An insurer that advises on a pension transfer today, settles a personal injury claim for a child, or underwrites a life policy on the strength of a medical report is taking on a confidentiality obligation that will outlast most of the IT estate that handles it. The policy administration platform will be replaced. The broker portal will be rebuilt. The data, and the duty to protect it, stays.
That gap is what makes harvest now, decrypt later an insurance problem rather than a general one. An adversary who records encrypted traffic or copies encrypted files today does not need to read them today. If the key exchange protecting that data relies on RSA or elliptic-curve cryptography, a future cryptographically relevant quantum computer could recover it, and the data will still matter when that happens.
This article looks at why insurers sit at the long end of the confidentiality spectrum, what regulators have already written down about cryptography, and why a one-off cryptographic inventory is not enough. It then sets out what continuous monitoring looks like in practice, using Quantum Sentinel as the worked example. For the general mechanics of the attack, see our explainer on post-quantum threats.
Why insurers keep secrets for decades
Most retention schedules in an insurer are driven by how long a claim or complaint could still arrive, and in insurance that can be a very long time. Three examples from UK primary sources show the shape of it.
- Pension advice records. The FCA’s COBS 9.5.2R requires suitability records relating to a pension transfer, pension conversion, pension opt-out or FSAVC to be kept “indefinitely”. Records relating to a life policy or personal pension carry a five-year minimum.
- Personal injury claims. Under section 11 of the Limitation Act 1980, a personal injury action can be brought within three years of the cause of action or the claimant’s “date of knowledge”, if later. Where the injured person was a child at the time, section 28 delays the start of that period until the disability ends. Liability insurers therefore hold claim files, medical evidence and correspondence for long periods after the event.
- Health and underwriting data. Medical underwriting, protection claims and occupational health evidence are data concerning health, a special category under Article 9 of the UK GDPR. A diagnosis disclosed on a 2026 application is no less sensitive in 2040.
The legal periods are only part of it. A long-tail liability book, a closed life fund or an annuity portfolio keeps customer data live for as long as the liabilities run. Reinsurers receive much of the same data through ceded business. None of these have a neat end date that falls before the quantum risk window opens.
Where harvesting happens in an insurance estate
The final NIST IR 8547 (October 2026) is plain about the threat: adversaries “collect encrypted data now with the goal of decrypting it once quantum technology matures”, and the risk “underscores the necessity of acting immediately, especially for data with long-term sensitivity, such as government secrets or medical records.” It singles out TLS as “a primary target for ‘harvest now, decrypt later’ collection”.
The attack targets public-key key establishment. NIST says its symmetric standards are “significantly less vulnerable to known quantum attacks” and that it does not expect to transition away from them as part of the PQC migration. In an insurer, that points attention at the links where RSA or elliptic-curve key exchange protects long-lived data in transit:
- customer and broker portals, and the APIs behind aggregator and comparison sites;
- file transfers and APIs carrying bordereaux, claims and policy data to delegated authorities, TPAs, reinsurers and brokers;
- site-to-site VPNs to outsourced claims handlers and data centres;
- replication and backup traffic between data centres and cloud regions;
- email and document exchange carrying medical reports and claim files.
On timing, avoid both complacency and alarm. FINMA’s 2026 guidance states that cryptographically relevant quantum computers “do not yet exist”. The cost estimates keep falling, though: in May 2025 Craig Gidney estimated that a 2048-bit RSA integer could be factored in under a week by a machine with fewer than a million noisy qubits, against 20 million in an estimate he co-published in 2019. For data that must stay confidential into the 2040s, the arrival date matters less than the fact that interception is happening before it.
What regulators already say to insurers about cryptography
No regulator yet requires insurers to have finished a post-quantum migration. Several already expect them to manage cryptography as a changing risk, and some name quantum directly.
EU: DORA and its ICT risk RTS
DORA (Regulation (EU) 2022/2554) has applied since 17 January 2025, and Article 2(1)(n) lists “insurance and reinsurance undertakings” in scope, while Article 2(3) excludes the small undertakings covered by Article 4 of Solvency II. The supporting Commission Delegated Regulation (EU) 2024/1774 is the more specific text. Article 6(4) requires the encryption policy to include “provisions for updating or changing, where necessary, the cryptographic technology on the basis of developments in cryptanalysis”, and firms that cannot do so must adopt “mitigation and monitoring measures” and record why (Article 6(5)). Recital 9 tells firms to “remain abreast of relevant developments in cryptanalysis” and to follow “a flexible approach, based on risk mitigation and monitoring” to cryptographic threats, “including threats from quantum advancements.” Article 7(4) adds a register of certificates and certificate-storing devices, at least for ICT assets supporting critical or important functions, which must be kept up to date.
UK: the FCA, the Bank of England and the G7
The most direct statement so far is from the FCA, which supervises insurers. In a November 2026 article it says it expects firms to act now to have “a thorough and accurate understanding of the cryptography in place across systems and information” and to implement migration plans and “monitor, test and validate the changes”. It is trialling a supervisory self-assessment tool with an initial sample of firms, and says it will consult on cyber resilience guidance that includes its expectations for how firms should address PQC risks.
The Bank of England’s October 2025 report on AI, DLT and quantum computing warns that malicious actors “may be harvesting encrypted data today” and says the Bank has developed risk scenarios for post-quantum threats, including encrypted data harvesting, and is “piloting a supervisory briefing on quantum risks.” The G7 Cyber Expert Group roadmap (January 2026), published in the UK by HM Treasury, is explicitly not guidance or a regulatory expectation, but it points to 2035 as an overall target, suggests critical systems could be addressed in 2030 to 2032, and ends with a “Validation and Monitoring” phase of “continuous validation and ongoing improvement.” The NCSC’s migration timelines set the same 2035 end point.
New York and Switzerland
23 NYCRR Part 500 covers entities licensed under New York’s Insurance Law. The regulation itself does not mention quantum, but NYDFS’s September 2026 industry letter on Part 500 risk assessments lists “advances in quantum computing that may affect future cryptographic protections” among the emerging risks firms should consider. Section 500.15 requires an encryption policy that “meets industry standards”, and section 500.13(a), with compliance due from 1 November 2025 under the second amendment’s two-year transition, requires a “complete, accurate and documented asset inventory” with a defined frequency to update and validate it. In Switzerland, FINMA Guidance 05/2026 (9 July 2026), based on a survey that included insurance companies, recommends a PQC roadmap by mid-2027, a cryptographic inventory, and priority for data that needs long-term protection against “harvest now, decrypt later” attacks. It adds that an inventory “that is continuously updated to reflect the current situation contributes to the effectiveness of these measures.”
Why a one-off inventory goes stale
Read together, these texts describe a process, not a deliverable. DORA’s RTS asks for a policy that changes cryptography as cryptanalysis changes. FINMA wants a continuously updated inventory. NYDFS wants a defined frequency for validating the asset inventory. The FCA wants migration changes monitored, tested and validated. The G7 roadmap ends in monitoring.
The practical reason is drift. An insurer’s estate changes every week: a broker portal is re-platformed, a load balancer profile is copied from an old template, a TPA connection is stood up in a hurry for a catastrophe event, a certificate is reissued with a smaller key because that is what the request template said. Each change can quietly reintroduce the cryptography a migration has just removed. A point-in-time inventory taken for a board paper in March tells you little about September.
Monitoring also changes what the migration evidence looks like. Under Article 6(5) of the DORA RTS, a firm that cannot yet update a system has to record the mitigation and monitoring it adopted and explain why. That record is much easier to defend when it is produced by repeated discovery rather than a spreadsheet updated by hand. Our guide to the Cryptographic Bill of Materials covers how that inventory is structured.
How Quantum Sentinel supports continuous monitoring
Quantum Sentinel is an Enterprise Cryptographic Exposure Management platform and one of our technology partners. Its insurance solution page describes three things aimed at this problem: an Enterprise CBOM covering the cryptography behind policy, claims and actuarial systems, a Cryptographic Exposure Score to “[s]ee which long-retained data carries the greatest harvest-now-decrypt-later risk”, and a Quantum Readiness Score to “track progress over time”. Here is what that monitoring consists of, as QS describes it.
- Scheduled re-discovery. Its automated discovery runs on five-field cron schedules in a chosen timezone, as one-off runs, or on demand through the REST API. Each schedule has blackout windows and a maximum runtime, so scans stay out of change freezes and trading windows.
- Crypto drift events. When a rescan finds an entry has got worse (a weaker algorithm, a lower TLS version, a changed cipher suite, a smaller key or a higher quantum risk score), QS records a drift event with old and new values side by side. Alert rules can watch drift and policy violations, scoped by asset or asset group and minimum severity, and send them to Slack, Teams, email or a SIEM over syslog.
- A trend line. Daily snapshots of the risk-tier distribution, and posture monitoring that tracks the Quantum Readiness Score over time, give the board a direction of travel rather than a single reading.
- Per-entry HNDL context. Each CBOM entry carries a 0 to 100 quantum risk score weighted by the criticality, environment and exposure of the asset it sits on, plus a harvest-now-decrypt-later flag, NIST and CNSA status and a recommended replacement. Report types include quantum readiness, NIST IR 8547 and DORA.
Discovery is agentless and read-only by design, collects cryptographic metadata rather than keys or application data, and the platform runs on-premises or air-gapped, according to QS’s Security & Trust statement. The one exception to agentless is Kubernetes, where QS adds a read-only sidecar to the pod. That matters in an estate full of medical and financial data.
QS is also clear about the limits, and an insurer should plan around them:
- It is scheduled scanning at one-minute resolution, which QS says is “not continuous real-time monitoring”.
- Drift events are raised for degradation only. An improvement updates the inventory without an event, and there is no new-versus-removed delta report between scans.
- Software package discovery (OS packages, language dependencies, SBOMs) is not covered.
- Scoring is weighted by asset criticality, environment and exposure. Knowing that a given file transfer carries pension transfer records with indefinite retention is classification your team supplies.
A practical sequence for an insurer
- Map retention to systems. Take the retention schedule and tag each system and data flow with the longest-lived data it holds or carries. Pension transfer advice (where you gave it), long-tail liability claims and medical underwriting will rise to the top.
- Discover what is actually running. Build the cryptographic inventory from the estate rather than from architecture diagrams, starting with external portals, partner links, VPNs and replication paths.
- Rank by exposure, not by algorithm. An RSA key exchange on a public broker API carrying claims data matters more than the same algorithm on an isolated test box. Quantum Sentinel’s exposure assessment scores findings in business context for this reason.
- Set the monitoring cadence and alert rules. Decide how often each segment is rescanned, which drift events go to the SOC, and which policy rules (for example, banned algorithms or minimum key sizes) you hold the estate to.
- Bring third parties in. FINMA reminds firms that outsourced functions remain their responsibility. Ask TPAs, delegated authorities, brokers and reinsurers for their post-quantum plans and make crypto-agility a contract term.
- Report the trend. Give the board and the regulator a readiness trend, the drift caught and fixed, and a reasoned record of what cannot change yet.
If you are comparing tools for the discovery and monitoring layer, our guide to choosing a cryptographic exposure management platform sets out the questions to ask.
How Sinevis Can Help
Sinevis runs post-quantum readiness assessments that start from an insurer’s retention obligations and regulatory footprint, whether that is DORA, NYDFS Part 500, FINMA or PRA and FCA supervision. We work with Quantum Sentinel, one of our technology partners, to build the cryptographic inventory and keep it under scheduled monitoring, and we help your teams set the alert rules, classification and reporting that turn scan results into decisions.
For a quick external baseline, try our free PQC compliance check. To discuss your estate, contact the Sinevis team.
Sources
- NIST, NIST IR 8547, Transition to Post-Quantum Cryptography Standards (final) (October 2026)
- Financial Conduct Authority, FCA Handbook, COBS 9.5 Record keeping and retention periods for suitability records (last updated 1 October 2020)
- legislation.gov.uk, Limitation Act 1980, section 11 (revised text as of 17 November 2026)
- legislation.gov.uk, Limitation Act 1980, section 28 (revised text as of 17 November 2026)
- legislation.gov.uk, UK GDPR, Article 9
- Craig Gidney, arXiv, How to factor 2048 bit RSA integers with less than a million noisy qubits (arXiv:2505.15917) (21 May 2025)
- Official Journal of the European Union, Regulation (EU) 2022/2554 on digital operational resilience for the financial sector (DORA) (14 December 2022)
- Official Journal of the European Union, Commission Delegated Regulation (EU) 2024/1774 (RTS on ICT risk management framework) (13 March 2024)
- Bank of England, The Bank of England’s approach to innovation in AI, DLT and quantum computing (15 October 2025)
- HM Treasury / G7 Cyber Expert Group, G7 Cyber Expert Group Statement on Advancing a Coordinated Roadmap for the Transition to Post-Quantum Cryptography in the Financial Sector (13 January 2026)
- UK National Cyber Security Centre, Timelines for migration to post-quantum cryptography (20 March 2025)
- Financial Conduct Authority, Act now to be resilient in the quantum future (4 November 2026)
- New York State Department of Financial Services, Industry Letter: Guidance on How to Conduct and Use Risk Assessments Required by the DFS Cybersecurity Regulation (10 September 2026)
- New York State Department of Financial Services, 23 NYCRR Part 500, Second Amendment (effective 1 November 2023)
- FINMA, FINMA guidance on quantum computing (Guidance 05/2026) (9 July 2026)

