The bank consortium stablecoin: 21 of the world's largest banks committed to a dollar — and not one sentence to merchant acceptance

What the 21 banks actually committed to on September 1

On September 1, 2026, twenty-one banks and asset managers announced a commitment to form a jointly owned company in the second half of 2026 whose purpose is to issue a US dollar stablecoin. The formation is subject to closing conditions — the standard language for "this is an intention, not an incorporation". The token itself is targeted for the first half of 2027, with a euro-denominated version named as the next priority after the dollar, and other G7 currencies flagged as later ambitions.

The participant list, per reporting on the announcement, spans four regions:

• North America (10): Bank of America, Capital One, Citi, Fidelity Investments, Goldman Sachs, PNC Financial Services, Scotiabank, TD Bank Group, Wells Fargo, WisdomTree.
• Europe (8): Banco Santander, BBVA, Commerzbank, Crédit Agricole, Deutsche Bank, Lloyds Banking Group, Rabobank, UBS.
• East Asia (1): MUFG Bank.
• Middle East & Africa (2): Sirius International Holding, Standard Bank.

The stated use cases are cross-border payments, settlement of digital-asset transactions, wholesale transfers between banks and large corporates, institutional settlement for tokenized securities, and retail payments. The group said it intends to be compliant with the US GENIUS Act — 1:1 reserve backing, and no yield paid to holders simply for holding — and with the EU's MiCA regime where applicable.

This is not a standing start. An initial group of ten banks disclosed in October 2025 that they were exploring a 1:1 reserve-backed digital payment asset issuable on public blockchains. That group has now more than doubled. Notably absent: JPMorgan, which has reportedly been weighing something of its own, separate from JPM Coin.

The market read it as competitive pressure on incumbent issuers — Circle's stock fell roughly 6% on the news. That reaction is about who captures the float. It is not about whether any business will be able to get paid in the thing.

Read the blanks, not the names

Announcements like this are usually judged by the logos. The more informative exercise is to list what twenty-one institutions with a combined balance sheet in the tens of trillions did not decide before going public.

As of the September 1 release, there is no company name. No token name. No decision on which blockchain networks the token will live on. No named reserve custodian. No published governance structure. No final redemption terms — meaning the answer to "how, exactly, do I turn one of these back into a bank dollar, how fast, and at what cost" does not exist yet.

None of that is scandalous. Twenty-one competitors agreeing to co-own an issuer is genuinely hard, and the announcement is a coordination device: it locks the participants in publicly so the negotiation can proceed. But it does set the epistemic status of "H1 2027". That is not a ship date. It is a target for a company that does not exist yet, to issue a token that does not have a name, on a chain that has not been chosen, redeemable on terms that have not been written.

A business deciding how to get paid in the next four quarters should file this accordingly.

Issuance is now crowded. Acceptance is still empty.

Here is the structural point, and it is the whole reason this announcement matters to a merchant rather than to a trading desk.

Count the organised efforts to issue a regulated dollar or euro on a public blockchain as of September 2026:

• This 21-bank consortium — USD, H1 2027, euro to follow.
• Open USD — 140+ companies, with Visa and Mastercard backing.
• Qivalis — 37 European financial institutions building a regulated euro stablecoin.
• Japan's three megabanks — a joint yen stablecoin targeting March 2027.
• US state banking associations — a shared bank blockchain, also with a 2027 date.
• And the incumbents already live: USDC, USDT, PYUSD, RLUSD and a long tail behind them.

Now count the organised, funded, multi-institution efforts to solve acceptance — the problem of a business being able to take one of these at a checkout, price in it, reconcile it, refund it and keep it. There is no consortium for that. There is no twenty-one-logo press release. There are payment companies, and there is a market.

The asymmetry shows up in the data. Research from Flagship Advisory Partners puts direct stablecoin acceptance at roughly 4% of top-50 merchants in the US, about 8% of large merchants in Europe and 12% in Latin America. Retail stablecoin payment volume was around $70 billion globally in 2025, growing 83% year on year — real growth, and still a rounding error against the $50 trillion-plus the card networks clear annually.

Twenty-one of the largest financial institutions on earth spent September 1 announcing more supply into a market whose binding constraint is on the other side of the counter.

Why banks are building a stablecoin they can't pay interest on

It's worth understanding the incentive, because it explains what the product will and won't be optimised for.

Under the GENIUS Act framework — now working through Treasury rulemaking — a compliant payment stablecoin is fully reserved and cannot pay a yield to holders purely for holding it. So the consortium is not building a deposit substitute that competes on interest. What it is building is a settlement asset the participating banks co-own.

The logic is defensive and fairly clear. If tokenized dollars are going to carry an increasing share of cross-border corporate payments, securities settlement and treasury movement, then every dollar that migrates to USDC or USDT is a dollar that leaves the banking system's own rails and sits with a non-bank issuer. Twenty-one banks co-owning an issuer keeps the settlement relationship — and the reserve float — inside the club.

That is a legitimate business objective. It is also, importantly, an objective about where money sits and how institutions move it between each other. Read the use-case list again in that light: cross-border payments, digital-asset settlement, wholesale bank-to-bank, tokenized securities, and then retail. The first four are interbank and corporate treasury. Retail is last on the list, and "retail payments" in a bank press release generally means a consumer can hold and send it, not a hardware store can price a sale in it and keep the proceeds.

Nothing about this venture is designed to make it cheaper or safer for a business to get paid. That is not a criticism. It is a scope observation, and scope determines what a merchant should expect from it.

The four questions a merchant needs answered — and this announcement answers none of them

Strip away the logos and a business accepting payment has exactly four questions about any new dollar token. Here they are, against what was published on September 1.

1. Who will accept it, and where? A payment instrument with no acceptance network is a wire transfer with extra steps. The announcement contains no distribution plan, no merchant programme, no acquirer partnerships and no consumer wallet strategy. It is not clear whether ordinary people will hold the token directly or only access it through their bank.
2. What will it cost to accept? Unknown, because there is no acceptance layer to price. If it eventually reaches merchants through the same acquiring relationships that price card acceptance today, the merchant discount rate is set by the same parties that set it now. A new settlement asset behind an unchanged acceptance layer does not change the number on your statement.
3. Is a payment final? Unanswered, and it is the question that matters most. On-chain settlement is final by construction. But if a consumer pays a merchant with a bank-issued token through a bank-issued interface, the dispute rights attached to that interface are a policy decision, not a property of the chain. Reversibility is the single largest hidden cost of card acceptance, and nothing in this announcement retires it.
4. Where does the money land? Also unanswered. If proceeds accrue to a balance held by an intermediary and then settle to you on their calendar, you have counterparty exposure and a settlement delay, on a blockchain. The token being on a public chain does not by itself mean the money is yours the moment it moves.

Four questions, zero answers, with a target date at least nine months out that is contingent on a company being formed first. That is the honest state of play.

The fragmentation problem lands on the merchant

There is a second-order consequence that gets almost no coverage, and it is the one that will actually reach a shop counter.

Every new consortium dollar adds an entry to a list that a merchant is eventually asked to support. In 2023 the practical question was "do you take USDT or USDC?". By 2027, on current trajectory, the same conversation involves a bank consortium dollar, an Open USD, a euro from Qivalis, a yen from the Japanese megabanks, PYUSD, RLUSD, and whatever else ships in the interim.

A business does not want to make that choice. A business wants to take what the customer is holding, and end up with a dollar it controls. Those are two different jobs, and the merchant only cares about the second one.

This is precisely why the acceptance layer has to be token-agnostic and chain-agnostic, and why the interesting infrastructure question of 2027 is not "which dollar wins" but "what sits between the customer's wallet and the merchant's". Whoever wins the issuance race, the merchant's requirement is unchanged: take whatever arrives, settle it into an asset I recognise, put it somewhere nobody else can reach, and don't let anyone pull it back later.

The one thing this announcement genuinely settles

It would be dishonest to treat September 1 as noise. It settles something real.

When Bank of America, Citi, Goldman Sachs, Wells Fargo, UBS, Deutsche Bank, Santander, BBVA, Crédit Agricole, Commerzbank, Lloyds, Rabobank, MUFG, TD, Scotiabank, PNC, Capital One, Fidelity, WisdomTree and Standard Bank jointly commit capital and reputation to issuing dollars on public blockchains, the "is this real?" conversation is over. Not because banks are always right, but because this is the most conservative cohort in finance choosing the same settlement substrate that a merchant taking USDC on Base is already using today.

For a business that has been putting off the decision because it felt early, the signal is straightforward: the incumbents have concluded that dollars on public chains are how a meaningful share of money will move. They have set themselves a 2027 date to participate. You do not have a 2027 date. You have a checkout, and the rail is already there.

How Payzum fits — the acceptance side, available now

Payzum is a non-custodial, crypto-only payment processor. The mechanic that matters against everything above is one sentence: funds go directly to wallets the merchant controls. There is no Payzum balance, no pooled account, no settlement calendar. The settlement is the payment.

Mapped against the four unanswered questions:

• Acceptance exists today. Hosted checkout (redirect, modal or inline), no-code payment links and buttons, invoices with expiry and overpayment detection, subscriptions, donations, a drop-in plugin for your existing store, and a POS that issues a fresh QR per sale and turns any phone into a terminal, with PIN-protected cashier accounts.
• Cost is a network fee, not a percentage of the ticket. Confirmation is roughly 0.4 seconds on Solana and about 2 seconds on Base and Polygon, at fees measured in cents regardless of amount. No acquirer, no card-network fees. // confirmar pricing actual
• Finality is structural. On-chain settlement means no chargebacks. Nobody reverses a shipped order 90 days later.
• The money is yours on arrival. Non-custodial settlement to your own wallet, with optional auto-conversion to USDC or USDT so a dollar-priced sale stays a dollar. Supported networks: Bitcoin, Ethereum, Solana, Polygon, Base, Arbitrum, Optimism, BNB Chain and Avalanche.

And the money-out side is the mirror image: mass payouts by CSV for BTC, LTC and DOGE, plus EVM stablecoin payouts on Polygon, Arbitrum, Optimism, Base, BNB and Avalanche.

How it works, step by step

1. Connect wallets you control. You provide the destination addresses. Payzum never holds the funds, so there is no balance for anyone — a consortium, an acquirer or a bank — to hold, freeze or reconcile.
2. Choose how you get paid. Online: hosted checkout, payment links, the drop-in plugin, invoices, subscriptions. In person: the POS, with a new QR per sale and per-cashier analytics. B2B: invoices that carry your own reference and detect overpayment.
3. Pick network and settlement currency. Nine chains supported. Turn on auto-convert to USDC or USDT if you want every sale to land as a dollar, whatever the customer paid with.
4. Reconcile automatically. Signed webhooks fire on payment, a REST API with API keys exposes the same events, and a full audit log plus per-cashier and per-terminal analytics tells you which sale settled and when. If you also sell an API, x402 lets AI agents pay USDC on Base per call, straight to your wallet.

What this looks like for three real businesses

Three shapes of business for whom H1 2027 is not a plan they can run on:

• The importer whose supplier already asked. A distributor in Medellín buys from a manufacturer in Shenzhen who now quotes in USDT. Waiting for a bank consortium token means waiting at least four quarters for an instrument whose redemption terms are unwritten. Paying and getting paid in USDT today means the invoice settles in seconds, at a network fee, with the transaction hash as the receipt — and the counterparty relationship keeps working.
• The online store bleeding on disputes. A DTC brand with a meaningful share of foreign orders pays cross-border interchange, an FX markup, and a dispute rate concentrated in exactly those orders. Nothing in the September 1 announcement addresses reversibility. Adding a stablecoin checkout alongside cards moves part of the basket onto a rail where settlement is final on arrival.
• The agency paying people in six countries. A studio bills clients in dollars and pays contractors in Buenos Aires, Lagos and Manila. The consortium's cross-border ambitions are aimed at bank-to-bank flows, not at a twelve-person contractor payroll. A single stablecoin payout batch replaces six remittance corridors, each with its own fee, delay and compliance queue.

A bank consortium stablecoin vs a non-custodial acceptance rail

Comparing like with like: what the announced venture is scoped to deliver, against what a merchant can switch on this week.

Fair objections

"Twenty-one global banks beats a startup. Shouldn't I just wait?"

Wait for what, specifically? The venture is scoped to interbank settlement, cross-border corporate flows and securities. Even in the best case where it ships on time and reaches consumers, it is a new dollar — it is not a checkout, not a POS, not an invoice, not a payout batch, and not an answer to chargebacks. Those are different products. If the consortium token becomes widely held, businesses will still need something that accepts it and settles it somewhere safe. Building that muscle now costs you nothing and works today.

"Doesn't a bank-issued token make USDC and USDT less relevant?"

Possibly, over years — that is what the market reaction was pricing. But this is an argument about issuer market share, and a merchant's exposure to it is small if the acceptance layer is token-agnostic. Auto-conversion to USDC or USDT is a setting, not an architecture. If the landscape shifts, the thing that has to adapt is the processor, not your checkout.

"We'd be taking on crypto volatility."

Only if you choose to. Optional auto-conversion to USDC or USDT means a customer pays with whatever they hold and you settle in a dollar-denominated stablecoin. What you adopt is the settlement mechanism — seconds, cents, final, non-custodial — not a position in a volatile asset.

"Isn't self-custody a burden we don't want?"

It's a trade you're already making in the other direction. Today an acquirer holds your money for one to three days, can impose a rolling reserve, and can close the account. Non-custodial settlement moves that responsibility to you: you hold keys, you set 2FA, you read the audit log. Many businesses run a dedicated operating wallet and sweep on a schedule. It is a real operational change, and it is the one that removes the party who can freeze you.