Nvidia's announcement on accelerating quantum computing stirred up the crypto community with talk of a possible Bitcoin hack. We break down the technical essence of the development and assess how justified the concerns are.
At the SC26 conference, Nvidia presented two projects in the field of quantum computing. The first is a joint platform with IBM and the American Fermilab laboratory, where Nvidia DGX B200 systems will operate alongside IBM quantum computers. The second is the TCBP-JOBS algorithm, which accelerates error correction in quantum computers.
It was the second project that caused a stir. Qubits — the basic elements of a quantum computer — are unstable and often make errors. For a machine to compute correctly, it needs a decoder that calculates the location of the failure. TCBP-JOBS is exactly such a decoder.
According to Nvidia, it runs up to 13,000 times faster than leading analogues. One calculation for studying new quantum computer designs was reduced from 12 weeks to 45 minutes. The acceleration applies to research tasks in which scientists test variants of correction codes. There is no mention of breaking encryption in the announcement.
Why it's too early to bury Bitcoin
The panic-mongers do have grounds for alarm. A quantum computer of sufficient power would be able to compute a private key from a public one and drain other people's coins. Earlier, the Google Quantum AI team published calculations on how many resources this would require. The result was fewer than 500,000 physical qubits — 20 times less than previous estimates.
After that, researchers named timelines. Justin Drake, a researcher at the Ethereum Foundation and co-author of the work, estimated the probability of a private key being hacked by 2032 at at least 10%. Craig Gidney, a Google researcher, cited the same risk as early as 2030 and advocated a transition to quantum-resistant cryptography by 2029.
But a real attack is still far off. These estimates were voiced even before Nvidia's announcement, and at that time experts explained why there is no need to rush to conclusions:
- modern quantum computers operate with roughly a thousand physical qubits, while hundreds of thousands are needed for a hack;
- Adam Back, head of Blockstream, noted that Google's work describes an improvement in algorithms, not hardware;
- Bernstein analysts stated that the risk is not existential, not new, and does not concern cryptocurrencies alone;
- Alex Thorn, head of research at Galaxy Digital, assessed the chances of an attack on Bitcoin in the next five years as low.
Adam Back called current quantum systems extremely primitive due to limitations in error correction. Nvidia's decoder accelerates research in precisely this area. But it is a tool for scientists, not a ready-made machine for hacking.
Investors' coins are not under threat right now. The risk applies to addresses whose public key has already been revealed — it becomes visible after the first spend from the address. Migrating Bitcoin to new cryptography will take years, so it is worth watching how quickly developers prepare a network upgrade.
Cryptalist analysis: Nvidia's announcement reflects progress in error correction, not in cryptanalysis, and does not change the fundamental assessment of the quantum threat to Bitcoin over the coming years. Institutional investors would be wise to view this risk as a long-term structural factor requiring monitoring of the pace of migration to post-quantum cryptography, rather than as a trigger for immediate action.