Policy · Stablecoins
Twenty-One Banks Have a Stablecoin Plan, a Launch Window, and No Name
Bank of America, Citi, Goldman Sachs, Deutsche Bank, UBS and sixteen others say they will bring a dollar token to market in the first half of 2027. The company does not exist yet. The chain has not been chosen. The only parties named in the announcement are the two consultancies, and both are noted as unable to bind anyone.
On September 1, twenty-one of the largest financial institutions in the world announced they intend to issue a dollar stablecoin. The list is not a fringe one. Bank of America, Citi, Goldman Sachs, Wells Fargo, Capital One, PNC, TD, Scotiabank, Fidelity and WisdomTree from North America. Santander, BBVA, Commerzbank, Crédit Agricole, Deutsche Bank, Lloyds, Rabobank and UBS from Europe. MUFG Bank in East Asia, Sirius International Holding in the Middle East, Standard Bank in Africa.
That roster is the announcement. Almost everything else in it is a placeholder.
Per Banking Dive, the group will form a new company in the second half of 2026, with a name that "will be announced in due course." The token is targeted at the first half of 2027, for wholesale, institutional and retail use including cross-border payments and digital asset settlement. Other G7 currencies may follow, starting with the euro. The venture says it will comply with the US GENIUS Act and, where applicable, the EU's MiCA.
What has not been settled, per Blockhead, is the company name, the reserve composition, the technical architecture and the distribution partners. No individual executive is quoted. The two organisations named in the announcement are Boston Consulting Group and Brunswick Group, as advisors — and the announcement notes they have no authority to bind the members.
Where the twenty-one come from
The geographic spread is the part that is genuinely hard to assemble, and it is worth looking at as a distribution map rather than a guest list.
The consortium by region
Eighteen of the twenty-one are North American or European. The three institutions covering East Asia, the Middle East and Africa are one apiece. For a product whose stated purpose includes cross-border payments, the corridors that are hardest to serve are the ones represented by a single member each.
The calendar is the argument
The launch window is the most informative number in the announcement, because of what it sits next to. Banking Dive gives the GENIUS Act's expected effective date as January 18, 2027. The consortium's product is due in the first half of 2027.
That is not a coincidence and nobody is pretending it is. The venture is timed to arrive as the American rulebook switches on, which is the sequencing a bank does and a startup usually cannot afford. It also means the group has roughly sixteen months to incorporate a company, choose a chain, settle a reserve model and build distribution — while the rules it is building against are still being written by regulators rather than read off a page.
The GENIUS Act's substance explains the shape of the thing. Per Brookings, the Act requires "at least a 1-to-1 reserve backing," admits permissible reserve assets beyond cash and Treasury bills including uninsured bank deposits and repo borrowing, and — the decisive provision — "explicitly prohibits payment of interest" by payment stablecoin issuers. Brookings reads the Act as clearly intending payment stablecoins "to serve as a medium of exchange."
Strip out yield and a stablecoin cannot compete on returns. What is left to compete on is trust, compliance, and the number of counterparties who already have an account with you. That is a contest designed for incumbents, and twenty-one incumbents have now entered it together rather than separately.
What has actually progressed
This group did not appear last week. Blockhead notes an initial ten-bank version announced in October 2025. Eleven months later the membership has slightly more than doubled, and the deliverables — a company, a name, a chain, a reserve model — remain where they were.
That is not automatically damning. Assembling twenty-one competitors across five regions into a single issuing entity is a legal and antitrust exercise before it is a technical one, and the participants have every reason to sign the roster before they argue about ownership stakes. But it does mean the honest description of the September announcement is that the consortium got larger, not that the product got closer.
The unchosen blockchain deserves particular attention. The token is meant for "digital asset settlement," which is a use case defined entirely by which ledger the assets settle on. A public chain has to be picked, and that choice determines finality guarantees, censorship exposure, fee volatility and which of the members' existing custody arrangements survive contact with it. It is the load-bearing decision in the entire venture, and as of this week it has not been made.
None of the twenty-one has published a technical specification. This desk found no named chief executive, chief technology officer or board for the venture, because the venture does not yet exist as a company.
The rulebook is still a draft
The consortium's timing argument rests on the GENIUS Act switching on. Since this story first ran, the shape of that switch has become clearer, and it is less settled than a single effective date implies.
On August 18 the Treasury Department published a notice of proposed rulemaking under docket TREAS-DO-2026-0496, implementing section 3 of the Act — the part governing, in Treasury's words, "the statutory prohibitions and limitations on payment stablecoin issuance, offer, and sale in the United States." It is a proposed rule, not a final one. Comments close on October 19, 2026.
That date sits inside the window in which the consortium says its company will be formed. The venture is due to incorporate in the second half of 2026; the public comment period on the rule that governs who may issue a payment stablecoin in the United States does not close until nearly the end of it. The members will be standing up a legal entity while the rules defining its core permission are still open for objection, including objections from the members themselves.
The effective date is conditional too. OCC Bulletin 2026-3 states the Act takes effect on "the earlier of 18 months after the enactment date (July 18, 2025) or 120 days after the primary Federal payment stablecoin regulators issue final regulations." Eighteen months from enactment is January 18, 2027. The other branch cannot now arrive first: final regulations cannot issue before comments close on October 19, and 120 days from that date falls in mid-February. So the January 18 date reported at the time of the announcement is, as of this month, the operative one by default rather than by design — the regulators ran out of runway to make it earlier.
Two calendars, one window
| Date | Regulator | Consortium |
|---|---|---|
| Jul 18, 2025 | GENIUS Act enacted | — |
| Oct 2025 | — | Initial ten-bank group announced |
| Aug 18, 2026 | Treasury publishes proposed rule on issuance | — |
| Sep 1, 2026 | — | Twenty-one members announced |
| Oct 19, 2026 | Comment period closes | — |
| H2 2026 | — | Company to be formed |
| Jan 18, 2027 | Act takes effect | — |
| H1 2027 | — | Token targeted |
None of this makes the launch window implausible. Banks routinely build against proposed rules, and the section 3 proposal is an implementation of statutory text the members have been reading since 2025 rather than a surprise. But it does sharpen what the consortium is actually betting on. The roster was assembled against a rulebook that exists in draft, and the one decision the group has publicly deferred — which chain — is the one no regulator is going to make for it.
The Take
Take the announcement at face value and it is still significant: getting Goldman, Deutsche Bank, UBS and MUFG to agree on anything with a launch date attached is real work, and the GENIUS timing shows people who have read the statute. But it should be reported as what it is, which is a letter of intent with a very good letterhead. The tell is not the missing name — names are easy and come last. The tell is that the chain is unchosen while the use case is settlement, because that is the decision everything else depends on, and a committee of twenty-one competitors is the worst possible instrument for making it. Every member has a custody stack, a compliance posture and a set of existing partners that argues for a different answer. Somebody will lose that argument, and until we see who, "first half of 2027" is a hope rather than a schedule. Meanwhile the interest prohibition tells you exactly what this is for. A token that cannot pay yield is not a savings product competing with Tether and Circle for reserves; it is plumbing, aimed at the settlement traffic these banks already move for each other. That is a smaller and more plausible ambition than the roster implies — and the roster, one suspects, is the point of announcing before the company exists.