Google Reduces Quantum Magic-State Errors 40-Fold on Willow Processor
Google achieves near-perfect quantum state preparation
Google Quantum AI announced Oct. 15 that it reduced errors in preparing "magic states" by roughly 40 times on its Willow quantum processor. The result, published in Science, achieved a state fidelity of at least 0.9999(1)—about 99.99%.
Magic states are special quantum states used in error-corrected quantum computing. They are consumed during operations needed for universal quantum computation. Producing high-quality magic states has been a major bottleneck because they typically require large amounts of a quantum computer's resources.
The researchers reported retaining about 8% of their preparation attempts after filtering out errored results. The benchmark included two additional cultivation rounds and three quantum error-correction cycles to verify the state's quality.
Key numbers from the study
- 40-fold reduction in magic-state preparation errors
- State fidelity of at least 99.99%
- 8% of preparation attempts retained after filtering
- Injected state error rate: about 0.44%
- Cultivated state error upper bound: about 0.01%
- Willow processor contains 105 physical qubits
What the research shows
The technique works like a filter: researchers inject a noisy quantum state, perform coordinated measurements, and discard attempts where errors are detected. This process is called "magic-state cultivation." Google's earlier preprint described the method, and the new Science paper reports the experimental validation on Willow.
The team also demonstrated transferring quantum states between two different error-correction encodings—moving from a distance-3 color code to a distance-5 surface code. This is a step toward more flexible quantum architectures.
However, the researchers acknowledged that scaling remains unfinished. Google's Emma Rosenfeld and Craig Gidney wrote that the team still needs to demonstrate magic-state cultivation at a larger scale to make it realistic for practical quantum computing.
Why this matters for crypto
The advance strengthens the experimental basis for resource estimates used in quantum computing threat models. Google's own March study modeled breaking Bitcoin and Ethereum signatures using Shor's algorithm and already assumed magic-state cultivation in its calculations.
That study estimated an attack on secp256k1 signatures would need between 1,200 and 1,450 logical qubits and either 70 million or 90 million Toffoli gates—a type of quantum logic operation. A faster modeled attack could run in roughly nine minutes after public-key exposure, using fewer than 500,000 physical qubits on superconducting hardware, assuming advance computation of part of the algorithm.
Project Eleven CEO Alex Pruden noted that cultivation was already assumed in earlier quantum resource estimates. The 40-fold improvement measures state error, not a 40-fold reduction in qubit count or attack runtime.
Pruden said on X: "This doesn't necessarily move Q-Day timelines forward. But it does experimentally validate one of the key assumptions underlying those estimates, making them more credible."
What is confirmed
Google achieved 99.99% state fidelity in magic-state preparation on Willow. The result validates a technique already used in quantum resource estimates for cryptographically relevant attacks. The experiment prepared a single logical qubit using 105 physical qubits. The researchers demonstrated transfer between two error-correction codes.
What is still unclear
The technique has not yet been demonstrated at a larger scale. It remains to be shown whether magic-state cultivation can be sustained as quantum computers grow. No cryptographic attack was demonstrated.
Official source
The study was published in Science on Oct. 15. Google also released a research post describing the cultivation technique and a preprint with additional details.
What happens next
Google researchers said the next step is demonstrating magic-state cultivation at a larger scale to make it realistic for practical quantum computing.
Sources
Newisty Editorial Team covers technology, cryptocurrency, digital products, online platforms, developer tools and the wider digital economy. Our content is researched from official sources, company announcements, public documentation, market data and other primary or reputable sources. Articles are reviewed and edited before publication for clarity, accuracy and useful context.
Related Posts
Russia's Largest Exchange Plans Bitcoin Trading This December
Trezor and BitBox warn users of phishing emails mimicking security alerts
Solana Passes Historic Inflation Vote as Schwab Adds SOL to Crypto Platform
