Why
Two banking systems answered the same question in one week and picked opposite sides. Thirty-nine US state bankers associations launched a network the industry would own, precisely so member banks would not depend on somebody else's rail. Shinhan signed to verify issuance, remittance and redemption on Visa's platform. Neither is wrong. What makes the pair worth one card is that Visa is the finished form of the choice the American banks are making — it began as a bank-owned association and ended as the company banks now pay. The Korean group is renting from the outcome; the American associations are re-running the beginning.
The consortium answer exists because of a cost that has nothing to do with Basel. tokenized-money-banks works out what each product family costs a bank's balance sheet — tokenized deposits cheapest, first-party stablecoins expensive, third-party cheapest of all because the claim leaves entirely. What it does not ask is whether a bank with four branches can issue any of them. The answer has never been about capital treatment. It is that the fixed cost of building and running the infrastructure does not scale down, and a consortium is the only structure that has ever solved that for small banks.
So the interesting part is not the technology, it is the ownership clause — and that clause has a track record. A consortium forms because no member can afford the rail alone, the rail succeeds, the rail's economics stop resembling its members' economics, and ownership separates. Visa and Mastercard are the loudest case. That is not a cynical prediction about this alliance; it is the base rate, and a card that writes down the base rate is more useful than one that repeats the press release.
The detail that says how early this is: no technology partner has been selected. What launched is a governance structure, not a system — thirty-nine associations agreeing on who owns the thing before anyone has decided what the thing is. For a network whose entire premise is industry-owned, that order is arguably correct, and it also means the 2027 target is a target for something not yet specified.
On the renting side, the three verbs in the Korean announcement are the demo and card settlement is the business. Issuance, remittance and redemption are what every stablecoin pilot verifies, and verifying them proves almost nothing. The line worth reading twice is the pilot for card settlement, because that is where a card issuer's money and float actually sit: a multi-party, multi-day netting chain from acquirer to network to issuer, much of it cross-border. If stablecoins do anything measurable to a bank's P&L in the near term, it is there and not in retail payments. Note the shape shared with japan-t0-settlement: the value is in compressing a settlement interval, and the cost is whatever netting that interval was providing.
And the timing is a legal fact, not a strategic one. Who may issue a won stablecoin is one of the contested clauses korea-digital-asset-act is built around. Verifying issuance on someone else's platform is what a bank does while waiting to learn whether it will be allowed to issue on its own — a position taken before a rule exists, which is the only kind available right now.
Underneath all three answers sits the same unglamorous mismatch. An ISO 8583 authorization is decades old, message-based and reversible; on-chain settlement is final. Mapping where the two meet is mapping where the reversibility gap has to be absorbed by somebody — and that somebody is a business decision, not a technical one. Build, rent or own, you inherit it either way.
How it works
Three answers, one question
| Answer | Who | What it costs | What it risks |
|---|---|---|---|
| Own it together | 39 US state bankers associations, BankChain Alliance | Building a rail from nothing, 2027 target, partner not yet chosen | Ownership historically does not stay where it started |
| Rent it | Shinhan × Visa, MOU 2026-08-24 | Nothing to build | The rail's economics are set by someone whose interests diverge over time |
| Build it yourself | The implicit third option | Fixed cost that does not scale down | Only available to large institutions |
Each side's risk is the other side's premise. Filling in the two-date table below is what turns that sentence from a slogan into an estimate.
The two-date table, and why it is the whole card
| Network | Launched industry-owned | Ownership ended or diluted | Years |
|---|---|---|---|
| Card associations (Visa, Mastercard) | |||
| Shared ATM networks | |||
| Bankers' banks | |||
| ACH operators |
An afternoon of reading produces a number nobody publishes: how long industry ownership has historically lasted once the rail succeeded. It is not a prediction — it is the prior a prediction would have to argue against, and it applies to every consortium chain, not only this one.
What the alliance says it will carry
| Service | What it needs that a single community bank lacks | Where it lands in tokenized-money-banks |
|---|---|---|
| Tokenized deposits | A ledger other banks also honour | Cheapest Basel treatment, narrowest reach — reach is exactly what a consortium adds |
| Stablecoins | Reserve management, redemption ops, an issuer entity | The expensive conversion: the claim stops being a deposit |
| Smart payments | Counterparties on the same rail | New ground |
| Automated settlement | Shared timing and dispute rules | The plumbing half, and the part nobody demos |
The unstated fact that decides the Korean side's cost
| If it turns out to be | The claim becomes | Basel treatment | Who holds the reserve |
|---|---|---|---|
| A tokenized deposit | Still a deposit, wrapped | Cheapest — 25% LCR runoff, 50% NSFR ASF (permissioned form) | The bank |
| A first-party stablecoin | A redemption contract | Expensive — LCR runoff 100%, NSFR ASF 0% | A segregated pool |
| A third-party stablecoin via a partner | Leaves the balance sheet entirely | No issuance cost, no claim either | Someone else |
The announcement is compatible with all three, which is why classification is the first task rather than a formality.
Why card settlement is the line to watch
| Leg of a card transaction | Where the delay is | What a stablecoin could compress |
|---|---|---|
| Authorization | Seconds | Nothing — already instant |
| Clearing | Same day to next day | Little; this is messaging |
| Interchange and settlement between acquirer, network and issuer | Days, netted in batches | The float, and the prefunding held against it |
| Cross-border leg | Correspondent banking and FX | The part with the largest spread |
The measurable question is the one cross-border-rail-interop asks of corridors: how much prefunded balance does the current arrangement require, and what does compressing the interval free up. A group that owns both a bank and a card company can see both ends of the chain.
The mismatch all three inherit
| Card rail (ISO 8583) | On-chain settlement | |
|---|---|---|
| Age | Decades | New |
| Model | Message-based | State-based |
| Reversibility | Authorization, capture, chargeback windows | Final |
| Who absorbs the gap | A business decision, not a technical one | — |
A flow map of where the card leg meets an on-chain settlement leg, with the touch points marked, is the concrete artefact here — no development item attached, and useful under any of the three answers above.
The confidentiality problem the consortium answer creates
On a shared ledger among thousands of institutions, keeping existing bank regulation and security standards bites in one place: a bank's balances and counterparties must not be visible to its competitors, and a naive shared chain publishes exactly that. So the real engineering question is not throughput — it is which parties see which fields, which is the subject what-encryption-does-not-hide circles. Any answer here will be an answer to that, whether or not the announcement frames it that way. Read alongside cross-border-rail-interop, where the same lesson appears in a different costume: governance, not throughput, decides which rail wins.
Where it lands in Jayverse
- Bridge: write down the failure state for a stuck mint. The Anvil⇄Sepolia lock-and-mint relayer sits exactly on the mismatch this card describes — a lock is final onchain, so document now what happens when the mint leg fails after the lock succeeds, since no card-style chargeback exists to fall back on.
- Devnet: check whether "own it" still scales down. Devnet is Jayverse's own-it-together answer, a hosted Anvil forked from Sepolia; watch whether its running cost stays flat as more services depend on it, and decide the point at which renting a public testnet for some workloads beats owning devnet for all of them.
- Verex: model the settlement leg, not just the authorization leg. Stripe onboarding and CLOB matching are the reversible, message-based leg; onchain settlement is final. Write out the netting/float window between them explicitly, the way the card's card-settlement leg does, since that is where money actually sits before finality.
Key expressions
| Expression | 뜻 · 쓰이는 자리 |
|---|---|
| run this play | 이미 시도된 전략을 그대로 실행하다 · 과거 사례를 반복한다는 뜻으로 쓰임. "four bank-owned networks that already ran this play" |
| work out | 계산해서 알아내다, 산출하다 · 비용·구조를 따져볼 때. "works out what each product family costs a bank's" |
| scale down | 규모가 줄어들다, 축소되다 · 고정비가 소규모로 잘 줄지 않는다는 맥락. "the infrastructure does not scale down" |
| have a track record | 실적·이력이 있다 · 반복돼 온 패턴을 가리킬 때. "that clause has a track record" |
| base rate | 기저율, 비교의 기준이 되는 비율 · 예측이 따져야 할 사전 확률. "it is the base rate, and a card" |
| float | 결제 대기 중인 부동자금 · 카드 정산 과정에 묶여 있는 자금을 가리킴. "the float, and the prefunding held against it" |
| absorb | 손실·비용을 떠안다, 흡수하다 · 누가 위험을 감당하는지 말할 때. "has to be absorbed by somebody" |
| either way | 어느 쪽이든 · 선택지와 무관하게 결과가 같음을 강조. "you inherit it either way" |
| prior | 통계상 사전 확률, 미리 정해둔 기준값 · 예측이 반박해야 할 출발점. "the prior a prediction would have to argue against" |
| free up | 자금·자원의 여유를 만들다, 풀어주다 · 시간 단축이 무엇을 절약하는지 물을 때. "what does compressing the interval free up" |
| bite | 문제가 실제로 발목을 잡다, 영향을 미치다 · 규제가 구체적으로 걸림돌이 되는 지점. "existing bank regulation and security standards bites in one place" |
| ISO 8583 | 카드 결제 메시지 표준(International Organization for Standardization 8583) · 카드 인증 메시지 규격을 가리키는 오래된 표준. "An ISO 8583 authorization is decades old, message-based" |
| ACH | 자동화청산소(Automated Clearing House) · 은행 간 소액 자동이체를 처리하는 미국 결제망. "the bankers' banks, the ACH operators" |
| MOU | 양해각서(Memorandum of Understanding) · 신한-Visa 간 협약 체결을 가리킴. "Shinhan-Visa MOU signed 2026-08-24" |
| LCR | 유동성커버리지비율(Liquidity Coverage Ratio) · 바젤 규제상 단기 유동성 요구 비율. "25% LCR runoff, 50% NSFR ASF" |
| NSFR ASF | 순안정자금조달비율의 가용안정자금(Net Stable Funding Ratio, Available Stable Funding) · 장기 자금조달 안정성을 재는 바젤 지표. "50% NSFR ASF (permissioned form)" |
| Basel | 바젤 은행자본규제 체계(Basel framework) · 은행 자본·유동성 규제의 국제 기준. "decides the Basel treatment" |
| BankChain Alliance | 미국 39개 주 은행협회가 공동소유로 만든 네트워크 · 카드 컨소시엄 대안 사례로 언급. "39 US state bankers associations, BankChain Alliance" |