China's Neutral-Atom Labs Trap 11,000+ Atoms, Lead in Gate Fidelity

Chinese neutral-atom quantum labs have trapped more than 11,000 atoms in a single array and demonstrated two-qubit gate fidelities above 99.8%, according to a research review published on arXiv and reported by Quantum Computing Report. The survey, authored by researchers at Information Engineering University, maps a hardware ecosystem spanning USTC, Tsinghua University and a cluster of commercial startups that has closed much of the physical performance gap with Western neutral-atom leaders like QuEra, Atom Computing and Pasqal.

According to the review, Pan Jianwei's group at USTC used an AI-driven spatial light modulator to assemble defect-free 2D and 3D arrays of 2,024 atoms in 60 milliseconds, hitting a two-qubit gate fidelity of 99.84%, single-qubit fidelity of 99.97%, and readout fidelity of 99.92%. In 2026 the same team reportedly added mid-circuit non-destructive readouts with feedback control, a step toward active error correction cycles.

The review also credits Tsinghua University's Zhai Hui group with trapping an average of 11,022 atoms across an 18,225-site grid using a 2-centimeter metasurface chip placed outside the vacuum chamber, paired with a graph neural network rearrangement algorithm ('Zhui Feng') that sorts 10,000-atom arrays in under 20 milliseconds. Fudan University and the Innovation Academy for Precision Measurement Science and Technology in Wuhan are named as additional academic hubs working on reduced circuit depths and fiber-based atom addressing.

On the commercial side, Quantum Computing Report details Zhongke Kuyuan's Hanyuan 2, a dual-cavity system supporting more than 500 optical tweezers, coherence times above 300 milliseconds, and over 80% domestic supply-chain localization, running inside standard server racks under 7 kilowatts. Shanghai-based Taiyi Liangsheng, pursuing ytterbium-171 qubits, closed more than ¥400 million (~$56 million USD) in angel and Pre-A funding in 2026, while Beijing's Liangyi Wanxiang has demonstrated a 500-plus atom platform with gate fidelities above 99.5% and is targeting a 1,000-qubit fault-tolerant system by 2027.

What Happened

The review's central finding is a split verdict: China has caught up or pulled ahead on raw hardware numbers, array sizes, rearrangement speed, gate fidelity, but still trails leading US programs on system-level quantum error correction. The paper points to QuEra and Harvard's transversal logical gates across 48 logical qubits, Atom Computing's multi-round syndrome extraction on toric codes, and Infleqtion's precompiled logical Shor's algorithm work as benchmarks China has not yet matched, noting the country's labs are only now moving from mid-circuit measurement primitives toward active, multi-round error correction cycles.

Why It Matters

This review is notable because it treats hardware scaling and commercialization as parallel tracks rather than sequential ones. Zhongke Kuyuan shipping a rack-mounted, under-7kW commercial system in parallel with academic world records suggests China's neutral-atom sector is organizing itself less like a research pipeline and more like a manufacturing supply chain, with over 80% domestic component localization already claimed. That pattern echoes workforce investments elsewhere in the sector, including New Mexico's recent Job Training Incentive Program awards to IonQ and Bandelier Technologies and California's $5.74 million in Workforce Accelerator Fund grants, of which more than $4.26 million specifically targets sub-doctoral quantum hardware technician pipelines. Both signal that governments view qubit scaling as a staffing and supply-chain problem as much as a physics one. For readers tracking the commercial side of this race, the investment implications of quantum computing advances extend well beyond any single qubit modality, touching funding flows into startups like Taiyi Liangsheng and the broader capital formation around hardware-stage quantum companies.

Background

Hardware milestones like atom-array scaling matter partly because they feed into timelines for when quantum computers could threaten current encryption standards, a concern addressed in MachineHours' Post-Quantum Cryptography entry. The further neutral-atom platforms push toward fault-tolerant logical qubits, the more urgent that migration becomes for organizations still relying on RSA and elliptic curve cryptography.

What to Watch

The review notes Liangyi Wanxiang is targeting a 1,000-qubit fault-tolerant platform by 2027, while Zhongke Kuyuan is still building its 1,000-qubit Hanyuan 3 and Taiyi Liangsheng is pursuing logical qubit demonstrations, with the overall roadmap heading into 2027 without specific dates confirmed for the latter two. Whether Chinese labs can translate their lead in raw gate fidelity and atom counts into multi-round error correction, the area where QuEra, Atom Computing and Infleqtion currently hold an edge, will determine how quickly the physical-to-logical gap closes. Parallel workforce programs in New Mexico and California suggest the US is also betting on technician-level hiring, not just research funding, to keep pace on the hardware side.