Google Cuts Quantum Error Rates 40-Fold With 'Magic State Cultivation'

Google's Quantum AI division announced on October 15, 2026, that it had demonstrated a technique called magic state cultivation on its Willow quantum processor, reducing errors in a specific class of quantum states by about 40 times compared with uncultivated states, according to a post on Google's official blog. The work was published the same day in the journal Science.

What Google announced

The experiment produced so-called logical T-states — a type of 'magic state' required for certain quantum operations — with a logical fidelity of 99.99%, which Google describes as a record for this kind of state. The paper was authored by Google researchers including Emma Rosenfeld and Craig Gidney, according to the company's blog post. The underlying cultivation technique was originally developed in 2024 by Google researchers Craig Gidney, Noah Shutty and Cody Jones, according to reporting by Interesting Engineering.

How magic state cultivation works

Magic states are highly accurate quantum states needed to run complex calculations that go beyond what classical computers can do, and each one can only be used once, per Interesting Engineering's account of the Google research. Cultivation works as a filtering process: a noisy quantum state is injected into a small region of the processor, coordinated measurements check it for errors, flawed attempts are discarded, and the states that pass are expanded for use in larger computations.

According to figures reported by the German outlet Blocktrainer, roughly 92% of attempts were rejected during the experiment — meaning only about 8% of runs produced a usable, high-fidelity magic state. Google's team also said it addressed practical interference problems, such as drift in laboratory electrical equipment and temperature changes, by improving processor controls and developing a new benchmarking method based on the cultivation process itself, according to Google's blog post.

Why magic states are a bottleneck

Google has previously estimated that producing error-corrected magic states could consume up to 95% of a quantum computer's total capacity, making their efficient creation one of the field's most severe resource constraints, as reported by Interesting Engineering and repeated in Google's own blog post. Reducing the error rate and the overhead needed to generate reliable magic states is seen as a prerequisite for building quantum computers capable of useful, large-scale computation.

What it means — and what it doesn't

Google's blog post describes the result as the first demonstration of a practical solution to the magic-state bottleneck, while cautioning that the technique still needs to be shown to work at a larger scale before it can support real-world quantum applications. The company did not provide a timeline for when cultivation might be deployed in a full-scale, fault-tolerant quantum computer.

The announcement has also drawn attention from cryptocurrency commentators because magic states and Shor's algorithm are theoretically linked to breaking certain cryptographic systems. The German outlet Blocktrainer reported that while the advance could, in principle, help make Shor's-algorithm calculations more efficient in the future, researchers remain far from actually breaking the ECDSA and Schnorr signature schemes that secure Bitcoin, and that the development instead underscores the importance of continuing post-quantum cryptography research.

Background

The announcement adds to a series of milestones Google has reported for its superconducting Willow chip as part of its broader quantum error-correction research program. The latest result focuses specifically on reducing the cost and error rate of magic-state production rather than on demonstrating a complete, application-ready quantum computer.