AI-assisted research halves quantum resource estimate for breaking Bitcoin and Ethereum keys
Regulation & Gov ·
A new paper from the Ethereum Foundation, Theta Labs, and StarkWare shows AI coding agents cut the estimated quantum computing resources needed to attack secp256k1 cryptography by over 50%.
The researchers used AI coding agents to optimize key calculations within Shor's algorithm, the quantum method theorized to break the elliptic-curve cryptography underlying both Bitcoin and Ethereum. The resulting optimized circuit brings the requirement down to 1,151 logical qubits, according to the paper, marking a reduction of more than 50% compared with a benchmark Google published in March, as detailed in a report from wublockchain.xyz.
The finding matters because it shows the threat timeline for quantum attacks on crypto networks can shrink through algorithmic refinement alone, without any leap in quantum hardware. Reducing the qubit count needed for an attack lowers the bar future machines would need to clear, even as physical quantum computers capable of running such circuits do not yet exist. The Block also covered the halved benchmark for this key operation in a prospective Bitcoin and Ethereum attack.
Multiple accounts of the research converge on the same core figures: a reduction exceeding 50% in resource requirements and the 1,151-logical-qubit target, underscoring that the result has been corroborated across separate write-ups rather than appearing in isolation.
The researchers were explicit that no quantum machine currently in operation can break Bitcoin's or Ethereum's cryptography today. Their warning centered on the lead time needed elsewhere: post-quantum upgrades to these networks take years to design, test, and deploy, and they argued that proactive action is required well before hardware catches up to the newly lowered resource bar.
What remains unresolved is how quickly hardware development could close the gap the algorithmic optimization has narrowed, and what concrete timeline Bitcoin and Ethereum developers will set for implementing post-quantum protections. Whether further AI-assisted optimizations could push the qubit requirement down again, and how the broader research and validator communities plan to respond, are open questions the current paper does not answer.