A week of AI coding cut a quantum-safe bitcoin transaction estimate from $320 to $66
Reported by CoinDesk · AI-assisted summary by ChikoCorp AI News Desk
AI-generated summary based on the linked source; not independently verified. This is not investment advice. Verify market-moving details at the original publisher before acting. See our editorial policy, AI content policy, and financial disclaimer.
Summary
StarkWare reported that AI-assisted coding reduced the estimated GPU cost of creating a quantum-resistant Bitcoin transaction from $320 to $66 by speeding up the underlying search process. This quantum-safe transaction method could enable eligible Bitcoin holders to move funds if quantum computers threaten exposed public keys, and it does so without needing a Bitcoin network upgrade. The $66 cost is an estimate from computational tests and has not been demonstrated with another mined transaction.
Why it matters
The development offers an emergency option to protect coins against a potential quantum computing threat before Bitcoin adopts a broader quantum-resistant solution. Lowering the computing cost from $320 to $66 may make this protective method more practical for users. However, the source does not elaborate further on the broader market or user impact.
Key context
The initial $320 estimate was based on a prior transaction mined in August, which took about 3,100 GPU hours across roughly 100 graphics processors before being submitted to Bitcoin. StarkWare's method uses hash-based protections expected to withstand attacks from quantum computers and fits within Bitcoin's existing protocol rules, thus not requiring a network upgrade. Transactions created with this method must be sent directly to miners and do not protect coins whose public keys are already exposed.
Key numbers and entities
StarkWare is the company behind the research. The original transaction cost estimate was $320, reduced to a $66 estimate after AI-assisted coding improvements. The starting code checked roughly 146 million candidates per second on a benchmark GPU, while the improved submission checked 881 million candidates per second. The methods require about 3,100 GPU hours on 100 GPUs for the initial transaction.
What remains unclear
It is not clear how replicable the $66 cost figure is since the improved code has not been demonstrated preparing another mined Bitcoin transaction. Details on which specific improvements bring the estimate down to $66 are not provided. The source also does not clarify the practical effectiveness or limitations of the method once widely adopted, beyond the need to send transactions directly to miners and its inability to protect already exposed public keys.