# ML-KEM

Module-Lattice Key Encapsulation Mechanism (FIPS 203), formerly known as Kyber. The NIST-standardized post-quantum replacement for ECDH and RSA key exchange.

ML-KEM is a KEM (Key Encapsulation Mechanism), not a NIKE (Non-Interactive Key Exchange). The distinction matters: KEMs require a round trip (one party encapsulates a shared secret to the other's public key, the other decapsulates), while NIKEs let both parties derive a shared secret from each other's public keys without interaction. Protocols that depend on NIKEs (some messaging protocols) need architectural changes to use ML-KEM.

## How it works

Based on the Module Learning With Errors (MLWE) problem. The sender generates a ciphertext that encapsulates a shared secret using the recipient's public key. The recipient decapsulates with their private key. An eavesdropper would need to solve MLWE — finding a short vector in a noisy lattice — which no known quantum algorithm can do efficiently.

Three security levels:
- **ML-KEM-512** — roughly equivalent to AES-128 security
- **ML-KEM-768** — roughly equivalent to AES-192 security (recommended default)
- **ML-KEM-1024** — roughly equivalent to AES-256 security

## Deployment status

Chrome and other browsers already ship ML-KEM in hybrid mode (X25519 + ML-KEM-768) for TLS 1.3 key exchange. Two years of deployment (starting as Kyber) has built significant confidence in the algorithm's correctness and the underlying math.

Per [[crqc-timeline]], any non-PQ key exchange should now be treated as a potential active compromise vector.