Nature Casts Doubt on Microsoft's Topological Qubit

Quantum computing rests on qubits, and Microsoft's qubits rest on particles that physicists have yet to observe directly — Majorana zero modes. In February 2025 the company claimed it had recorded the signatures required for a topological qubit in hybrid structures of indium arsenide and aluminum. On June 24, 2026, the journal Nature published a formal critique of that conclusion by Henry Legg of the University of St Andrews — a paper with the dry title “On the reliability of topological gap detection by transport measurements.”

The objection strikes right at the load-bearing point: the topological gap protocol (TGP), which Microsoft proposed precisely to separate a genuine signal from a false positive, is in Legg's view insufficient to rule out a far more mundane explanation — the so-called Andreev bound states, a long-known source of confusion in nanowire experiments. Put another way: what the company took to be the first breath of a topological qubit may well be noise.

Microsoft was not to be outdone: granted a right of reply, the company published on the same day in Nature its disagreement with Legg's critique, calling its own work sound; and shortly before that, in early June, it had unveiled Majorana 2 — a next-generation chip whose development, by the company's own account, involved its own AI agent. None of this settled the dispute: as of mid-2026, there is still no peer-reviewed proof that the Majorana modes in Microsoft's device are not a measurement illusion — that they are signal and not noise.