The IMF's warning about stablecoins in emerging markets is a merchant opportunity, not a merchant risk

What the IMF actually said on August 7, and what it didn't

The event was not a press release. Dan Katz, First Deputy Managing Director of the International Monetary Fund, delivered a lecture titled "Stablecoins: Promise, Risks, and Policy Choices for Emerging Markets" at the University of Cape Town on August 7, 2026, hosted by the Faculty of Commerce's School of Economics and its Financial Innovation Hub. It is the most substantive statement the IMF's second-in-command has made on stablecoins as a payments phenomenon rather than as a financial-stability curiosity.

Start with what he conceded, because it is unusual coming from the Fund. Katz said dollar-backed stablecoins attract users through superior liquidity, powerful network effects and seamless cross-border acceptance, and that they deliver real utility for remittances, trade settlement and financial inclusion in places where local banking infrastructure is unreliable. That is not a grudging aside. It is the IMF acknowledging that the demand is rational — that people in emerging economies reach for digital dollars because digital dollars solve a problem their banking system doesn't.

Then came the counterintuitive part, and it is what every outlet led with. A growing number of central banks and banking groups are exploring local-currency stablecoins — a rand token, a peso token, a won token — partly on the theory that a domestic digital alternative will keep users off dollar rails. Katz's argument is that it may do the reverse: local tokens, he said, "might even accelerate the adoption of FX stablecoins."

Note the hedging, because it matters and most coverage stripped it out. He framed this as a risk contingent on country circumstances — the strength of the macroeconomic framework, whether currency substitution is already widespread, the structure of local financial markets — not as an inevitability. On South Africa specifically he said it was "too early to draw firm conclusions." This is a hypothesis with a mechanism, not a forecast.

The dollarization paradox, in plain terms

Here is the mechanism, stripped of the economics vocabulary.

Today, if someone in Johannesburg, Buenos Aires or Lagos wants to move from local currency into dollars, they go through an institution: a bank, an authorised dealer, a bureau, an exchange with a fiat on-ramp. That institution is licensed, reports flows, applies limits, and gives the central bank a place to look and a lever to pull. The friction is the policy instrument.

Now put a rand-linked token and a dollar-linked token on the same chain. The conversion between them is a swap. It happens in a liquidity pool, on a decentralised exchange, or peer-to-peer between two wallets — in seconds, at any hour, with no counter to walk up to. Katz's point is that this migrates foreign-exchange activity away from banks and dealers, and with it the visibility and the levers. The domestic token, intended as a defensive move, ends up building the on-ramp to the dollar token it was meant to displace.

And there is a demand asymmetry underneath. He noted that users gravitate toward dollar stablecoins for liquidity, network effects and cross-border acceptance — three properties a purely domestic token structurally cannot match, because they come from being accepted elsewhere. A rand token is useful in South Africa. A dollar token is useful everywhere, which is exactly why someone holds it.

The South African data he cited makes the point better than the theory does. Rand trading volumes against dollar stablecoins rose from under 4 billion rand in 2022 to roughly 80 billion rand in the first ten months of 2025 — a twentyfold increase. Meanwhile dollar stablecoins have still gained only limited traction in the domestic economy, and rand-denominated tokens attracted even less demand than that. Building the local alternative did not create local-token users. It is not obvious it ever will.

The number a merchant should read twice

Katz put the stablecoin market at roughly $300 billion in capitalisation — flat over the past year, after nearly tripling between 2021 and 2025. That plateau has been read by sceptics as the end of the story. It isn't the number that matters here.

The number that matters is the composition. About 99% of that $300 billion is dollar-denominated. And citing a Bank for International Settlements study, Katz said that more than 70% of cumulative net fiat inflows into dollar stablecoins came from non-dollar currencies.

Read that again with a merchant's eyes. Seven out of every ten dollars that entered the digital-dollar system came from someone converting out of their own currency. This is not American crypto traders moving between exchanges. It is people in emerging economies — and increasingly businesses in emerging economies — deliberately holding a dollar balance because the alternative loses value or is hard to move.

The rest of the scale he described sets the ceiling. Stablecoin transaction volume exceeded $30 trillion in 2025, of which payment-related flows were around $390 billion — the same figure independent payments research has converged on. Cross-border payments overall move on the order of $1 quadrillion a year. And remittances still cost a global average of about 6.5%, a number that has barely moved in a decade of promises.

So the honest reading of the IMF's own framing is this: digital dollars are already a meaningful store of value in emerging markets, and barely a payment method. The store-of-value side was built by exchanges, wallets and P2P markets. The payment side needs somebody on the receiving end — and that's the part nobody has built, because it can only be built one business at a time.

Where the policy response is aimed — and where it isn't

The University of Cape Town summarised Katz's five recommendations for harnessing stablecoin innovation while protecting monetary and financial stability:

1. Build robust macroeconomic foundations so a country can resist unwanted currency substitution in the first place.
2. Close data gaps through better measurement of digital-asset flows — the IMF is working through the G20 Data Gaps Initiative.
3. Revise policy tools and bring comprehensive regulation to the stablecoin trade-offs.
4. Tailor the response to the adoption channel, particularly where capital controls are concerned.
5. Strengthen international cooperation to prevent regulatory arbitrage and offshore migration.

In the operational detail he was more specific: bring on-ramps, off-ramps and on-chain exchange points inside the regulatory perimeter, and pay attention to the venues where domestic and dollar tokens are freely swapped. On timing, he told central banks plainly: "Data collection should not wait for perfect regulation."

Now notice what is not on that list. Not one of the five points, and none of the operational detail, is about a business accepting payment for goods or services. The regulatory attention is on issuance (who mints the token and what backs it) and on conversion (where fiat becomes crypto and where one token becomes another). A bakery taking USDC for a birthday cake is not a conversion gateway. Neither is a clinic invoicing a patient, an agency billing a foreign client, or a hotel taking a deposit.

That distinction has a practical consequence. The more of your payment flow that runs through custodial intermediaries — an exchange account, a wallet provider that holds your balance, a ramp that converts on your behalf — the more of it sits inside the perimeter the IMF is asking regulators to tighten, and the more counterparties you depend on. A payment that goes from your customer's wallet straight to yours has one fewer moving part than any alternative. That is not a regulatory-avoidance argument — your tax, invoicing and reporting obligations are unchanged, and you should meet them. It is a resilience argument, and we have watched it play out when a custodial provider went down overnight and merchants couldn't reach their money.

The accelerant the IMF flagged: agents

The line that got least coverage may age best. Katz pointed to artificial intelligence as a potential accelerant of stablecoin adoption — and used it as an argument for why data collection can't wait.

This isn't an offhand remark; it follows the Fund's own published work. IMF Note 2026/004, "How Agentic AI Will Reshape Payments" (April 2026, by Sonja Davidovic and Hervé Tourpe), describes the shift from human click-to-pay to agent-mediated decide-to-pay, and separates payments into three layers: intent, authorisation and settlement. Its recommendation is that agentic AI belongs in the upstream intent-and-orchestration layer while authorisation and settlement stay deliberately rule-based and deterministic — payment rails, in the authors' framing, should remain "dumb." The note also observes that as stablecoins spread in cross-border payments, agents transacting over stablecoin rails could become important infrastructure in their own right.

For an API provider or a software business, this stops being macroeconomics and becomes a product decision. An AI agent doesn't have a card, a bank account or a signup flow. It has a wallet. If you want it as a customer, you need a way to be paid per request. That is precisely what x402 does: the agent hits your endpoint, gets an HTTP 402 Payment Required with a price, pays in USDC on Base, and the call goes through — settled directly to the provider's wallet.

Payzum's role there is worth stating precisely, because it's frequently misdescribed. Payzum is the middleware/proxy in front of your existing API — you configure your endpoint, your API key and a price in the dashboard, and Payzum publishes an x402 URL, returns the 402, settles the payment through an external facilitator (currently Coinbase's), then proxies the paid call to your real endpoint with your key. No protocol to implement, no code to write. Payzum is not the facilitator.

What a merchant in an emerging market actually does about this

The IMF's question is whether a central bank can keep monetary control while its citizens hold digital dollars. That's a real question, and it isn't yours. Yours is narrower and entirely actionable: if a customer holds USDC or USDT, can they pay you with it?

Today, for almost every shop, clinic, studio and agency on earth, the answer is no — not because of regulation, but because nobody ever published an address to pay to. That's a configuration gap, and it's cheap to close.

Payzum is a non-custodial, crypto-only payment processor. Money moves from your customer's wallet to a wallet you control. Payzum never holds, pools or routes it — the settlement is the payment, and there is no Payzum balance to freeze, delay or reconcile.

Across the counter, you run POS with a new QR code per sale: any phone or tablet becomes a terminal, cashiers get PIN-protected logins, and you get analytics per cashier and per terminal. No acquirer, no card-network fees, and no chargebacks to defend four months later.

Online, you can take payment through hosted checkout (redirect, modal or inline), no-code payment links and buttons, invoices with expiration and overpayment detection, donation or tip-jar pages, recurring subscriptions, or a drop-in plugin next to your existing stack. A REST API with signed webhooks is there when you want it wired into your own systems.

On settlement, you accept crypto and receive crypto, with optional auto-conversion to USDC or USDT so a day's takings keep their unit of account. Payments run across nine networks — Bitcoin, Ethereum, Solana, Polygon, Base, Arbitrum, Optimism, BNB Chain and Avalanche — with typical confirmations around 0.4 seconds on Solana and about 2 seconds on Base and Polygon.

And when you need to pay out — suppliers, contractors, affiliates, prize pools — there are mass payouts by CSV and EVM stablecoin payouts from the same wallet you were paid into.

How you'd set it up, step by step

1. Create the account and point it at your wallet. You supply the destination address — yours, on the chains you want to accept. Because it's non-custodial, that's where the first payment lands and every one after it. There is no intermediate balance.
2. Turn on auto-convert if you want your revenue in dollars. Accept whatever the customer holds and settle to USDC or USDT, so a week of takings doesn't move overnight with an asset price.
3. Open the door your business actually uses. Physical counter: install the POS app on any phone, create PIN logins for staff, and start generating a fresh QR per sale. Online: drop in hosted checkout, generate payment links, send invoices with expiry, or connect via the REST API and signed webhooks. API provider: configure your endpoint and price and let agents pay per call over x402.
4. Test with one real payment, then leave it open. Send yourself a small USDC payment and watch it confirm in seconds into your own wallet. Then keep your existing rails exactly as they are — this is an additional door, not a migration, and it costs nothing to keep open while adoption catches up.

Three businesses this is already true for

The IMF was describing behaviour that has already happened. These are the shapes it takes in a normal business:

• An agency, studio or freelancer billing foreign clients. The client is abroad, the invoice is in dollars, and the wire takes days and loses a slice to correspondent fees and FX spread. A payment link settles in seconds into your own wallet — the exact scenario in getting paid from abroad without a bank account and in cross-border acceptance.
• A local shop, clinic or restaurant in a high-inflation economy. Customers who already hold USDT to protect savings currently have to sell it to spend it, taking the spread twice. A QR at the register lets them pay you from the balance they already hold, and you keep the full ticket instead of handing 2–4% to an acquirer — the same dynamic behind Argentina's peso-stablecoin push.
• A money-transfer, travel or import business moving value across a border. Against a 6.5% average remittance cost, settlement that clears in seconds for network fees is not a marginal improvement — see stablecoin remittances for money transfer businesses.

Same digital dollar, three ways it can reach you

Fair objections, answered

"The IMF is warning about this. Isn't accepting stablecoins the wrong side of the trade?"

The IMF's concern is monetary and macroprudential: currency substitution, capital-flow visibility, offshore migration of FX activity. Those are questions for finance ministries and central banks, and Katz addressed them to central bankers. None of the five recommendations proposes restricting a business from being paid for goods or services. Read the operational detail and the target is consistently issuance and conversion venues. A merchant accepting payment isn't either of those.

"My country hasn't finished regulating this. Shouldn't I wait?"

Rules will keep changing, and you should follow the ones that apply to you — including tax, invoicing and any local reporting. But note what Katz told the central bankers themselves: don't wait for perfect regulation to start measuring. The same logic applies commercially. Opening a stablecoin door is a configuration, not a construction project: no acquirer contract, no hardware to buy, no monthly minimum, no revenue share committed. If it turns out you don't need it, you've lost an afternoon.

"If my central bank launches a local stablecoin, doesn't this all change?"

That's the exact scenario Katz was analysing, and his answer was that it may increase dollar-token usage rather than replace it. In South Africa, dollar stablecoins gained limited traction and rand-denominated tokens gained even less. A merchant who can accept an on-chain payment is indifferent to which token wins: you accept what the customer holds and auto-convert to the stablecoin you want to keep.

"I don't want to hold a volatile asset."

Then don't. Stablecoins are dollar-denominated by design, and optional auto-conversion settles what you accept into USDC or USDT so your revenue keeps its unit of account. The volatility objection is about accepting Bitcoin at a café; it doesn't apply to a digital dollar that stays a dollar. Payzum is crypto-only — it accepts and settles in crypto, and does not settle to a fiat bank account.