The BIS-led prototype has moved tokenised wholesale payments from theory toward real-value testing. The harder question is whether a shared programmable architecture can solve the legal, liquidity, compliance and governance frictions that technology alone cannot remove.
The central test: Can tokenisation compress a multi-bank, multi-currency payment chain into one coordinated transaction without sacrificing settlement finality, monetary sovereignty, financial-crime controls or operational resilience?
Cross-border payments are a strange anomaly in modern finance. A domestic transfer can move in seconds, yet an international wholesale payment can still pass through a chain of correspondent banks, separate ledgers, compliance checks, foreign-exchange processes and settlement windows. The Financial Stability Board’s 2025 progress report concluded that substantial policy work had not yet translated into tangible improvements for end-users at the global level and warned that satisfactory progress was unlikely to arrive on the original 2027 timetable. That diagnosis matters because the problem is no longer a lack of initiatives. It is the difficulty of making multiple national systems behave like one transaction.
Project Agorá is the most ambitious attempt yet to test whether tokenisation can change that architecture rather than merely speed up individual steps. In May 2026, the Bank for International Settlements and the Institute of International Finance published the project’s prototype report, after work involving seven central banks and more than 40 regulated financial institutions. The prototype combined tokenised commercial bank deposits with tokenised central bank reserves and demonstrated atomic multi-currency settlement: the payment legs settle together on an all-or-nothing basis instead of moving sequentially through a chain.
That is an important proof of feasibility. It is not yet proof that the model can operate at production scale. The BIS announcement on 27 May 2026 says the next phase will include real-value transactions for selected currencies and participants, with the Bank of Canada joining the project. The distinction is crucial. Agorá has shown that a new settlement design can work in a prototype. It has not yet shown that it can become the default global infrastructure for payments measured in trillions of dollars, under real market stress, across every relevant legal and regulatory regime.
Why Cross-Border Payments Are Still Hard to Fix
The basic correspondent-banking model is resilient because it is modular. Banks keep accounts with other banks, payment instructions travel over messaging networks and each institution applies its own controls before updating its own ledger. The weakness is that the same modularity creates duplication. Information is checked repeatedly, balances are reconciled repeatedly and settlement depends on operating hours and liquidity in different jurisdictions. Every hand-off adds latency and creates another point at which a payment can be stopped, repaired or delayed.
The G20’s targets make the gap measurable. The FSB target for wholesale payments is for 75% of cross-border wholesale payments to be credited within one hour by the end of 2027, with the remainder within one business day. Yet the CPMI’s May 2026 monitoring brief, based on responses from 82 jurisdictions, said it was unlikely that the G20 targets would be fully met by 2027. Its prescription is revealing: longer operating hours, wider access, interoperability by design, harmonised data and coordinated legal reform. In other words, the bottleneck is an ecosystem, not one piece of software.
This is why Agorá matters. It asks whether some of those separate activities can be composed into a single programmable workflow. Rather than sending a message and later reconciling the resulting ledger movements, the instruction, conditions and settlement can be linked to the same transaction logic. The potential gain is not simply faster messaging. It is fewer sequential dependencies.
What Project Agorá Actually Built
The phrase “unified ledger” can sound as though Agorá places all central-bank and commercial-bank money on one giant blockchain. The architecture is more careful than that. The BIS Annual Economic Report 2026 describes a shared cross-border platform with a unifying ledger for tokenised commercial bank deposits and separate, jurisdiction-specific ledgers for tokenised central bank reserves. That separation allows central banks to retain control over their own reserve money and domestic rules while still participating in an interoperable transaction.
The core mechanism is atomic settlement. Once validation is complete and the required balances are locked, the different currency legs settle together. If a necessary condition fails, the transaction does not partly complete. This reduces principal risk in the payment chain and can also reduce the reconciliation burden created when one part of a transaction has moved but another has not.
Programmability adds a second layer. Smart-contract logic can express conditions that today may be handled through separate messages, manual interventions or back-office workflows. The Project Agorá report says the platform can embed workflow logic, compliance requirements and conditional payment triggers. In principle, this opens the door to always-on wholesale transfers, payments released when contractual conditions are met, and tighter coordination between payment, foreign-exchange and trade-finance processes.
Why the Money Design Matters More Than the Blockchain
Agorá’s most consequential design choice is monetary rather than technological. It does not attempt to replace bank money with a new private digital asset. Commercial bank deposits remain claims on regulated banks, while central bank reserves remain the final settlement asset. The BIS’s May 2026 findings explicitly state that tokenisation does not change the legal character of central bank reserves or commercial bank deposits.
That preserves the two-tier monetary system: banks continue to provide credit, customer relationships and payment services, while central banks anchor settlement and the singleness of money. The BIS Annual Economic Report argues that this institutional arrangement is central to trust because private money can be redeemed and settled at par in central bank money. Agorá therefore represents a strategy of modernising bank money rather than routing around it.
For banks, this is strategically attractive. Stablecoins demonstrated the demand for programmable, 24/7 value transfer, but tokenised deposits allow banks to offer similar functionality without giving up the deposit relationship. For central banks, the model offers innovation without ceding control over the ultimate settlement asset. The trade-off is that the platform inherits the banking system’s governance, compliance and access constraints instead of escaping them.
The Prototype Solves Settlement Risk Better Than It Solves Everything Else
Atomic settlement is powerful, but it can be overstated. It can make a multi-leg transaction indivisible; it cannot make foreign-exchange spreads disappear, guarantee deep liquidity in every currency or remove the cost of complying with different national laws. A payment can settle instantly and still be expensive. It can be technically available 24/7 and still require a bank to hold more liquidity outside traditional operating windows. It can carry structured data and still encounter incompatible privacy, sanctions or data-localisation rules.
The project itself acknowledges these boundaries. Its legal work found that settlement finality could be achieved across the seven jurisdictions in the prototype, but the BIS press release says further technical, operational and contractual work is required to align implementation with each jurisdiction’s legal framework. Privacy was demonstrated at balance and transaction levels, but production deployment would still require governance over who can see what data, under which legal authority and how information is shared when financial-crime investigations cross borders.
Similarly, the BIS describes future enhancements in anti-money-laundering, counter-terrorist-financing, sanctions compliance and fraud detection as capabilities that can develop as regulatory and data-sharing frameworks evolve. That wording is important. Agorá can create a better technical surface for compliance automation, but it does not by itself harmonise the laws that institutions must apply.
The Real-World Test Will Be Liquidity, Governance and Failure Handling
A prototype can assume that participants are available and correctly configured. A production infrastructure must cope with failed nodes, cyber incidents, disputed instructions, insufficient balances, emergency sanctions changes, legal stays, operational outages and volatile FX markets. The BIS 2026 assessment identifies resilience, governance, access rules and interoperability with legacy systems as preconditions for scaling tokenised rails safely.
Liquidity is especially important. Today, correspondent banks manage liquidity around established cut-off times and settlement windows. Always-on settlement changes that rhythm. Treasury teams may need to monitor and mobilise liquidity overnight and through weekends, or develop automated mechanisms for reserving, pricing and transferring liquidity. Faster settlement can reduce some exposures while increasing the operational need for money to be in the right place at the right second.
Governance may prove even harder than technology. A cross-border platform needs rules for membership, software changes, cyber standards, dispute resolution, liability, access to data and emergency intervention. It must also preserve monetary sovereignty: a central bank must be able to control its currency and reserve accounts even when the payment instruction is coordinated across a shared international platform. Agorá’s layered architecture is designed to preserve that autonomy, but real-value testing is where those boundaries begin to have economic consequences.
Agorá Is Not Racing Against Swift So Much as Converging With It
The most important competitive development since Agorá’s prototype report came from the incumbent messaging network. On 9 July 2026, Swift announced that its blockchain-based ledger was ready for initial use, with 17 banks from six continents preparing to pilot live cross-border transactions using tokenised deposits. Swift’s design is not identical to Agorá: it provides an orchestration layer for bank-issued tokenised deposits on banks’ own ledgers, while final settlement can still complete through existing systems. But the direction of travel is strikingly similar—regulated bank money, shared digital coordination and 24/7 capability.
This weakens the idea that the future will be a clean contest between “blockchain” and the traditional banking network. Swift itself says 75% of payments on its network now reach beneficiary banks within ten minutes, often in seconds. The remaining friction increasingly sits beyond the message: beneficiary availability, compliance checks, liquidity, FX, local clearing, operating hours and account structures. Tokenisation is therefore likely to be layered into existing networks as much as it replaces them.
The inference is that interoperability may become more valuable than any single ledger. Banks will not want separate liquidity pools and incompatible tokens for Agorá-type infrastructures, Swift’s ledger, domestic tokenisation platforms and individual bank networks. The infrastructure that can coordinate these environments while preserving legal finality and bank-grade controls may capture more value than a closed network that is technically elegant but difficult to connect.
Europe Is Already Building the Adjacent Infrastructure
Project Agorá is also influencing domestic and regional payment design. The European Central Bank’s May 2026 summary of the project says its findings will feed into Pontes and Appia. Pontes is intended to connect market DLT platforms to TARGET Services, with a planned go-live in September 2026, while Appia explores a longer-term European tokenised financial ecosystem. As of 10 August 2026, that September go-live remains a planned milestone, not a completed deployment.
The Eurosystem’s broader payments strategy is explicit that central bank money should remain at the core of wholesale settlement while regulated private settlement assets, including tokenised deposits, can complement it. The Bank of England has taken a similar direction: in a May 2026 speech on modernising money and markets, Deputy Governor Sarah Breeden described Agorá as part of a wider effort to support tokenised deposits and central bank money in cross-border payments for financial institutions and corporates.
These projects suggest that Agorá is best understood as one piece of a broader infrastructure transition. Central banks are experimenting with how reserve money reaches programmable platforms; banks are tokenising deposits; existing networks are adding shared ledgers; and standards bodies are trying to make data interoperable. The final system may be a network of connected infrastructures rather than one universal ledger.
ISO 20022 Still Matters in a Tokenised World
Tokenisation does not make payment standards obsolete. In February 2026, the CPMI updated its harmonised ISO 20022 data requirements for cross-border payments, warning that inconsistent implementation could undermine the benefits of richer structured data. That issue does not disappear on a programmable ledger. Smart contracts are useful only when the data they consume is reliable, consistently structured and legally usable.
This is one reason the transformation will be slower than technology demonstrations imply. Cross-border payment infrastructure must connect legal identities, account information, purpose-of-payment data, sanctions screening, FX instructions and settlement records. A shared ledger can coordinate these elements, but industry standards determine whether institutions interpret the data in the same way. In practice, tokenisation and ISO 20022 are complements: one changes the execution environment, while the other improves the common language.
What It Means for Banks, Fintechs, Regulators and Investors
For banks, Agorá changes the strategic question from whether tokenisation will matter to where the bank wants to sit in the stack. Global transaction banks can use tokenised deposits to defend the deposit franchise and make corporate liquidity more mobile. Smaller banks may benefit from shared infrastructure that reduces the need for long correspondent chains, but access rules and integration costs will determine whether the network genuinely broadens competition or simply strengthens the largest institutions.
For fintechs, the opportunity is less likely to be the creation of a parallel currency than the software around interoperability, treasury automation, identity, compliance analytics, liquidity routing and programmable workflows. If bank money becomes natively usable on digital platforms, many functions that were previously bundled inside correspondent banking can become modular services.
For regulators, the challenge is to avoid solving speed by creating new concentrations of operational risk. Shared infrastructure can become systemically important quickly. Authorities will need clarity on governance, access, outsourcing, cyber resilience, data location, recovery arrangements and the treatment of transactions when participants enter resolution. The FSB’s March 2026 implementation summit underscored that the cross-border payments agenda has entered an implementation phase in which public-private collaboration and jurisdiction-level action matter more than additional high-level principles.
For investors, the near-term signal is that payments infrastructure is becoming capital-intensive again. The beneficiaries may include banks with strong global liquidity networks, technology providers that can connect multiple ledgers and institutions capable of monetising always-on treasury services. The losers are more difficult to identify. Tokenisation can compress fees in parts of the payment chain, but it can also create new premium services around liquidity, compliance and programmable cash management.
The Counterargument: Existing Rails May Improve Faster Than New Ones Scale
There is a serious case against assuming that Agorá-type platforms will replace correspondent banking. Existing rails are improving. Swift is accelerating delivery, payment systems are extending operating hours, ISO 20022 is becoming more consistent and jurisdictions are linking fast-payment infrastructures. The CPMI’s 2026 monitoring work argues that these conventional measures—wider access, longer hours, interoperability and better standards—can deliver meaningful cross-border improvements without requiring every institution to migrate to a new tokenised architecture.
That counterargument is strongest where current systems already work well. Replacing a reliable ten-minute bank-to-bank flow with a complex new ledger does not create value unless it improves the end-to-end outcome, reduces liquidity usage, automates costly processes or unlocks a service that existing infrastructure cannot offer. Tokenisation needs a business case, not merely a technical one.
The response from Agorá’s supporters is that the real opportunity lies in transactions that are difficult precisely because they involve multiple steps and institutions—multi-currency settlement, conditional payments, tokenised assets, trade finance and coordinated compliance. In those cases, making the entire transaction programmable may be more valuable than making one message faster.
Conclusion: Tokenisation Can Fix the Settlement Layer, Not the Politics of Money
Project Agorá has already answered one question. Atomic wholesale settlement across multiple currencies and jurisdictions using tokenised commercial bank deposits and tokenised central bank reserves is technically feasible within a structure that preserves central-bank control and bank money. That is a meaningful advance from earlier blockchain experiments, because it tests new technology against the institutional foundations of the monetary system rather than outside them.
The bigger question remains open. Cross-border payments are expensive and slow not only because ledgers are fragmented, but because laws, operating hours, liquidity, compliance regimes, business incentives and national monetary systems are fragmented. Agorá can coordinate more of the transaction and reduce settlement and reconciliation risk. It cannot make those institutional differences disappear.
The next phase—real-value testing—will therefore be more important than the prototype. If participants can prove resilience, manage 24/7 liquidity, achieve legal finality, automate compliance without compromising privacy and interoperate with existing rails, Agorá could become a blueprint for a new generation of correspondent banking. If those conditions prove too costly or politically difficult, tokenisation may still reshape payments, but through a patchwork of regional platforms and interoperable bank networks rather than one shared global marketplace.
The most defensible conclusion today is neither that Agorá will replace correspondent banking nor that it is another laboratory exercise. It has moved the debate to a more demanding stage. The technology has shown it can coordinate money across borders. The financial system must now show that it can coordinate the institutions around it.
References
3. Bank for International Settlements — Project Agorá project page (updated 27 May 2026)
6. Financial Stability Board — G20 Targets for Enhancing Cross-border Payments
7. Financial Stability Board — New implementation phase for cross-border payments (12 March 2026)
13. Bank of England — Sarah Breeden: Modernising money and markets (19 May 2026)
Editorial note: Project Agorá remains an experimental public-private initiative. References to possible benefits describe prototype findings or analytical implications, not a claim that a production system has already achieved those outcomes at global scale.