France Races to Replace RSA-2048 as “Harvest Now, Decrypt Later” Threat Forces Post-Quantum Shift

French government agencies and major industries are stepping up preparations for a future in which a sufficiently powerful quantum computer could break RSA-2048, one of the world’s most widely used encryption standards.

In 2026, France is organizing what officials and companies describe as a heavy technical transition toward post-quantum cryptography—new algorithms intended to withstand quantum attacks—across everything from state services and banks to telecom networks. The concern isn’t confined to academic research: it directly affects the security of communications and the protection of sensitive data over multi-year time horizons.

The risk, French stakeholders warn, is already operational because encrypted information sent today can be intercepted, stored, and decrypted later if quantum computing advances far enough. The strategy has a name: “harvest now, decrypt later,” and it is driving organizations that must keep secrets for long periods—government, defense, health care, and industry—to act sooner rather than later.

RSA-2048 under pressure as “harvest now, decrypt later” shifts the threat to today

RSA-2048 encryption remains broadly used to protect communications, sign software, and establish secure sessions. Its strength comes from the difficulty of factoring a very large integer. But a powerful enough quantum computer could use Shor’s algorithm to drastically reduce that difficulty.

At this stage, the operational question in France is not pinpointing the day a break happens. It is treating quantum decryption as a structural risk that has to be built into security management now.

The central danger is the lifespan of data. Industrial secrets, medical records, diplomatic exchanges, and court archives can retain value for years. If they are encrypted with schemes that may become vulnerable, an interception today can become exploitable later. “Harvest now, decrypt later” effectively moves the threat into the present, because an attacker only needs to capture and store traffic—even without the ability to read it immediately.

That logic forces organizations to sort assets by how long they must remain confidential. Long-lived data becomes the top priority in migration plans, especially in regulated sectors where retention and traceability stretch across many years. The issue also hits authentication and digital signatures: if a signature can be broken retroactively, it can call into question evidence, software updates, or control procedures.

France’s planners also emphasize that the technical tradeoffs go beyond swapping one algorithm for another. Key sizes, CPU cost, network latency, energy use, and compatibility with older equipment all shape what is feasible. Heavier cryptography can degrade critical applications, particularly when millions of connections must be established.

The broad push is toward hybrid strategies during a transition period—systems that combine classical and post-quantum mechanisms to reduce risk while limiting compatibility breaks. The challenge is deploying those solutions at scale without creating new weak points, since added complexity can increase configuration errors.

ANSSI and French agencies map a phased migration to post-quantum standards

In France, preparation is being structured around public-sector cybersecurity actors and government agencies that run sensitive systems. The goal is to define a credible migration path to post-quantum cryptography while accounting for budget constraints, existing infrastructure, and the need to keep services running.

A key objective is avoiding fragmented migrations in which each organization picks its own non-interoperable components. Governance starts with a detailed inventory of cryptographic usage—something many organizations lack because cryptography is often buried inside libraries, network appliances, smart cards, or cloud services.

The first step, as described in the French approach, is a census: which algorithms are used, where, for what data, and with what retention periods. Without that visibility, any plan remains theoretical.

Agencies also need to separate external traffic—such as public-facing portals—from internal flows like inter-agency exchanges or links with operators. Migration can begin with pilot environments and expand as technical maturity improves, a gradual approach meant to surface performance impacts, interoperability failures, and deployment risks.

Identity management and signatures are another pressure point. If certificates and public key infrastructures rely on RSA-2048 or other exposed algorithms, France’s agencies must anticipate renewing chains of trust. That means changes on servers and on the client side—potentially across millions of devices or applications. In the public sector, the diversity of hardware, operating systems, and software versions complicates planning.

Industrial control systems and embedded devices pose a particularly sensitive challenge. Equipment installed for 10 or 20 years cannot be updated like a web browser. Post-quantum migration planning therefore includes workarounds such as network segmentation, encrypted edge tunnels, gateways, or gradual upgrades during maintenance cycles—an approach framed as pragmatic and driven by priorities and credible attack scenarios.

French industry tests NIST-track algorithms across networks and cloud systems

The post-quantum transition is also playing out in the private sector, where major digital service and infrastructure providers must guarantee confidentiality and authenticity. Telecom companies, hosting providers, security software vendors, and systems integrators are evaluating families of algorithms that are standardized or nearing standardization through international selection efforts, including work associated with the U.S. National Institute of Standards and Technology (NIST).

Pilot deployments often focus on key-establishment mechanisms (KEMs) and digital signatures intended to replace or complement classical methods.

The biggest difficulty reported in pilots is performance. Some post-quantum algorithms increase key sizes or the size of exchanged messages, which can weigh on heavily used protocols like TLS. On mobile networks or high-latency environments, a few extra bytes may seem minor, but across millions of daily connections the impact becomes measurable in bandwidth and server load.

Industry also has to contend with software supply chains. Cryptographic libraries, hardware accelerators, HSMs (hardware security modules), and runtime environments do not all support the same primitives. Choosing an algorithm is not enough: implementations must be robust, audited, and resistant to side-channel attacks. In some cases, the question is less whether it can be done than how much trust can be placed in code that will run at massive scale.

Cloud providers and security vendors are frequently pursuing hybrid approaches that combine a classical mechanism with a post-quantum one to preserve compatibility and reduce dependence on any single algorithm. That strategy comes with costs—more complex configuration and harder incident diagnosis—but it offers a safety net while not all customers are ready to switch.

Interoperability is becoming central across supply chains. A company may update its own portal but still depend on a partner that has not migrated, or on network equipment that blocks a protocol extension. French industrial players—often embedded in European and global supply chains—are therefore aligning roadmaps around shared standards to avoid security plans colliding with commercial and technical constraints.

Banks, health care, and defense prioritize long-lived data and archival systems

Sectors that handle sensitive information over long periods are among the first affected. Banks and insurers must protect transaction histories, proof of operations, and identity data. Health care manages medical records whose value can extend for decades. Defense and certain strategic industries must preserve operational secrets and technical plans. In each case, exposure to “harvest now, decrypt later” is presented as a reason to act without waiting for a definitive “break date.”

Data mapping, in this framework, goes beyond labeling information as sensitive or not. It must incorporate how long a leak would matter, legal impact, reputational risk, and how an adversary could exploit it. An archive database might be accessed less often than a transactional app, yet be more critical if it contains persistent identifiers or information that can reconstruct behavior.

Organizations are also being pushed to scrutinize backup and archiving systems. Stronger transport encryption is not enough if archives encrypted with vulnerable algorithms are stored for years. That leads to strategies for progressive re-encryption under tight constraints: high volumes, limited maintenance windows, and traceability requirements. Costs are not only hardware-related; they also come from engineering, testing, and auditing.

Long-term software and document signatures are another critical issue. Organizations need to prove the integrity of a document over time. If a signature based on RSA-2048 is weakened, the evidentiary value of a signed document can be challenged. Sealing, timestamping, and evidentiary preservation mechanisms must therefore incorporate post-quantum considerations, including renewable chains of trust.

Finally, France’s approach stresses crisis exercises. Security and risk leaders are urged to consider scenarios in which a cryptographic vulnerability is announced or a technological leap suddenly reduces safety margins. Even if an operational quantum attack has not been publicly demonstrated, an organization could be forced into rapid migration if risk perception shifts. Business continuity planning benefits from switch-over options, clear priorities, and internal or vendor capacity to deploy updates at volume.

Frequently asked question

Why is post-quantum cryptography becoming urgent in 2026? Because data encrypted today can be intercepted and stored to be decrypted later if quantum capabilities progress. Organizations that must protect long-lived information—health care, finance, or the state—have an incentive to plan a gradual migration rather than wait for a technological break.