What It Costs to Make a Payment Look Instant

A cross-border payment can arrive in seconds because the money behind it was parked in another country days earlier. The customer sees a fee and an exchange rate. What they do not see is the cost of funding that money while it waits.

In the UAE, exchange houses sent AED 188.8 billion abroad in 2025. If that entire flow needed two days of advance funding, it would imply roughly AED 1 billion tied up. That is a scenario, not a measured industry balance. But it gives the hidden cost some scale.

It also points to what comes next. The UAE has licensed dirham payment tokens for payments at home. Using them across borders at scale is still ahead, and the likely customers are already visible.

The World Bank puts the UAE second only to the United States for personal remittance outflows: about $58.5 billion in 2024. That is an economy-wide measure, not the narrower UAE exchange-house series used below. On exchange-house corridors, India took 28% of outward transfers in 2024, Pakistan 13% and Bangladesh about 8.5%.

On those corridors, sending money home is already cheap and fast. Digital services charge 1% to 2.5%. At an exchange-house counter, 3% to 4% is common. India's UPI, Pakistan's Raast and Egypt's InstaPay pay out the local leg in seconds.

In July I argued that the fast part of a cross-border payment is rarely the expensive part. The cost sits in the money somebody parks in advance so the payment can look instant. I ended that piece with three words: watch the warehouse.

A remittance can look instant because the money was already there. An exchange house promising payment in seconds in Karachi has usually bought rupees and placed them with its payout partner before the customer walks in. The customer's payment and the institution's settlement are two separate events.

UAE exchange houses sent AED 188.8 billion abroad in 2025, about AED 517 million a day. Fund that entire flow two days ahead and you tie up about AED 1 billion. Five days ahead, about AED 2.6 billion. Actual requirements vary by corridor, netting and liquidity arrangements, and there is no published industry-wide measure of how many days providers fund ahead. One public reference point: a listed UAE exchange group says more than $150 million can be provisioned for nostro funding at any given time.

That money is not free. At an illustrative annual funding cost of 6% to 20%, a continuously outstanding AED 1 billion balance would cost AED 60 million to AED 200 million a year. In practice, the cost lands somewhere: in the fee, in the exchange rate or in the provider's margin.

So every cross-border payment has two prices. One is on the receipt: the fee and the exchange-rate margin. The other sits on the provider's balance sheet: how much it commits before the payment, and for how long. The sender sees the first. Only the treasurer sees the second.

On the UAE's biggest corridors, competition has already pushed the first price down. The second is where the next gains are.

The picture of a migrant worker sending wages home is still right. It is getting less complete.

The central bank splits exchange-house outflows by purpose. In 2025, AED 115.7 billion was personal remittances and AED 63.3 billion was trade payments, about a third of the total. Since 2022, personal transfers have grown 9% and trade payments 122%. But the rise in trade was not smooth: it was just AED 26.9 billion in 2024, before jumping in 2025. That is a change worth watching, not a trend line to extrapolate confidently.

That changes what matters. For a family sending AED 1,000, the fee and the rate are everything. For a company paying a supplier $10 million, timing matters about as much as price. A rate 5 basis points better saves $5,000. Getting the same $10 million there three days sooner, at a 5% cost of funds, saves about $4,100.

In 2009, when China's trade with Africa had just reached $107 billion a year, Ben Simpfendorfer, then chief China economist at the Royal Bank of Scotland, told The National: "It makes more sense to service that trade from Dubai than it does to Africa." Last year China's trade with Africa was $348 billion.

The UAE now has trade worth more than $100 billion a year with each of China ($111.5 billion of non-oil trade in 2025), India ($101 billion in 2025-26) and Africa (about $158 billion of non-oil trade in 2025, up from $112 billion a year earlier).

Those trade flows put Dubai close to a hard problem: exporters want reliable payment in dollars or another usable currency, while many African importers struggle to source dollars when they need them.

Some have stopped waiting for a bank. Afriex, a payments firm citing a TechCabal report, describes importers in Lagos, Accra and Nairobi paying Chinese suppliers in USDT instead of by wire, cutting waits of up to a week to minutes. Some freight forwarders now list USDT as a payment option on their invoices.

Stablecoin payments are still a small share of the total: McKinsey and Artemis put genuine stablecoin payments at about $390 billion a year by December 2025, roughly 1% of reported stablecoin volume. But $226 billion of that was business to business, up 733% in a year. That is why business settlement, rather than retail remittances alone, is worth watching.

This is where I go further than the report.

The UAE did the regulatory work early. The central bank's payment-token rules have been in force since 2024 and license dirham tokens for payments on the mainland. Several are approved, the first in December 2024. Dollar stablecoins are limited to narrow uses, such as buying virtual assets, and converting between tokens is itself a licensed activity.

The bigger opportunity is across borders, and that part is still to come. In my view, it is one reason dirham tokens are still early: their biggest job has not started at scale yet.

A family in Kerala does not need a cross-border dirham token; their money already lands in seconds over UPI. The importer in Lagos paying a supplier in Guangzhou through Dubai is a different customer: large payments, slow banks, scarce dollars, and a UAE provider prefunding the gap.

The economic case is easy to see from Abu Dhabi. The dirham has been pegged at 3.6725 to the dollar since 1997. A dirham-denominated token therefore carries relatively stable dollar exchange-rate exposure while that peg holds.

There is official ambition here too. When the central bank unveiled the new dirham symbol in 2025, it said the symbol contributes to its efforts to establish the dirham as an international currency. A regulated dirham-token corridor that businesses actually use would be one practical expression of that ambition.

Doing it well takes groundwork, and much of it is already being laid. A token that crosses a border needs supervision that travels with it, a licensed partner at the other end to turn it into naira or shillings, and a working relationship with the receiving country's central bank, which has its own rules. A corridor is only as fast as its slowest licensed step, so corridors are likely to open one at a time.

Whatever the rules allow, the evidence so far says the incumbents will carry it. Exchange houses hold the licences, the branches, the payroll relationships and the payout partners. Banks hold the balance sheets. New settlement infrastructure is more likely to sit underneath them than to replace them.

In September, First Abu Dhabi Bank and Citi ran live dollar transactions through Swift's ledger: the tokenised money stayed on the banks' balance sheets, and settlement ran through correspondent accounts. The state's own rails are moving too. The Digital Dirham made its first government payment over mBridge in November 2025, and Aani, the UAE's instant payment platform, had 12.5 million users by April.

The report's base case for 2030 is what it calls an invisible hybrid: banks, exchange houses and fintechs use dirham tokens, tokenised deposits and public rails behind familiar apps, and customers rarely see a token. I agree, with one addition. The first place the hybrid shows up will not be a cheaper remittance. It will be a smaller balance sitting in a payout partner's account, and a supplier paid the same day.

That also tells you what to measure. A payment moved onchain is not proof of value by itself. The useful numbers are the money held in advance by corridor, the time to final settlement, failed payments and total cost, before and after.

My prediction, with moderate confidence: as regulated cross-border corridors using dirham tokens reach meaningful scale, their first substantial volumes will be businesses paying suppliers, not workers sending wages home. Banks and exchange houses are more likely to carry that business than standalone crypto platforms.

If the first commercially meaningful cross-border dirham-token service turns out to be a mass-market app for workers' remittances, I have this wrong. In the meantime, watch the trade share of exchange-house outflows in the central bank's next financial stability report, likely in 2027. If it keeps rising beyond a third, the demand side of the business-payment case is getting stronger.

The full report, with the corridor data, the dirham token landscape and three scenarios for 2030, is available here.

See you next week.

James Smith

Disclosure: this piece draws on the UAE Cross-Border Payments Outlook 2027, a report I wrote at ADI Foundation. ADI Chain is the infrastructure that DDSC, one of the UAE's licensed dirham tokens, is issued on, and ADI has an agreement with the AfCFTA Secretariat to build digital trade infrastructure in Africa. Where I go beyond the report, the views are my own.