Post-Quantum Cryptography: What Is Actually Changing, and When
Which algorithms break, which do not, what NIST standardized, the published migration deadlines, and what a normal organization should do this year.

Most writing about quantum computing and encryption gets the shape of the problem wrong in the first paragraph, usually by implying that everything breaks at once on an unknown date. The real situation is more specific and considerably more actionable: a known subset of algorithms is affected, replacements have been standardized since August 2024, and two national agencies have published migration deadlines you can plan against. The uncertainty is about when a capable machine arrives, and that uncertainty has been deliberately removed from the planning question.
What breaks and what does not
There are two kinds of cryptography in everyday use and they are affected very differently.
Public-key cryptography is the problem. RSA, Diffie-Hellman and elliptic-curve algorithms rest on the difficulty of factoring large numbers or computing discrete logarithms. A sufficiently large, sufficiently reliable quantum computer running Shor's algorithm solves both efficiently. This is the mathematics underneath TLS key exchange, certificate signatures, SSH host and user keys, code signing, VPN handshakes and most document signing.
Symmetric cryptography is largely fine. AES, ChaCha20 and hash functions like SHA-256 face a different algorithm, Grover's, which offers a square-root speedup on brute-force search. The practical answer is to use larger keys. AES-256 remains comfortable. This is why "quantum breaks all encryption" is wrong: the bulk encryption protecting your data at rest and your file transfers is not in the collapse category.
So the migration is a public-key migration. That narrows the work enormously, and it also explains why the work is hard: public-key cryptography is embedded in protocols, certificates, hardware roots of trust and device firmware rather than sitting in one configuration file.
The standards exist now
NIST ran a multi-year public competition and published three final standards in August 2024.
- FIPS 203, ML-KEM, derived from CRYSTALS-Kyber. A key encapsulation mechanism, which is what replaces Diffie-Hellman and RSA key transport in protocols like TLS. This is the one that matters most and soonest.
- FIPS 204, ML-DSA, derived from CRYSTALS-Dilithium. The general-purpose digital signature replacement.
- FIPS 205, SLH-DSA, derived from SPHINCS+. A hash-based signature scheme with larger signatures and a different and very conservative security basis, intended as a hedge in case lattice problems turn out to be weaker than believed.
NIST has continued work on additional signature schemes to diversify the mathematical foundations. The practical reading for an organization is that the algorithms you will deploy are settled for key exchange and for mainstream signatures, and that waiting for more standards is not a reason to delay inventory work.
Two engineering consequences are worth knowing before you plan. Post-quantum keys and signatures are substantially larger than their elliptic-curve equivalents, which matters for constrained devices, for protocols with packet size limits, and for anything that embeds many signatures. And some of these schemes have different performance characteristics, so "drop in replacement" is optimistic in embedded and high-throughput contexts.
The timelines that have actually been published
This is where speculation can be replaced by citation.
NIST's draft internal report on the transition to post-quantum cryptography standards sets out a deprecation schedule for classical public-key algorithms: the 112-bit security level algorithms, which includes RSA-2048 and comparable elliptic curve parameters, deprecated after 2030 and disallowed after 2035.
The UK National Cyber Security Centre has published migration timelines with three checkpoints: by 2028, complete a discovery exercise and build a migration plan; by 2031, complete the highest-priority migration work and refine the plan for the rest; by 2035, complete migration across all systems.
CISA's Post-Quantum Cryptography Initiative and its quantum-readiness guidance say the same thing in a different register, with emphasis on inventory first.
Notice what these documents have in common. None of them predicts when a cryptographically relevant quantum computer will exist. All of them set deadlines based on how long migration takes. That is the correct way to plan under this kind of uncertainty, and it is also permission to stop arguing about timelines in your own organization: the deadline is 2035 regardless of what the hardware does, because that is what the guidance you will be audited against says.
What is already happening without you
A reassuring amount of this is being handled by the ecosystem. Major browsers and large content networks have deployed hybrid key exchange, combining a classical algorithm with ML-KEM so that the session is secure if either component holds. A substantial share of ordinary web traffic now negotiates it by default. Operating systems, SSH implementations and TLS libraries have been adding support over several releases.
This means that for a typical organization whose sensitive traffic is public web and standard protocols with up-to-date software, a meaningful part of the key exchange problem resolves itself through patching. That is an argument for keeping software current, which was already the argument. Our note on keeping business devices updated covers the unglamorous version.
Where it does not resolve itself: anything with a long-lived embedded root of trust, bespoke protocols, hardware security modules, payment terminals, operational technology, VPN appliances near end of support, and any product whose vendor has no published post-quantum roadmap.
What to do this year
For an organization that is not a government, a bank or a critical infrastructure operator, the honest list is short and mostly consists of preparation rather than deployment.
- Build a cryptographic inventory. Where is public-key cryptography used, in which products, for what, with which algorithms and key sizes, and who owns the system. This is the task every agency puts first and the one everyone underestimates. Start with internet-facing services, VPNs, code signing and anything handling long-lived secrets.
- Identify long-confidentiality data. Which of your data would still be damaging if decrypted in ten or twenty years. That subset is where key exchange migration is urgent rather than scheduled, for reasons covered in harvest-now-decrypt-later.
- Put post-quantum questions into procurement. Ask every vendor of a security-relevant product for their post-quantum roadmap and the version in which they will support FIPS 203. The answers tell you which suppliers will be a problem in 2031. The assessment framing in our third-party risk guide applies directly.
- Work toward crypto agility. Remove hardcoded algorithm choices, move cipher selection into configuration, and make certificate and key management capable of handling new key types. This has value independent of quantum computing, because it is what you need the next time an algorithm is deprecated for ordinary reasons.
- Keep patching. Much of your key exchange migration will arrive in library and operating system updates. The organizations that struggle in 2031 will be the ones running software that stopped receiving them.
What not to do: buy a product promising quantum-safe security without checking which standardized algorithm it implements, deploy a non-standard scheme, or treat this as urgent enough to displace multi-factor authentication, backups and patching. An organization without tested backups has a far more likely catastrophe ahead of it than a cryptographic one.
What is speculation, and what would settle it
Three claims circulate that deserve labels.
The claim that quantum computers will break encryption by a specific near year. Speculation, with no reliable basis. Nobody outside a well-informed guess can date this, and the people who publish standards pointedly decline to. Treat any specific year as marketing.
Prediction, labeled as such: no publicly known quantum computer will factor a 2048-bit RSA key before 2031. The reasoning is the gap between current qubit counts and error rates and the millions of high-quality operations required. Falsification condition: a peer-reviewed or independently reproduced demonstration of factoring an RSA key of cryptographically relevant size, or credible disclosure of one. A demonstration on a small, specially chosen number does not count, and several have been misreported as if they did.
Prediction: the 2035 deadline will drive more real work than the hardware will. Most organizations that complete this migration will do so because a customer questionnaire, a regulator or a government procurement rule required it, not because the threat materialized. Falsification condition: a credible public demonstration of quantum factoring at scale before 2033, which would compress every timeline in this article.
Checking something suspicious? Paste the message or link into CyberWatch AI Scan for a free check.
Keeping it in proportion
Post-quantum migration is a large, slow, genuinely necessary infrastructure project with published deadlines and finished standards. It is not an emergency, and it is not a reason to buy anything exotic. The organizations that handle it well will be the ones that already know where their cryptography lives, keep their software current and ask their suppliers hard questions, which describes a competent security program doing its ordinary work. If yours is not there yet, the sequencing in our small business cybersecurity guide is a better use of the next six months than anything on this page, and the encryption basics in what end-to-end encryption does and does not protect are worth having straight first.
Frequently asked questions
Does quantum computing break all encryption?
No. It breaks public-key cryptography based on factoring and discrete logarithms, which means RSA, Diffie-Hellman and elliptic curve. Symmetric encryption such as AES and hash functions such as SHA-2 are affected much less, and doubling symmetric key sizes is generally considered sufficient.
What did NIST actually standardize?
In August 2024 NIST published FIPS 203 for key encapsulation (ML-KEM, from CRYSTALS-Kyber), FIPS 204 for digital signatures (ML-DSA, from CRYSTALS-Dilithium) and FIPS 205 for stateless hash-based signatures (SLH-DSA, from SPHINCS+). These are finished standards, not drafts.
When do I have to migrate?
Published guidance gives dates. NIST's draft transition report sets out deprecation of 112-bit classical public-key algorithms around 2030 and disallowing them after 2035. The UK NCSC's timelines expect discovery and planning by 2028, high-priority migration by 2031 and completion by 2035.
Does a quantum computer capable of this exist?
Not publicly, and not close. Current machines are far from the scale and error rates required to run Shor's algorithm against real key sizes. Nobody credible can tell you the year, which is exactly why the official timelines are driven by migration effort rather than by a predicted breakthrough.
Why migrate now if the threat is years away?
Two reasons. Migration of key exchange protects data being recorded today against decryption later, which is the harvest-now-decrypt-later problem. And the work itself takes years in large estates, because much of it is in embedded devices and vendor products you do not control.
What is crypto agility?
The ability to change algorithms without rebuilding a system. In practice it means no hardcoded algorithm choices, configuration-driven cipher selection, certificate and key management that can handle new types, and a documented inventory of where cryptography is used.
Do I need to replace my TLS certificates now?
Not urgently. Signatures protect authenticity at the time of use, so a certificate that expires before a quantum computer exists is not retroactively broken. Key exchange is the priority, because recorded traffic can be decrypted later.
What is hybrid key exchange?
Combining a classical algorithm with a post-quantum one so the session is secure if either holds. Major browsers and servers already negotiate hybrid key exchange, which means a meaningful share of web traffic is already protected without anyone doing anything.
Sources
- Post-Quantum Cryptography, NIST Computer Security Resource Center
- FIPS 203, Module-Lattice-Based Key-Encapsulation Mechanism Standard, NIST
- FIPS 204, Module-Lattice-Based Digital Signature Standard, NIST
- FIPS 205, Stateless Hash-Based Digital Signature Standard, NIST
- IR 8547, Transition to Post-Quantum Cryptography Standards, NIST
- Timelines for migration to post-quantum cryptography, UK National Cyber Security Centre
- Quantum-Readiness: Migration to Post-Quantum Cryptography, CISA
- Post-Quantum Cryptography Initiative, CISA
- Software Bill of Materials (SBOM), CISA


