Algorithm
ML-DSA-65 (CRYSTALS-Dilithium)
ML-DSA-65 · FIPS 204 · Patent Pending · Filed 2026-05-08
smshPQ™ · What it is, and why now
The inference log you sign today is the only one that still verifies in 2035.
smshPQ is a post-quantum signed-inference receipt rail. Every model call gets an ML-DSA-65 signed receipt — FIPS 204 finalized, NSA CNSA 2.0 listed. Sub-millisecond signing. $0.00018 per receipt. Built for OpenAI, Anthropic, Google, Kimi, Manus, vLLM, SGLang, and any team self-hosting open weights.
01 · What it is
One primitive. One receipt. Every call.
Every inference call routed through Hive returns a structured receipt, signed with ML-DSA-65 inside the inference boundary, before the response returns to the buyer. The signature names the model, the prompt hash, the response hash, the route taken, the cache state, and the policy applied. The receipt is small, the verification is sub-second, and the algorithm is on the only NIST signature standard the NSA has cleared for national-security workloads through 2035.
Standard
FIPS 204 · Finalized Aug 2024
Mandate
NSA CNSA 2.0 listed
Receipt size
~3.3 KB · sub-ms sign
Per-call price
$0.00018 per receipt
Audit lifetime
Unbounded against Q-Day
The Lead
Ed25519 and ECDSA are not getting a second life. Every inference log signed with them today is harvestable — captured by an adversary now, forged at leisure once a cryptographically-relevant quantum computer lands. An audit log signed in 2026 with ML-DSA-65 is the only one that still verifies in 2035.
02 · Why now · three live deadlines
The window to sign with the right algorithm is closing on a schedule that's already on paper.
Aug 13, 2024 · Already past
FIPS 204 finalized
NIST published the Module-Lattice Digital Signature Standard. ML-DSA-65 is the named successor for general-purpose signatures. Every government cryptographic procurement is now obligated to plan migration.
csrc.nist.gov/pubs/fips/204/final2025 onward · In effect
NSA CNSA 2.0 listed
The Commercial National Security Algorithm Suite 2.0 names ML-DSA as the digital-signature standard for National Security Systems. Software signing required from 2025. Full sector compliance by 2030–2033.
media.defense.gov · CNSA 2.0Aug 2, 2026 · 9 weeks out
EU AI Act · Article 12 logging
High-risk AI systems must keep automatically-generated logs sufficient to trace operation across the lifecycle. Article 12 specifies "logs", but Articles 13 and 50 demand integrity, transparency, and machine-verifiability. A signed receipt is the simplest way to satisfy all three.
artificialintelligenceact.eu · Article 12Three deadlines, three jurisdictions, one cryptographic answer. Every other signature scheme on every other AI audit log shipping in 2026 is on the wrong side of these three lines.
03 · The PQ gap
Ed25519 / ECDSA versus ML-DSA-65, on the dimensions that matter for an audit log.
| Dimension | Ed25519 / ECDSA | ML-DSA-65 (smshPQ) |
|---|---|---|
| Signature scheme family | Elliptic-curve, pre-quantum | Module-lattice, post-quantum |
| NIST FIPS status | FIPS 186 (legacy) | FIPS 204 · Finalized Aug 2024 |
| NSA CNSA 2.0 | Excluded | Listed · Required for NSS |
| Quantum-forgeable | Yes · by Shor's algorithm | No · lattice-hard |
| Harvest-now-decrypt-later exposure | Full · adversary stores today, forges later | Zero |
| Sign latency | ~0.05 ms | ~0.6 ms · sub-millisecond |
| Receipt size | 64 B | ~3.3 KB |
| Audit-log lifetime against Q-Day | Bounded · expires before the model retires | Unbounded |
The 64-byte signature looks lighter on the wire. It also expires the moment a cryptographically-relevant quantum computer arrives. The 3.3 KB receipt is the one your CISO can defend in front of an auditor in 2032.
04 · Why not wait
Harvest-now-decrypt-later is already happening to the data you are generating today.
The logs are being captured now.
Adversaries with state-level resources are storing signed traffic today on the assumption that a cryptographically-relevant quantum computer arrives inside the audit-log retention window. AI inference logs — which name the model, the prompt, and the output — are among the highest-value targets for retroactive forgery.
The migration takes years.
NIST's own published guidance assumes a multi-year cryptographic transition. Google's published timeline targets 2029 for full migration. Software-signing requirements under CNSA 2.0 begin in 2025. The inference layer is the one place still routinely signing with Ed25519 — which means it is the one place still creating fresh harvestable traffic every second.
blog.google · PQC migration timeline05 · The economics
The math works at the bottom. Free isn't part of the conversation.
Default scenario for an Anthropic-scale buyer: 300M calls/month at 1,200 tokens per call, $3.50 per million tokens. Current spend is $1.26M/month. smshPQ-routed spend, with cache-proof and verified-routing savings, lands at $772K/month — net annual savings $5.85M, 9× ROI, three-day payback. One click in the calculator reroutes the assumptions to OpenAI-scale or hyperscale.
Per receipt
$0.00018
Net annual · Anthropic-scale
$5.85M
Net annual · OpenAI-scale
$19.51M
Net annual · Hyperscale
$195M
Payback
3 days
ROI
9×
Scope a pilot.
One scoping call. We bring the receipts. You bring your call volume, your model mix, and the audit constraint you've been quietly worried about since the EU AI Act passed.
[email protected]