The Google Quantum AI team has successfully implemented a magic state cultivation method on its Willow processor, as detailed in the journal Science. This approach reduces error rates by a factor of 40 compared to operations performed without it.
Magic states are essential for universal quantum operations but previously consumed up to 95% of a quantum computer's processing power. The new technique acts as a filter, discarding noisy states and retaining clean ones through coordinated measurements.
During the experiment, the team achieved a record 99.99% accuracy in preparing T-states. Researchers also demonstrated the first successful transfer of a cleaned state between two distinct error correction methods, specifically color and surface codes, on physical hardware.
This development validates a theoretical concept introduced in 2024 and reduces the hardware scale required for practical quantum computing. The team now aims to demonstrate the method at a larger scale to build fault-tolerant systems.
