China's Quantum Startup Just Broke a World Record — and Raised Hundreds of Millions
A Tsinghua University spinout called Qosmos Technology has trapped 11,000 atoms — shattering the previous record of 6,100 set at Caltech. Now it's raising hundreds of millions to scale up, in a move that signals how fast China is closing the gap in the quantum race.
The number that should unsettle American quantum leaders
Qosmos Technology just trapped 11,000 atoms and held them steady. The previous record — set at Caltech — was 6,100. By a wide margin. This isn’t a marginal improvement; it’s a leap that reshapes the competitive map for neutral-atom quantum computing.
What makes this story more than a lab bench achievement is the company behind it. Qosmos spun out of Tsinghua University less than two years ago. Its founding team traces back to a cold-atom research group that started work around 2000. The company’s chairman, Shi Qi, put it bluntly: “We want to compete shoulder to shoulder with the top teams at Caltech, Harvard, and MIT.”
The timing of their funding round — hundreds of millions of yuan raised in a Series A extension — suggests the Chinese investment community is already betting that the gap between Chinese and American quantum labs is narrowing faster than most Western observers realized.
How the record actually matters
Neutral-atom quantum computing works by trapping individual atoms with lasers and using them as qubits. The harder problem isn’t catching atoms — it’s catching many atoms, keeping them stable long enough to perform computations, and wiring them together with gate operations that don’t fall apart.
Caltech’s 6,100-atom record was itself a milestone. Qosmos’ 11,000 doesn’t just beat it; it pushes into a regime where error correction and multi-qubit gate fidelity become the real bottleneck, not raw atom count. That’s where the company’s second generation of technology comes in: a self-developed optical tweezer platform that dynamically reconfigures rubidium atoms into programmable arrays of 800 to 1,000 qubits with no missing atoms.
Shi Qi was explicit about what separates the serious contenders from the rest: “You can’t build a competitive advantage by breaking through in just one technology. Scale, precision, and error-correction capability all have to be there.”
That framework is useful for reading the field. It also helps explain why Qosmos’ approach — combining Tsinghua’s theoretical foundation with its own engineering team — might actually land rather than dissolve into another lab demo that never ships.
The money trail
Legend Capital led the latest funding round, with Costone Capital participating. Existing investors Shunwei Capital and AI company iFlytek added more. The exact amount was reported as “several hundred million yuan” — roughly tens of billions of yen. That’s substantial for a company that’s barely two years old.
But the investors tell you more than the amount. iFlytek, one of China’s largest AI firms, isn’t a passive check-writer here. In March 2026, Qosmos and iFlytek spun up a joint venture called Intelligent Quantum Inception, focused on merging quantum computing with AI. The subsidiary Tian Ce Liang Wu will handle measurement and control systems and a quantum cloud platform.
This is the pattern: deep-tech startups in China are increasingly bundled with AI majors, defense suppliers, and state-linked universities. The lines between “startup,” “research institute,” and “strategic partner” are blurring — and that’s intentional.
What Western press has missed
American coverage of China’s quantum push tends to focus on superconducting qubits and the familiar names — Google, IBM, Rigetti,PsiQuantum. Neutral atoms get a mention when a paper drops in Nature. The Chinese ecosystem moves in silence. Tsinghua’s cold-atom group has been publishing quietly for decades. The spinout culture is newer, but the research base isn’t.
What the English-language media hasn’t fully tracked is the velocity of translation from lab to product in China’s quantum sector. Qosmos developed nearly 10 components in its first two years — feedback systems, atom sources, optical metasurfaces — and is already selling some of them to universities and research institutes. That’s not a prototype company. That’s a supply-chain play.
The university-to-company model is also worth watching. Tsinghua handles the fundamental research; Qosmos handles commercialization. They co-ran the Beijing Key Laboratory for Cold Atom Quantum Computing. It’s a structured pipeline, not an ad hoc partnership. Shi Qi described it precisely: theory validation at the lab, engineering at the company.
Who wins, who loses
If Qosmos scales this to meaningful qubit counts with error correction, the winners are clear: Chinese quantum hardware, Chinese cloud-access models, and the Chinese AI firms that piggyback on them. iFlytek’s involvement suggests the first applications will be AI-adjacent — optimization, simulation, pattern recognition.
The losers are less immediate but real. US neutral-atom teams — MIT, Harvard, Caltech — now face a competitor with state-aligned funding, a massive domestic research base, and a company structure that converts papers into products faster. The advantage wasn’t insurmountable before, but it’s thinner now.
Globally, the implication is that quantum computing is becoming less a US-led race and more a multipolar one, with China’s scale advantage in talent, funding, and integration across sectors creating a different kind of moat than raw patents.
What happens next
The real test isn’t the atom count. It’s whether Qosmos can convert 11,000 trapped atoms into a machine that runs an algorithm better than a classical computer. The company has gate fidelities it calls “world-class” and a reconfigurable array platform. Those are necessary but not sufficient conditions.
Watch for three things: first, whether Qosmos announces a working quantum processor with error-corrected logical qubits; second, whether the iFlytek joint venture ships a usable quantum-AI product; third, whether US or European teams respond with their own neutral-atom advances or shift strategy.
The record is real. The funding is real. The question is whether either of those translates into a quantum computer that changes anything — or just changes the leaderboard.
Shi Qi’s own words capture the stakes: “Even if you achieve breakthroughs in individual technologies, you can’t build a comprehensive competitive advantage without all three elements.”