Key facts
- AI agents and human participants reduced the resource score for a quantum attack on Bitcoin and Ethereum cryptography by 86%.
- The resource score was cut from 10.75 billion to 1.496 billion.
- The leading quantum circuit design used 1,151 logical qubits and approximately 1.3 million Toffoli gates.
- A later design brought the gate count below 1 million.
- The competition ECDSA.Fail was launched by Eigen Labs in late May.
- NIST proposes deprecating classical public-key algorithms after 2030 and disallowing them after 2035.
Researchers leveraging AI agents have significantly reduced the computational resources required for a quantum attack on the cryptography securing Bitcoin and Ethereum. The ECDSA.Fail competition, launched by Eigen Labs, saw over 100 participants develop quantum circuits for secp256k1, the elliptic curve used in transaction signatures. These efforts slashed the resource score, a benchmark for the computing power needed to uncover a wallet's private key, by 86% from 10.75 billion to 1.496 billion. The leading design utilized 1,151 logical qubits and approximately 1.3 million Toffoli gates, with a subsequent design achieving under 1 million gates. While the exact arrival of quantum computers capable of such an attack, known as 'Q-Day,' remains uncertain, the findings highlight the growing need for migration away from vulnerable cryptography. NIST is already standardizing post-quantum replacements and plans to deprecate classical public-key algorithms after 2030. The competition also served as a case study for 'Open Autoresearch,' a process involving AI agents and human participants iteratively improving solutions. This work is part of broader industry efforts, with firms like Galaxy Digital and BlackRock investing in quantum defense research.