Learn · Migration

Planning your post-quantum migration

A practical, step-by-step guide: governance, inventory, prioritisation, algorithm choices, remediation layer by layer, testing and timelines.

In short: post-quantum migration is a multi-year programme, not a single project. Every national framework starts the same way: name an owner, find where public-key cryptography is used, rank systems by risk, and turn that into a funded roadmap aligned with regulator deadlines around 2030 and 2035.

Last reviewed: 14 October 2026

1Govern2Discover3Prioritise4Plan5Remediate6ValidateA repeating programme: keep the inventory current and stay crypto-agile as standards evolve.
Six phases of a post-quantum migration programme. Discovery and remediation are the largest pieces of work.

Phase 1 · Govern

Quantum risk is a business risk with a long lead time, so it needs ownership at senior level.

  • Name an executive owner and a programme lead, and brief the board.
  • Add quantum risk to the risk register, with links to existing cyber, data-protection and resilience frameworks such as ISO 27001.
  • Agree the scope: business units, subsidiaries, outsourced services and critical suppliers.
  • Set a reporting rhythm and success measures, for example the share of high-risk systems migrated.
  • Budget for discovery, testing, hardware refresh and skills, not only software licences.

Phase 2 · Discover

Data-centre aisle with an overlay of network connections
Cryptography is spread across applications, networks, devices and suppliers, which is why discovery comes before any upgrade. Image: Canva

You cannot replace cryptography you cannot find. A cryptographic inventory records which algorithms, key sizes, certificates, libraries and protocols are used, where, and by which business service. A cryptographic bill of materials (CBOM) is a structured, machine-readable form of that inventory that can be kept up to date.

Where to look

AreaExamples
Network and protocolsTLS on websites, APIs and load balancers; VPNs; SSH; email gateways
Certificates and PKIPublic and internal certificate authorities, certificate lifecycles, root and intermediate keys
Applications and codeCryptographic libraries, hard-coded algorithms, third-party components
Keys and hardwareHardware security modules, key management systems, smart cards, tokens, trusted platform modules
Data at restDatabases, backups, archives and file encryption
Suppliers and cloudSaaS, cloud platforms, managed services and the vendor products you run
Devices and OTIoT, industrial control systems, medical devices, network equipment with long lifecycles

Combine several methods: automated discovery tools that scan code, network traffic and certificates; configuration reviews; and questionnaires for suppliers. Automated tools find far more than interviews alone, but no single tool sees everything.

Phase 3 · Prioritise

Rank systems so the riskiest move first. A simple scoring model works well:

FactorQuestionRaises priority when…
Data shelf lifeHow long must the data stay confidential?More than five to ten years
ExposureDoes traffic cross the internet or untrusted networks?Internet-facing or partner links
Trust roleDoes the system sign code, firmware, certificates or transactions?It is a root of trust for others
DependenciesHow many systems break if it changes?Shared platforms and PKI
Upgrade pathCan it be fixed in software, or must hardware change?Hardware replacement with long lead times
RegulationIs it covered by a regulator’s expectations?Banking, government, critical infrastructure

Apply Mosca’s test to each system: if the years the data must stay secret plus the years the migration will take exceed the years until a quantum attack is possible, the system is urgent.

Phase 4 · Plan

Choose the algorithms

UseQuantum-vulnerable todayPost-quantum replacementNotes
Key establishmentRSA, ECDH, Diffie-HellmanML-KEM (FIPS 203)ML-KEM-768 is widely used; ML-KEM-1024 is required for US national security systems under CNSA 2.0
General digital signaturesRSA, ECDSA, EdDSAML-DSA (FIPS 204)Signatures are larger, for example 3,309 bytes for ML-DSA-65
Conservative signaturesRSA, ECDSASLH-DSA (FIPS 205)Based only on hash functions; larger and slower
Firmware and code signingRSA, ECDSALMS or XMSS (NIST SP 800-208), ML-DSAStateful hash-based signatures need careful key-state management
Symmetric encryptionAES-128AES-256For data that must stay secret long term
HashingSHA-256SHA-384 or SHA-512 where requiredSHA-256 remains widely accepted for general use

Decide where to use hybrid mode

Hybrid key exchange combines a classical algorithm with a post-quantum one, so a connection stays secure as long as either holds. It is the most common first step on the internet: the hybrid group X25519MLKEM768 is supported by major browsers and is a default in OpenSSL 3.5. Regulators differ on hybrid use, so check the guidance that applies to you.

Design for crypto-agility

Crypto-agility means changing algorithms through configuration rather than rewrites. Centralise cryptography in shared services and libraries, avoid hard-coding algorithms and key sizes, record what is used where, and make sure certificate and key lifecycles can be rotated quickly.

Build the roadmap

Turn the priorities into a sequenced, funded plan with milestones. A roadmap aligned with current regulator expectations might look like this:

ByMilestoneReference points
End of 2026Governance in place, inventory under way, first plan agreedEU roadmap first steps; ASD refined plan
2028Inventory complete, priorities set, migration of critical systems startedUK NCSC plan by 2028; ASD critical systems by end 2028
2030High-risk systems migratedEU high-risk use cases; ASD; HKMA readiness target; NIST IR 8547 deprecation
2031Highest-priority migrations completeUK NCSC
2035Migration completeUK NCSC; EU; NIST IR 8547 disallowance

Phase 5 · Remediate

LayerWhat to doWatch out for
Web and API traffic (TLS)Move to TLS 1.3 and enable hybrid key exchange (X25519MLKEM768) on servers, load balancers and CDNsOlder middleboxes that fail on larger handshake messages
Certificates and PKIPlan for post-quantum certificates, test CA and client support, shorten certificate lifecyclesLarger certificate chains; limited post-quantum certificate support so far
VPNs and remote accessAsk vendors for hybrid or post-quantum IKE and TLS support and upgrade pathsHardware appliances that need replacing
SSHUse current OpenSSH: version 9.0 (2022) made the hybrid sntrup761x25519 key exchange the default, and version 10.0 (2025) switched the default to ML-KEM-based mlkem768x25519Legacy servers and embedded devices
Code and firmware signingAdopt LMS/XMSS or ML-DSA for new devices and update channels firstDevices already in the field that cannot verify new signatures
HSMs and key managementConfirm firmware support for ML-KEM and ML-DSA and its certification statusLead times and validation schedules
Applications and librariesUpgrade cryptographic libraries (for example OpenSSL 3.5 or later), remove hard-coded algorithmsThird-party components you do not control
Data at restUse AES-256 for long-lived data and plan re-encryption of archives where neededBackup and archive formats
Suppliers and cloudAdd post-quantum and crypto-agility requirements to contracts and renewalsVendors without a published roadmap

Phase 6 · Validate and stay agile

  • Test performance. Post-quantum keys and signatures are larger: an ML-KEM-768 public key is 1,184 bytes and its ciphertext 1,088 bytes. Measure latency, bandwidth and handshake size.
  • Test compatibility. Check older clients, partners, middleboxes and devices before switching defaults.
  • Monitor. Keep the inventory current and track which systems still use vulnerable algorithms.
  • Watch the standards. New algorithms (such as FN-DSA and HQC) and new guidance will keep arriving.

Questions to ask your vendors

  • When will your product support ML-KEM and ML-DSA, and in which versions?
  • Will you support hybrid modes during the transition?
  • Can algorithms be changed by configuration?
  • How will installed devices and licences be upgraded, and at what cost?
  • Is your implementation validated, or scheduled for validation?

Common pitfalls

  • Treating post-quantum migration as a one-off technical project rather than a programme
  • Starting remediation before the inventory is good enough to set priorities
  • Forgetting suppliers, outsourced services and shadow IT
  • Underestimating hardware replacement lead times and budget
  • Switching defaults without testing older clients and partners

A 12-point checklist

  1. Executive owner named and board briefed
  2. Quantum risk on the risk register
  3. Scope agreed, including suppliers
  4. Discovery tooling selected
  5. Cryptographic inventory (CBOM) started
  6. Long-lived data identified
  7. Systems ranked by risk
  8. Algorithm and hybrid policy agreed
  9. Roadmap with milestones to 2030 and 2035
  10. Vendor roadmaps requested and tracked
  11. Pilot of hybrid TLS completed
  12. Crypto-agility requirements in new projects

Need help?

OptimaTech, an IBM Silver Partner, helps organisations in Hong Kong and New Zealand assess readiness, discover cryptography with IBM Quantum Safe tools, and plan and deliver migration. See our solutions.

Sources

  • NIST, FIPS 203, 204 and 205 (August 2024) and SP 800-208
  • NSA, Commercial National Security Algorithm Suite 2.0 (CNSA 2.0)
  • UK NCSC, Timelines for migration to post-quantum cryptography (March 2025)
  • NIS Cooperation Group, Coordinated Implementation Roadmap (June 2025)
  • ASD, Planning for post-quantum cryptography
  • OpenSSL 3.5.0 release, April 2025