U.S. Bank mints its own dollar on a public chain, with a clawback

What happened

U.S. Bank said on 9 September 2026 that it had completed a live pilot payment in USBDC, its own dollar-backed token, between U.S. Bank entities in North America and Europe. The release is datelined Minneapolis and carries the sub-heading 'Launch demonstrates the bank's ability to move funds on-chain while leveraging traditional banking system controls and risk management'. It says the transaction 'was completed on the Stellar blockchain, demonstrating the ability for U.S. Bank to transfer value on-chain while maintaining integration with the bank's core finance, risk, compliance and operations infrastructure', and that 'As one of the first bank-issued stablecoins deployed on a public blockchain, USBDC promises the potential of leveraging a new technology to bridge gaps in the global banking system with 24/7 transaction capabilities'. The list of what was exercised is the substance of the announcement: 'The pilot, which evaluates minting, payment redemption, freezing and clawback capabilities, also validates U.S. Bank's internally developed Digital Asset Platform, which serves as the foundation for issuing, managing and moving tokenized assets and enables seamless interaction between traditional banking infrastructure and blockchain networks.' Gunjan Kedia, chairman and chief executive officer at U.S. Bank, said 'This live pilot demonstrates our ability to accelerate global cash management and money movement capabilities' and 'We are excited to create value for our clients and harness the power of a new technology within the banking system.' Jamie Walker, head of digital assets and money movement, said 'This pilot is another step forward in our broader digital asset strategy' and 'Our focus remains on delivering solutions that solve real client challenges while maintaining the safety, security and reliability that clients expect from U.S. Bank.' The release places the work on a relationship with the Stellar Development Foundation, calls Stellar 'a leading institutional-grade blockchain for regulated financial services with global reach, near-instant settlement and sub-cent costs', and names 'enhanced liquidity management, collateral mobility, cross-border treasury operations, and other institutional use cases' as applications under exploration. It gives no amount, no settlement time, no reserve arrangement, no counterparty count and no timetable, and it does not say whether USBDC will be available to clients or to other institutions. The reversibility the pilot tested is not the bank's engineering, and this is where the reports go beyond the release. The Industry Spread points out that Stellar has carried issuer clawback as a native flag for years; the desk read the protocol proposal, CAP-0035, which is dated 14 December 2020, is marked final against protocol version 17, and states its purposes as recovering assets that have been fraudulently obtained, responding to regulatory actions if required, and letting proofed persons recover an enabled asset after loss of key custody. Stellar's developer documentation confirms the two flags and their dependency: AUTH_REVOCABLE_FLAG lets an issuer revoke an existing trustline's authorisation, freezing the holding and cancelling that account's open orders, and AUTH_CLAWBACK_ENABLED_FLAG lets the issuer burn a holder's balance, and it 'requires that revocable is also set'. The same report corrects a claim repeated in the coverage, saying that U.S. Bank is not the fifth-largest American bank but the sixth by consolidated assets, at $683.4bn against Goldman Sachs Bank USA at $751.8bn on the regulator's own data for 31 March 2026. Its group figures cross-check against the filings: U.S. Bancorp reported total assets of $725.9bn at 30 June 2026 in its quarterly report, and consolidated deposits of $522.2bn at 31 December 2025 in its annual report. Ledger Insights notes that the November 2025 work with the Stellar Development Foundation and PwC was an issuance capability rather than a live transaction, that the transfer is intra-group rather than a client payment, and that U.S. Bank is a part owner of Zelle, which is separately building a dollar token. This desk ran one check of its own: a query of Stellar's public Horizon asset index for the code USBDC returned no records at 13:13 UTC on 10 September, in either case, while the same endpoint returns records for USDC. That is not evidence the token does not exist. Horizon's asset index is built from trustlines, so an asset held only by the issuer and one affiliate, or a token implemented as a contract rather than as a classic asset, would not appear in it.

Why it matters

A bank-issued token on a public chain with a freeze and a clawback attached is not a stablecoin in the sense the word has been used for a decade, and pretending otherwise misses what has been demonstrated. The property that made public-chain money interesting was that a transfer, once made, was final and could not be undone by the issuer. What U.S. Bank has proved is that it can run the settlement on Stellar's rails and keep the issuer's power to reverse: freeze a holding, cancel its open orders, and burn a balance without the holder's consent. That is a bank ledger with a public transport layer, and for a corporate treasurer it may be the more attractive product, because the reason treasurers have stayed away from tokenized cash is not settlement speed but the absence of anyone to call when a payment goes to the wrong address. The interesting part of the story is therefore not that the bank built something clever; the primitive is nearly six years old and the chain was chosen because it was already there. It is that a bank has demonstrated its own risk, compliance and operations stack can fire that primitive end to end, which is the thing a supervisor asks about. The second point is what the choice of chain reveals about the market's direction. There is a live argument in this corpus between tokenized deposits, which keep the claim inside the bank and its existing rules, and stablecoins, which move it to a bearer instrument. U.S. Bank has taken a third position that looks like the first wearing the clothes of the second: a token that only its own entities have so far held, backed by nothing it has disclosed, transported over a permissionless network and revocable at the issuer's instruction. If that becomes the pattern, the public chain ends up as a messaging and settlement layer between bank ledgers rather than a place where value sits, which is the same conclusion Swift's ledger pilots have been converging on and the same one the thirty-odd institutions on the Cari network reached in public on 2 September. Third, and least remarked, the sequencing is competitive. Twenty-one banks and asset managers, four of the largest American banks among them, said on 1 September that they would establish a company in the second half of 2026 and go to market in the first half of 2027; that company does not exist yet, and the bank ranked sixth by consolidated assets has already minted, moved, frozen and clawed back its own token. An intra-group transfer with no disclosed size is a demonstration and not a business, but it is a demonstration that the consortium cannot yet match, and in a market where the supervisory question is whether a bank can control what it issues, being first to show the controls work is worth something.

What is not settled

The release contains no numbers, so almost everything a reader would price is missing: the size of the payment, the settlement time, the fee, the reserve arrangement behind the token, the entities involved beyond North America and Europe, and whether any of it is available to a client. Whether USBDC is a payment stablecoin under the GENIUS Act, and if so which issuance route the bank intends, is not addressed at all, which matters because the Treasury rules are still being written and a bank-issued token held only by affiliates may not be a payment stablecoin yet. Nothing is said about who holds the reserve, whether it is segregated, or whether the token is redeemable by anyone other than the bank itself. The clawback control is disclosed as a tested capability and not as a policy: there is no statement of when the bank would use it, who authorises it, what recourse a holder would have, or how the power is reconciled with settlement finality for a counterparty who has given value against a received token. Whether third parties will ever be able to hold USBDC is unanswered, and until they can the freeze and clawback are untested against the case that matters, which is an unaffiliated holder. The desk could not confirm the token on the public ledger: Stellar's Horizon asset index returned nothing for the code, which is consistent with a token that has no third-party trustlines and equally consistent with an implementation the index does not cover, and the release does not publish an issuer address. And the relationship to the bank's other token work is unstated, both to Zelle's planned dollar token, in which the bank is a part owner, and to the tokenized deposit services its larger peers are building.