Google Quantum AI demonstrates new method to reduce quantum computing errors

Google Quantum AI has experimentally demonstrated a new method called ‘magic state cultivation’ to address a long-standing bottleneck in quantum computing. The technique, published in the journal Science, was tested on the company’s Willow quantum processor.

For two decades, researchers have struggled with the high resource cost of creating ‘magic states,’ which are essential for quantum systems to outperform classical computers. Previous methods, such as ‘magic state distillation,’ were extremely resource-intensive, potentially consuming up to 95% of a quantum computer’s total computational power.

Unlike traditional distillation, the cultivation method acts as a filter. It involves injecting noisy states into small areas of the quantum chip and repeatedly performing measurements to discard failed attempts, effectively ‘growing’ high-quality states. In experiments, this approach reduced errors by approximately 40 times compared to not using cultivation, achieving a fidelity of at least 99.99%.

The team also successfully implemented ‘grafting,’ a process of moving states created with a color code error correction method into a larger surface code area. While Google noted that technical challenges remain and larger-scale demonstrations are required, this achievement marks a significant step toward realizing practical, error-corrected quantum computers.