A six month old Shanghai startup has raised about $44 million to build a quantum computer that uses atoms held in place by lasers — a “neutral atom” approach — using ytterbium instead of the more common rubidium, backed by an unusual coalition including SAIC's venture arm, a…
A six-month-old Shanghai startup has pulled roughly $44 million from an unusual coalition of Chinese industrial investors, betting that a rarely used rare-earth element, ytterbium, is the right material to build a commercial quantum computer.
Taiyi Quantum's oversubscribed Pre-A round, formally 300 million yuan, closed this month and was co-led by Gaorong Ventures and IDG Capital. The more telling number is who joined them. SAIC Financial and Hengxu Capital, the venture arms of state-owned automaker SAIC Motor, signed on, alongside a fund tied to CETC, the defense electronics conglomerate. iFlytek, the AI speech company, came in through its venture arm. Jinko Solar, the photovoltaic manufacturer, invested too, as did Galbot, the humanoid-robot maker, and Boyuan Capital, the industrial arm associated with Bosch's China operations. The list reads less like a venture capital syndicate than an industrial capability survey.
For a startup that incorporated in January 2026 and has fewer than 50 employees, the round is a statement of intent. Taiyi is committing to a particular technical path within quantum computing, a path most of its Chinese peers have not chosen. The field has consolidated around three hardware approaches: superconducting circuits (the technology behind IBM's and Google's machines), trapped ions (IonQ's specialty), and neutral atoms, where individual atoms are held in place by laser beams and used as quantum bits, or qubits. Inside the neutral-atom camp, most teams work with rubidium, an alkali metal that is cheap, well understood, and the default in academic labs. Taiyi is betting on ytterbium-171, a different kind of atom with an extra electron shell.
That extra electron shell is the whole point. Ytterbium belongs to a class of atoms called alkaline-earth metals, and its electronic structure lets it occupy two internal states cleanly. The result, in the thesis Taiyi is funding, is that when an atom is lost from the array, a routine failure in any neutral-atom machine, the error is physically detectable. Physicists call this an "erasure channel": errors announce themselves, which makes the error-correction machinery simpler and less qubit-hungry than in a system where every error has to be inferred.
The investor list points to industrial positioning rather than venture-style return chasing. SAIC's venture arm has a manufacturing and simulation use case in view. CETC, the defense electronics group, has a strategic-capability mandate. iFlytek's venture arm is hedging across AI hardware modalities. Jinko Solar's interest is plausibly tied to materials simulation for next-generation photovoltaics.
The bet is large relative to Taiyi's age, small relative to the incumbents. The startup has a roughly 50-person team in Shanghai's Xuhui district, a roughly 1,000 square meter cleanroom rated to class 10,000, and two optical-tweezer systems already built. Its claimed roadmap is aggressive: logical qubits by the end of 2026, and a single platform with a 50,000-atom array that the company says will convert to 300 logical qubits. Logical qubits are the error-corrected units that would actually be useful for commercial work; a 300-logical-qubit machine is not yet a cryptographically meaningful computer, but it would be a competitive neutral-atom milestone. The cryptography use case, RSA-2048 factoring, would require a far larger machine and is not on this timeline.
Taiyi is also a deliberate outlier in the Chinese landscape. Most domestic quantum hardware companies have built around superconducting circuits (the lineage that includes Origin Quantum in Hefei) or trapped ions (Turing Quantum in Shanghai). The notable domestic neutral-atom work sits inside the USTC / Jiuzhang group in Hefei, which is closer to academic research than commercial product. Taiyi is positioning itself as China's first dedicated integrated-system startup on the alkaline-earth neutral-atom path.
The cross-industry syndicate reflects a wider shift in how the technology is being financed. Quantum was named the top future industry in China's 15th Five-Year Plan, and Chinese quantum financing roughly quadrupled year on year in 2025, according to industry coverage. Internationally, the consensus that the eventual quantum computer will be superconducting is eroding: Google has brought in Adam Kaufman, a veteran neutral-atom physicist, to lead its hardware effort; IBM has begun publishing neutral-atom work; NVIDIA invested in QuEra, a Harvard spinout using rubidium. The hardware race is no longer one-modal.
Whether the ytterbium path delivers is the open question. Taiyi's lead technical voice is CTO Liu Hongbin, a former Microsoft Chief Quantum Architect who, according to Chinese tech press, led Microsoft's neutral-atom strategic investment and its partnership with Atom Computing, a U.S. company that has built 50-plus logical-qubit systems and made the first commercial sale of a Level-2 logical quantum computer. CEO Fang Zhenghao spent roughly fifteen years as a managing partner at Xiaomiao Langcheng, a frontier-tech investment firm. The two are Fudan University classmates. Taiyi has academic collaborations with Peking University, Fudan, and East China Normal University.
The honest caveats matter. Taiyi is six months old. The company's erasure-conversion thesis is a physics argument, not a shipped product capability. Total funding since founding is roughly $59 million (over 400 million yuan including a 100-million-plus-yuan angel round in March 2026), which is significant for a Chinese hardware startup but a rounding error against the capex IBM and Google are directing at superconducting. The 8-day, 2-billion-yuan investment-intent figure that circulated in Chinese press points to oversubscription, not closed capital. And the cross-industrial investor list is a strategic bet, not a customer pipeline. SAIC is not buying quantum computers from Taiyi; its venture arm is making an option purchase on a modality.
The 300-million-yuan round is large enough to fund Taiyi through the next milestone, which is demonstrating a small logical-qubit array by the end of 2026. If that milestone slips, or if the ytterbium path turns out to need more qubits than the company forecasts, the next round will tell the reader whether the industrial money sticks around.