# Aztec Says Proving-System Flaw Leaves Alpha V5 Funds Exposed Until V6

An internal AI-assisted audit found that an attacker could build a proof the network accepts for a transaction it should reject, and the same team now targets a full post-quantum migration by the end of 2030.

- Published: 2026-10-21T05:37:25.162Z
- Canonical: https://polylog.news/crypto/2026-10-21/aztec-says-proving-system-flaw-leaves-alpha-v5-funds-exposed
- Publisher: Polylog (Crypto desk)
- Section: crypto
- Sources: [Aztec Network (V5 vulnerability)](https://aztec.network/blog/alpha-v5-proving-system-vulnerability), [Aztec Network (Alpha V5)](https://aztec.network/alpha-v5), [Aztec Network (Post-Quantum Aztec)](https://aztec.network/blog/post-quantum-aztec), [The Defiant](https://thedefiant.io/news/blockchains/aztec-targets-full-post-quantum-migration-by-end-of-2030)

Aztec, the privacy-focused Ethereum layer-2 network, has [disclosed a critical vulnerability in the proving system of its Alpha V5 release](https://aztec.network/blog/alpha-v5-proving-system-vulnerability). The flaw lets an attacker construct a zero-knowledge proof that passes verification for a transaction the network should reject. Core contributors found it on July 27 through internal AI-assisted auditing. Aztec tells users to treat funds, applications and contract state on [V5](https://aztec.network/alpha-v5) as exposed to a protocol-level failure until incident response is complete. The fix is planned for V6.

Aztec says reviewers have found no other high-severity or critical V5 issues, and that internal and external human audits are complete. The team had already published a disclosure for a [critical vulnerability in Alpha V4](https://aztec.network/blog/critical-vulnerability-in-alpha-v4). Two consecutive alpha releases have now had critical proving-layer defects.

A soundness bug of this kind is different from an ordinary smart-contract exploit. Most rollups rely on validators re-executing transactions as a fallback check. That check does not work when the proof itself is wrong, because the proof is what everyone relies on. On a privacy network the problem is worse. Transaction contents are hidden, so outside observers have fewer ways to notice that state has been altered.

Separately, Aztec [published a post-quantum roadmap](https://aztec.network/blog/post-quantum-aztec) that aims for full migration by the end of 2030. The Defiant [reports](https://thedefiant.io/news/blockchains/aztec-targets-full-post-quantum-migration-by-end-of-2030) that the first deliverable, hardened developer libraries, is due in February 2027. Selection of new signature schemes follows in March 2027, public-key encryption in July 2027 and zero-knowledge proof systems in November 2027. Aztec says the final stage depends on changes to Ethereum's data-availability layer, which Aztec does not control.

## What this means

Privacy layer-2s are pitching confidential execution to institutions. Their security, though, rests on proving code that is young and that only a small number of people can audit. Exposure sits with early depositors and builders on Aztec, and more broadly with any rollup that relies on a single proof system. The consensus view treats audits as a finished step. Aztec's own account says human audits were complete when AI-assisted review found the flaw. This incident shows that AI-assisted review can also serve a defensive purpose. In 2019 Zcash disclosed a counterfeiting flaw that its developers had found in its original Sprout proving system and fixed quietly. That disclosure made multi-proof designs and escape hatches standard requests in later designs. The likelier outcome within months is that Aztec and its peers point to proof-system diversity and forced-exit paths in their mainnet plans, and that institutional pilots wait for at least one clean release cycle. That depends on V6 shipping without a new critical finding. A second soundness disclosure after V6 would show the gap between audits and AI review is not closing.

## What to watch

- The V6 release and whether its audit reports describe independent verification of the fixed prover. That release decides whether exposed V5 users have a safe place to move.
- Whether other zero-knowledge rollups publish their own AI-assisted audit results. A cluster of disclosures would confirm that the defect backlog is industry-wide.
- Ethereum developer discussions on post-quantum data availability, since Aztec says its 2030 target depends on them.