AI Challenge Cuts Quantum-Safe Bitcoin Build Cost 79% to $67
The estimate reflects GPU computation, not a Bitcoin transaction fee. A later dashboard listed $63 under a simulated 100-RTX 3090 setup.
An open optimization challenge cut the modeled GPU-compute cost of constructing a Quantum-Safe Bitcoin (QSB) transaction by 79% in one week, making a limited quantum-defense fallback less expensive to prepare.
The estimate fell from about $320 to roughly $67 on Sept. 23. The figure represents computing resources used before broadcast, not a Bitcoin transaction fee or a fixed production price. A challenge dashboard accessed Sept. 27 listed a later estimate of $63 for a simulated fleet of 100 RTX 3090 GPUs.
StarkWare, Yukon Research and Eigen Labs launched the Quantum-Safe Bitcoin Optimization Challenge on Sept. 16. StarkWare offered $20,000 in prizes, with an additional prize from Yukon Research, and invited developers, researchers and artificial intelligence agents to improve two workloads: transaction pinning and subset selection.
The benchmark measures SHA-256 hashing, secp256k1 public-key recovery and searches involving nine omissions from 150 signature pushes. Transaction-pinning throughput increased from 146.09 million to more than 820 million verified candidates per second on an RTX 4090 benchmark. The competition had 62 promoted submissions across both tracks as of Sept. 23.
The dashboard accessed Sept. 27 listed a GPT-labeled submission at 979.2 million verified candidates per second on Sept. 26 at 9:21 a.m. ET (13:21 UTC). An Opus 5×2 submission reached 960.8 million at 2:56 a.m. ET (06:56 UTC) the same day. A GPT-6 Astra submission recorded 881.3 million at 4:43 p.m. ET (20:43 UTC) on Sept. 23.
The first QSB transaction was mined on Bitcoin’s mainnet Aug. 26 without a soft fork or protocol upgrade. Constructing it required about 3,100 GPU-hours across roughly 100 GPUs and cost approximately $320 in computing resources, excluding the network fee.
QSB adds a hash-based protection layer and uses an off-chain search to find a transaction whose hash fits the format Bitcoin normally expects for a signature. This shifts the expense into computation rather than a network charge. The benchmark does not itself create live Bitcoin transactions.
The method is designed for coins whose public keys have not already been exposed and protects a specific transaction construction, not Bitcoin as a whole. QSB transactions are nonstandard, bypass the ordinary mempool relay path and currently require direct submission to a miner.
A sufficiently capable quantum computer running Shor’s algorithm could derive private keys from exposed public keys and forge transactions. QSB does not make Bitcoin broadly quantum-safe.
StarkWare CEO Eli Ben-Sasson said Avihu Levy’s work had shown that “Bitcoin has no expiration date.” Ben-Sasson added, “A soft fork is still the better long-term answer, as StarkWare has argued for consistently, but it is no longer the only option.”
The Sept. 23 estimate continues earlier coverage of the falling QSB preparation cost, while the specialized format and computational requirements continue to limit practical use.
