Computational cryptography › Not live yet
Trust or structural assumption

Proof systems are sound and bug-free

In plain words

The bet that complex proof systems have no bugs that would let someone prove something false.

Why it matters: Matters more and more as Ethereum plans to rely on proofs instead of re-running blocks.

The details

On the map: zkVM circuits and provers, formal verification in progress, k-of-n proofs tolerate one bad prover

Zero-knowledge virtual machines are large, complex systems, and soundness bugs (a prover convincing a verifier of something false) have been found in production provers[1].

Plans that let Ethereum validators rely on execution proofs assume the proof systems are sound and correctly implemented. Today's ZK rollups and privacy apps already rely on the same kind of bet for their own circuits (the "ZK circuits" box in the live section). Requiring proofs from several independent systems (for example 2 of 3)[2] tolerates a bug in one. Formal verification of leanVM and zkEVMs is underway[1].

Rests on

If any of these breaks, this breaks too.

If it falls

If this breaks, these break with it.

Everything that falls with it (5 boxes)

References

  1. Ethereum Foundation. Shipping an L1 zkEVM #2: The Security Foundations. Ethereum Foundation blog, 2025. blog.ethereum.org/en/2025/12/18/zkevm-security-foundations
  2. Kevaundray Wedderburn, Justin Drake, Ignacio Hagopian et al. EIP-8025: Optional Execution Proofs. Ethereum Improvement Proposals (Draft), 2025. eips.ethereum.org/EIPS/eip-8025