RESOURCE
Google reports in Science that a technique called magic state cultivation works on its Willow processor, producing T states at 99.99 percent logical fidelity, roughly a 40 fold error reduction. Magic states are the scarce ingredient in error corrected computation, so cheaper magic states lower the hardware a useful machine would need.
Fabric Networks
Start with a conversation about your cryptographic posture.
On October 15, 2026 Google Quantum AI described an experiment, published in Science, in which it prepared high quality T states (the so called magic states that let an error corrected quantum computer run non trivial operations) on its Willow chip. The team reports 99.99 percent logical fidelity, about a 40 fold reduction in error compared with earlier approaches, and it moved the prepared states between two forms of error correction as a first test of fitting the method into a larger architecture.
Google is careful to say the resource problem is not solved at practical scale and that larger tests and better storage of prepared states are still needed. The reason the result matters is that producing magic states was expected to consume most of a future machine's capacity. If that cost falls, so does the size of machine needed for the calculations that matter to cryptography.
Every one of these results trims the assumed distance to a cryptographically relevant quantum computer, and none of them restores it. The data at risk was collected last year. Canada's deadlines (high priority systems by the end of 2031, all systems by 2035) were set with a margin; that margin is what is shrinking. Assess this fall, budget in 2027, secured by January 2028.