AI-Boosted Quantum Attack on Bitcoin Could Be Just Around the Corner

Researchers working with AI coding agents have made significant progress in reducing the computing resources needed to carry out a potential quantum attack on Bitcoin and Ethereum. In an open competition called ECDSA.Fail, more than 100 participants used AI agents to lower the resource score for a key step in a quantum attack from 10.75 billion to 1.496 billion, an 86% reduction.

The leading circuit used 1,151 logical qubits and approximately 1.3 million Toffoli gates. A later design reduced the gate count below 1 million, with results excluding the hardware costs of a full attack. The findings suggest that while the timing of a potential quantum attack remains uncertain, migration away from vulnerable cryptography is already underway.

The researchers noted that NIST has standardized post-quantum replacements and proposed deprecating classical public-key algorithms at the 112-bit security level after 2030. They also highlighted the importance of continued research in this area, with companies such as Galaxy Digital committing up to $5 million to quantum defense efforts.