AI competition slashes cost of quantum-resistant Bitcoin transactions
An open competition cut the estimated GPU computing cost of a quantum-resistant Bitcoin transaction from about $320 to $67 in one week. StarkWare, Yukon Research, and Eigen Labs organized the challenge, and AI models produced the best benchmark results. The method remains limited and does not yet provide complete protection against quantum threats.
The challenge, run by StarkWare, Yukon Research, and Eigen Labs, sought faster ways to create Bitcoin transactions designed to withstand quantum attacks. The first such transaction was mined on mainnet last month. Its original construction required roughly 3,100 GPU hours and about $320 in computing costs. The competition reduced that estimate to around $67.
The costly step involves brute-force searching before blockchain data is recorded. A hash is placed where the network normally expects a signature, and a valid result appears only about once in 70 trillion attempts. On an RTX 4090, benchmark throughput rose from 146 million to over 820 million candidates per second, with 62 improvements. AI models produced top results.
This cost reduction could make quantum-resistant Bitcoin transactions less resource-intensive for miners, developers, and wallet providers exploring future safeguards. It may also accelerate AI-assisted optimization in security research. However, because the approach remains non-standard, requires direct miner submission, and protects only coins whose public keys are not yet exposed, it may not meaningfully protect most users against quantum threats. Its broader impact could depend on whether such methods become standardized or require a network software change.