Daily Crypto Briefs 
NEW YORK, September 24, 2026
StarkWare says an AI-assisted coding contest cut the estimated computing expense of preparing a quantum-resistant Bitcoin transaction by 79%, to about $67 from $320, making an experimental emergency safeguard cheaper to test while Bitcoin weighs broader defenses.
The September 23 research update reports faster code for two GPU-intensive searches in its Quantum-Safe Bitcoin, or QSB, method. The improvement is a benchmark result, not a new Bitcoin network upgrade or a second optimized transaction confirmed on-chain.
Bitcoin traded near $83,413 at about 12:04 UTC Thursday, down about 2.4% over 24 hours, with a market capitalization near $1.68 trillion and roughly $43.9 billion in daily volume, according to CoinGecko’s market page. Those price moves are market context; the research announcement does not establish that it moved BTC.
StarkWare said the contest’s dashboard applies the measured speedups to the actual GPU expense of its August transaction. Its published $67 figure is a modeled compute estimate and excludes the separate fee for recording a transaction on Bitcoin. The live challenge dashboard displayed about $66 at 12:05 UTC September 24 as later submissions changed the estimate.
The result advances StarkWare researcher Avihu Levy’s earlier no-fork wallet proposal from a working but costly proof of concept toward a cheaper fallback. That is a material new benchmark and public contest result, not evidence that ordinary Bitcoin wallets have acquired quantum protection.
GPU benchmark cuts Bitcoin fallback estimate
The original QSB transaction, mined in August, required roughly 3,100 GPU-hours across about 100 graphics processors. StarkWare put that off-chain computing expense at approximately $320. Its September 16 challenge, run with Yukon Research and Eigen Labs, invited developers and AI agents to optimize the search code rather than change Bitcoin’s rules.
In the September 23 update, StarkWare said the transaction-pinning workload passed 820 million verified candidates per second on the benchmark RTX 4090, up from 146.09 million at baseline. The separate subset-selection workload reached about 623.5 million candidates per second, up from roughly 62 million. Together, the two tracks had 62 promoted submissions from 23 pinning solvers and 17 subset-selection solvers at that time.
Those numbers measure how quickly the software checks possible answers to the puzzle needed to construct a QSB spend. StarkWare said submissions are checked against a CPU reference implementation so an apparent speed gain cannot come from skipping required work. Faster searches can lower GPU-hours, but the dollar estimate also depends on hardware and cost assumptions.
The distinction is visible in the challenge methodology: its current estimate models a fleet of 100 RTX 3090 GPUs, applies measured improvements to the earlier transaction’s cost components and retains an unallocated cost portion. The site also notes that newer RTX 4090 records extend beyond a measurement cutoff. A dashboard figure should therefore be read as an estimate that can change, not a demonstrated invoice for the next user’s transaction.
Bitcoin
BTCQuantum-safe Bitcoin still has narrow reach
QSB aims to let an eligible holder move coins using a hash-based protection within Bitcoin’s existing consensus rules. A sufficiently powerful future quantum computer could attack the elliptic-curve signatures used to authorize ordinary Bitcoin spends. The method is designed as an emergency path for a coin whose public key has stayed hidden until it is spent.
It does not protect coins with public keys that are already exposed, a group at the center of the earlier debate over dormant Bitcoin. Nor can a QSB transaction currently travel through the ordinary peer-to-peer transaction relay: StarkWare says it must be delivered directly to a miner. Those are limits on who can use the method, independent of the GPU bill.
StarkWare also cautions that the competition’s synthetic benchmark reproduces the expensive search core but does not itself assemble live Bitcoin transactions. The latest code must still be tested in the production implementation and in another mined spend before its practical end-to-end cost can be established.
The research is about preparedness for a possible future quantum threat, not a reported theft, active wallet compromise or evidence that a cryptographically relevant quantum computer exists today. A cheaper fallback could be useful for some holders if the threat emerges before a broader network change, but it would not settle Bitcoin’s migration problem.
Bitcoin’s broader quantum upgrade remains open
StarkWare says a protocol-level soft fork remains the preferred long-term route for efficient, widely usable protection. The QSB workaround is valuable precisely because it works under today’s consensus rules while that larger design and adoption process continues.
Other proposals, including the BIP-360 discussion, address a wider migration path for vulnerable wallet types. They have different requirements and would need community review and deployment before changing how Bitcoin users spend. A benchmark gain in QSB does not approve or activate any of them.
Market sentiment remains a separate measure. The Alternative.me index registered 71, or “Greed,” on September 24, but that broad Bitcoin sentiment gauge does not measure confidence in quantum defenses.
Fear & Greed Index
September 24, 2026The next concrete test is whether the faster kernels can build and confirm another QSB transaction at a documented total compute expense. Until then, the $67 company estimate and the dashboard’s changing $66 reading describe benchmark-based cost modeling, while the August $320 transaction remains the mined reference point.
Stay up to date
Get the latest crypto insights delivered to your inbox
Primary sources and further reading
| Source | Title |
|---|---|
| | StarkWare: AI research competition cut Quantum-Safe Bitcoin costs by 79% in a week |
| | Yukon: Quantum Safe Bitcoin challenge dashboard and methodology |
| | StarkWare: Quantum-Safe Bitcoin optimization challenge launch |
| | StarkWare: first quantum-safe Bitcoin transaction mined |
| | CoinGecko: Bitcoin price, market capitalization and volume |
| | Alternative.me: Crypto Fear and Greed Index |
Fact-checked by: Daily Crypto Briefs Fact-Check Desk
Related Articles
Frequently Asked Questions
Does the $67 figure mean a Bitcoin transaction fee fell?
No. It is StarkWare's September 23 estimate of the GPU computing expense needed before submitting one Quantum-Safe Bitcoin transaction. It excludes the separate Bitcoin network fee and is not a market price or guaranteed user cost.
Has the optimized quantum-resistant transaction been mined?
Not yet, according to StarkWare's September 23 report. Its August demonstration was mined, but the latest speed gains came from a synthetic GPU benchmark that has not produced another confirmed Bitcoin transaction.
Does the method protect Bitcoin whose public keys are already exposed?
No. StarkWare says this fallback works for coins whose public keys have remained hidden. It does not retroactively secure already exposed keys, and its transactions must currently be submitted directly to a miner.
Is Bitcoin now safe from a future quantum attack?
No. The research concerns a narrow emergency transaction construction. StarkWare says a Bitcoin protocol upgrade remains the preferred longer-term path to broad quantum protection.