FAQ on Microsoft’s topological qubit thing
Q1. Did you see Microsoft’s announcement?
A. Yes, thanks, you can stop emailing to ask! Microsoft’s Chetan Nayak was even kind enough to give me a personal briefing a few weeks ago. Yesterday I did a brief interview on this for the BBC’s World Business Report, and I also commented for MIT Technology Review.
Q2. What is a topological qubit?
A. It’s a special kind of qubit built using nonabelian anyons, which are excitations that can exist in a two-dimensional medium, behaving neither as fermions nor as bosons. The idea grew out of seminal work by Alexei Kitaev, Michael Freedman, and others starting in the late 1990s. Topological qubits have proved harder to create and control than ordinary qubits.
Q3. Then why do people care about topological qubits?
A. The dream is that they could eventually be more resilient to decoherence than regular qubits, since an error, in order to matter, needs to change the topology of how the nonabelian anyons are braided around each other. So you’d have some robustness built in to the physics of your system, rather than having to engineer it laboriously at the software level (via quantum fault-tolerance).
Q4. Did Microsoft create the first topological qubit?
A. Well, they say they did! [Update: Commenters point out to me that buried in Nature‘s review materials is the following striking passage: “The editorial team wishes to point out that the results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices. The work is published for introducing a device architecture that might enable fusion experiments using future Majorana zero modes.” So, the situation is that Microsoft is unambiguously claiming to have created a topological qubit, and they just published a relevant paper in Nature, but their claim to have created a topological qubit has not yet been accepted by peer review.]
Q5. Didn’t Microsoft claim the experimental creation of Majorana zero modes—a building block of topological qubits—back in 2018, and didn’t they then need to retract their claim?
A. Yep. Certainly that history is making some experts cautious about the new claim. When I asked Chetan Nayak how confident I should be, his response was basically “look, we now have a topological qubit that’s behaving fully as a qubit; how much more do people want?”
Q6. Is this a big deal?
A. If the claim stands, I’d say it would be a scientific milestone for the field of topological quantum computing and physics beyond. The number of topological qubits manipulated in a single experiment would then have finally increased from 0 to 1, and depending on how you define things, arguably a “new state of matter” would even have been created, one that doesn’t appear in nature (but only in Nature).
Q7. Is this useful?
A. Not yet! If anyone claims that a single qubit, or even 30 qubits, are already useful for speeding up computation, you can ignore anything else that person says. (Certainly Microsoft makes no such claim.) On the question of what we believe quantum computers will or won’t eventually be useful for, see like half the archives of this blog over the past twenty years.
Q8. Does this announcement vindicate topological qubits as the way forward for quantum computing?
A. Think of it this way. If Microsoft’s claim stands, then topological qubits have finally reached some sort of parity with where more traditional qubits were 20-30 years ago. I.e., the non-topological approaches like superconducting, trapped-ion, and neutral-atom have an absolutely massive head start: there, Google, IBM, Quantinuum, QuEra, and other companies now routinely do experiments with dozens or even hundreds of entangled qubits, and thousands of two-qubit gates. Topological qubits can win if, and only if, they turn out to be so much more reliable that they leapfrog the earlier approaches—sort of like the transistor did to the vacuum tube and electromechanical relay. Whether that will happen is still an open question, to put it extremely mildly.
Q9. Are there other major commercial efforts to build topological qubits?
A. No, it’s pretty much just Microsoft [update: apparently Nokia Bell Labs also has a smaller, quieter effort, and Delft University in the Netherlands also continues work in the area, having ended an earlier collaboration with Microsoft]. Purely as a scientist who likes to see things tried, I’m grateful that at least one player stuck with the topological approach even when it ended up being a long, painful slog.
Q10. Is Microsoft now on track to scale to a million topological qubits in the next few years?
A. In the world of corporate PR and pop-science headlines, sure, why not? As Bender from Futurama says, “I can guarantee anything you want!” In the world of reality, a “few years” certainly feels overly aggressive to me, but good luck to Microsoft and good luck to its competitors! I foresee exciting times ahead, provided we still have a functioning civilization in which to enjoy them.
Update (Feb 20): Chetan Nayak himself comments here, to respond to criticisms about Microsoft’s Nature paper lacking direct evidence for majorana zero modes or topological qubits. He says that the paper, though published this week, was submitted a year ago, before the evidence existed. Of course we all look forward to the followup paper.
Follow
Comment #1 February 20th, 2025 at 4:43 am
Scott – great post. Thanks. Just a small reminder – we published our work on non abelian
Anyons in 2023. Exact same outcome in the sense of a topological qubit. Think of our gate fidelities as MSFT energy gap 🙂 more to come on this from Quantinuum this year. This work sits alongside our more established QCCD architecture and our QEC work
Comment #2 February 20th, 2025 at 5:00 am
> one that doesn’t appear in nature
that we know of.
Comment #3 February 20th, 2025 at 5:48 am
Informative as always, many thanks!
Regarding Q9, also Nokia Bell Labs is developing topological qubits, although from a somewhat different angle. https://www.nokia.com/bell-labs/research/air-lab/data-and-devices/topological-quantum-computing/
Comment #4 February 20th, 2025 at 5:50 am
Scott,
Are there any in principle advantages for Topological qubits vs other approaches that give them a fundamental edge ? Yes, engineering them would be different story, however if we are sure they are better in a way that transistors were for vacuum tubes, then it would involve only redirecting engineering resources towards that approach ?
Comment #5 February 20th, 2025 at 5:54 am
Q11: How much resilience can they show with that single topological qubit?
Comment #6 February 20th, 2025 at 6:25 am
Seems to me that in order to claim that a qubit exists, you should have coherence times and some gates with fidelities. All I see in the paper is observed switching rates between two states under measurement.
Comment #7 February 20th, 2025 at 8:44 am
Hello, first of all, I am not an expert and still learning QC. So please read me with the appropriate caution.
I found someone online raising an interesting question: https://www.reddit.com/r/QuantumComputing/comments/1ite220/majorana_1_did_anyone_read_the_fine_print/
He quotes the editorial team:
“The editorial team sought additional input from Reviewers #2 and #3 after the second round of review to establish this manuscript’s technical correctness. Their responses proved satisfactory enough to proceed to publication. The editorial team wishes to point out that the results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices. The work is published for introducing a device architecture that might enable fusion experiments using future Majorana zero modes.”
It seems to me from reading the reports that the referees do not agree on the demonstration of a qubit’s presence.
From ref#1 (on first review):
https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-024-08445-2/MediaObjects/41586_2024_8445_MOESM2_ESM.pdf
“Here the authors seem to imply that in this manuscript a qubit (and its parity readout) is demonstrated, which is not true.”
And from the Nature publication (I add the link just to save you some time):
https://www.nature.com/articles/s41586-024-08445-2
“1DTSs²,³,⁴ are a promising platform for building topological qubits.”
All these elements confuse me:
On the one hand, I found online that their new chip had 8 qubits:
https://www.techtarget.com/searchdatacenter/news/366619479/Microsoft-unveils-quantum-chip-Majorana-1-for-future-advances#:~:text=Today%2C%20Majorana%201%20has%20only,said%20in%20a%20blog%20post.
On the other hand, some referees from the peer review do not seem to acknowledge the presence of a qubit nor the proof of the existence of Majoranas.
The majority of the referees still seem to recognize this work as being worthy of publication.
My main confusion is that they seem to claim to have an 8-qubit chip while still discussing if even one has been made.
Disclaimers:
I mostly do not understand the physics involved.
I have been a silent reader for the past few months and only have ~1 year of QC background.
I have a simplified view of what is happening, and I am only asking for clarifications on things that appear to me as contradictory statements.
I am not saying anything posted is wrong or sensationalist, I am just confused.
Feel free to get rid of the disclaimer parts to keep the conversation structured on the blog
Comment #8 February 20th, 2025 at 8:54 am
Philip Reinhold #6
> Seems to me that in order to claim that a qubit exists, you should have coherence times and some gates with fidelities
… and in order to claim that the qubit is based on topological Majorana modes, the coherence time should increase exponentially when temperature is decreased and when the length of the wire is increased.
Interestingly, the paper itself makes a much more modest claim: “These measurements do not, by themselves, determine whether the low-energy states detected by interferometry are topological”, and even Nature’s editorial team isn’t fully aboard the hype train this time: “The editorial team wishes to point out that the results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices” (though this sentence is hidden away in the peer review file).
Comment #9 February 20th, 2025 at 10:00 am
If the claim holds up it sure sounds like a breakthrough to me. This sort of resilience was always the unachievable holy grail for QC so it seems a lot more likely to me that we get useful QC in a decade than it did a few days ago.
Comment #10 February 20th, 2025 at 10:15 am
Hi Scott,
I was surprised to hear that they were claiming they had a topological qubit… I thought this was some misinterpretation in the conversion from scientific result to science writing. But here the scientists are indeed claiming “yes, we have a topological qubit.” That would be amazing, but it seems like not what’s claimed in Nature? What do you make of this disclaimer in the peer review file (available here https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-024-08445-2/MediaObjects/41586_2024_8445_MOESM2_ESM.pdf):
“The editorial team wishes to point out that the results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices. The work is published for introducing a device architecture that might enable fusion experiments using future Majorana zero modes.”
So I guess my question is, if the qubit thing is not in the Nature paper, is it a claim we should be waiting on more results for? Seems like there’s a disconnect somewhere.
Comment #11 February 20th, 2025 at 10:30 am
Hi Scott, I am a bit surprised that your post seems to have missed an opportunity to clarify what seems to be the most crucial point: Microsoft’s claim to have created a topological qubit using two nanowires does not appear in their Nature paper, only in their press release. As far as I can tell (and despite how the news was reported by the New York Times, the Financial Times and probably countless other outlets), there are no qubits in the paper.
Comment #12 February 20th, 2025 at 10:34 am
So did they do it or did they not? Did they find the Majoran particle or did they not?
Comment #13 February 20th, 2025 at 10:34 am
@Mikael Johansson – to be run on a 5110 I hope 🙂
Comment #14 February 20th, 2025 at 10:52 am
Why are topological qubits so difficult to create?
Why does Microsoft continue to bet on topological qubits if other quantum computing technologies already exist?
How likely is it that topological qubits will become the dominant technology in quantum computing?
What does ‘creating a new state of matter’ with topological qubits imply?
What impact would the creation of a single topological qubit have on quantum computing?
Comment #15 February 20th, 2025 at 11:28 am
Hi Scott, did Chetan mention when/where they will report the data on their “topological qubit that’s behaving fully as a qubit”? It’s not in the Nature paper….
Comment #16 February 20th, 2025 at 12:12 pm
Ilyas #1: Thanks. I do want to maintain the distinction between creating nonabelian anyons “in software”—what Quantinuum did in 2023, by simulating the anyons in a system of trapped-ion qubits—and creating them “in hardware,” which is what Microsoft claims to have done.
Comment #17 February 20th, 2025 at 12:15 pm
Martin #7: Thanks!! I’ve added a note about that to the post.
It’s clear that Microsoft is claiming to have created both majorana zero modes and topological qubits. It’s also clear that, if they once again haven’t, they’ll have egg all over their faces; I don’t know if their topological QC effort could recover from two high-profile retracted claims. Beyond that, I certainly don’t know any more than the device physicists do, and people should feel free to debate it here!
Comment #18 February 20th, 2025 at 12:15 pm
Microsoft is not the only commercial player working on Majorana fermions. Nokia is also working in that field, although quietly. In Murray Hill, New Jersey.
Comment #19 February 20th, 2025 at 1:53 pm
So they created something and packaged it in a chip in 20 years that professional physicists working on particle colliders only identified experimental evidence of in recent years (https://en.wikipedia.org/wiki/Anyon#Experiment)?
Is this possible?
Comment #20 February 20th, 2025 at 3:04 pm
How good are topological qubits at performing non-Clifford gate operations?
Their device roadmap paper seems to focus almost entirely on implementing Clifford gates, but it says in passing that the T gate “is not a topologically-protected operation in Majorana-based qubits. However, a noisy implementation of the T gate together with low-noise Clifford operations can be used to distill low noise resource states from which fault-tolerant T gates can be implemented … Going beyond Clifford operations to obtain a computationally universal gate set requires additional, non-topologically-protected operations, [which] are not the focus of this paper.”
Are they sweeping a potentially big issue under the rug here? Or are they correct in their implication that protecting non-Clifford gate operations isn’t that big of a deal? Does the additional overhead required for protecting non-Clifford gates risk negating most of the advantage of the topological protection of the Clifford gates?
Comment #21 February 20th, 2025 at 4:21 pm
Where is the fervor that we saw when other smaller companies made outrages claims, such as having realized quantum annealing on a chip?
But I get it. There’s bigger chip to fry when the entire federal government goes up in flames.
Comment #22 February 20th, 2025 at 4:44 pm
This thread by Sergey Frolov makes for dire reading. He pretty much calls this academic fraud.
https://bsky.app/profile/spinespresso.bsky.social/post/3limkxk7kmk2f
Comment #23 February 20th, 2025 at 4:48 pm
Anyonimous #19 Particle accelerators deal with elementary particles. These topological qubits build on quasiparticles. Excitations in solid-sate systems that behave like particles.
For instance sound excitations can create phonons in condensed matter.
https://en.wikipedia.org/wiki/Quasiparticle
Comment #24 February 20th, 2025 at 5:33 pm
Henning #21: The crucial difference is, D-Wave made claims that I knew were bullshit, because they contradicted almost everything I understood about quantum algorithms.
By contrast, I know of no deep reason why Microsoft couldn’t have made a topological qubit. It’s all an engineering question.
In this post, I carefully declined to take a position on whether Microsoft’s latest claim will stand or fall. Certainly, if it stands that’s great, while if it falls, many people will have had it with Station Q, and won’t take further claims from them about topological qubits seriously.
Comment #25 February 20th, 2025 at 6:00 pm
Prasanna #4:
Are there any in principle advantages for Topological qubits vs other approaches that give them a fundamental edge ?
That depends entirely on what you mean by words like “in principle” and “fundamental.” From a high enough theoretical remove, everything that can be done using topological qubits, can be done just as well with fault-tolerant non-topological qubits: they give rise to the same complexity class BQP. On the other hand, the built-in robustness provided by nonabelian anyons is a matter of great interest to theoretical physicists.
Comment #26 February 20th, 2025 at 6:18 pm
Scott #24:
In a sense it’s great to see mastery of quantum information become gradually insufficient for sanity checking press releases, as the field drifts from theory towards practice.
Comment #27 February 20th, 2025 at 6:42 pm
Philip Reinhold #6:
Seems to me that in order to claim that a qubit exists, you should have coherence times and some gates with fidelities. All I see in the paper is observed switching rates between two states under measurement.
I’m sure that Chetan Nayak gave me some coherence numbers when we talked, but I don’t remember them, and wouldn’t want to share without permission even if I did. Maybe we can get him or someone else from Microsoft Station Q to comment here.
Comment #28 February 20th, 2025 at 6:45 pm
Scott #24, fair enough.
Just to be clear, I really hope that Frolov is wrong. There are lot of great people at Station Q and it would be awful to see them all discredited if he turns out to be right.
Comment #29 February 20th, 2025 at 6:50 pm
Davide Castelvecchi #11:
I am a bit surprised that your post seems to have missed an opportunity to clarify what seems to be the most crucial point: Microsoft’s claim to have created a topological qubit using two nanowires does not appear in their Nature paper, only in their press release.
Often a post like this functions as a “stone soup,” where I blog whatever I know, then I add crucial additional facts as the commenters bring them to my attention.
In this case, my main source for what Microsoft was claiming was neither the press release nor the Nature paper, but simply what Chetan Nayak told me personally. Nowhere did I assume that Microsoft’s claims coincide with what the research community will ultimately accept as true.
At present, though, I think it would be fair to characterize Microsoft’s topological qubit claims as neither accepted nor rejected by Nature’s review process, but simply not ruled on by that process at all.
Comment #30 February 20th, 2025 at 6:58 pm
Michael Vassar #9 (and others): It’s important to understand that, while topological qubits could reduce the need for active error correction, they almost certainly wouldn’t eliminate it entirely. Indeed in the near-term, the coherence properties could even be worse than with traditional qubits. So even if the creation of topological qubits is confirmed, it will still remain an open question whether this approach will win the race to practical QC.
Comment #31 February 20th, 2025 at 7:58 pm
Scott #27:
Readers of our Nature paper may have noticed that the paper was submitted on March 5, 2024 and published on February 19, 2025. We have continued to make progress in the intervening year. I showed you these new results during our call, and I presented them in detail to more than 100 researchers from across industry and academia at the Station Q meeting this week. I’ll discuss them during my talk at the APS March Meeting.
We have fabricated a two-sided tetron (in the terminology of Phys. Rev. B 95, 235305 (2017)). Both nanowires were tuned into the topological phase via the topological gap protocol, as in Phys. Rev. B 107, 245423 (2023). This is the topological qubit configuration: there are 4 Majorana zero modes (MZMs), one at each end of each topological nanowire.
We have performed both Z and X measurements. These are the basic native operations in a measurement-based topological qubit. In the Z measurement (which measures the fermion parity of one of the nanowires) we see errors occurring on a time scale of ~10 milliseconds, and we attribute them to “poisoning” by above-gap quasiparticles. In the X-measurement (which measures the fermion parity between MZMs on two different nanowires), we see errors on a 5 microsecond time scale, and we attribute them to the overlap between the MZMs at the opposite ends of a wire.
Comment #32 February 20th, 2025 at 8:21 pm
Chetan Nayak #31: Thank you so much! The fact that your majorana zero mode and topological qubit claims are based on progress that occurred after the submission of your paper to Nature a year ago is immensely clarifying. Given the degree of skepticism that commenters are expressing here, on Twitter, on Hacker News, etc., I’ll eagerly await peer review of the MZM and topological qubit claims. As I said in the post, I’m rooting for Station Q to succeed!
Comment #33 February 20th, 2025 at 8:39 pm
Chetan Nayak #31
> In the Z measurement (which measures the fermion parity of one of the nanowires) we see errors occurring on a time scale of ~10 milliseconds, and we attribute them to “poisoning” by above-gap quasiparticles. In the X-measurement (which measures the fermion parity between MZMs on two different nanowires), we see errors on a 5 microsecond time scale, and we attribute them to the overlap between the MZMs at the opposite ends of a wire.
Are these attributions based on the appropriate exponential scaling with temperature and wire length, respectively?
Comment #34 February 20th, 2025 at 8:52 pm
I’m not closely following Microsoft’s efforts, but I support their general plan. If anyone else remembers, from about 2005 to 2015, Kitaev’s toric code and its close relative the surface code were always of interest but not so much in favor for the fast track to fault-tolerant QC, instead there was more attention on the idea of concatenating small CSS codes like the Steane code in order to maximize the threshold for achieving fault tolerance. That was also the era of negative results for the many-body Hamiltonian version of the surface code, such as Bravyi and Terhal’s no-go theorem for 2D self-correcting memories (2009), as everyone came to understand the issues with string-like logical operators as an obstacle to passive quantum memories. Various generalizations of Kitaev’s toric code ideas, like the 2D and 3D color code and the subsystem/gauge codes with non-commuting checks, made the toric code seem like a respectable grandfather of quantum memories but not the path most people expected to take to fault tolerance. But that outlook turned around in the past decade 2015-2025: the surface code has had a resurgence, thanks to better decoding algorithms that improved its estimated threshold, and thanks to transmon qubits and being championed by major players like Google. Deep reasons that perhaps no one could predict, but that’s how it is with the most visionary ideas. No one knows for sure which fault-tolerant scheme will be used for the first useful quantum advantage, but the surface code is at least in the running (“the original and still the best”).
I had my own brush once with trying to improve on a construction due to Kitaev, in the different subfield of Hamiltonian complexity. Everyone who has taught the Quantum Cook-Levin theorem in courses knows that in a “NO” instance, the ground state energy of the Feynman-Kitaev history state Hamiltonian with its little hopping particle on a line to represent a clock, is pushed up by an amount Omega(T^{-3}) for a circuit with T gates (the part that is traditionally proven with the “geometrical lemma”). I noticed a small thing: by tweaking the history state to be non-uniform, to put a constant amount of amplitude on the beginning t = 0 and the end t = T, we could improve this increase in the no instance to Omega(T^{-2}) [because in the geometrical lemma you care about the overlap with H_in and H_out]. This doesn’t have any complexity implications (it establishes a polynomial promise gap either way) so no one much cared. The fun part of the story is that we later discovered the geometrical lemma was not tight; the original construction due to Kitaev also has an Omega(T^{-2}) increase, and you can prove this with more effort, beyond the geometrical lemma. Then we considered: can we do better by considering arbitrary unitary labeled graphs with arbitrary weights and phases on vertices and edges, is there anything out there generalizing the original construction that can beat this T^{-2}? And we proved a no-go for that. “The original is still the best.” So two times I have noted that Kitaev’s proposals turned out to still be optimal decades later, for reasons that no one could have foreseen. So in conclusion I believe there is some meta-historical principle at work, that both braiding nonabelian anyons and storing quantum states in unpaired fermions in a wire will eventually turn out to be optimal and useful in some way.
Comment #35 February 21st, 2025 at 12:23 am
Scott #30 :
“It’s important to understand that, while topological qubits could reduce the need for active error correction, they almost certainly wouldn’t eliminate it entirely.”
Thanks Scott. That is why I wanted to understand. I got the impression while reading the news release that error correction was not needed for topological qubits.
Comment #36 February 21st, 2025 at 12:35 am
I think you would agree that the for something to be considered a qubit we need to be able to store, manipulate and read out the quantum information. I’m not sure how this is “behaving fully as a qubit” until they can do all three. This is definitely step forward, but without gates, especially non-Cliffords which are hard for Majoranas, I wouldn’t call it a qubit.
Arguably, there are Majoranas for example in non-Abelian quantum Hall states, and there is some ongoing work on me, but we don’t call that qubits, because we yet to store+manipulate+measure them
Comment #37 February 21st, 2025 at 1:39 am
Hi Scott,
Thanks a lot for your post. I would like to draw your attention to the movie published yesterday by Microsoft https://www.youtube.com/watch?v=wSHmygPQukQ where Chetan, Mathias, and others “explain” Majorana 1. Yes, the one in which they “imagine” a battery that never discharges! In my opinion this movie crosses the thin line between overselling (which we all do) and deceiving the public. Why should we care? Because if DARPA is spending taxpayers’ money on this “science”, I’m not surprised the new administration wants to pull the brake. I understand why people at Microsoft do this, but I feel that the academia has the duty to discern what is true and what isn’t.
Best regards,
Emanuele
Comment #38 February 21st, 2025 at 4:55 am
[…] for Microsoft is to provide more concrete experimental evidence. Nayak, responding to criticisms, noted that Microsoft’s paper was submitted a year before additional supporting evidence was gather…. He also wrote that more research would be […]
Comment #39 February 21st, 2025 at 5:10 am
Speaking of Phys. Rev. B 107, 245423 (2023) in #1 and #31:
There was an Editorial: “In this issue of Physical Review B, Aghaee et al. [1] report on an advancement towards the goal of topological quantum computing. While Physical Review readers are well aware that the many minutiæ of procedures, computations, and synthesis may be omitted in any particular dispatch, in this publication the intellectual property of the authors’ employer has prevented the release of some parameters of the studied devices that may be needed in order to reproduce them. As a reflection of the traditional values of the scholarly community, this is not in accordance with the usual norms of the Physical Review journals. Nonetheless, through the rigors of the review process, the published version of this paper reveals much more information than originally had been provided and the authors have agreed to release the remaining device parameters by the end of 2024.”
– Just wonder if the device parameters have been released as promised…
Comment #40 February 21st, 2025 at 6:59 am
Chetan Nayak #31
Thank you for this clarification, the case is much more clearer now.
Nevertheless, it is actually IMHO a very good practice, when the original research becomes available for the community (via arxiv or Nature paper), and only after that the achievements are covered in press. First things first.
Basically, saying that the research is presented only privately to a small number of (probably, very respectable and distinguished) people is not good since all the others are reading breathtaking press releases and even not having a clue what is going on in your lab.
Comment #41 February 21st, 2025 at 7:15 am
Chetan Nayak #31
Ok, the qubits and majorana claims (from the science team) aren’t based on the nature paper ? Please correct me if my understanding is wrong.
But the marketing team/ Microsoft news team seems to say otherwise :
“The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves.”
https://news.microsoft.com/source/features/innovation/microsofts-majorana-1-chip-carves-new-path-for-quantum-computing/
Ok, maybe the data shared at the Station Q meeting (wich I did not see) is strong enough to be conclusive.
But,
Isn’t the announcement too early then ? Shoulden’t the full power communication (with the terms qubit and majorana) wait for the new advances to be written in an article and also peer reviewed and accepted ? (maybe there is already a peer reviewed paper I missed, if so, please enlighten me )
In the hope that your team (or an other) succeeds in bringing QC to a utility scale and fiability.
Comment #42 February 21st, 2025 at 7:46 am
Alex #40 ‘s comment is the view i share.
But but he expressed it much better than I did.
So you can ignore my last comment : Martin#41
Just adding this :
“The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves.”
https://news.microsoft.com/source/features/innovation/microsofts-majorana-1-chip-carves-new-path-for-quantum-computing/
Comment #43 February 21st, 2025 at 8:01 am
I can’t think of any serious group that just puts out a press release saying that they’ve achieved a major milestone, misleadingly couples it to a paper that does not show said milestone, and then shows 0 evidence.
(Their press release says “The Nature paper marks peer-reviewed confirmation that Microsoft has not only been to create Majorana particles, …”)
It’s actually the duty of serious scientists to push back on this type of BS.
Contrast this to Google’s surface code scaling result:
– August 2024: After presentations to experts, Google put out their paper on arXiv
– December 2024: Paper is peer reviewed and published in Nature. PR is timed with publication, and mostly reflects content of paper.
Finally, it’s not a matter of suspecting there to be a fraud: I sincerely hope that the team has been able to demonstrate what they are claiming.
It’s a matter of principle: Major scientific claims need to be peer-reviewed. If they just claim they’ve done it, what’s the point of peer review?
Comment #44 February 21st, 2025 at 8:39 am
“Anyons … behave neither as fermions nor as bosons.”
You are wrong. All particles are excitations of either fermionic or bosonic degrees of freedom—the celebrated spin-statistics theorem, which is taught in every introductory course on quantum field theory.
Every projective representation of SO(3) is a faithful representation of SU(2), and the classification of irreducible representations of SU(2) is an elementary problem in every introductory QFT course. Irreducible representations of SU(2) are classified by half-integer weights. By the spin-statistics theorem, particles in the integer representation have bosonic statistics, and particles in the other representations have fermionic statistics.
Having familiarized ourselves with this elementary result, we note that it is mathematically impossible for any particle to “behave neither as a boson or a fermion.” Then what representation would it transform in? It doesn’t make any sense.
Indeed, Majorana modes are fermionic. I’m not sure where you got this idea that they have “impossible statistics” from, and it sounds like it derives from a poorly written popular science article or such. But it is not correct, and I’m curious who said this.
All the best.
Comment #45 February 21st, 2025 at 8:47 am
Scott #24, Regarding D-Wave, I see Scott still views D-Wave very negatively and I asked Steve Jurvetson who is a big proponent of D-Wave on why Scott hates it even now (I see Scott mentioning other names like Rigetti but never D-Wave), he response below on X to me:
“Theorists become much less interesting when engineers build actual machines. Also, there were some early mistakes made by the company deserving of criticism. But after a decade of criticism, cognitive dissonance and identity politics can capture the mind.”
Maybe it’s possible D-Wave has improved after criticism by Scott and deserves attention
Comment #46 February 21st, 2025 at 8:52 am
Physics anon #44: Nope, the spin-statistics theorem holds only in 3-dimensional space. Anyons famously evade it by being confined to 2-dimensional media. Wikipedia explains this. Engage in good faith or stay in the moderation queue.
Comment #47 February 21st, 2025 at 9:38 am
Nikhil Kamma #45: Steve Jurvetson is not just a “proponent” of D-Wave, but a major investor in them. Did you know that, and if so, why didn’t you mention it?
I’ve been perfectly clear that D-Wave has improved a lot since the Geordie Rose era, and has published interesting annealing physics, although they still also make cringey claims about helping Volkswagen and other customers solve practical optimization problems (without giving evidence for any improvement over classical). In the meantime, though, I’ve personally found the rapid experimental progress on gate-based QC to be much more exciting so have spent more time covering that.
Comment #48 February 21st, 2025 at 9:51 am
Nikhil Kamma #45
I have been playing around with Dwave on their plateform. I am yet to see a convincing article on a real optimization problem.
I do believe it can give interresting results on pure Ising/Max Cut/QUBO (they are very similar) problems.
https://arxiv.org/abs/2403.00910
But, combinatorial optimisation problem turned into a qubo, I am yet to see it beat classical heuristics (in a fair comparison).
Comment #49 February 21st, 2025 at 10:48 am
So, I’ve done some more reading, and I admit I may have been wrong and/or slightly confused about statistics in the two dimensional case. That said, you should have been clearer in your original post so that this misunderstanding wouldn’t have occurred. I encourage you to edit it to that effect.
All the best.
Comment #50 February 21st, 2025 at 11:15 am
Physics anon #49: I linked to the Wikipedia article so you or anyone else could read the details of how anyons work for themselves. You are a troll.
Comment #51 February 21st, 2025 at 11:39 am
Hi Scott,
Thanks for adding the hyperlink for me, I appreciate it.
That said, I’m upset with how you’re admonishing me here. I resent being called a troll. Yes, I was wrong about particle statistics in two dimensions. Quantum mechanics, symmetry representations and particle statistics are really technical, counter-intuitive subjects that are confusing to learn. I’m just a student. I read this brief part of your post, and found it confusing, because it contradicted what my professor reiterated many times about spin statistics and symmetries. I had never learned about anyons in my course. So, I wrote something up to contribute my knowledge about QM and have a good-natured conversation or debate about this. Maybe I was too confident in how I wrote it, was that the problem? Yes, I was wrong, but you’ve said there’s nothing wrong with saying something wrong or being confused as a student, because it’s part of learning! Well: now you’re shaming a student who was genuinely confused about a really hard topic.
No hard feelings, I just want to know why you’re doing this?
Comment #52 February 21st, 2025 at 11:47 am
It’s quite counterintuitive that “things” confined in a 2d space can be “braided” around one another. Unless maybe the space is a mobius strip or something twisty like that?
Comment #53 February 21st, 2025 at 12:11 pm
Great FAQ Scott.
thanks for sharing.
Comment #54 February 21st, 2025 at 12:26 pm
Thank you for your response Scott and Martin.
No, I was not trying to hide that Jurvetson is an investor, but yes proponent was not the right word.
Comment #55 February 21st, 2025 at 12:37 pm
fred #52: See, e.g., Figure 4 of this paper for an example of what anyonic braiding looks like.
Comment #56 February 21st, 2025 at 12:44 pm
Physics anon #51:
Yes, I was wrong, but you’ve said there’s nothing wrong with saying something wrong or being confused as a student, because it’s part of learning!
Next time, if you have questions, please try asking them in a spirit of genuine curiosity, rather than opening with “You are wrong” (spoiler alert … as you now agree, I wasn’t!), then lecturing me on physics with approximately one additional snide remark per sentence.
I do all this as a free service to the world, and I ask the world only for good vibes, rather than miserable sneering ones, as my fuel to keep going.
Comment #57 February 21st, 2025 at 12:55 pm
Scott #55
thanks! and from the wiki illustrations it seems that the braiding is happening in spacetime, so 3d.
https://en.m.wikipedia.org/wiki/File:Particle_exchange_2d_anticlockwise.gif
makes sense.
Comment #58 February 21st, 2025 at 1:26 pm
Thanks. I’m sorry for my behavior. I was snide. I’m doing poorly in my program, and I’m not sure if I can succeed in physics. I am afraid about my quals. I’m not doing well in my courses (passing them, but not well). I feel like everyone else is smarter than me. I’m struggling. Being snide is a defense mechanism I use to feel better about myself. I pick one advanced topic I know about (in this case, spin representations) and try to use it to feel smarter than other people. Making other people feel unknowledgable makes me feel better about myself, I am very sorry to say. It’s something I’m not proud of. I want to stop doing this. But I don’t know how else to cope with being in a program where it feels everyone else is more successful than I am. I thought I found something that a bigshot professor (you) misunderstood, which made me feel good about myself and my intelligence in physics. I was obviously wrong. In truth I am miserable. Sometimes I feel PhD was a mistake. I don’t know where I’m going with my life. What would you recommend?
Comment #59 February 21st, 2025 at 2:35 pm
Physics anon #58: Therapy?
In any case, sorry, this is too far off topic; nothing further on this in the thread.
Comment #60 February 21st, 2025 at 3:16 pm
Can anyone explain to me in simple terms what are MZMs and why they are useful in quantum computation? I know they are more fault tolerant but am not sure why. Is this because they are their own antiparticles?
I have only taken some intro quantum computing and quantum classes. It’s mostly an interest for me. I just really want to understand why MZMs would be a big breakthrough.
An explanation on the distinct between MZMs and Andreev states would be great as well. Why can qubits be built on MZMs and not Adreev states? thanks!
Comment #61 February 21st, 2025 at 3:46 pm
QC enthusiast #60: Good q! The best place to start here is understanding exchange statistics.
Imagine I have two indistinguishable electrons. If I swap their position in space, the quantum state is identical to what it was when I started (maybe fudging some phase factor — regardless, not measurably different). This is a function of their Abelian exchange statistics. MZMs have non-Abelian exchange statistics; exchanging their position in space performs a logic operation. This allows for “inherent fault tolerance” because the information is stored in the exchange process, not locally in some state that is susceptible to decoherence (e.g., atomic energy level). It also allows operations to be performed extremely reliably… in theory.
The difference between Andreev bound states and Majorana bound states is a little complicated and someone else is welcome to explain in more detail (I am procrastinating lol) but for the purposes of what I just discussed above, ABSs don’t have interesting exchange statistics.
Comment #62 February 21st, 2025 at 4:23 pm
Amir #43
We adhered to standard practices in our announcement. Like other top-tier companies in our field, we coupled the publication of our Nature paper with new unpublished results presented at a scientific conference. This follows a long tradition of presenting new findings at scientific conferences before publishing them in a paper.
We appreciate your interest in our work and encourage you to attend Chetan Nayak’s talk at the APS March Meeting to learn more about our latest, breathtaking results.
Comment #63 February 21st, 2025 at 5:25 pm
Emanuele, #37
DARPA spent two years evaluating our program with a large team. They hired 40 people to review our results, architecture, and engineering plans. They analyzed our roadmap including the underlying science and engineering, milestones, budgets, hiring plans, supply-chain issues. We have shared the latest measurement results and proprietary information with them under NDA. Their scientists are in our labs, measuring our devices. I don ‘t think that watching a promotional video aimed at general non-technical audiences is a reliable way to judge whether they are wasting their money.
We shared the latest data on both X and Z measurements on topological qubits at the Station Q meeting this week. Chetan Nayak will share it with a broader audience at the APS Global Physics Summit: https://summit.aps.org/events/MAR-F14/1 . We invite everyone to attend and discuss with us.
Comment #64 February 21st, 2025 at 6:06 pm
zulfi #62,
So you are defending the press release saying the peer reviewed nature paper confirms what it does not in fact confirm?
Comment #65 February 21st, 2025 at 7:08 pm
Adam Treat #64:
We are defending Microsoft Quantum Majorana findings. Here is what the Microsoft Quantum blog post says: “Research published today in Nature, along with data shared at the Station Q meeting demonstrate our ability to harness a new type of material and engineer a radically different type of qubit that is small, fast, and digitally controlled. ”
https://azure.microsoft.com/en-us/blog/quantum/2025/02/19/microsoft-unveils-majorana-1-the-worlds-first-quantum-processor-powered-by-topological-qubits/?msockid=0f4f6db5a254609603ec782ea33a61ad
Comment #66 February 21st, 2025 at 7:41 pm
zulfi #62:
Sorry, no, this is absolutely not standard practice.
Also you’re conflating two entirely different things:
1 – Presenting unpublished results at conferences – yes, this is standard practice and perfectly fine.
2 – Claiming that a Nature paper provides “peer-reviewed confirmation” of results that are explicitly not in that paper: This is absolutely not standard practice, and it’s exactly what you did with your press release.
It definitely shows a failure of integrity on the part of the team to be issuing press releases to a general audience, saying they have “Peer-reviewed confirmation of topological qubits,” when the actual, peer-reviewed text says something significantly more modest and never confirms that milestone.
From outside, this looks like a PR synergy campaign with little care for scientific truth: (“We got something into Nature .. let’s also announce the next step in the same time so it gets coverage!”). It leads to the widespread preception that the scientific community (ie. nature’s peer review process) has now signed off on topological qubits at Microsoft which it explicitly did not.
To repeat myself, the more correct way which would have demonstrated greater integrity would have been to do e.g. what Google AI did:
1- They presented to experts
2- Put their paper on arXiv in August 2024
3- Waited for peer review and publication (December 2024)
4- Issued PR that (more or less) reflected the paper’s content
This whole things matters to me because what Microsoft has done here undermines the integrity of the peer review process, misleads, and sets a terrible precedent for how major scientific claims are communicated.
Please do better next time.
Comment #67 February 21st, 2025 at 8:44 pm
Bell Labs. Now there is a name I haven’t heard in a while. In another world where the breakup didn’t happen, I’d be curious to see what their research group on quantum computing would have looked like
Comment #68 February 21st, 2025 at 9:28 pm
Evan #67: It was actually Bell Labs, before the Lucent and Avaya breakups, that produced both Shor’s algorithm and Grover’s algorithm.
Comment #69 February 22nd, 2025 at 12:43 am
Amir # 66
Thank you for being such a dedicated steward of the scientific review process.
However, in this particular case, your analysis seems extremely selective. Rather than discussing our colleagues from partner companies, whom we hold in the highest regard, I would like to present the facts as I see them:
1. We published our results in Nature on the X measurement.
2. Following a thorough due diligence process, we signed a definitive agreement with the United States Government based on our results.
3. We showcased our latest X and Z measurements at Station Q.
4. We shared our claims externally.
We would be delighted to further discuss our findings with you at APS. In the meantime, it seems that those spreading loose accusations are simply seeking attention.
Comment #70 February 22nd, 2025 at 1:34 am
Scott #68: I never knew that. And knowing it now makes me actually sad to think about what could have been.
Comment #71 February 22nd, 2025 at 6:20 am
Evan #67,
Having seen the workings of Bell Labs from close quarters, those times were surely a contrast to today’s rushing out the door with heavy PR approach of the corporate world (with obvious benefits to be gained from stock price gains). In fact the demise of the standing of Bell Labs is some part due to underplaying their research and getting out products on time, in spite of the fact that they could count at least 5 Noble Prizes as their direct contribution, in addition to too numerous to mention tech breakthroughs. If Bell Labs and their parent AT&T had adequately capitalized on the innovations it would be as big as the top 5 companies combined of today without doubt.
Comment #72 February 22nd, 2025 at 7:29 am
For the love of God, can _someone_ from Microsoft just do the absolute minimum of acknowledging that putting out in a press release:
“The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves.”
Was a mistake?
Think of this as a “blink twice” to make sure you’re not all being held hostage by your Microsoft-PR overlords.
Suge, zulfi, anyone?
Scott, #8 and #9 would appear to be in violation here of your cancelation recommendations. I’m so tired of the insidious nature of PR campaigns that cannot admit even the tiniest bit of fault on behalf of large corporations. It is like they take delight in gaslighting. That is definitely what it feels like Microsoft-stooges are doing here in this thread.
Comment #73 February 22nd, 2025 at 7:46 am
Chetan Nayak #31
> We have performed both Z and X measurements. These are the basic native operations in a measurement-based topological qubit.
Obviously you need non-clifford operations too, but fine for now I suppose.
> In the Z measurement (which measures the fermion parity of one of the nanowires) we see errors occurring on a time scale of ~10 milliseconds, and we attribute them to “poisoning” by above-gap quasiparticles.
If I understand, this number will not improve with greater separation of the mzms, Is there any point to “topological protection” then?
Comment #74 February 22nd, 2025 at 8:36 am
Matthias Troyer #63
For me it looks like every new attempt to respond the reasonable criticism regarding PR policy of your team does make your position much weaker and less clearer to the community.
DARPA is military funding agency, do you really think that those 40 people are doing valuable expertise on quantum computing hardware platform? Or how you say, do they even dare to measure you devices? As an experimentalist I’m surprised to hear this (though I have never worked in the US so maybe there is are very clever experimental physicists in military funds, you’ll never know). All this arguments look like a complete nonsense.
Your meeting in station Q doesn’t allow you to create such a strong press release, this is so basic, why do we even wasting time discussing that?! Also, X and Z measurement is nice progress, but it is not a full qubit operation, period. My position is that you need at least single qubit non-cliffords to say that “my platform is potentially scalable”.
Comment #75 February 22nd, 2025 at 9:05 am
A comment on the spin/statistics discussion.
Strictly speaking, the terms “boson” and “fermion” refer to the Grassmann parity of an operator or state. In turn, these control how two decoupled quantum systems can be combined into a new system: bosonic operators in one system will commute with all operators in the other system, while fermionic operators will anti-commute with fermionic operators in the other system. For fermionic Majorana modes, this leads to the result that the 2n modes associated to n decoupled nanowires will form a Clifford algebra Cliff(2n) with 2n generators rather than n commuting copies of Cliff(2).
These considerations are separate from the topological properties of anyons. A topological phase of matter in 2+1d made out of bosonic spins can have anyons which braid as particles of spin 1/2, but they will still be “bosons”, as you cannot really produce a fermion out of bosonic degrees of freedom. On the other hand, certain topological phases of matter are only available if you have fermions in the system.
This topological version of the spin/statistics relation is more mysterious than the standard spin-statistics theorem, which assumes relativistic invariance and is relatively easy to prove.
Comment #76 February 22nd, 2025 at 9:08 am
zulfi #69:
Dismissing legitimate concerns as “loose accusations” from those “seeking attention”, isn’t a good look. This is about norms on how major breakthroughs are validated and communicated.
Let’s be precise:
1. Your press release claimed the Nature paper provided “peer-reviewed confirmation” of Majorana particles
2. Nature’s editorial team explicitly stated it does not confirm this, as did the paper itself.
3. This is a direct factual contradiction, not a matter of interpretation.
And just so we don’t go in circles, if you or anyone else from Microsoft respond, please at least answer this question:
Do you agree that this sentence is factually incorrect: “The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles…”?
https://news.microsoft.com/source/features/ai/microsofts-majorana-1-chip-carves-new-path-for-quantum-computing/?ocid=FY25_soc_omc_br_x_QuantumMajorana
(This is one of several instances where the claim is made)
Moreover, none of your points address this core issue:
– Government agreements and funding are irrelevant to scientific peer review. I’m glad you got funding. Not relevant to discussion.
– Private presentations on the day of the announcement at Station Q are not peer review
– “Sharing claims externally” — which has yet to occur, at least beyond a “we did it, trust us” — is not peer review
Here’s another example of how things should be done: after 30 something years, LIGO detected GWs in fall of 2015. It was pretty clear to the whole team & others in the field who knew about the results. Yet they waited, they submitted to PRL, and timed the PR with publication in 2016.
This scientific process, the institutions and norms around it are damn important for self-correction and progress of the scientific project. They also exist to prevent claims of major breakthroughs from resting solely on private demonstrations and corporate announcements. If you have achieved these results, submit them for peer review. Claiming peer-reviewed confirmation of results that haven’t been peer-reviewed is just lying to the public.
Your response suggests you think that scientific peer review is a business inconvenience rather than an important part of the science.
Will you correct your press release’s mischaracterization of what has and hasn’t been confirmed by peer review?
Comment #77 February 22nd, 2025 at 10:09 am
Amir #76
Yes lets be precise.
This statement is factually correct, as is the one on the Microsoft Quantum blog:
Research published today in Nature, along with data shared at the Station Q meeting demonstrate our ability to harness a new type of material and engineer a radically different type of qubit that is small, fast, and digitally controlled. ”
In our 2023 PRB paper, we demonstrated that we could reproducibly create topological superconductivity and MZMs with very high confidence (which we quantified in the paper; please take a look there for the details). In our Nature paper, we showed how to measure the fermion parity shared by a pair of MZMs. This demonstration also placed even more constraints on any non-topological explanation of our data. Here is the relevant statement in our paper:
“This effect completely washes out the flux-dependent bimodality unless the coupling between the ‘hidden’ Majorana modes and the visible MZMs is less than 1 neV or the hidden Majorana modes are effectively gapped out, as shown in Supplementary Fig. 4.”
You might have read the sentence: “These measurements do not, by themselves, determine whether the low-energy states detected by interferometry are topological,” without noticing the one that followed: “However, our data tightly constrain the allowable energy splittings in models of trivial Andreev states.” If so, you might have missed that the combined results of the 2023 PRB and this week’s Nature paper are more significant than the interferometry measurements alone. Additionally, you may not have realized just how stringent the constraints have become.
It would be helpful if stewards of the scientific process would actually take the time to read and understand before pontificating.
Reasonable people can disagree about the standard of proof (95% confidence level, 99% confidence level, etc.). We would be happy to have that discussion with you, if that is the discussion that you wish to have. We believe that we have set a very high bar and exceeded it in these two publications, and this is accurately reflected in the Microsoft Quantum Blog post and the Microsoft Source story that you referenced.
Comment #78 February 22nd, 2025 at 11:33 am
zulfi #77:
Sorry just looking for clarification… is the claim
1) that topological qubits were achieved in the 2023 PRB paper?
2) that topological qubits were achieved by combining the results of PRB and nature?
3) that topological qubits wrere achieved by combining 2023 PRB and the Station Q data?
I don’t mean to be too invective, but at best this is very confusing. If the claim for topological qubits rests on the Station Q data, then it has not been peer reviewed, full stop. Personally I don’t put too much stake in “formal” peer review, but these results should be on the arXiv soon so that the community can vet them.
Whether or not you agree with it, I think it’s clear that the release has been met with considerable skepticism within the academic community. You can argue about whether that is a fault of Microsoft’s PR or academia, but clearly something is not working here.
FWIW I very much hope that this is real. But I really wish I could be reading an arXiv paper right now to decide for myself, given the magnitude of the claims.
Comment #79 February 22nd, 2025 at 11:52 am
zulfi #77:
Your response makes even less sense in light of your colleagues statements. If, as you now argue, the press release is justified solely by the published results in PRB 2023 and Nature 2025, why would Chetan and Matthias need to emphasize unpublished results, private demonstrations, and government evaluations? Why discuss X and Z measurements with specific error rates that appear nowhere in these papers?
Your team is alternating between two positions:
1 – “Trust us, we have more results coming”: pointing to private demonstrations, government contracts, and unpublished measurements
2 – “No no, the published results alone justify our claims”: arguing the papers themselves validate your press release
I push back strongly on claims of type (1) because they undermine the norms and practices essential to scientific progress. You call this “pontificating” but since we clearly have different backgrounds and perspectives on the role of peer review in science, we can leave it at that.
I disagree with your new statements above (which fall under 2). The Nature paper’s editorial team (and paper itself) explicitly stated the result “does not represent evidence for the presence of Majorana zero modes.” Yet Microsoft’s press release claims “The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles.” These statements directly contradict each other. If we can’t agree on something basic like this, then there’s no real point on continuing debate, I let the Scott and readers of blog come to their own conclusions.
Comment #80 February 22nd, 2025 at 12:26 pm
Msft press release says:
“A new paper published Wednesday in Nature outlines how Microsoft researchers were able to create the topological qubit’s exotic quantum properties and also accurately measure them, an essential step for practical computing.”
(https://news.microsoft.com/source/features/ai/microsofts-majorana-1-chip-carves-new-path-for-quantum-computing/?ocid=FY25_soc_omc_br_x_QuantumMajorana)
As Amir pointed out, that’s factually incorrect.
Zulfi is referring to a blogpost by Chetan Nayak, which says:
“research published today in Nature, along with data shared at the Station Q meeting, demonstrate our ability to harness a new type of material and engineer a radically different type of qubit”
(https://azure.microsoft.com/en-us/blog/quantum/2025/02/19/microsoft-unveils-majorana-1-the-worlds-first-quantum-processor-powered-by-topological-qubits/)
That’s not factually wrong (though, as Amir also mentioned, not the right way to make a big scientific announcement).
What is written in the blogpost doesn’t invalidate Amir’s point that the press release *is* factually incorrect.
Comment #81 February 22nd, 2025 at 2:43 pm
You can always just say “yes, some of the press release was wrong. We apologize for that. The topological qubit claim has not yet been peer reviewed. But we do believe that we have a topological qubit, and are excited to release that data for peer review soon.”
Always an option. I personally care more about whether this is real or not.
Comment #82 February 22nd, 2025 at 3:12 pm
Adam Treat #72
You object to the following sentence in the Microsoft Source story:
“The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves.”
So let’s break it apart:
“… Microsoft has **not only** been able to create Majorana particles …”
In our 2023 PRB, we showed that we could create the topological phase and Majorana zero modes. In the Nature paper, we showed that we could **not only** do that, but also could do more. Hence, the first half of the sentence is true.
You may have technical criticisms of that paper. If so, let’s discuss your criticisms. We stand behind that peer-reviewed paper 100%. You can choose to disbelieve any published paper out there. But that’s a completely separate issue than whether Microsoft PR is misleading. So, to re-iterate, the first half of the sentence is true.
Agreed, Adam?
The second half says “… but can also reliably measure …”
Indeed, we showed in the Nature paper that we can measure the fermion parity of a topological superconducting wire. So the second half is true as well.
Agreed, Adam?
MOREOVER, our fermion parity measurements strengthen the evidence that we have a topological phase and MZMs (as you might expect — the evidence should become stronger as you do more different types of measurements in more complex devices). Hence, it again strengthens the first half of the sentence.
So if we both agree that the sentence is true, why are you arguing?
Comment #83 February 22nd, 2025 at 3:52 pm
Suga #82:
Just to clarify, you are saying that Microsoft confirmed the existence of Majoranas in the 2023 PRB paper. So all this business about a new phase of matter is two years old?
Comment #84 February 22nd, 2025 at 5:29 pm
Amir #79
Yes, it’s best for readers to draw their own conclusions.
You chose to focus on a specific sentence in the Microsoft Source story, not me. That sentence is fully supported by the published papers. It seems there’s a reluctance to acknowledge that you might have been mistaken about the initial issue you raised. Further discussion on new arguments seems unnecessary, as it feels like you’re arguing just for the sake of it.
Comment #85 February 22nd, 2025 at 5:35 pm
Still confused #78
The specific statement in the Microsoft Source story that Amir #76 criticized is:
“The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves.”
Let’s break it apart:
“… Microsoft has **not only** been able to create Majorana particles …”
In our 2023 PRB, we showed that we could create the topological phase and Majorana zero modes. In the Nature paper, we showed that we could **not only** do that, but also could do more. Hence, the first half of the sentence is true.
The second half says “… but can also reliably measure …”
Indeed, we showed that we can measure the fermion parity of a topological superconducting wire. So the second half is true as well.
MOREOVER, our fermion parity measurements strengthen the evidence that we have MZMs (as you might expect — the evidence should become stronger as you do more different types of measurements in more complex devices).
Hence, the statement is true.
And the reason you were still confused is that you didn’t consider option 4:
4) The 2025 Nature paper built on the 2023 PRB paper. Building on both of those results, we have made topological qubits.
Let’s be perfectly clear. This is a new result, building upon previous published papers. You need topological superconducting nanowires (PRB 2023). You need a way to measure fermion parity (Nature 2025). Then you need to put two wires together and do a fermion parity measurement across the wires. We’ve done this.
It was presented at a conference, as new results often are.
This is fairly common. I can give you many examples, if you would like. And once something is presented at a conference, it won’t stay secret for long, and it is bound to be reported in the press.
Looking forward to sharing more at APS and thank you for rooting for our success in this field.
Comment #86 February 22nd, 2025 at 5:42 pm
Scott, at this point I think we need you to step in or at least issue a new update as referee. This PR campaign being waged in your comment section would seem to be a case of flat out lying. It would seem Chetan Nayak’s update is now not the operative word out of Microsoft because numerous people _from_ MS are now telling us in this comment section that the press release was not confusing and was perfectly fine. No follow-up paper is needed in this new telling.
What is going on? Has Microsoft had peer review confirming their finding that they were able to create Majorana particles or not? What say you?
Comment #87 February 22nd, 2025 at 6:02 pm
Adam Treat #86: No, I’m not a “referee,” just a messenger and a moderator.
I think it’s obvious to everyone at this point that Microsoft Station Q can either publish a research paper with direct evidence for Majorana zero modes and topological qubits, or else lose credibility with the research community. On the other hand, Station Q now ought to be given a chance to do that. So please let’s calm down, and I hope we’ll learn more one way or the other over the next few months.
Comment #88 February 22nd, 2025 at 6:17 pm
Scott #87,
Okay, it would seem they claim to have already published a peer reviewed paper with direct evidence for Majorana zero modes. See the press release they continue to defend and Suga #87’s comment.
Seeing as how you say they _still_ need to accomplish what
they say they’ve _already_ accomplished, they’ve lost all credibility in my eyes and I had nary a concern in the fight one way or the other before this blog post. Good job Station Q/Microsoft!
Comment #89 February 22nd, 2025 at 6:49 pm
Zulfi #85: Thanks for explaining, I now understand your position. Majorana fermions were established assuming the topological gap protocol in 2023, and now you have the capabilities to manipulate those MZMs as qubits.
I think the disconnect was that the existence of MZMs seemed like the big step externally, whereas internally it seems that MS already accepts that MZMs have been observed. Hopefully with the additional capabilities presented at the conference, existence of MZMs makes its way out of MS and into mainstream acceptance.
As Scott said, I think the proof will be in the pudding of the upcoming publication. Certainly, excitement has been raised sky high. Like Scott, I sincerely hope that direct evidence of a topological qubit will be delivered.
Comment #90 February 22nd, 2025 at 8:16 pm
Hi Scott, thanks for sharing your expertise on this, very much appreciated as always.
What probability would you assign to MSFT being able to scale this to thousands of error-tolerant qubits over the next 5 years as MSFT PR claims? And as a theorist, how much do you trust your own judgment on this?
I’m asking because in your answer to Q8 you’re equating this milestone with where other qubit technologies were 20-30 years ago. Obviously that’s true from a number of qubits perspective (if the result holds) but MSFT clearly claims in their PR that these milestones aren’t comparable, presumably bc the topological 1 qubit is a far more advanced milestone on the way to scalable QC. You seem unconvinced. Where’s the discrepancy?
ie that there is no cause to believe MSFT will have scaled
Comment #91 February 22nd, 2025 at 8:34 pm
A B #90: I have no idea, and before even trying to give probabilities for scaling this up, I’d first like clearly to establish the truth about what’s already happened!
Comment #92 February 23rd, 2025 at 6:23 am
In the paper, the authors identify two devices: Device A (a gated superconductor-semiconductor nanowire) and Device B (a triple quantum dot interferometer). The paper describes a general experimental setup comprising two main components:
1. A 1D Topological superconducting nanowire (with Majorana Zero Modes at its ends): This part of the experiment involves a nanowire engineered to host Majorana modes at its endpoints when tuned into a topological phase.
2. Quantum dot interferometer (likely a TQDI): This part involves using quantum dots in an interferometric loop to measure fermion parity, which indirectly probes the quantum state of the Majorana modes at the wire’s ends.
Measurements are taken from Device A and Device B, showing consistent behavior. Still, the specific roles of these devices in the context of the experiment are not clearly outlined in the paper. I suppose both devices contain components from the setup described above, possibly with variations to test different configurations (e.g., variations in quantum dot tuning or coupling parameters).
Thus, based on the context of the experimental results, both devices appear to contain parts of the gated superconductor-semiconductor nanowire with Majorana modes and an associated quantum dot interferometric setup.
**From a philosophical point of view**, if the authors report results from two separate and different devices—Device A and Device B—showing similar behavior, the observed phenomena might be reproducible across different samples rather than being a one-off occurrence in a single device. If so, it shows consistency across the two devices, which gives researchers confidence that the experimental results are reliable. It suggests that the effects—such as the flux-dependent h/2e-periodic bimodality, the random telegraph signal (RTS) in the quantum capacitance, and the inferred fermion parity Z of MZMs in the wires—are intrinsic to the device architecture and the underlying physics (e.g., Majorana zero modes), rather than due to a fabrication anomaly or measurement artifact in one sample.
Observing the same effects in two different Device A and Device B **that were fabricated and measured independently** strengthens the statistical basis of the claims, showing that the results are not just due to quantum noise or isolated imperfections.
To prove robustness, Device A and Device B should be produced in separate fabrication runs and be physically distinct chips. The point is that the devices A and B share the same design. Showing reproducibility in devices with the same design doesn’t yet prove robustness across entirely different designs. For complete robustness, one must demonstrate that the effect persists under variations in design parameters, materials, or even across different device architectures!
Comment #93 February 23rd, 2025 at 8:24 am
Suga #82
Just how stupid do you think this blog’s readers are?
<You object to the following sentence in the Microsoft Source story:
<"The Nature paper marks peer-reviewed confirmation that Microsoft has not only been able to create Majorana particles, which help protect quantum information from random disturbance, but can also reliably measure that information from them using microwaves."
Actually, many of us are objecting to this sentence because of its Trumpian, black is white, white is black nature.
<So let's break it apart:
<… Microsoft has **not only** been able to create Majorana particles …"
Ok, let's, but let's also not omit the opening phrase "The Nature paper marks peer-reviewed confirmation that …" (did you think we wouldn't notice such an omission?). The objection repeated ad nauseam throughout this thread is that "no such peer-reviewed confirmation" was provided at all precisely because Nature's editorial team went out of its way to explicitly opine in its, shall we say, less-than-full-throated endorsement to proceed to publication:
“The editorial team wishes to point out that the results in this manuscript do not represent evidence for the presence of Majorana zero modes in the reported devices. The work is published for introducing a device architecture that might enable fusion experiments using future Majorana zero modes.”
One might wonder why the editorial team felt it was so important to nervously emphasize that the results do not imply the presence of Majorana zero modes but anticipating your efforts, we don’t have to read much further to find out: (did you think we would not read these? did you think that we would not notice you consisting ignore the actual peer-review?).
Reviewer Comments:
“What I do NOT like is the way the article is written which, sometimes subtly and sometimes more crudely, uses a language and wording that at all times leads the reader to think that we are
dealing with a measurement that demonstrates parity in a topological qubit based on Majorana states. The examples are many and here I highlight only a few:
… Given that this experiment cannot firmly exclude a fine-tuned zero-energy state, the abstract’s statement that “these results are consistent with … Majorana …” without mentioning alternatives could be misleading. Especially considering that this may be the take-home message in press release and media coverage. The authors should revise their abstract to prevent potential misinterpretation.”
“Therefore, the novelty of this manuscript does not lie in providing stronger evidence for MZMs, but in its methodological approach”
“All referees are in agreement that the central result here is not a definitive yes/no on the Majorana question.”
…
So we see that the very existence of Majorana was a central point of contention for all reviewers and in fact two recommended the paper *not* be published in Nature at all with reviewer 4 particularly caustic:
“I would like to stress the fact that the present manuscript comes after a decade of focused and well-funded research. The Microsoft team has been producing a fairly large amount of work often published in high-visibility journals, including many from the npj family. Most of the claims made are still controversial, a number of papers have raised significant criticism, and some were eventually retracted due to proven scientific misconduct. As a result, a clear, unambiguous observation of topological superconductivity and Majorana edge modes is still missing.”
Finally both the abstract (as insisted by a reviewer) and conclusion of the Nature paper make clear that that Majorana zero modes have not been demonstrated but that instead
“Here we introduce a device architecture … These measurements do not, by themselves, determine whether the low-energy states detected by interferometry are topological. However, our data tightly constrain the allowable energy splittings in models of trivial Andreev states.”
In conclusion, our findings represent substantial progress towards the realization of a topological qubit based on measurement-only operations… ”
None of this involves any clear assertion whatsoever about the creation of a topological qubit or a Majorana (it is not, for example, a case of reasonable people disagreeing) and further, we know from the reviewer reports and an earlier arxiv version any such claims that even hinted at such creation were excised including the removal of the following from the original abstract:
“These results are consistent with a measurement of the fermion parity encoded in a pair of Majorana zero modes”.
All of which makes suga’s gaslighting even more disingenuous and unconscionable:
<In our 2023 PRB, we showed that we could create the topological phase and Majorana zero modes. In the Nature paper, we showed that we could **not only** do that, but also could do more. Hence, the first half of the sentence is true.
Plainly then, this (and any variant in this thread) is simply a pathetic, bald-faced lie that is unbefitting for a blog in which I had once assumed that good-faith argumentation was de-rigour.
Suga, you may believe that you helped show the existence of Majorana in the Nature paper, the other Microsoft lackys activated here may all believe they helped showed the existence of Majorana in the Nature paper and heck, another Nature paper may even one day demonstrate a genuine “peer-reviewed confirmation” of the existence of Majorana (although I doubt it) but no amount of gaslighting can ever change the plain fact that there was NO such “peer-reviewed confirmation” of the existence of Majorana in that Nature paper.
<So if we both agree that the sentence is true, why are you arguing?
Actually I think you'll find most of us don't agree and we are arguing because, truth and the process of attaining truth matters. Microsoft's claims laid out in this thread are Trumpian on so many different levels: The initial brazen lie, the Roy-Cohen-like doubling down, the snide characterization of opponents, the track record of fraud, the bad-faith omissions but for mine, the most contemptible of all is gaming the (peer-review) system at the expense of those respecting it. Trying to co-opt the imprimatur of a supposedly prestigious publication like Nature to dishonestly gain further funding is, in the end, done at the expense of researchers of integrity and the discoveries they might have made.
Comment #94 February 23rd, 2025 at 10:24 am
on a related note, today we have the 1,000,000-th breakthrough about to give us free power “real soon now”:
https://www.sciencealert.com/new-record-reactor-crosses-crucial-milestone-in-achieving-nuclear-fusion?utm_source=ScienceAlert+-+Daily+Email+Updates&utm_campaign=bde5bad12b-RSS_EMAIL_CAMPAIGN&utm_medium=email&utm_term=0_fe5632fb09-bde5bad12b-366065442
Comment #95 February 23rd, 2025 at 3:30 pm
Thanks to the stewards of the scientific process for making it clear that “the presence of Majorana zero modes in the reported devices” was neither accepted nor rejected by Nature’s review process (and similar for the topological qubit claims).
Thanks also to the people from Microsoft Station Q who comment here. Your comments made the actual situation much clearer. I strongly disagree with anybody who dares to compare your clarifications with Trumpian alternative facts. We may not like it, but most of us react to criticism in such ways. This is not Trumpian, but just normal human behavior. (Some readers may be convinced that they never react like that, but I believe Matthew 7:3 is right: “You can see the speck in your friend’s eye, but you don’t notice the log in your own eye.”) And it is good, because it means that you are convinced that you really created topological qubits, and are confident to be able to build on this and create something usefull from it.
Comment #96 February 23rd, 2025 at 6:14 pm
DedicatedStewardOfTheScientificReviewProcess #93 — I completely agree with you and I am also exasperated as you are, even though I would not have expressed my exasperation so bluntly
Suga, Zulfi and all Microsoft people — like gentzen #95 I too thank you for your engagement here. I suspect your stubbornness and insistence on refusing to see the point of your critics comes from an a honest belief that you do have shown publicly what you claim you have. I also suspect that you have that belief given that you know things which we don’t know, and therefore *to you* the very thin and vague clues that are actually shown in those two papers look like a huge smoking gun — which is not the case for a majority of the readers.
I also accept the fact that marketing might be putting a huge pressure on what you can say (or even “double wink” here), as that department does on all of your competitors — which is something that I hate, but I don’t have a solution for.
As such, I think it’s a moot point to continue this discussion: MS has convinced who could be convinced, and if we continue MS will make the unconvinced (like myself) even more hostile than the current hostile ones. The APS is not that far, so we can give them the benefit of the doubts until then (even though we don’t agree with the process they followed)
Comment #97 February 23rd, 2025 at 7:43 pm
Del #96: The view that Station Q should be condemned as frauds if they can’t immediately prove to the world that they have a topological qubit, strikes me as a weird mirror-image of the view that their claim should just be immediately accepted on their say-so. What both views have in common is impatience. The scientific process is really, really good at sorting out exactly this sort of dispute, but it does so on the timescale of months, not at Internet speed. What it’s possible to figure out quickly is just that we are in this sort of situation.
Comment #98 February 23rd, 2025 at 8:13 pm
Scott #97: It seems odd to characterize the Station Q position as ‘can’t immediately prove’ when they are explicitly saying ‘we have already proved’. This feels similar to the abc conjecture situation in some sense. Except here, in addition to Mochizuki [MS] saying ‘yes, I’ve solved this problem,’ he’s also saying ‘oh and by the way Scholze and Stix [Nature] agree with me’. If someone makes an unproven claim and then also mischaracterizes the point of view of those who disagree with the claim…perhaps they are a fraud in some sense. No?
Comment #99 February 23rd, 2025 at 11:17 pm
I’d like to hear a response to Philip Reinhold’s #73 point: who cares whether or not the qubit is “topological” if it suffers from errors in the same way that any other qubit does? If this “poisoning” produces errors on the order of 1% that can’t be beat down any faster than errors found in other qubits, then this entire result seems astoundingly uninteresting. Maybe interesting to a condensed matter physicist, but from a quantum computing perspective: a single qubit with 1% error was interesting maybe 15 years ago.
Can anyone from Microsoft comment on this? Do you have any reason to believe that your errors are going to decrease fast enough to overcome a 2 decade gap in development between you and ions/SC qubits?
Comment #100 February 24th, 2025 at 3:23 am
DedicatedStewardOfTheScientificReviewProcess #93
It’s hard to see how insulting people would be helpful. It seems that anyone who disagrees with you is labeled a Microsoft lackey. I’m not sure if name calling is what you meant by labeling yourself a ‘DedicatedSteward.’
From our perspective the paper was published in Nature after undergoing their peer-review process.
1. The published paper contains several statements similar to this one: ‘This effect completely washes out the flux-dependent bi-modality unless the coupling between the “hidden” Majorana modes and the visible MZMs is less than 1neV or the hidden Majorana modes are effectively gapped out, as shown in Fig.S4.’
2. The paper clearly states that it is nearly impossible for the observed effect to be anything other than MZMs, and neither the referees nor the editor asked us to remove those statements.
3. A previous peer-reviewed paper (2023 PRB) also showed direct evidence for MZMs, but the Nature paper builds on that and strengthens the evidence considerably.
4. Reasonable people may disagree about our conclusions on our data. We are presenting our understanding, which has been published via the peer-review process. We will also be presenting more at APS.
The facts are: (a) the editors allowed us to clearly state just how difficult it is for the observed effect to be anything other than MZMs, (b) the editors added their editorial comment, and (c) we chose to make that editorial comment and the referee reports public.
Comment #101 February 24th, 2025 at 4:06 am
It’s a testament to the quality and importance of this blog, that we have at least three co-authors of the MS paper engaging in this conversation and giving pertinent answers. This is the way blogs should work.
But can we *please* address the elephant in the room, a question already raised by Philip Reinhold: what is the point of topological protection if coherence times are still limited to 5 microseconds, and we therefore need error correction after all? It’s good that we are now moving from the realm of theoretical considerations to experimental data, but doesn’t this all imply that fully-protected topological qubits need a more elaborate implementation? Can a knowledgeable person here please answer this question?
Comment #102 February 24th, 2025 at 5:48 am
Alex #74:
DARPA does in fact fund quantum science research (as does the US DoD in general). DoD actually funds a huge amount of research in the US, around $8 billion with $70 million specifically for quantum (for reference, the NSF budget is around $10 billion). A number of quantum research programs are listed on DARPA’s website, such as their Quantum Benchmarking Initiative. The program manager of this initiative is a former professor and has published many papers on photonics based quantum information. It’s fair to say they are familiar with academic research on quantum computing, and would know how to solicit expert opinions. And the DoD is not just a funding source. The US Air Force even has its own qubit laboratory.
Of course for those of us on the outside of Microsoft’s NDA, it’s hard to determine exactly what this means Re:Majoranas, but all this is to say that DARPA and the US military do in fact have interest and experience with contemporary quantum research. It is extremely naive to think that quantum computing research would be of no interest to the military.
Comment #103 February 24th, 2025 at 7:47 am
suomynona #102
Thank you very much for your response. Some of the information you provided is indeed new to me.
Comment #104 February 24th, 2025 at 8:09 am
zulfi alam #100:
All due respect, half of the referees (2/4) recommended that the paper NOT be published. This is not indicative of an endorsement of the paper and its written contents. Asking the authors to specifically remove those statements would be akin to rearranging deck chairs on the Titanic.
Comment #105 February 24th, 2025 at 9:01 am
I second the questions of Ostrogradsky #101 and Philip Reinhold #73, and I reiterate my question in comment #20 about exactly what benefit (if any) topological protection provides for non-Clifford gate operations.
Comment #106 February 24th, 2025 at 9:18 am
In keeping with Scott’s wisdom that when you see someone making a mistake it is wise to always offer a way out I’d like to extend the following:
If the MS authors will acknowledge that they have not received “peer reviewed *confirmation*” inline with the editors notes and refereed comments listed above I will try and have the patience that Scott recommends and keep an open mind that you will publish such peer reviewed confirmation in the future.
However, if you continue to insist that this has already been accomplished which is not in keeping with the updated comment from your co-author, then I’ll _still_ offer the way out for you to name a single **peer** who is independent of Station Q and who has publically and full-throatedly endorsed the claims you are making in these comments.
OTOH, if you can’t do either of these I’ll be forced to think you’re lying and will not admit mistake when caught in a bald faced one.
In short, we’re all human and I can see you’re proud of your work. But claiming peer reviewed confirmation is really a bright line. Just name an independent peer; an actual individual. Simple as that.
Comment #107 February 24th, 2025 at 9:42 am
It seems the core of the disagreement surrounds these lines in the conclusion of the Nature paper: “These measurements do not, by themselves, determine whether the low-energy states detected by interferometry are topological. However, our data tightly constrain the allowable energy splittings in models of trivial Andreev states.”
The MS folks are basically claiming “come on, these are MZMs.” The editors and referees are claiming “NO, this is NOT conclusive evidence for MZMs.” Could an expert weigh in on this? Would a reasonable scientist think these are probably MZMs? In analogy I’m asking for 2-sigma “likelihood” rather than 5-sigma “assurance”. Or are the concerns deeper than adding a few sigma onto the certainty?
Comment #108 February 24th, 2025 at 9:52 am
Anyone working in a publicly traded corporation knows it’s the time of the year for the “performance review”, determining new salary/bonus/internal project funding… obviously the only “peer review” that really matters! :_D
Comment #109 February 24th, 2025 at 10:09 am
zufi alam #100
I have no expertise or vested interest, but your responses in this thread have been incredibly frustrating for me to read. As a scientist (if indeed you are one) would encourage you examine the ethics of your behavior.
Indeed, the facts are:
(1) The peer review process (as documented by the opinions of the reviewers, the editorial board, and the broader scientific community) has not yet confirmed the claims of MZM’s.
(2) The press release claims that it has been confirmed
(3) You believe that there is sufficient evidence for MZM’s.
(4) Additional evidence is forthcoming.
Comment #110 February 24th, 2025 at 10:22 am
Hi Scott, thanks for opening this discussion here!
I have a few questions regarding the Microsoft paper as a grad student not too familiar with the more applied details.
If the Microsoft claims of having a chip with 8 topological qubits and having the ability to perform X and Z based measurements are really true, what other capabilities would they need to perform a measurement-based braid?
Would demonstrating such a braid conclusively exclude any other possibilities (such as the trivial Andreev states) causing the observations by the Microsoft team, other than MZMs?
Are the errors in the measurements induced by above-gap quasiparticle poisoning inherent to the measurement or the system itself? E.g. if the braiding was implemented through transport of the MZMs instead of the measurements, would these errors still be present?
Comment #111 February 24th, 2025 at 6:45 pm
Adam Treat #106
How about 5 Peers ….
1. Abhisek Khole, a physicist in Julich (LinkedIn): “Despite all the skepticism and criticism, I believe the Microsoft team’s work is truly remarkable. I genuinely think they have succeeded in measuring the parity flipping of Majorana fermions… Each figure in the main paper alone could be standalone papers, given the sheer amount of physics involved.”
2. Jay Sau, Professor at U. of Maryland (Press): “Majorana 1 is a significant achievement because it shows evidence for a significant level of coherence from a topological qubit.”
3. Steven Simon, Oxford theoretical physicist (Press): “Would I bet my life that they’re seeing what they think they’re seeing? No, but it looks pretty good.”
4. Prineha Narang, Professor at UCLA (LinkedIn): “Congratulations to the Microsoft Station Q team! Chetan Nayak gave a compelling talk yesterday discussing the breakthrough. Impressive to see the science, engineering, and relentless innovation that made it happen.”
5. Nature referee reports, referee #2: “This is still the highest bar reached so far in terms of tests for Majorana physics and, as the authors state, places strong constraints on alternative hypotheses.”
I hope this list suits your needs!
Comment #112 February 24th, 2025 at 7:15 pm
Sankar Das Sarma, at University of Maryland, asked me to post the following comment from him:
I have been following the comments in this discussion. As someone who has been involved actively in MZM research since 1999, I can say with some authority that the response from the MSFT people has been reasonable and sound. The experimental physics here is not binary, such as ‘yes here are the MZMs and the topological qubit’ and ‘no, there are no MZMs and no topological qubit’. It is much more nuanced. It is a process with many steps. It is engineering. For example, ‘zero energy’ immediately leads to the question: ‘how close to zero energy is it correct to call the object a Majorana zero mode?’ There is no yes/no unique answer to this question as it depends crucially on all the engineering details. In real devices, nothing can be zero energy! Microsoft has made a disruptive advance, and we should all be grateful to them for putting their considerable resources to an abstract physics/math/CS topic creating the foundation for future TQC. I simply do not understand the vehement negativity expressed by some in these comments. This is not a simple problem where a binary yes/no paradigm applies. I know as I have been at it for more than 25 years.
Comment #113 February 24th, 2025 at 7:21 pm
Zulfi, the first reference sufficed. Given that along with Scott’s last post I’ll grant that you have received some form of peer reviewed confirmation even if it wasn’t Nature. Mea culpa.
Comment #114 February 25th, 2025 at 2:35 am
zulfi Alam #111
I’m not an expert in condensed matter / Majoranas; I’m just trying to follow the conversation here and elsewhere. But I feel compelled to jump in because I think some additional quotes and context should be added to the list of enthusiastic peers provided by Zulfi.
1. As far as I can tell from this article and his LinkedIn profile, Abhisek Kole is still a PhD student.
2. The same article where Jay Sau’s quote is taken from also quotes him saying the following:
But there are additional quotes from him elsewhere that are more damning:
3. There’s another quote from Steven Simon in the same article:
Comment #115 February 25th, 2025 at 8:15 am
Ostrogradsky #101
Of course these results are just the first step. The team is working on better device designs, larger topological gap, lower disorder. Based on simulations and ongoing materials and fab developments there is a clear path to better performance.
Comment #116 February 25th, 2025 at 2:15 pm
Matthias #115
May I ask how large a topological gap would be needed to achieve a significant advantage over, say, existing superconducting qubits, and what gap can we realistically achieve in InAs/Al nanowires (or in other manufacturable nanowires)?
Comment #117 February 25th, 2025 at 8:23 pm
Matthias Troyer #115
This could just as well be said for any of the dozens of quantum efforts in the world. The field is a sea of “promising” qubits that need a little more work, a little less error. Do you believe your qubit is qualitatively different in this regard? Setting aside the academic value of having achieved a MZM, do you have good reason to believe your qubit will significantly outpace the decades-advanced progress of other well-established qubits, or are you going to get stuck in the same exponential-effort-for-logarithmic-gain grind as everyone else?
Comment #118 February 25th, 2025 at 8:50 pm
@Ostrogradsky #116
Our roadmap paper https://arxiv.org/abs/2502.12252 provides the criteria for the required gap and how to calculate it. We see these target as achievable.
Comment #119 February 26th, 2025 at 2:44 pm
Matthias # 118,
Thank you for the link to the article. If I understand correctly, the measured topological gaps (20-60 ueV) are indeed large enough for the “qubit lifetime” to reach ~10ms. On the other hand, Chetan Nayak #31 mentioned that MZMs at opposite ends of the wire are likely interacting and limiting coherence time to only 5 microseconds, which implies that the wires should be longer, assuming a fixed MZM localization length. The requirement for longer wires is in turn complicated by the fact that the experiment is designed is such a way that the wire lengths needs to be close to (or shorter than) a certain coherence length, as explained in the PRB paper. Could you possibly comment on the path to suppressing this type of error resulting from the MZM interaction? This seems to be a fundamental limiter in the current device architecture.
Comment #120 March 2nd, 2025 at 1:30 am
As an attendee of the upcoming APS March Meeting, I am eager to see the data. I am excited about the possibility of a useful, fault-tolerant quantum computer and want Microsoft to succeed. However, even if there is conclusive T1/T2 data from one device, without multiple devices showing a scaling improvement that could push these coherences to fault-tolerant levels, I will remain unconvinced that this architecture provides any benefit over others.
Additionally, I feel as if it would be helpful to note in this thread that there was a recent criticism of the 2023 PRB paper posted on arxiv: https://arxiv.org/abs/2502.19560. I have copied the abstract below.
As the above pre-print has multiple, distinct criticisms, I would love to hear rebuttals from the various Microsoft-affiliated commenters watching this blog. Another paper retraction would not be ideal.
Comment #121 March 2nd, 2025 at 4:52 am
There is a 1-hour video by Mourik and Frolov, essentially saying that the PRB’23 paper is flawed (to the extent of doubting the superconductance of the device); the evidence in the new Nature paper is unconvincing; the software to test the presence of Majorana is not doing the right thing and has been trained on an invalid mix of synthetic and measured data. They also claim that the whole concept of anyon-based computing is not great, with only Clifford operations being topologically protected and, moreover, the superconducting-regime operations being subject to essentially the same noise mechanisms as transmons, thus impossible to be fundamentally better than today’s transmon-based systems that have hundreds of qubits already. https://youtu.be/9Ag-L3hZiXo?si=2DP9S26NjC71NUJo
There is also a correspondence with PRX published by Mourik on a request by PRX whether to send a version to peer review: https://pubpeer.com/publications/4C7F25D6DFE392856EF4171E379D67
(I am myself a stupid Computer Scientist and could not explain why there must be two Majoranas and whether the plot we see tell us that the materials used within the device are well- or poorly-manufactured; just saw those and thought they are useful.)
Comment #122 March 7th, 2025 at 10:13 am
[…] announcement. But the company has subsequently indicated to Nature’s news team, and in a comment online, that it created the topological qubits using the TGP. “Since the TGP is flawed, the very […]
Comment #123 March 18th, 2025 at 8:11 pm
Vincent Mourik thinks it’s all flawed, *especially* after hearing today’s talk and additional data from M$:
https://www.linkedin.com/posts/vincent-mourik-8188379_comments-on-microsoft-qubit-claims-aps-mm-activity-7307793712217030658-BN4M/
tl;dr;
> In summary, it is impossible from a physics point of view that
> Microsoft has demonstrated a topological qubit.
Comment #124 March 21st, 2025 at 8:03 pm
Hey, Scott!
Thank you for the drill down into the Microsoft paper.
I wonder if you got a chance to see another paper from the quantum space, a paper published in Science regarding Quantum Annealing:
Andrew D. King et al. ,Beyond-classical computation in quantum simulation.Science0,eado6285DOI:10.1126/science.ado6285
I would love to hear your opinion about it.
Thanks in advance!
Yours,
A fan
Comment #125 March 23rd, 2025 at 11:21 pm
[…] exposed by Microsoft at the Terminal Q conference reveals their gadgets stay error-free for 5 microseconds, however they think this can be enhanced. (For contrast, a typical superconducting qubit in […]
Comment #126 March 24th, 2025 at 2:57 pm
Is there any update on this now that it has presented at March Meeting? I wasn’t in attendance but this sentiment doesn’t seem to be that the Station Q team unambiguously demonstrated a qubit https://www.nature.com/articles/d41586-025-00829-2