Nature has published a peer-reviewed challenge to the evidence behind Microsoft's Majorana 1 quantum claims, putting more than a year of scientific skepticism into the formal record.
The critique, written by Henry F. Legg of the University of St. Andrews and published June 24 in Nature's "Matters Arising" section, targets the 2025 paper Microsoft used to support its Majorana 1 announcement. Microsoft disputes the critique in a reply published in the same issue, arguing that its measurements still support a topological interpretation.
The dispute lands at a sensitive moment. Microsoft announced Majorana 2 on June 2, saying the newer chip is 1,000 times more reliable than Majorana 1 and keeps the company on track for a scalable quantum computer by 2029. Legg argues that the newer announcement does not resolve the central question: whether Microsoft has shown convincing evidence that its devices contain the Majorana zero modes needed for topological qubits.
Why Microsoft's quantum claim drew scrutiny
Microsoft unveiled Majorana 1 in February 2025, describing it as a quantum processing unit powered by a Topological Core and saying it was designed to scale to a million qubits on a single chip. In the company's Majorana 1 announcement, Microsoft also said it was on track to build a fault-tolerant prototype "in years, not decades."
That public framing went further than what the underlying paper could settle. Microsoft's 2025 Nature paper reported single-shot parity measurement in InAs–Al hybrid devices, but it also said the measurements did not, by themselves, determine whether the low-energy states detected by interferometry were topological. That distinction matters because topological qubits depend on Majorana zero modes, and ordinary non-topological states can mimic some of the expected signals.
Microsoft's design uses an indium arsenide semiconductor joined to a superconductor. The goal is to create Majorana zero modes, quantum states that could make qubits more stable and easier to scale than many competing approaches. Whether Microsoft has shown those modes clearly enough is the central dispute.
What Legg says the data shows
Legg's critique makes two main arguments: one about the physics and one about the analysis workflow.
On the physics, Legg argues that Microsoft's transport data does not show the robust superconducting gap needed to support the company's interpretation. Instead, he says the relevant regions of the device appear disordered or gapless, leaving open the possibility that the observed signals came from non-topological mechanisms such as quantum dots or trivial Andreev states.
On the analysis workflow, Legg says software issues affected what reviewers could see. One issue allegedly limited a plot to the largest region passing Microsoft's Topological Gap Protocol. Another allegedly processed bias-voltage data by array index rather than physical value. If correct, those issues could mean reviewers did not see the full range of data that passed the same test.
Microsoft rejects that interpretation. In its Nature reply, the company argues that Legg focuses too narrowly on transport measurements and does not account for the quantum-capacitance measurements Microsoft considers central to its case. Microsoft says those measurements strongly indicate a topological origin and sharply constrain non-topological explanations.
The disagreement is narrow but consequential. The question is not whether topological qubits are worth pursuing. It is whether Microsoft has shown enough evidence for the specific physics behind its Majorana 1 claim.
Majorana 2 raises the stakes
Microsoft announced Majorana 2 on June 2, saying the chip uses a new materials stack that delivers a 1,000-fold reliability improvement over the prior generation. The company said the chip has a mean qubit lifetime of 20 seconds, with some instances lasting as long as one minute, and that it now expects to achieve a scalable quantum computer by 2029.
Legg argues the newer announcement does not settle the dispute because stable parity is not the same thing as demonstrated quantum superposition. A classical bit can hold a stable state for years; that alone does not make it a qubit. Microsoft disputes that framing and says parity lifetimes translate into qubit lifetimes in its architecture.
The Majorana 2 results were released as a June 2 preprint, not a peer-reviewed paper. That makes the 2029 target dependent not only on Microsoft's engineering pace, but also on whether outside researchers are persuaded by the evidence behind the company's topological-qubit claims.
The tests that could settle the dispute
Nature has not retracted Microsoft's 2025 paper, and the new critique does not end the debate. It does, however, give researchers a clearer record to evaluate: Legg's formal challenge, Microsoft's formal reply and the original paper behind the Majorana 1 announcement.
Three developments would move the discussion forward. Independent peer review of the Majorana 2 results would test Microsoft's newer claims under the same scrutiny now applied to Majorana 1. A convincing superposition measurement would address a key gap raised by critics. Broader public data access would allow outside teams to test whether Microsoft's interpretation holds up beyond the company's own analysis.
Until then, the central question remains open: whether Microsoft's published evidence supports the confidence of its public quantum roadmap.
Comments
Be the first, drop a comment!