and Lean Technologies, which supplies open banking APIs across the UAE and Saudi Arabia.

Geidea

Geidea sells card terminals, a payment gateway, payment links and soft POS, plus restaurant and retail software such as kiosks, kitchen screens and loyalty tools. Its Saudi site says it serves more than 530,000 businesses and operates under SAMA supervision from Riyadh, per Geidea. In March 2021 it became the first non-bank in the region to receive an acquiring licence from SAMA, The Fintech Times reported, and in September 2026 it announced national payment gateway accreditation for e-commerce and certification for its SoftPOS, with Tap to Pay on iPhone still awaiting SAMA approval, per Zawya. Its pitch is bundling: the terminal is the door, and the software and settlement keep the merchant.

HALA

HALA offers SME and freelancer business accounts with an IBAN, terminals and soft POS, expense cards, SADAD bill payments, payroll and POS-based financing, under a SAMA licence held by Hala Payments Company. In September 2025 it raised $157 million in a Series B led by TPG's Rise Fund and Sanabil Investments, and said it served more than 140,000 businesses and processed over $8 billion of transactions a year, according to TPG; its own site now cites more than 150,000 customers. Same-day settlement of card sales into the merchant's own account is the feature that matters most for a small shop's cash flow.

Lean Technologies

Lean provides account verification, financial data and pay-by-bank APIs and lists Saudi Arabia and the UAE as its markets. Its website cites more than AED 16 billion processed, over 2 million accounts connected and 350+ clients across MENA, including Tabby. It raised a $67.5 million Series B led by General Catalyst in November 2024, when it reported more than $2 billion of total transaction volume, per FinTech Futures. The firm is a bet that open banking, already set out in Saudi Arabia, becomes a mainstream way to pay. Our open banking explainer compares models.

Which digital banks and wallets lead in Saudi Arabia?

STC Bank, licensed by a Council of Ministers decision in June 2021 and supervised by SAMA, is the most prominent digital bank in Saudi Arabia in 2026. It offers accounts, cards, financing and international transfers through an app. Wallets such as Tweeq, now part of Tabby, show the other path into banking: start with a stored-value licence, then add credit.

STC Bank

STC Bank's licence dates from Council of Ministers Decision No. 671 of 22 June 2021, and it reports paid-up capital of SAR 6.35 billion, according to its website. It serves individuals, SMEs and larger businesses with prepaid and credit cards, savings products, personal and working capital financing, merchant POS and domestic and international transfers. In April 2025 its chief executive said it had about 1.5 million customers and would roll out its full range of services later that year, after stc injected $800 million; stc owns 85% and Western Union 15%, Argaam reported, citing Bloomberg. Its link to the country's largest telecoms group gives it distribution that a standalone startup cannot buy.

Tweeq

Tweeq is a SAMA-licensed digital wallet that Tabby agreed to acquire in 2024, as Finance Middle East reported. It matters less as a standalone company than as an example of how licences, rather than technology, have become the scarce asset in Saudi fintech.

How do Saudi Arabia's leading fintechs compare?

Metrics marked "company" are self-reported on company websites and are not audited.

Worked example: what a monthly remittance from Saudi Arabia costs

Take a worker sending SAR 750, about $200 at the riyal's dollar peg, home each month. The World Bank's global average cost of sending $200 was 6.36% in Q3 2025. At that average, each transfer would cost about SAR 47.70 in fees and exchange-rate margin, or SAR 572.40 over a year. That is close to an extra month's transfer handed to intermediaries.

Gulf corridors are generally priced well below the global average because of competition, so the true cost for most senders will be lower. Scale it up anyway: if the SAR 165.5 billion that expats sent in 2025 had carried an all-in cost of just 2%, fees and FX margins would total about SAR 3.3 billion, roughly $880 million, in a single year. Every percentage point that digital competition removes is worth about SAR 1.65 billion to senders' families.

In Saudi Arabia, the fintech prize is not the purchase. It is the payday transfer that leaves the country every month.

Illustrative calculation using the World Bank global average and SAMA's 2025 expat remittance total; corridor prices vary by provider, amount and destination. Our cost of sending money home report tracks corridor prices.

What to watch next in Saudi fintech

Several dated events over the next year will show which way the sector is moving. Anyone considering a Saudi fintech licence, investment or product should take advice from a qualified local adviser.

• Monthly SAMA data: remittance figures for the rest of 2026, which will show whether March's record was a Ramadan effect or a trend, and whether Saudi nationals' outflows keep falling.
• SAMA approval of Tabby's Series F: the September 2026 round is subject to regulatory sign-off, and Tabby's CEO has said there is no rush to list.
• Phone wallets and SoftPOS approvals: SAMA's decision on Tap to Pay on iPhone for acquirers such as Geidea, a test of how quickly the domestic network opens to new acceptance methods, important for tourism and Hajj and Umrah spending.
• Fintech count against target: 371 firms by August 2026 against 525 by 2030; the next Fintech Saudi annual report will show whether licensing keeps that pace.
• Cross-border infrastructure: after leaving mBridge in May 2025, SAMA's next cross-border payments partnership will shape remittance rails. See our UAE fintech guide for how its neighbour is approaching the same question.
• Instalments into credit: whether BNPL leaders' new SAMA consumer and SME finance licences lead to larger, longer-term lending, which would bring more regulatory scrutiny.

The Saudi model is distinctive. The state owns the rails, licences are scarce, and the big private winners are consumer finance companies rather than payment apps. Remittances, by contrast, are still mostly a counter business. That gap between a modern domestic stack and an older outbound one is where the next set of Saudi fintech companies will be built. Our explainer on Gulf payroll cards and the wage protection system shows where that money starts its journey.

This list is editorial research, not investment advice or an endorsement of any company. Company figures are as reported by the linked sources.

In Balangir district of western Odisha, some rural job-scheme workers told surveyors in 2025 that they had waited more than five months for wages the law said should reach them in fifteen days. The money itself moved in seconds once it moved. Everything before that moment was the problem, and that is the real story of how the Indian state pays the people who work for it.

Government payroll in India runs on two very different tracks. One is the salaried track: roughly 50 lakh central government employees and 69 lakh central pensioners, plus several crore state employees, paid monthly through treasury systems that have been digitising for two decades. The other is the daily-wage track, dominated for twenty years by the rural employment guarantee, which paid crores of casual workers through a chain of muster rolls, wage lists and fund transfer orders. On 1 July 2026 that second track changed its legal basis entirely. This piece looks at what the data says about how well both tracks work, and where the next failures are likely to sit.

The rails are almost fully digital already

The headline numbers on digitisation are not in dispute. According to a Ministry of Rural Development overview published by PIB in August 2025, the share of rural job-scheme wages paid through the electronic fund management system rose from 37% in 2013-14 to 99.94% in 2024-25. Aadhaar numbers seeded in the scheme database went from about 76 lakh in January 2014 to 13.45 crore, and 13.05 crore workers were linked to the Aadhaar Payment Bridge System, the NPCI mapper that routes a credit to whichever bank account a person's Aadhaar number currently points at.

That scale is enormous. The same document puts person-days generated in 2024-25 at 290.60 crore across 15.99 crore registered households, with women supplying 58.15% of the work, up from 48% in 2013-14. The budget allocation for 2025-26 was Rs 86,000 crore, against Rs 33,000 crore in 2013-14.

So when people talk about "digitising government wages" in India, they are describing something that has largely happened. The more useful question is what the digital rails did and did not fix.

Where the delay actually lives

A wage payment under the old scheme passed through roughly five hand-offs. A muster roll recorded attendance at the worksite. A wage list was generated after the roll closed. A fund transfer order (FTO) was created and digitally signed by two officials at the block or panchayat level. The FTO then went to the Centre, which released money through PFMS, and finally the credit landed in the worker's account through the Aadhaar bridge or an account-based transfer.

Official reporting measures the first half of that chain. In a Lok Sabha reply on 22 July 2025, the ministry said over 96% of FTOs had been generated within 15 days of muster roll closure in each of the previous five years, and that 100% of wage payments went directly to beneficiaries through DBT. The PIB overview put on-time FTO generation at 97.81% by March 2025.

Those figures are probably accurate, and they are also the wrong yardstick. An FTO generated on day 12 says nothing about when the Centre actually pushed the money, or whether the credit bounced because a bank account was no longer correctly mapped. Researchers at LibTech India have argued for years that the second stage, between FTO signing and credit, is where most delay accumulates, and that it was historically excluded from delay compensation calculations. Their 2024-25 national tracker collected worker testimony of waits of up to three months in Karnataka and Uttar Pradesh and over five months in parts of Odisha.

The money that never arrived is measurable too. Unpaid wages under the scheme rose from Rs 512.75 crore in 2020-21 to Rs 974.38 crore in 2024-25, according to ministry data compiled by IndiaSpend as of March 2025. The ministry told Parliament in July 2025 that all pending wage liabilities up to 2024-25 had since been cleared. Both statements can be true: liabilities get cleared in lumps, after the household has already borrowed to cover the gap.

A payment system is only as fast as its slowest signature, and in government payroll the slowest signature is usually a fund release, not a bank.

ABPS: the identity layer that cut both ways

The Aadhaar-based payment system has existed in the scheme since 2017, but the government made it mandatory for wage payments from 1 January 2024, after several extensions from an original February 2023 date. The logic is sound on paper. A worker's Aadhaar becomes their financial address, so a changed or closed bank account no longer breaks the payment as long as the NPCI mapper is updated. It also makes duplicate or fictitious beneficiaries harder to sustain.

The cost was eligibility. A worker whose Aadhaar was not seeded, whose name was spelled differently on the job card, or whose bank had not completed mapping could not be paid through ABPS at all. Coverage reported by Rau's IAS at the time noted that 12.7% of active workers were still ineligible in December 2023 and cited about 7.6 crore job-card deletions since April 2022, often linked to such mismatches. By April 2025, LibTech's tracker put ABPS ineligibility at 27.5% of all registered workers but only 1.5% of active workers.

That gap between "all" and "active" is the whole debate in one number. The government reads it as a clean-up success: the people actually working are almost all payable. Critics read the same figure as evidence that people who could not get linked simply stopped being counted as workers. The trend in participation is consistent with both readings. Workers who availed employment fell from 11.17 crore in 2020-21 to 7.88 crore in 2024-25, per the Lok Sabha reply, though the 2020-21 peak was inflated by pandemic-era demand.

The new law: 125 days, shared bills, a weekly clock

Parliament passed the Viksit Bharat Guarantee for Rozgar and Ajeevika Mission (Gramin) Act in December 2025, and it replaced the 2005 Act from 1 July 2026. According to a PIB explainer from May 2026, the guarantee rises to 125 days per household per year, with states allowed to notify a pause of up to 60 days during peak sowing and harvest. Wages are to be paid weekly, or within 15 days of muster roll closure, with compensation of 0.05% of unpaid wages per day of delay beyond the 16th day. The central share for 2026-27 is Rs 95,692.31 crore, and the total outlay is expected to exceed Rs 1.51 lakh crore once state shares are counted. The administrative expenditure ceiling goes from 6% to 9%.

The change that matters most for payments is the funding split. A Drishti summary of the Act describes a 60:40 Centre-state ratio for most states and 90:10 for northeastern and Himalayan states, where the old scheme had the Centre carrying the entire unskilled wage bill. In practice, every wage payment now needs two treasuries to have cash ready at the same time.

Worked example: what a late wage is worth

Take a worker on a notified wage of Rs 300 a day who completes 12 days on a fortnightly muster roll. The wage due is Rs 3,600. Suppose the money arrives 46 days after the muster closes, roughly the kind of wait workers described to LibTech in Odisha and Karnataka, and not an extreme one.

• Days of delay counted for compensation: 46 minus 16 = 30 days.
• Compensation rate: 0.05% per day x 30 days = 1.5%.
• Compensation owed: 1.5% of Rs 3,600 = Rs 54.

Rs 54 for a month-long delay works out to an annualised rate of about 18%. That sounds respectable until you compare it with what a rural household actually pays to bridge a month without wages from an informal lender, which is frequently far higher. And it assumes the compensation is calculated, approved and credited, which has historically been the weakest link. Our view: the rate is too low to change official behaviour on its own. What will change behaviour is whether the weekly clock gets measured end to end, from muster closure to credit, and published by state.

Salaried payroll: quieter, but not finished

On the salaried side, digitisation looks less dramatic because it happened earlier, but the numbers are large. Karnataka's HRMS, integrated with its Khajane-II treasury, covers about 5.03 lakh government employees and generates around 7.9 lakh pay bills a year, according to a Karnataka presentation to a CAG conference of state finance secretaries in September 2025. From 1 June 2025 the state stopped sending printed copies of roughly 7.5 lakh salary-bill vouchers a year to the Accountant General. Telangana is still entering records for about 3.48 lakh employees into its e-HRMS, with a 15 October 2026 target set in a review this month. Tamil Nadu's IFHRMS, sanctioned at Rs 288.91 crore in 2018, was designed for about 29,000 drawing and disbursing officers and 9 lakh digitised service registers.

Pensions show what end-to-end digitisation can do. The defence ministry's SPARSH system, launched in October 2020, handles both sanction and disbursement on one platform and had onboarded about 31 lakh of 32 lakh defence pensioners by April 2025, according to reporting on defence ministry figures. It replaced an arrangement involving three agencies and over 45,000 public interfaces. That is the model wage schemes have not reached: one ledger owning the entire path from entitlement to credit.

The fiscal pressure behind all of this is rising. The CAG's decade review found states' committed spending on salaries, pensions and interest grew 2.49 times between 2013-14 and 2022-23, to about Rs 15.64 lakh crore, with salaries the largest item in 19 states, as Business Standard reported. When the payroll is that large a share of the budget, a treasury that cannot see its liabilities in real time ends up delaying whatever is easiest to delay. Historically, that has been wages for people with the least power to complain.

Why this matters beyond India

Many governments are now trying to replicate India's model of identity-linked payments; we cover the broader pattern in G2P payments around the world. The Indian lesson is uncomfortable but useful. Identity-based routing removes one class of failure, the stale bank account, and introduces another, the identity mismatch that blocks a payment entirely. It also tends to make delays look like someone else's problem, because each stage reports its own on-time rate. Gulf states solved a different version of this with mandated wage-file monitoring, described in our explainer on wage protection systems and payroll cards, which put the employer's payment date itself under surveillance.

For readers following India's welfare plumbing more broadly, our guides to direct benefit transfer and India's digital public infrastructure cover the shared rails, and how state treasuries move money explains why the new 60:40 split makes state cash management the critical variable.

What to watch next

• 15 October 2026: Telangana's deadline for completing e-HRMS records for its roughly 3.48 lakh employees. Slippage would be an early signal of how hard state payroll consolidation still is.
• Q3 and Q4 of 2026-27: the first full quarters of VB-G RAM G payments. Watch whether states publish end-to-end payment times, not just FTO generation rates, and whether any state's 40% share arrives late enough to trigger compensation at scale.
• Around April 2027: the 8th Central Pay Commission, whose terms of reference the Cabinet approved in late October 2025 with an 18-month deadline, is due to report. Its recommendations, expected to take effect from 1 January 2026 per The Tribune, will mean arrears calculations for about 50 lakh employees and 69 lakh pensioners, a large stress test for payroll systems.
• 1 February 2027: the Union Budget for 2027-28 will show whether the central allocation for rural employment holds near Rs 95,692 crore once actual state co-funding behaviour is known.

Figures are from government releases and independent trackers as cited; scheme data is revised frequently, so check the latest ministry dashboards for current values.

A fishmonger at Kirkgate Market in Leeds takes card payments on a terminal from one fintech, banks her takings with a second, pays her Norwegian supplier through a third, and keeps her rainy-day pot in a savings app run by a fourth. Fifteen years ago every one of those jobs belonged to a high-street bank. In 2026 the UK has the deepest bench of independent fintech companies in Europe, several now profitable and fully licensed, even as fresh venture money into the sector has fallen to its lowest level in a decade.

How did we choose the leading UK fintech companies for 2026?

We selected UK fintech companies for 2026 on measurable criteria as of October 2026: customers, revenue and profit in the latest reported year, payment or lending volume, licences held, and valuation where disclosed. We grouped them by category rather than ranking them, because a challenger bank, a payment processor and a business lender earn money in very different ways.

The funding climate is the backdrop. UK fintech investment fell to £1.8 billion across 205 deals in the first half of 2026, down about two-thirds from £5.0 billion a year earlier and the lowest since KPMG began the series in 2016, according to KPMG's August 2026 release. The largest deal was a £543 million private equity investment in Ebury. Thin funding makes profits matter more, and most of this list now has them.

Which UK digital banks are largest in 2026?

By revenue and customers, Revolut is the UK's largest fintech in 2026, with £4.5 billion of 2025 revenue, £1.7 billion of pre-tax profit and 68.3 million retail customers worldwide. Monzo is second, with £1.7 billion of revenue and 15.2 million customers in the year to March 2026, and Starling is the most profitable per pound of revenue among the three.

Revolut

Revolut's 2025 revenue rose 46% to £4.5 billion and profit before tax rose 57% to £1.7 billion, with 68.3 million retail and 767,000 business customers, according to Euronews. Its last funding round valued it at $75 billion. The licence saga is over: Revolut won a restricted UK banking licence in July 2024 and the Prudential Regulation Authority let it leave mobilisation in March 2026, after about 20 months, as Tech.eu reported. It has about 13 million UK customers.

Monzo

Monzo's revenue rose 39% to £1.7 billion in the year to March 2026, with adjusted profit before tax of £172.6 million, 15.2 million customers and deposits of £25.7 billion, up 55%, per Payment Expert. It launched in Ireland in April 2026 on a Central Bank of Ireland licence and closed its US operations.

Starling Bank

Starling is the exception to the growth story. Revenue fell to £887 million in the year to March 2026 and pre-tax profit dipped 3% to £217 million as interest income dropped with rates, while accounts rose to 6.2 million, according to Tech.eu. Its software arm, Engine, signed Scotiabank's Tangerine as its first North American client.

Which UK payments companies lead in 2026?

Wise is the UK's largest listed payments fintech in 2026, moving $243 billion across borders in the year to March 2026 and switching its primary listing to Nasdaq in May. Checkout.com is the largest private UK payments processor by volume, handling more than $300 billion in 2025. GoCardless and TrueLayer lead bank-to-bank payments.

Wise

Wise began trading on Nasdaq under the ticker WSE on May 11, 2026, keeping a secondary London listing. For its fiscal year to March 2026 it reported $243 billion of cross-border volume, up 31%, net revenue of $2.5 billion and $39 billion of customer balances, according to Crowdfund Insider. Our explainer on mid-market rates and zero-fee transfers covers how its pricing compares.

Checkout.com

London-based Checkout.com processes online card payments for large merchants. It said in February 2026 that processed volume passed $300 billion in 2025, up 64%, with net revenue up more than 30% and full-year adjusted EBITDA margin above 10%, per its newsroom. A September 2025 staff share buyback set a $12 billion valuation.

GoCardless

GoCardless runs direct debit and open banking collections for subscription and invoice businesses. In the year to June 2025 its revenue rose about 18% to £155.5 million, payment volume doubled to £79.2 billion after the Nuapay acquisition, and its loss narrowed to £25.5 million, according to Tech.eu. It posted its first adjusted EBITDA-positive quarter in the final quarter of that year and expects a first EBITDA-positive full year in 2026.

TrueLayer

TrueLayer provides open banking payment APIs in the UK and Europe. It processed 188 million transactions in 2024, up 84%, on revenue of £20.3 million and a pre-tax loss of £38.6 million, as Tech.eu reported from its accounts. Its numbers are small next to card processors, which is the point: account-to-account payments in the UK are still early.

Which UK lenders, savings and banking infrastructure firms stand out?

By profit, OakNorth is the largest UK fintech lender in 2026, earning £223 million before tax in 2025 on a £7.2 billion loan portfolio. Zopa reported £65 million of underlying profit, Moneybox reached £23.2 billion in assets under administration by mid-2026, and ClearBank, which provides clearing to other fintechs, posted a third year of UK profit.

OakNorth

OakNorth lends to growing mid-sized businesses. In 2025 it earned £223 million before tax, grew new originations 33% to £2.8 billion and expanded its loan facilities portfolio to £7.2 billion, with the US accounting for 40% of new lending, according to Crowdfund Insider.

Zopa

Zopa, a lender turned bank, reported 2025 revenue of £377.1 million, up 24%, and underlying profit before tax of £65 million, with 1.7 million customers and £6.4 billion of deposits, per the bank's own release. Its Biscuit current account signed up hundreds of thousands of customers within months of launch.

Moneybox

Moneybox offers savings, investing, pensions and Lifetime ISAs. By its second quarter of 2026 it had £23.2 billion of assets under administration, up 40%, 1.9 million funded customers and quarterly revenue of £41.2 million, up 52%, per its shareholder update.

ClearBank

ClearBank is a clearing bank for other fintechs and payment firms. In 2025 its normalised group revenue rose 34% to £121.6 million, UK pre-tax profit rose 53% to £12.2 million and payment scheme volumes jumped 57% to 262 million, though the group made a statutory loss of £16.7 million as it invested in Europe, according to its results.

How do the leading UK fintech companies compare?

As of October 2026, the UK's fintech leaders split between profitable licensed banks (Revolut, Monzo, Starling, Zopa, OakNorth), payments firms reporting volume (Wise, Checkout.com, GoCardless, TrueLayer) and specialist platforms (Moneybox, ClearBank). The table shows each one's latest dated metric.

What does revenue per customer reveal? A worked example

Dividing revenue by customers in the latest year shows Revolut earns about £66 per customer, Monzo about £112 and Starling about £143. Revolut's lower figure reflects a huge international base of lighter users, yet it converts revenue into profit far more efficiently, with a margin near 38% against roughly 10% at Monzo.

The arithmetic: Revolut's £4.5 billion over 68.3 million retail customers is about £66 each. Monzo's £1.7 billion over 15.2 million is about £112. Starling's £887 million over 6.2 million accounts is about £143. On profit, Revolut's £1.7 billion is 38% of revenue, Starling's £217 million is about 24% and Monzo's adjusted £172.6 million is about 10%.

Two lessons follow. Starling earns the most per customer because many of its accounts are main business accounts, and 56% of its SME customers use it as their primary bank. But it is also the most exposed to falling rates, which is why its revenue fell while the others grew. Monzo, still investing in Europe and mortgages, has the most room to widen margins if those bets work.

The UK's three big digital banks have the same app-shaped front end and three entirely different businesses behind it.

How is the UK regulating stablecoins in 2026?

The UK published final stablecoin rules in 2026. The FCA regulates all UK issuers, opening its authorisation gateway on September 30, 2026, and the regime goes live on October 25, 2027. The Bank of England adds stricter rules for issuers deemed systemic, including a £40 billion guardrail per coin, having dropped earlier proposed holding limits.

Under the FCA's policy statement, issuers must fully back coins, keep at least 10% in core assets such as on-demand deposits and short-term government debt, and redeem by the end of the next business day, according to Payment Expert. Systemic issuers must hold 30% at the central bank and 70% in short-dated gilts. ClearBank's chief executive argued the rules still make bank issuance hard to run profitably. Our detailed guide to the UK stablecoin regime covers the full rulebook.

What should you watch in UK fintech next?

The key dates for UK fintech are the FCA stablecoin gateway, open from September 30, 2026 to February 28, 2027, the regime's go-live on October 25, 2027, and Revolut's aim of 100 million customers by mid-2027. Whether funding recovers from a decade low will decide which younger London fintechs reach the scale of this list.

• Now to February 28, 2027: FCA authorisation gateway for stablecoin issuers and crypto firms.
• October 25, 2027: UK stablecoin and cryptoasset regime goes live.
• Mid-2027: Revolut's target of 100 million customers, alongside its pending US banking licence application filed in March 2026.
• Listings: Wise's Nasdaq move is a test case. A strong run would encourage Revolut, Monzo or Starling to list in New York rather than London.
• Remittance pricing: Competition from Wise and Revolut keeps pushing down the cost of sending money from Britain, as our UK to India money transfer guide shows.

For comparison, see our lists of leading US fintech companies and Canadian fintech companies.

This list is editorial research, not investment advice or an endorsement. Company figures are as reported by the linked sources; profit measures differ between companies (statutory, adjusted or underlying).

Here is a question we hear often from people abroad: if India already has UPI, which handled more than 24 billion transactions in September 2026 alone, why did the Reserve Bank of India bother building a digital rupee? It sounds like a country building a second road next to an empty-ish highway. The answer is that they are not the same road, and they are not even carrying the same cargo.

The one-line difference

UPI is a set of rails that moves money between bank accounts. The e₹ is the money itself, issued by the central bank and held in a digital wallet. When you pay a chai stall with UPI, your bank debits your account and the stall's bank credits theirs. When you pay with e₹ from wallet to wallet, a digital token issued by the RBI moves from your phone to theirs, and no bank account is touched.

The RBI's FAQ spells this out: wallet-to-wallet e₹ payments settle instantly, while an e₹ app that scans a UPI QR code settles on UPI's timelines.

The scale gap

According to NPCI data reported by Outlook Money, UPI moved ₹29.37 lakh crore in September 2026, with volume up 23% on the year. The retail e₹, by comparison, had crossed 120 million transactions in total by December 2025, according to RBI Deputy Governor T Rabi Sankar. At September's pace, UPI clears that many payments in under four hours.

So this is not a race. Nobody at the RBI expects the e₹ to replace UPI for buying groceries.

Side by side

What changes for a regular user

Honestly, not much on day one. If you live in Bengaluru and pay your rent, electricity and food delivery through UPI, the e₹ adds a second balance you have to top up. You get no interest on it, and wallets offered by non-banks such as MobiKwik and Cred cap transfers at ₹10,000 each and ₹50,000 a day, according to Business Standard.

Where it changes things is when someone else sends you e₹ with conditions. A company might pay a delivery rider's fuel allowance in e₹ that only works at petrol pumps. A state government might pay a welfare benefit that has to be spent within a window. The user sees the rules in the wallet. For the sender, it removes a lot of audit work.

What changes for a merchant

For the shopkeeper, the friction is low because e₹ wallets can pay a standard UPI QR code. A kirana store in Lucknow does not need new hardware. If the customer pays from wallet to wallet, the merchant receives central bank money that is final at once, rather than waiting on bank-to-bank settlement. Merchant numbers in the pilot are still modest: RBI data showed about 0.42 million merchants by June 2024.

UPI made moving money cheap. The e₹ is a bet that money with instructions attached is worth having too.

Why the RBI keeps both

• Resilience. UPI depends on banks being online. A central bank token that can work offline is a backup for outages, disasters and dead zones.
• Programmability. Rules at the money level are hard to bolt onto account-to-account transfers. They are native to a token.
• Cash costs. Printing, transporting and replacing notes is expensive. A digital equivalent of cash, held directly against the central bank, is the logical replacement for some of that.
• Cross-border. Central banks are testing how their digital currencies could settle with each other, a theme we cover in our mBridge explainer. UPI's international links follow a different path, through linkages such as Project Nexus.

A month in two wallets

To make this concrete, picture Ravi, a delivery rider in Hyderabad. His salary arrives in his bank account, and he pays rent, his phone bill and his daily tea through UPI, as he always has. Nothing about that changes.

Now suppose his employer starts paying the monthly fuel allowance in e₹ instead of cash reimbursements. The money lands in an e₹ wallet with a rule attached: it can only be spent at fuel stations, and it expires at the end of the month. Ravi does not have to keep receipts. His employer does not have to audit them. At the pump, he scans the same QR code he would use for UPI. If the station is in a patch of poor signal on the edge of the city, the offline capability the RBI has been piloting is meant to let the payment go through anyway.

For Ravi, the e₹ is not a replacement for anything. It is a narrow tool that appears when someone else wants to send him money for a specific purpose. That is a realistic picture of how most Indians will meet the digital rupee: through a scheme, an allowance or a subsidy, rather than by choosing it at the checkout.

The same logic applies to merchants. A fuel station that already shows a UPI QR does not have to do anything new to accept that allowance. The cost of supporting the e₹ is close to zero for them, which is precisely why the RBI made the two interoperable.

Our view

We think the two-track approach is sensible, and the comparison people make between them is mostly the wrong one. UPI answered the question of how to move a rupee cheaply between any two accounts in India. The e₹ answers a different one: what if the rupee itself could carry instructions, work without a network, and be held without a bank in the middle? Most households do not need that today. Government departments, employers and, eventually, cross-border corridors might.

For the basics of the digital rupee itself, see our e₹ explainer. For how India's bill payment network fits around both, see Bharat Connect explained.

A payroll manager in Lisbon pays forty contractors across three continents every month, and since last year she has been tempted to do it in dollar stablecoins. Before she can, she has to answer a question that did not exist two years ago: which stablecoin is legal where? In the EU, the platform she uses had to stop offering USDT spot trading. In Brazil, paying a contractor in a stablecoin counts as a foreign exchange operation. In Singapore, the coin needs a licence before anyone can call it MAS-regulated. Stablecoin regulation is no longer a vague future risk. It is a patchwork of real laws, and it decides which coins can move where.

This guide compares the rules in twelve jurisdictions as of October 2026, from the United States to India. We focus on four questions for each: what the law is, who supervises issuers, how far it has got, and how it treats foreign coins like USDT and USDC. Total stablecoin supply is about $306 billion as of October 2026 according to DefiLlama, with USDT at roughly $184 billion and USDC at about $73 billion, so the treatment of those two coins matters more than any local design choice.

Four models for regulating stablecoins

Strip away the acronyms and most countries fall into one of four camps.

• Licensed issuer regimes. A dedicated licence for issuers, with 1:1 reserves, redemption at par and a ban on paying interest. The US, EU, Hong Kong, Japan, Singapore and Canada are here.
• Central bank as backstop. The UK adds a second layer for systemic coins, with backing partly held at the central bank and a cap on issuance.
• Gatekeeping through FX and payments law. Brazil treats stablecoin payments as foreign exchange. The UAE lets foreign coins be used mainly to buy other crypto, reserving payments for licensed dirham tokens.
• No regime yet. India, South Korea and Australia have no live issuer framework, though each is somewhere on the road to one.

Our view is that the first model has won the argument. Nearly every jurisdiction that has legislated has landed on the same core: full reserves in short, safe assets, a legal redemption right and no yield to holders. The real differences lie at the edges, in how foreign coins get in and how much of the reserve income an issuer gets to keep.

The comparison table

United States and Canada: the dollar bloc writes its rules

The GENIUS Act was signed on 18 July 2025. It requires 1:1 reserves in cash, short-dated Treasuries and similar instruments, monthly public disclosure of reserve composition, and bans issuers from paying interest or yield. Issuers with no more than $10 billion outstanding can opt for state supervision if the state regime qualifies. Foreign issuers can be designated non-compliant by Treasury, which would bar service providers from facilitating secondary trading in their coins.

Implementation has slipped. The statute set a one-year deadline for rules, 18 July 2026, and none had been finalised by then. Regulators had issued around ten proposals, including the OCC's main reserves and capital proposal on 25 February 2026 and a Treasury proposal on state-regime similarity in April. The Chapman tracker notes the Act takes effect on the earlier of 18 January 2027 or 120 days after final rules. In practice that means January. We track the rulemaking in our GENIUS Act implementation guide.

Canada moved faster than many expected. Its Stablecoin Act was folded into the Budget Implementation Act, Bill C-15, which received Royal Assent on 26 March 2026, Fasken reports. Issuers will register with the Bank of Canada, hold 1:1 high-quality liquid reserves in the reference currency and offer redemption at par. The Bank's mandate took effect at Royal Assent, but the detailed framework will take years to build.

Europe: MiCA reaches its hard deadline

The EU's Markets in Crypto-Assets Regulation started applying to stablecoins in June 2024 and to all crypto service providers by the end of a transition that closed on 1 July 2026, with ESMA refusing extensions. MiCA splits stablecoins into e-money tokens referencing one currency and asset-referenced tokens referencing baskets, and requires issuers to be authorised in the EU. Our MiCA EMT and ART explainer covers the categories in detail.

The practical result is that foreign coins without an EU issuer are being squeezed. Tether has no MiCA authorisation, and Coinbase, Kraken, Crypto.com and Binance's EU entity have delisted or restricted USDT spot trading for EU users, according to a review of the July deadline. USDC and EURC are authorised through Circle's French e-money subsidiary. ESMA's interim register of e-money tokens listed 50 entries from 25 issuers as of 30 September 2026.

United Kingdom: two regulators and a £40bn ceiling

The FCA published final crypto rules on 30 June 2026, including PS26/10 on stablecoin issuance: full backing on statutory trust, redemption by the end of the next business day and no interest to holders. Authorisation becomes mandatory on 25 October 2027. The Bank of England adds a regime for systemic sterling coins. Its June 2026 policy dropped proposed £20,000 per-person holding limits and replaced them with a temporary £40 billion issuance guardrail per coin, with 30% of backing held as unremunerated deposits at the Bank and 70% in gilts of up to six months. No other major jurisdiction caps issuance of coins in its own currency.

Asia-Pacific: Hong Kong, Singapore, Japan, Korea and Australia

Hong Kong has the most advanced Asian regime in practice. Its Stablecoins Ordinance took effect in 2025, requiring licensed issuers to keep reserves fully backed by high-quality liquid assets, segregated and independently audited, with redemption within one business day. On 10 April 2026 the HKMA granted the first two licences, to HSBC and to Anchorpoint Financial, a joint venture of Standard Chartered, HKT and Animoca Brands, Charltons reports. HSBC plans a Hong Kong dollar coin in the second half of 2026.

Singapore set out its single-currency stablecoin framework in August 2023 but is only now legislating it. A MAS consultation on amendments to the Payment Services Act 2019, published on 1 September 2026, closes on 16 October 2026, Gibson Dunn notes. Only licensed issuers could call their coins MAS-regulated, with fines of up to S$250,000 for entities that falsely claim the label. Reserves must at all times at least equal the par value in circulation.

Japan was one of the first major economies to legislate, through a June 2023 revision that requires stablecoin issuers to be registered under banking or funds transfer rules. JPYC registered as a fund transfer provider in August 2025 and began issuing the first yen stablecoin on 27 October 2025, fully reserved in yen deposits and government bonds, The Block reported.

South Korea remains stuck. The Financial Services Commission and the ruling Democratic Party are drafting a consolidated Digital Asset Basic Act, but about ten digital asset bills sit before the National Assembly and lawmakers disagree over whether won stablecoin issuers should be controlled by bank-led consortiums, Sumsub reported in July 2026.

Australia passed the Corporations Amendment (Digital Assets Framework) Bill in 2026, requiring digital asset platforms and tokenised custody platforms to hold an Australian Financial Services Licence, with commencement about 12 months after Royal Assent. It does not create a dedicated stablecoin issuer regime.

The fight is no longer about whether stablecoins are allowed. It is about whose stablecoins.

The Gulf, Latin America and India

The UAE runs a dirham-first policy. Under the Central Bank's Payment Token Services Regulation, dirham tokens need a full licence and can be used for payments, while foreign coins from registered issuers can be used onshore mainly to buy virtual assets. Federal Decree-Law No. 6 of 2025 then extended Central Bank licensing to DeFi and to promotion, with criminal fines of up to AED 500 million for unlicensed activity. The financial free zones ADGM and DIFC keep their own token lists. Our Asia and Gulf rules comparison sets this against Singapore and Hong Kong.

Brazil took the most unusual route. Banco Central do Brasil Resolutions 519, 520 and 521, in force since 2 February 2026, treat buying, selling or paying with fiat-pegged tokens, including cross-border transfers, as foreign exchange. Transfers with unlicensed foreign counterparties are capped at $100,000 each, and providers must identify the owners of self-custodied wallets they send to or receive from, Cointelegraph reported. The central bank's president, Gabriel Galipolo, has said about 90% of Brazil's crypto activity involves stablecoins. For the domestic payments context see our Pix and Drex explainer.

India has no stablecoin framework. Crypto gains are taxed at 30% with 1% TDS on transfers, platforms must register with FIU-IND under anti-money-laundering law, and the RBI remains opposed to private stablecoins, preferring its own e-rupee. The Finance Ministry's discussion paper on crypto, in drafting since 2022, had not appeared as of October 2026.

A worked example: the same coin under three rulebooks

Take a hypothetical issuer with 10 billion units in circulation and see how three regimes treat it.

1. US dollar coin under GENIUS. All $10 billion can sit in cash, short Treasuries and similar instruments. For every 1 percentage point of yield, gross reserve income is about $100 million a year, none of which may be paid to holders as interest.
2. Sterling coin recognised as systemic in the UK. £3 billion (30%) must sit at the Bank of England earning nothing; £7 billion can be in short gilts. Each percentage point of yield earns about £70 million, roughly 30% less than the US case. The coin can grow to £40 billion before hitting the guardrail.
3. A Brazilian importer using the dollar coin. Paying a $250,000 invoice to an unlicensed offshore counterparty would breach the $100,000 per-transfer cap, so the payment has to go through a licensed FX institution or an authorised virtual-asset provider, and is reported as an FX operation.

Identical coins carry very different economics depending on where they are issued and used. Issuer profitability favours the US model; user friction is highest where stablecoins are treated as FX. For how issuers turn reserves into revenue, see our explainer on stablecoin issuer economics.

What to watch next

• 16 October 2026: MAS consultation on Singapore's stablecoin legislation closes.
• Second half of 2026: HSBC's planned Hong Kong dollar stablecoin launch under its new licence.
• End of 2026: Bank of England aims to finalise its Code of Practice for systemic stablecoin issuers.
• 1 January 2027: South Korea's 20% tax on crypto income above 2.5 million won is scheduled to start.
• 18 January 2027: the GENIUS Act takes effect in the US if final rules have not triggered it earlier.
• 28 February 2027 and 25 October 2027: the FCA's application deadline for the saving provision and the UK regime's go-live.

The direction of travel is clear. By the end of 2027 most major financial centres will have a live issuer licence, and the question for USDT in particular will be how many of them it is willing to apply for. For a deeper look at the two coins that dominate supply, read our USDT vs USDC comparison.

This overview summarises laws and proposals as of October 2026 and is not legal advice. Rules in several jurisdictions are still being finalised; firms should seek advice from qualified local counsel.

Every month, millions of people do the same small ritual. Payday arrives, they open an app or walk to a counter, and they send part of what they earned to someone they love in another country. The amounts are small. The fees, added up, are enormous.

Where the money actually goes

On a $200 transfer at the global average, roughly $12.70 never reaches the family. It disappears into three places:

1. The visible fee. The flat charge on the receipt. This is the part providers compete on in their ads.
2. The exchange-rate margin. The gap between the rate you're given and the real market rate. It's often bigger than the fee, and it's rarely shown in plain numbers.
3. The plumbing. Correspondent banks, pre-funded accounts sitting idle in every country, compliance checks done three times over. You never see these costs, but they're baked into the first two.

Digital is already cheaper, but not cheap enough

The data is clear on one thing: going digital helps. In the same World Bank data, fully digital services averaged about 4.59%, against 7.30% for traditional, cash-heavy ones. Some corridors are much better than the average. Sending to India through non-bank providers runs closer to 3.3% for a $200 transfer, thanks to fierce competition and fast domestic rails at the receiving end.

But "digital" mostly means the front end got nicer. Behind the app, money still moves through the same old network of banks that was designed for large, infrequent transfers between companies, not $200 sent every payday.

India: the world's biggest receiver

India received about $150.7 billion in 2025, more than any other country and roughly 39% of everything flowing into Asia and the Pacific. That number has more than doubled since 2016. Behind it are nurses in the Gulf, engineers in California, drivers in Toronto, students in Melbourne, and the families back in Kerala, Punjab, Maharashtra, Haryana and Delhi who depend on them.

And here's the part the statistics don't capture: a large slice of that money isn't spent on "anything". It goes to specific things, on specific dates. School fees. Electricity. A loan instalment. Medicine. About three-quarters of remittances go to immediate household needs.

A remittance is rarely just money. It's a school fee, an electricity bill or a hospital deposit wearing a currency costume.

How stablecoins change the maths

Stablecoins replace the slowest, most expensive middle stretch of the journey. Instead of a chain of banks and pre-funded accounts, value moves as a digital dollar that settles in seconds, any day of the week, for a tiny fraction of a cent. Genuine stablecoin payments reached about $390 billion in 2025, by McKinsey and Artemis's estimate, and Visa named remittances among the areas where stablecoins fit best.

That doesn't make the first and last mile free. Someone still has to convert local currency into a digital dollar at one end and back out at the other, under proper licences and with real compliance. But it means the remaining costs are honest ones, and there are fewer hands in the middle taking a slice.

What to check before you send

• Compare the final amount received, not the fee. That one number includes everything.
• Check the rate against the mid-market rate you see on Google. The difference is a cost.
• Ask whether the money can go straight to the bill. Paying the electricity company directly saves a second hop through someone's bank account, and removes the "did it arrive?" phone call.

The 3% target was set years ago. It's still not met. Closing that gap will probably take new settlement rails and honest exchange-rate pricing, rather than another small cut to the visible fee.

Picture a weekly haat in a hill block of Uttarakhand, where the nearest mobile tower sits behind a ridge and a UPI payment spins for a minute before failing. Cash still rules there, not out of habit but because it works without a network. Offline central bank digital currency is an attempt to copy that one property of cash into a phone or a card, and it turns out to be the hardest engineering problem in the whole CBDC project.

Why central banks keep coming back to offline

An online payment, whether UPI, a card or an online CBDC wallet, depends on a server confirming that the money exists and has not already been spent. Remove the network and that check disappears. Offline CBDC is the set of tricks that lets two devices settle a payment between themselves and trust the result.

Central banks rate the feature highly. A survey by the BIS Innovation Hub Nordic Centre for its Project Polaris handbook on offline CBDC, published in May 2023, found 49% of central banks saw offline payments with a retail CBDC as vital and another 49% saw them as advantageous. The reasons cited were resilience, inclusion, privacy and resemblance to cash. The same handbook warns that there is no one-size-fits-all design.

The appeal is easy to see in India. The RBI's own explanation of its February 2024 decision, as reported by Business Standard, was to enable transactions in areas with poor or limited internet connectivity, testing proximity and non-proximity solutions across hilly, rural and urban areas. The RBI's 2024-25 annual report presented offline capability as a highlight aimed at rural, disaster-hit and low-connectivity areas, which is also why it matters for digital cash in disaster relief.

How a device-to-device payment works

Strip away the branding and most offline designs follow the same five steps.

1. Load while online. The user moves value from a bank account or online wallet into an offline store on the device. The European Central Bank's first digital euro progress report describes pre-funding via the internet or at a designated ATM.
2. Hold value in protected hardware. The balance lives in a secure element, a tamper-resistant chip similar to the one in a bank card or SIM, so the user cannot simply edit the number.
3. Pay by proximity. Two devices talk directly, typically over NFC or Bluetooth. The sender's chip reduces its balance and signs a transfer; the receiver's chip checks the signature and increases its balance.
4. Keep the record local. In the ECB design, transaction data stays on the two devices and is not sent to a central system.
5. Reconnect and reconcile. When a device goes back online, it can refresh its keys, report counters and, for the receiver, move value back to an online account.

Every serious design worries about step three. If a chip is cracked, a fraudster could pay the same tokens twice to two shopkeepers who have no way of knowing. As a Ledger Insights summary of Polaris put it, double spending and counterfeiting are easier in the absence of online checks. That is why every offline design comes with limits.

What the RBI is testing for the offline e-rupee

India's thinking goes back to the RBI's October 2022 concept note, which called offline functionality important for inclusion and proposed to contain double-spending through technical measures and monetary limits on offline transactions. Since 2024, the pilot has tested several approaches. An RBI HaRBInger post, cited by Upstox in September 2026, invites solutions that enable offline e-rupee transactions using Bluetooth and NFC. One approach uses NFC on merchant point-of-sale terminals; another works without internet but still needs telecom signal, which helps where mobile data is weak but voice and SMS work.

The RBI has not published offline limits specific to the e-rupee. It has, though, a template. Its framework for small-value offline digital payments, which covers cards, wallets and mobile devices and was last updated in December 2024, caps offline payments at ₹500 per transaction and ₹2,000 outstanding per instrument, with UPI Lite at ₹1,000 and ₹5,000. Replenishing the limit requires an online transaction with additional authentication. It would be surprising if offline e-rupee ended up far from that range, at least at first.

There is also a link to programmability. The same February 2024 announcement that brought offline e-rupee brought programmable use cases, and the rural schemes where offline matters most are often the same ones where a state wants to fence money. We cover those in programmable e-rupee use cases.

The ECB's offline digital euro: the privacy pitch

Europe's offline design is built around privacy rather than connectivity. The euro area has good mobile coverage; what it has lacked is a digital payment that feels as private as a banknote. The ECB's progress report says that for offline payments only the two parties would know the personal transaction details, with no data shared with payment service providers, the Eurosystem or support providers. It calls this a cash-like level of privacy.

Politics has kept offline at the centre. The Council of the EU agreed its negotiating position in December 2025, backing a digital euro usable online and offline with a high degree of privacy, according to a Lewis Silkin note. Parliament's lead negotiator, Fernando Navarrete, had gone further in an October 2025 report, suggesting an offline-only digital euro unless the private sector failed to deliver a pan-European payment solution. The ECB rejected that idea, with an adviser arguing that offline alone would not let consumers shop online or send money over distance, Cyprus Mail reported. Parliament's committee position of June 2026 kept offline payments in the design and made them free of fees.

The ECB is candid that the hardware question is open. Offline phone payments depend on access to the NFC antenna and the secure element, which on some phones is controlled by the handset maker, and the progress report says that access depends on the legislation. The ECB is also studying cards with no battery that work through a bridge device. For the broader euro project, see the digital euro explained.

China's e-CNY: "touch one touch" since 2020

China has run offline e-CNY payments in public for longer than anyone. In a December 2020 Suzhou pilot, a group of recipients of 200 yuan red envelopes could test an offline wallet and make "dual offline" payments, where neither side has a connection, by tapping phones together, according to NFCW. Only designated phone models could take part.

Hardware came next. A 2021 China Daily report described e-CNY hardware wallets from a Shenzhen cash-machine maker in three forms, a display card, a shaped wallet and a SIM-based film, which pay by tapping a terminal. The People's Bank of China's white paper said these rely on security chips and can be built into IC cards, phones, wearables and other devices. That breadth matters because, as UCL's Geoffrey Goodell notes in a paper on hardware wallets, the global median smartphone cost US$92.59 in 2024, about 11% of global median income per person. A card is cheaper than a phone. We cover the wider programme in China's e-CNY.

The three designs side by side

Worked example: why the limits are low

Suppose an offline e-rupee wallet launched with the same ceiling the RBI uses for other offline instruments: ₹2,000 held offline and ₹500 per payment. These numbers are an assumption for illustration, not an RBI rule for CBDC.

• A farm worker loads ₹2,000 into the offline wallet before leaving town. At the haat he pays ₹40 for tea, ₹180 for vegetables and ₹60 for a shared jeep. His chip now shows ₹1,720. Nothing has touched a server.
• The vegetable seller takes 14 such payments during the day, averaging about ₹120, so she ends the day with roughly ₹1,680 received offline. If her own offline cap is also ₹2,000, she must reconnect and sweep funds to an online wallet before she can take much more.
• Now the attack. If someone cracks a chip and spends its ₹2,000 twice, the most a single device can create is ₹2,000 of bad money. Even an industrial fraud on 100,000 devices would cap out at ₹20 crore (100,000 × ₹2,000), detectable when receivers reconnect.

That is the trade. Each extra rupee of offline limit makes the wallet more useful and the worst case more expensive. It also explains why designers want devices to reconnect regularly: counters can be checked, compromised keys revoked, and bad tokens traced to the device that minted them. A design that lets a device stay offline for weeks with a large balance is a design that accepts a bigger fraud budget.

Offline CBDC is cash with a fraud budget. The size of that budget is a policy choice, not a technical detail.

Privacy and double-spend controls pull against each other

Here is the uncomfortable part. The strongest defence against double spending is to see every transaction eventually, which is how online systems work. The strongest privacy is to see none. Offline designs sit somewhere in the middle, and where they sit is a choice.

The Polaris research found that 44% of central banks believe offline CBDC should offer more privacy than online. The ECB has gone furthest by saying offline payment details stay between the two parties. India's concept note frames privacy as "managed anonymity", with small payments more private than large ones, which fits naturally with low offline caps: if each offline payment is small, more anonymity is easier to defend. For the full argument, read our piece on the CBDC privacy debate.

Liability is the other quiet question. If a merchant accepts a token from a cracked device, who eats the loss: the merchant, the wallet provider, or the central bank? Europe has flagged offline liability as one of the issues the ECB must settle before launch. Users deserve a clear answer in plain language before they are asked to trust an offline wallet with real money.

What to watch next

• Late 2026: EU trilogue negotiations on the digital euro regulation aim to conclude by the end of the year. The final text will set how offline holding limits are fixed and whether phone makers must open NFC and secure element access.
• Second half of 2027: the ECB's planned 12-month pilot, if the regulation is adopted in 2026, is the first chance to see offline digital euro payments with real users.
• 2026-27 in India: the RBI plans to extend CBDC pilots to more DBT schemes and retail use cases. Watch whether offline limits for the e-rupee are published, and whether any state uses offline wallets in a scheme that reaches places UPI does not.
• 2029: the ECB's target for a potential first issuance of the digital euro, assuming legislation is in place.

We set the three currencies against each other in e-rupee vs digital euro vs e-CNY.

On 28 September 2026, two judges of the Supreme Court of India read out a number that most citizens never see on their phones: Rs 49,194.05 crore in traffic fines issued through e-challans across the country, of which only Rs 26,175.05 crore had been recovered. Roughly Rs 23,000 crore was simply sitting there, owed to the state and unpaid. The bench responded by ordering that vehicles with unpaid challans can be impounded and that their owners can be locked out of registration, fitness and licence services until they clear their dues.

That order is a useful lens on a part of payments that gets far less attention than remittances or stablecoins: the money citizens send to their own government. Property tax, water charges, GST, court fees, passport fees, traffic fines. India has built excellent rails for collecting this money. What it has not fully built is the habit, or the enforcement, that makes people use them. This piece looks at where digital government collections work, where they leak, and why the hardest problem is no longer the payment button.

The collection gap that technology did not close

The e-challan system is a triumph of digitisation on the issuing side. Cameras and handheld devices generate fines, the record lands on the national Parivahan back end, and a vehicle owner can pay with UPI, a card or net banking in under a minute. Yet according to the figures placed before the court and reported by LiveLaw, the recovery rate works out to about 53%. Almost half the value of fines issued has not come in.

Even government departments are part of the problem. A report summarised by Angel One noted that government vehicles in Karnataka had racked up 82,208 traffic violations over three and a half years, with Rs 19.24 crore in unpaid fines, and 3,366 of those violations involved police vehicles. When the state does not pay the state, it is hard to argue that the missing ingredient is a better checkout page.

Our view: the court has correctly identified that friction at the point of payment is not the bottleneck. The bottleneck is the absence of consequence. Digital payment rails reduce the cost of complying, but they do nothing about the cost of not complying. The September order changes that second number by tying fines to services people actually need, such as renewing a registration certificate or transferring ownership.

What the order actually requires

The directions are worth listing because they turn a payments question into an identity and records question. The court asked authorities to impound vehicles found with unpaid challans during random checks, freeze vehicle-related online services until dues are cleared, refuse fitness and pollution certificates, block ownership transfers by flagging vehicles on the Parivahan portal, and suspend licences for repeat violators. It also asked the transport ministry to publish a monthly public dashboard under Section 136A of the Motor Vehicles Act and to examine a national road safety system linking VAHAN, SARATHI, e-Challan and accident databases.

Notice what this does. It converts an unsecured fine into something closer to a lien on a digital record. The same logic is already familiar from FASTag, where a toll is deducted before the boom barrier lifts. It is also the logic that makes property tax collections in well-run cities so high: you cannot sell or mortgage a flat cleanly with arrears on the municipal record.

A payment rail lowers the cost of paying. Only a linked record raises the cost of not paying.

Property tax: the quiet success story

If traffic fines show the gap, municipal property tax shows how far digital collection can go when the habit is in place. Pune Municipal Corporation collected Rs 1,395.86 crore in the first quarter of 2026-27 from 7,40,282 property owners, up from Rs 1,292.94 crore a year earlier, according to Punekar News. Of that, Rs 927.02 crore came online from 5,41,500 owners. Cheques and demand drafts brought Rs 354.46 crore and cash Rs 114.38 crore.

Navi Mumbai tells a similar story. Its civic body recorded Rs 876 crore in property tax in 2025-26, its highest ever, from 2,25,576 property owners, with about 61% paying online, as reported by the Free Press Journal. It also ran an amnesty in March that waived half of late payment penalties, which tells you something about what moves the laggards: a deadline and a discount, not a new app.

Worked example: who still pays by cheque

The Pune numbers let us do a small piece of arithmetic that the headline figures hide. Online, Rs 927.02 crore came from 5,41,500 owners, an average of about Rs 17,100 per owner. Offline, the remaining 1,98,782 owners paid Rs 468.84 crore through cheques, drafts and cash, an average of roughly Rs 23,600 each.

So by count, about 73% of Pune's paying owners went digital in the quarter. By value, only about 66% of rupees did. The offline group pays larger bills on average, which points to commercial properties, institutions and older owners writing cheques for big assessments. That matters for any city chasing a "100% digital" target. The last third of value is concentrated in a smaller group with larger, often more complicated bills, and it will not move because someone adds another QR code to the counter. It moves when the assessment, the dispute process and the receipt are all online too.

GST: the biggest bill most citizens never see

For businesses, the largest regular payment to government is GST, and here the digital share is effectively total because the system was built online from day one. Gross GST collections reached Rs 2,03,521 crore in September 2026, up 14.7% on a year earlier, with net collections after refunds of Rs 1,76,520 crore, according to provisional figures summarised by A2Z Taxcorp. For April to September, gross collections were Rs 12.46 lakh crore, up 11.6%.

The interesting payments detail is on the edges. Businesses generate a challan on the portal and pay through net banking, NEFT or RTGS, or over the counter for small amounts. From January 2024, GSTN began allowing credit and debit card payments in 10 states and union territories, with any transaction charges added on top of the tax, as noted in a Taxo summary of the GSTN advisory. For a small trader, that charge question is real. A card fee of even one or two percent on a quarterly liability is money that never reaches either the business or the exchequer. Readers working out which method suits them should check current charges with their bank and a tax adviser.

Bharat Connect and the long tail of municipal billers

The other quiet piece of plumbing is Bharat Connect, the interoperable bill payment system formerly known as BBPS. Its public category list shows 111 municipal tax billers, 68 water billers and 98 electricity billers, alongside a striking 20,410 education fee billers, according to the Bharat Connect category page. There is no separate category for traffic challans.

Read that list as a coverage map. India has thousands of urban and rural local bodies. Only a little over a hundred appear as municipal tax billers on the shared network. That means most local taxes are still paid through each body's own portal, a counter, or a payment gateway the city contracted separately. Every one of those is a separate integration, a separate receipt format and a separate reconciliation job for a municipal accounts officer. Getting more local bodies onto a shared biller network would do more for citizen convenience than any number of new apps, because it lets a resident pay from whichever bank or UPI app they already use.

At the central level, the equivalent is Bharatkosh, the Non-Tax Receipt Portal run by the Controller General of Accounts. The CGA describes it as a single online point for depositing fees, fines and other payments into the government account across civil ministries. It is a sensible design that few citizens have heard of.

What the UK learned running one payment layer for government

Britain offers a useful benchmark because it chose a shared platform early. GOV.UK Pay turned ten in September 2026 and has processed nearly GBP 10 billion over its life, according to the Government Digital Service. The first billion took about four years. It now handles that much in roughly five months. Almost 1,800 services across more than 600 organisations use it, including central departments, the NHS and police forces.

The weak spot looks familiar. Only 157 UK local authorities, 41% of the total, use the platform. Councils had their own contracts, their own gateways and their own reasons for staying put. India's version of this problem is larger by an order of magnitude, but the lesson transfers: a shared rail spreads fastest where the centre can mandate it, and slowest at the local level, where most citizen bills actually originate.

Why linking records beats building apps

There is a pattern across all of these examples. Collection rates rise when the payment is tied to something the citizen needs from the state. FASTag works because the barrier will not lift. Property tax works in cities that block transfers with arrears. GST works because input tax credit depends on it. The e-challan system lagged because, until September 2026, an unpaid fine had almost no effect on daily life.

This is also where the risks sit. Linking payment status to licences, registrations and utility connections gives the state a lot of leverage over individuals, and it amplifies the damage from bad data. A wrongly issued challan, a camera error or a fine attached to a sold vehicle becomes a blocked service. The court's direction for a monthly public dashboard is welcome. Equally important is a fast, online dispute channel with deadlines, so that enforcement and correction move at the same speed. The India Stack approach works best when the grievance layer is as well built as the payment layer.

The same rails that collect money also carry it outwards, through direct benefit transfers. A government that can both pay and collect through linked digital records can, in principle, net one against the other. That is a powerful idea and one that deserves careful public debate before anyone tries it.

What to watch next

• Late October 2026: the transport ministry's first steps on the monthly e-challan dashboard ordered by the Supreme Court on 28 September. The key metric is whether the recovery rate climbs from 53%.
• Early November 2026: October GST figures. Watch whether gross collections hold above Rs 2 lakh crore for a fourth straight month.
• Through Q3 2026-27: growth in Bharat Connect's municipal tax biller count from its current 111. Every new local body on the network is one less standalone portal.
• 31 March 2027: the financial year-end rush for property tax, when cities like Pune and Navi Mumbai typically run penalty waivers. Compare their online share against this year's 61% to 66%.

Citizens paying government is the least glamorous corner of payments, and also the one where the rails are most clearly ahead of the institutions. For a wider view of how other countries handle the outgoing side, see our coverage of government-to-person payments.

Picture a freelance designer in Hanoi paid in dollar stablecoins by a client in Toronto, or a software tester spending a month in Lisbon on a balance earned from American startups. For people like them, a stablecoin card looks like the obvious bridge: income arrives in digital dollars, and the card turns it into dong or euros at the till. It can work well. It can also leak money in small, avoidable ways. This guide walks through the parts that matter, without pointing you to any particular product.

First, understand what you are actually holding

A stablecoin card is a Visa or Mastercard debit or prepaid card whose balance is funded from stablecoins, usually USDC or USDT. If you want the full mechanics, our explainer on how crypto cards work covers issuers and settlement. For travellers, three questions come first:

• Who holds the coins? A custodial provider, or your own wallet? If the provider freezes your account while you are abroad, a backup card from a bank becomes essential.
• Which coin, on which blockchain? Sending USDC on the wrong network to a top-up address is one of the most common and least recoverable mistakes.
• Where is the card licensed? Programmes are issued under specific jurisdictions, and some restrict use or onboarding by country of residence.

The product category is growing fast. Visa reported more than 160 stablecoin-linked card programmes live in its fiscal second quarter of 2026, according to The Block. More choice is good, but it also means terms vary a lot from one card to the next.

The FX decision happens at the terminal

When you pay abroad in local currency, there are usually two conversions: stablecoin to dollars (often close to one to one) and dollars to the local currency. The second one is where most of the cost sits. Some programmes use the card network's daily rate plus a markup; others apply their own rate.

The bigger trap is dynamic currency conversion, when a terminal or ATM offers to charge you in your home currency instead of the local one. It sounds helpful. In practice the merchant's bank sets the rate, and it is rarely better than the one your card would use. The European Union has required providers offering this service to disclose the total charge since April 2020, and since April 2021 card issuers must send electronic reminders of their currency conversion charges, as FXC Intelligence summarised. Outside the EU, disclosure can be thinner.

When a machine asks whether you want to pay in dollars or local currency, local currency is almost always the cheaper answer.

ATMs, limits and the cash problem

Cash withdrawals are where crypto cards differ most. Some include a small free monthly allowance, others charge a percentage on every withdrawal, and the local ATM operator may add its own fee on top. Before a trip, check:

• the daily and monthly ATM limits, which are often lower than spending limits;
• whether the card supports contactless withdrawals and chip and PIN in the country you are visiting;
• how the card handles hotel and car rental pre-authorisations, which can lock up a slice of your balance for days.

Withdraw larger amounts less often if fees are per transaction, and keep a second card on a different network. A freelancer whose only card is suspended during a compliance review in another country has very few quick options.

Getting paid in stablecoins: the freelancer's side

For remote workers, the appeal is receiving dollar value from clients without waiting days for a wire. If you are comparing that with traditional transfers, our guide to the cost of sending money across borders shows how much fees and FX margins add up on bank routes.

A few habits keep it clean:

1. Invoice in a fiat amount. State the dollar value and the stablecoin and network you accept, so there is no argument about the rate on the day.
2. Keep a client-facing wallet address separate from the wallet that funds your card.
3. Hold a working float, not your savings, on the card. A custodial balance is not the same as a bank deposit.

Tax records: the part people skip

Income received in stablecoins is still income. Most tax authorities expect you to record its value in local currency on the day you received it. Then, depending on where you live, converting or spending crypto may itself count as a disposal.

In the United States, custodial brokers began issuing Form 1099-DA for tax year 2025, reporting gross proceeds from digital asset sales and exchanges. For assets acquired on or after 1 January 2026 and held at the same broker, cost basis is reported as well, according to Credit Karma's summary of the form. Older holdings, or coins moved between platforms, still depend on your own records.

A simple spreadsheet works: date, amount received, fiat value, source, and later the date and value of each conversion. Export your card statements every month; some apps keep only a limited history.

Pitfalls we see again and again

• Topping up on the wrong network. Always send a small test amount first.
• Holding volatile crypto as a spending balance. A card funded with ether can turn a coffee into a taxable event and a budget into a guess.
• Ignoring residency rules. Moving countries can change whether your card remains available to you.
• Single point of failure. One card, one phone, one app. Carry a backup card and keep your recovery details offline.
• Treating rewards as income. Token rewards can be taxable and can lose value.

When a crypto card makes sense

For someone already earning in stablecoins and spending in several currencies, a well-run card can remove a conversion step and a few days of waiting. For someone who earns in a local currency and travels twice a year, a low-markup bank card may be just as good with fewer moving parts. The honest answer is to compare the full route, from income to the coffee you buy, and pick the one with the fewest expensive hops.

This guide is general information, not financial, legal or tax advice. Rules differ by country; check with a qualified adviser where you are tax resident.

A backend developer in Indore sends her October invoice to a startup in Austin: $4,000, net 30. When the money lands, her bank will want a purpose code, her GST return will want proof that the payment arrived in foreign currency, and from this month the RBI wants a single new export form filed for all her software invoices. None of this is difficult. All of it is easy to get wrong, and the cost of getting it wrong is usually cash stuck with the tax department or a bank entry that never closes.

India sells a remarkable amount of services to the world. The RBI's survey of software and IT-enabled services exports, released on 18 September 2026, put 2025-26 exports at $221.4 billion, up 8.2 percent. The United States took 54.1 percent and Europe 31.8 percent. And 91.7 percent was delivered off-site, which is to say from desks in Indian cities, by companies large and small and by a long tail of individual freelancers. The rules that govern how that money comes home changed on 1 October 2026. This guide walks through them in the order an exporter meets them.

The new rulebook that started on 1 October

For a decade, India's export payment rules lived in a 2015 Master Direction patched by well over a hundred circulars. In January 2026 the RBI replaced them with the Foreign Exchange Management (Export and Import of Goods and Services) Regulations, 2026 and accompanying Directions, effective 1 October 2026. According to an EY regulatory alert and a Mondaq summary, the new framework supersedes the 2015 export and 2016 import Master Directions and 167 circulars. The changes that matter most to smaller exporters are these:

• Realisation period. Proceeds must be realised and repatriated within nine months, counted from the shipment date for goods and the invoice date for services. Exports invoiced in rupees get 12 months.
• One form instead of several. A single Export Declaration Form (EDF) replaces the old SOFTEX form for software and services. Software exporters can file one EDF covering all of a month's invoices, due within 30 days after month-end. Other service exporters may file on or before the date payment is received.
• Small reductions on self-declaration. For invoices up to Rs 10 lakh, banks may accept a reduction in the amount realised, for example a client's short payment, on the exporter's own declaration.
• Set-off. Goods export receivables can be set off against service import payables and the other way round.
• A sharper penalty for chronic delay. If proceeds stay unrealised more than a year past the due date, the exporter may ship only against full advance payment or an irrevocable letter of credit.
• Bank conduct. Banks must publish their policies and may not charge customers for delays caused by regulatory processes.

The 15-month window itself predates October. A 13 November 2025 amendment had already stretched realisation from nine months to fifteen and extended the deadline for shipping goods against an advance from one year to three. The 2026 regulations first kept the longer window, then a September 2026 amendment restored nine months from 1 October, with 12 months for rupee-invoiced exports. The rule has moved three times in a year, so check the RBI's current text before you plan around it.

Step one: be paid through the right channel

Export proceeds must arrive through an authorised dealer bank, in convertible foreign currency or in rupees where the RBI permits it. In practice exporters use one of four routes: a SWIFT wire straight into an Indian bank account; a foreign-currency collection account that receives local transfers abroad and remits to India; a regulated cross-border payment aggregator; or a card or wallet platform that ultimately pays out through an Indian bank. Each route must end with a bank credit that is tagged as export income and documented.

When you compare routes, compare four things, not just the headline fee. The exchange margin against the mid-market rate is usually the largest cost and the least visible. Fixed fees per transfer matter more on small invoices. Intermediary deductions on SWIFT wires can shave an unpredictable amount off the top, a problem we cover in our explainer on correspondent banking. And the route must produce the right paperwork: an inward remittance advice with your name, the foreign remitter, the amount and the purpose code. A cheaper route that leaves you without that document is not cheaper once your GST refund stalls.

Step two: get the purpose code right

Every inward remittance is classified with an RBI purpose code. For services exporters the common ones, as listed in a purpose code directory, include P0802 for software consulting and implementation, P0807 for off-site software exports, P0803 for data processing, P1006 for business and management consulting, P1007 for advertising and market research, and P1008 for research and development services. Pick the code that describes the work, not the size of the payment.

The most expensive mistake is letting a business receipt be coded as a personal remittance. It may still reach your account, but it will not count as an export of services, which can unravel GST zero-rating and leave you explaining a mismatch between your returns and your bank's reporting. If a client's payment arrives with the wrong code, ask your bank to correct it promptly rather than at year-end.

Step three: the documents, and who issues them

Keep all of it. A guide for Indian freelancers recommends retaining invoices, contracts, delivery emails, bank credit advices and remittance advices for at least six years and reconciling them monthly. That sounds like overkill until a refund officer asks about an invoice from three years ago.

The paperwork is not the obstacle. The obstacle is doing it in March for a year of invoices.

Step four: GST, the LUT and why it decides your cash flow

Exports of services are zero-rated under GST, but only if five conditions in section 2(6) of the IGST Act are all met. As a GST explainer sets out: the supplier is in India; the recipient is outside India; the place of supply is outside India; payment is received in convertible foreign exchange, or in rupees where the RBI allows it; and the two parties are not merely establishments of the same entity. An Indian subsidiary billing its foreign parent counts as a separate person, but a branch billing its own head office does not.

If you are GST-registered you then have a choice. File a Letter of Undertaking and invoice without tax, or pay 18 percent IGST on each export invoice and claim it back. According to an export compliance guide, LUT approval is usually immediate or takes a day or two, while IGST refunds nominally target 60 days and in practice often take 45 to 90. One trap catches many consultants: if your service falls under the place-of-supply rule for intermediaries, the place of supply may be in India, and zero-rating may not apply at all. That is a question for a qualified tax adviser.

A worked example: one developer, one quarter

Return to the Indore developer. She bills $4,000 a month and her client pays 30 days after invoice. We use the rupee's 9 October 2026 close of 96.73 for illustration.

1. Rupee value per invoice: $4,000 x 96.73 = Rs 3,86,920.
2. With an LUT on file: she invoices without GST. Her bank credits about Rs 3,86,920, less the exchange margin and any fees of her chosen channel.
3. Without an LUT: she must pay IGST at 18 percent, Rs 69,646 per invoice, from her own funds when she files her return, then wait for a refund. With three monthly invoices in the refund queue at a 90-day turnaround, about Rs 2.09 lakh of her money sits with the government at any moment.
4. Realisation deadline: an invoice dated 15 October 2026 must be realised by 15 January 2028. Had she invoiced in rupees, the deadline would be 15 April 2028.
5. EDF deadline: her October software invoices go on one EDF due by 30 November 2026.
6. If the client short-pays by $150 on a disputed sprint, the reduction is far below the Rs 10 lakh self-declaration threshold, so her bank can close the entry on her declaration.

The difference between steps two and three is pure working capital. For a solo exporter, Rs 2 lakh tied up in a refund queue is often a month of rent and living costs. Filing the LUT in April each year is the highest-return ten minutes in her calendar.

Currency choice and the October hedging rules

The same RBI survey shows that 72.6 percent of India's software exports were invoiced in dollars, 9.6 percent in euros, 6.3 percent in rupees and 6.0 percent in sterling. The new 18-month window for rupee invoices is a nudge toward the rupee, but most foreign clients will still prefer their own currency.

Exporters who hedge should note the RBI's 10 October 2026 measures. The limit for hedging without documenting the exposure fell from $100 million to $5 million, and cancelled contracts can no longer be rebooked. The new 20 percent reserve applies to contracts where the customer buys foreign currency, so an exporter selling dollars forward is not directly affected. Our guide to FX risk for small businesses works through hedging choices with rupee numbers.

What to watch next

• 30 November 2026: the first monthly EDF deadline for software exporters' October invoices under the new regime. Expect banks and STPI centres to iron out filing glitches.
• Late 2026: banks must publish their export-import policies and standard procedures under the new Directions. Read your bank's version; it tells you what documents it will actually ask for.
• Rupee hedging costs: how banks implement the October undertaking requirement and the rebooking ban for exporters with forward books.
• 31 March 2027: every LUT filed for 2026-27 expires. File the 2027-28 undertaking before your first April invoice.
• September 2027: the RBI's next annual software exports survey will show whether rupee invoicing moved off its 6.3 percent share.

For the wider picture of how money moves into and out of India, see our guides to FEMA rules in 2026 and to cross-border B2B payments. This article explains rules in general terms; it is not tax or legal advice, and exporters should confirm their position with a chartered accountant or their bank.

Ask anyone who sends money home what it's for, and you'll rarely hear "general expenses." You'll hear: Amma's electricity bill. My brother's college fees. Papa's insurance premium is due on the 10th. The money already has a name before it leaves.

The hidden second journey

Today, most of that money takes two journeys. First it crosses a border into someone's bank account. Then, a few days later, that person has to log in, find the biller, remember the account number and pay it. Sometimes they do it the same day. Sometimes the bank is slow, the parent isn't comfortable with apps, or the money gets used for something more urgent and the bill goes overdue.

Every extra hop is a chance for something to go wrong, and an extra phone call from abroad asking, "Did you pay it?"

What families actually pay for

Research on remittances keeps finding the same pattern: about three-quarters of the money goes to immediate household needs. In practice that means a small, very predictable list:

These aren't random transfers. They're recurring obligations, the same ones, every month, often on the same date. That makes them perfect for a better system.

Why paying the bill directly is better

• It arrives where it's meant to. The biller gets paid, full stop. No one has to forward it.
• It's on time. No late fees, no disconnection notices, no stress for parents who don't do online banking.
• It's proof. A receipt from the biller is better than "I sent it" on a group chat.
• It's dignity. Many older parents feel awkward asking for money. A paid bill just shows up as paid.

Money is an emotion. A paid bill says "I'm looking after you" in a way a bank transfer never quite does.

Countries are already building for this

India's Bharat Connect network has allowed Non-Resident Indians to pay Indian bills directly since 2022, covering things like utilities, school fees, municipal taxes and insurance. Banks and fintechs in the Gulf, the UK and Canada have plugged into it. We cover how that works in our Bharat Connect guide. Global bill-payment networks reach tens of thousands of billers in other countries too, from mobile top-ups in Kenya to electricity in the Philippines.

Coverage is still patchy, though. Each service supports a different set of countries and billers, and exchange-rate margins on the cross-border leg vary widely.

Why the framing matters

Statistics usually count remittances as transfers between people. Seen from the household's side, much of that money behaves more like a standing order: rent, school, power, medicine, a parent's phone plan, due on predictable dates. Measuring it that way would make it easier to see where delays and late fees actually hurt.

It also changes what a good service looks like. The useful questions are whether the biller was paid on time, what exchange rate was applied, and whether the family got a receipt they can check. Those are worth asking of any provider, old or new.

For a decade, every fintech pitch deck had a slide that said SWIFT is slow. The network's own data now tells a more awkward story: most payments cross borders in under an hour, and the delay sits at the end of the chain, inside domestic banks. That reframing explains almost everything SWIFT has done since, from retiring its old message formats to experimenting with a shared ledger.

What SWIFT is, and what it is not

SWIFT does not move money. It is a messaging cooperative that more than 11,500 financial institutions use to tell each other where money should go. The money itself moves across correspondent accounts that banks hold with one another. On an average day the network carries over 53 million FIN messages, covering payments, securities and trade instructions.

It is also owned by its members, which shapes its incentives. SWIFT earns fees on messages, so it has every reason to keep banks on its network rather than see them drift to rival systems, and its member banks have a say in how fast change arrives. That cooperative structure helps explain why upgrades take years and why, once agreed, they tend to stick across the whole industry.

That distinction matters because it shapes what SWIFT can fix. It can standardise messages, track payments and set service rules for its members. It cannot force a bank in Lagos to credit a customer's account faster, and it cannot make nostro accounts cheaper to fund. Our guide to correspondent banking covers that plumbing in detail.

The speed numbers, read carefully

In October 2024 SWIFT reported that 90% of cross-border payments on its network reached the destination bank within an hour. Speeds into the euro area hit 92%, the Middle East 88% and Africa 87%. Measured against the G20 goal of settling 75% of cross-border payments within an hour by 2027, SWIFT said it was above target for all but two of its 40 largest country routes.

The catch is in the same release. Only 43% of payments reached the end customer's account within an hour. The bottleneck is the receiving bank: batch processing, manual compliance reviews, cut-off times and weekends. A worker in Riyadh might see the money leave in minutes while the family in Dhaka waits until the next business day.

The Financial Stability Board's own measurement is harsher. Its 2025 progress report found that only about 35% of cross-border payments met the one-hour goal globally and concluded that satisfactory improvement by 2027 was unlikely, as The Paypers summarised. The two figures measure different things, and the gap between them is the last mile.

ISO 20022: the end of MT

On 22 November 2025, the coexistence period for cross-border payments ended and SWIFT retired the old MT103 and MT202 instructions in favour of ISO 20022 messages, according to Payment Expert. Coexistence had begun in March 2023 after an earlier delay. In the United States, the Federal Reserve moved Fedwire Funds to ISO 20022 in July 2025.

Why should anyone outside a bank operations team care? Because the old MT format was a fixed set of short text fields. Names were truncated, addresses were stuffed into free text, and the purpose of a payment often vanished along the way. ISO 20022 carries rich, structured data: full party details, remittance information, purpose codes. That makes sanctions screening more accurate and reduces the false positives that hold payments in compliance queues, which is exactly where the last-mile delay lives.

The migration is not fully done. Translation tools let lagging banks keep running MT internally, and critics argue those converters strip out the structured data that was the point. The next deadline is November 2026, when unstructured postal addresses stop being accepted. Banks have to send structured or hybrid addresses, and the ones that relied on converters will feel it.

SWIFT's speed problem moved from the network to the receiving bank. Its strategy is to chase it there.

A blockchain ledger, from the incumbent

At Sibos in September 2025, SWIFT announced it would add a blockchain-based shared ledger to its infrastructure, starting with real-time, round-the-clock cross-border payments. Consensys built the conceptual prototype, and more than 30 institutions, including Santander, Bank of America, BNP Paribas and HSBC, signed up to give feedback and help implement it, The Block reported.

The ledger is designed to record transactions between banks, enforce rules through smart contracts and handle tokenised value while staying interoperable with existing systems. No go-live date has been published. Read plainly, it is a defensive move. If banks start settling with stablecoins or tokenised deposits on public networks, SWIFT wants to be the trusted layer that connects those networks, rather than the legacy system they route around.

Retail payments get their own rulebook

The other notable 2026 change is aimed at consumers. In March 2026 SWIFT launched a retail framework under which participating banks commit to fees known upfront, full value delivered to the recipient and the fastest available speed. More than 25 banks were scheduled to go live by the end of June across corridors including India, Bangladesh, China, the UK, the US and Canada, with over 50 banks supporting the scheme.

Full value delivery is the quiet headline. For decades, an intermediary bank could deduct a fee in transit, so a student's tuition payment arrived short and nobody could say why. A rule that bans that, if enforced, removes one of the oldest complaints in cross-border payments.

What to watch through 2027

• The customer-credit number. If the 43% figure climbs toward the bank-to-bank 90%, the last-mile problem is being solved. If it stalls, expect regulators to lean on receiving banks.
• Ledger pilots. The first live transactions on the shared ledger, and which tokens it carries, will show whether SWIFT is serious about public-chain interoperability.
• Structured data in practice. November 2026 will reveal how many banks still depend on converters.

SWIFT in 2026 is neither the dinosaur of fintech folklore nor a finished product. It is a network that fixed much of its own core and is now trying to drag its weakest members forward. Whether that is enough against blockchain-native rivals depends less on SWIFT than on the banks at the far end of each payment.

A spare-parts dealer in Kano's Sabon Gari market buys gearboxes from a supplier in Guangzhou, pays a cousin's logistics firm in Cotonou, and collects from mechanics across northern Nigeria. Almost none of it touches a card network. Customers pay by bank transfer on Nigeria's instant system, the Benin leg runs on a mix of cash and mobile money, and the China invoice is increasingly settled in dollar stablecoins bought through an over-the-counter desk. That is how payments in Africa look in 2026: not one system, but a stack of mobile money, national instant rails, a pan-African settlement network still finding its feet, and a growing layer of dollar tokens filling the gaps the banks leave.

This guide works through each layer with the most recent data, and argues that the big change of the next few years is not more mobile money. It is the fight over who settles cross-border trade.

Mobile money is still the base layer

Sub-Saharan Africa is the centre of the world's mobile money industry. The GSMA's State of the Industry Report 2025 counted 1.1 billion registered accounts in the region in 2024, up 19 percent and about half the global total of 2.1 billion. Transaction value in the region reached $1.1 trillion, around two-thirds of the global $1.68 trillion. There were 165 live services in Sub-Saharan Africa alone.

Kenya remains the reference market. The Central Bank of Kenya reported 95.31 million registered mobile money accounts in August 2026, a count inflated by the fact that many people hold several wallets and SIM cards, served by 563,605 active agents. Those agents handled 227.41 million cash-in and cash-out transactions worth KSh 761.13 billion in that single month. The agent network is the clue to what mobile money really is: a bridge between cash and digital, not a replacement for cash. Our M-Pesa explainer covers how the model works.

Mobile money also carries cross-border flows. The GSMA put international remittances through mobile money at $34 billion in 2024, up 22 percent, across about 270 million transactions.

The averages show what kind of money this is. Globally, the GSMA counted around 108 billion mobile money transactions in 2024 worth $1.68 trillion, which works out to roughly $15.60 per transaction. In Kenya's August 2026 data, the average agent cash transaction was about KSh 3,350. These are airtime top-ups, school fees, market purchases and wages for day labourers. Mobile money is a high-volume, low-value system, and that is precisely why it reached people banks never served.

It also has limits that matter for the next phase. Wallet balances are usually capped, so they are poorly suited to business payments of any size. Interest on balances is rare. Cross-border use depends on bilateral deals between operators. And because most wallets are denominated in the local currency, they offer no protection when that currency falls, which is where the dollar layer described below comes in.

Instant payment systems are catching up

Alongside mobile money, central banks have built national instant payment systems that link banks, and increasingly wallets. The AfricaNenda Foundation's SIIPS 2025 report counts 36 live instant payment systems in Africa. Eleven now support cross-border transactions, up from six a year earlier, and ten support government payments to citizens. Nigeria's NIBSS Instant Payments, launched in 2011, is the continent's oldest large-scale example and is the rail most Nigerian businesses actually use day to day.

The policy direction is interoperability. A wallet user should be able to pay a bank account holder or a user of a competing wallet without friction. Where that has happened, volumes have grown quickly. Where mobile money operators and banks still run closed loops, users carry several wallets and move money between them through agents. For the global comparison, see instant payments around the world.

PAPSS and the intra-African trade problem

Here is the core inefficiency. A Ghanaian importer paying a Kenyan exporter has traditionally had to buy US dollars, send them through correspondent banks in New York or London, and have the Kenyan bank convert them back into shillings. Two conversions, several intermediaries, and days of delay for a trade that never involved America. Each conversion carries a spread, and each intermediary bank can hold the payment for compliance checks. For a small exporter, the result is working capital locked up in transit and an FX cost that is hard to see on the invoice.

The Pan-African Payment and Settlement System, launched publicly in January 2022 by Afreximbank and the African Union, is meant to fix that by settling in local currencies across borders. Its network page, last updated in July 2026, shows 21 central banks including the regional BEAC, among them Nigeria, Ghana, Kenya, Egypt, Morocco, Zambia and Zimbabwe, plus a dozen national switches.

PAPSS has solved the plumbing. It has not yet solved the liquidity.

Our view is that PAPSS is promising but under-used. Connecting central banks is the easy part. The hard part is FX liquidity between thinly traded currency pairs, such as the Ghanaian cedi and the Kenyan shilling, without falling back on the dollar. Until there is a deep market for those pairs, many banks will continue routing through dollars, and businesses will keep looking for faster dollar alternatives. Correspondent banking's role is explained in our correspondent banking guide.

Stablecoins: the dollar layer nobody planned

That gap is where stablecoins have grown. Chainalysis's 2025 Sub-Saharan Africa report found the region received over $205 billion in on-chain value between July 2024 and June 2025, up roughly 52 percent, making it the third fastest-growing region in the world. Nigeria alone received over $92.1 billion, nearly triple South Africa in second place. Kenya, Ethiopia and Ghana rank in the regional top five.

The use cases are specific. Chainalysis describes stablecoins supporting regular multi-million-dollar transfers tied to trade between Africa, the Middle East and Asia, and a growing business-to-business sector in Nigeria and South Africa. At the retail end, over 8 percent of the region's transfer value was in amounts under $10,000, against about 6 percent elsewhere. USDT serves as a practical dollar substitute where official dollar access is rationed and the gap between official and parallel exchange rates is wide. Chainalysis links spikes in activity to naira devaluations, including one that drove monthly regional volume close to $25 billion.

Regulators are responding with licensing regimes rather than bans, though approaches differ widely. We track them in stablecoin regulation by country.

Country snapshot

Funding figures are total equity and debt from Partech; on-chain data is from Chainalysis for July 2024 to June 2025.

Where the money is going: fintech funding

African tech funding recovered in 2025. Partech's 2025 Africa report recorded about $4.1 billion in total equity and debt, up 25 percent from $3.25 billion, across 570 deals. Debt reached a record $1.64 billion, around 41 percent of the total. Kenya, South Africa, Egypt and Nigeria took about 72 percent of the money.

Fintech, long the dominant sector, went the other way. It raised $1.49 billion in total, down 12 percent, and only $769 million in equity, down 43 percent. Investors are not abandoning payments. They are getting pickier, preferring revenue-backed debt over equity, and favouring infrastructure and B2B flows over another consumer wallet. That fits the trade-settlement story above.

A worked example: the cost of moving $200

Sub-Saharan Africa has long been the world's most expensive region to send money to. AfricaNenda, citing World Bank data, put the 2024 average at 8.45 percent of a $200 transfer. Here is what that looks like in practice.

1. A worker in Johannesburg sends $200 to family in Lilongwe. At 8.45 percent, the cost is about $16.90, leaving roughly $183 to arrive.
2. The GSMA's figures imply an average mobile money international remittance of about $126 ($34 billion over 270 million transactions). At the same 8.45 percent, that transfer would cost about $10.65.
3. Compare the 6.36 percent global average in Q3 2025: on $200, the cost would be $12.72, about $4.18 less per transfer than the African average.
4. For a family receiving $200 a month, that gap alone is about $50 a year, before any FX margin hidden in the rate.

The cheapest routes now tend to be wallet-to-wallet, where both ends are digital and no agent cash-out is needed. The expensive part is the last mile, which is why mobile money agents remain so important. For more on corridor costs see why sending money to Africa costs so much.

What to watch next

• Monthly CBK data: Kenya's agent cash-out volumes are the best real-time indicator of whether digital money is staying digital.
• Late 2026 and 2027: new central banks joining PAPSS and, more importantly, whether its local-currency FX market develops real depth.
• December 2026: the World Bank's annual remittance estimates, which will show whether Egypt keeps its lead over Nigeria.
• Early 2027: the GSMA's next State of the Industry report on mobile money and Partech's 2026 funding data.
• Ongoing: stablecoin licensing in Nigeria, Kenya and South Africa, which will decide whether the dollar layer moves into the regulated system or stays in OTC markets. Businesses weighing this can start with stablecoin payments for business.

This article describes market data and is not investment advice. Rules on crypto assets differ by country; check with a qualified adviser before acting.

For most of their history, stablecoin issuers were private companies that published reserve reports and little else. Circle's listing on the New York Stock Exchange in June 2025 changed that. For the first time, investors could vote every day on what a regulated dollar-token business is worth. Sixteen months later, the verdict is in, and it is more nuanced than either the euphoric debut or the slump that followed.

The debut: a stablecoin company priced like a tech darling

Circle priced its offering at $31 a share, above the range it had marketed, and sold about 34 million shares to raise roughly $1.1 billion. When trading opened on 5 June 2025 the stock started at $69.50 and closed the day at $83.29, a 168% first-day gain that left the company worth around $16.7 billion, according to Crunchbase News.

The rally kept going. Within three weeks the shares hit a record of about $299, more than nine times the offer price. At that peak the market was valuing Circle well above the amount of USDC it had in circulation, which was a striking thing to pay for a business whose main income is interest on Treasury bills.

The comedown: rates, distribution costs and a sceptical BIS

The turn came quickly. On the day the Bank for International Settlements published its 2025 annual report arguing that stablecoins fail key tests of sound money, Circle's shares fell more than 15%, as The Block reported. By early July the lead underwriter, JPMorgan, had cut its rating to neutral, and the stock had dropped by more than a third from its peak.

Three worries did the damage, and they are still the right lens for the company:

• Interest-rate exposure. Nearly all of Circle's revenue is reserve income, so every Federal Reserve cut lowers it. Circle's reserve return rate was 3.48% in the second quarter of 2026, down 66 basis points from a year earlier.
• Distribution costs. Circle pays a large share of its income to Coinbase and other partners. In the second quarter of 2026, distribution, transaction and other costs were $412 million against $701 million of revenue.
• Flat supply. USDC circulation was $73.3 billion at the end of June 2026, down 5% from the previous quarter even though it was up 19% on the year.

By December 2025 the shares were down nearly 70% from their high, according to The Motley Fool. For the full year 2025, Circle reported revenue of $2.7 billion, up 64%, but a net loss from continuing operations of $70 million, largely because of $424 million of stock-based compensation tied to the listing.

The 2026 rebound: charters and a cheaper starting point

This year has been a story of regulatory progress lifting a battered stock. Circle won a national trust charter in early July 2026 and a New York trust charter at the end of the month. Shares rose 52.6% in August, helped by bullish analyst notes and a 20% rise in bitcoin, according to an S&P Global Market Intelligence figure cited by Nasdaq. Even after that rally, the stock was still about 61% below its 2025 peak in early September.

On 9 October 2026 Circle closed at $84.51, giving a market value of about $21.6 billion. Its 52-week range ran from $49.90 to $147.44, and the shares were down roughly 44% over twelve months, according to MarketScreener data. Circle's Arc blockchain, which went live in September with BlackRock, Visa and Mastercard among its validators, has given the bulls a new story about network revenue.

Public markets have decided that a stablecoin issuer is a rate-sensitive financial company first and a technology platform second.

What the share price is really telling us

Strip out the noise and four messages stand out.

1. The float is valued like a deposit base. At about $21.6 billion, Circle trades at roughly 0.3 times its USDC in circulation. At the mid-2025 peak it was above one. Investors now price USDC much as they would a bank's low-cost deposits: valuable, but only as valuable as the spread they earn.
2. Distribution is the weak spot. Markets reward businesses that own their customers. Circle's reliance on partners for reach shows up directly in its margins, which we unpack in how stablecoin issuers make money.
3. Regulation is a premium, not a moat. Charters and licences lifted the stock, but they are available to competitors too, including the bank consortia and card networks now issuing their own tokens.
4. Crypto beta has not gone away. Circle still moves with bitcoin. Short interest was around 13% of shares in early September 2026, a sign of how divided investors remain.

The wider listing wave

Circle's debut opened the door for others. The exchange operator Bullish raised about $1.11 billion in its August 2025 New York listing, and Gemini raised about $425 million in September 2025 after pricing at $28 a share. Each has since traded on its own fundamentals, but together they gave public investors their first broad set of regulated crypto businesses to compare.

The most interesting comparison sits outside the stock market. Tether, far larger and more profitable than Circle, explored raising money at a $500 billion valuation in late 2025 before reportedly scaling back the plan after investor pushback. The gap between that ambition and Circle's roughly $21 billion market value shows how hard it still is to agree on what a dollar-token franchise is worth. For context on the two coins, see USDT vs USDC.

What comes next

Circle reports third-quarter results on 4 November 2026. The numbers to watch are USDC circulation, the reserve return rate after further rate cuts, the share of revenue paid to distributors, and any early income from Arc. The bigger question is the one public markets have been asking since June 2025: can a stablecoin issuer become a payments network that earns fees, rather than a money-market fund that earns spread? The share price will keep swinging until that is answered. Our 2026 market report tracks the supply side of the story.

Share prices and market values are as of the dates stated. This article is for information only and is not a recommendation to buy or sell any security.

At one minute before midnight on 31 March 2025, Binance switched off spot trading pairs for nine stablecoins for users in the European Economic Area, among them USDT, FDUSD, TUSD, DAI and PAXG. The world's largest stablecoin had not been made illegal to hold. It simply had no authorised issuer under the EU's Markets in Crypto-Assets Regulation, and platforms that wanted to keep their own licences could no longer list it. Fifteen months later, with MiCA's last transitional window having shut on 1 July 2026, that sorting is complete.

MiCA is often described as one stablecoin law. It is really two regimes with different rules, different supervisors and very different commercial outcomes. Getting the category right is the first question any issuer, exchange or corporate treasurer in Europe has to answer.

Two boxes: e-money tokens and asset-referenced tokens

MiCA's stablecoin rules, in Titles III and IV, have applied since 30 June 2024. Everything that aims to hold a stable value lands in one of two boxes.

• E-money tokens (EMTs) reference a single official currency. USDC, EURC and the bank-issued euro coins sit here. Only an authorised electronic money institution or a credit institution may issue one, and holders get a claim on the issuer at par.
• Asset-referenced tokens (ARTs) reference anything else: a basket of currencies, a commodity such as gold, other crypto-assets or a mix. A gold-linked token would be an ART.

The EMT box is the one the market has chosen. As of mid-2026, no ART issuer had been authorised anywhere in the EU, according to independent trackers of the ESMA register. The ART regime is heavier, closer to running a small bank's treasury, and without a clear payments use it has drawn no takers.

The reserve and own-funds figures come straight from Article 54 and Article 45 of the regulation. Readers comparing this with the American approach should start with our GENIUS Act and MiCA explainer.

The significance test, criterion by criterion

The European Banking Authority moves into the frame once a token becomes "significant". Article 43 sets seven criteria, and the same test is applied to EMTs. A token is classified when at least three are met:

1. More than 10 million holders.
2. Value issued, market capitalisation or reserve size above EUR 5 billion.
3. Average daily transactions above 2.5 million in number and EUR 500 million in value.
4. The issuer is a designated gatekeeper under the Digital Markets Act.
5. Significant international activity, including use in payments and remittances.
6. Interconnectedness with the financial system.
7. The issuer offers at least one other token or crypto-asset service.

Read that list as a large dollar coin would. A token with EUR 5 billion-plus in circulation already trips criterion two. A global issuer offering both a dollar and a euro token trips criterion seven. International use trips criterion five almost by definition. In other words, any big global EMT is likely to be significant on qualitative grounds alone, well before it reaches 10 million EU holders. The gatekeeper criterion is aimed squarely at the largest technology platforms.

Worked example: what significance costs

Take a hypothetical asset-referenced token with an average reserve of EUR 6 billion over the past six months. Under Article 35, own funds must be the highest of EUR 350,000, 2% of the average reserve, or a quarter of last year's fixed overheads. Assume overheads of EUR 80 million, so a quarter is EUR 20 million.

• Floor: EUR 0.35 million
• 2% of EUR 6 billion: EUR 120 million (the binding figure)
• Quarter of overheads: EUR 20 million

If the token is classified as significant, Article 45 lifts the percentage to 3%, so own funds jump to EUR 180 million, an extra EUR 60 million of capital that earns nothing for holders. The draft liquidity standards must also require a minimum deposit in each referenced currency of no less than 60% of the amount referenced in that currency.

For an EMT the arithmetic is different but just as concrete. An issuer holding EUR 1 billion of funds from token holders must keep at least EUR 300 million in separate accounts at credit institutions under Article 54, with the remainder in high-quality liquid assets. Because EU banks hold those deposits, and because issuers cannot pay interest to holders, the margin left after costs depends heavily on deposit rates. Our piece on how stablecoin issuers make money shows why that 30% floor squeezes European economics relative to a US issuer holding mostly Treasury bills.

These are stylised numbers to show the mechanics. Firms considering issuance should take advice from qualified EU counsel.

The cap that worries dollar coins

The clause that gets less attention but matters most for payments is Article 23. If an ART's estimated quarterly average use as a means of exchange inside a single currency area exceeds 1 million transactions and EUR 200 million a day, the issuer must stop issuing and present a plan within 40 working days to bring usage back below those levels. Article 58 applies the same mechanism to EMTs denominated in a currency that is not an official currency of a member state, which means dollar EMTs.

MiCA welcomes dollar stablecoins as trading tools and quietly caps them as money.

The policy intent is plain: the EU does not want a dollar token becoming everyday money inside the euro area. A dollar coin can be held, traded and used for settlement, but if it starts powering millions of daily retail payments in the eurozone, its issuer hits a wall. For merchants and payment firms planning on-chain acceptance in Europe, that is a strong argument for euro EMTs at the checkout.

How USDT left European order books

The delisting wave followed a clear sequence. On 17 January 2025, ESMA and the European Commission told national authorities to bring crypto-asset service providers into line on non-compliant ARTs and EMTs as soon as possible and no later than the end of the first quarter. Crypto.com set a 31 January 2025 date for delisting USDT and nine other tokens for European users, as The Defiant reported. Binance followed on 31 March, allowing sell-only trading until the cut-off, per BeInCrypto via TradingView.

Two details are often lost. First, holding USDT in a self-custody wallet in the EU remains lawful; the rules bite on offers to the public and on services provided by licensed platforms. Second, the restriction is about the issuer, not the asset design. USDT has no MiCA-authorised issuer, so it does not appear in the EMT register at all. Our USDT vs USDC comparison covers why one of the two big dollar coins chose to come inside the regime and the other did not.

Who is inside the tent

The authorised side is now a recognisable roster: Circle with USDC and EURC, Societe Generale-Forge with EURCV and USDCV, Paxos's EU entity, Banking Circle's EURI, Monerium's EURe, StablR, Quantoz and AllUnity among them. Most tokens are euro-denominated, with dollar, sterling and Swiss franc tokens in the mix.

Scale is the problem. The ECB's April 2026 Macroprudential Bulletin, cited by Forbes, put euro stablecoins at about EUR 450 million in January 2026, up from EUR 50 million two years earlier but still roughly 0.2% of a stablecoin market worth around $300 billion. A ninefold rise sounds impressive until you set it next to the dollar side. MiCA has made the EU the most legally certain place to issue a euro coin; it has not yet made anyone want one at scale. For newer entrants, see beyond USDT and USDC.

Platforms: the end of grandfathering

The issuer rules have been live since 2024. What changed on 1 July 2026 was the platform side. Member states could let existing crypto firms keep operating under national rules for up to 18 months while they sought a MiCA authorisation. Some cut that short: the Netherlands, Finland, Latvia, Hungary and Slovenia closed their windows on 30 June 2025 and Sweden on 30 September 2025. The final outer limit was 1 July 2026, according to Elliptic, which counted 213 authorised CASP entries across 23 jurisdictions on ESMA's interim register just before the deadline. December 2025 was the busiest month, with 41 authorisations. Every one of those licensed platforms is bound to list only compliant stablecoins. For the licence types involved, see CASP, VASP and MTL licences explained.

What to watch next

• Second half of 2026: the Qivalis consortium of European banks, which started with nine banks in September 2025 and added BNP Paribas by December, targets launch of a euro stablecoin. It is the first real test of whether bank-issued euro EMTs can grow.
• EBA significance decisions: the first formal classification of a large EMT as significant would bring EBA into day-to-day supervision and six-monthly independent audits under Article 58.
• Digital euro legislation: the ECB moved the project to its next phase on 30 October 2025. A pilot is possible from mid-2027 and issuance no earlier than 2029, if the regulation passes in 2026. See our digital euro explainer.
• Transatlantic equivalence: the US GENIUS Act requires Treasury to judge whether foreign regimes are comparable. MiCA is the obvious first candidate, which matters for euro EMTs seeking US distribution.

Our view: MiCA did what it set out to do on legal certainty and consumer claims, and its usage caps are a deliberate monetary-sovereignty tool rather than an accident. The open question is commercial. If euro coins stay below one percent of the market, Europe will have built a fine regulatory house that mostly hosts dollar tenants. For a wider view across jurisdictions, see stablecoin regulation by country.

If you measured the future of money by how many central banks are studying it, CBDCs would look unstoppable. Almost every major economy has a team on it. If you measured by how many ordinary people use one, you would conclude they barely exist. Both readings are correct, and the gap between them is the most interesting thing about digital currencies in 2026.

The headline numbers

Those figures come from the Atlantic Council's CBDC Tracker, which is the most widely cited count. As of its May 2026 update the tracker records 41 pilot projects, down from the record 49 pilots Axios reported in July 2025, and 77 jurisdictions in an advanced stage that spans development, pilot and launch. The three launched currencies belong to the Bahamas, Jamaica and Nigeria. None of them has yet shown sustained mass use.

The Bank for International Settlements tells a similar story from the inside. Its 2024 survey, published in August 2025, found that 91% of the 93 central banks that responded were exploring a retail CBDC, a wholesale CBDC or both. More than one in three said stablecoins and other crypto assets had pushed them to speed up.

Research is cheap; launching is hard

Why the gulf between 146 and 3? Launching a retail CBDC forces a central bank to answer questions it can comfortably avoid in a research paper. Who runs the wallets? What happens to bank deposits if people move savings into central bank money? How much privacy should a payment have? What does the central bank do when a phone is lost?

Every serious project has hit at least one of these. The euro area's answer is holding limits and no interest. India's answer is a wallet that pays nothing and caps transfers in some apps. China, as we explain below, has just changed its answer entirely.

The two giants: China and India

The largest pilots by far are China's e-CNY and India's e₹. China's central bank said that by the end of September 2025, the e-CNY had processed 3.32 billion transactions worth a cumulative 14.2 trillion yuan, with 225 million personal wallets opened. From 1 January 2026, e-CNY wallet balances became interest-bearing liabilities of commercial banks, a structural shift that turns digital cash into something closer to a deposit.

India's retail e₹ had a little over 8 million users and more than 120 million transactions by December 2025. It is growing, but at nothing like the pace of UPI. See our digital rupee explainer for the figures.

Europe is the one to watch

The digital euro is the most consequential retail CBDC still on the drawing board, because it would cover a large, rich, cash-heavy economy. The European Parliament voted 416 to 169 in July 2026 to open negotiations on the legislation, with a deal targeted by the end of 2026. The ECB has pencilled in a pilot for 2027 and possible issuance in 2029. We walk through the details in digital euro explained.

The CBDC story of 2026 is less about coins for consumers and more about pipes between banks.

The quiet shift to wholesale

The BIS survey noted that wholesale CBDC work is, overall, further along than retail. That matches what is happening on the ground:

• mBridge, the multi-central-bank platform linking China, Hong Kong, Thailand and the UAE, had passed about $55.5 billion in cumulative cross-border value by January 2026, according to Atlantic Council data cited by Reuters. The PBOC's own count, almost 500 billion yuan by the end of 2025, is higher.
• Project Agorá, led by the BIS with seven central banks and more than 40 financial firms, reported in May 2026 that its prototype could settle tokenised cross-border payments in seconds.
• India's wholesale e₹ is being used to test settlement of government bonds, call money and tokenised certificates of deposit.

Wholesale projects dodge the hardest political questions. Nobody worries about bank runs when the only users are banks.

The ones going the other way

Not everyone is moving forward. In the United States, Executive Order 14178, signed on 23 January 2025, bars federal agencies from establishing or promoting a retail CBDC. The tracker now classifies the US as inactive on retail. Brazil, once a poster child with its Drex project, told participants in November 2025 that it would shut down the blockchain infrastructure behind it and refocus on a narrower collateral project. We cover that turn in Pix and Drex.

A quick tour by region

• Caribbean: the Bahamas and Jamaica were early launchers and remain the best evidence that a small, well-run CBDC can go live, and that going live does not guarantee people will use it.
• Africa: Nigeria's eNaira is the continent's only fully launched retail CBDC. Elsewhere, mobile money services already meet many of the needs a CBDC would target.
• Asia: China and India run the largest pilots, Hong Kong and Thailand sit inside mBridge.
• Gulf: the UAE remains a core mBridge participant, while the Saudi Central Bank ended its membership after completing a proof of concept.
• Europe: the euro area is legislating, with a deal targeted for the end of 2026.
• Americas: the US has stepped back from retail, and Brazil has narrowed Drex.

How to read the tracker

A few things to keep in mind when you see a big number quoted:

1. Exploring is a low bar. It can mean a single research paper or a working group.
2. A pilot is not a product. Many pilots are closed tests with a handful of banks.
3. Launched does not mean used. All three launched retail CBDCs have struggled with adoption.
4. Wholesale and retail are different animals. Counting them together hides where the real progress is.

Our reading is that CBDCs are becoming infrastructure that most people will never see directly. The digital money most consumers touch will more likely be bank deposits and, increasingly, stablecoins and tokenised deposits, with central bank money settling underneath.

Country counts are from the Atlantic Council CBDC Tracker as reported in 2025 and early 2026; the tracker is updated periodically and counts may shift.

If you've paid an electricity bill in India through your bank app, Paytm, PhonePe or a corner shop in the last few years, you've almost certainly used Bharat Connect without knowing it. It's one of the most important pieces of India's payment system that nobody talks about at dinner.

What is Bharat Connect?

Bharat Connect is India's national bill-payment network. It used to be called the Bharat Bill Payment System, or BBPS, and was renamed at the Global Fintech Fest in 2024. It's run by NPCI Bharat BillPay, a subsidiary of the National Payments Corporation of India, the same organisation behind UPI.

Think of it as a common language between billers and the apps people pay from. An electricity company plugs in once. After that, every bank, wallet and payment app connected to the network can fetch its bills and pay them, with instant confirmation and a standard receipt.

What can you pay through it?

The list has grown well beyond utilities. Today it covers electricity, water, piped gas, LPG bookings, mobile postpaid, broadband, DTH, insurance premiums, loan repayments, credit cards, FASTag, school and college fees, municipal taxes, housing society dues, subscriptions and more. Recurring payments can be set up as standing instructions, so things like insurance or a mutual fund instalment pay themselves.

How it works, in four steps

1. You choose a biller in any connected app and enter your customer number.
2. The app fetches the bill live from the biller: amount, due date, name on the account.
3. You pay by UPI, card, net banking or wallet.
4. The biller gets confirmation instantly, and you get a reference number that both sides recognise if anything goes wrong.

That last point matters more than it sounds. Because every payment carries a standard reference, complaints and refunds follow a common process instead of depending on each company's customer service.

Paying Indian bills from abroad

Since September 2022, Non-Resident Indians have been able to pay Indian bills directly from overseas through approved partners, without first moving money into a relative's bank account. Banks and exchange houses in the Gulf, the UK, Canada and elsewhere have built services on top of this, and the network expanded the scope of cross-border transactions in 2024.

For families, this is a quiet revolution. A daughter in Dubai can pay her parents' electricity bill in Kochi directly. A son in Toronto can clear his sister's college fees in Pune before the deadline. The money doesn't stop in anyone's account along the way.

The best thing about Bharat Connect is that most people don't know they're using it. That's what good infrastructure looks like.

What's still hard

• The cross-border leg. Bharat Connect handles the Indian side beautifully. Getting money from abroad into India still runs through the traditional remittance system, with its fees and exchange-rate margins.
• Visibility for families. Most apps abroad let you pay one bill at a time. Few show you a household's full picture: what's due, what's paid, what's coming.
• Fragmented choice. Each overseas provider supports a different set of countries and categories.

Where this is heading

India has shown the world what a national bill network can do. The next step is joining it to fast, low-cost global rails, so that paying a bill in India from anywhere feels as simple as paying one at home. Other countries are watching closely, and several are building their own versions.

Uttarsanda, a village of about 12,000 people in Gujarat's Kheda district, has a modern bus stand, upgraded classrooms with projectors and CCTV, a beautified pond and new crematoriums. Most of it was paid for by people who no longer live there. Under the state's Vatan Prem scheme, emigrant families have put roughly Rs 9 crore into the village, including Rs 40 lakh from one donor for a crematorium. Drive an hour in almost any direction in central Gujarat, coastal Kerala or Punjab's Doaba belt and you will find the same pattern: a diaspora that sends money home not only to relatives but to the place itself.

Diaspora philanthropy is large, fast-growing and badly measured. It also runs into one of India's most tightly policed laws, the Foreign Contribution (Regulation) Act. This piece sets out what the data does show, how the main giving channels work, and why the same Rs 5 lakh gift can be routine or regulated depending on which passport the giver holds.

How big is diaspora giving?

No official series tracks diaspora donations to India. The best estimates come from the diaspora's largest and richest branch. A 2025 study by Dalberg for Indiaspora estimated that Indian Americans gave $4-5 billion in 2024, nearly triple the 2018 baseline, and that the gap between their giving potential and actual giving had narrowed from $2-3 billion to about $1 billion. The same report projects a $2 trillion intergenerational wealth transfer within the community over twenty years. Note that this is total giving, much of it to US causes such as universities; the study does not split out what goes to India.

For scale, Bain's India Philanthropy Report 2026 puts all private philanthropy inside India at Rs 1.43 lakh crore, or about $16 billion, in FY25, with families contributing about 42% through personal giving and family-business CSR. Bain projects private giving to grow 9-11% a year to FY30, well short of the 25%-plus it says would be needed to stop the social-sector funding gap from widening. It also flags the diaspora, which it puts at about 34 million people, as an under-tapped source of money, time and expertise.

Put those numbers next to India's $135.46 billion in annual remittances and the picture is clear: formal philanthropy is a thin slice of what the diaspora sends. A lot of giving is informal, routed through relatives, temple and mosque committees, alumni groups and caste or village associations, and appears in the statistics as family maintenance. Our analysis of what remittances pay for shows how blurred that line is.

The passport test at the heart of FCRA

The single most misunderstood rule in diaspora giving is that FCRA turns on citizenship, not residence. An Indian citizen living in Dallas or Dubai is generally not a "foreign source" under Section 2(1)(j) of the Act, so a donation from that person to an Indian charity is ordinarily treated as domestic money. A person of Indian origin who holds a US, UK or Canadian passport, including an OCI cardholder, can be a foreign source, and the same gift becomes a foreign contribution under Section 2(1)(h), even if it is paid from an NRO account in India.

That matters because only organisations with FCRA registration or prior government permission may accept foreign contributions, and the rules for those that do were tightened sharply in 2020. Under the current framework, as summarised in this guide to FCRA as amended up to June 2026:

• All foreign contributions must first land in a designated FCRA account at the State Bank of India's designated FCRA main branch.
• Administrative expenses are capped at 20% of foreign contributions, down from 50% before 2020.
• Foreign contributions cannot be passed on to another NGO; the penalty is Rs 1 lakh or 10% of the amount, whichever is higher.
• Registration lasts five years and requires a track record of at least three years and at least Rs 15 lakh of programme spending.
• Gifts from relatives abroad do not need FCRA registration, but if they exceed Rs 10 lakh in a financial year the recipient must file an intimation within three months.

Enforcement has been heavy. The Home Ministry has cancelled or refused renewal for more than 16,000 NGOs since 2020, according to data tabled in Parliament, and an FCRA Amendment Bill, 2026 is pending. Many small village trusts have simply stopped accepting money from foreign-citizen donors because compliance is not worth it.

In diaspora giving, the passport matters more than the postcode. Two siblings can send the same cheque to the same school and trigger two different legal regimes.

Channels compared

Matching schemes: Gujarat's Vatan Prem model

Gujarat launched the Vatan Prem Yojana on 7 August 2021, replacing an earlier 50:50 scheme. Non-resident Gujaratis contribute 60% of the cost of a village project and the state pays the remaining 40%, across a menu that includes smart classrooms, libraries, primary health centres, anganwadis, CCTV, drainage, sewage treatment, bus stands and solar street lights. The governing body proposed works worth Rs 1,000 crore at launch. Reporting from Kheda district shows what it looks like on the ground: Rs 72 lakh from emigrants for a new school serving nearly 400 children in Khadal, and the Rs 9 crore package in Uttarsanda.

The appeal for donors is leverage and accountability. The state's share stretches each rupee, and the asset is built through public works procedures rather than by an unknown local committee. The trade-off is speed and control: government execution can be slow, and donors cannot choose contractors. We think more states with large emigrant populations should copy the model, with published project-level dashboards so donors can see delivery, not just announcements.

Worked example: two siblings, one village school

Consider two siblings in New Jersey whose parents came from a Kheda village. Anjali still holds an Indian passport. Her brother Rohit is a US citizen with an OCI card. Each wants to give Rs 5 lakh to upgrade the village school.

1. Through a private village trust. Anjali's Rs 5 lakh is ordinarily domestic money; the trust issues a receipt and spends it. Rohit's Rs 5 lakh is a foreign contribution. If the trust lacks FCRA registration, it cannot accept it. If it has registration, the money must arrive in the trust's SBI designated main-branch FCRA account, no more than Rs 1 lakh of it may go on administration, and none can be passed to the school's own management committee if that is a separate entity.
2. Through the matching scheme. If the siblings pool Rs 10 lakh as the diaspora's 60% share, the project budget becomes about Rs 16.67 lakh, with the state adding roughly Rs 6.67 lakh. That is two-thirds more school for the same money, delivered through public works.
3. Through their mother in India. If both siblings send money to their mother, who then donates locally, gifts from relatives below Rs 10 lakh a year need no FCRA filing. But the money legally becomes hers first, which has tax and inheritance consequences in both countries.

None of this is legal or tax advice. FCRA is enforced strictly, and both donors and recipients should confirm the current rules with a qualified adviser. Our guides to FEMA rules for NRIs and tax on money sent to India cover the adjacent questions.

Zakat, sadaqah and the Gulf corridor

For millions of Muslim workers in the Gulf, giving is a religious obligation with a fixed rate. Zakat is calculated at 2.5% of savings and assets held for a lunar year above the nisab threshold. Zakat al-Fitr, paid before Eid, is small per person but universal: the UAE Fatwa Council set it for Ramadan 2026 at AED 25 in cash or 2.5 kg of rice per person.

Much of this money is distributed locally in the Gulf, but a significant share flows home through families to madrasas, orphanages and needy relatives in Kerala, Uttar Pradesh, Bihar and Hyderabad. Institutional zakat is also scaling. UNHCR's Islamic philanthropy programme received more than $23 million in zakat in 2025, distributed in full to over 579,000 eligible refugees and displaced people across 17 countries, alongside about $16 million in sadaqah. Rules on routing zakat to Indian institutions face the same FCRA passport test when the giver is a foreign citizen. For sending mechanics on the corridor, see our guide to UAE to India money transfer.

Crowdfunding and the trust problem

Village WhatsApp groups now raise money for borewells, temple roofs, flood relief and a classmate's chemotherapy within days. That speed is the strength and the weakness. Crowdfunded appeals are hard to verify, platforms vary in how they check beneficiaries, and the same FCRA rules apply to the recipient organisation whatever app collected the money. Fake appeals during floods and medical emergencies are a known scam pattern; our guide to money transfer scams covers the warning signs.

Three habits separate effective diaspora givers from frustrated ones: they give through a registered entity with published accounts, they ask for photos and receipts tied to specific milestones, and they pool money through alumni or village associations rather than writing many small cheques. Giving circles, which Dalberg found interest 24% of next-generation donors and a third of next-generation women, formalise exactly that.

What to watch next

• Winter Session of Parliament, late 2026: whether the pending FCRA Amendment Bill, 2026 is taken up, and whether it tightens rules further for small NGOs or for OCI donors.
• 2026-27: implementation of the FCRA Rules now in force, including purpose- and state-based registration and new utilisation thresholds for renewal.
• 7-9 January 2027: Pravasi Bharatiya Divas in Kolkata, a natural venue for new state-level matching schemes modelled on Gujarat's.
• Ramadan 2027 (around February-March): the next Zakat al-Fitr rate announcements from Gulf authorities.

Our view: the diaspora's appetite to give is not the constraint. The constraints are trust in recipients, the compliance cost FCRA imposes on small organisations, and the absence of good public data. States that build transparent matching schemes, and charities that make compliance visible rather than hidden, will capture a growing share of the money. For the investment side of diaspora engagement, see diaspora bonds explained.

Every June and July, a caregiver in Hong Kong or a seafarer docking in Rotterdam sends a little more home than usual. School is starting in the Philippines, and tuition, uniforms and enrolment fees come due at once. The central bank's monthly data shows the pattern clearly, and it tells a bigger story: remittances to the Philippines are still growing, but slowly, while the way families receive and spend that money has changed faster than at any time in the past decade.

The numbers: a record year, a slower pace

According to Bangko Sentral ng Pilipinas data reported by The Philippine Star, cash remittances sent through banks reached a record $35.63 billion in 2025, up 3.3% from 2024. Personal remittances, which also count informal channels and goods in kind, came to $39.62 billion. December alone set a monthly record of $3.52 billion.

Cash remittances were equal to 7.3% of GDP and 6.4% of gross national income. Land-based workers sent $28.49 billion and sea-based workers $7.14 billion. Growth in 2026 has stayed in low single digits: cash remittances for January to July totalled $19.54 billion, up 2.3%, with July coming in at $3.24 billion, according to the Daily Tribune. RCBC's chief economist called July's pace among the slowest since May 2022, while noting that school-opening costs had pushed the monthly level up.

One reason families have not felt the slowdown as much: a weaker peso. With the dollar near 62.80 pesos in mid-2026, each dollar stretched further at home, even if higher local prices took some of that back.

Where the money comes from

The United States remains dominant, accounting for 39.7% of cash remittances in both 2025 and the first seven months of 2026. Singapore (7.1% in January to July 2026), Saudi Arabia (6.3%), Japan (5%) and the United Kingdom (4.7%) follow. Some of the US share reflects how money is routed: transfers from workers elsewhere sometimes pass through US-based banks and are recorded there.

The Gulf matters more than its headline share suggests, because Filipinos there tend to send smaller amounts more often. That makes per-transfer fees especially important. A domestic worker in Kuwait sending 50 dinars every payday feels a flat fee far more than a nurse in California sending a larger sum once a month.

Seasonality is just as strong. December is the peak, driven by Christmas spending and the year-end bonuses many overseas workers receive, and June to August rises again with the school calendar. Families plan around those spikes: enrolment, uniforms and the Noche Buena table are budgeted months ahead, often against an expected transfer rather than money already in hand. That is one reason a slower growth rate matters even when totals are at records. Families who have built their budgets on rising transfers feel any slowdown at once.

What it costs to send

The Philippines is one of the cheaper places in the world to send money to. World Bank Remittance Prices Worldwide data for the third quarter of 2025 shows the average total cost of sending the equivalent of $200 from the UAE to the Philippines at 3.04%, and from the United States at 4.52% for transparent offers. Both sit well below the global average, which the World Bank put at 6.36% in the same period.

Averages hide a wide spread, though. A cash pickup at an agent with a poor exchange rate can cost several times more than an app transfer straight to a wallet. Our breakdown of the real cost of sending money home explains how FX margins often outweigh the visible fee.

GCash, Maya and the end of the queue

A decade ago, receiving money often meant a trip to a pawnshop or remittance counter. Today a growing share lands directly in e-wallets. GCash says about 81% of Filipinos have an account, and it has added real-time inbound transfers from the US and Canada through licensed money transmitters, as reported by Gulf News in 2025. Maya and bank apps compete for the same flows.

The wider payments system has moved in the same direction. The BSP said digital payments made up 64.7% of retail payment volume in 2025, up from 57.4% a year earlier and inside its 60% to 70% target, according to the Daily Tribune. QR Ph processed 2.47 billion transactions worth P1.16 trillion and overtook card payments for the first time. A central bank circular now requires transfer fees between banks and e-wallets to stay close to same-institution fees, and several large banks waived InstaPay and PESONet charges.

The money still crosses the ocean the old way; what changed is the last mile, which now ends on a phone.

From remittance to bills

For many families, the remittance is effectively a bill payment that took a detour. Money arrives in a wallet, and within days most of it goes to electricity, water, tuition, internet and loan instalments. Wallets have built bill-pay menus for exactly this reason. We argued in bills are the real remittance that paying a household's obligations directly from abroad could cut out a conversion and a few days of float.

A sender in Riyadh who pays her family's internet and water bills directly, rather than sending cash and hoping they are paid on time, gains something hard to measure: certainty.

What to watch next

• Instant cross-border links. The Philippines is one of the founding members of Project Nexus, the BIS-incubated scheme to connect national instant payment systems, with InstaPay and PESONet among the domestic systems to be linked. Our piece on Project Nexus covers the timeline.
• Seafarer and Gulf demand. Sea-based remittances grew 2.9% in 2025, a little slower than land-based ones. Shipping and Middle East labour conditions will shape the next two years.
• Exchange rate. A weak peso flatters peso-denominated totals; a rebound would squeeze families even if dollar flows hold steady.

The headline growth numbers may look sleepy. The plumbing beneath them is anything but.

In February 2026, ration-card holders in parts of Gujarat began receiving part of their food subsidy as digital rupees that could only be spent at a fair price shop or a designated merchant. Try to buy a phone recharge with that money and the wallet simply declines. That small refusal is the whole idea behind a programmable e-rupee, and it is the corner of India's central bank digital currency that has moved fastest in 2025-26, even as the headline circulation number went the other way.

What "programmable" means in the RBI's version

The Reserve Bank of India set out its design in the Concept Note on CBDC of October 2022. The retail digital rupee (e₹-R) was to be a token, closer to a banknote than to a bank balance, distributed through banks in a two-tier model and paying no interest. Programmability was not the headline then. It arrived in February 2024, when Governor Shaktikanta Das told the policy press conference that the pilot would add use cases built on programmability and offline functionality, as Business Standard reported.

The examples he gave were concrete. A government agency could ensure that money reaches only a defined benefit. A company could fence spending to something like employee business travel. Rules could also set a validity period, or restrict the geographic area in which the token works. In practice a programmable token carries a small rulebook: who may receive it, which merchant categories may accept it, where, and until when. When the conditions are met, the merchant receives ordinary e-rupee that it can use like any other money.

That last point matters for the economics. Deputy Governor T Rabi Shankar has argued that a restriction is temporary, saying programmability "just puts it on hold" as far as fungibility goes, a framing quoted in a Finshots explainer. The rupee in a citizen's wallet is conditional; the rupee in a shopkeeper's wallet is not.

It is easy to confuse this with other Indian tools. e-RUPI, for instance, is a voucher that runs over the UPI rails and is issued by a bank. A programmable e₹ is central bank money with a rule attached. We cover the voucher route separately and the broader concept in purpose-bound money explained.

From banknote to rulebook: how the pilot got here

The wholesale pilot (e₹-W) started on 1 November 2022 with nine banks settling secondary-market government securities trades. The retail pilot followed on 1 December 2022 in closed user groups. By December 2023 the retail pilot had hit a target of one million transactions a day, a number the RBI and the pilot banks had set out to reach.

Programmability and offline payments were added during 2024. At the RBI@90 conference in August 2024, the Governor described a pilot of more than 5 million users and 16 banks, and named two farm-sector use cases that we return to below, according to a Digital Pound Foundation summary of the event. By March 2025, the RBI's annual report counted 17 banks and about 6 million users, and non-bank firms such as MobiKwik and CRED had been allowed to run CBDC wallets, as MediaNama's digest of the report noted. Reported user numbers reached roughly 7 million across 19 banks by October 2025.

Then came the awkward number. The RBI Annual Report 2025-26, released at the end of May 2026, put CBDC-Retail in circulation at ₹771.66 crore on 31 March 2026, down about 24% from ₹1,016.46 crore a year before, as Outlook Money reported. In the same report, the RBI described a year in which CBDC pilots moved into direct benefit transfer schemes of the Centre and the states. Fewer rupees sitting in wallets, more rupees doing specific jobs. We think those two facts are connected.

What is live: the use cases the RBI has actually run

The list below is limited to use cases that the RBI or the governments involved have described publicly as run or piloted. Most are small. None is a national rollout.

The Odisha case is the clearest example of a government using the rail at scale. Subhadra pays eligible women ₹10,000 a year for five years, and the state paid 88,000 of them in digital rupees under the pilot. The PDS pilots, reported by OrissaPOST from the 2025-26 annual report, are a stronger test of the "programmable" part, because the rule is narrow: food, at a specific class of shop.

Agriculture: the strongest case for programmable money

Subsidy fencing gets the attention. The more interesting idea is credit. Many tenant farmers cultivate land they do not own, so they cannot pledge a land title, and banks have long been reluctant to lend to them under the Kisan Credit Card. The Governor's argument at RBI@90 was simple: if loan money can only be spent on seeds, fertiliser and other inputs, a bank has more assurance that the loan funds a crop and not something else, and that can make it willing to lend without the title.

That is a credit-risk argument, and it is a good one. It does not depend on surveillance of the farmer's whole financial life, only on the narrow question of whether loan money reaches an input dealer. It also creates a natural end point: once the dealer is paid, the money is ordinary e-rupee in the dealer's wallet.

The practical hurdles are just as specific. Input dealers in rural blocks need to accept e-rupee, which means wallets, QR codes and a reason to bother. Connectivity is patchy in exactly the places where tenant farming is common, which is why the RBI tied programmability to offline functionality in the same announcement. We look at how that might work in our explainer on offline CBDC payments. And the rule has to be coded somewhere: a list of eligible merchant categories, kept current, with a route for a farmer to complain when a legitimate dealer is wrongly blocked.

Programmable money is most defensible when the rule protects a lender or a scheme, and least defensible when it simply second-guesses the person holding it.

Worked example: one welfare scheme, two kinds of plumbing

Take Subhadra's published numbers. A beneficiary is entitled to ₹10,000 a year for five years, or ₹50,000 in total. If all 88,000 women paid in digital rupees received a full year's entitlement that way, the state would have pushed about ₹88 crore of e-rupee into wallets in a year (88,000 × ₹10,000). Set that against the ₹771.66 crore of retail e-rupee in circulation at end-March 2026 and one state scheme alone would be equal to roughly 11% of the whole stock. That is how small the pilot still is.

Now compare two ways of delivering the same ₹10,000.

• Bank-account DBT. The money lands as a deposit. It is fully fungible from the first second. The state can see that it was credited; it cannot see or control what it was used for. Leakage, if any, happens after credit and is invisible.
• Programmable e-rupee. The money lands as tokens with a rule. Suppose, purely as an illustration, a ₹5,000 instalment carries a rule that it can be spent only at registered merchants in the beneficiary's district. If she spends ₹3,200 at approved shops in the first month, the merchants receive ₹3,200 of unrestricted e-rupee and the remaining ₹1,800 stays conditional in her wallet.

Two things fall out of the arithmetic. First, programmable welfare money is designed to move quickly to merchants, who are likely to convert it back to bank deposits. That pushes circulation down even as use rises. Second, every rule has a cost to the beneficiary if the list of approved shops is short. A ₹1,800 balance that cannot be spent at the nearest shop is worth less to her than ₹1,800 in a bank account. The design of the merchant list is the scheme.

The Subhadra rule set used in this example is illustrative; Odisha has not published the exact conditions coded into its e-rupee payments.

Why circulation fell while use cases grew

A 24% fall in a year looks bad, and it should prompt questions. But circulation is a stock, not a flow. It measures what is sitting in wallets on one day. Several things can shrink it while the pilot is, in a narrow sense, working.

Some early activity was target-driven. The pilot had a stated goal of one million transactions a day by December 2023, and usage built to meet a target is not the same as balances people choose to keep. The e-rupee also sits in a separate wallet that competes with UPI, which was handling about 802 million payments a day by September 2026. For most people, a separate wallet with no interest and no merchant advantage is a harder sell than a UPI app linked to the account they already use. We compare the two in e-rupee vs UPI.

And programmable money, as the worked example shows, is built to be spent. A scheme that pays ₹88 crore a year through e-rupee might leave only a sliver of that in wallets at any month end. If the RBI wants circulation to rise, it needs reasons for people to hold e-rupee. If it wants use to rise, it needs schemes and merchants. The 2025-26 numbers suggest it has chosen the second path, at least for now.

The risks: coercion, fungibility and who writes the rules

The objections to programmable money are not abstract. Restricted money can force spending a person would rather defer. Expiry dates, an idea sometimes floated for stimulus payments, are a tax on people who save. Rules can be changed. And the same plumbing that limits a subsidy to food can, in principle, limit anything. The Bank of England has said it will not pursue government or central bank-initiated programmable functions for a digital pound, which is a useful marker of how differently central banks read the same technology.

The RBI's concept note leans on "managed anonymity", with small payments more private and large ones traceable. That is a privacy design, not a programmability rule, and the two interact: a rule that restricts merchant category requires the system to know the merchant category of every restricted payment. We go deeper in the CBDC privacy debate.

Our view is that the safeguards matter more than the technology. Rules should be tied to the law or scheme guideline that funds the transfer, published in plain language, and narrow. Restrictions should drop away once money reaches a merchant. Beneficiaries need a grievance route when a payment is wrongly declined. And the default for ordinary e-rupee, the kind a person buys with their own deposit, should be no rules at all.

What to watch next

• 2026-27 pilot expansion. The RBI says it will extend CBDC pilots to more DBT schemes, domestic retail and business applications during the financial year ending March 2027. Watch which central schemes sign up; that is where volumes would move. Our guide to direct benefit transfer explains the scale involved.
• The CAT sandbox. The RBI plans a framework for testing products through a CBDC and Asset Tokenisation sandbox, which is where corporate allowance and supply-chain use cases are likely to be tried.
• Cross-border pilots. The bank signed a digital asset collaboration MoU with the Monetary Authority of Singapore in 2025-26, held talks with the Central Bank of the UAE on a cross-border CBDC pilot, and joined BIS Innovation Hub work including Project Rialto and the second phase of Project Mandala.
• End-May 2027. The RBI Annual Report 2026-27 will give the next circulation figure and, ideally, transaction counts for the programmable pilots. If the RBI wants the public to judge programmability on its merits, publishing use-case-level numbers would help.

For a plain-English primer on the instrument itself, start with the digital rupee explained.

Ask five payments engineers which blockchain is best for stablecoins and you will get five answers, each correct for a different job. A freelancer in Manila being paid by a client in Sydney cares about fees and whether her local exchange supports the network. A treasury team moving millions between subsidiaries cares about finality and audit trails. A card network cares about uptime. This guide compares the main options on the numbers that matter in 2026.

Where the stablecoins actually are

Liquidity is the first filter, because a cheap network is useless if nobody on the other side can receive or cash out on it. DefiLlama data cited on 25 September 2026 put Ethereum at about $145.9 billion of stablecoins, Tron at about $94.1 billion, Solana at a record $16.6 billion, BNB Chain at about $13.3 billion and Base at about $5.1 billion, according to Solana Compass. Together that is about 90% of all stablecoin supply, which you can see in context in our 2026 stablecoin market report.

The comparison table

Fee and speed ranges for Ethereum, Solana, Base and Arbitrum are mid-2026 estimates from Spark's comparison; they move with congestion and token prices. For layer-2 networks like Base and Arbitrum, the fast confirmation is from the network's sequencer, while final settlement back to Ethereum takes longer, though exchanges usually credit deposits within minutes.

Ethereum and its layer-2s: the settlement layer

Ethereum's share of stablecoins has fallen from above 90% in 2020 to around half, yet the dollar amount has grown many times over. It remains where large issuers mint first, where institutional funds hold balances and where DeFi liquidity is deepest. Its weakness is cost: a few dollars per transfer is fine for a $50,000 invoice and absurd for a $40 remittance.

Layer-2 networks fix most of that. Base, run by Coinbase, has become the default for consumer checkout; Shopify's USDC payments run on it. Arbitrum and Optimism carry a lot of exchange and DeFi traffic. All of them inherit Ethereum's security over time.

Tron: the people's dollar rail

Tron is the network most people in emerging markets actually use, almost entirely for USDT. It became the default because it was cheap and simple when Ethereum was neither. In August 2025 Tron's validators halved the price of the "energy" that transactions consume, cutting a typical USDT transfer from about $2.47 to roughly $0.72, as Unchained reported. That is still far more than Solana, but Tron's real advantage is network effects: the exchange in Lagos, the OTC desk in Dubai and the shopkeeper's cousin all already accept it.

Solana and Stellar: built for many small payments

Solana charges fractions of a cent and confirms in about two seconds, which is why fintech apps and card programs like it. Visa settles USDC with US banks over Solana, and Western Union chose Solana for its planned USDPT stablecoin. Its stablecoin supply hit a record in September 2026.

Stellar is smaller by supply but was designed for payments from the start. Franklin Templeton's tokenised money fund records ownership on it and settles in five to seven seconds, and MoneyGram, which used USDC on Stellar for years, launched its own MGUSD stablecoin there in June 2026 for a customer base of more than 60 million. We cover it in depth in our Stellar explainer.

The best chain for a payment is usually the one the person at the other end can cash out on.

The new payment chains: Tempo, Arc and Plasma

The biggest shift of the past year is the arrival of networks built specifically for stablecoins, often by companies that already move money at scale.

• Tempo. Incubated by Stripe and Paradigm, Tempo opened a public testnet in December 2025 with design partners including Mastercard, UBS and Kalshi, after raising a $500 million Series A. Its mainnet went live on 18 March 2026 alongside a Machine Payments Protocol aimed at AI agents paying each other. Klarna's KlarnaUSD stablecoin was slated to launch on it.
• Arc. Circle's layer-1 opened its public mainnet on 16 September 2026. Fees are paid in USDC rather than a volatile token, finality is under a second, and the 11 founding validators include BlackRock, DTCC, ICE, Mastercard and Visa, according to The Block.
• Plasma. A Bitfinex-backed network, Plasma launched its mainnet beta in September 2025 with zero-fee USDT transfers and more than $2 billion of stablecoin liquidity on day one.

The incumbents are hedging. Visa's settlement pilot expanded to nine chains in April 2026: Avalanche, Ethereum, Solana and Stellar, plus Arc, Base, Canton, Polygon and Tempo. The new Open USD stablecoin launched on Ethereum, Solana, Base and Tempo at once.

So which is best?

Our honest answer, by use case:

• Small cross-border transfers into emerging markets: Tron for USDT reach, Solana or Stellar where local partners support them.
• Online checkout and consumer apps: Base and Solana, for low fees and wallet support.
• Large B2B and treasury movements: Ethereum, or a payment chain like Arc or Tempo if your bank or processor is already there.

The trend is toward users never having to choose. Circle's cross-chain transfer protocol settles in seconds between networks, and most payment apps now hide the chain entirely. The more the network fades into the background, the more the contest becomes about off-ramps, compliance and price, which is exactly where it should be. For the business side of this, see our 2026 payments trends.

Fees and supply figures change constantly. Check a block explorer or your provider before sending large amounts.

For most of the 2020s, Ripple was known mainly for a court case. That chapter closed in 2025. What's emerged since is a very different company: less about one token, more about owning the boring, profitable plumbing that moves money between businesses.

The lawsuit is over

The US securities regulator's case against Ripple, filed in 2020, ended in 2025 when both sides dropped their appeals. Whatever you thought of the case, its end removed the single biggest reason banks and payment companies hesitated to work with Ripple in the US.

RLUSD: the stablecoin that ties it together

Ripple launched its own dollar stablecoin, RLUSD, at the end of 2024. It's aimed squarely at institutions: banks, payment providers, brokers. By the end of September 2026 its supply had reached roughly $2.49 billion, up more than 86% in nine months, with about 56% on Ethereum and 44% on the XRP Ledger. Its reserves are held with a New York-chartered trust company, and it won approval in Japan this year.

That growth is modest next to USDT or USDC. But RLUSD isn't really competing for retail wallets. It's the settlement asset inside Ripple's own network.

The acquisitions tell the story

In 2025 Ripple went shopping. It bought Hidden Road, a prime brokerage; Rail, a stablecoin payments platform; and GTreasury, a corporate treasury software company, for about $1 billion. Put together, the strategy is clear:

• Payments to move money between companies and countries.
• Treasury software where big companies already manage their cash.
• Brokerage to provide liquidity and credit to institutions.
• A stablecoin running underneath all three.

Ripple wants to be where a CFO already works, then offer stablecoin settlement as a better option for cross-border flows.

Ripple's bet is simple: the biggest stablecoin opportunity isn't in wallets. It's inside corporate treasury.

What about XRP?

XRP still exists and still has a passionate community, and the XRP Ledger remains a fast, cheap network with a built-in exchange. But in 2026 Ripple talks far more about RLUSD, payments and custody than about the token. For businesses that's good news: they can use the rails without needing a view on any token's price.

What it means for cross-border payments

Ripple's moves make life easier for anyone building on stablecoin rails. More licensed payout partners, more liquidity in more currencies, more institutions comfortable holding digital dollars. Its stated focus markets for 2026 include the UAE, Nigeria and Turkey, all large remittance and trade corridors.

For households, the effect is indirect but real. Every time a big, regulated player builds payout routes into a new country, the cost of getting money there on stablecoin rails falls a little, and that's money that stays with families instead of disappearing into the pipes.

We don't think one network will win everything. Ripple, Stellar and the big general-purpose chains are each strong in different places. The products that win will be the ones that use all of them without asking the user to care.

The United States now has two instant payment networks, a $10 million limit on both, and more than a thousand banks connected to each. In the second quarter of 2026 the Federal Reserve's FedNow service settled about 5 million payments. India's UPI does that in under ten minutes. The comparison is unfair in some ways, but it captures the puzzle: the pipes are built, and Americans still mostly wait for ACH.

Two networks, two owners

The RTP network came first. The Clearing House, owned by the largest US banks, launched it in 2017. FedNow arrived in July 2023, run by the Federal Reserve and open to every bank and credit union that holds a Fed account. Both settle in seconds, around the clock, every day of the year. Both now carry a per-transaction ceiling of $10 million: RTP raised its cap from $1 million in February 2025, and FedNow followed in November 2025.

Neither network sells to consumers. They sit behind banks, payroll firms and payment companies, which decide whether to offer instant transfers at all. That middleman layer is the single biggest reason adoption looks the way it does.

The growth is real

Look at the trend lines and the story is healthy. The Fed's quarterly statistics show FedNow settling 97,424 payments in the first quarter of 2024 and 4,997,811 in the second quarter of 2026. Value went from $31 million to $274.7 billion over the same period. Across 2025, FedNow moved $853.4 billion, connected about 1,600 institutions and averaged almost 30,000 payments a day, PaymentsJournal reported.

RTP is bigger. It processed more than $1.3 trillion in 2025 with roughly 1,135 participating institutions. On 1 May 2026 it set a single-day record of 2.27 million transactions worth $8.62 billion, and monthly value had climbed from $28 billion in January 2025 to $195 billion in April 2026, according to Digital Transactions.

Look at the average payment

The most revealing figure is the size of a typical transfer. In 2025 the average FedNow payment was about $101,435. In the first quarter of 2026 it was $99,414. In the second quarter it fell to $54,957 as volume nearly doubled, which suggests smaller payments are starting to arrive, but these are still not grocery-sized numbers. RTP's average in 2024 was $719, much closer to everyday life, but its value surge since then has been driven by corporate treasury moves and large time-sensitive transfers.

In other words, US instant rails are being used first as a cheaper, always-open alternative to wire transfers. A property company settling a closing on a Saturday, a corporate treasurer sweeping cash between banks at night, a payroll provider paying gig workers on demand. Those are valuable use cases. They are not the mass consumer habit that Pix built in Brazil or UPI built in India.

America built instant rails for businesses first and is waiting for consumers to find them.

Why consumers have not shown up

ACH is good enough for most bills. Nacha reports that the ACH network carried 35.2 billion payments worth $93 trillion in 2025, and Same Day ACH alone handled 1.4 billion payments worth $3.9 trillion, up 16.7% in volume, according to Nacha. For a utility bill due in ten days, same-day settlement is plenty. Instant only matters when timing is tight.

Cards and wallets already feel instant. A customer tapping a card at a store sees an immediate approval and earns rewards. The merchant waits days for settlement and pays fees, but the customer does not feel that. Peer-to-peer apps already show money arriving instantly, even when the underlying settlement is slower.

There is no front door. Pix and UPI succeeded partly because the central bank or national operator set rules that made every bank offer the same QR codes and addresses under one recognisable brand. FedNow and RTP have no consumer brand, no shared QR standard and no rule forcing banks to send, not just receive. Many smaller institutions joined on the receiving side first because it is cheaper and lower risk.

Fraud fear. Instant and irrevocable is a scammer's dream. Banks remember what authorised push payment fraud did in the UK, and many keep low default limits on consumer instant transfers. That caution is reasonable, but it also keeps volumes small. Our guide to money transfer scams explains why.

Fragmentation. Two networks that do not fully interoperate mean a bank on FedNow cannot always reach a bank only on RTP. Some institutions join both, many join one. That creates gaps a consumer cannot see but a product team has to work around.

What could change the curve

Three developments are worth watching. First, request for payment: if billers can send a request that the customer approves in a banking app and settles instantly, instant rails could start to eat card-on-file and ACH debit for household bills. Second, payroll and earned wage access, where speed has an obvious value for the worker. Third, tax refunds: RTP said refund payments through its network rose 78% in the first four months of 2026, a sign that government and tax-preparer disbursements can pull consumers onto the rails without any marketing campaign.

A fourth, less certain driver is competition from stablecoins. US dollar tokens already offer 24/7 settlement for businesses that hold them, and the GENIUS Act gave them a clearer legal footing. If banks see deposits leaking toward regulated stablecoins, offering instant bank transfers becomes a defensive necessity rather than a nice extra.

Our read

US instant payments are not failing. They are following a different path from Brazil and India: business first, consumer later, and no mandate to speed things up. The numbers are doubling, the limits now cover almost any commercial payment, and both networks are past a thousand participants. What is missing is a reason for an ordinary customer to care. Until a bank app makes instant the default and a recognisable button sits on every bill, ACH will keep winning by being familiar.

FedNow figures cover settled customer credit transfers only, as published by the Federal Reserve. RTP figures are as reported by The Clearing House via trade press.

The treasurer of a mid-sized freight forwarder in Rotterdam gets a simple question from her CFO: a Brazilian customer wants to pay $2 million in USDC, so can they just hold it for a few weeks until the next supplier run? The operational answer takes ten minutes. The accounting answer, until very recently, took a technical memo, a call with the auditor and a shrug. Is a dollar stablecoin cash, an investment, or something else? On 18 August 2026 the US Financial Accounting Standards Board published a proposal that finally starts to answer, and it turns on details most treasury teams have never had to think about.

Why treasurers suddenly care

For most companies, stablecoins arrive from the outside: a customer wants to pay in them, a marketplace settles in them, or a subsidiary in a dollar-short market has been receiving them informally. The scale of the instrument makes it hard to ignore. Total stablecoin supply was about $306 billion as of October 2026, and genuine stablecoin payments reached roughly $390 billion in 2025, according to McKinsey and Artemis estimates.

Corporate use is still early. In an EY-Parthenon survey of 350 firms conducted in June 2025, 9% of corporates and 23% of financial institutions said they used stablecoins. Only 8% of corporates accepted them as payment. Among current users, 41% reported savings of at least 10%, mostly on dollar cross-border B2B payments. The striking number for treasury is a different one: 79% of corporates converted stablecoins to fiat immediately after a transaction. Companies are using stablecoins as a rail, not yet as a place to keep money.

The reason is not mysterious. Regulatory uncertainty was the top concern at 73%, with accounting and tax treatment second at 38%. If you cannot tell your board whether an asset counts as cash, you do not hold it overnight.

The accounting gap: ASU 2023-08 does not cover most stablecoins

US GAAP got its first crypto-specific standard in December 2023, ASU 2023-08, which requires qualifying crypto assets to be measured at fair value through net income. That fixed the old problem of bitcoin being carried as an impaired intangible. It did not fix stablecoins.

The standard's scope has six criteria, one of which is that the token must not give the holder enforceable rights to underlying goods, services or other assets. As Indiana University's Vivian Fang explained in Bloomberg Tax in October 2025, reserve-backed redemption rights are exactly such a claim, so many stablecoins fall outside ASU 2023-08. Falling outside does not make them cash. Companies have instead had to apply other guidance, and practice has varied.

The cash-equivalent test in ASC 230 asks for short-term, highly liquid investments readily convertible to known amounts of cash and so near maturity that rate risk is insignificant. A stablecoin has no maturity date at all, which is why auditors have hesitated.

FASB's August 2026 proposal: three attributes

FASB added a digital assets project to its agenda in 2025, after the GENIUS Act became law in July of that year, and on 18 August 2026 it issued a proposed update to Topic 230. Deloitte's Heads Up summary sets out what it does and does not do. The definition of a cash equivalent would not change. Instead, FASB adds illustrative examples built around three attributes, considered together:

1. An on-demand contractual right to redeem for cash at the holder's election, without significant fees.
2. A direct redemption right against the issuer, rather than having to rely on an exchange or market maker to get known amounts of cash.
3. Segregated reserves of short-term, highly liquid assets held at least one-for-one against tokens in circulation.

Classification would be a policy election, not a requirement. A token backed by crypto assets or gold fails, per one of the proposal's cases. And every entity presenting cash equivalents would have to disclose the significant components each year, whether those are Treasury bills, money market funds or stablecoins. Comments are due by 19 November 2026; the effective date is not yet set, and early adoption would be permitted.

The proposal makes the relationship with the issuer, not the token, the thing auditors will test.

The second attribute is the one that will catch companies out. A business that receives USDC from a customer and holds it in a self-custodied wallet, with no account at the issuer, may not have a direct redemption right at all. To get dollars it must sell through an exchange or OTC desk. Under the proposal's logic, the same token could be a cash equivalent in one company's books and not in another's.

And under IFRS?

International standards have no stablecoin-specific guidance. The IFRS Interpretations Committee concluded in 2019 that cryptocurrency holdings are not cash and generally fall under intangible assets or inventory. A May 2026 IASB staff paper on intangible assets notes strong stakeholder support for better treatment of crypto, with cryptoassets added to the list of potential projects, but no decisions and no timetable for a stablecoin answer. Groups reporting under IFRS should expect to wait longer than US filers. A qualified accountant should confirm the treatment for any specific holding.

The GENIUS Act draws a line around issuers

The GENIUS Act, signed in July 2025, creates the federal framework for payment stablecoins: licensed issuers, one-for-one reserves in cash and short-term Treasuries, monthly reserve reports, no interest paid to holders and priority claims for holders in an issuer's insolvency. For treasury, its most important feature is the one Deloitte flags: once effective, the Act generally prohibits treating a payment stablecoin from a non-permitted issuer as cash or a cash equivalent. It takes effect at the earlier of 18 months after enactment or 120 days after final implementing rules, which puts the outside date in January 2027.

That gives treasurers a clean policy rule: approved issuers are permitted issuers, full stop. It also means the largest stablecoin by supply, USDT at about $184 billion, sits in a different category for US-reporting companies unless its issuer's US arrangements change. Our pieces on GENIUS and MiCA and GENIUS Act implementation follow the rule-writing.

Banking access: better, still uneven

Holding stablecoins is useless if your bank will not touch the fiat leg. On 7 March 2025 the OCC issued Interpretive Letter 1183, confirming that national banks may provide crypto custody, hold deposits that back stablecoins and engage in certain stablecoin payment activities without first seeking supervisory non-objection, as Sullivan & Cromwell summarised. At the time the FDIC and Federal Reserve had not matched it, so what a bank will do still depends partly on its regulator.

In practice, a treasury that wants to hold stablecoins needs three things lined up: a bank that will receive wires from the issuer or an exchange without freezing them for review, a custodian or wallet setup that internal audit will sign off, and ideally a direct issuer account so that redemption at par does not depend on a third party. Many companies still route through a specialist payments partner instead. EY's survey found 56% of corporates prefer embedded APIs in existing treasury platforms and 70% would be more inclined to adopt with ERP integration.

Risk policy: what March 2023 taught

The case every treasury committee should study is the weekend Silicon Valley Bank failed. Circle disclosed that $3.3 billion of USDC reserves, nearly 8% of about $40 billion, sat at SVB; USDC fell below 87 cents on secondary markets, CNBC reported, before recovering once US authorities backstopped SVB depositors. A token can be fully reserved and still trade well below par when its reserves are temporarily inaccessible. Our history of depegs covers the other episodes.

A sensible stablecoin treasury policy borrows from money market fund policy and adds crypto-specific controls:

• Approved issuers only, tied to legal status (GENIUS-permitted or MiCA-authorised) and published reserve reports.
• Concentration limits per issuer and a maximum share of operating liquidity in stablecoins.
• Holding-period caps, for example a rule that receipts are converted within a set number of business days unless a payment out is scheduled.
• Custody standards: multi-approval wallets, whitelisted addresses and segregation of duties, covered in our guide to self-custody and custodial wallets.
• Chain and contract risk: which blockchains are approved, and how sanctioned-address screening is done.
• Redemption drills: periodically redeem a small amount to prove the route works.

Worked example: the $2 million USDC receipt

Return to Rotterdam. The figures for yield and spread below are illustrative assumptions, not market quotes. The company receives 2,000,000 USDC and needs the dollars in 21 days.

• Hold as USDC. No yield, since GENIUS-compliant issuers cannot pay holders interest. If the company had instead parked $2 million in a money market fund at an assumed 4% annual yield, it would have earned about $4,600 over 21 days ($2,000,000 x 4% x 21/365). That is the opportunity cost.
• Tail risk. In a repeat of March 2023, a 13% secondary-market discount would show a temporary mark of about $260,000 if the company had to sell that weekend. With a direct issuer redemption account, it could wait for par; without one, it is exposed to the market price.
• Convert immediately. If an OTC desk or payments partner converts at an assumed all-in 0.10%, the cost is $2,000. The company then holds dollars in a bank account it already understands.

Unless the company expects to pay suppliers in USDC within days, converting is usually cheaper and simpler, which is exactly what most corporates in the EY sample did. The calculation changes for firms paying stablecoin-accepting suppliers, where holding avoids two conversions. Firms should take advice from their auditor and a qualified adviser before setting policy.

What to watch next

• 19 November 2026: comment deadline on FASB's Topic 230 proposal. Watch whether the direct-redemption attribute is softened for holders using regulated intermediaries.
• By 18 January 2027: the latest date the GENIUS Act takes effect, if final rules are not finished sooner, bringing the bar on non-permitted issuers into force for cash classification.
• Bank regulators: whether the FDIC and Federal Reserve align with the OCC on stablecoin activities, which would widen corporate banking access.
• Tokenised deposits: bank-issued alternatives, discussed in our comparison of stablecoins, CBDCs and tokenised deposits, may give treasurers on-chain dollars that are already deposits on the balance sheet.

The shift in 2026 is that the stablecoin question has moved from "is this allowed?" to "is this cash?". For treasurers, the answer will depend less on the token than on the paperwork behind it: who issued it, under which law, and whether the company can walk up to the issuer and get its dollars back.

A coconut seller on Boa Viagem beach in Recife does not carry a card machine. She has a laminated QR code taped to her cooler, and almost every customer pays by Pix. On Friday 6 June 2025, Brazil's instant payment system processed nearly 280 million transactions in a single day, moving about R$135.6 billion, according to the central bank figures reported by Agencia Brasil. India's UPI now does roughly three times that every day. The United States' FedNow, meanwhile, settled about 5 million payments in the whole of the second quarter of 2026.

All three are instant payment systems. They move money between bank accounts in seconds, around the clock. Yet their usage differs by a factor of thousands. This piece compares the main systems, explains why design and policy matter more than technology, and sets out what comes next as countries start linking them across borders.

The BIS says that in more than 70 countries, domestic instant payments now settle in seconds. ACI Worldwide's most recent Prime Time for Real-Time report counted 266.2 billion real-time transactions globally in 2023, up 42.2 percent on the year, equal to 19.1 percent of all electronic transactions. It projects about 575 billion by 2028, when real-time would exceed a quarter of electronic payments.

That headline growth is heavily concentrated. A handful of national systems, led by India and Brazil, account for most of the world's instant payment volume. In many rich economies, by contrast, instant rails exist but remain a minority choice.

Five systems side by side

The numbers are not perfectly comparable. Some are daily averages, some are record days, some are quarterly totals. But the orders of magnitude are clear, and they tell a story about how each system was built.

India and Brazil: the central bank as product manager

UPI and Pix share three features. They are free or nearly free to consumers. They work through a single national QR standard that any bank or app must accept. And the public authority pushed hard on adoption, with the central bank in Brazil and the NPCI and RBI in India acting less like referees and more like product owners.

Pix had over 175 million users by the end of May 2025, including 160 million individuals, in a country of about 210 million. Brazil's prior daily record of 252.1 million transactions came on 20 December 2024, the day many employers paid the second instalment of the 13th-month salary. That detail matters. Pix is not just a way to split a restaurant bill; it has become payroll infrastructure. Our Pix and Drex explainer covers how it was designed.

UPI's scale is harder to grasp. At about 802 million transactions a day in September 2026, it processes roughly 557,000 payments every minute. Read more in UPI explained.

Instant rails succeed when someone is made responsible for adoption, not just for settlement.

Europe: the regulator steps in

Europe launched SEPA Instant Credit Transfer in November 2017. According to the European Payments Council, a transfer must complete within nine seconds. For years, though, banks treated it as a premium service, and uptake was slow.

The EU's Instant Payments Regulation changed the incentives. Eurozone credit institutions have had to meet its requirements since 9 October 2025. The effect is showing up. The ECB's payment statistics for the second half of 2025 show that instant transfers made up 25 percent of the number and 8 percent of the value of credit transfers processed by euro area retail systems. Cards still dominate overall, at 57 percent of all non-cash transactions. Our guide to the SEPA instant regulation explains the rules in detail.

The United States: built, not yet adopted

FedNow shows what happens when a central bank builds the rail but leaves adoption to the market. The Federal Reserve's quarterly statistics show 4,997,811 settled payments in Q2 2026, worth about $274.7 billion. That is up 83 percent on a year earlier, which sounds impressive until you set it against India.

The average FedNow payment in Q2 2026 was about $54,957. This is not a coffee-and-rent system. It is used mostly by businesses and institutions moving large sums, often to fund accounts or settle obligations quickly. The private RTP network run by The Clearing House operates alongside it, and consumers mostly use card-linked apps that sit on older rails. For the full US picture, see FedNow and RTP explained.

A worked example: putting volumes on one clock

It helps to put the figures on a single timescale.

1. UPI processes about 802 million payments a day. Divided by 1,440 minutes, that is about 557,000 a minute.
2. FedNow settled about 5.0 million payments across the 91 days of Q2 2026.
3. Divide 5.0 million by 557,000 and you get about 9. In other words, UPI processes a full quarter of FedNow volume in roughly nine minutes.
4. Now look at value. Pix's record day averaged about R$484 per transaction (R$135.6 billion over roughly 280 million payments). FedNow averaged nearly $55,000. Euro area instant transfers were a quarter of credit transfers by number but only 8 percent by value.

The lesson is that instant systems are serving very different jobs. In India and Brazil, they replaced cash for everyday payments. In Europe, they are replacing slower account transfers, mostly small ones. In the US, they are a faster pipe for business treasury flows. Judging FedNow by UPI's standard misses that it was never asked to do UPI's job.

Five design choices that decide adoption

Looking across these systems, the technology is broadly similar. Messages move between banks, a central operator settles, and the recipient's account is credited within seconds. What differs is a short list of choices made by policymakers, and they explain most of the gap in usage.

• Price to the consumer. UPI and Pix are free for individuals. Thailand made PromptPay transfers up to 5,000 baht free. Where banks charged for instant transfers, as many in Europe did before the regulation, people kept using the slow, free option.
• Mandatory participation. When every bank must join and must accept incoming instant payments, a sender never has to wonder whether a transfer will work. Voluntary systems, such as FedNow, grow only as fast as the slowest banks decide to sign up.
• A shared front end. A single national QR code and simple aliases, such as a phone number or a virtual address, let small merchants accept payments without a terminal. That turned the Recife coconut stall into a digital merchant.
• Room for non-bank apps. UPI's growth came largely through third-party apps that built consumer products on top of the rail. Systems that keep the interface inside bank apps tend to grow slower.
• A reason to switch. Card rewards are a powerful habit in the US. In India and Brazil, the alternative to instant payments was often cash, and cash has no loyalty points.

None of these choices is free. Zero pricing means someone else pays for the infrastructure, usually banks or the state. That is why the business model of UPI remains a live debate in India even as volumes hit records.

The risk side: fraud moves at the speed of the rail

Every instant system inherits a fraud problem. A payment that settles in seconds and cannot be reversed is ideal for a scammer who has persuaded a victim to send money. The UK, an early mover with Faster Payments, now requires banks to reimburse most scam victims and splits the cost between sending and receiving banks. We look at that model, and the role AI now plays in scams, in AI fraud in payments. Any country scaling an instant rail without a liability framework is, in effect, subsidising fraud.

Cross-border: linking the national rails

The next phase is connecting systems across borders. Bilateral links already exist across Asia, including UPI connections abroad and links involving Thailand's PromptPay. The multilateral version is Project Nexus. According to the BIS, the central banks of India, Indonesia, Malaysia, the Philippines, Singapore and Thailand incorporated Nexus Global Payments in 2025 to take the scheme into live operation, and the BIS has handed the project over to it. Instead of building dozens of bilateral links, each system connects once to a shared hub.

If it works, a sender in Singapore could pay a recipient in India in roughly the time it takes to send a domestic transfer, with the FX rate shown upfront. That would put real pressure on the cost of remittances in Asia's busiest corridors. More in our Project Nexus explainer.

What to watch next

• Monthly NPCI data: whether UPI's daily average holds above 800 million through the October-November festival season.
• October 2026: FedNow Q3 2026 statistics from the Federal Reserve, the first test of whether the Q2 jump was a one-off.
• January 2027: the ECB's payment statistics for the first half of 2026, showing the first full half-year after the EU instant payment mandate took effect for eurozone banks.
• 2026-2027: Nexus Global Payments' first live connections between member systems.
• Ongoing: Pix's next record days, which tend to fall on salary dates and the December 13th-month payment.

A retired teacher in Lisbon still pays at the market with coins. Her grandson in Munich has not touched a banknote in months, and every tap of his phone runs through a card network headquartered in the United States. The digital euro is Europe's attempt to give both of them a public, euro-native way to pay, and in 2026 it finally moved from theory into legislation.

Where things stand

The European Central Bank's Governing Council decided on 29 October 2025 to close the preparation phase and move into the next stage, focused on building technical capacity and getting ready for pilot transactions, as reported by TechRepublic. The ECB's target is to be ready for possible issuance in 2029, but only if EU law is in place.

That law moved forward in two steps this summer. The Parliament's economic and monetary affairs committee backed its position on 23 June 2026 by 43 votes to 14, with one abstention. Then, after three groups forced a plenary vote, MEPs approved the negotiating mandate on 9 July 2026 by 416 to 169, with 22 abstentions, according to Brussels Signal. Talks with the Council, which has already adopted its position, aim for a deal by the end of 2026.

Why Europe wants one

The argument that keeps coming back is sovereignty. ECB figures cited in the Parliament debate show Visa and Mastercard handling 61% of card payments in the euro area and nearly all cross-border card transactions. As cash use falls, that share of everyday spending running through non-European networks grows. A digital euro would give Europe a payment option that works across all euro countries, built and governed at home.

There is a resilience argument too. If a card network or a large processor goes down, as happens from time to time, a public option that also works offline gives shops and customers a fallback that does not depend on one private company.

The second argument is about cash itself. The same legislative package strengthens the legal status of banknotes and coins and would require member states to keep cash accessible. The pitch is a pairing, not a replacement.

Holding limits: the real fight

The question that has slowed everything is how much digital euro one person can hold. Banks fear that if people can move unlimited savings into central bank money, deposits will drain away, especially in a crisis. So there will be a cap.

The European Commission would set the individual limit after ECB advice and review it regularly. No figure is in the law yet. ECB analysis has found that limits of up to €3,000 per person would not threaten financial stability even in extreme scenarios, and Eurogroup finance ministers agreed in September 2025 on how the limits should be governed. Businesses would generally not be allowed to hold large balances, but could keep incoming payments for up to 24 hours. The digital euro would pay no interest and charge none, so it is a way to pay, not a place to save.

The holding limit is where monetary policy meets the kitchen table.

Privacy and offline payments

Privacy is the issue most voters raise. The draft framework, as summarised by Crowdfund Insider, follows privacy by design and by default. The ECB would not see who is paying whom; identifying data stays with the bank or provider that runs the wallet.

Offline payments go further. They would run device to device using secure local storage, with no internet connection, and are meant to offer something close to the privacy of handing over cash. They would also be free.

There was a moment this year when the whole design nearly shrank to offline only. The lead rapporteur, Fernando Navarrete, had pushed that idea before dropping it in March 2026, Euronews reported, clearing the way for a version that works both online and offline.

What it costs, and who pays

• Users: basic services such as opening a wallet, holding funds and paying would be free.
• Merchants: most would have to accept it, with exemptions for very small businesses that take no other electronic payments. Fees would be capped and should not exceed what retailers pay for comparable electronic methods.
• The Eurosystem: roughly €1.3 billion to build the infrastructure and about €320 million a year to run it after launch.
• Banks: estimated investment of €4 billion to €5.8 billion across the sector.

Distribution would run through banks, e-money firms, post offices and certain regulated crypto-asset service providers.

What changes at the till

For a café owner in Lyon or a bakery in Valencia, the most important line in the draft is that merchant fees should not exceed what they pay for comparable electronic payments, with caps in place. Card acceptance is one of the larger costs for small European retailers, and they have little bargaining power with the international schemes. A public alternative with capped fees gives them leverage, even if most customers keep using cards.

Acceptance would be compulsory for most businesses, which is how a new payment method avoids the chicken-and-egg problem of shops waiting for customers and customers waiting for shops. The exemption for micro-businesses that accept no other electronic payment protects the market stall that only takes cash.

The timeline

1. End of 2026: target for a political deal between Parliament, Council and Commission.
2. Mid-2027: earliest start of pilot transactions with payment providers and merchants.
3. At least 24 months: rollout period for banks, merchants and the public once the law is in force.
4. 2029: the ECB's stated target for first issuance.

Our take

We think the digital euro's success will depend less on technology and more on whether it feels useful on an ordinary Tuesday. A free, euro-wide payment option that works offline is a real improvement for travellers, small shops and anyone without a credit card. A capped balance with no interest is unlikely to excite savers, and it is not meant to. For a broader comparison with other forms of digital money, see stablecoins vs CBDCs vs tokenised deposits, and for Europe's crypto rules, our MiCA explainer.

Most blockchains want to be everything: a casino, a computer, an art gallery and a bank. Stellar has spent more than a decade wanting to be one thing, a way to move money across borders cheaply. For a long time that sounded unambitious. In 2026 it looks like good timing.

Built for one job

Stellar was created in 2014 with remittances and financial inclusion in mind. It's designed around assets issued by real institutions, such as dollars, euros or tokenised funds, and around "anchors": licensed companies that let people move between local currency and the network. Transactions settle in a few seconds and cost around one hundredth of a cent.

That focus means fewer fireworks. It also means fewer surprises, which is exactly what you want from payment rails.

Figures are as reported by the Stellar Development Foundation in its Q2 2026 update; the volume figure counts stablecoin transfers on the network in April to June 2026.

The MoneyGram story

MoneyGram first connected to Stellar in 2021 to settle with USDC, letting people move between cash and digital dollars at its locations. In 2026 it went further and launched its own regulated digital dollar, MGUSD, on the network. When one of the world's best-known remittance brands puts its own dollar on a blockchain, it tells you how seriously the industry now takes this.

Not just remittance any more

The other big shift this year was institutions. The DTCC, which sits underneath much of the US securities market, picked Stellar as one of the first public blockchains for tokenisation. Franklin Templeton runs a tokenised money-market fund on it. Bermuda chose it for its plan to put its economy on-chain. These aren't speculative projects; they're organisations whose whole business is being careful.

Stellar's advantage was never speed or hype. It was patience: more than a decade of building the boring parts properly.

Why it matters for families sending money

• Cost. When a transfer costs a fraction of a cent to settle, the network is no longer the expensive part of sending money.
• Cash still works. Through anchors and cash networks, people who don't have bank accounts can still send and receive.
• Regulated dollars. USDC, EURC and now MGUSD give families digital dollars issued by accountable companies.

The honest caveats

Stellar is smaller than Ethereum or Tron by stablecoin supply, and its ecosystem of apps is narrower. The quality of the experience depends heavily on the anchors in each country: a great network with a weak local partner still gives a weak result. And, as with every chain, the user should never have to think about which network their money is on.

Still, if you were designing a blockchain specifically for the problem of a construction foreman in Abu Dhabi paying his parents' bills in Cebu or Thrissur, it would look a lot like this one.

On a Friday evening in Dubai's Al Karama, the queue outside an exchange house can stretch down the pavement. Inside, electricians, drivers and shop assistants hand over dirhams and walk out with a receipt that becomes rupees in Kochi, taka in Sylhet or rupees in Pokhara within minutes. Behind each receipt is a household budget: a mortgage instalment in Thrissur, a younger brother's college fees in Sylhet, a new roof in a village outside Pokhara. This corridor, from the six Gulf states to South Asia, is one of the densest flows of household money on earth. In 2025 and 2026 it set records almost everywhere, and the way it moves is starting to change.

The scale of the flow

The receiving countries tell the story. On the RBI's fiscal-year count, India took in a record $135.46 billion in remittances in 2024-25 (IFAD's calendar-year estimate for 2025, on a different definition, is $150.7 billion), according to RBI data reported by All India Radio. Bangladesh received $35.56 billion in fiscal 2025-26, up from $30.33 billion a year earlier. Nepal's central bank reported Rs 2,363 billion, or $16.19 billion, in fiscal 2025-26, a 28.1% rise in dollar terms. Sri Lanka passed $8 billion in workers' remittances for the first time in 2025, reaching $8.07 billion, up 22.8%.

On the sending side, the Central Bank of the UAE reported that outward transfers through the country's exchange houses rose 27.8% to Dh188.8 billion in 2025, with personal remittances the largest slice at Dh115.7 billion, as covered by Khaleej Times.

India: the Gulf is still huge, but no longer number one

The RBI's latest remittance survey, covering 2023-24, found that the GCC states together accounted for 38% of India's inward remittances, with the UAE alone at 19.2%. The United States, at 27.7%, has now overtaken the UAE as the single largest source, and the UK has climbed to 10.8%, according to Business Standard. That shift reflects a more skilled, higher-earning diaspora in Western economies. It does not mean the Gulf is fading: Indians in Gulf countries still make up around half of all Indian migrants worldwide, and for many households in states such as Kerala, Gulf wages remain the main income.

Why the cost gap is so wide

South Asia is the cheapest receiving region in the World Bank's dataset, with an average cost of 4.80% for $200 in the first quarter of 2025, against a global average of 6.49% that quarter (6.36% by Q3 2025). But the averages conceal big differences between corridors.

Sending $200 from the UAE to India cost 3.72% on average in the third quarter of 2025, according to World Bank Remittance Prices Worldwide. From Saudi Arabia to Bangladesh, the average was 10.20%, and more than half of that came from the exchange rate margin rather than the fee. Private data points the same way: FXC Intelligence's 2025 rankings placed Bangladesh seventh among the costliest of 35 major receiving markets, while Nepal and Sri Lanka sat among the ten cheapest.

The reason is mostly structural. The UAE to India corridor is crowded with banks, exchange houses and apps competing on a liquid currency pair. Bangladesh has had a managed exchange rate and a long history of informal hundi channels, which keeps formal margins wider. Our piece on the mid-market rate explains why a zero-fee headline can still cost more.

On the same street in Riyadh, two workers sending the same $200 home can pay a threefold difference, depending only on where home is.

From counters to apps

Exchange houses built this corridor, and many are now competing with their own apps. A Visa survey of UAE residents published in September 2025 found nearly two in three preferred mobile apps and digital platforms over visiting a branch, with ease of use the most cited reason and high cost the most common complaint.

The receiving side is going digital too. India's UPI has linked with several overseas systems, and in June 2026 India and Nepal went live with a person-to-person connection between UPI and Nepal's National Payments Interface, bringing the number of countries where UPI is live to 10. Our guide to UPI's global expansion covers the other links. For a worker in Doha whose family banks in Nepal, those connections could eventually mean cheaper and faster routes with fewer intermediaries.

Nepal and Sri Lanka: dependence and resilience

Nepal is among the most remittance-dependent economies in the world, and the 2025-26 surge was striking. One detail deserves attention: first-time foreign employment approvals fell to 406,519 from 505,957, while renewals rose. More money is coming from workers already abroad, which may not repeat if fewer people leave.

Nepal is also a reminder of how much of this corridor is a Gulf story. Qatar, Saudi Arabia, the UAE and Kuwait have long been among the main destinations for Nepali workers on labour permits, and the money they send pays for school fees, house construction and loan repayments in hill districts with few other sources of cash income.

Sri Lanka's record year marks a recovery from the crisis years, when much money moved informally. December 2025's $879.1 million was the highest monthly inflow on record, according to the Central Bank of Sri Lanka figures reported by Ada Derana.

What would actually lower costs

• Price transparency on FX. Showing the margin against a reference rate, not just the fee, would expose the most expensive offers.
• Fast-payment links between Gulf instant payment systems and South Asian ones, so fewer correspondent banks take a cut.
• Direct bill payment. A large share of this money pays school fees, electricity and loans. Paying those directly from the Gulf, as we argue in bills are the real remittance, removes a cash-out step.

The corridor's volume is not in question. Its price is, and that is where the next few years of competition will be fought, one exchange-house receipt at a time.

The message usually arrives at a bad hour. A hospital billing desk in Coimbatore wants an advance before an angiogram; a cousin in Iloilo says the caregiver has quit and Lola cannot be left alone tonight. The child is in Calgary or Riyadh, eight or five time zones away, with a card that works abroad and a parent whose phone keeps asking for an OTP nobody can read. Long-distance care is often described as an emotional problem. In practice, much of it is a payments problem: getting the right amount to the right counterparty in India or the Philippines, quickly, without handing a frail parent a pile of cash.

This guide looks at what that care actually costs in the two countries with the largest overseas families, which public schemes now carry part of the load, and how to structure the money so that the 2 a.m. call is a bill to settle rather than a crisis.

Ageing at home, children abroad

India is ageing faster than most people realise. The UNFPA's India Ageing Report found that people aged 60 and over were 10.5% of the population in 2022 and are projected to reach 20.8% by 2050, with the elderly outnumbering children under 15 by around 2046. About one-fifth of older Indians already live alone or only with a spouse, 71% live in rural areas, and nearly a third of older women have never worked and have no income of their own.

Those numbers overlap heavily with migration. Kerala, which received about a fifth of India's remittances in the RBI's last survey, is also one of the oldest states, with women at 60 expecting around 22 more years of life. The pattern is familiar: one child in the Gulf or North America, another in Bengaluru, and parents in the family home who are fine until, suddenly, they are not.

The Philippines has a younger population but an even heavier reliance on overseas earnings. Personal remittances reached a record $39.62 billion in 2025, with cash remittances equal to 7.3% of GDP. The United States supplied 39.7% of cash remittances, followed by Singapore at 7.3% and Saudi Arabia at 6.6%. A large share of that money pays for medicine, maintenance drugs and someone to sit with an elderly parent. Our explainer on OFW remittances covers the corridor itself.

Who pays the hospital: public cover is finally growing

For decades, Indian families paid most health costs out of their own pockets. That is changing, slowly. The government's National Health Accounts show out-of-pocket spending fell to 39.4% of total health expenditure in 2021-22, from 62.6% in 2014-15, while government spending rose to 48% of the total. Even so, almost two rupees in every five spent on health still come straight from households, and for families with a child abroad, that often means a transfer.

The biggest recent change for older parents is the Ayushman Vay Vandana card. Since October 2024, every resident Indian aged 70 and above qualifies for Rs 5 lakh of hospital cover under AB PM-JAY, regardless of income. By 31 December 2025, about 96.7 lakh cards had been issued and 10.3 lakh hospital admissions authorised, worth Rs 2,154.37 crore, across more than 33,000 empanelled hospitals. That works out to roughly Rs 20,900 per admission, which suggests the scheme is mostly paying for routine procedures and short stays rather than long ICU episodes. Enrolment is done online with e-KYC, which a child abroad can often walk a parent through on a video call.

In the Philippines, the shift happened earlier. Under Republic Act 10645, every Filipino aged 60 and over who is not already a PhilHealth member through work is automatically covered, with premiums paid by the national government from sin-tax revenue. Coverage includes room and board, professional fees, medicines, outpatient consultations, the Konsulta primary-care package and dialysis, and zero-balance billing has expanded in government hospitals. Indigent seniors also receive a social pension of P1,000 a month, doubled under RA 11916, reaching more than 4 million people by end-2025 with over a million more on the waitlist.

Public cover now handles the big hospital bill more often than before. What it does not handle is the Tuesday afternoon: the caregiver, the pharmacy run, the electricity bill.

India and the Philippines side by side

Sources: PIB National Health Accounts, Outlook Money, Gulf News, GMA News, Samarth, RBI, BSP via Philstar. Care costs are indicative ranges, not quotes.

Worked example: budgeting care for a parent in a Tier-2 Indian city

Take a 72-year-old widow living alone in Coimbatore, with a son in Calgary. She is mobile but has diabetes and needs help with cooking, bathing and medicines. Here is what a sensible monthly structure looks like, using published ranges.

• Care: Samarth's 2025 survey of home-care pricing puts part-time attendants at Rs 12,000-25,000 a month and basic live-in care at Rs 35,000-50,000, with Tier-2 cities running 15-20% below metros. A live-in attendant in Coimbatore might therefore cost around Rs 28,000-42,500. Budget Rs 35,000.
• Hospital risk: She qualifies for the Vay Vandana card, which covers up to Rs 5 lakh at empanelled hospitals. That does not remove the need for private cover if the family prefers non-empanelled hospitals, but it changes the size of the emergency fund.
• Recurring bills: electricity, water, phone, gas cylinder, property tax and medicines. These are best paid directly by the son, not routed through her account. Systems like Bharat Connect let many of these billers be paid from one interface.
• Emergency buffer: At Rs 20,900 per average Vay Vandana admission, a buffer of Rs 1-1.5 lakh held in an account the son can operate covers deposits, non-covered consumables and transport for several episodes.

Total run-rate: roughly Rs 35,000 for care plus Rs 8,000-12,000 for bills and medicines, or about Rs 5.2-5.6 lakh a year, with a one-off buffer of Rs 1-1.5 lakh. Moved as a single monthly transfer, that is a manageable number. Moved as panic transfers each time something breaks, it costs more in fees, worse exchange rates and stress.

If the son still files Indian returns under the old tax regime, health insurance premiums paid for senior-citizen parents can qualify for a deduction under Section 80D, within a combined ceiling of up to Rs 75,000 or Rs 1 lakh depending on ages, provided the premium is paid by any mode other than cash. The deduction is not available under the new regime. Tax treatment for NRIs depends on individual circumstances, so check with a qualified adviser; our guide to tax on money sent to India covers the basics.

The Philippine version

Now take a 68-year-old father in Quezon City with a daughter working in Riyadh. He is automatically covered by PhilHealth. The household's main cost is a live-in kasambahay, whose legal minimum in Metro Manila rose to P7,800 a month from 7 February 2026, up from P7,000. Twelve months at the minimum is P93,600 a year before food, lodging and benefits, and anyone with caregiving skills will expect more. If he receives an SSS pension, it gets a 10% increase each September through 2027, which helps cover maintenance medicines.

Pay the counterparty, not the parent

The most useful shift in long-distance care is structural. Instead of sending a lump sum to a parent and hoping it is spent well, pay the people and institutions directly wherever possible.

1. Caregivers: pay by bank transfer or UPI on a fixed date, with a written agreement. Cash handed over by a frail parent invites disputes and, sometimes, exploitation.
2. Utilities and subscriptions: pay billers directly. Our guide to bill payments for NRIs walks through the options.
3. Hospitals: register the parent's insurance or Vay Vandana card with the nearest large empanelled hospital in advance, so admission can be cashless. Ask the hospital which international payment methods its billing desk accepts before you need it.
4. Insurance: pay premiums from your own account and keep receipts, both for continuity and for any tax claim.
5. Emergency money: keep a defined buffer in an account you can access remotely, with a trusted local relative or friend who can act on the ground.

Paying directly also protects parents from the fastest-growing threat to their savings: scams. Older people who receive regular transfers are prime targets for fake bank calls, KYC-update links and impersonation. See our guide to money transfer scams before setting anything up.

The paperwork that saves a week

Every family that has handled a hospital stay remotely says the same thing: the money was not the hard part, the authority was. Banks will not talk to a child about a parent's account without a mandate. Hospitals want a local signatory. Insurers want original documents. A few hours spent now, while parents are well, removes most of that friction.

• A registered power of attorney or bank mandate for at least one account.
• Copies of ID, insurance cards, Vay Vandana or PhilHealth details, prescriptions and recent test results, stored where the child and a local contact can reach them.
• Nominee and joint-holder details updated on bank accounts and deposits.
• A named local person, with a written list of who to call and what they can approve.

None of this is legal advice, and rules on powers of attorney and joint accounts differ between India and the Philippines, so a local lawyer is worth an hour of fees.

What to watch next

• Through September 2027: the Philippines' SSS pension increases continue, with retirement pensions rising 10% each September. The September 2026 step-up reached about 3.8 million pensioners; the final one is due in September 2027.
• 2026-27 Philippine Congress: the Universal Social Pension Bill, which would extend at least P1,000 a month to all seniors, has cleared a House panel but still needs House plenary and Senate approval.
• Ongoing: Vay Vandana enrolment and claims data from the National Health Authority. Watch whether average claim values rise as older beneficiaries use the cover for larger procedures.
• December 2026: the World Bank's next migration and remittances brief, which will show whether health and care spending is pushing remittances to India and the Philippines to new highs.
• February 2027: India's Union Budget, the usual venue for changes to health-scheme funding and senior-citizen tax provisions.

Our view: public cover in both countries has improved enough that the catastrophic hospital bill is less likely than it was a decade ago. The real cost of caring from abroad is now the steady, unglamorous stream of wages, medicines and household bills, and the families who handle it best are the ones who treat it like a small payroll rather than a series of emergencies. For more on how remittances are actually spent, see what remittances pay for.

The stablecoin industry has a credibility problem with its own statistics. Headlines quote trillions of dollars in transfers, but most of that is trading bots and money shuffling between exchange wallets. Strip it out and you are left with a smaller, more interesting number: the payments real businesses and people make. In 2026 that number is finally big enough, and growing fast enough, to take seriously.

Start with the honest number

The clearest measure comes from a joint analysis by McKinsey and Artemis Analytics, which counted about $390 billion of genuine stablecoin payments in 2025. That is roughly 0.02% of global payments. Tiny, in other words, but concentrated in exactly the places where existing rails are slowest and most expensive.

Two things stand out. Businesses dominate by value. But stablecoins have their deepest penetration in person-to-person transfers, where they already handle about 0.37% of the market, which is mostly cross-border family money. This year the pace picked up: identified stablecoin payments reached at least $401 billion between January and August 2026, up 42% on the year before.

Trend 1: B2B is the engine

An importer in Dhaka paying a supplier in Shenzhen does not care about blockchains. She cares that a wire takes three days, costs a fixed fee plus a currency spread, and sometimes gets stuck at an intermediary bank. A stablecoin transfer settles in minutes at any hour. Artemis and Stablecon estimated that B2B stablecoin payments grew more than 730% during 2025, led by small and mid-sized firms rather than multinationals.

The infrastructure behind this is consolidating into big hands. Stripe closed its $1.1 billion purchase of the stablecoin platform Bridge in February 2025, and Mastercard completed its acquisition of BVNK on 3 August 2026, describing cross-border B2B payments, payouts and treasury flows as the target, according to Mastercard's announcement. Background on why legacy rails are slow sits in our correspondent banking explainer.

Trend 2: card networks are settling in stablecoins

The most underrated shift is happening behind the card swipe. When a cardholder spends from a stablecoin balance, the merchant still gets paid in local currency, but the settlement between the card program and Visa can now happen in USDC. Visa's stablecoin settlement passed a $20 billion annualized run rate in September 2026, up more than fifteenfold in a year, with more than 160 stablecoin-linked card programs live, according to The Block. As recently as April the run rate was $7 billion.

Spending on crypto-linked cards has grown alongside. Artemis put it above $18 billion in early 2026, close to the volume of direct peer-to-peer stablecoin transfers. If you are new to these products, start with our guide to crypto cards.

Trend 3: checkout is opening up

Merchant acceptance moved from experiment to default in some places. Shopify began rolling out USDC payments in June 2025, built with Coinbase and Stripe on the Base network, switched on by default for eligible merchants and with up to 0.5% cash back for merchants who accept it. Stripe now uses Open USD as its default stablecoin for business transactions after the token's launch on 30 September 2026. Open USD is jointly owned by Visa, Mastercard, Stripe, Shopify and Coinbase, and more than 200 companies signed up by launch day.

Will shoppers in Toronto start paying for groceries in stablecoins? Probably not soon. The real use is cross-border e-commerce, where a small brand selling to customers in 40 countries can avoid card fees and currency conversion on every order.

Stablecoins are winning where money crosses a border, not where it crosses a counter.

Trend 4: payroll and contractor payouts

Paying remote workers is a natural fit. Deel, which runs payroll in more than 150 countries, launched stablecoin salary payouts for full-time employees on 20 May 2026 on the Polygon network, starting in the US and the eurozone. Employees choose what share of their net pay arrives as stablecoins. Specialist platforms got there earlier: Rise passed $1 billion of payroll volume in November 2025, and a case study cited by BeInCrypto said more than half of its users opt for stablecoin payouts.

Picture a software contractor in Kathmandu paid by a startup in San Francisco. A dollar balance that arrives the same day, which she can hold or convert when the rate suits her, is a meaningful upgrade on a monthly wire.

Trend 5: remittance companies are issuing their own coins

The incumbents of family remittance have stopped waiting. MoneyGram launched its own MGUSD stablecoin on Stellar in June 2026 for its base of more than 60 million customers and nearly 500,000 locations, and Western Union announced its USDPT stablecoin on Solana, issued by Anchorage Digital Bank, for launch in 2026. Both still rely on cash agents and bank partners at the receiving end, which is where most of the cost of sending money still sits. Our cost of sending money home analysis covers that last mile.

What could slow it down

• Off-ramps. Converting stablecoins into rupees, pesos or shillings still runs through licensed local partners, and their spreads can wipe out the savings.
• Rules on rewards. US regulators are still writing GENIUS Act rules, including how far partners can share reserve income with users.
• Fragmentation. More issuers and more chains mean more conversion steps unless interoperability keeps improving.

There is also a measurement problem. Much of the volume that looks like payments is still a business moving its own money between accounts, and the line between a treasury transfer and a real purchase is blurry. Analysts at Artemis and elsewhere are getting better at separating the two, but readers should treat any single payments figure as an estimate with wide error bars.

None of these is a reason to doubt the direction. In 2026 stablecoins stopped being mainly a trading tool and became a quiet settlement layer for businesses, card networks and remittance firms. The next step, machines paying machines, is already being built; see our piece on agentic payments.

For years, the honest answer to "are stablecoins safe?" was "it depends who issued it and whether you trust them." In 2025 and 2026, the US and the EU replaced that shrug with law. It's the single biggest reason stablecoins are moving from crypto curiosity to everyday money.

The US: the GENIUS Act

Signed in July 2025, the GENIUS Act created the first federal rulebook for payment stablecoins in the United States. The headline rules:

• Full reserves. Issuers must hold reserves at least equal to every token in circulation, at all times.
• Safe assets only. Reserves are limited to cash, central-bank deposits, short-dated US Treasuries, Treasury-backed repos and money-market funds that hold only those things, kept separate from the issuer's own money.
• Fast redemption. Holders must be able to redeem for dollars within two business days in normal conditions.
• Checks. Monthly third-party attestations on reserves, plus full audits for the very largest issuers.
• No interest for holding. Payment stablecoins can't pay you yield just for keeping them.

Regulators spent 2026 writing the detailed rules. The OCC published a 376-page proposal in February, and other agencies followed. The law takes full effect in January 2027.

Europe: MiCA

The EU's Markets in Crypto-Assets regulation got there first. Its stablecoin rules applied from mid-2024, and the transition period for crypto firms ended on 1 July 2026. Dollar and euro stablecoins offered in Europe must be "e-money tokens" issued by an authorised institution.

The consequences were visible. USDC and EURC, issued under Circle's French licence, kept flowing. USDT, which isn't authorised, was delisted or restricted for EU users on the big exchanges. At the end of September, Europe's securities regulator recommended going further and stopping licensed firms from servicing non-compliant stablecoins at all.

Regulation didn't kill stablecoins. It gave them the one thing they were missing: a reason for ordinary people and cautious institutions to trust them.

What this means for you

1. The dollar promise has teeth. A regulated stablecoin now has legal reserve, redemption and reporting rules behind it.
2. Where you live changes which coins you can use. Europe and the US are drawing lines; Asia, the Gulf and Africa are writing their own.
3. Fewer cowboys. Apps and exchanges that serve you must be licensed, which is good news for anyone who just wants to pay a bill without worrying.
4. Savings and spending are separating. Because payment stablecoins can't pay interest, yield products are becoming distinct, regulated offerings.

The road ahead

The rest of the world is catching up. The UAE, Hong Kong, Singapore and Japan all have stablecoin regimes in place or on the way, and India continues to study the space carefully. The direction is the same everywhere: stablecoins are being treated less like crypto and more like a new form of electronic money.

For users, that is a good outcome. Clear rules are what let a family trust that money sent every month will keep arriving, year after year.

Europe had an instant payment scheme from 2017. For years it was optional, often more expensive than a normal transfer, and used for a small slice of payments. Then Brussels stopped asking. The Instant Payments Regulation turned instant transfers into the default for euro accounts, and the share of credit transfers sent instantly has roughly doubled in two years.

Why the EU legislated

The SEPA Instant Credit Transfer scheme, known as SCT Inst, lets a euro payment land in another account within ten seconds, at any hour. The technology worked. Adoption did not. In the first quarter of 2024, only 17.3% of SEPA credit transfers were instant, according to a European Payments Council update to the ECB's retail payments board. Many banks did not offer it, and those that did often charged a premium.

For EU policymakers, that was a strategic problem as much as a consumer one. Without a common instant rail, European retail payments leaned on international card schemes and big tech wallets. The regulation, adopted in March 2024, was meant to fix the base layer.

The four obligations

The rules, summarised by the European Central Bank, boil down to four requirements for payment providers that offer ordinary euro credit transfers.

1. Receive instant payments. Euro area banks had to be reachable for SCT Inst by 9 January 2025.
2. Price parity. From the same date, an instant transfer can cost no more than a standard one. If a normal transfer is free, so is an instant one.
3. Send instant payments. Euro area banks had to offer sending by 9 October 2025.
4. Verification of payee. Also from 9 October 2025, banks must check, free of charge, whether the name the customer types matches the account holder behind the IBAN, for standard and instant transfers alike. The answer comes back as match, close match, no match or not possible.

Banks also have to screen their customers against EU sanctions lists at least daily, instead of checking every single instant payment, which removes a technical reason for delays. Member states outside the euro have until January and July 2027, and electronic money and payment institutions have until April 2027.

What happened to volumes

The regulation worked on its own terms. EPC data presented to the ECB in May 2026 shows instant transfers reached 35.6% of SEPA credit transfers in the first quarter of 2026, roughly double the 2024 level. SCT Inst processed 10.1 billion transactions in 2025, up 73% on the year before. Some 2,817 payment providers had signed up to the scheme by April 2026, covering all 21 euro area countries.

Price parity did the heavy lifting. Once a bank could no longer charge a premium for speed, keeping a separate slow product made little commercial sense, and many banks simply switched their apps to send instantly by default. For customers the change was often invisible: the transfer button looked the same, but the money arrived before they had closed the app.

That still leaves almost two-thirds of credit transfers going the slow way. Part of that is corporate batch payments, such as salaries and supplier runs, which run on a schedule where seconds do not matter. Part of it is habit. And instant credit transfers remain a small share of all non-cash payments, about 7% in the euro area at the end of 2025, because cards still dominate at the till.

Price parity did more for instant payments than seven years of voluntary adoption.

Verification of payee: the sleeper feature

The name check may end up being the most noticeable change for ordinary people. Before October 2025, a bank transfer in most of Europe went through on the IBAN alone. Type a wrong digit, or get tricked by a fake invoice, and the money went wherever the number pointed.

Early data is encouraging. A Natixis review found that across December 2025 and January 2026, 66% of checks returned a match, 13% a close match and 19% no match. In Latvia, the central bank said 57.4 million checks ran in the first five months, and half of surveyed people who saw the check said it helped them avoid a wrong transfer, according to Latvijas Banka.

A 19% no-match rate sounds alarming, but many of those are innocent: nicknames, married names, company trading names. The real test is whether customers learn to stop when they see a warning, or click through out of habit. Natixis also cautioned that the check can create a false sense of security, since a fraudster can open an account in a name that matches.

Who benefits next

Instant rails at scale create room for new products. Wero, the wallet from the European Payments Initiative, runs on SCT Inst and reported more than 43 million registered users across Germany, Belgium and France by August 2026, with merchants such as Lidl accepting it. The regulation also opened access for payment and e-money institutions to central bank settlement systems, letting fintechs plug in more directly.

Next comes the payments package that replaces PSD2. A provisional deal in November 2025 on the new Payment Services Regulation extends name and identifier checks and shifts more fraud liability onto providers, covered in our piece on open banking. And the ECB's digital euro project, if it goes ahead, would sit alongside these rails rather than replace them.

What it means for people sending money

For a freelancer in Lisbon invoicing a client in Frankfurt, the change is simple: payment arrives in seconds, at the same price as before, with a name check that catches mistakes. For someone sending money outside Europe, nothing changes yet, because SCT Inst is euro-only and stops at the SEPA border. Linking Europe's instant rail to systems such as India's UPI or Brazil's Pix is the next frontier, and it is the subject of projects like Nexus.

The lesson for other regions is blunt. Voluntary schemes grow slowly when banks earn more from the old way. A rule that sets price parity and a hard date moved Europe further in two years than the previous seven.

On 25 July 2025, Galaxy announced that it had sold more than 80,000 bitcoin for a client it described only as a Satoshi-era investor, a trade worth over $9 billion at the prices of the day. Nobody watching the public order books saw a single order of that size. That is the point of an over-the-counter desk: to move a position large enough to wreck a market without the market finding out until it is done. The same machinery, at a far smaller scale, now handles a Lagos importer converting stablecoins into dollars and a Dubai trading house swapping dirhams for USDT before paying a supplier in Shenzhen.

What an OTC desk does, in plain terms

An OTC desk is a dealer or broker that trades directly with a client rather than through a public order book. The client asks for a price on a specific size; the desk quotes a single all-in price for the whole lot; if the client accepts, the two sides settle bilaterally. There is no partial fill, no visible order, and no slippage after the quote is accepted.

Desks come in two broad types. Principal desks take the other side of the trade onto their own balance sheet and manage the risk afterwards, which is what the large trading firms do. Agency desks find the other side, or work the order across exchanges and liquidity providers, and charge for the service. Many exchanges run both an order book and an OTC desk, and the desk routes some flow into the book over time.

The minimums tell you who the service is for. Kraken's published OTC terms set a floor of $100,000 equivalent, offer automated executable quotes for orders between $100,000 and $1 million, take larger trades by chat, and require all trade obligations to be settled within 24 hours. Those numbers are typical of exchange-affiliated desks. Pure dealer desks often start higher.

How a block trade and an RFQ actually work

The request-for-quote (RFQ) is the basic unit of OTC trading. It has five steps, and the economics hide in the second.

1. Request. The client names the asset, side and size: sell 5 million USDT for US dollars, say, or buy 250 BTC against USDC.
2. Quote. The desk prices the trade off the live market, then adds a spread that reflects size, volatility, the asset's liquidity, how quickly the desk can hedge and how much it trusts the client to settle.
3. Accept. Quotes are live for seconds. On chat desks, a trader confirms the lot and price in writing; on RFQ screens and APIs, a click locks it in.
4. Settle. Cash and crypto move between the two sides, either from balances already held with the desk or by wire and on-chain transfer.
5. Confirm. Both sides get a trade confirmation for books and compliance.

For really large trades, a desk may "work" the order over hours or days instead, executing in slices across venues and charging a fee or a guaranteed average price. The Galaxy sale was described as a completed sale, and the firm did not say how it was executed. Its fourth-quarter 2025 results show what that kind of business looks like from the inside: 1,620 trading counterparties at year-end, up from 1,532 a quarter earlier, and a loan book of about $1.8 billion.

The market in 2025: stablecoins took over