The amount of computing power potentially needed to target Bitcoin and Ethereum with a quantum computer has been reduced significantly, according to new research shared with CoinDesk. Researchers working with AI coding agents have improved a key calculation used by Shor’s algorithm, producing a result more than 50% below Google’s benchmark from March.
The research involved scientists from the Ethereum Foundation, Theta Labs, StarkWare and several other organizations. Their quantum circuit uses 1,151 logical qubits and approximately 1.3 million Toffoli gates.
These measurements describe two different aspects of the required quantum system. The number of logical qubits indicates the approximate size of the machine, while the number of Toffoli gates represents the amount of computation it must perform.
The primary circuit achieved a score of about 1.5 billion, compared with roughly 3 billion in Google Quantum AI’s March result. The researchers warned that the figures should not be treated as a perfectly direct comparison because the two approaches use different interfaces and accounting conventions.
A separate version designed to more closely reflect the way the calculation would be used within Shor’s algorithm produced a score of approximately 1.96 billion. That was still below Google’s reported result.
AI and researchers optimize quantum calculations
The team focused on point addition, a mathematical operation repeatedly used by Shor’s algorithm.
The significance for cryptocurrencies comes from the algorithm’s ability, in principle, to undermine elliptic-curve cryptography. A sufficiently advanced quantum computer could potentially derive a private key from an exposed public key, allowing an attacker to create transaction signatures associated with the targeted wallet.
Both Bitcoin and Ethereum rely on secp256k1, the cryptographic system examined in the research.
More than 100 people participated in the optimization effort over roughly eight weeks through ECDSA.Fail, an open challenge established by Eigen Labs. The project generated more than 400 accepted submissions, with each successful improvement providing a new baseline for subsequent work.
AI coding agents handled much of the implementation, testing and smaller-scale optimization. Human researchers generally determined which directions to pursue and made larger changes to the circuit design.
The paper does not quantify the individual contribution of humans and AI to the final improvement.
Nevertheless, the results add another dimension to the cryptocurrency industry’s quantum timeline. Quantum computers do not necessarily have to become more powerful at the same pace if researchers can continue making the underlying computations more efficient.
Bitcoin and Ethereum remain safe from current machines
The latest results should not be interpreted as evidence that Bitcoin or Ethereum can be attacked by quantum computers today.
The researchers optimized only one major calculation within a potential attack. Their work does not include physical error correction, the complete Shor algorithm or all of the hardware-specific costs that would be involved in executing the computation on an actual quantum machine.
No existing quantum computer can use the reported results to break Bitcoin or Ethereum.
The efficiency gains have also continued after the paper’s July cutoff. A later design reduced the score to approximately 1.26 billion, while another lowered the machine requirement to 813 logical qubits but required substantially more computation.
Jieyi Long, the study’s lead author and CTO of Theta Labs, told CoinDesk that the research does not mean an attack is about to occur. The concern, he explained, is that protecting against such a threat requires preparation that could take years and cannot simply be applied after an attack becomes possible.
The broader quantum race is also gaining government support. The U.S. Commerce Department this week finalized CHIPS Act awards of up to $100 million apiece for Rigetti, D-Wave and Quantinuum, while taking minority stakes in each company. The funding is intended to help develop larger fault-tolerant quantum machines.
For the crypto industry, the latest research is therefore less about an immediate vulnerability and more about preparation. As researchers reduce the amount of computation needed for quantum attacks, blockchain networks may have less time than previously assumed to transition toward quantum-resistant security.
