← All articles

What the ISO 20022 migration teaches us about post-quantum cryptography

On this page

1. What ISO 20022 was, in one paragraph
2. Lesson 1: the dates will move, and they're not yours to set
3. Lesson 2: coexistence is long, expensive and where the risk hides
4. Lesson 3: the standard was the easy part
5. Lesson 4: size limits bite where you least expect
6. Lesson 5: a fixed end date is what finally moves everyone
7. Where post-quantum is harder
8. What to take from it
9. Sources and further reading

The financial industry has recently completed a major phase of its move to ISO 20022, a richer, structured standard for payment messages. Swift's coexistence period for cross-border payments ended in November 2025, while domestic payment systems followed their own schedules. The migration wasn't about cryptography, but it's a useful case study in what post-quantum migration will demand: coordinated change across institutions, shared infrastructure, suppliers and systems that can't all switch at once.

I've spent 15 years in banking, most recently on payments systems. These are the lessons I believe carry across to post-quantum migration. They're my interpretation, based on public timelines and industry evidence rather than any one institution's experience.

What ISO 20022 was, in one paragraph

For decades, banks sent cross-border payment instructions over the Swift network in compact formats called MT messages. ISO 20022 replaces them with richer, structured messages (MX) that carry far more data about who is paying whom, and why. Because every bank and every market infrastructure had to understand the new format, nobody could switch alone. The industry agreed a period in which old and new formats ran side by side, called coexistence, followed by an end date for the old ones.

Lesson 1: the dates will move, and they're not yours to set

The cross-border migration was first planned for November 2021. It moved to the end of 2022, then to March 2023, when coexistence finally began. National systems moved with it: Canada's Lynx activated its ISO 20022 release in March 2023 to align with Swift's change, and the euro area's new T2 system went live on 20 March 2023. In the United States, the Fedwire Funds Service's own ISO 20022 implementation moved from March to July 2025.

The delays reflected how hard it is to coordinate a shared change across institutions, market infrastructures and vendors. A network-wide migration depends not just on each bank's readiness, but on agreed timelines, interoperability testing and the readiness of critical participants. Post-quantum migration has similar dependencies. A bank can upgrade some systems on its own, but shared certificate infrastructure, payment networks, hardware and third-party services need coordinated change.

Plan to be ready early, and treat the network's date as a dependency you track, not a deadline you control.

Lesson 2: coexistence is long, expensive and where the risk hides

From March 2023 to the end of coexistence for payment instructions in November 2025, banks ran both formats. That meant translating between them, testing every combination and supporting two ways of doing the same thing. Translation had a known weakness: the new messages carry more data than the old ones can hold, so converting a rich message back into the old format could lose information along the way.

Post-quantum migration will have its own coexistence challenges. Some connections will use hybrid key establishment, combining a traditional and a post-quantum mechanism; done correctly, that stays secure even against an attacker who can break only the traditional part. Others will still need to talk to partners and systems that haven't adopted post-quantum cryptography at all, and those links remain exposed until they move.

The challenge is to keep security and interoperability throughout that period, check that hybrid mechanisms are implemented correctly, and decide when vulnerable legacy-only options can be switched off. Supporting a new algorithm is one milestone; retiring the old one is a separate task that needs evidence, ownership and coordination. That's why NIST's draft transition plan, and roadmaps like Canada's for its federal systems, set end dates for the old algorithms rather than only start dates for the new ones.

Lesson 3: the standard was the easy part

Publishing the message standard didn't migrate anything. The work was finding every system that created, read, stored, screened or reported on a payment message, often across hundreds of applications, many owned by different teams and suppliers. Fields that one system ignored turned out to matter to another downstream.

Post-quantum cryptography is the same, only more so. The new standards, ML-KEM and ML-DSA, were published in August 2024. Publishing them didn't move a single bank. The hard part is a cryptographic inventory: finding and validating how cryptography is used across the enterprise. And discovering an algorithm isn't the same as understanding the risk it creates. To prioritise, each record needs:

• the system or service using it, and what it's for: key establishment, encryption, signing or authentication;
• the data or business process it protects, and how long that must stay confidential;
• the owner, vendor and dependencies;
• available post-quantum support, testing needs and what has to happen before the old method can be retired.

My illustrative inventory of a wire-payment flow shows how many places that can be in a single flow.

Lesson 4: size limits bite where you least expect

Richer messages meant bigger messages, and systems built around the old, compact formats had to be checked for field lengths, storage and screening limits. Post-quantum signatures can be much larger than today's: an ML-DSA signature is a few kilobytes, against a few hundred bytes for RSA, with the exact figures depending on the algorithm, parameter set and protocol. In 2025, the second phase of the Bank for International Settlements' Project Leap replaced conventional signatures with post-quantum ones on liquidity transfers between central banks, and reported on the performance, interoperability and crypto-agility challenges involved.

So test more than whether the cryptography works. Check certificate and message sizes, network limits, storage and database fields, hardware security modules, firmware, APIs, logging, performance and downstream processing. A system can support a new algorithm in principle and still fail when larger keys or signatures pass through its existing interfaces.

Lesson 5: a fixed end date is what finally moves everyone

For years, ISO 20022 was something most banks planned to do. The fixed end of coexistence turned it into something they had to do. Post-quantum migration has no single universal deadline for every bank and payment network yet. The Canadian Centre for Cyber Security's roadmap sets milestones for the Government of Canada's own systems: high-priority migrations by the end of 2031, the rest by the end of 2035. NIST's transition plan is still a draft. These are useful planning context, not binding deadlines for Canadian banks. OSFI has set no separate post-quantum deadline; it expects quantum risk to be managed within banks' existing technology and cyber-risk obligations.

In practice, a bank's own timeline has to come from its regulatory expectations, threat exposure, how long its data must stay confidential, system lifecycles, and dependencies on vendors and market infrastructure. Suppliers and payment networks will, meanwhile, plan around the public roadmaps.

Where post-quantum is harder

The analogy has limits, and they all point the same way.

• Scope. ISO 20022 changed payment messages. Post-quantum touches almost every system that agrees keys or checks signatures: websites, apps, cards, ATMs, internal links, hardware security modules and suppliers.
• No single coordinator. Swift and the market infrastructures could set dates for their networks. No one body sets the post-quantum timetable for everything a bank runs.
• The cost of waiting. Delaying ISO 20022 mostly postponed its benefits. Delaying post-quantum key exchange lets exposure build up silently, because data recorded today can be decrypted later.

They're also different kinds of problem. ISO 20022 changed message formats and how they're processed. Post-quantum migration changes cryptographic mechanisms, and can affect confidentiality, authenticity, integrity and long-lived trust relationships.

ISO 20022 isn't a blueprint for post-quantum migration. It's evidence that ecosystem-wide change takes years of coordination, testing and dependency management, even when the target standard and a transition date are known.

What to take from it

1. Start the inventory now; it's the long pole, and it doesn't depend on anyone else.
2. Track network and supplier timetables as dependencies, with dates and owners.
3. Plan for a long hybrid period, and budget for running two ways of doing things.
4. Test size limits early, especially in cards, message formats and hardware.
5. Build crypto-agility, so the next change is cheaper than this one.

The banking guide turns these into a plan: building the inventory, tracking vendors and payment networks and a phased roadmap. For the cryptography inside a payment message itself, read what post-quantum cryptography changes in payment messages.

Sources and further reading

Tags show what kind of source each one is. A standard or government guidance is an official document; a peer-reviewed paper has been checked by other experts; a preprint has not been peer-reviewed yet; an experiment reports a real-world demonstration; a company announcement is the company's own account. Dates and figures were checked against these sources on 12 October 2026. Spotted an error? Email hello@plainquantum.com and it will be corrected, with a note.

1. Industry guidanceISO 20022 for financial institutionsSwift
2. ReferenceSWIFT ISO 20022 adoption for cross-border payments pushed backTreasury Management International, 2020
3. Industry guidanceUpdate on implementation timing of Lynx Release TwoPayments Canada, 2022
4. Government guidanceT2 go-live on 20 March 2023European Central Bank, 2023
5. Industry guidanceWires ISO 20022 implementation rescheduled to July 14, 2025Federal Reserve Financial Services, 2025
6. StandardFIPS 204: Module-Lattice-Based Digital Signature Standard (ML-DSA)NIST, 2024
7. ExperimentProject Leap Phase 2: results of testing quantum-resistant cryptography in real-world payment systemsBanca d'Italia, 2025
8. Government guidanceNIST IR 8547 (Initial Public Draft): Transition to Post-Quantum Cryptography StandardsNIST, 2024
9. Government guidanceRoadmap for the migration to post-quantum cryptography for the Government of Canada (ITSM.40.001)Canadian Centre for Cyber Security, 2025
10. Government guidanceGuideline B-13: Technology and Cyber Risk ManagementOffice of the Superintendent of Financial Institutions (OSFI), 2022