The Bank of England published three Project Meridian FX APAC reports on 13 November 2026. The experiment connected a synchronisation-operator prototype to three simulated wholesale settlement environments with different architectures and processing capabilities.

The project demonstrated bilateral payment-versus-payment settlement and multilateral net FX settlement across those test environments. In both cases, the design sought to make linked payment legs complete together or not proceed, reducing the risk that one currency is delivered without the other.

These are experimental results, not a production launch, regulatory standard or legal conclusion about settlement finality. The Bank says the test did not reproduce the full FX trade lifecycle, live participants or production operating conditions.

A common workflow across different systems

The reports describe a synchronisation operator that received and matched settlement instructions, coordinated the reservation of funds, monitored readiness and instructed settlement or unwind actions.

The experiment did not require the three test systems to adopt the same technology. Instead, system-specific adapters and operating arrangements addressed differences in access, funds reservation, status reporting and processing.

That flexibility matters, but it also changes where responsibilities sit. Where a settlement system supported authorised third-party access and earmarking, the operator could coordinate settlement without holding participants' funds. In one test environment, however, the prototype used a dedicated escrow account to reproduce funds reservation. The Bank stresses that this was specific to the experiment and is not an inherent feature of synchronisation or a proposed future model.

Multilateral settlement and operational dependencies

The project also demonstrated multilateral net FX payment-versus-payment settlement across the three environments. The reported result indicates that a coordinated group of eligible transactions could be settled atomically after calculating offsetting positions.

The potential liquidity benefit remains conditional. The operator had to coordinate a common transaction set while funds stayed reserved across the connected systems. Differences in the information available from each system affected how quickly failed or ineligible transactions could be identified and removed.

The experiment tested selected exceptions, including insufficient funds, invalid account identifiers and temporary infrastructure outages. The results provided evidence that linked settlement could either complete or unwind in those scenarios. They do not establish performance under every real-world condition. The Bank identifies non-overlapping operating hours and the expiry of funds reservations outside operating hours among the issues that would require further work.

Technical atomicity is not legal finality

The technical report distinguishes the point at which a test transaction became technically committed from legal settlement finality. The policy report says technical atomicity alone does not create a common, legally recognised point of finality across participating systems.

A live arrangement would therefore need jurisdiction- and design-specific analysis of the operator's functions, access model, governance, oversight and information-sharing arrangements. If an operator held funds, calculated settlement positions or performed functions closer to clearing, its legal characterisation and risk profile could differ from a non-custodial coordination model.

Questions for a prospective operating model

For teams assessing a cross-border settlement design, the reports support several bounded questions:

  • Which functions sit with the synchronisation operator and which remain with connected infrastructures?
  • Can funds be reserved without the operator holding or controlling them?
  • How are instructions matched, readiness monitored and exceptions resolved across systems?
  • What happens when operating hours, status information or reservation periods differ?
  • Where does legal finality occur in each system, and how are conflicting outcomes addressed?
  • Which entity owns end-to-end resilience, incident management and cross-border governance?

These are design and governance questions, not new obligations created by the reports.

Status and sources

The Bank of England published the executive summary, policy report and technical report on 13 November 2026. They report research and experimental results; they do not announce a live service, rule, consultation or implementation deadline. The Bank's stated ambition to develop a live synchronisation service in 2028 is its own programme direction and does not amount to an equivalent commitment by other participating central banks.

This article provides general information, not legal or regulatory advice. A live settlement model requires analysis of the relevant systems, jurisdictions, permissions, contracts and operational arrangements.

← All insights