On 3 September 2026, under the French G7 presidency, France's national cybersecurity agency (ANSSI) and its counterparts in the G7 cybersecurity working group published a joint document entitled Preparing for the Post-Quantum Era: A Call to Action. Signatories: seven national agencies, supported by the European Commission and ENISA as observers. This document is not one more set of good practice guidelines. It is a coordinated policy instruction, addressed simultaneously to governments and private organisations, which sets deadlines that many teams are not yet aware of.
What this document changes compared with existing recommendations
Since 2022, ANSSI has published recommendations for a gradual migration to post-quantum cryptography. Since August 2024, the NIST standards FIPS 203, 204 and 205 have been the international technical reference. These documents set the technical framework. The G7 document of 3 September does something different: it raises the subject to the level of a governance obligation, with official milestones.
Two deadlines deserve attention. From 2027, ANSSI will stop certifying security products that do not incorporate encryption resistant to quantum attacks. For organisations that rely on certified products in their public tenders or cyber insurance contracts, this date is already part of their current purchasing cycle. From 2030, post-quantum security will become mandatory for certain categories of security products on the European market. NIST, for its part, sets 2035 as the horizon for retiring vulnerable algorithms (RSA, ECC, ECDH) from critical government systems.
Between the publication of the NIST standards in 2024 and this September 2026 document, the question has changed category. It is no longer only technical. It is budgetary, contractual and regulatory.
The five G7 priorities and their practical scope
The document identifies five priority areas for action.
Awareness. The document asks that quantum risk be treated as an active risk within risk management cycles, rather than as a future threat to be reviewed in five years' time. "Harvest now, decrypt later" (HNDL) attacks have been active for several years: data encrypted today with RSA or ECC is collected so that it can be decrypted once quantum computers reach sufficient power. Data whose confidentiality must hold beyond 2030 is already at risk.
National strategies and procurement. G7 member governments commit to integrating PQC requirements into their public procurement rules. For State suppliers or subcontractors in regulated sectors, this will translate into eligibility criteria in tender documents. This priority is the most directly actionable for private companies: if your main client is public or heavily regulated, you will be subject to these requirements by contract before being subject to them by law.
Research and deployment. The G7 asks for PQC implementations to be tested in real systems, not only in laboratories. Hybrid schemes, which combine classical and post-quantum algorithms, are the recommended approach during the transition period: they maintain security if the PQC implementation is compromised, and provide protection if a quantum computer emerges.
Public-private coordination. Agencies commit to sharing information on critical cryptographic dependencies with publishers and operators. For organisations, this means mapping the third-party products they rely on and tracking their PQC roadmaps: a provider that does not publish a PQC roadmap is a continuity risk from 2027.
Integration into cybersecurity requirements. The document asks for crypto-agility and PQC support to be included among the evaluation criteria when purchasing security solutions. In practice: your next tender for a VPN, an EDR or a PKI solution should include a FIPS 203/204/205 support criterion.
The cryptographic inventory: the first non-negotiable action
Before any migration, any purchasing decision or any transition plan, one question must be answered: where does your organisation currently use algorithms vulnerable to quantum attacks? The honest answer, in most organisations, is that they do not know precisely.
RSA, Diffie-Hellman and elliptic curve algorithms (ECDH, ECDSA) are present in layers that are rarely audited: TLS certificates on internal servers, VPN sessions, SSH protocols, email signatures, authentication tokens, code signing certificates. Third-party dependencies add to this: every SaaS product or embedded software uses its own cryptographic layer, often invisible in commercial specifications.
The G7 explicitly asks organisations to start with this inventory: identify systems that contain critical data, map cryptographic assets, establish dependencies. Without this inventory, it is not possible to prioritise the migration or to assess real exposure to HNDL attacks. The technical foundations of this work are described in our article on the fundamentals of post-quantum migration.
A concrete example: buying a VPN in October 2026
An IT department that launches a tender today to renew its VPN infrastructure has a deployment horizon of 12 to 18 months and a product lifecycle of five years. If it does not include a PQC criterion in its specifications, it will find itself in 2028 with a product that is potentially incompatible with ANSSI's 2027 certification requirements, and will have to plan for a further replacement. The cost of leaving out a PQC criterion in a 2026 tender is higher than the cost of including one.
This calculation applies to any acquisition of a security solution with a lifecycle of more than two years: PKI, HSMs, electronic signature solutions, authentication platforms.
The sequence to follow now
The G7 document recommends a four-step approach. The first is the cryptographic inventory described above. The second is prioritisation by data lifespan: information whose confidentiality must hold for ten years is more urgent than ephemeral transactions. The third is updating purchasing criteria to include PQC support and crypto-agility. The fourth is testing hybrid implementations in non-critical systems, to gain operational experience before migrating sensitive systems.
The G7 call of 3 September 2026 marks a turning point: post-quantum migration is no longer a subject for cryptography research teams. It is entering governance decisions, budget cycles and contracts. Organisations that treat this as a signal have two to three years ahead of them to build a solid position. Those that wait for the regulatory constraint of 2030 will migrate under pressure, with less room to do it well.