Microsoft unveils revised Majorana 2 quantum chip – skepticism remains in the scientific community

Principle of Operation: Tetrons and Parity Storage

According to Microsoft, the AI platform Microsoft Discovery played a central role in the development. Agent-based AI controlled workflows, automated measurements, optimized manufacturing processes, and detected previously unnoticed errors. This is particularly relevant because generating a topological state requires the coordination of hundreds of parameters – a process that would take weeks manually.

Microsoft also states that it has halved the original roadmap and is now aiming for the year 2029 for a scalable, commercially usable quantum computer. Competitors such as Google with the Willow chip and IBM with the Nighthawk chip are also working towards this goal, relying on a growing number of superconducting qubits with improved error correction. IBM is announcing investments of around ten billion dollars for this purpose in the coming years.

Lingering Doubts in the Scientific Community

At the same time, the scientific dispute over Microsoft's topological approach remains open. The central question is whether the measured signals unequivocally indicate Majorana zero modes or can also be explained by more conventional effects. As early as February 2025, Microsoft had caused a stir with Majorana 1; the predecessor chip caused considerable doubt among many physicists. The existence of Majorana particles is not considered unequivocally proven within the scientific community.

An earlier Nature article by a Microsoft team from 2018, which claimed the first observation of a Majorana state, was retracted in 2021 because the original data analysis did not meet scientific quality standards. Against this background, scientists are calling for Microsoft to make its results publicly accessible so that they can be independently verified. Microsoft justifies its reluctance with the protection of trade secrets.

External Validation by DARPA and Competition in Europe

However, Microsoft has confidentially shared data with the US Defense Advanced Research Projects Agency (DARPA), which is currently evaluating various technological approaches for building quantum computers. DARPA has included Microsoft as one of only two companies in the final phase of its Quantum Benchmarking Program. In this phase, Microsoft is to develop a fault-tolerant prototype based on topological qubits. DARPA's evaluation provides an external framework but does not replace independent scientific verification of Microsoft's central claims.

Henry Legg, a lecturer in quantum physics at the University of St. Andrews in Scotland, sharply articulated the criticism. "Microsoft can use as much lead as it wants. That doesn't protect the company from the fundamental scientific principle that results must be reproducible," he said. Zander, on the other hand, expressed his conviction: "We have delved deeply enough into the physics. I wouldn't spend the money if I felt we were wrong."

Quantum computers are also being developed in Europe. Finnish company IQM and German startups EleQtron, SaxonQ, and PlanQC are considered pioneers. Munich-based semiconductor manufacturer Infineon is also involved in the field. Google, Amazon, and some Chinese companies, alongside Microsoft, IBM, and European firms, are among the players aiming to pave the way to a commercially usable quantum computer. The coming years will show whether Microsoft's topological approach can meet the expectations placed upon it.

Details about Majorana 2 are published in the paper "20 Second Parity Lifetime in an InAs–Pb Tetron Device." In it, the team describes the switch from aluminum to lead as a superconductor and the update of the active semiconductor region to a combination of indium arsenide and indium arsenide antimonide. According to Microsoft, the topological energy gap in Majorana 2 is more than twice as large as in the predecessor model. However, lead is water-soluble and is washed out in conventional manufacturing processes for quantum chips, making the material change a significant intervention in terms of process engineering.