During Build 2026, Microsoft announced its next-generation topological quantum chip, "Majorana 2." This chip not only achieves significant breakthroughs in hardware materials but also, for the first time, deeply integrates with Microsoft's own "Microsoft Discovery" proxy AI platform, improving the reliability of qubits compared to its predecessor."Majorana 1"A dramatic increase of 1000 times.
From Aluminum to Lead: The Hardware Secrets Behind the Majorana 2's 1000-Fold Increase in Lifespan
Last year, Microsoft's Majorana 1 first demonstrated the possibility of creating entirely new states of matter using "topological superconductors," thereby stabilizing quantum computing. This year's Majorana 2, however, involves extremely bold adjustments to the stacking of materials.
Microsoft Technical Fellow Chetan Nayak pointed out that the original Majorana superconductor used aluminum, but in the new generation of chips, the team switched to lead, which has extremely strong radiation protection capabilities. In the extremely sensitive environment of quantum computers, lead superconductors can effectively protect fragile qubits from cosmic rays and other interference that could cause instability.
This material change has led to dramatic results. In the past, the industry typically measured the lifetime of qubits in "microseconds"; but Majorana 2 offers an average qubit lifetime of up to 20 seconds, and in some cases it can even last for nearly a minute.
Microsoft describes this as equivalent to increasing a phone's battery life from one day to nearly three years on a single charge.
With a 1000-fold increase in reliability, extremely fast computing speed measured in microseconds, and an extremely small size of only about one-hundredth of a millimeter, Microsoft firmly believes that this development path will enable quantum computers to officially step out of the laboratory in 2029 and solve complex human problems such as global health, sustainable development, and energy production.
Microsoft Discovery is now fully open: AI agents are becoming the most powerful tool for interdisciplinary scientists.
The biggest contributor to Microsoft's ability to halve the timeline for the advent of quantum computers is the "Microsoft Discovery" advanced research and development platform, which was also announced to be fully open today.
This is a platform designed specifically for scientific research, integrating an autonomous AI agent team with a dedicated Discovery Engine. Simply put, it allows scientists to deploy a whole group of human-guided AI assistants to automatically sift through massive amounts of data spanning up to 20 years, thereby generating hypotheses, optimizing experimental parameters, and even automatically performing extremely difficult quantum measurements.
Zulfi Alam, Corporate Vice President of Microsoft's Quantum Business, stated that designing quantum devices is a precise engineering process "atom by atom." In the past, a massive number of experiments were needed to find the correct material ratios; now, through simulations and predictions by proxy AI, AI can accurately pinpoint the target locations most likely to succeed, and ideally, the correct material ratios can be quickly determined with just one experiment.
Furthermore, when faced with a large development team spanning multiple countries and disciplines (including physics, mechanical engineering, and process engineering), the AI agent acts as a perfect "interdisciplinary translator" and data sorter, capable of processing large amounts of fragmented professional information in parallel and automatically finding connections that humans cannot perceive.
Microsoft has also launched an early preview version of the Microsoft Discovery application with core functions. Users can download it for free through their GitHub Copilot account and run it locally, thereby significantly lowering the barrier for businesses and academia to enter AI-driven scientific research.
