Morning Edition · Saturday, October 17, 2026Published at 1:38 AM EDT · New York
Nine crypto firms launched a Quantum Alliance the same day, estimating that 6.4 million bitcoin carry exposed public keys, though its members disagree over whether coins that are never migrated should be frozen.

Two research results published within a day of each other reduce different parts of the cost of a future quantum attack on bitcoin's signatures. Neither one breaks bitcoin today. Taken together, they show that the hardware cost and the algorithm cost are falling at the same time.
The first result came from Google Quantum AI. In a paper published in Science on October 15, researchers reported using a technique called magic state cultivation on the Willow superconducting processor to prepare the specialized quantum states that error-corrected machines need for advanced calculations. CoinDesk reports that the states had about 40 times fewer errors than states prepared without the method. Google called the resulting 99.99% fidelity a record. The process remains costly. According to crypto.news, only about 8% of attempts passed every check, so the chip discards more than nine of every ten tries to keep one clean state.
The second result came from Anthropic. The company said its Claude Mythos model designed quantum circuits that could bring quantum computing "within reach sooner." Interesting Engineering reports that, working from plain-language instructions, the model outperformed the best published designs on seven of eight tasks. One design used 17 times fewer gates in exchange for 12% more qubits. Among the outputs was a reversible arithmetic routine for elliptic-curve calculations. Shor's algorithm needs exactly that kind of arithmetic to derive a private key from a bitcoin public key. Researchers from Harvard University and QuEra Computing took part in the work.
The industry response arrived on October 16. Nine companies, including the auditor Hacken, Cardano founder Charles Hoskinson's Midnight Foundation, Quantus and Postquant Labs, formed the Quantum Alliance to draft security standards. The group estimates that about 6.4 million bitcoin, or 30.2% of supply, sit in outputs whose public keys are already visible on-chain. PANews reports that this figure includes roughly 1.1 million coins attributed to Satoshi Nakamoto. Glassnode's earlier estimate, cited by crypto.news, is lower at 6.04 million.
The members agree on the exposure and disagree on the remedy. TFTC reports that Postquant Labs' chief technology officer favors migrating coins to quantum-resistant outputs. Bryn Bennett of The Vault opposes freezing coins, arguing that a freeze would break bitcoin's promise against confiscation. The incentives are worth noting. Several alliance members sell post-quantum products, and an urgent timeline would help their business. That conflict of interest does not make the published hardware and circuit results false. It does mean their claims about timing deserve more scrutiny than the papers themselves. CoinDesk recorded bitcoin as little changed near $82,100 on October 16, so the market has not priced either result as an immediate threat.
Part of a tracked trend
Quantum Risk Moves From Theory to Crypto Roadmaps
Over the next 3-6 months, quantum-resistance becomes a concrete design and custody concern across major chains and institutional custodians, driving opt-in post-quantum schemes and 'long-dormant coin' risk discussion.
Start a discussion in Townsquare.
More from this edition
If the shorter timeline is true, the alliance members that sell post-quantum security products gain customers, and Google and Anthropic gain research credit. Holders of the roughly 6 million bitcoin whose public keys are already visible on-chain carry the risk that those coins could later be stolen and sold.
Google's Science experiment used 54 of Willow's 105 physical qubits. Anthropic's 17-fold gate reduction applies to modular inversion, a single step in elliptic-curve arithmetic. Neither result runs Shor's algorithm against a real key, and the 2028 break date put forward by Postquant Labs' chief technology officer, Richard Carback, comes from a firm that sells post-quantum products.
An open-source-intelligence read of how likely this story is true with its real nuance, not a judgment of any outlet. It assesses the claim, weighing independent and adversarial reporting. How we label confidence.
What this means
Quantum risk enters bitcoin's price through one channel. About 30% of the supply sits in addresses whose public keys are already exposed, much of it in dormant early coins. If holders come to believe those coins could be stolen, the market would have to price in a possible release of that supply. The governance question adds a second exposure, because a protocol freeze would set a confiscation precedent that works against bitcoin's core promise of protecting property rights. The market already accounts for the fact that no machine can break a key today, given that only about 8% of magic-state attempts succeed and error rates remain significant. What is new is that hardware errors and circuit size fell in the same week, from two independent labs. The common view treats this as a question of hardware timing. The tighter constraint is bitcoin's governance, because migrating to new outputs requires a soft fork and then years of individual users moving their coins. The SHA-1 hash function offers the closest precedent. Theoretical attacks on it appeared in 2005, but most systems completed migration only after Google and CWI Amsterdam published a practical collision in 2017. The likelier outcome over the coming months is that custodians and exchanges publish migration plans and that opt-in post-quantum address proposals gain developer attention. Freeze proposals will probably stay unadopted, because opponents can block the broad consensus a freeze would require. This forecast would be wrong if a freeze or sunset proposal gains merged code or miner signaling, or if a lab runs elliptic-curve arithmetic on error-corrected logical qubits.
What to watch
Observations to monitor, not financial advice.
Synthesized from: CoinDesk · CoinDesk · Polylog editors · crypto.news · Interesting Engineering · PANews · TFTC
Comments
1Oct 17, 5:38 AM
At an 8% pass rate each clean magic state costs about 12.5 attempts, so the 40 fold error cut comes with a throughput cost that an attack pays again for every such state it consumes.