Skip to content
Wiki
HOME / USE CASES / Project Agorá
Teardown

Project Agorá

GLOBALcbdcpilotOperator: BIS Innovation HubIssuance details unconfirmed

BIS Innovation Hub flagship cross-border tokenisation programme bringing together seven central banks and more than 40 private-sector institutions to test tokenised commercial-bank deposits settling on a unified ledger alongside tokenised central bank money. The most institutionally weighted post-mBridge cross-border tokenisation experiment to date.

Operating details

How it actually settles
Architecture
Solid = liveDashed = design-intent
1 · Two-layer unified ledger
SOLID = BUILT IN PROTOTYPEDASHED = FUTURE / NOT BUILTUNIFYING LEDGER · TOKENISED COMMERCIAL-BANK DEPOSITSUnifying ledgerTokenised commercial-bank deposits · accessible to all participantsPayment coordinatorworkflow contract: sequences the life cycle,advancing only when all legs are readyalso hosts payment-leg contracts + workflow stateorchestrates only · no bridgecentral-bank ledgers stay sovereignJURISDICTIONAL LEDGERS · TOKENISED CENTRAL-BANK RESERVESOne ledger per currency area · each operated under its own central bank's authorityEURBdFreservesJPYBoJreservesKRWBoKreservesMXNBanxicoreservesCHFSNBreservesUSDFRBNYreservesGBPBoEreservesCADBoCnext phaseDomestic RTGS / core-banking linksdesigned for, not testedDvP / tokenised securitiesextensibility · futureAtomic = coordinator-governed commit-or-cancel across legs, not synchronous cross-ledger execution.Permissioned EVM · QBFT / Proof-of-Authority · prototype user-tested; no real value settled.
Project Agorá · two ledgers, not one: tokenised deposits on a shared unifying ledger, tokenised central-bank reserves on one sovereign ledger per currency, coordinated without a bridge · BIS prototype report, 27 May 2026
2 · Correspondent banking, re-plumbed
SOLID = IN USE / BUILTDASHED = NOT YET LIVETODAY · CORRESPONDENT BANKING (PRESERVED, THEN RE-PLUMBED)Debtor bankoriginatingCorrespondentnostro / vostroFX correspondentCLS · T+2Beneficiary bankdestinationSequential hops, Swift messaging, compliance re-checked at each hop, FX via CLS (T+2), multi-daySAME MODEL, RE-PLUMBED, NOT REPLACEDON AGORÁ · FIVE-STAGE FLOW (PROTOTYPE, USER-TESTED)1 · CoPconfirm payeeprivacy group2 · Pathdiscoverybilateral3 · Validatecompliancein parallel4 · Lock& delegateon-ledger5 · Settlecommit/cancelin secondsDeposits on the unifying ledger · reserves on the jurisdictional ledgers · PvP links the legsCompliance runs in parallel, not sequentially · information aligned before funds are committedsettles in seconds · 24/7 · atomic commit-or-cancel removes the settlement-risk windowDemonstrated in the prototype, user-tested across all seven jurisdictions.No real value has settled · no production timeline · CoP = Confirmation of Payee
Project Agorá · the correspondent-banking model preserved but re-plumbed: a five-stage flow that runs compliance in parallel and settles atomically, demonstrated in the prototype with no real value yet moved
3 · The consortium
SOLID = PROTOTYPE COHORTDASHED = JOINING NEXT PHASEGOVERNANCEProject Agorá Committee · co-chaired by BIS + IIFworkstreams: business · technology · legalCENTRAL BANKS · ONE PER CURRENCY AREAEURBdFJPYBoJKRWBoKMXNBanxicoCHFSNBUSDFRBNYGBPBoECADBoCSeven central banks built the prototype; Bank of Canada (CAD) joins for the next phasePRIVATE-SECTOR COHORT · 42 FIRMS CONVENED BY THE IIFUS banks (USD)JPMorgan · Citi · BNY ·TD Bank · FNBOUK / EuropeanHSBC · Standard Chartered ·NatWest · Lloyds · BBVA ·Santander · BNP Paribas ·Crédit Agricole · BPCE ·Deutsche Bank · Commerzbank ·CaixaBank · Banco BVJapanese (JPY)MUFG · Mizuho · SMBC ·SBI ShinseiKorean (KRW)Hana · IBK · KB Kookmin ·NongHyup · Shinhan · WooriSwiss (CHF)UBS · PostFinance · Sygnum ·SDX · BC Vaudoise ·Basler KB · Amina BankMexico + networksBanorte · Monex (MXN)Swift · Visa · Mastercard ·Euroclear · Eurex Clearing42 named institutions convened by the IIF · grouped by currency leg · clusters illustrative
Project Agorá · seven central banks (one per currency) plus 42 IIF-convened private firms, governed by the BIS + IIF Project Agorá Committee · Bank of Canada joins for the next phase

The 27 May 2026 prototype report (BIS publ othp110) settles the architecture: Agorá is two layers on two ledger surfaces, not one shared platform. Tokenised commercial-bank deposits sit on a single shared unifying ledger that also hosts the payment-coordinator contract; tokenised central-bank reserves sit on one independent jurisdictional ledger per currency area, each operated under its own central bank so reserves stay under domestic control. The payment coordinator sequences the payment life cycle but never directly invokes the jurisdictional ledgers, and there is no cross-ledger bridge: coordination runs through deterministic events and participant-operated middleware. The model preserves the correspondent-banking backbone and re-plumbs it rather than replacing it, with each institution running its own sanctions and AML screening independently and in parallel ahead of an atomic commit-or-cancel settlement. Consortium-style: the seven central banks set the perimeter and the 42 IIF-convened firms build the use cases. The prototype was built and user-tested across all seven jurisdictions; no real value has settled and there is no production timeline.

Participants

BIS Innovation Hub, Federal Reserve Bank of New York, Bank of England, Bank of France, Bank of Japan, Bank of Korea, Bank of Mexico, Swiss National Bank, Institute of International Finance (IIF), Over 40 private-sector firms including JPMorgan, Citi, HSBC, Mastercard, Visa, MUFG, SMBC, BNP Paribas, Société Générale-FORGE, Société Générale, Santander, Sumitomo Mitsui Trust, DBS, Standard Chartered, BNY

Scale

Seven participating central banks (Fed New York, BoE, BdF, BoJ, BoK, Banxico, SNB) and 41 private-sector firms named in the September 2024 cohort announcement. Workstream phases through 2025-2026 cover technical experimentation, architectural design, and use-case demonstrations across cross-border payments and tokenised wholesale settlement.

Regulatory wrapper

Sits inside the BIS Innovation Hub research programme, with each participating central bank retaining sovereign authority over its CBDC issuance and each private-sector firm operating under its existing licences. No new multilateral regime is being created; Agorá is an architectural test of the unified-ledger blueprint laid out in the BIS Annual Economic Report 2025 Chapter III.

Known limits

Research and architectural pilot rather than production. No live cross-border settlement of real value across the corridor as of early 2026. The programme deliberately avoids the geopolitical complications that surrounded mBridge by anchoring to the G7-plus-Korea-and-Mexico cohort rather than including PBoC or PBoC-adjacent participants.

In depth

Project Agorá is the cross-border tokenised-commercial-bank-money project led by the BIS Innovation Hub together with 7 central banks (Banque de France for the Eurosystem, Bank of England, Bank of Japan, Bank of Korea, Bank of Mexico, Federal Reserve Bank of New York representing the USD leg, and Swiss National Bank) and a private-sector cohort of more than 40 firms convened by the Institute of International Finance (IIF). The project explores a unified-ledger architecture in which tokenised commercial bank money (i.e. tokenised deposits) settles atomically against tokenised central bank money on a single programmable platform, with the structural design of replacing the conventional correspondent-banking model for cross-border wholesale payments. For an institutional tokenisation operator, Project Agorá is the most important live cross-border tokenisation programme bringing together major non-USD G7 central banks and a deep TradFi private-sector cohort, and the worked example of the unified-ledger thesis that BIS senior research has been articulating since the 2023 BIS Annual Economic Report (Project Agorá BIS reference page).

What it is

Project Agorá was announced on 3 April 2024 as a joint initiative of the 7 participating central banks and the BIS Innovation Hub Switzerland Centre. The project's core premise is that the conventional cross-border wholesale payments architecture, where commercial banks transfer funds across borders through a chain of correspondent-bank relationships with multiple intermediation steps, is materially inefficient and that a unified programmable ledger architecture can replicate the function with materially lower latency, lower cost, and lower operational risk.

The architectural design point is the unified ledger thesis, a single programmable platform on which tokenised central bank money (the wholesale CBDC layer) and tokenised commercial bank money (the tokenised-deposit layer) coexist, with smart-contract-mediated atomic settlement across the two layers. On this platform, commercial-bank participants hold tokenised-deposit balances, central-bank participants hold tokenised-CBDC balances, and the settlement primitives allow commercial-bank tokenised deposits to settle atomically against central-bank reserves, replicating the conventional two-tier system in tokenised form on a single programmable substrate.

The 7 participating central banks were announced as Banque de France (representing the Eurosystem), Bank of England, Bank of Japan, Bank of Korea, Bank of Mexico, Federal Reserve Bank of New York (representing the USD leg), and Swiss National Bank. The currency mix is materially significant, since USD, EUR, JPY, GBP, CHF, KRW, and MXN cover most of the wholesale cross-border payment corridors that the conventional correspondent-banking architecture currently serves.

Operating model

Project Agorá's operating model has been articulated through public BIS Innovation Hub material, IIF private-sector cohort coordination, and adjacent academic and policy literature, with the live experimentation cycle running across 2024-2026. The private-sector cohort was convened by the IIF and announced in May 2024 with more than 40 firms participating, including (per BIS and IIF disclosure) Mastercard, Visa, SWIFT, Citi, HSBC, Deutsche Bank, MUFG, State Street, and a broader set of major commercial banks, financial-market infrastructures, and payment-network operators (BIS Project Agorá participant disclosure).

The separation of central-bank and private-sector roles follows the conventional two-tier system. Central banks operate the central-bank-money layer of the unified ledger and supervise the participating commercial-bank cohort. Commercial banks operate the tokenised-deposit layer, holding tokenised-deposit balances as liabilities of their own balance sheets but settling atomically against tokenised central-bank money on the same platform. The financial-market infrastructure participants (SWIFT, the card networks) bring messaging, identity, and settlement-network expertise into the design conversation.

The conceptual release published in 2025 set out the unified-ledger architecture in detail and validated the design pattern across the participating central banks and private-sector cohort. The live experimentation work through 2025-2026 has tested specific architectural choices (programmable settlement triggers, identity and KYC framework integration, asset segregation across commercial-bank participants, atomic-settlement primitives across central-bank and commercial-bank money) without yet committing to a production-platform timeline. The current read is that Agorá is in the design-and-validation phase rather than the production-deployment phase, with the question of whether a production platform follows still open.

The architectural relationship to mBridge (the multi-CBDC cross-border platform with HKMA, Bank of Thailand, PBoC Digital Currency Institute, Central Bank of UAE, and Saudi Central Bank as participants) is complementary by design rather than competitive. mBridge operates exclusively at the wholesale-CBDC layer with commercial-bank participants on each currency leg; Agorá operates at the unified central-bank-and-commercial-bank-money layer with the 7 G7-and-emerging-market central banks. The 2 platforms cover overlapping use cases (cross-border wholesale settlement) under different architectural patterns. Whether the two converge, interoperate, or remain separate is not in current public coverage.

Why it matters

Project Agorá's significance rests on the central-bank cohort it assembled, the depth of its private-sector participation, and what it tests about the unified-ledger thesis itself.

  • The central-bank participation is the clearest signal, since Agorá brings together the Federal Reserve Bank of New York (USD), Banque de France (Eurosystem), Bank of Japan (JPY), Bank of England (GBP), Bank of Korea (KRW), Bank of Mexico (MXN), and Swiss National Bank (CHF), the most consequential central-bank cohort to commit to a single tokenisation platform exploration outside the BIS Annual Economic Report's framing exercises, and a combination that complements mBridge's APAC-and-Middle-East-anchored participant set; together the 2 platforms cover most of the global wholesale cross-border payment flow.
  • The private-sector cohort also runs deep, with the 40-plus IIF-convened participants including the largest US, European, Japanese, and APAC commercial banks, the leading global financial-market infrastructures (SWIFT), the largest card networks (Visa, Mastercard), and leading custody and asset-servicing franchises (State Street, among others), a depth that materially distinguishes Agorá from earlier multi-stakeholder cross-border tokenisation experiments that struggled to assemble comparable private-sector participation.
  • The unified-ledger thesis itself is on trial, since the BIS senior research on tokenisation since the 2023 Annual Economic Report has positioned the unified-ledger architecture as the design point for next-generation programmable money infrastructure, and Project Agorá is the live test of whether that thesis holds operationally with 7 central banks and a deep private-sector cohort committed to the experimentation; the implications for the broader tokenisation conversation (including the Kinexys BDA model, Progmat-Japan tokenised-deposit pilots, the Project Ensemble HKD wholesale tokenisation, and the various national tokenised-deposit consortium platforms) carry real weight, since a validated unified-ledger architecture becomes the reference design pattern for cross-border tokenised wholesale settlement, while a failed test leaves the conventional correspondent-banking model in its current position.

The competitive map is partly mBridge on the multi-CBDC side, partly the SWIFT-led tokenised-payments interoperability work that has been positioned as a non-unified-ledger alternative, and partly bilateral central-bank settlement bridges (the BoE-MAS bridge, the Banque de France-MAS bridge, others) that test similar concepts at smaller scale.

Recent moves

  • 2025-2026. Live experimentation work continued through workstream-specific testing across the participating cohort, with the question of whether a production-platform timeline follows still open.
  • 25 November 2025. Tommaso Mancini-Griffoli appointed Head of the BIS Innovation Hub starting 1 March 2026, joining from the IMF where he was Assistant Director, Payments, Currencies and Infrastructure. The remit overlap with Project Agorá's cross-border tokenisation portfolio has been read as a continuity-and-emphasis signal for the project.
  • 12 November 2025. Andréa Maechler, then Acting Head of the BIS Innovation Hub, delivered "How deposits can harness tokenisation" at the Singapore Fintech Festival, the most recent set-piece on tokenised deposits from the Hub leadership and closely adjacent to the Agorá conceptual framing.
    1. Conceptual release published, setting out the unified-ledger architecture in detail and validating the design pattern across the participating central banks and private-sector cohort.
  • May 2024. The private-sector cohort convened by the Institute of International Finance announced with more than 40 participating firms including Mastercard, Visa, SWIFT, Citi, HSBC, Deutsche Bank, MUFG, and State Street (BIS).
  • 3 April 2024. Project Agorá announced by the 7 participating central banks (Banque de France, Bank of England, Bank of Japan, Bank of Korea, Bank of Mexico, Federal Reserve Bank of New York, Swiss National Bank) and the BIS Innovation Hub Switzerland Centre.

Open questions

  • Whether a production-platform timeline follows the conceptual-release-and-experimentation phase, and on what governance terms. The current design phase is clearly distinct from a production-deployment phase; the transition path is not in current public coverage.
  • The interaction with mBridge. The 2 platforms are complementary in design, with overlapping but non-identical participating central banks; whether they remain parallel rails, integrate, or one absorbs the other is not in current disclosure.
  • The Federal Reserve Bank of New York's continued participation across the political-economic cycle in the United States. The FRBNY participation was confirmed at announcement; subsequent political-economic developments around US central-bank engagement with multilateral cross-border tokenisation work have been actively debated in trade press but the formal participation status is unchanged in current public material.
  • The regulatory and supervisory framework for the participating commercial-bank cohort. Tokenised deposits issued by Citi, HSBC, MUFG, Deutsche Bank, and others on a unified ledger across 7 currencies require coordinated supervisory treatment that has not yet been publicly framed.
  • APAC implications. Bank of Japan and Bank of Korea are participating central banks; the integration with Progmat-Japan tokenised-deposit pilots and the broader regional tokenisation infrastructure has been thinly disclosed.
  • Agentic commerce posture. A unified-ledger architecture with tokenised commercial bank money and tokenised central bank money is naturally compatible with AI-agent-controlled wallet designs at the commercial-bank tokenised-deposit layer; whether the project's design framework explicitly accommodates agentic-commerce primitives has not been publicly addressed.

Related

Recent news2 curated · not yet covered