The post-quantum privacy chain.

Every layer, not just the signature.

One ring, one parameter set. The cryptography is published research; the integration is ours. Two things we would rather state than have you find. HAWK-512 was withdrawn from NIST standardisation on 29 July 2026, and we have since removed it from every path. The chain produces no block proof at all today.

Stated layer by layer, because a single headline number would hide more than it tells you. Spending authorization on the testnet is set by the signature layer, ML-DSA-44, with ML-DSA-65 on blocks and validator identity; both are FIPS 204 at NIST categories 2 and 3. This replaced HAWK-512 outright after the July 2026 cryptanalysis cut its key-recovery cost to no more than 2¹⁰⁸. The state-commitment layer is independently estimated at roughly 280-bit classical and 250-bit quantum by the standard lattice estimator as of 21 June 2026, and it is enforced on every block. The chain produces no block proof today, so there is nothing there to count toward live security either way. Parameters are published, not asserted. There are no real funds on the testnet.

There isn't one right now. The folding prover was removed from the node. It took about a second to fold a single block against a sub-second block time, so it could never keep pace, and each proof ran to 2.29 MB against a 50 KB design target. Validators confirmed state by re-execution regardless, and no block header ever committed to a proof, so nothing on the live chain depended on it. Re-introducing a proof layer, from a construction that verifies directly rather than by folding, is tracked in the progress log, and we will say so here when it lands rather than before.

About 2,161 transactions per second on a single validator, measured 23 August 2026 under ML-DSA-44 with 1,000 independent-UTXO transactions and 200 ms blocks. Thinner block policies measure far lower, around 200 tx/s on a strictly-chained workload, so the ceiling is block policy and admission, not signing. Blocks tick well under a second with deterministic finality. That measurement and its date are in the progress log.

Yes. Multi-validator consensus runs the testnet today. Header aggregation for fast light-client verification is an open milestone on the chain and consensus track.

SHA3-256 and SHAKE-256 are FIPS 202. HAWK was withdrawn from the NIST additional-signatures process on 29 July 2026, after cryptanalysis that recovers a key from the public key alone; its authors judged that larger parameters would make the scheme uncompetitive. Our published contingency confined a move to ML-DSA (FIPS 204) to the signature layer alone. That move is done: ML-DSA-44 authorizes spends and ML-DSA-65 signs blocks and validator identity, which also improves the CNSA 2.0 posture. Separately we are moving spending authorization to a proof of key knowledge, so that the chain stops depending on any one signature scheme.

That withdrawal is also why authorization is moving off signatures altogether. When the chain checks a proof relation instead of running one named algorithm, changing scheme is a wallet update rather than a rewrite of addresses, balances and consensus rules. Post-quantum cryptography is young enough that this will happen again, so the useful question is not whether a scheme breaks but what breaking one costs.

Settlement only today. No untrusted smart-contract VM. Programmability comes after the cryptography reaches audit-grade.

Yes. Run Nixeon as a sidecar settlement layer, or integrate the full stack.

Auditable privacy, not anonymity. The privacy lane hides amounts, and viewing keys let a holder disclose to counterparties, regulators and auditors. Amounts are public today. Hiding them is in build, and viewing keys are still at the research stage. Status per track is in the progress log.

Not yet, and we say so plainly. This is a public testnet alpha. An external post-quantum security audit is the gate before production use with real funds.