Japan's Hitachi/Infleqtion/Institute for Molecular Science consortium turned on Shunkai, a 50 qubit room temperature quantum computer that uses neutral atoms held and steered by lasers rather than superconducting chips, with a 2031 fault tolerance
Japan's new quantum computer runs at room temperature on neutral atoms that are physically moved with laser beams during the calculation. Shunkai, a 50-qubit machine built by a Hitachi/Infleqtion/IMS consortium under the Japanese government's Moonshot R&D Program, is an architectural bet, not a count war.
The system went live on 24 August at the Institute for Molecular Science (IMS), part of Japan's National Institutes of Natural Sciences (NINS), under Goal 6 of the Cabinet Office's Moonshot Research and Development Program, a goal-oriented research effort where Goal 6 specifically targets a fault-tolerant quantum computer. The project is led by Professor Kenji Ohmori. The consortium pairs Hitachi, which built the software stack and compiler, with Infleqtion, which built the quantum processing unit (QPU) hardware. Together, they have produced what the program calls Japan's first "full-stack" neutral-atom quantum computer, meaning the qubits, the control electronics, the compiler, and the orchestration all come from a single delivery.
The architectural difference is real. Where most quantum hardware in the news runs superconducting qubits near absolute zero inside cryogenic dilution refrigerators, Shunkai uses neutral rubidium atoms: single atoms held in 2D optical-tweezer arrays, the tightly focused laser beams that trap and move individual atoms. Microwave and laser pulses drive the qubits at room temperature. Readout happens by fluorescence, the same family of imaging used in molecular biology. There is no cryogenic fridge in the loop, and no etched superconducting chip to manufacture.
That difference reshapes what scaling has to solve. Because the atoms are physically movable, the system can run what the consortium calls "dynamic atom transport," physically shuttling qubits to other qubits during a calculation, which gives the machine all-to-all connectivity and lets the circuit topology be reconfigured on the fly. Static superconducting machines route signals along fixed wires; Shunkai's wiring is whatever the algorithm needs it to be. For algorithms that need dense qubit interaction, including quantum chemistry, certain optimization problems, and parts of the error-correction stack, that flexibility is the whole point.
The trade is qubit count. Shunkai's published roadmap is staged: Phase 1 lands at roughly 50 physical qubits, where it sits today; Phase 2 reaches 500; the March 2031 target is 10,000 fault-tolerant qubits. That is a six-year climb from a small base, and the consortium is asking whether the neutral-atom modality and dynamic transport, not raw qubit count, will carry the system to fault tolerance. "Fault-tolerant" means the system can keep computing accurately as its individual components error out, the standard bar for a "useful" quantum computer. Stage 2 of the Ohmori Moonshot Project runs through that date, and partial cloud access to academic and external users is expected along the way.
For now, the system is announcement-stage hardware. There are no published gate-fidelity benchmarks, no external-access service-level agreement, and no third-party validation of the architecture. The bet, instead, is architectural: neutral-atom hardware and dynamic transport, with Hitachi doing the orchestration and Infleqtion doing the atoms, can scale to a fault-tolerant machine on a six-year horizon.
The name carries the older bet with it. "Shunkai" (春海) honors Harumi (Shunkai) Shibukawa, the Edo-period astronomer who designed Japan's first indigenous calendar from celestial observations, a fitting namesake for a national-caliber instrument.
The watch items are concrete. The 50-qubit system needs a credible fidelity benchmark before the 500-qubit Phase 2 is a story rather than a slide. Infleqtion's neutral-atom hardware has to ship in quantity, on schedule, inside a Japanese national program. The March 2031 target is a commitment; the current milestone is operational. The next one is whether the architecture's room-temperature, dynamic-transport claim survives contact with benchmarks.