The experiment addresses a challenge in building fault-tolerant quantum computers: preparing a key resource with sufficiently low errors. Its focus is that preparation method, rather than a demonstration of a complete, large-scale quantum computer.
The researchers also tested whether the method could be combined with a surface code, a form of quantum error correction. The result is promising, but larger-scale tests and further technical work remain.
What Did the Researchers Find?
The researchers reported that magic state cultivation reduced state-preparation error by a factor of 40. The cultivated state had a reported fidelity of at least 0.9999(1), or about 99.99%.
Fidelity describes how closely a prepared quantum state matches the intended one. The team retained 8% of attempts for the reported result, so the high-fidelity outcome came from a minority of attempts.
How Did the Team Test It?
The team implemented magic state cultivation on superconducting quantum hardware, rather than presenting cultivation only as a theoretical proposal. The experiment then used code-switching to transfer the cultivated state into a surface code. The paper was published in Science on 15 October 2026.
A surface code is a method of organising quantum information so that errors can be detected and corrected. Code-switching connected the cultivation procedure with that error-correction approach.
The researchers also developed a fault-tolerant measurement protocol to place a bound on the state’s fidelity.
Why Do Quantum Computers Need Magic States?
Fault-tolerant quantum computing needs a universal gate set: a collection of operations capable of carrying out general quantum computations. The challenge is that non-Clifford gates can carry a significant resource cost in many error-correction architectures.
Magic states are a quantum-computing resource that can help address that cost. Preparing them with low error is therefore important to approaches that aim to make quantum computations reliable despite hardware errors.
Magic state cultivation is presented as an alternative to resource-intensive magic state distillation. The Willow experiment tested cultivation as a way to prepare the resource, while also connecting it to a surface code.
Why Does the Result Matter?
The experiment demonstrated magic state cultivation on a quantum processor and combined it with code-switching into a surface code. That makes the result more than a proposal: the researchers tested the approach on superconducting quantum hardware.
The experiment established magic state cultivation as a viable solution to an important quantum-computing challenge. The measured reduction in preparation error, alongside the surface-code test, provides experimental evidence for the method’s potential role in fault-tolerant computing.
However, a successful experiment does not by itself show that the method is ready for a useful, large-scale quantum computer. Google said the approach still needs larger-scale testing before it can be considered practical.
What Is Still Unknown?
The experiment retained 8% of attempts for its reported result. That means the demonstrated high-fidelity outcome was not obtained from most attempts in this test.
Google said cultivation still needs to be demonstrated at a larger scale. The company also said technical challenges remain before the method can support a useful, error-corrected quantum computer.
The result therefore establishes a promising experimental approach, not a finished route to large-scale fault-tolerant computing. How well cultivation performs in larger tests, and whether the remaining technical challenges can be overcome, are still open questions.





