In June 2026, a research team from Tsinghua University and a Beijing startup called Liangyi Wanxiang published a preprint on arXiv with a result that broke a record held for more than two years. Using a single optical metasurface roughly two centimeters in diameter, the team generated an array of 18,225 optical tweezers and trapped an average of 11,022 individual rubidium atoms across repeated trials, with a standard deviation of about 160 atoms.
The previous record, set by Caltech in 2024, was 6,100 atoms trapped in roughly 12,000 tweezer sites. The Chinese result nearly doubled the number of trapped atoms and pushed the usable qubit resource on a neutral-atom platform past the 10,000 mark for the first time.
Four months later, on October 23, 2026, the Treasury Department pulled the same lever on quantum information technologies, publishing a proposed rule that widens restrictions on U.S. capital flowing into Chinese quantum companies. Officials sold it as a security matter; coverage slotted it into the wider contest over technology.
The two events are connected, though not as the policy debate frames them. The Treasury proposal restricts capital flows. The Tsinghua result is a physics result. One cannot be undone by restricting the other.
What 11,022 Atoms Actually Means
A neutral-atom quantum computer uses lasers to trap individual atoms in an array, then manipulates their quantum states to perform computation. The number of atoms is the number of physical qubits. The more atoms a system can trap and control, the larger the computational space it can address.
The trapping itself is the bottleneck. Optical tweezers are tightly focused laser beams, and generating a large, uniform array of them requires complex optical systems — spatial light modulators, acousto-optic deflectors, and high-numerical-aperture microscope objectives. Those systems have practical limits: finite pixel counts, limited laser power efficiency, and narrow fields of view. Caltech's 6,100-atom result was widely regarded as close to the ceiling of what that approach could deliver.
The Tsinghua team replaced the conventional optical system with a metasurface — a flat, nanofabricated optical element covered with sub-wavelength structures that can shape an incoming laser beam into a precise array of focal spots. The metasurface was placed outside the vacuum chamber, at a working distance of about 1.5 centimeters, eliminating the need for an in-vacuum microscope objective. A single 30-watt laser produced the entire 18,225-tweezer array, with roughly 15 watts of effective trapping power reaching the atoms.
The significance is not just the number. It is the architecture. The metasurface approach decouples the size of the trap array from the complexity of the optical system. Adding more tweezers no longer requires stacking more modulators or building a larger objective. It requires a larger metasurface, which is fabricated with the same lithographic processes used to make semiconductor chips.
That is the part of the result that matters for scaling. A platform that can trap 11,000 atoms with a single optical element has a path to trapping more. The path is manufacturing, not experimental physics.
The Rest of the Ecosystem
The Tsinghua result is the most visible data point, but it is not the only one. A comprehensive review published on arXiv in August 2026 mapped China's neutral-atom ecosystem across academic labs and commercial startups, and the picture is broader than a single record.
The University of Science and Technology of China's Pan Jianwei team has demonstrated 2,024-atom defect-free arrays assembled in 60 milliseconds, using AI-driven parallel rearrangement, with two-qubit gate fidelities of 99.84 percent and single-qubit fidelities of 99.97 percent. That result addresses a different problem than trapping: it is about rearranging a randomly loaded array into a perfect, defect-free geometry fast enough to be useful, and doing so with gate fidelities that are at or near the threshold for error correction.
Fudan University has built a 1,000-atom cavity-coupled Rydberg platform with a circuit depth reduced from O(√n) to O(log₂ n), and a parallel instruction set with a loss rate below 1 percent. The Chinese Academy of Sciences' Wuhan institute has demonstrated dual-isotope trapping and heteronuclear Rydberg CNOT gates, and is developing fiber-end microlens arrays for compact parallel control.
The commercial layer is emerging alongside the academic work. Zhongke Kuyuan delivered its first commercial full-stack system, the Hanyuan 1, in 2025 and launched the dual-cavity Hanyuan 2 in 2026, with more than 500 tweezers, a coherence time above 300 milliseconds, a cabinet consuming under 7 kilowatts, and more than 80 percent domestic supply-chain localization, according to company disclosures. A 1,000-qubit Hanyuan 3 is planned.
Taiyi Liangsheng, a Shanghai startup founded in January 2026, raised 400 million yuan (about $56 million) across angel and pre-A rounds, according to Quantum Computing Report's review, and has since raised additional capital in a Series A, bringing cumulative funding above 1 billion yuan. The company is developing ytterbium nuclear-spin qubits with telecom-band photonic interconnects. Other startups in the sector, including Atomic Matrix in Hangzhou, have raised close to 1 billion yuan in cumulative financing over four rounds in a single year.
The ecosystem is not yet at the stage of fault-tolerant quantum computing. According to Quantum Computing Report's review of the arXiv survey, Western groups have pushed ahead in multi-round active quantum error correction and topological codes, while Chinese teams have achieved leading performance in core physical parameters — trapping, gate fidelity, and rearrangement speed. Those are the parameters that determine whether a platform can scale. They are not the same as a working error-corrected computer.
What the Treasury Proposal Actually Does
The U.S. Treasury's proposed rule would implement the Comprehensive Outbound Investment National Security Act, known as the COINS Act, which was passed as part of the Fiscal Year 2026 National Defense Authorization Act.
The COINS Act authorizes the Treasury Secretary to define "prohibited" and "notifiable" technologies in five sectors: semiconductors and microelectronics, artificial intelligence systems, quantum information technologies, high-performance computing and supercomputing, and hypersonic systems. The proposed rule adds biotechnology to that list, exercising the Secretary's authority to add technologies beyond those specifically enumerated in the statute.
The rule would require U.S. persons to notify Treasury of certain transactions involving covered foreign persons engaged in notifiable technologies, and to refrain from certain transactions involving prohibited technologies. The public comment period runs through December 7, 2026.
The proposal expands a framework that already covered quantum information technologies. Quantum was included in the original outbound investment executive order issued in August 2023, alongside semiconductors and AI. The new rule adds detail — technical parameters that define which quantum technologies are covered, and which transactions require notification or are prohibited.
The policy logic is that American capital should not help a strategic competitor develop technologies that could be used against the United States. The Treasury Secretary's statement framed the rule as a measure to prevent adversaries from using U.S. investment to develop critical technologies that threaten national security.
Why the Two Instruments Operate on Different Planes
The Treasury rule and the Tsinghua result are both about quantum computing. They operate on entirely different planes.
An investment restriction changes the flow of capital. It makes it harder for American investors to fund Chinese quantum startups. It may raise the cost of capital for those companies, slow their hiring, or force them to seek funding from domestic sources.
A physics result changes what a machine can do. It demonstrates that a particular architecture can trap a particular number of atoms with a particular optical element. It does not depend on where the funding came from. It depends on whether the experiment worked.
Taiyi Liangsheng raised its early capital from Chinese investors, according to the Quantum Computing Report review, not from American ones. The Tsinghua experiment was conducted at a Chinese university with Chinese collaborators. The metasurface was fabricated with domestic lithography. The capital restriction, even if it had been in place when these projects began, would not have stopped them.
That is not an argument that investment restrictions are irrelevant. Capital controls can slow an ecosystem by limiting the availability of risk-tolerant funding, by restricting access to certain types of expertise, and by raising the cost of the long, uncertain research cycles that quantum computing requires.
But the restriction targets the funding channel, not the physics. The physics is happening in laboratories that are funded domestically, staffed domestically, and equipped with components that are increasingly sourced domestically. Zhongke Kuyuan's Hanyuan 2 already reports more than 80 percent domestic supply-chain localization.
The investment ban is a policy instrument aimed at a problem that the physics result suggests is not primarily a funding problem.
The Part That Is Hard to Restrict
There is a specific feature of the neutral-atom approach that makes it resistant to the kind of control the Treasury rule represents.
Neutral-atom quantum computing does not require the most advanced lithography. It does not require extreme ultraviolet scanners or the most advanced semiconductor fabrication nodes. The atoms are trapped by lasers. The optical elements are metasurfaces, which can be fabricated with mature photolithography. The vacuum systems, lasers, and control electronics are commercially available.
The technology is not dependent on a single chokepoint. It is dependent on a combination of laser physics, optical engineering, atomic physics, and control software — fields in which Chinese universities and companies have deep expertise. The Tsinghua team's innovation was an optical design, not a manufacturing process. The metasurface was fabricated on a single chip, but the fabrication did not require a leading-edge fab.
That distinguishes neutral atoms from superconducting qubits, which depend on advanced lithography and specialized fabrication facilities, or from semiconductor chips, which depend on extreme ultraviolet lithography tools that are themselves subject to export controls.
The neutral-atom platform is not immune to supply-chain pressure. It still uses commercial lasers, vacuum pumps, and electronics, some of which may be imported. But the core physics — trapping atoms with light — is not gated by a single tool or a single supplier. It is gated by the quality of the optics and the control of the experiment. Those are things a well-funded domestic research ecosystem can build.
What the Numbers Support
The 11,022-atom result is a record. It is also a data point in a broader trend.
In 2024, the record was 6,100 atoms. In 2026, it is 11,022. The number of trapped atoms has nearly doubled in two years. The USTC team has demonstrated 2,024-atom defect-free arrays in 60 milliseconds. Fudan has built a 1,000-atom platform with a logarithmic circuit depth. Commercial systems with 500 tweezers are being delivered. The ecosystem has moved from laboratory demonstrations to deployable hardware, and the hardware is increasingly sourced domestically.
The Treasury proposal is a response to the strategic implications of that trend. It is a legitimate policy instrument, and it may slow the flow of American capital into Chinese quantum startups. But it does not address the physics. It does not un-trap a single atom.
The 11,022 atoms are already trapped. The metasurface is already fabricated. The experiment has already been published. The capital restriction changes the cost of the next experiment. It does not change the result of the last one.
Sources: Quantum Computing Report review of arXiv:2608.05010 (October 25, 2026); Beijing Economic-Technological Development Area announcement on the Tsinghua–Liangyi Wanxiang result (June 8, 2026); Photon Box analysis of the 11,022-atom record (June 3, 2026); U.S. Department of the Treasury press release on the COINS Act proposed rule (October 22–23, 2026); East Asia Brief on the six-sector expansion (October 24, 2026); People's Daily commentary on the outbound investment proposal (October 26, 2026); BlueQat News on Taiyi Liangsheng financing (June 27, 2026); Jiefang Daily on Atomic Matrix financing (August 20, 2026); 36Kr on Taiyi Liangsheng cumulative financing (October 10, 2026).
Disclaimer
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations
The atom-trapping result is an arXiv preprint and has not completed peer review.
Qubit counts measure trapped atoms, not error-corrected logical qubits or useful computation.
Startup funding figures are drawn from Chinese trade press and company disclosures, and may include undisclosed rounds.
The Treasury rule is a proposal with a comment period that runs through December 7, 2026; final terms are not settled.
Reporting standards differ between Chinese and Western outlets, and figures cited here reflect the primary sources as published.
Market size and shipment estimates vary by methodology; where estimates conflict, the source used is named in the text.
The sectors discussed are changing quickly, and some figures may be revised after publication.
Sources
- Quantum Computing Report review of arXiv:2608.05010 (October 25, 2026)
- Beijing Economic-Technological Development Area announcement on the Tsinghua–Liangyi Wanxiang result (June 8, 2026)
- Photon Box analysis of the 11,022-atom record (June 3, 2026)
- U.S. Department of the Treasury press release on the COINS Act proposed rule (October 22–23, 2026)
- East Asia Brief on the six-sector expansion (October 24, 2026)
- People's Daily commentary on the outbound investment proposal (October 26, 2026)
- BlueQat News on Taiyi Liangsheng financing (June 27, 2026)
- Jiefang Daily on Atomic Matrix financing (August 20, 2026)
- 36Kr on Taiyi Liangsheng cumulative financing (October 10, 2026).
The information provided in this article is for general informational and educational purposes only. It does not constitute legal, financial, or professional advice. The author and publisher are not responsible for any actions taken based on the content of this article. Readers should consult qualified professionals for advice specific to their situation. All trademarks and references to third-party products, services, or organizations are the property of their respective owners. The performance data and benchmarks discussed are based on specific research studies and may not generalize to all use cases or environments. As of the publication date, the AI landscape continues to evolve rapidly, and readers should verify current information independently.
Limitations: The atom-trapping result is an arXiv preprint and has not completed peer review.; Qubit counts measure trapped atoms, not error-corrected logical qubits or useful computation.; Startup funding figures are drawn from Chinese trade press and company disclosures, and may include undisclosed rounds.; The Treasury rule is a proposal with a comment period that runs through December 7, 2026; final terms are not settled.; Reporting standards differ between Chinese and Western outlets, and figures cited here reflect the primary sources as published.; Market size and shipment estimates vary by methodology; where estimates conflict, the source used is named in the text.; The sectors discussed are changing quickly, and some figures may be revised after publication.