A peer-reviewed critique published in Nature has challenged Microsoft’s landmark 2025 claim of having created the first topological qubits, alleging that the evidence was shaped by coding errors and selective data presentation.
The critique, written by University of St Andrews lecturer Henry Legg, was published in Nature’s “Matters Arising” section, the journal’s formal venue for challenging previously published findings. It targets Microsoft’s Topological Gap Protocol (TGP), an automated test designed to eliminate human bias and prevent false positives in quantum device validation.
Legg identified two specific coding errors. The first was a hardcoded filter that forced the TGP to “display only the single largest region,” concealing other results from phase maps. The second involved a Python array reversal by index position rather than by actual physical bias voltage values, transforming the data in ways that favoured the desired conclusion.
When peer reviewers asked whether other regions had passed the protocol, Microsoft reportedly stated they had investigated the only such region, a claim Legg says was not correct.
The TGP was supposed to prevent exactly the kind of false positives that forced Microsoft to retract Majorana-related papers in both 2018 and 2021. Legg’s overall assessment is blunt: Microsoft is “centuries, not decades away” from a working topological quantum computer.
Microsoft is not conceding. Chetan Nayak, Technical Fellow and Corporate Vice President of Quantum Hardware, stated: “We stand by our results and our roadmap.” The company characterised the issues as minor bugs and pointed to DARPA’s evaluation of the program as support for its results. Separately, University of Pittsburgh physicist Sergey Frolov suggested the original Nature paper “likely needs to be retracted”, a sharper verdict than Legg himself issued.
The challenge marks the third time Microsoft has faced serious scientific scrutiny of its Majorana research, following retractions in 2018 and 2021.
Sources: A new paper argues Microsoft exaggerated its quantum claims a year ago (The Verge, June 24, 2026); Microsoft Majorana Quantum Claim Challenged in Nature Critique (AI Weekly, June 24, 2026)