A research team from Australia is raising more questions about the kind of topological quantum computing being pursued by Microsoft. In a paper on the pre-print archive arXiv (Decoherence of Majorana Qubits by 1/f Noise), the researchers argue the noise mechanism they have identified imposes fundamental limitations on the use Majorana quasiparticles as qubits in semiconductor nanowire, which is the architecture being pursued by Microsoft.

“We note that the decoherence time is substantially shorter than the time needed to perform a qubit measurement, which is currently 32.5 µs as well as the target measurement time of 1 µs on the Microsoft roadmap,” write the researchers.

Microsoft vigorously disagrees. (See response comments below).

The controversy is around so-called topological qubits, based on Majorana quasiparticles, which are predicted to be extremely resistant to errors. Microsoft caused a stir last February with introduction of its Majorana 1 quantum chip, based on leveraging Majoranas, and also issued a roadmap to fault-tolerant quantum computing. (See HPCwire coverage, Microsoft’s Big Bet on Majorana Pays Off with New Topological Quantum Chip). Targeting topological qubits was and is a risky yet bold move by Microsoft.

The Microsoft announcement at the time was greeted with both enthusiasm and skepticism.

Travis Humble, director of the Quantum Science Center (QSC), ORNL, said, “The Microsoft results are important progress in the field of quantum computing. They demonstrate results for the measurement of parity in a topological superconducting nanowire design. This is a first step toward validating topological protection, and more work should seek to demonstrate the expected non-Abelian statistics.” QSC is one of five national quantum research centers and a big part of its mission is investigation materials for topological quantum computing.

This latest work is by A. Alase from the University of Sydney, M. C. Goffage, M. C. Cassidy, and S. N. Coppersmith, all from the University of New South Wales. Their work focuses on trying to use Majorana quasiparticle in semiconductor-based superconducting material (nanowires) and suggests some fundamental limitation in trying to implement Majorana-based qubits.

The abstract to their paper provides a good summary of the work:

“Qubits based on Majorana Zero Modes (MZMs) in superconductor-semiconductor nanowires have attracted intense interest as a platform for utility-scale quantum computing. These qubits have been predicted to show extremely low error rates due to topological protection of the MZMs, where decoherence processes have been thought to be exponentially suppressed by either the nanowire length or the temperature. However, here we show that 1/f noise, which is ubiquitous in semiconductors, gives rise to a previously unexplored mechanism for Majorana qubit decoherence.

“The high frequency components of this noise causes quasiparticle excitations in the bulk of the topological superconductor, which in turn result in qubit errors that increase with the length of the nanowire. We calculate the probability of quasiparticle excitation for disorder-free nanowires in the presence of 1/f noise and show that this mechanism limits the decoherence times of the MZM qubits currently being developed to less than a microsecond even for perfectly uniform nanowires with no disorder. This decoherence time is significantly shorter than the time to implement quantum gates using this technology and is also shorter than the decoherence times of qubits in other leading solid-state architectures.”

Microsoft is Forging Ahead

Not surprisingly, Microsoft has a very different view. Chetan Nayak, Technical Fellow, Microsoft Quantum, told HPCwire by email, “The preprint paper is about quasiparticle poisoning, induced by high-frequency noise, in topological qubits. This has been well-studied for a decade in theory papers, and we have both direct and indirect experimental measurements in our devices, showing that the associated error rates are small, as noted in our Nature paper and recent analysis on ArXiv.

“Microsoft is committed to advancing topological quantum computing as a promising pathway to achieving fault-tolerant quantum systems at scale. By leveraging Majorana zero modes, Microsoft’s topological qubit design intrinsically protects quantum information from local sources of decoherence. This approach aims to dramatically improve stability and scalability, making it a compelling foundation for building a practical quantum computer.”

Shown below is the Microsoft roadmap to achieving fault-tolerant quantum computing which it released at the same time it announced the Majorana 1 chip.

It’s perhaps too early for a deep critique of the Australian researchers’ paper, but it is likely to further stir the heated conversation around Microsoft’s approach and its Majorana chip.

Analysts are taking a measured approach:

  • Mark Horvath, VP analyst for Gartner notes, “If the result of the paper holds up, it would certainly make the already challenging path to topological qubits significantly harder. Currently, many investigators are working to independently verify Microsoft’s February 2025 results, and it takes time. Similarly, these results would also need to be verified, and if true would be a huge setback. Science is the art of being wrong in the right way, and it takes time to do it properly. Topological qubits are a very tempting target for quantum computing and would solve a huge number of problems, but we need to let the science work itself out one way or another.”
  • Bob Sorensen, SVP research, Hyperion Research said, “At this stage Microsoft quantum research is still in the experimental, exploratory stage and to date I have not seen any definitive claims from them about development of a production ready Majorana-based quantum system. If/when that happens, that will be the time to delve into the overall capabilities of the system, but at this early stage I think that just about any qubit modality under development could face similar scrutiny from a not insignificant number of proponents and opponents. That said, if this paper materially contributes to advancements in Majorana research that then I am all for it; it not, then I hope that it does not contribute to the relative confusion within the QC end user base as to the overall promise of QC systems to address real-world use cases.”

It will be interesting to see how the broad quantum community reacts. Among other things, the Australian researchers suggest that “significant materials breakthroughs would be needed to enable the large increases in the superconducting gap and/or the large decreases in the level of 1/f noise that would enable error rates that are low enough to enable efficient scaleup of tetron qubits to be achieved.”

They further suggest other architectures might better support the topological approach to quantum computing: “In more spatial dimensions the quasiparticles can remain far from the topological defects being employed as qubits for much longer times and so topological protection is more likely to improve qubit coherence in architectures that utilize topologically nontrivial condensates with more spatial dimensions.”

As always it’s best to read the paper directly.

Stay tuned.

Link to paper (Decoherence of Majorana Qubits by 1/f Noise), https://arxiv.org/abs/2506.22394