A new paper argues Microsoft exaggerated its quantum claims a year ago
A new paper argues Microsoft exaggerated its quantum claims a year ago
一篇新论文指出微软一年前夸大了其量子计算成果
A peer-reviewed critique of Majorana 1 claims Microsoft did not conclusively demonstrate a working topological qubit. Microsoft disagrees. 一篇针对 Majorana 1 的同行评审批评文章指出，微软并未确凿地证明其制造出了可工作的拓扑量子比特。微软对此表示异议。
A critique published in Nature Wednesday calls the basic technology behind Microsoft’s “breakthrough” quantum computing chip, the Majorana 1, into question. Microsoft unveiled the chip in February 2025 and said it featured a brand-new technology known as a topological qubit. Topological qubits, they said, would be the “building blocks” for their future quantum computer. Microsoft announced the next generation chip Majorana 2 at Build earlier this month. 周三发表在《自然》（Nature）杂志上的一篇评论文章，对微软“突破性”量子计算芯片 Majorana 1 背后的基础技术提出了质疑。微软于 2025 年 2 月发布了该芯片，并称其采用了名为“拓扑量子比特”的全新技术。他们表示，拓扑量子比特将成为未来量子计算机的“基石”。本月初，微软在 Build 大会上发布了下一代芯片 Majorana 2。
But in a peer-reviewed article, Henry Legg, a physicist at the University of St Andrews, reanalyzed Microsoft’s data on their device and argued that the company’s researchers did not conclusively demonstrate a working topological qubit in the first place. 然而，圣安德鲁斯大学的物理学家亨利·莱格（Henry Legg）在一篇经过同行评审的文章中，重新分析了微软关于该设备的数据，并指出微软的研究人员从一开始就没有确凿地证明其制造出了可工作的拓扑量子比特。
Microsoft’s design, unique among quantum computing companies, involves a tiny wire, thinner than a human hair, made of the semiconductor indium arsenide stuck to a superconductor. Theory predicts that the electrons in this wire behave in a collective pattern known as a Majorana particle, for which the chip is named. Microsoft wants to encode information in the properties of the Majorana particle. (A topological qubit is to a Majorana particle as a transistor is to silicon.) 微软的设计在量子计算公司中独树一帜：它使用一根比头发丝还细的砷化铟半导体导线，并将其贴合在超导体上。理论预测，导线中的电子会以一种被称为“马约拉纳粒子”（Majorana particle）的集体模式运动，该芯片也因此得名。微软希望利用马约拉纳粒子的特性来编码信息。（拓扑量子比特之于马约拉纳粒子，正如晶体管之于硅。）
Proponents of the Majorana particle think it is promising qubit material because theory predicts that when formed into topological qubits, the Majorana should compute with fewer errors than competing materials, such as superconducting circuits pursued by IBM. This suggests that ultimately, fewer topological qubits are needed to scale up to a useful quantum computer. 马约拉纳粒子的支持者认为它是一种很有前景的量子比特材料，因为理论预测，当其形成拓扑量子比特时，其计算错误率应低于竞争材料（如 IBM 所采用的超导电路）。这意味着，最终只需较少数量的拓扑量子比特即可扩展成实用的量子计算机。
That is, if Microsoft has actually made a Majorana particle. “They haven’t convincingly shown that they have Majoranas,” Legg told The Verge. “You can’t make a qubit if you don’t have the Majoranas.” 前提是，微软确实制造出了马约拉纳粒子。“他们并没有令人信服地证明他们拥有马约拉纳粒子，”莱格告诉《The Verge》，“如果你没有马约拉纳粒子，就不可能制造出量子比特。”
In Legg’s critique, he writes that what Microsoft claims as a signature of the Majorana particle could actually be from the formation of quantum dots, which are electron-containing structures, in the device. Quantum dots would not be useful for building the quantum computer. He also writes that Microsoft cherry-picked their data. 在莱格的批评文章中，他写道，微软声称的马约拉纳粒子特征，实际上可能源于设备中形成的“量子点”（一种包含电子的结构）。量子点对于构建量子计算机并无用处。他还指出，微软在数据选择上存在“樱桃采摘”（即只挑选支持自己结论的数据）的行为。
Microsoft’s team published a rebuttal in Nature disputing Legg’s interpretation of their data. Legg’s critique “does not constitute a substantial scientific challenge to our findings,” the Microsoft team wrote. Legg has not “proposed an alternative model that fits all of our data,” Chetan Nayak, a physicist leading Microsoft’s quantum team, told The Verge. 微软团队在《自然》杂志上发表了反驳文章，质疑莱格对其数据的解读。微软团队写道，莱格的批评“并未对我们的研究结果构成实质性的科学挑战”。领导微软量子团队的物理学家切坦·纳亚克（Chetan Nayak）告诉《The Verge》，莱格并没有“提出一个能解释我们所有数据的替代模型”。
Legg first posted his critique on the online physics repository arXiv on March 11, 2025, within a month of Microsoft’s Majorana 1 announcement. It took a year for Nature to conduct a peer review and publish his article. 莱格于 2025 年 3 月 11 日（即微软发布 Majorana 1 不到一个月后）首次将他的批评文章发布在在线物理学预印本平台 arXiv 上。《自然》杂志花费了一年时间进行同行评审并发表了他的文章。
Meanwhile, on June 2, Microsoft announced a new chip, the Majorana 2, featuring what they claimed was the next generation of their topological qubits. The company says they can build a “scalable quantum computer” by 2029. “We 100% stand behind our results,” Nayak told The Verge. “We stand by our roadmap. We stand behind our long-standing commitment to scientific rigor and dialogue.” 与此同时，微软在 6 月 2 日发布了新芯片 Majorana 2，并声称其采用了下一代拓扑量子比特。该公司表示，他们可以在 2029 年前构建出“可扩展的量子计算机”。“我们 100% 支持我们的研究结果，”纳亚克告诉《The Verge》，“我们坚持我们的路线图，并坚持我们对科学严谨性和学术对话的长期承诺。”
Legg says the company’s characterization of Majorana 2, which Microsoft wrote in a non-peer reviewed manuscript, suffers from similar problems he pointed out a year ago. “Nothing in this [manuscript] resolves the fundamental issues that so many scientists have with this company’s previous claims,” Legg told The Verge. 莱格表示，微软在一份未经同行评审的手稿中对 Majorana 2 的描述，存在他一年前指出的类似问题。“这份（手稿）中的任何内容都无法解决许多科学家对该公司此前声明所持有的根本性质疑，”莱格告诉《The Verge》。