Researchers from Eigen Labs, Trail of Bits, StarkWare, Theta Labs, MultiVM Labs, and the Ethereum Foundation have demonstrated a substantial reduction in the resources needed to execute a quantum attack on Bitcoin and Ethereum's cryptographic security.
The team, working with AI coding agents, decreased a key benchmark for quantum attacks on the secp256k1 elliptic curve from 10.75 billion down to 1.496 billion between late May and July 26. This represents an 86% reduction in the computational cost required for this step of a potential quantum attack.
The competition, called ECDSA.Fail and run by Eigen Labs, scores circuit designs by measuring logical qubits and Toffoli gates—expensive quantum operations. A lower score indicates fewer resources needed to crack the cryptography. The leading design in the competition used 1,151 logical qubits and roughly 1.3 million Toffoli gates, while a later submission reduced gates below one million.
Bitcoin and Ethereum both rely on secp256k1 to secure signatures. A sufficiently powerful quantum computer could theoretically reverse this math and derive a private key from a public one. The researchers employed what they call Open Autoresearch, with humans and AI agents iterating against a shared measurable target with a verifier in the loop.
The timing coincides with an accelerating industry effort to develop quantum-resistant defenses. The Ethereum Foundation set a hard deadline of December 2029 to make transactions, validators, and storage quantum-resistant. StarkWare pushed the first quantum-safe Bitcoin transaction to mainnet, and Ethereum developers have proposed rebuilding the validator deposit contract. Ripple is hardening the XRP Ledger, while Galaxy committed up to $5 million in July, and nine firms including BlackRock and Coinbase pledged $15 million over three years for quantum security research.
The researchers noted that the National Institute of Standards and Technology has proposed deprecating classical public-key algorithms at the 112-bit level after 2030 and disallowing them after 2035.
The significance of the benchmark improvement lies in how quickly it was achieved. Previous timelines for quantum threats assumed attack research would proceed at human speed. This demonstration shows that AI-assisted research can substantially accelerate progress in attack methodology within weeks, not years.
