Post-quantum cryptography¶

WaveLedger uses NIST-standardized post-quantum primitives for every piece of in-band cryptography. There is no classical-crypto fallback.

Why¶

Shor's algorithm, run on a sufficiently large quantum computer, breaks all widely-deployed asymmetric primitives in polynomial time. That includes:

RSA — used in TLS, code signing, and most certificate authorities
ECDSA — used in Bitcoin, Ethereum, and most other chains
Ed25519 — used in many newer chains and SSH
Diffie-Hellman (classical + elliptic-curve variants)

Symmetric primitives (AES, SHA-3) are weakened by Grover's algorithm, but only quadratically — a 256-bit key becomes "128-bit-equivalent", which remains infeasible to brute-force. NIST classifies AES-256 and SHA3-512 as PQ Category 5.

A chain that must keep proving past transactions valid 40 years from now cannot afford to use cryptography breakable by a quantum computer 20 years from now.

The primitive set¶

Use	Primitive	NIST standard
Transaction signatures	ML-DSA-87	FIPS 204
Key encapsulation (planned: encrypted handoff, p2p session keys)	ML-KEM-1024	FIPS 203
Hashes (block hash, merkle root, tx id, address derivation)	SHA3-512	FIPS 202
Wallet backup encryption	AES-256-GCM + Argon2id	SP 800-208 (Cat 5)

ML-DSA, ML-KEM, and SHA3 are all NIST-standardized as of 2024-25. There are no proprietary primitives in the chain.

Sizes¶

PQ keys and signatures are larger than classical equivalents.

Quantity	Bytes	vs Ed25519
ML-DSA-87 public key	2,592	81× larger
ML-DSA-87 signature	4,627	72× larger
ML-KEM-1024 public key	1,568	—
ML-KEM-1024 ciphertext	1,568	—

A transaction with one signature is ~5 KB on the wire. Block size is not artificially capped; the effective limit is MAX_TRANSACTIONS_PER_BLOCK = 100 (network parameter, raisable on mainnet).

Scope and caveats¶

Symmetric crypto is not post-quantum replaced — it is post-quantum safe by virtue of being symmetric. AES-256 has 128 bits of quantum security; SHA3-512 has 256 bits of quantum collision resistance.
Side-channel resistance is not in scope. ML-DSA implementations have published timing-attack mitigations; the reference Python implementation (dilithium-py) is not hardened.
Long-term forward secrecy of historical data depends on the ciphers remaining unbroken. If ML-DSA is broken in 2055, every transaction ever signed becomes forgeable in retrospect. The same caveat applies to every signature scheme ever invented.

Future extensions¶

Hardware-backed signing.

Crypto agility¶

The chain provides a runtime signature scheme registry — see verify_sig(scheme, ...) in Fourier and SLH-DSA-SHA2-128s as the second registered scheme. Contracts select which scheme to require, and new schemes can be added without a hard fork.

The intent: if any single scheme is broken or deprecated, the chain keeps running on the others during migration.