Seven central banks and more than 40 regulated financial institutions completed a two-year prototype demonstrating that tokenized central bank reserves and commercial bank deposits can settle wholesale cross-border payments atomically, around the clock, across multiple currencies. The Bank for In...
"Once you know you have everything to run the transaction, you settle it in one go." — Andréa Maechler, Deputy General Manager, Bank for International Settlements
Seven central banks and more than 40 regulated financial institutions completed a two-year prototype demonstrating that tokenized central bank reserves and commercial bank deposits can settle wholesale cross-border payments atomically, around the clock, across multiple currencies. The Bank for International Settlements published the 97-page Project Agorá report on May 27, 2026, confirming settlement finality across all seven participating jurisdictions and announcing a transition to real-value transaction testing.
The prototype compresses what currently takes one to five business days through correspondent banking chains into seconds-level settlement once funds are locked. It does so without stablecoins, without public blockchain tokens, and without bypassing existing regulatory frameworks — a design choice that positions it as a direct institutional alternative to both private crypto payment networks and prior CBDC bridge experiments.
The cross-border payments market processes over $150 trillion annually through SWIFT's 11,500-member network, according to SWIFT data. Average remittance fees remain at 6.5% as of March 2025, per the World Bank — more than double the UN's 3% global target. Project Agorá addresses the wholesale layer of this market, where correspondent banking frictions impose delays, counterparty risk, and operational costs that compound across intermediaries.
Project Agorá uses a two-layer architecture designed around a non-negotiable requirement from participating central banks: each jurisdiction must retain full sovereign control over its reserve currency ledger.
Layer 1 — Unifying Ledger. Tokenized commercial bank deposits from all participating institutions sit on a shared ledger. This layer handles workflow orchestration, compliance checks, and payment coordination across jurisdictions. Smart contracts embed anti-money-laundering verification, sanctions screening, and conditional payment triggers directly into transactions.
Layer 2 — Jurisdictional Ledgers. Tokenized central bank reserves reside on independent ledgers operated under each central bank's authority. The unifying ledger coordinates workflows but never directly triggers actions on jurisdictional ledgers — a deliberate separation that preserves monetary sovereignty.
The prototype demonstrated three core capabilities:
The BIS qualitative assessment found that nine priority friction points identified by participants were "materially or partially addressed" by the prototype.
The project represents the largest BIS Innovation Hub initiative to date. Participating central banks:
| Central Bank | Jurisdiction | |---|---| | Federal Reserve Bank of New York | United States | | Bank of England | United Kingdom | | Banque de France | Eurosystem | | Bank of Japan | Japan | | Swiss National Bank | Switzerland | | Bank of Korea | South Korea | | Bank of Mexico | Mexico | | Bank of Canada (joined May 2026) | Canada |
The seven original central banks represent five major reserve currencies. The Bank of Canada joined on May 27, 2026, as the project transitions to real-value testing.
Private-sector participants include JPMorgan, HSBC, Deutsche Bank, Citi, BNP Paribas, Santander, UBS, MUFG, Mizuho, Sumitomo Mitsui Banking Corporation, Standard Chartered, BNY Mellon, SWIFT, Mastercard, and Visa — among others. Total private-sector participation exceeded 40 licensed financial institutions.
The project was co-convened by the BIS and the Institute of International Finance (IIF), whose head Tim Adams stated: "It will benefit the entire financial system."
The prototype targets four wholesale payment inefficiencies that persist in the current correspondent banking model:
1. Speed. Cross-border wholesale payments currently settle in one to five business days, depending on the number of intermediary banks involved. SWIFT data shows 50.6% of wholesale transactions clear end-to-end within an hour, but nearly half do not. The G20 target — 75% of payments within one hour and 98% within one day — remains unmet. Project Agorá's prototype settles transactions "in seconds once funds are locked," according to the BIS report.
2. Compliance efficiency. Current correspondent banking requires sequential compliance checks at each intermediary. A payment from Mexico to Japan might pass through three or four compliance screening steps, each adding hours or days. The prototype executes compliance checks in parallel rather than sequentially, using smart contracts to run anti-money-laundering and sanctions screening simultaneously across jurisdictions.
3. Transparency. Payment status visibility in correspondent banking is limited. Sending institutions often cannot track where a payment sits in the intermediary chain. The shared unifying ledger provides real-time payment status visibility for all parties in the settlement chain.
4. Counterparty risk. Sequential settlement across multiple intermediaries creates windows where one party has released funds but the counterparty has not. Atomic settlement eliminates this window entirely — the transaction either completes in full or reverts.
Project Agorá is the third major BIS cross-border payment initiative, following Project Helvetia and Project mBridge. Each took a materially different approach.
Project mBridge (launched 2021, MVP reached mid-2024) built a shared multi-CBDC platform among four central banks — the People's Bank of China, Bank of Thailand, Central Bank of the UAE, and Hong Kong Monetary Authority. mBridge uses a single shared ledger where all central banks operate, raising sovereignty concerns that limited adoption by Western central banks.
Project Agorá solves the sovereignty problem through its two-layer architecture. Central bank reserves never leave jurisdictional ledgers. The unifying ledger coordinates workflows but cannot trigger reserve movements — central banks retain full autonomy.
SWIFT is pursuing its own blockchain-based shared ledger, which reached MVP status on March 30, 2026, enabling 24/7 cross-border payments through tokenized commercial bank deposits. SWIFT's approach adds blockchain capability to its existing 11,500-institution network rather than building a new platform. However, SWIFT's shared ledger does not tokenize central bank reserves — it operates at the commercial bank deposit layer only.
Project Agorá's dual-layer model positions it between these approaches: more sovereignty-preserving than mBridge, more architecturally ambitious than SWIFT's overlay, and compatible with existing legal frameworks in all participating jurisdictions.
The prototype's architecture carries a specific message for private cross-border payment networks built around bridge tokens or stablecoins: institutional infrastructure is adopting blockchain rails while deliberately avoiding public crypto assets as settlement instruments.
Ripple's XRPL was designed to serve as a bridge currency for cross-border payments. Stellar's network targets remittance corridors and underserved markets. Both positioned their native tokens (XRP and XLM, respectively) as liquidity bridges between fiat currencies.
Project Agorá achieves the same objective — atomic, multi-currency, near-instant settlement — using tokenized versions of existing regulated money. No bridge token required. No liquidity pool needed. Settlement occurs in central bank money, not private tokens.
This does not render XRP or XLM irrelevant. Ripple has expanded into stablecoins (RLUSD reached $1.4 billion market cap by early 2026) and institutional infrastructure. Stellar recently secured a connection to DTCC's tokenized securities platform. Both networks serve corridors and use cases outside Project Agorá's wholesale institutional scope.
But the value proposition narrows. If central banks and the world's largest commercial banks can settle cross-border wholesale payments atomically in central bank money, the addressable market for bridge-token-based alternatives contracts to retail remittances and emerging-market corridors — segments with lower margins and higher regulatory complexity.
The transition from prototype to real-value testing represents the project's first contact with production-grade constraints. Several open questions remain:
Scalability. The prototype demonstrated functionality with a limited number of simulated transactions. Processing volumes comparable to SWIFT's 68 million daily messages at peak would require infrastructure orders of magnitude larger.
Legal harmonization. Settlement finality was confirmed across seven jurisdictions, but production deployment requires harmonizing technical protocols and contractual frameworks with domestic legal requirements in each jurisdiction.
Operational hours. The prototype supports around-the-clock settlement, but participating central banks currently operate on business-day schedules. Extending central bank operations to 24/7 introduces staffing, risk management, and policy questions that remain unresolved.
Timeline. The BIS has not announced a production deployment date. Carolyn Rogers, Bank of Canada Senior Deputy Governor, noted the technology has "the potential to make these payments faster, cheaper and more efficient and secure" — but IIF head Tim Adams emphasized there is no set timeline to "get this right." Mid-2026 marks the first checkpoint for real-value testing performance.
Project Agorá represents the most technically complete attempt by central banks to apply tokenization to wholesale cross-border payments. The two-year prototype answered the threshold question — can tokenized central bank reserves and commercial bank deposits achieve atomic settlement across multiple jurisdictions without compromising monetary sovereignty? The answer, according to the BIS, is yes.
What remains unanswered is whether the prototype can scale to production volumes, whether central banks will extend operational hours to support 24/7 settlement, and whether the legal harmonization required across eight (and potentially more) jurisdictions can be achieved within a commercially relevant timeframe.
The project's trajectory also clarifies the institutional position on how blockchain enters cross-border payments: as infrastructure for existing regulated money, not as a platform for new monetary instruments. For the $150 trillion annual cross-border payment market, that distinction carries significant implications for every participant — from central banks to commercial banks to private crypto networks seeking the same market.