The Only Problem With Topological Quantum Computing

November, 2026

Eighty-nine years after Majorana’s theoretical work, the decisive experimental proof remains the question that marketing cannot answer.

A beautiful theory is not a working computer

In 1937, Ettore Majorana described a remarkable possibility: a fermionic particle that is its own antiparticle. Almost nine decades later, the idea remains one of the most fascinating in theoretical physics. It has also become the intellectual foundation for an ambitious commercial proposition: topological quantum computers whose information is protected by exotic quantum states rather than by ever-larger layers of conventional error correction. The vision is elegant. The difficulty is that an elegant mathematical possibility is not the same thing as a verified physical device.

There is an important distinction in the terminology. A fundamental Majorana fermion would be an elementary particle; the Majorana zero modes sought in semiconductor–superconductor devices are emergent quasiparticle excitations. Finding one would not be the same as discovering a new fundamental particle. But the distinction does not rescue an unproven hardware claim. A company advertising a topological qubit still has to demonstrate, convincingly and reproducibly, the non-Abelian topological physics that supposedly makes that qubit special.

The elusive signature

The appeal of Majorana zero modes is that quantum information could, in principle, be stored nonlocally, making it less vulnerable to certain local disturbances. In a simple nanowire picture, the desired modes appear at separated ends of a superconducting system. Researchers look for indirect signatures in tunnelling conductance, energy spectra, parity measurements and, ultimately, operations that establish the predicted non-Abelian behavior.

But the field has confronted an uncomfortable experimental problem for years: ordinary, non-topological Andreev bound states and disorder can produce signals resembling the signatures attributed to Majoranas. A zero-bias conductance peak, however intriguing, is not a certificate of a topological quantum computer. The critical issue is not whether a device produces an interesting signal. It is whether alternative explanations have been excluded and whether the defining topological properties have actually been measured.

The 2018 breakthrough that disappeared

The most instructive episode arrived in 2018, when a high-profile Nature paper reported quantized Majorana conductance in semiconductor–superconductor nanowires. It was precisely the kind of experimental milestone that could transform a long-standing theoretical aspiration into a credible engineering roadmap. In March 2021, however, the authors retracted the paper.

The retraction identified unreported corrections to charge jumps, a mislabeled axis and a recalibration that shifted the supposed quantized plateau. Replotting the broader data undermined the original claim. TU Delft subsequently reported that independent experts found selective presentation of data favoring the sought-after phenomenon, while finding no evidence that this had been intentional. This was not a minor dispute about publicity language. It was the withdrawal of a flagship experimental result in a field where the central phenomenon was already difficult to establish.

The lesson is stark: when a research program depends on extraordinarily subtle experimental signatures, confidence must be earned through complete data, independent scrutiny and reproducibility—not through the prestige of a journal, laboratory or corporate sponsor.

Majorana 1: the announcement and the evidence

In February 2025, Microsoft unveiled Majorana 1 and promoted a route toward topological quantum computing. The announcement generated extraordinary attention, but the accompanying Nature publication did not demonstrate the fully established topological qubit that the broader publicity invited readers to imagine.

The American Physical Society’s Physics reporting noted that the peer-review discussion explicitly concluded that the published results did not constitute evidence for Majorana zero modes in the reported devices. Nature and Science News subsequently documented substantial skepticism among physicists. This gap between a company’s public framing and the narrower claims substantiated in a peer-reviewed paper is exactly where scientific communication becomes commercial theater.

A chip name is not a measurement. A roadmap to a million qubits is not a million-qubit machine. And a proposed method for reading out a device is not proof that the underlying state is topological. Those distinctions are not pedantic; they determine whether the promised fault tolerance has any demonstrated foundation.

The 2026 challenge deepens the problem

The controversy did not end with the 2025 announcement. In a 2026 Nature Matters Arising article, Henry F. Legg examined transport data underlying Microsoft’s topological-gap detection approach. The critique argued that purported parity readout occurred in parameter regions displaying substantial disorder and apparently gapless behavior, raising the possibility that non-topological mechanisms explained the observed signals.

Separately, reporting on Microsoft’s newer device results in 2026 described continued disagreement over whether the measurements established a working topological qubit. The central question remained the same: have researchers demonstrated the distinctive quantum operations and protection that justify the term topological, or have they demonstrated sophisticated but ambiguous device behavior?

This is why successive press releases cannot substitute for decisive experiments. Each announcement may describe genuine laboratory work. Yet genuine laboratory work and proof of the advertised computational architecture are very different things.

The economics of a promise

Topological quantum computing is marketed as a shortcut around one of quantum engineering’s hardest problems: error correction. If protected qubits could be built and controlled reliably, the architecture might reduce some of the formidable overheads associated with fault-tolerant computation. That conditional promise is the entire commercial attraction.

But conditional is the operative word. A commercial roadmap that assumes the key physical ingredient before conclusively establishing it is not an engineering forecast in the ordinary sense. It is a chain of hypotheses: that the relevant modes are present, that their topological character is established, that their operations can be controlled, that error rates are advantageous, and that all of this can scale manufacturably. Failure at the first links renders the later projections speculative.

The danger is that investors, customers and policymakers may hear the final projected qubit count while overlooking the unresolved experimental premise. In that environment, the language of a revolutionary platform can get years ahead of the physics.

What would count as proof?

The standard should be demanding but straightforward: unambiguous evidence of the relevant topological phase; convincing exclusion of trivial look-alikes; reproducible demonstrations across devices and laboratories; and computational operations that establish the expected protection, including the characteristic non-Abelian behavior. Claims about useful quantum computing should then be supported by measured logical performance, not extrapolation alone.

These are not arbitrary obstacles invented by skeptics. They are the measurements required to connect a theoretical mechanism to an actual technology. Until that connection is secure, a proposed topological qubit remains a research objective rather than a proven commercial advantage.

The only problem

Topological quantum computing has a magnificent theoretical story, a compelling engineering dream and some of the most powerful marketing machinery in technology. Its only problem is the one that matters most: proving that the physical resource on which the dream depends has been realized and controlled in the way the architecture requires.

Nearly ninety years after Majorana’s proposal, the industry should be careful not to confuse the persistence of an idea with the verification of a machine. Microsoft and others are entitled to pursue bold research. They are not entitled to have publicity treated as experimental proof. Science advances when a claim survives scrutiny; marketing advances when a claim attracts attention. Topological quantum computing must ultimately be judged by the former.

References and further reading

Majorana, E. (1937), “Teoria simmetrica dell’elettrone e del positrone.”
https://doi.org/10.1007/BF02961314

Zhang et al., Retraction Note: Quantized Majorana conductance, Nature (2021).
https://www.nature.com/articles/s41586-021-03373-x

TU Delft, Retraction Nature article “Quantized Majorana conductance” (2021).
https://www.tudelft.nl/en/2021/tu-delft/retraction-nature-article-quantized-majorana-conductance

APS Physics, Experts Weigh in on Microsoft’s Topological Qubit Claim (2025).
https://physics.aps.org/articles/v18/57/

Nature, Microsoft quantum-computing claim still lacks evidence (2025).
https://www.nature.com/articles/d41586-025-00829-2

Legg, On the robustness of topological gap detection via transport, Nature (2026).
https://www.nature.com/articles/s41586-026-10567-8

Science News, Microsoft’s quantum chip got an upgrade. Critics are still skeptical (2026).
https://www.sciencenews.org/article/microsoft-quantum-chip-upgrade-majorana