A group of researchers has more than halved the computational complexity estimate for a key stage of a potential future quantum attack on Bitcoin and Ethereum, compared to Google’s benchmark, CoinDesk reports.
Our very own CTO @jieyilong is the lead author on a paper that cuts the estimated quantum cost of the core operation in breaking Bitcoin and Ethereum’s cryptography by more than half. https://t.co/2K3AzwuoBL
— Theta Network (@Theta_Network) September 10, 2026
The study involved over 100 specialists and AI agents.
Estimate Reduced from 3 Billion to 1.5 Billion
Experts from the Ethereum Foundation, Theta Labs, StarkWare, and other organizations developed a quantum scheme for one of the key operations of Shor’s algorithm. It uses 1151 logical qubits and about 1.3 million Toffoli gates.
The authors compare these parameters to the size of the machine and the workload: fewer qubits mean a more compact quantum computer, and a reduction in the number of gates means less computation.
The final estimate of the scheme is about 1.5 billion arbitrary units—more than 50% lower than the approximately 3 billion reported by Google Quantum AI in March. However, researchers note that the comparison is not entirely accurate due to differences in interfaces and accounting rules.
A version of the scheme adapted closer to real-world use in Shor’s algorithm computation achieved a score of about 1.96 billion.
AI Involved in Optimization
Researchers optimized the elliptic curve point addition operation, which is repeatedly performed in Shor’s algorithm.
Over 100 participants worked for about eight weeks on the task as part of the open ECDSA.Fail challenge created by Eigen Labs. During this time, they made over 400 accepted improvements, each serving as a starting point for subsequent participants.
AI agents were actively used in the work: they helped write code, conduct numerous tests, and perform small optimizations. Humans primarily chose research directions and made more significant architectural changes. The authors did not separately assess the contributions of humans and AI to the final result.
Thus, progress in approaching a quantum attack may not only come from creating more powerful computers. Improving algorithms and software optimizations can also reduce hardware requirements.
Bitcoin Breach Still Distant
Researchers did not optimize the entire attack. The presented scheme covers one of the main computational operations and does not include physical error correction, the full calculation of Shor’s algorithm, and several hardware costs that would arise when running on a real quantum computer.
Therefore, existing systems do not allow these results to be used to breach Bitcoin or Ethereum.
Moreover, calculations continue to improve. The study presents a later design with a score of about 1.26 billion, and another variant allows reducing the requirement to 813 logical qubits, but at the cost of significantly more computation.
Bitcoin and Ethereum use secp256k1—a cryptographic scheme on elliptic curves, which is the target of the discussed attack. A sufficiently powerful fault-tolerant quantum computer could theoretically use Shor’s algorithm to derive a private key from a public key and sign a transaction on behalf of the owner.
The authors of the study do not believe that a quantum attack on cryptocurrencies is inevitable in the near future. In their view, the problem lies elsewhere: transitioning to quantum-resistant cryptography could take years, and already compromised keys cannot be retroactively protected.
According to IonQ’s estimate, a fault-tolerant quantum computer with about 20,000 physical qubits would require 26 days to break secp256k1 cryptography.
