Microsoft's Majorana 2 Hits 1,000x Reliability Improvement, Cuts Scalable Quantum Computer Timeline to 2029

Started by Sophie86, Yesterday at 06:01 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: Microsoft's Majorana 2 Hits 1,000x Reliability Improvement, Cuts Scalable Quantum Computer Timeline to 2029   Views(Read 36 times)

Sophie86

Microsoft unveiled Majorana 2, its next-generation topological quantum chip, built with assistance from the company's own Microsoft Discovery agentic AI platform. The new chip's materials stack delivers a 1,000-fold improvement in qubit reliability over the prior generation, with a mean qubit lifetime of 20 seconds and individual instances persisting as long as one minute, a figure technical fellow Chetan Nayak compared to inventing a phone battery that lasts nearly three years on a single charge instead of dying in a day. Most competing qubit approaches measure lifetime in microseconds rather than seconds, making the scale of the reported improvement particularly striking if it holds up under independent scrutiny.

With this progress, Microsoft now expects to achieve a scalable, commercially valuable quantum computer by 2029, cutting its previously stated timeline in half. The underlying technology, topological superconductors built on Majorana zero modes, represents a fundamentally different approach to qubit construction than the superconducting transmon or trapped-ion methods pursued by most competitors, theoretically offering intrinsic protection against certain error types built into the physics of the qubit itself rather than requiring as much external error correction overhead. Majorana 2's qubits also operate at one-hundredth of a millimetre in physical size and execute operations in roughly one microsecond.

The development process itself is part of the story Microsoft is telling. The quantum team used Microsoft Discovery's agentic AI capabilities throughout development to manage experimental workflows, automate measurements, optimise fabrication processes, identify previously unnoticed material flaws and propose new solutions, with Nayak describing agentic AI as having become a natural part of the team's daily workflow rather than a novelty. Microsoft Discovery, the broader platform behind this capability, is now generally available, including a free local app version for individuals with a GitHub Copilot account, positioning the same AI-accelerated research methodology Microsoft used internally as a product other organisations pursuing materials science and engineering breakthroughs can now access directly.