Google Quantum AI announced on Thursday that its Willow processor has experimentally demonstrated “magic state cultivation,” a technique that could remove one of quantum computing’s most serious bottlenecks. Magic states are special quantum states required for universal quantum computing; producing high-quality, error-corrected versions has been so resource-intensive that some estimates suggested up to 95% of a machine’s capacity could be consumed by making them.
Google researchers Craig Gidney, Noah Shutty, and Cody Jones began developing the approach in 2024 as a kind of filter. The team described injecting a noisy state into a small region on a quantum chip, then running highly coordinated measurements to check for errors and remove flawed attempts. On the Willow processor, magic state cultivation achieved a 40-fold reduction in errors and T-states with a record logical fidelity of 99.99%. Google cautioned that larger-scale demonstrations remain necessary before the method can support practical fault-tolerant quantum computing.
The announcement comes as the crypto industry works on post-quantum cryptographic protections. A sufficiently powerful quantum computer could theoretically break digital signature schemes used to authorize cryptocurrency transactions, as well as public-key cryptographic systems across the internet and financial infrastructure. Earlier this week, Coinbase published technical priorities for Bitcoin’s post-quantum transition, urging the community to begin preparations because the process would be complex and multi-year. Google has previously set an internal goal to transition to post-quantum cryptography by 2029, while some observers have predicted “Q-day” as early as 2028. Separate research estimates have suggested that around 1,200 logical qubits could threaten Bitcoin.
Castle Island Ventures co-founder Nic Carter said quantum computing timelines are generally compressing, with AI-assisted improvements contributing to acceleration. “There’s no imminent threat, and there remain engineering challenges to building a scaled QC, but it is another reminder that AI is meaningfully accelerating the rate of progress,” Carter said. Project Eleven CEO Alex Pruden added that the discovery does not necessarily move Q-day timelines because the technique was already assumed in prior resource estimates, but it does validate those assumptions and makes the current Q-day timeline more certain.
