Skip to content

Why quantum computing threatens today's cryptography

A sufficiently large, error-corrected quantum computer running Shor's algorithm would break RSA and elliptic-curve cryptography. Symmetric ciphers and hash functions are weakened rather than broken, and are addressed by increasing parameter sizes. No such machine exists today, and the migration is urgent for reasons of timescale rather than imminence.

The two algorithms that matter

Shor's algorithm solves integer factorisation and discrete logarithms efficiently on a quantum computer. RSA rests on the first and elliptic-curve cryptography on the second, so both fall to the same result. This is a break, not a weakening.

Grover's algorithm gives a quadratic speed-up on unstructured search, which effectively halves the security level of a symmetric cipher. AES-128 drops toward 64 bits of quantum security; AES-256 remains comfortable. This is why symmetric cryptography needs larger parameters rather than replacement.

What is affected, concretely

  • Broken by Shor — RSA encryption and signatures, ECDH, ECDSA, EdDSA, Diffie-Hellman. Every TLS handshake and almost every certificate hierarchy in production.
  • Weakened by Grover — AES, SHA-2, SHA-3. Move to 256-bit symmetric keys and longer hashes and the problem is addressed.
  • Unaffected — the vast majority of your application logic, which is where most breaches actually originate.

Where the machine actually is

Current quantum computers have limited qubit counts and error rates far above what running Shor's algorithm at cryptographically relevant scale would require. Published resource estimates for breaking RSA-2048 call for large numbers of high-quality logical qubits, each built from many physical ones.

Estimates have moved in both directions as error-correction techniques improve and as hardware progresses. Nobody credible claims a date, and anyone who does is telling you something about themselves.

Why the timescale still forces action

Three durations added together. How long your data must stay confidential, plus how long migration takes, plus how long your hardware lives. If that total exceeds the time until a capable machine exists, you are already late — this is Mosca's formulation and it is the honest framing.

For most organisations the migration duration and the hardware lifetime dominate, not the confidentiality horizon. Where the confidentiality horizon genuinely matters is a narrower case than the marketing suggests.