Stablecoin payments in Japan: what merchants actually get for 0.98%

What happened in Japan in August 2026

Three weeks of Japanese retail news did something the rest of the market has mostly avoided: it turned stablecoin acceptance from a demonstration into a priced product with a merchant on the other side of the invoice.

The sequence matters. On August 6, Lawson ran the trial we covered here — a yen stablecoin, JPYC, through the ordinary register at its Takanawa Gateway City store, with a non-custodial wallet on the customer's side. That was an engineering result: the coin fit inside a checkout ritual measured in seconds.

What followed is a commercial result. On August 17, a different operator ran a proof-of-concept at a different Lawson store — Gate City Osaki Atrium, after a preliminary run on August 3 — this time accepting USDC, USDT and JPYC. Per crypto.news, the customer displays a payment barcode from a wallet app — MetaMask, in this run — and the store's existing POS register scans it, with no dedicated crypto terminal on the counter. The supported networks were Solana, Morph and Polygon for the dollar-denominated coins and Polygon for the yen one. Participation was restricted to related personnel; this is not open to the public, and Lawson's more than 14,000 stores are not switching on next quarter.

The part that deserves more attention than it got: the service being trialled is already for sale. Cointelegraph reported in mid-July 2026 that the operator had launched it commercially and opened applications from merchants wanting to accept multiple stablecoins — most of them able to use their existing payment terminals — with the fee set at 0.98%. Prior deployments included two shops at Haneda Airport between January and February 2026 and a trading-card store in Himeji from April.

And then the clause that decides everything, quoted in substance: merchants handle pricing, sales records and settlement in yen even when the customer pays with a dollar-denominated stablecoin, removing the need to hold crypto or manage exchange rates.

The number that matters isn't the chain list — it's 0.98%

Every stablecoin retail story of the last two years has been reported as a technology story: which coin, which chain, how many seconds. Those are the least interesting variables. A merchant deciding whether to switch on a payment method asks three questions, in this order: what does it cost me, when do I have the money, and who can take it back?

Japan has now answered the first one out loud. 0.98%. That is a real, published, merchant-facing price for stablecoin acceptance — and it is a genuinely good number in a country with strong domestic card and QR competition. It is also, and this is the uncomfortable part, approximately the number a decent domestic card deal already produces in a developed market. If your existing blended rate is 1.2% and your new rate is 0.98%, you have bought a modest discount, not a different category of payment.

Which is fine! A modest discount at national-chain volume is worth having. But it clarifies what is being sold. Compare the same decision in the markets where card acceptance actually hurts: a Latin American merchant paying 3.5% plus a cross-border surcharge plus an FX spread plus 30 days to settle is not choosing between 1.2% and 0.98%. They are choosing between a rail that costs them a tenth of their margin and one that costs cents in network fees.

So the honest read of Japan's number is: stablecoin acceptance has been productised, and where card rails are already efficient, the productised version competes on price at the margin. The interesting economics live elsewhere — and so does the interesting question, which is about the settlement clause rather than the fee.

Three ways to "accept stablecoins" — and only one puts the token in your wallet

Over the past month this blog has now looked at three structurally different things, all marketed with the same phrase. It is worth laying them side by side, because a merchant evaluating a proposal needs to know which one is on the table.

The three are not ranked by virtue. They are ranked by what changes hands. In the first two, the stablecoin is a piece of upstream plumbing that improves someone else's cost of funds; the merchant receives the same instrument they always received — a claim on an intermediary, denominated in local currency, payable later. In the third, the merchant receives the money.

What 0.98% buys, honestly

It would be easy — and wrong — to dismiss the converted model. It delivers three real things, and pretending otherwise would be marketing rather than analysis.

It is genuinely cheaper than cross-border card acceptance, and it removes the FX question from the shopkeeper's day. A Tokyo store that takes a USDC payment from a visiting tourist gets yen at a known rate without operating a treasury function. That is not nothing; it is most of what a small merchant actually wants.

It removes the chargeback, or at least the card-scheme version of it. The customer's payment leg is an on-chain transfer, which no scheme can reverse 120 days later on the cardholder's say-so. Whatever the merchant's contract with the intermediary says about disputes, the structural mechanism that produces friendly fraud on card rails simply is not present.

It requires almost nothing of the merchant, which is why it will win a lot of shops. Existing terminals in most cases. Prices in yen. Books in yen. No wallet to secure, no key to lose, no new category on the tax return.

Now the other column. The convenience is purchased with a re-introduced middleman, and three specific properties come back with him.

The finality stops before it reaches you. The customer's transfer is final on-chain, but that finality belongs to the intermediary's wallet, not yours. Your leg is an ordinary payable, with an ordinary settlement lag, that clears when their operations clear. The single most valuable property of the technology — the settlement is the payment — is consumed one step upstream.

Counterparty risk returns in full. Between the sale and your bank credit there is a balance sheet that is not yours. We wrote about this at length in the counterparty-risk piece and watched it play out concretely when a crypto payment provider shut down overnight. Everything the industry says about not trusting custodians applies to an intermediary that holds your Saturday takings until Tuesday, regardless of how well regulated it is.

And you don't get the dollar. This is the quiet one. If you are in Tokyo and your costs are in yen, being paid in yen is exactly right. If you are in Buenos Aires, Caracas, Lagos or Istanbul, the reason your customer is holding USDT is the same reason you want to keep it — and a service whose core promise is "we'll convert it to local currency for you" has solved a problem you don't have while removing the benefit you were after.

Why Japan's version has this shape

None of this is a criticism of the operators involved. Japan's model looks the way it looks because of its regulatory perimeter, not because anyone chose the more custodial design for fun.

Japan regulates fiat-referenced stablecoins as a distinct category under its Payment Services Act, with a defined set of permitted issuer types and a registration regime for the intermediaries that buy, sell, exchange, transfer or hold them on someone else's behalf — the framework and the current rulebook are published by the Financial Services Agency. In a market organised that way, the natural commercial product is exactly what shipped: a registered intermediary standing between the token and the shop, absorbing the regulated activity so that the shop performs none of it.

That is a coherent, defensible design for Japan. The mistake would be reading it as the design. The merchant-side question is not "what does Japanese law require of a Japanese intermediary" but "in my jurisdiction, what do I actually receive, and from whom?" A merchant who is paid directly by their customer, in an asset they choose to hold, to an address they control is in a different position — commercially and structurally — from one who is paid by an intermediary, later, in local currency, at a percentage.

How Payzum fits — and what it deliberately doesn't do

Start with the honest limitation, because it defines the choice. Payzum is crypto-only. It does not settle to a bank account and does not deliver yen, euros or pesos. If what you want is local currency in a local bank, the converted model is the right shape and this is the wrong tool.

What Payzum does is the third row of that table. It is a non-custodial crypto payment processor: the customer's payment goes directly to a wallet address the merchant controls. Payzum never holds, pools or controls the money. There is no Payzum balance, no settlement schedule, no payout queue and no intermediary between the sale and the funds. The settlement is the payment.

Mapped onto the specific things the Japanese trials put on the table:

• The register problem. The Lawson runs solved it by integrating with the store's POS — an achievement that requires a chain, a telecom operator and a processor. Payzum routes around it: any phone becomes a terminal, showing a fresh QR per sale. Nothing has to talk to your existing register, so there is no vendor roadmap to wait for.
• Multi-coin, multi-chain, same as the trial. USDC and USDT across Solana, Base, Polygon, Arbitrum, Optimism, BNB Chain and Avalanche, plus Bitcoin and Ethereum. Typical confirmation times are around 0.4 seconds on Solana and about 2 seconds on Base and Polygon — the same order of magnitude the Tokyo trials were measuring at the counter.
• Volatility, without giving up the asset. If a customer pays in something other than a stablecoin, optional auto-conversion to USDC or USDT handles the price risk. The result is still a dollar-denominated asset in your wallet — not a local-currency claim on a company.
• Staff and accountability. PIN-protected cashiers and per-cashier and per-terminal analytics, so a multi-till shop reconciles a shift the way it always did.
• Fees. No acquirer and no card-network fees. What remains is the network fee, measured in cents on Base, Polygon or Solana rather than as a percentage of the basket.
• Chargebacks. There are none, because there is no scheme to file one with. The customer's transfer is final, and that finality lands with you rather than one step upstream.

Online, the same principle applies through hosted checkout, no-code payment links and buttons, invoices with expiry and overpayment detection, recurring subscriptions and a drop-in plugin — plus a REST API with signed webhooks if your own systems need to know the second a payment confirms.

How it works, step by step

1. Connect the wallet you already control. You provide the receiving addresses — your own, on the chains you want to accept. Payzum never takes custody of them, and there is no balance held on your behalf.
2. Choose your assets and networks. USDC and USDT are the usual answer for a shop that wants dollar stability; add Bitcoin or Ethereum if your customers ask for them, with optional auto-convert to a stablecoin so the price you quoted is the value you keep.
3. Turn a phone into the terminal. Open Payzum POS on any phone or tablet at the counter, key the amount, and it shows a fresh QR for that single sale. Add cashiers with their own PINs so each shift is attributable.
4. The customer pays from their own wallet. They scan and confirm. Confirmation comes back in roughly a second or two on Solana, Base or Polygon, and the funds are at your address — not queued at a processor.
5. Reconcile, and wire the rest of your stack. Per-cashier and per-terminal analytics for the end-of-shift count; signed webhooks so your POS, inventory or accounting system marks the sale paid automatically; a full audit log, 2FA and encrypted secrets on the account itself.

Where this actually matters for a real business

The Japanese trials are useful precisely because they prove customer demand exists at an ordinary till. Here is where the direct version of that demand is worth more than the converted one.

• A shop in a tourist district. The visiting customer's card is foreign — cross-border interchange, an FX spread and a non-trivial decline rate on a card issued three time zones away. A QR at the counter takes USDC from their wallet at the same speed a domestic QR scheme would, and the sale that used to fail at the terminal simply completes.
• A business whose customers already hold dollars. In Argentina, Venezuela, Nigeria or Turkey, USDT is not a payment novelty — it is how people keep savings. A merchant who prices in dollars and gets paid in dollars, into their own wallet, has removed the conversion round trip that the converted model would reinstate.
• A multi-till operation that needs accountability, not integration. A restaurant, a clinic or a retail chain with four registers doesn't need its POS vendor to ship a stablecoin module. It needs a QR per sale, a PIN per cashier and a report per shift — which is a phone, today, on any counter.
• An online seller with a chargeback problem. Hosted checkout or a payment link produces a final payment. For digital goods, courses, bookings and subscriptions, that removes the single largest source of revenue leakage on card rails.

Converted acceptance vs non-custodial acceptance

Objections worth taking seriously

"Being settled in my local currency is exactly what I want. Isn't the converted model just better?"

For a lot of merchants, yes — and we would rather say so than pretend otherwise. If your costs, payroll and taxes are all in a stable local currency and your bank relationship is healthy, converting at the point of sale is a clean answer, and 0.98% is a fair price for it. The decision flips when one of three things is true: your customers hold dollars because your local currency is losing value; your card rates are three to four times Japan's; or you have already been de-risked by a bank or an acquirer and the problem is access, not price.

"Doesn't holding the stablecoin myself just move my risk to the issuer?"

It moves some of it, and that is a fair trade to state plainly. Holding USDC or USDT is an exposure to that issuer's reserves and redemption process. What it removes is the layer nobody talks about: the operating company that holds your takings between the sale and the payout, which has its own solvency, its own banking relationships and its own ability to freeze an account. One exposure you can research and choose; the other you inherit by default. And you can hold it for minutes rather than months — nothing about non-custodial settlement obliges you to keep a balance.

"My customers don't pay in stablecoins, so this is theoretical."

That is the honest state of most counters today, and the Japanese trials do not refute it — they were closed to the public. What they do establish is that the acceptance side is now solved well enough to be sold as a product with a price list, at a national chain, on three chains at once. The demand side has its own evidence, and it is growing faster than the acceptance side: consumers are already spending stablecoins through cards funded by them. Switching on a QR costs you nothing when nobody uses it, and captures the sale the day someone does.

"Doesn't taking crypto directly create a compliance problem for me?"

It creates obligations you need to look at with your own advisers, and they differ sharply by country — Japan's regime, for instance, deliberately routes this activity through registered intermediaries. What non-custodial settlement gives you on that front is a better record, not a workaround: an on-chain payment is a timestamped, independently verifiable transfer of an exact amount to an address you control, matched to an invoice, alongside a full audit log. That is more evidence than cash produces, not less. It is not legal advice, and it is not a substitute for checking your local rules.