Skip to content

PQC migration timelines, NIST, ANSSI, BSI and CNSA 2.0 compared

NIST published the first post-quantum standards in August 2024 and has proposed deprecating current public-key algorithms around 2030 and disallowing them around 2035. CNSA 2.0, ANSSI and BSI each set their own expectations, and they do not align — the applicable one depends on your sector and jurisdiction.

Dates and statuses change. Check the current published guidance from the authority that applies to you before making a commitment on the strength of a summary.

NIST

FIPS 203, 204 and 205 were finalised in August 2024. NIST has separately proposed a transition in which today's public-key algorithms are deprecated around 2030 and disallowed around 2035, in draft guidance on transitioning cryptographic algorithms.

The word to hold onto is proposed. It signals direction clearly and it is the anchor most other timelines are read against.

CNSA 2.0

The NSA's suite for US national security systems, which sets out when PQC becomes required across categories of equipment, with software and firmware signing moving earliest and general networking later. It is directive for the systems it covers.

For vendors it matters beyond its formal scope, because equipment sold into national security systems has to meet it, and that shapes product roadmaps for everyone.

ANSSI

The French authority has set out a phased position. In practical terms it means qualification considerations arriving from 2027 and purchasing expectations by 2030.

It is worth stating precisely what that is and is not. It is guidance shaping qualification and procurement, not a universal legal prohibition on classical cryptography, and a vendor describing it as a ban is overstating it.

BSI

The German authority has published its own recommendations, has been notably supportive of hybrid constructions during the transition, and maintains its own evaluation methodology for the components underneath — AIS 20/31 for random number generators.

What to do with the disagreement

Take the earliest date that applies to you as your planning date, then assume the details will change again before you get there. That combination — act now, expect revision — is the practical case for building for change rather than for a target.

The first concrete step is not choosing an algorithm. It is knowing where cryptography is used at all.