AI agents have drastically reduced the resource costs for potential quantum attacks on Bitcoin and Ethereum by 86%, signaling a shift in cryptographic security.
Overview
- AI agents helped reduce the resource score for a quantum attack from 10.75 billion to 1.496 billion.
- The leading quantum circuit used 1,151 logical qubits and around 1.3 million Toffoli gates.
- A new design brought the gate count below 1 million, excluding hardware costs.
A recent study highlighted a significant breakthrough in the potential quantum threat to Bitcoin and Ethereum, with researchers utilizing AI agents to decrease the resource requirements for a critical phase of a quantum attack by 86%. This research was published on Wednesday and indicates a major reduction in computing power needed to compromise wallets and steal assets.
The timeline for the emergence of powerful quantum computers capable of executing these attacks, termed "Q-Day," remains uncertain. Nevertheless, researchers emphasize that a transition away from existing cryptographic methods is already underway, with the National Institute of Standards and Technology (NIST) standardizing alternatives. A draft suggests that public-key algorithms at the 112-bit security level may be phased out post-2030 and banned by 2035.
This research stemmed from the ECDSA.Fail competition, launched by Eigen Labs in late May, which involved over 100 participants working on circuits for secp256k1, the elliptic curve securing Bitcoin and Ethereum transactions. Key contributors included teams from Theta Labs, MultiVM Labs, and the Ethereum Foundation.
The study successfully designed a quantum circuit capable of performing calculations necessary to breach Bitcoin’s security, achieving an 86% reduction in resource requirements. However, the tests confirmed the circuit's calculations without actually cracking any Bitcoin private keys.
The challenge assessed the quantum memory and computational requirements of each circuit, calculating the resource score by multiplying the logical qubits by the number of Toffoli gates used. A lower resource score indicates fewer resources are needed for potential future attacks. By late July, the resource score had dropped significantly, reaching 1.496 billion, which is about half of the score reported by Google Quantum AI earlier this year, although direct comparisons are complex due to different methodologies.
Researchers noted, “ECDSA.Fail exemplifies a case study of Open Autoresearch: a process where human participants and AI agents collaboratively generate, implement, test, and share enhancements against a common goal.” This development is part of a broader trend within the cryptocurrency sector to invest in research addressing quantum threats, with firms like Galaxy Digital committing significant funds to advance this area.
In July, Galaxy Digital announced a pledge of up to $5 million for research in this field, while nine companies, including BlackRock and Coinbase, collectively invested $15 million over three years to enhance Bitcoin security, including defenses against quantum attacks.
