---
title: Fujitsu Bets on Diamond-Spin Qubits in the Race to Scale Quantum Computers
description: Fujitsu's prototype diamond-spin quantum computer, developed with QuTech, bets on linking modules with light. Where it sits in the 2026 quantum race.
author: Darie Nani (Editor-in-Chief)
date: 2026-09-08T10:12:32.374Z
updated: 2026-09-08T10:12:32.387Z
canonical: https://www.sovereignmagazine.com/article/fujitsu-diamond-spin-quantum-modular-scaling
image: https://cdn.nanimediahouse.com/pexels-a-minimalist-view-of-sleek-skyscrapers-in-katowice-showcasin-39063072.jpg
categories: Science & Tech
content_type: News
region: Global
publication: Sovereign Magazine
schema_type: Article
---

Fujitsu said on 8 September 2026 that it has developed a working prototype of a diamond-spin quantum computer, a machine that stores quantum information in defects inside diamond rather than in the superconducting circuits used by most of the industry's best-known players. The company describes it as the world's first working prototype of its kind. It works by embedding tin-vacancy (SnV) centres, single atoms sitting in a diamond lattice, into photonic integrated circuits, and Fujitsu says it runs at -271.6 degrees Celsius, a little warmer than the roughly -273.13 degrees at which typical superconducting quantum computers operate.

The temperature gap is small: both machines sit within a degree or two of absolute zero, and the difference between them changes nothing a user would notice. What sets the diamond-spin qubit apart is that it holds its quantum state for far longer than many rivals and can be linked to other qubits using light. Those two properties, long coherence times and optical connectivity, are what Fujitsu's bet rests on.

## The Qubits Hold Their State Longer and Link With Light

Most of the attention in quantum computing has gone to raw qubit counts, but the harder problem is wiring many qubits together without the whole system falling apart. Fujitsu presents its prototype as a step toward a modular quantum computer, an architecture in which smaller units are joined into a larger machine, and it calls that one of the more promising routes to scaling. Because SnV centres can be coupled optically, modules can in principle be connected with light rather than crammed onto a single chip. Fujitsu points to high fidelity and efficient optical connectivity as the reasons it favours the approach.

The science underneath the announcement is not Fujitsu's alone. The work grew out of joint research begun in 2020 with Delft University of Technology and QuTech, its quantum research institute, and the partners [established a joint quantum lab in January 2024](https://qutech.nl/2024/01/25/fujitsu-and-delft-university-of-technology-establish-new-quantum-lab/). The physics of efficiently coupling tin-vacancy centres to photonic cavities was published independently by QuTech in the journal Physical Review X in June 2026, three months before this week's announcement. Fujitsu's contribution is turning that result into an integrated, working system.

The academic partner is candid about how far there is to go. Kees Eijkel, general director of QuTech, frames a scalable diamond-spin machine as a long-term goal rather than a near one.

> "Demonstrating the scalability expected of diamond spin quantum computing remains a long and challenging journey."
> — Dr Kees Eijkel, General Director, QuTech

## Fujitsu's Roadmap Sits Alongside Rivals Aiming Higher

Fujitsu has put dates to its ambitions. The company says it is targeting 250 logical qubits by fiscal 2030 and 1,000 logical qubits by fiscal 2035. Logical qubits are the useful, error-corrected units that many physical qubits combine to produce, and they are the number that matters for real computation.

Those figures are modest next to the field's scale leaders. [IBM's public roadmap](https://www.ibm.com/roadmaps/quantum/) points to hundreds of logical qubits around 2030 and on the order of 10,000 by 2035, all on superconducting hardware. On raw numbers, then, Fujitsu's plan runs in line with or behind the front-runners. Its case rests on the architecture instead, and Vivek Mahajan, Fujitsu's corporate executive officer and chief technology officer, framed the diamond-spin approach as one that "not only offers exceptional scalability in its own right, but also has the potential to be integrated with superconducting quantum computers to further extend their capabilities." In Mahajan's telling, the two technologies could end up complementary rather than rivals.

The wider field in 2026 is genuinely split. Superconducting qubits anchor IBM and Google; trapped ions, fast-scaling neutral atoms from firms such as QuEra and Atom Computing, and photonics each have serious backers. Modular, optically connected designs are a mainstream strategy pursued by several of them, not a Fujitsu invention. Even within diamond, there is more than one path: Quantum Brilliance already ships room-temperature nitrogen-vacancy systems commercially, a different bet from Fujitsu's cryogenic tin-vacancy route.

For now the prototype is a demonstration rather than a product, though Fujitsu says it can be accessed through its Hybrid Quantum Computing Platform without specialist knowledge. Whether diamond-spin proves out at scale is, as Eijkel notes, a question that years more work will answer.

## FAQ

**Q: What is a diamond-spin qubit?**
It is a qubit that stores quantum information in the spin of an atomic-scale defect inside a diamond crystal. Fujitsu's prototype uses tin-vacancy centres, where a tin atom sits alongside a gap in the diamond lattice, coupled to photonic circuits so the qubits can be read and linked using light.

**Q: What is the difference between a physical qubit and a logical qubit?**
A physical qubit is a single hardware element, and it is error-prone. A logical qubit combines many physical qubits with error correction to behave as one reliable unit. Roadmap figures such as Fujitsu's 250 logical qubits by fiscal 2030 refer to these corrected units, which is why they are much smaller than physical-qubit counts.

**Q: What does a modular quantum computer mean?**
It is a design that joins a large machine together from smaller connected units rather than one enormous chip. Fujitsu links its modules optically, which it argues is a more practical way to scale than packing every qubit into a single device.

**Q: Who leads the quantum computing race in 2026?**
No single company leads outright. The field is divided across superconducting qubits (IBM, Google), trapped ions, neutral atoms (QuEra, Atom Computing) and photonics, and the leaders differ depending on whether you measure raw qubit count, error rates or architecture. On sheer scale, IBM's published targets are among the most aggressive.