# Aztec Discloses Critical Proving-System Flaw in Its Live Alpha V5 Privacy Network

Contributors found through internal AI-assisted auditing that an attacker could build a proof that verifies for a transaction the network should reject, with the fix deferred to V6.

- Published: 2026-09-25T13:11:38.958Z
- Canonical: https://polylog.news/crypto/2026-09-25/aztec-discloses-critical-proving-system-flaw-in-its-live-alp
- Publisher: Polylog (Crypto desk)
- Section: crypto
- Sources: [Aztec Network](https://aztec.network/blog/alpha-v5-proving-system-vulnerability), [Aztec Network](https://aztec.network/alpha-v5), [Aztec Network](https://aztec.network/blog/dark-forest-aztec-game-goes-live), [The Defiant](https://thedefiant.io/news/defi/aztec-v4-withdraw-june-25-v5-upgrade-security-vulnerability)

Aztec, the privacy-focused layer-2 network on Ethereum, told users that its live Alpha V5 release carries a [critical vulnerability in the proving system](https://aztec.network/blog/alpha-v5-proving-system-vulnerability). Contributors identified the defect on 27 July 2026 through internal auditing assisted by artificial intelligence (AI). The failure is a soundness break: an attacker may be able to construct a proof that passes verification for a transaction the protocol should have rejected. Funds, applications and contract state on V5 should be treated as exposed to a protocol-level failure until operators complete the required network actions, and the corrective release is planned for V6.

This is the second consecutive version of the network to ship with a proving-system flaw disclosed after launch. Aztec [disclosed a critical vulnerability in Alpha V4 in March](https://thedefiant.io/news/defi/aztec-v4-withdraw-june-25-v5-upgrade-security-vulnerability), urged users to withdraw before a governance vote made the details public, and delivered the patch in V5. The team frames these findings as the intended output of an alpha phase, in which the audit process is meant to surface exactly this class of defect before wider deployment.

The disclosure lands while Aztec pushes usage forward. The network has been promoting applications built on private execution, including [a live version of the Dark Forest game](https://aztec.network/blog/dark-forest-aztec-game-goes-live) and documentation on how [gas works on Aztec](https://aztec.network/blog/how-gas-works-on-aztec), and it markets V5 as the current [Alpha network release](https://aztec.network/alpha-v5).

A soundness bug is a different risk category from a smart-contract exploit. Validator re-execution, the fallback most rollups rely on to catch bad state transitions, cannot detect a proof that verifies correctly against a flawed circuit. The user's real protections are conservative deposit limits, the escape hatch, and the team's willingness to disclose the flaw before an attacker finds it independently.

## What this means

The exposure sits with anyone holding value inside a zero-knowledge rollup whose security reduces entirely to circuit correctness, because a soundness defect lets an attacker mint or move state that the verifier accepts as valid, and no honest node can object. Privacy rollups carry the sharpest version of this, since encrypted state makes anomalous balances harder for outside observers to spot early. The precedent is Aztec's own V4 cycle: disclosure timed to an upgrade, a user withdrawal window, and a patch one release later, with no loss reported. The likelier outcome over the coming months is that competing privacy and general-purpose rollups respond with proof-system diversity, multi-prover verification and explicit exit guarantees, and that institutional users demand these before committing size, which slows deposit growth into single-prover networks. The read depends on the V6 fix arriving without an incident in between. It would be proven wrong if an attacker exploits the V5 flaw before the patch, which would push the question from design practice to loss recovery.

## What to watch

- Whether Aztec publishes the V6 timeline and a specification of the defective circuit, which determines how quickly outside auditors can check for the same pattern elsewhere.
- Whether other zero-knowledge rollups announce second proof systems or multi-prover verification, the practical response to a risk that re-execution cannot catch.
- Whether AI-assisted auditing keeps producing critical findings inside teams rather than from external researchers, which changes who controls disclosure timing.