Microsoft Quantum Doubts Deepen
Microsoft's quantum computing claims face renewed scrutiny from Nature and UK physicist Legg
Key Points
Two Microsoft-backed papers have been retracted by Nature, and editors flagged alerts about possible research problems in two more papers (one in Nature and another in Science), highlighting ongoing scrutiny of Microsoft’s quantum claims.
Nature Matters Arising pieces feature Henry Legg’s critique that Microsoft’s topological qubit claim may be noise, with Legg’s response published in Nature and a Microsoft rebuttal also appearing in Nature.
The February 2025 paper at the center of the debate is described as not being retracted despite scrutiny; Legg characterized the mechanism as potentially a 'chaotic jumble of mismatched parts' suggesting the result could be noise rather than a breakthrough.
The U.S. policy backdrop underscores the stakes: the Trump administration has invested about $2 billion in quantum computing and aims to have a national quantum system by 2028.
A peer-reviewed critique published in Nature on June 24, 2026 accuses Microsoft of building its claimed quantum computing breakthrough on "basic Python errors" and cherry-picked data. The paper, written by UK physicist Dr. Henry Legg of the University of St. Andrews, targets Microsoft's February 2025 research introducing the "Majorana 1" chip — a device Microsoft said marked the dawn of topological quantum computing.
The dispute lands at a high-stakes moment. Just two days earlier, President Trump signed executive orders — including the "Quantum Genesis Initiative" — setting a 2028 deadline for a national quantum system, backed by roughly $2 billion in federal investment, according to Nextgov/FCW.
Legg Says Microsoft Found a 'Chaotic Jumble,' Not a Breakthrough
Microsoft's 2025 paper claimed its Majorana chip used topological qubits — a theoretically stable type of qubit based on exotic "Majorana zero modes" — to perform calculations far more reliably than rivals. Legg's rebuttal, published as a "Matters Arising" piece in Nature, says the software tool Microsoft built to verify those qubits contains coding errors. He argues the results are likely noise, not a genuine signal.
"Last year Microsoft claimed they had built the equivalent of a precision Swiss watch," Legg said, according to Mirage News. "I found what looked like a chaotic jumble of mismatched parts." He adds that simply shifting a measurement window in the software can flip a result from "failed" to "passed" — a finding that cuts at the heart of Microsoft's verification method.
Two Retractions Already on the Books, Skeptics Still Unsatisfied
This is not Microsoft's first collision with Nature over Majorana research. A 2018 paper from a Microsoft-funded team at TU Delft was retracted in March 2021 after investigators found data was selectively presented. A second related paper was retracted in 2022, according to Crypto Briefing. Authors of the 2021 paper blamed being "caught up in the enthusiasm of the moment."
Physicist Sergey Frolov of the University of Pittsburgh, who helped expose the 2021 retraction, told Scientific American that Microsoft's track record is not a "solid foundation." At the American Physical Society summit in March 2025, Frolov and Legg both described Microsoft's presented data as "random noise" — a "Rorschach test" that supporters see as a breakthrough and skeptics see as nothing at all.
Microsoft Fires Back, Unveils a New Chip 1,000x More Stable
Microsoft has not backed down. The company published a rebuttal in Nature and its corporate VP Zulfi Alam dismissed earlier criticism from Legg, calling him a "pontificator" who misread the data, according to PCMag. Microsoft's technical fellow Chetan Nayak argues that skepticism is a normal part of science and that progress vindicates the approach.
At its Build conference on June 2, 2026, Microsoft unveiled the "Majorana 2" chip. The company claims it is 1,000x more reliable than Majorana 1, with qubits lasting a mean of 20 seconds — up from 12 milliseconds on the previous chip, according to Tom's Hardware. Microsoft also moved its target for a scalable quantum computer up from 2033 to 2029, citing AI-assisted materials discovery as the key accelerant.
Washington Bets $2 Billion While the Science Remains Unsettled
The political stakes make the scientific dispute harder to ignore. On May 21, 2026, the Trump administration's Department of Commerce announced a $2 billion investment in nine quantum firms under the CHIPS and Science Act, taking equity stakes in each, according to Forbes. IBM alone received $1 billion to build a quantum chip foundry.
The core fear driving that spending is that China could develop a quantum computer capable of breaking modern encryption before the U.S. does. Trump's "Quantum Genesis Initiative" sets a 2028 deadline for a national quantum system — a goal that now runs on a collision course with unresolved questions about whether Microsoft's foundational science actually works, according to The Register.
Sentiment
Mostly NegativePublishers
16
Articles
82
Reach
98
