Japan has started operating its first room-temperature optical quantum computer at RIKEN’s research campus in Wako, near Tokyo. The system uses light instead of chilled superconducting circuits to process quantum information. RIKEN and Nippon Telegraph and Telephone, usually called NTT, developed the machine together. They announced operations on November 14, 2024, and opened the platform for research through cloud access.
The launch marks a technical milestone, but it does not create an all-purpose replacement for conventional computers. Instead, it gives researchers a programmable platform for testing optical quantum methods and potential applications. Its room-temperature design also removes one major obstacle facing many quantum systems. However, demanding optics, control electronics, and error management remain essential.
A milestone within Japan’s quantum program
RIKEN’s Center for Quantum Computing led the research alongside scientists and engineers from NTT. The partners built upon years of work involving squeezed light, entanglement, optical measurements, and high-speed feedback. Earlier experiments demonstrated extremely large cluster states containing many connected optical modes. The operational machine combines those foundations within a remotely usable computing system.
Japan already operated superconducting quantum computers before this launch. RIKEN introduced a domestically developed superconducting machine in March 2023, using 64 qubits and cryogenic equipment. The new platform follows a different physical approach rather than simply adding another superconducting processor. That distinction explains why officials emphasize optical operation and room-temperature capability.
How the optical computer handles information
A conventional bit stores either zero or one. A quantum system can represent combinations called superpositions, while entanglement connects outcomes across separate quantum states. Optical computers carry this information through carefully controlled properties of light. This machine specifically uses continuous variables, rather than assigning every logical unit to one individual photon.
Continuous variables and squeezed light
Continuous-variable computing encodes information in light’s field quadratures, which resemble amplitude and phase coordinates. Engineers create squeezed states that reduce quantum uncertainty along one coordinate while increasing it along another. They then entangle many optical modes within a cluster state. That cluster supplies the resource for measurement-based quantum computation.
The processor measures selected modes and quickly adjusts later measurements according to earlier results. This feed-forward process changes the intended calculation as optical signals move through the apparatus. Time-domain multiplexing lets successive light pulses share optical paths and components. Consequently, designers can handle many modes without constructing a separate physical circuit for each one.
Why room-temperature operation matters
Superconducting qubits usually need dilution refrigerators that hold processors near absolute zero. Those refrigerators consume space, power, and specialized engineering effort. Optical signals can travel and interfere without placing the main processor inside such a refrigerator. RIKEN therefore describes the optical machine as operating at room temperature.
Room-temperature operation does not mean the system needs ordinary office conditions alone. Stable lasers, low-loss fibers, accurate detectors, vibration control, and fast electronics still matter. Some photonic platforms also use cooled detector technologies, depending on their measurement requirements. The important difference concerns the processing architecture and its avoidance of deep cryogenic cooling.
Cloud access creates a shared research platform
The launch connected the optical processor with control software and a cloud-based programming environment. Authorized researchers can submit jobs remotely, then examine outputs without standing beside the hardware. Cloud delivery lets multiple teams test algorithms on the same scarce instrument. It also helps developers compare theoretical models with results from a physical processor.
A general-purpose label refers to programmability across different calculations, not unlimited practical performance. The system aims to support universal operations required for varied quantum algorithms. That ambition separates it from fixed optical experiments designed for one sampling task. Nevertheless, available scale, accuracy, and software determine which programs can run meaningfully today.
Research opportunities and possible applications
Potential studies include quantum simulation, optimization, machine learning, and methods for solving differential equations. These fields contain problems that may map effectively onto continuous-variable operations. Researchers must still prove useful speed, accuracy, or energy advantages over classical alternatives. The launch itself does not establish such an advantage.
Near-term work may focus more heavily on hardware behavior than commercial applications. Users can study noise, loss, compilation, measurement strategies, and hybrid quantum-classical workflows. They can also develop benchmarks comparing optical processors with superconducting, trapped-ion, and classical systems. Such evidence will clarify where photonics offers genuine value.
Technical limits remain substantial
Photons interact weakly, which helps preserve information during transmission. However, weak interaction makes certain gates and entangling operations difficult. Optical loss can erase signals, while imperfect squeezing adds noise to every calculation. Finite detector efficiency and delayed electronic feedback create further constraints.
Large, fault-tolerant quantum computers will require effective error correction. For continuous variables, researchers study bosonic encodings, including grid states known as Gottesman-Kitaev-Preskill states. Producing those non-Gaussian resources reliably remains difficult. The Japanese platform provides a valuable test bed, not a finished fault-tolerant machine.
Current quantum processors also face competition from rapidly improving classical hardware. Classical simulators can reproduce many small or noisy quantum experiments. Researchers therefore need transparent benchmarks, repeated trials, and fair comparisons. Claims about supremacy or immediate industrial transformation would exceed the announced evidence.
Supporting Japan’s wider technology strategy
The project fits Japan’s broader effort to build domestic quantum capability. National programs support hardware, software, communications, sensing, education, and industry partnerships. Japan now hosts superconducting, trapped-ion, neutral-atom, and photonic research efforts. Maintaining several approaches reduces dependence on one unproven technology.
Training a specialized workforce
The machine may also train engineers who need practical experience beyond computer simulations. Students and developers can learn how real noise changes algorithm design. Cloud access broadens participation because every research group cannot build a complex optical processor. Wider access could create shared software tools and common testing standards.
Understanding what “first” means
The word “first” needs a precise boundary. Japan had quantum computers before 2024, and other room-temperature quantum approaches exist worldwide. This milestone concerns Japan’s first operational general-purpose optical quantum computer, according to the organizations behind it. It does not mean every component works without specialized environmental controls.
Different quantum technologies measure qubits, modes, and computational scale in different ways. A direct qubit-count comparison can therefore mislead readers. Continuous-variable machines use optical modes and squeezing levels, while superconducting machines usually report discrete physical qubits. Performance requires broader measures covering fidelity, loss, speed, programmability, and usable circuit depth.
The next stage of development
RIKEN and NTT can now improve the platform using evidence from actual users. Priorities will likely include lower loss, better non-Gaussian resources, faster feedback, and more effective error correction. Software teams must also translate practical problems into operations that the optical hardware performs well. Each improvement needs reproducible testing rather than promotional claims.
Integration presents another challenge and opportunity. Photonic chips could eventually replace some bulky optical assemblies, improving stability and manufacturability. Fiber links could also connect processors, memories, or quantum communication networks. However, scaling an experiment into reliable infrastructure requires consistent components and automated calibration.
Japan’s room-temperature machine therefore represents a beginning, not a completed quantum revolution. It turns longstanding optical research into a programmable national resource. The platform gives scientists a new way to test claims, expose limitations, and refine engineering choices. Its lasting importance will depend on measurable progress in reliability, scale, access, and useful computation.
