Microsoft's quantum claims face skepticism despite technical advances

In February, Microsoft Quantum announced it had built a processor called Majorana 1, claiming it contained the world's first topological qubits. The chip held eight of them, and the company said future versions would scale to a million. The same day, the team published its findings in Nature. But almost immediately, a gap opened between what Microsoft's press release promised and what the paper actually showed.

The Nature paper was notably more cautious. The researchers described building devices from indium-arsenide and aluminium, and they demonstrated a single-shot measurement of parity—whether the number of electrons in a quantum state was odd or even. They called this measurement a necessary ingredient for topological quantum computing. But the scope of what they had actually demonstrated fell short of what the company's public announcement suggested they had achieved. The discrepancy raised a question in the scientific community: was Microsoft claiming credit for a commercially valuable technology before proving it worked?

To understand the stakes, you need to know what topological qubits are supposed to do. Quantum computers promise to solve certain problems far faster than classical machines, but quantum bits are fragile. A qubit stores information in a quantum state that can be disrupted by stray heat or magnetic fields. Most quantum computers today use superconducting circuits, atoms, ions, or light. Microsoft Quantum has been pursuing a different path: qubits based on Majorana zero modes, exotic quasiparticles that physicists Sankar Das Sarma, Roman Lutchyn, and Jay Sau predicted in 2010 would appear at the ends of specially engineered superconducting wires. The key insight is that two of these modes, separated by distance, can store information together—both particles holding the data jointly rather than each holding half. This arrangement means small disturbances affecting only one end cannot destroy the information, yet both ends are needed to access it. The space between them is crucial.

But here lies the central problem: there is no way to directly observe a Majorana zero mode. Scientists cannot point a microscope at a device and declare success. Instead, they must measure various properties and infer whether the modes are actually present. Many impostor states are easier to create but lack the unusual properties that make Majorana modes valuable. The inference has to be airtight. In Microsoft's case, critics argue it is not.

On June 2, Microsoft Quantum announced Majorana 2, a new processor with a redesigned architecture. The team replaced aluminium with lead, switched to a gallium-antimonide substrate, and restructured the semiconductor stack to reduce disorder. The reported improvements were striking: the topological gap—the energy needed to excite the system out of its protected state—jumped from roughly 30 microelectron-volts to 70. The parity switching time increased from milliseconds to around 20 seconds. Microsoft presented these numbers as evidence that the encoded information was more robustly protected, and on that basis announced it would build a scalable quantum computer by 2029.

Yet skepticism persisted. Sergey Frolov, a professor at the University of Pittsburgh, said Microsoft's track record made him unwilling to treat the new work as resting on solid ground. The history of the field itself offered reason for caution. In 2018, researchers from TU-Delft published a Nature paper claiming they had detected Majorana zero modes. Three years later, the paper was retracted after independent scientists found problems in the data analysis. The lesson was clear: detecting these modes is extraordinarily difficult, and past confidence has proven premature.

Microsoft's 2025 paper relied on a protocol from 2023 that used electrical conductance data to identify conditions consistent with gapped Majorana modes, then applied radio-frequency interferometry to check one region where the team said parity switching occurred. At an American Physical Society meeting in July, physicists remained unconvinced. Eun-Ah Kim from Cornell called the data weak. Kartiek Agarwal at Argonne National Laboratory said the claims were premature. Henry Legg from the University of St. Andrews noted that the headline results came from a small number of measurements on a single device.

On June 24, Nature published a rebuttal from Legg alongside Microsoft's response. Legg argued that Microsoft's device may not have been qualified to host Majorana modes at all, that the electrical current data appeared consistent with a gapless, highly disordered system, and that Microsoft's protocol was sensitive to how researchers chose to analyze the data rather than agnostic to it. Microsoft's physicist Chetan Nayak disputed Legg's logic and said the company's claims rested on radio-frequency measurements, not the protocol alone. Many physicists acknowledge Microsoft's work is promising and praise its data transparency. But many remain unconvinced that Majorana 1 and Majorana 2 actually contain working topological qubits.