Researchers used a AI encoding proxy to significantly reduce the benchmark for quantum computing resources related to the cryptographic security of Bitcoin and Ethereum. According to the latest paper, the resource rating for key computational steps on the secp256k1 curve has been reduced from 10.75 billion to 1.496 billion, a decrease of about 86%.
This result does not mean that the research team has cracked Bitcoin’s private key, nor does it indicate that quantum computers already possess the practical capability to carry out such attacks. It reflects that among the potential quantum attack pathways, the theoretical resources required for a certain critical computational step have been further reduced.
Competition Focus: secp256k1
Relevant results come from the ECDSA.Fail open competition. This project was initiated by Eigen Labs at the end of May, with the goal of optimizing the quantum circuit design for secp256k1. secp256k1 refers to the elliptic curve used for transaction signatures in Bitcoin and Ethereum.
The authors of the paper come from multiple institutions, including Theta Labs, MultiVM Labs, Eigen Labs, Trail, of, Bits, StarkWare, as well as the Ethereum Foundation. The research team stated that over 100 participants worked collaboratively with AI agents to continuously generate, test, and improve candidate circuit solutions.
Resource rating dropped to 1.496 billion
The method used in the competition is to multiply the number of logical quantum bits by the number of Toffoli gates. The latter represents a type of quantum operation that is relatively more costly. The lower this score, the fewer comprehensive resources are required to complete the calculation.
- Uses 1,151 logical quantum bits
- Approximately 1.3 million Toffoli doors are required.
- The subsequent plan has reduced the number of doors to less than 1 million.
The paper also mentions that this score is approximately only Google Quantum AI of the relevant benchmark in March of this year. However, researchers also emphasize that due to different testing conditions and counting methods, the two cannot be directly compared.
There is still some time before the actual attack occurs.
The research team stated that this test verified the results of quantum circuit computations themselves and did not actually crack any Bitcoin private keys. More importantly, the paper did not take into account the hardware costs required for a complete attack, so there is still a significant gap from a feasible attack in reality.
Nevertheless, this study has once again drawn attention to quantum risks in the crypto industry. The paper points out that the National Institute of Standards and Technology (NIST) in the United States has completed the standardization of quantum cryptography and proposed in a public draft that some traditional public key algorithms will be phased out gradually after 2030, and the use of certain classical solutions with a 112-bit security level will no longer be allowed after 2035.
Additional information:In recent months, several crypto institutions have begun to increase their investment in this area. In July, Galaxy Digital pledged to invest up to 5 million US dollars to support such research; BlackRock, Coinbase, Strategy, and nine other institutions also committed to investing a total of 15 million US dollars over three years in broader Bitcoin security research, including quantum defense.
