Quantum computers made in Germany

Coordinated by Forschungszentrum Jülich, German research institutions and companies have developed a quantum computer comprising 30 superconducting qubits with low error rates – a significant contribution to Germany’s technological independence.

Here it is: QSolid demonstrator at Forschungszentrum Jülich

In 2021, the Federal Ministry of Research set out a clear objective: to have a competitive quantum computer based on domestic research that is accessible to users by 2027. But this was not the only goal: while the computer was being developed, technologies were to be transferred to industry and start-ups established. To realize this vision, the ministry launched the funding measure for quantum computer demonstration setups”.

Prof. Frank Wilhelm-Mauch from the Peter Grünberg Institute – Quantum Computing Analytics subsequently formed a consortium comprising 22 German research institutions and companies. It successfully applied for funding and launched the QSolid project in early 2022 with a budget of € 76.3 million. Since autumn 2026, the 30-qubit computer has undergone various tests, with its key strength being a particularly low error rate in the qubits.

“Even with the first prototype, the performance data showed that our approach of relying on qubits with low error rates is paying off.

Prof. Frank Wilhelm-Mauch, Project Coordinator and Director at the Jülich Institute for Quantum Computer Analytics

A key factor in the success of the project was the appointment of Prof. Rami Barends. The Dutch physicist had helped to develop the first quantum computer at the Google Quantum AI Lab in California that outperformed classical computers on a specific task. At Jülich, as director of the newly founded institute division “Functional Quantum Systems”, he took charge of developing and manufacturing the QSolid quantum processors.

On board from the very beginning: Jeferson R. Guimaraes, a doctoral candidate at the Jülich Institute for Functional Quantum Systems.

Jeferson R. Guimaraes was one of the team’s founding members. “At the start, there were only seven of us, and I was something of a jack-of-all-trades who stepped in wherever I was needed,” recalls the doctoral researcher. The task was not only to manufacture the quantum chip, but to set up the initial laboratory infrastructure, such as the cooling systems. Guimaraes also helped to design the mechanical components and the magnetic shielding for the quantum computer.

After around six months, he was able to devote himself to the main task: fabricating quantum chips at the Helmholtz Nano Facility, which has 1,200 m² of cleanroom space on the Jülich campus. “It was a lot of painstaking manual and mental work where you couldn’t afford to make any mistakes – and it meant standing for hours on end every day,” says Guimaraes. Meanwhile, QSolid partners manufactured and supplied further components for the quantum computer, including printed circuit boards, mechanical parts, and cables.

By the end of 2024, the project had reached its first milestone: the 10-qubit prototype was up and running. Physical qubits are the quantum units that a quantum computer can, ideally, control simultaneously and use for calculations. Compared with Google’s 53-qubit processor from 2019, for example, 10 qubits seem modest. “However, the sheer number of qubits does not determine the performance of a quantum computer,” says project coordinator Wilhelm-Mauch. Even the slightest disturbances can knock qubits out of their quantum state and destroy their ability to compute. That is why many qubits are used solely for error correction. “Even with the 10-qubit prototype, the performance data showed that our approach of relying on qubits with low error rates is paying off,” emphasizes Wilhelm-Mauch. Achieving a certain level of performance with fewer qubits paves the way for more cost-effective and compact quantum computers.

In the months that followed, the team succeeded in connecting the 10-qubit prototype to Jülich’s quantum computing infrastructure, JUNIQ, which is operated by the Jülich Supercomputing Centre (JSC). This enabled QSolid partners, such as those based in Berlin, to access it. At the same time, the QSolid team was working on building a 30-qubit quantum computer – the ultimate goal of the project.

30-qubit chip

Thanks to the intermediate step involving the 10-qubit system, the team was able to identify critical issues and implement improvements while the project was still underway. “A quantum computer consists of hundreds of components that must be assembled into a single system following design and manufacture,” explains Guimaraes. “It is important to check at regular intervals whether the entire development cycle, right through to the integration of all components into a single system, is functioning correctly.”

Wilhelm-Mauch also believes that without the lessons learnt during the development of the 10-qubit prototype, the team might not have been able to successfully build the error-reduced 30-qubit system. As the theoretical physicist puts it: “It has proved true that a large system is far more susceptible to external disturbances, such as magnetic fields and vibrations, and will not tolerate even the slightest deviations in the manufacture of the quantum chip.”

QSolid has also fulfilled the objective of bringing together German companies and research institutions in a quantum computing network and transferring technology to industry. “This applies in particular to software components and high-frequency cabling,” says Wilhelm-Mauch.

The Jülich-based scientist is already working on the follow-up projects, which aim to achieve true quantum utility. They fit in perfectly with High-Tech Agenda Germany published by the Federal Government in July 2025. This document lists quantum technology as one of six technologies to be prioritized for funding due to their enormous economic and societal potential. With the planned computer, Jülich could make a significant contribution to making Germany – and Europe – more independent and competitive in the field of quantum computing.

This article is part of effzett 2/2026. Text: Frank Frick

Last Modified: 27.10.2026