AI Agents Cut Quantum Attack Benchmark for Bitcoin and Ethereum by 86%
Key Takeaways
- •The ECDSA.Fail competition reduced the measured quantum resource requirement to 1.496 billion by July 26.
- •The leading circuit required 1,151 logical qubits and about 1.3 million Toffoli gates, while a later design used fewer than 1 million gates.
- •The study’s score was about half a March Google Quantum AI benchmark, although the researchers said differing methods prevent direct comparison.
- •NIST has standardized post-quantum alternatives, while a draft proposal would phase out vulnerable classical public-key algorithms after 2030 and prohibit them after 2035.
- •Galaxy Digital and a separate group of nine firms have committed funding toward quantum-attack defenses and broader Bitcoin security research.
More than 100 participants using AI coding agents reduced the resource score for a key step in a potential quantum attack on the cryptography securing Bitcoin and Ethereum from 10.75 billion to 1.496 billion, according to a paper posted Wednesday.
The work focused on the secp256k1 elliptic curve, which is used to secure transaction signatures on both networks. Researchers designed and tested a quantum circuit for a calculation needed to uncover a wallet’s private key. If carried out as part of a full attack, such a capability could allow an attacker to steal funds. However, the tests verified the circuit’s calculations and did not crack a Bitcoin private key.
The paper, available on arXiv, said the 86% reduction was achieved through ECDSA.Fail, an open competition launched by Eigen Labs in late May. More than 100 participants developed circuits for the challenge. The paper’s authors include researchers affiliated with Theta Labs, MultiVM Labs, Eigen Labs, Trail of Bits, StarkWare, and the Ethereum Foundation.
The competition measured the quantum memory and computational work required by each circuit by multiplying its number of logical qubits by its number of Toffoli gates. The Toffoli gate is a costly type of quantum operation invented by Tommaso Toffoli in 1980. A lower score indicates that the calculation requires fewer combined resources, potentially reducing the resources needed for a future attack.
By July 26, participants had reduced the score from 10.75 billion to 1.496 billion. The leading circuit used 1,151 logical qubits and approximately 1.3 million Toffoli gates. A later design reduced the gate count to below 1 million. The reported results do not include the hardware costs of carrying out a full attack.
According to the study, the leading score was roughly half of a March benchmark from Google Quantum AI. The researchers cautioned that differences in testing and counting methods prevent a direct comparison with Google’s result.
“Beyond the resulting circuits, ECDSA.Fail provides a case study of Open Autoresearch: a verifier-gated research process in which human participants and AI agents iteratively generate, implement, test, and share candidate improvements against a common measurable objective,” the researchers wrote.
The arrival date of quantum computers capable of carrying out a full attack remains uncertain, a milestone commonly known as “Q-Day.” The researchers wrote: “Although its timing remains uncertain, migration away from vulnerable cryptography is already under way. NIST has standardized post-quantum replacements, and the initial public draft of NIST IR 8547 proposes deprecating classical public-key algorithms at the 112-bit security level after 2030 and disallowing them after 2035.”
The findings come as crypto companies increase funding for research into defenses against quantum attacks. In July, Galaxy Digital committed up to $5 million to that work. Nine firms, including BlackRock, Coinbase, and Strategy, separately pledged a combined $15 million over three years for broader Bitcoin security research, including quantum defenses.
Source: https://decrypt.co/377925/ai-agents-slash-cost-quantum-attack-bitcoin
