Lloyds Banking Group has completed three live tokenized deposit transactions inside Project Agorá’s Real-Value Testing phase — becoming one of the first banks to publicly confirm real money moving through the Bank for International Settlements’ multi-central-bank settlement framework. The headline transaction, a sterling-for-Swiss-franc conversion, linked the FX exchange, payment instruction, and final settlement into a single simultaneous flow rather than routing them through separate systems in sequence. That “all-or-nothing” structure is what payments engineers call atomic settlement — and it is the first time a UK retail bank has publicly reported executing it across currencies in a live, regulated wholesale environment.
For the roughly 40 private-sector institutions and eight central banks participating in Project Agorá — convened by the BIS and the Institute of International Finance — today’s announcement is a specific data point that moves the project from demonstrated prototype to documented real-value execution. For the European Central Bank, which is scheduling its Pontes pilot — a system to link distributed ledger technology platforms directly into TARGET Services, the ECB’s settlement backbone — for the third quarter of 2026, those live data points feed directly into an infrastructure decision that is now weeks away.
Why Correspondent Banking Loses $120 Billion a Year — and Why Atomic Settlement Is the Fix
Cross-border wholesale payments still run, in most cases, the way they have for decades: through a chain of correspondent banks, each maintaining nostro and vostro accounts on behalf of other institutions. A payment crossing currencies involves separate systems handling the FX conversion, the payment instruction, and the final settlement — often across different institutions, different time zones, and different cut-off windows. The FX leg typically settles through CLS (Continuous Linked Settlement), the multilateral netting system built after the 1974 Herstatt Bank failure showed what happens when the FX delivery and receipt legs are not synchronized: the German bank was closed mid-day, between deutschmarks delivered and US dollars not yet received, leaving counterparties exposed to losses equal to the full principal of delivered currency.
That exposure — FX settlement risk, sometimes still called Herstatt risk — is not merely historical. The Bank for International Settlements has repeatedly flagged it as a structural feature of the existing correspondent system, not a failure mode. Industry estimates put the total annual cost of correspondent banking inefficiencies at roughly $120 billion, including intermediary fees that can run 1% to 3% of transaction value for large corporates and 5% or more for small businesses.
Tokenization addresses this at the architecture level. When the deposit itself is represented as a programmable token on a shared platform — carrying all the information needed to execute a transaction, including counterparty details, currency, and settlement terms — the separate steps that would ordinarily run in sequence can instead be bundled into a single smart contract instruction. Either everything executes at the same moment or nothing does. The conversion happens, the payment happens, and the settlement in central bank reserves happens simultaneously. No party is ever in the position of having delivered value without confirming receipt.
What Lloyds Actually Did — and Why the Three-Transaction Structure Matters
In the atomic GBP/CHF transaction, Lloyds’ Corporate Markets business converted Swiss francs into sterling, with Lloyds Bank providing the sterling settlement leg. The critical structural element, confirmed in the official announcement, is that the FX conversion, payment, and settlement were executed simultaneously rather than as sequential hand-offs between systems.
In two further transactions, Lloyds participated as a customer bank — receiving, rather than providing, the settlement leg — in Swiss franc and euro payment flows using tokenized deposits. This combination tests both sides of the transaction architecture: the bank providing the settlement asset and the bank receiving it. Together, the three transactions demonstrate that tokenized commercial bank money can move across GBP, CHF, and EUR within the regulated banking system without leaving the two-tier monetary structure — central bank reserves remain the settlement layer; commercial bank deposits remain the consumer-facing layer; nothing about that relationship changes under tokenization.
“Moving from prototypes to live transactions is an important step in understanding how tokenized deposits could work in real payment scenarios,” said Peter Left, Head of Digital and Markets Innovation at Lloyds Banking Group. “This testing has allowed us to bring together our payments, settlement and foreign exchange capabilities in a cross-currency transaction. It gives us valuable practical insight into how tokenization could help reduce friction and settlement risk in wholesale payments, while operating within the regulated banking system.”
How Project Agorá’s Architecture Differs From Swift’s Ledger — and Why That Gap Matters
Readers of TechTimes’ July 10 coverage of Swift’s blockchain shared ledger will recognize the surface similarity: both Swift’s 17-bank pilot and Project Agorá involve tokenized deposits and distributed ledger technology targeting cross-border payments. The architectural difference is fundamental.
Swift’s shared ledger, built on Hyperledger Besu with Chainlink CCIP interoperability, is a commercial-bank-layer system. When a tokenized deposit moves across the Swift ledger, the underlying cash still settles through correspondent banking rails when those systems reopen. Banks gain always-on liquidity management and the ability to pre-position funds overnight and on weekends — a genuine advance over today’s cut-off windows — but not atomic finality. The settlement window remains; it is just shorter and available outside business hours.
Project Agorá settles in tokenized central bank reserves, not only tokenized commercial bank deposits. When the GBP leg of Lloyds’ GBP/CHF transaction settled, it settled in Bank of England reserves on the shared platform — the same settlement asset that conventional RTGS systems use, moved into a programmable layer. That is what makes the settlement atomic in the full sense: the FX conversion, the commercial bank token movement, and the central bank reserve transfer execute together. There is no settlement lag to manage, no Herstatt exposure window, and no dependency on a legacy correspondent banking system reopening.
BIS Deputy General Manager Andréa Maechler articulated the distinction in January 2026, when Project Agorá entered its user-testing phase: “Atomic settlement could be a game changer for cross-border payments in a digital era,” she said in remarks accompanying the real-value testing announcement. The difference is not rhetorical — it is the difference between restructuring the timing of correspondent banking (Swift’s model) and structurally replacing the settlement-risk exposure correspondent banking creates (Agorá’s model).
Lloyds’ Digital Finance Track Record: From Gilts to Francs in Eighteen Months
Today’s announcement did not arrive in isolation. Lloyds has built a systematic portfolio of tokenization milestones since July 2025, each adding a new dimension to what tokenized commercial bank money can do.
In July 2025, the bank worked with Aberdeen Investments and Archax to use tokenized money market fund units and tokenized UK gilts as collateral for foreign exchange trades — the first such use of those instruments in the UK. Then, in January 2026, Lloyds completed what it described as the UK’s first public blockchain transaction using tokenized sterling deposits, issuing the tokens on the Canton Network — a privacy-enabled public blockchain designed for regulated financial markets — and using them to purchase a tokenized gilt from Archax.
Each transaction added a dimension. The July 2025 trade demonstrated tokenized assets as collateral. The January 2026 trade demonstrated tokenized deposits on a public blockchain in a domestic sterling context. The Project Agorá transactions add multi-currency, multi-central-bank, international settlement. That progression — collateral, domestic blockchain, international atomic settlement — traces the architecture of a tokenized wholesale payments system from its components to its completion.
What Structural Risk Atomic Settlement Actually Eliminates
The Herstatt Bank’s failure in June 1974 established the concept that FX settlement risk is not a management failure but a structural property of non-simultaneous settlement. When a bank delivers one currency and must wait — even hours — to receive the other, it carries full principal risk on the delivered amount during that window. The multilateral netting through CLS was the industry’s response: netting reduces settlement exposures rather than eliminating them, and CLS operates only during defined daily windows, in defined currencies.
Atomic settlement in a framework like Project Agorá does not reduce settlement exposure — it eliminates the window during which it exists. The BIS May 2026 prototype report confirmed that atomic settlement is achievable across all seven participating jurisdictions and that settlement finality — the legal certainty that a transaction cannot be unwound — is technically demonstrated. The report also flagged that “further technical, operational and contractual development” is needed to align the model with each jurisdiction’s legal frameworks — an honest acknowledgment that confirmed feasibility is not yet confirmed readiness for production at scale.
That distinction matters for readers assessing the timeline. What Lloyds completed today is a real-value demonstration inside a controlled testing environment: real money, real central bank reserves, real atomic finality, but not yet on the rails that will handle the full volume of global wholesale payments. The path from demonstrated to deployed runs through the regulatory and contractual work the BIS flagged, and through the Pontes pilot launch scheduled for this quarter.
The September Countdown: What Pontes Going Live Actually Means
The Eurosystem’s Pontes initiative — its plan to link market DLT platforms directly to TARGET Services, the ECB’s settlement infrastructure for T2 (payments), T2S (securities), and TIPS (instant payments) — is scheduled to launch its pilot in the third quarter of 2026. That quarter ends September 30. The ECB’s official materials describe the full go-live as planned for the first quarter of 2028, but the Q3 2026 pilot is the first live connection between DLT-based transactions and central-bank-money settlement inside the European financial infrastructure.
The ECB confirmed in May 2026 that insights from Project Agorá were feeding directly into Pontes and into Appia, the Eurosystem’s longer-term blueprint for a European tokenized financial ecosystem. Appia’s blueprint is due in 2028; Pontes is the bridge infrastructure that allows DLT-based transactions to settle in central bank money now, before the longer-term vision is finalized.
What that means for the Project Agorá participants: the real-value data produced by banks like Lloyds in July 2026 is the input that informs the Pontes pilot’s design choices. A successful atomic GBP/CHF settlement — documented, verifiable, with real money and real central bank reserves — is precisely the evidence that tells the Eurosystem’s engineers what the platform must be capable of handling and what contractual alignment is still required between jurisdictions.
Bank of Canada senior deputy governor Carolyn Rogers stated upon joining the project in May 2026: “We know that the Canadian economy could benefit from innovation in cross-border payments. Tokenization has the potential to make these payments faster, cheaper, and more efficient and secure.”
Does Atomic Settlement Really Work — or Is This Still a Prototype?
The BIS has been consistent in framing Project Agorá as experimental — “BIS Innovation Hub projects are experimental in nature,” its press releases note, “for the purpose of investigating technological and practical feasibility.” That framing should be understood accurately rather than dismissively. It means the project is structured to produce reliable evidence about what is possible, not to announce a product. “Experimental” in this context is a methodological designation, not a synonym for “uncertain.”
What the May 2026 prototype report found — and what today’s real-value transactions reinforce — is that the technological and cross-jurisdictional feasibility questions have been answered affirmatively. Atomic settlement works across currencies and jurisdictions. Settlement finality is legally achievable under the frameworks of all seven original participating central banks. Privacy protections are technically implementable at both the balance and transaction level.
What remains open are the questions that come after feasibility: the operational questions (24/7 availability requirements, liquidity management, system resilience), the contractual questions (aligning the model’s settlement finality with each jurisdiction’s legal definition of finality), and the policy questions (whether central banks and regulators in the broadest set of currencies will commit to the platform or develop parallel national alternatives). Those are not small questions. But the technological demonstration that the May 2026 report provided, and that Lloyds’ live transactions now supplement with real-money evidence, is no longer in question.
Tim Adams, head of the IIF, captured the condition for success when the real-value testing phase was announced: “Tokenization has the potential to reshape how value moves, but only if it can be effectively integrated with governance, compliance, and risk frameworks that regulators and markets can trust.” That integration is what the next phase of Project Agorá, the Pontes pilot, and the contractual harmonization work underway across jurisdictions are designed to achieve.
Frequently Asked QuestionsWhat is atomic settlement, and why does it eliminate FX settlement risk in a way that older systems cannot?
Atomic settlement means all legs of a transaction — the FX conversion, the payment instruction, and the final transfer of funds between central bank accounts — execute simultaneously as a single event governed by a smart contract. If any leg fails, the entire transaction reverts. No party is ever in the position of having delivered one currency and not yet received the other, which is the exposure that created the Herstatt risk problem in 1974 and has been managed (but not eliminated) by multilateral netting systems like CLS ever since. Atomic settlement does not reduce the exposure window — it eliminates it entirely. Project Agorá achieves this because both the commercial bank tokens and the central bank reserve tokens live on the same shared programmable platform: the FX conversion and the central bank settlement can execute together, rather than running through separate systems that complete at different moments.
How is Project Agorá different from Swift’s new blockchain ledger, and why do both exist?
Swift’s shared ledger, which entered production in July 2026 with 17 pilot banks, uses tokenized commercial bank deposits on a Hyperledger Besu network to enable always-on, cross-border liquidity management. Its key advance is eliminating the hard cut-off windows that prevent banks from moving funds overnight and on weekends. However, the underlying cash still settles through correspondent banking rails when those systems reopen — the settlement window is shorter, but it exists. Project Agorá uses tokenized central bank reserves as the actual settlement asset, achieving true atomic finality without dependency on a legacy correspondent system. Both are addressing the same structural problem but at different layers of the payment stack. They are expected to coexist and potentially interoperate as the infrastructure matures.
What happens next, and when might this kind of settlement become standard for international corporate payments?
The Eurosystem’s Pontes pilot, which will link distributed ledger platforms directly to TARGET Services for settlement in central bank money, is scheduled to launch in the third quarter of 2026 — meaning within the next two months. That pilot represents the first live connection between DLT-based transactions and the ECB’s settlement backbone. A full Pontes go-live is planned for the first quarter of 2028. For the broader adoption of atomic cross-border settlement as a standard feature of international corporate payments, two things must happen in parallel: the contractual and legal frameworks in each participating jurisdiction must be harmonized to ensure that DLT-based settlement finality is recognized as legally binding under local law, and participating banks must demonstrate that the platform can handle the throughput and 24/7 availability requirements of production-grade wholesale payment volumes. Both are active work streams in 2026 and 2027.
What does Lloyds’ involvement in Project Agorá mean for UK corporate banking clients specifically?
For now, it is a research and development signal rather than a product announcement. Project Agorá is experimental infrastructure, and Lloyds’ participation means it is accumulating the technical knowledge, internal process alignment, and cross-jurisdictional experience that would be prerequisites for offering atomic multi-currency settlement to corporate clients at scale. In practical terms: corporate treasurers at large UK companies with significant cross-border flows in GBP, CHF, and EUR should watch the Pontes pilot launch this quarter and the contractual harmonization progress across jurisdictions, since those two milestones define the path from Lloyds’ current demonstrated capability to a production service that a corporate treasury could rely on for its day-to-day payments.