May 21, 2026
Forschungszentrum Jülich (FZJ) is part of the new Unit NRW at ELLIS, the European Institute for Learning and Intelligent Systems. ELLIS is considered one of Europe’s leading networks for machine learning and artificial intelligence research. The new unit will bring together leading AI researchers and large-scale computing infrastructure from across North Rhine-Westphalia to develop open-source foundational models and datasets to power trusted AI applications for scientific and industrial applications.
The central focus of this unit is the development and research of open source, generalist-based models as core components of machine learning research. We investigate how supercomputers and open datasets can be used to train such models to frontier levels, increase their reliability, and safely adapt them to different application domains. The unit also strengthens research on trustworthy AI, especially for situations involving uncertainty, rare events, or sensitive real-world environments.
The second research track of the ELLIS NRW initiative focuses on moving machine learning techniques to application areas such as healthcare, sustainable agriculture, and embedded AI systems such as autonomous robots. Researchers in this unit cover areas such as robotics, computer vision, natural language processing, healthcare AI, and neural symbolic systems.
Within FZJ, the Jülich Supercomputing Center (JSC), the Institute for Advanced Simulation (IAS), and the Peter Grünberg Institute (PGI) are contributing to this effort. FZJ is an important contributor alongside Bonn and Aachen in shaping the scientific profile of the NRW force, especially in the area of open-based models and the datasets required for their development.
The Jülich infrastructure, including JAIF (JUPITER AI Factory), not only enables large-scale training of foundational models, but also supports the development, deployment, and management of large-scale datasets for future open AI systems. This effort relies on advanced high-performance infrastructure, including FZJ’s JUPITER supercomputer, and expertise in operating large-scale AI workloads.
“Forschungszentrum Jülich offers a unique combination of AI expertise, high-performance computing and scientific collaboration. Through ELLIS Unit NRW, we will help advance open-based modeling research in Europe and translate AI research into real-world applications.”
This effort will benefit from close collaboration with LAION, an international open research organization specializing in open-foundation models, large-scale AI training workflows, and open datasets. LAION is already conducting research on open foundation models using the resources of the Jülich Supercomputing Center (JSC), underscoring the strategic importance of Jülich’s high-performance computing infrastructure for open AI research in Europe.
“Through the new ELLIS Unit NRW, LAION will further strengthen its collaboration with ELLIS researchers in the field of open infrastructure models, building on collaborations such as ongoing European major projects including OpenEuroLLM, ELLIOT, Minerva and LLMs4EU,” explained Dr. Zhenia Zitsev, one of the co-directors of the unit, one of its scientific leaders and co-founder of LAION.
The NRW Unit is coordinated by the University of Bonn under the leadership of ELLIS Fellow Professor Dr. Jürgen Gall with a joint board of participating institutions including Dr. Zhenia Zitsev (JSC/FZJ and LAION). Additional members of the Forschungszentrum Jülich include Professor Emre Neft (PGI), Professor Christer Michelsen (JSC), Dr. Martin Schulz (JSC) and Professor Moritz Helias (IAS).
The unit includes approximately 40 principal investigators and more than 300 postdocs and postdoctoral fellows from across North Rhine-Westphalia and fosters joint research, doctoral training, the ELLIS Summer School, and cooperation with industry and public sector partners.
The establishment of ELLIS Unit NRW further strengthens the role of the Jülich Foundation and the Jülich Supercomputing Center as a major contributor to Europe’s open, high-performance, and trusted AI ecosystem.
/Open to the public. This material from the original organization/author may be of a contemporary nature and has been edited for clarity, style, and length. Mirage.News does not take any institutional position or position, and all views, positions, and conclusions expressed herein are those of the authors alone. Read the full text here.
