Why
The durable pattern is that operating hours and settlement technology are separate decisions, and the cheaper one usually comes first. Nothing here is onchain. This is a wire network with a longer clock, and it delivers the benefit foreign investors actually asked for — submit a won instruction at 10 a.m. in London — without waiting for a tokenised ledger to be production-ready. An engineer who reads "24-hour" as "blockchain" has misread the announcement.
"24 hours a day, excluding weekends and public holidays" is the sentence to keep. It is a window, not continuity. A rail that settles around the clock has no calendar; a window has one, and Korea's public-holiday calendar is long and moves with the lunar year. So a New York desk gains a usable morning on Tuesdays and still loses Chuseok. Any product built on this needs a holiday calendar as a first-class input, not a footnote.
The second pattern is access without an account, which is the same move as the KB Kookmin and Kinexys item: the plumbing changes, the legal relationship is re-routed rather than removed. An RFI-K is a registered intermediary, so a foreign investor is spared a Korean bank account but not spared an intermediary, its onboarding, or its own hours. "No account needed" is an access claim, not a disintermediation claim.
The Bank of Korea is also running two tokenised tracks in parallel, and they are not this network. Project Agorá, led by the Bank for International Settlements, completed real-value cross-border payment trials in July using tokenised central bank reserves and commercial bank deposits, with the won included. Project Hangang is a domestic pilot settling tokenised commercial bank deposits with a wholesale central bank digital currency. Keeping the three apart is the whole discipline: one is live plumbing, two are experiments, and only the experiments are onchain.
How it works
Three tracks, one central bank
| Track | What it is | Onchain | Status |
|---|---|---|---|
| 24-hour won settlement network | international wire network, extended hours | no | pilot with four banks; full operation January 2027 |
| Project Agorá | BIS-led cross-border trials, tokenised reserves and deposits | yes | real-value trials completed July 2026 |
| Project Hangang | wholesale CBDC settling tokenised deposits | yes | pilot |
What the pilot changes and what it does not
| Question | Before | With the pilot |
|---|---|---|
| When can a foreign desk submit? | Korean banking hours | home-country business hours, 24h |
| Weekends and Korean public holidays? | closed | still closed |
| Korean bank account required? | yes | no — an RFI-K account instead |
| Intermediary required? | yes | yes, the RFI-K |
| Settlement technology | wire | wire |
One useful test
- Fix one payer, beneficiary, amount and compliance outcome; vary only the submission time.
- Submit at 10:00 London on a Tuesday, at 02:00 Korean time, on a Saturday, and on a Korean public holiday.
- Record submission-accepted, settled, and beneficiary-usable as three separate timestamps.
- Repeat the whole matrix through a conventional correspondent route for a baseline.
- Report the hours gained and name every date the network is closed. A holiday calendar that is wrong by one day is an outage.
Where it lands in Jayverse
- Auditor: "a longer window is not a continuous rail". When a release says 24-hour, the rule asks for the calendar. The check is to name the days the service is closed; if the answer is "none", ask again, because a rail with no calendar is a different kind of system from a window with one. Same shape as the claim-boundary habit on the KB Kookmin and Kinexys item.
- Invest: one row per won venue, with its hours and its holidays. Submission window, settlement window, and the closed-days list side by side. The column that decides whether a strategy is executable from London is the holiday list, not the headline hours.
- Wallet and Rabbit: the holiday calendar is an input, not a constant. Any flow that quotes a settlement date for a won leg must read a maintained Korean public-holiday calendar, including the lunar-year moves, and must show the date it computed rather than a promise of speed.
- Chains: mark the rail, not the vibe. The capability matrix should say plainly which of these is onchain. The 24-hour network is not; Project Agorá and Project Hangang are. Three entries, one central bank, and only two of them are ledgers.
Claim boundary
The announcement supports the four launch banks, the January 2027 target, the 24-hour-minus-weekends-and-holidays window, the RFI-K access route, and the Bank of Korea's stated aim of strengthening the won's international standing. It does not state end-to-end settlement times, fees, volume limits, or which foreign banks will join. Treat "24-hour" as an operating window and nothing more.
Sources: Bank of Korea announcement as reported on 2026-09-21, pasted into chat by jay; URL not captured, so the article body is not copied into this repo. Primary references named in the piece: the Bank for International Settlements' Project Agorá and the Bank of Korea's Project Hangang.
Key expressions
| Expression | 뜻 · 쓰이는 자리 |
|---|---|
| operating window | 서비스가 열려 있는 시간대 · "It is a window, not continuity." |
| around the clock | 24시간 내내 · rail을 설명할 때. "A rail that settles around the clock has no calendar." |
| a cutoff | (당일 처리) 마감 시각 · "after the Korean cutoff" |
| a wire network | 전신 송금망 · 온체인이 아닌 기존 결제 인프라 |
| a domestic lender | 국내 은행 · 기사체에서 bank 대신 자주 |
| full operations | 본격 운영 · 시범(trial) 운영과 대비 |
| to settle a transaction | 거래를 결제(청산)하다 · "settle won transactions" |
| a registered intermediary | 등록된 중개 기관 · RFI-K를 일반 명사로 부를 때 |
| disintermediation | 탈중개 · "'No account needed' is an access claim, not a disintermediation claim." |
| onboarding | 신규 고객 개설 절차 · 계좌를 면제받아도 남는 것 |
| a wholesale CBDC | 기관용 중앙은행 디지털화폐 · 개인용(retail)과 구분 |
| tokenised reserves | 토큰화된 중앙은행 지급준비금 · Project Agorá에서 |
| international standing | 국제적 위상 · "strengthening the won's international standing" |
| a first-class input | 일급 입력값, 부수적이 아닌 필수 입력 · "a holiday calendar as a first-class input, not a footnote" |
| a claim boundary | 주장의 경계(발표가 뒷받침하는 범위) · 발표문을 읽을 때의 습관 |