Microsoft Refuses to Release the Data Behind Its 1,000x Quantum Bit

Microsoft's quantum chief has a short answer for the physicists who won't endorse his team's latest claim. "I don't really care," Zulfi Alam, Microsoft's corporate vice president of quantum, told WIRED when asked about the doubters. "Why should I?"

That shrug is the most visible edge of a private dispute over which mechanisms — peer review, independent replication, and access to raw data — are supposed to protect the public when a frontier-tech company asks the world to take a foundational physics claim on commercial-sensitivity grounds. The cost of getting the answer wrong is not a corrected paper. It is a quantum computer the field, the public, and the next generation of encryption standards may already be building on.

A quantum computer processes information in qubits, the basic unit of a machine that, if it works at scale, could one day simulate molecules for drug discovery, design new materials, or break the encryption that protects ordinary internet traffic. In June, the software giant said it had produced a qubit that is 1,000 times more reliable than its predecessor. Alam's team says it effectively split an electron into so-called Majorana quasiparticles — a kind of behavior in a material that Microsoft is betting its quantum design on. Majorana particles were first hypothesized nearly 90 years ago, and many physicists say no one has yet seen one in a way that survives independent scrutiny.

The company has not released the underlying data. It cites commercial sensitivity. Independent physicists, several of whom told The Verge the data already published in earlier Microsoft papers was inconclusive, say they have no way to check the new claim.

This is not Microsoft's first turn through the same loop. In 2021, the company retracted a Nature paper that had claimed evidence of Majorana particles after independent physicists identified problems in the data analysis. The retraction preceded Alam's tenure running the program, and Microsoft says the current work is built on different materials and a different measurement approach. The dispute today is not over whether the 2021 paper was wrong; it is over what gate is now standing between the company's new claim and the public.

The two gates science normally relies on are peer review and independent replication. Peer review asks other physicists to evaluate the work before publication. Replication asks someone else to run the experiment and see if it holds. Microsoft has published the current work in Nature, which is peer reviewed. The company is declining to share the raw measurements other labs would need to try to replicate it. Alam tells WIRED the publishing process is too slow and that the field's verification norms are not built for a company trying to ship a product on a deadline.

Public agencies are already planning for the day a useful quantum computer could break today's encryption, and that planning assumes the physics will eventually work. Microsoft's bet is on a particle that, by Microsoft's own history, has tripped the same program once before.

Microsoft is not alone in the race. A wave of well-funded startups are pursuing competing quantum approaches, including neutral atoms, trapped charged atoms, and photons. The Verge's coverage of the current dispute lays out how each design bets on a different physical quirk to make a working qubit. Microsoft's wager is on Majorana. So far, neither the company nor independent labs have produced evidence the rest of the field accepts as definitive.

Microsoft has said a scaled-up, fault-tolerant prototype based on the current design is targeted for the end of the decade. The first independent measurement, run on equipment the company has so far declined to share, would close the question. Microsoft has not said who runs that measurement, or when.