Chapter 3: NIST PQC Standards¶
3.1 Overview of Finalized Standards¶
NIST finalized the first three Post-Quantum Cryptography standards in August 2024 after an eight-year standardization process. These represent the mandatory baseline for quantum-safe cryptography in U.S. government systems and are being adopted globally.
The Four Published NIST PQC Standards
-
FIPS 203 (August 2024): ML-KEM — Module-Lattice Key Encapsulation Mechanism
-
FIPS 204 (August 2024): ML-DSA — Module-Lattice Digital Signature Algorithm
-
FIPS 205 (August 2024): SLH-DSA — Stateless Hash-Based Digital Signature Algorithm
-
HQC (March 2025): Code-based KEM — 5th algorithm, draft standard pending
A draft standard for FIPS 206 (FN-DSA, based on FALCON) is expected in 2025–2026.
3.2 FIPS 203: ML-KEM¶
3.2.1 Specification Summary¶
FIPS 203 standardizes ML-KEM, derived from CRYSTALS-Kyber. Key design choices:
-
Module dimension \(k \in \{2, 3, 4\}\) (for ML-KEM-512/768/1024).
-
Ring \(R_q = \mathbb{Z}_{3329}[X]/(X^{256}+1)\); \(q = 3329\) chosen for efficient NTT.
-
Uses Kyber.CPA-PKE as an internal primitive, then applies the Fujisaki-Okamoto (FO) transform to achieve IND-CCA2 security.
-
Key derivation uses SHAKE-256 (XOF) and SHA3-512 throughout.
-
Encapsulation key (public key) and decapsulation key (private key) renamed from Kyber conventions.
3.2.2 Changes from Kyber to ML-KEM¶
FIPS 203 introduced several changes from the CRYSTALS-Kyber Round 3 submission:
| Aspect | Kyber (R3) | ML-KEM (FIPS 203) |
|---|---|---|
| Terminology | pk, sk, ct | ek (encaps key), dk (decaps key), c |
| Algorithm names | Kyber-512/768/1024 | ML-KEM-512/768/1024 |
| Domain separation | Less explicit | Explicit domain separation added |
| Implicit rejection | Randomized | Deterministic (security fix) |
| Decaps check | Conditional re-encryption | Always re-encrypt (side-channel fix) |
3.2.3 ML-KEM Full Specification¶
| Parameter | ML-KEM-512 | ML-KEM-768 | ML-KEM-1024 | ||||
|---|---|---|---|---|---|---|---|
| NIST Security Level | 1 | 3 | 5 | ||||
| Module dimension \(k\) | 2 | 3 | 4 | ||||
| Modulus \(q\) | 3329 | 3329 | 3329 | ||||
| Polynomial degree \(n\) | 256 | 256 | 256 | ||||
| CBD parameter \(\eta_1\) | 3 | 2 | 2 | ||||
| CBD parameter \(\eta_2\) | 2 | 2 | 2 | ||||
| Compression \(d_u\) | 10 | 10 | 11 | ||||
| Compression \(d_v\) | 4 | 4 | 5 | ||||
| Encapsulation key (B) | 800 | 1184 | 1568 | ||||
| Decapsulation key (B) | 1632 | 2400 | 3168 | ||||
| Ciphertext (B) | 768 | 1088 | 1568 | ||||
| Shared secret (B) | 32 | 32 | 32 | ||||
| Security (classical bits) | \(\approx\)118 | \(\approx\)180 | \(\approx\)247 | ||||
| Security (quantum bits) | \(\approx\)107 | \(\approx\)164 | \(\approx\)226 |
3.3 FIPS 204: ML-DSA¶
3.3.1 Specification Summary¶
FIPS 204 standardizes ML-DSA (based on CRYSTALS-Dilithium), a Fiat-Shamir signature scheme with rejection sampling to prevent secret leakage.
Key design choices:
-
Modulus \(q = 8380417 = 2^{23} - 2^{13} + 1\) (NTT-friendly prime).
-
Polynomial degree \(n = 256\), supporting efficient NTT.
-
Rejection sampling ensures signatures reveal no information about the secret key.
-
Message hashing via SHAKE-256; deterministic signing by default in FIPS 204.
-
Three variants: ML-DSA-44 (Level 2), ML-DSA-65 (Level 3), ML-DSA-87 (Level 5).
3.3.2 ML-DSA Full Parameter Table¶
| Parameter | ML-DSA-44 | ML-DSA-65 | ML-DSA-87 |
|---|---|---|---|
| NIST Security Level | 2 | 3 | 5 |
| \((k, l)\) | \((4, 4)\) | \((6, 5)\) | \((8, 7)\) |
| Modulus \(q\) | 8380417 | 8380417 | 8380417 |
| \(\gamma_1\) | \(2^{17}\) | \(2^{19}\) | \(2^{19}\) |
| \(\gamma_2 = (q-1)/\alpha\) | \((q-1)/88\) | \((q-1)/32\) | \((q-1)/32\) |
| \(\tau\) (challenge weight) | 39 | 49 | 60 |
| \(\beta = \tau \cdot \eta\) | 78 | 196 | 120 |
| \(\omega\) (max hints) | 80 | 55 | 75 |
| Verification key (B) | 1312 | 1952 | 2592 |
| Signing key (B) | 2528 | 4000 | 4864 |
| Signature (B) | 2420 | 3309 | 4627 |
| Security (classical bits) | \(\approx\)122 | \(\approx\)178 | \(\approx\)247 |
3.4 FIPS 205: SLH-DSA¶
3.4.1 Specification Summary¶
FIPS 205 standardizes SLH-DSA, based on SPHINCS+. It is a stateless hash-based signature scheme — the most conservative of the three, relying only on hash function security.
3.4.2 SLH-DSA Full Parameter Sets¶
| Variant | Level | \(n\) | pk (B) | sk (B) | sig (B) | |
|---|---|---|---|---|---|---|
| SLH-DSA-SHA2-128s | 1 | 16 | 32 | 64 | 7856 | |
| SLH-DSA-SHA2-128f | 1 | 16 | 32 | 64 | 17088 | |
| SLH-DSA-SHAKE-128s | 1 | 16 | 32 | 64 | 7856 | |
| SLH-DSA-SHAKE-128f | 1 | 16 | 32 | 64 | 17088 | |
| SLH-DSA-SHA2-192s | 3 | 24 | 48 | 96 | 16224 | |
| SLH-DSA-SHA2-192f | 3 | 24 | 48 | 96 | 35664 | |
| SLH-DSA-SHAKE-192s | 3 | 24 | 48 | 96 | 16224 | |
| SLH-DSA-SHAKE-192f | 3 | 24 | 48 | 96 | 35664 | |
| SLH-DSA-SHA2-256s | 5 | 32 | 64 | 128 | 29792 | |
| SLH-DSA-SHA2-256f | 5 | 32 | 64 | 128 | 49856 | |
| SLH-DSA-SHAKE-256s | 5 | 32 | 64 | 128 | 29792 | |
| SLH-DSA-SHAKE-256f | 5 | 32 | 64 | 128 | 49856 |
The “-SHA2” variants use SHA-256/SHA-512; “-SHAKE” variants use SHAKE-128/SHAKE-256.
3.5 HQC: The Fifth PQC Standard¶
NIST selected HQC (Hamming Quasi-Cyclic) in March 2025 as the fifth PQC algorithm and the second KEM standard. HQC provides:
-
Diversity: Backup to lattice-based ML-KEM, based on code-based hard problems (not lattices).
-
Different mathematical foundation: If a breakthrough attack on MLWE is found, HQC remains secure.
-
Moderate key sizes: \(\sim\)2–7 KB public keys (much more practical than Classic McEliece’s 1.3 MB).
A draft FIPS for HQC is expected in 2025–2026.
3.6 NIST IR 8547: Transition Timeline¶
NIST Internal Report 8547 specifies the timeline for deprecated classical algorithms:
| Algorithm | Current Use | Deprecated | Disallowed |
|---|---|---|---|
| RSA (all key sizes) | Key exchange, signatures | 2030 | 2035 |
| ECDH (all curves) | Key exchange | 2030 | 2035 |
| ECDSA (all curves) | Digital signatures | 2030 | 2035 |
| DH (classic) | Key exchange | 2030 | 2035 |
| DSA | Digital signatures | 2023 (already) | 2030 |
| ECDH P-256 | TLS key exchange | 2030 | 2035 |
| ECDSA P-384 | Certificate signing | 2030 | 2035 |
| AES-128 | Symmetric encryption | Not deprecated | Retain w/ PQC |
| SHA-256 | Hashing | Not deprecated | Retain |
Action Required Before 2030
Systems using RSA, ECDH, ECDSA, or DH must migrate to NIST PQC standards (ML-KEM, ML-DSA, SLH-DSA) before 2030. For systems with long-lived keys or sensitive data, migration should begin immediately due to the HNDL threat.
3.7 NSA CNSA 2.0 Requirements¶
The NSA’s Commercial National Security Algorithm Suite 2.0 (CNSA 2.0, September 2022) mandates PQC for National Security Systems (NSS):
| Function | Algorithm | Parameters | Timeline |
|---|---|---|---|
| Key Establishment | ML-KEM | ML-KEM-1024 (Level 5) | Ready now |
| Digital Signature | ML-DSA | ML-DSA-87 (Level 5) | Ready now |
| Digital Signature | SLH-DSA | SLH-DSA-SHA2-256s | Ready now |
| Symmetric Encryption | AES | AES-256 | Ready now |
| Hashing | SHA-3 | SHA-384 or SHA-512 | Ready now |
CNSA 2.0 Timeline for NSS
-
2025: PQC software deployed and ready; begin use where possible.
-
2026: PQC used in all new NSS; no new classical-only deployments.
-
2030: Classical algorithms deprecated from NSS.
-
2033: All NSS fully migrated to PQC; classical algorithms removed.
3.8 NIST Additional Signatures Round 2 (2023–Present)¶
After selecting the primary standards, NIST launched a separate process to standardize additional signature schemes, specifically seeking:
-
Diversity from lattice-based schemes (preferring non-lattice approaches).
-
Shorter signatures (addressing SLH-DSA’s large signature sizes).
| Scheme | Family | Notes | pk + sig (B) |
|---|---|---|---|
| MAYO | Multivariate (UOV) | Small pk and sig; promising | 1168 + 321 |
| SNOVA | Multivariate | Very small; efficient | 1157 + 69 |
| QR-UOV | Multivariate | UOV over quotient rings | 55K + 264 |
| UOV | Multivariate | Conservative; large pk | 278K + 96 |
| FAEST | MPC-in-the-head | Hash-based; larger sig | 4.4K + 5.6K |
| SQIsign | Isogeny | Tiny keys; very slow | 64 + 177 |
| MiRitH | MPC-in-the-head | Fast verify; medium sig | 128 + 3K |
| MIRA | Rank-code based | New family | 96 + 346 |
| RYDE | Rank-code based | Based on LRPC codes | 133 + 2K |
| Wave | Code-based | Hash-and-sign; large pk | 3.6M + 822 |
| HAWK | Lattice (NTRU) | Compact; no Gaussian sampler | 1024 + 555 |
| EagleSign | Lattice | NTRU-based variant | 1312 + 666 |
3.9 FIPS 140-3 and Algorithm Compliance¶
FIPS 140-3 is the U.S. government standard for cryptographic module validation (CMVP — Cryptographic Module Validation Program):
-
All FIPS-approved modules must use only FIPS 140-3 approved algorithms.
-
ML-KEM, ML-DSA, and SLH-DSA are FIPS 140-3 approved as of 2024–2025.
-
Cryptographic modules must be validated by NIST-accredited labs (cost: $50K–$500K, timeline: 12–24 months).
-
Security Level 1–4 validation, with Level 2/3 requiring physical security.
CMVP Status 2025
As of early 2025, several implementations of ML-KEM and ML-DSA have undergone or are undergoing FIPS 140-3 CMVP validation. Key vendors: AWS (AWS-LC), Microsoft (SymCrypt), Botan, wolfSSL, OpenSSL (via providers). Check NIST’s CMVP module search at csrc.nist.gov/projects/cryptographic-module-validation-program.
3.10 International Standards Alignment¶
| Body | Standards / Activity | Status |
|---|---|---|
| NIST (USA) | FIPS 203/204/205; HQC selection | Finalized 2024–2025 |
| ETSI (Europe) | TS 103 744 PQC guidance | Published |
| ISO/IEC | ISO 18033 (encryption), 14888 (signatures) | Revisions in progress |
| IETF | RFC drafts for ML-KEM, ML-DSA in TLS 1.3 | Active drafts 2024–2025 |
| BSI (Germany) | Technical Guidelines TR-02102-1 | Updated for PQC |
| ANSSI (France) | PQC recommendations aligned with NIST | Published |
| CRYPTREC (Japan) | Cryptography Research and Evaluation | Monitoring NIST output |
| KPQC (Korea) | Indigenous algorithm competition | HAETAE, NTRUplus, Peregrine |
3.11 Summary¶
-
FIPS 203 (ML-KEM), FIPS 204 (ML-DSA), and FIPS 205 (SLH-DSA) are the three finalized NIST PQC standards as of August 2024.
-
HQC was selected in March 2025 as the fifth algorithm, adding code-based algorithmic diversity.
-
NIST IR 8547 mandates deprecation of RSA, ECDH, ECDSA by 2030 and disallowance by 2035.
-
NSA CNSA 2.0 requires ML-KEM-1024 and ML-DSA-87 (Level 5) for classified systems with earlier deadlines.
-
FN-DSA (FALCON) is expected in FIPS 206 for applications requiring compact signatures.
-
14 additional signature candidates are under evaluation; MAYO and SNOVA are among the most promising.