Fujitsu Unveils Diamond-Spin Quantum Computer Prototype

Fujitsu has announced a functional quantum computer prototype that stores qubits within atomic flaws in diamonds, operating at minus 271.6 degrees Celsius. This temperature is approximately 1.5 degrees warmer than the superconducting quantum machines currently prevalent. Though seemingly a small difference, in the field of cryogenics, every fraction of a degree above absolute zero significantly impacts cost and engineering complexity.

The announcement, made via press release on September 8, describes it as the “world’s first working prototype of a diamond-spin quantum computer” that incorporates tin-vacancy centers into photonic circuits. The company’s claim of “first” is noteworthy. The underlying physics builds upon years of collaborative research with Delft University of Technology and QuTech, with some findings published in journals such as Physical Review Applied and Nature Communications. However, the prototype itself has not yet been detailed in a scientific paper.

Understanding a Diamond Qubit

Diamond consists of carbon atoms arranged in a highly ordered lattice. By removing a carbon atom and replacing it with another element, a defect is created. These defects are known as color centers because they alter how the crystal absorbs and emits light. Fujitsu’s design utilizes a tin-vacancy center (SnV), where a single tin atom occupies a position between two empty sites.

The critical aspect is that the electron spin trapped within this defect functions as a qubit. It can be placed in a superposition, measured, and entangled with other qubits. Much of the prior diamond-spin research has focused on nitrogen-vacancy (NV) centers, which you may have encountered in quantum-sensing discussions. SnV centers offer greater symmetry, which makes them more resilient to electrical noise from their environment compared to NV centers. Reduced noise leads to more stable qubits, which is why Fujitsu opted for tin.

Controlling such a qubit is a complex process, which requires the coordinated application of light, microwaves, and radio-frequency signals. Fujitsu has developed software that translates standard quantum circuits (the gates typically drawn on a whiteboard) into this intricate sequence of physical pulses. This allows users of Fujitsu’s Hybrid Quantum Computing Platform to operate the system without needing a deep understanding of tin’s behavior at cryogenic temperatures, a level of expertise few possess.

The Challenges of Photonic Integration

The engineering achievements here are genuinely impressive. To measure a qubit, it’s necessary to capture the single photon emitted by the SnV center. Fujitsu developed a photonic integrated circuit, a chip that directs light similarly to how a conventional chip directs electrons, to capture and route these individual photons.

Furthermore, they had to bond diamond to alumina, a material that does not readily adhere to diamond. Subsequently, the process thinned the diamond from several hundred micrometers to several hundred nanometers, about a thousand-fold reduction. Achieving a chip-scale diamond of this thinness without compromising the internal color centers is a significant scientific and engineering feat.

Why is this relevant for scalability? Photons enable the connection of discrete modules. Instead of constructing one massive chip and hoping for success, the approach involves building smaller modules and linking them with light, even across separate cryostats. Fujitsu and Delft previously demonstrated entanglement between remote NV centers in different cryostats, published in Nature Communications in 2026. 

This exemplifies the modular concept: a quantum computer that expands by adding components rather than by cramming everything onto a single, impossibly complex slab.

The Advantage of Warmer Operation, and Cautious Optimism

The 1.5-degree temperature difference is more significant than it appears. Cryogenics represents a substantial cost for superconducting machines, and any qubit capable of operating in a warmer environment alleviates this constraint. Diamond spin also holds promise for high fidelity. Fujitsu and Delft reported two-qubit gates with error rates below 0.1% using NV center spins, which could translate to fewer physical qubits required per logical qubit in the future. Reducing the number of physical qubits per logical qubit is a critical factor for quantum error correction.

A candid perspective comes from Fujitsu’s partner, Dr. Kees Eijkel, General Director of QuTech at Delft University of Technology.

“Demonstrating the scalability expected of diamond spin quantum computing remains a long and challenging journey. However, by further strengthening our collaboration with Fujitsu, we are committed to tackling this ambitious and meaningful challenge and leading the development of next-generation quantum technologies.”

What This Prototype Represents

This is a single-module prototype. The announcement provides no qubit count, benchmarks, or claims of quantum advantage. The photonic integration, material bonding, and diamond thinning represent concrete fabrication achievements, and the control software functions in a test environment. These are the demonstrated elements.

Beyond this, everything is part of a roadmap. Fujitsu aims for a multi-module prototype by 2027, practical quantum computing by 2030, a 250-logical-qubit system by fiscal 2030, and 1,000 logical qubits by fiscal 2035. Fujitsu is also pursuing a parallel track with a 10,000-plus qubit superconducting machine and plans to eventually combine diamond and superconducting approaches.

What is certain: Fujitsu has developed a diamond-spin prototype that operates at warmer temperatures, features functional photonic readout, and includes a usable control stack. What remains aspirational: any claims involving “logical” qubits.