Microsoft called it a breakthrough. A landmark replication study calls it a dead end. Both cited the same data.
image from grok
Microsoft's Majorana 1, announced as the world's first topological qubit processor in February 2025 with a Nature paper and DARPA backing, has failed independent replication. A two-year effort led by Sergey Frolov at the University of Pittsburgh, published in Science, found that in every device tested, additional data resolved the promising Majorana-like signals into mundane physical explanations. The case illustrates a recurring pattern in the Majorana field: limited datasets produce signatures consistent with exotic topological physics, but fuller datasets consistently reveal trivial origins—a 2018 Nature paper on quantized Majorana conductance was similarly retracted in 2021 after findings could not be sustained.
When Microsoft unveiled Majorana 1 in February 2025, calling it the world's first topological qubit processor, the company backed the announcement with a peer-reviewed Nature paper and DARPA selection for a high-profile quantum benchmarking program. A year later, the most comprehensive independent replication effort yet conducted has concluded: the underlying physics does not hold.
A team led by Sergey Frolov, a professor of physics at the University of Pittsburgh, spent two years attempting to reproduce the results behind Microsoft's topological qubit claims. Their findings, published January 8, 2026, in Science (DOI: 10.1126/science.adk9181), are unsparing. In every case across multiple devices, additional data revealed mundane explanations for signals that had originally looked consistent with Majorana zero modes — the exotic quasiparticles Microsoft said its chips produced. As ScienceDaily reported, the paper spent a record two years under peer review before acceptance. The limited dataset looked like topological physics. The fuller dataset revealed trivial origins.
Microsoft announced Majorana 1 — described as the world's first quantum processor powered by a topological core — on the Azure Quantum blog, saying the topological core relied on a "new state of matter" that previously existed only in theory. The company said topological qubits could enable fault-tolerant computing with built-in error protection. DARPA subsequently selected Microsoft to advance to the final phase of its Underexplored Systems for Utility-Scale Quantum Computing (US2QC) benchmarking program.
The promise of topological qubits rests on a specific kind of physical protection. Information encoded in Majorana zero modes is theoretically protected from environmental noise — a property that, if real, would make topological qubits substantially more robust than the superconducting or trapped-ion alternatives currently on the market. The problem is that the signature signals associated with Majorana modes can be produced by other, more ordinary physics. Distinguishing genuine topological protection from a mundane false positive requires more data, not just more careful analysis.
The 2018 paper claiming quantized Majorana conductance in similar devices was retracted by Nature in 2021 after an investigation found the results could not be sustained. The authors apologized for insufficient scientific rigor. The 2025 Nature paper on Microsoft's InAs-Al hybrid devices, with more than 160 researchers listed, faced its own reckoning. Nature's editorial team sought additional review from two of the paper's original reviewers and concluded that the results do not represent evidence for the presence of Majorana zero modes.
The protocol Microsoft used to identify topological phenomena has a documented vulnerability. A separate analysis posted to arXiv found that the topological gap protocol "produces results that depend on measurement choices and input parameters rather than on underlying device properties, and is likely to deliver false positives." The protocol, in other words, can be fooled by the wrong input choices even when nothing topological is actually happening.
At the APS March 2025 meeting, Microsoft principal fellow Chetan Nayak presented data intended to silence critics. Eun-Ah Kim, a physicist at Cornell University who reviewed the plots, was not convinced. "I would have loved this to just come out screaming at me that there's only two," Kim told Science News, referring to the two Majorana modes the chip was supposed to harbor. "But I don't think that's what I see." She described the data plot as resembling random jitter.
Frolov's group replicated the original results across multiple device types and experimental configurations, working with co-authors from the University of Minnesota and Grenoble. Their conclusion in the Science paper: the limited dataset was consistent with four topological phenomena, but additional data identified the most likely origins of those patterns as trivial. The two-year peer review period — described by ScienceDaily as a record — reflects the depth of the scrutiny applied to what the journal recognized as consequential claims.
The authors of the Science paper are largely the same researchers who published the original topological qubit claims in top journals. That the replication now appears in Science rather than Nature or Physics Review Letters is itself notable: while the original papers claiming advances for quantum computing made their way into high-impact journals, the individual follow-up studies could not make it past editors at those same publications.
Frolov was direct in his assessment. "That chip cannot possibly work given what we saw today," he told Science News. Henry Legg, a physicist at the University of Basel, was blunter. "Any company claiming to have a topological qubit in 2025 is essentially selling a fairytale — and I think it's a dangerous fairytale," he told Science News.
Microsoft had not issued a formal public response at time of publication. The DARPA US2QC benchmarking program continues; its evaluation methodology may now incorporate the Science findings as a reference point for what constitutes genuine topological evidence versus a false positive.
The field of topological quantum computing now faces a reckoning distinct from the usual rhythm of incremental hardware advances. If the false positive rate for topological signatures is as high as the Science paper suggests, the entire literature built on early-stage Majorana measurements needs re-examination. Several groups beyond Microsoft have pursued topological approaches; the replicability problem, if it is systemic, does not belong to one company alone.
The durable lesson here is not that Microsoft failed. It is that the protocol used to identify topological milestones is prone to false positives by design — a methodological flaw documented in peer-reviewed work, not merely argued in critique. Any future claim of topological qubit evidence should be evaluated against this finding, not against the press release that accompanied it.
Story entered the newsroom
Research completed — 0 sources registered. Replication study (Frolov et al., Science 2026) examined original Majorana zero mode claims from Microsoft and others. Found limited data consistent w
Draft (922 words)
Approved for publication
Headline selected: Limited Data Looked Revolutionary. Full Data Revealed Trivial Physics.
Published (911 words)
Get the best frontier systems analysis delivered weekly. No spam, no fluff.
Quantum Computing · 17h 59m ago · 2 min read
Quantum Computing · 18h 56m ago · 4 min read
