AI Agents Optimize Quantum Attack Calculations
A research initiative involving over 100 researchers utilizing AI coding agents reduced the resource score for a critical component of quantum attacks against Bitcoin by 86.1%, according to a paper published on September 9. Eigen Labs introduced the ECDSA.Fail challenge in late May to improve the efficiency of secp256k1 point addition circuits, with performance measured by multiplying logical qubits used by the average number of Toffoli gates executed.
The optimization score dropped from 10.75 billion to 1.496 billion by July 26. The top-performing design used 1,151 logical qubits and approximately 1.3 million Toffoli gates, with later attempts reducing the gate count below one million and one design requiring only 813 qubits. The research explored mathematical efficiency rather than demonstrating an actual attack on Bitcoin wallets.
Full Attack Still Requires Non-Existent Hardware
Despite the optimization gains, executing a complete secp256k1 attack remains beyond current technological capabilities. IonQ estimates that a full attack would require approximately 1,457 logical qubits and 39 million Toffoli gates, translating to 19,397 physical trapped-ion qubits and about 25.7 days of processing time. IonQ indicates such hardware aligns with company plans for around 2028.
Google researchers calculated that either 1,200 logical qubits with 90 million Toffoli gates or 1,450 logical qubits with 70 million gates would be needed for an efficient complete attack, with their superconducting circuit design running in minutes using fewer than 500,000 physical qubits.
Bitcoin's Current Quantum Vulnerability
According to Glassnode analysis, 6.04 million BTC—30.2 percent of current supply—faces quantum attack risk because public keys are available on the blockchain. Of this total, 1.92 million BTC are exposed in the output category and 4.12 million BTC through behavior such as address reuse.
A second vulnerability emerges when a public key becomes publicly visible during a transaction. In theory, a sufficiently powerful computer could derive the private key before the transaction confirms.
Existing Mitigation Measures and Their Limitations
BIP 360 introduces Pay-to-Merkle-Root, a new output type that eliminates the quantum attack threat for Taproot's key-path spending option. This protects coins with visible public keys over extended periods but does not address quantum risk during the brief mempool window and cannot automatically migrate existing coins.
StarkWare demonstrated a quantum-safe Bitcoin transaction on mainnet, though the method requires hours of off-chain GPU work and uses a nonstandard miner-direct route. It cannot protect coins whose public keys are already exposed.
Migration Challenges Pose Greater Risk Than Technology Itself
The primary challenge lies in governance and migration logistics rather than quantum computing advancement. A report from the Coinbase Independent Advisory Board on Quantum Computing and Blockchain noted that approximately 1.7 million BTC spread across roughly 20,000 early P2PK public keys is stored in wallets believed controlled by Satoshi Nakamoto or users who have lost access. Setting migration deadlines risks freezing these coins or exposing them to compromise.
Ethereum targets quantum resistance across execution, consensus, and data by December 2029 and plans to prepare for the transition as early as 2030. The G7 Cyber Expert Group has called for coordinated post-quantum migration efforts across the financial sector.
Market risks and the ambiguity surrounding migration processes, custody methods, and dormant coins present challenges that may materialize before any quantum computer capable of breaking secp256k1 cryptography emerges.
