JUPITER AI Factory

JAIF is one of 19 AI factories set up by the European Commission to bridge the gap between AI research and real-world application by providing researchers in the European Union, industry (start-ups, SMEs, and large companies), and the public sector with access to AI infrastructure.

JAIF combines Europe's most powerful computing resources with unique AI expertise, domain-specific solutions, a novel, state-of-the-art inference infrastructure, and outstanding talent to solve real-world problems in the private sector, academia, and beyond.

The JAIF consortium network consists of four core partners: the Supercomputing Centre (JSC) at Forschungszentrum Jülich (JSC), the AI Center at RWTH Aachen University, Fraunhofer FIT, Fraunhofer IAIS, and hessian.AI, as well as two AI Factory Antennae, the Belgian AI Factory Antenna and the Hungarian AI Factory Antenna. Together, the partners are creating a one-stop shop for research and industry and offering in-depth support for the use of JUPITER. The focus is on strategically important fields of application: healthcare, energy, climate change, education, media, the public sector, and finance.

Fraunhofer FIT contributes its expertise and experience to the JAIF project in the areas of Data Science and Artificial Intelligence, Digital Energy, Digital Business and Human-Centred Engineering and Design.

Data Science and Artificial Intelligence: Processing, integration, organisation and analysis of a wide variety of data types through a systematic, flexible, effective and efficient approach that covers the entire lifecycle of knowledge-based value creation and decision support from data. In the context of JAIF, the following areas of expertise are contributed:

• Data Protection and Data Sovereignty: Trustworthiness and security of AI systems
• Smart Data Management and Analytics: Data management and exchange, ontologies and interoperability, as well as data spaces
• Knowledge-Enhanced Data Analytics: Graph RAG, model adaptation, agent-based AI and complex knowledge structures
• FAIR Data and Distributed Analytics: FAIR principles, semantics, interoperability, machine-readable data formats and powerful, privacy-friendly analytics
• Process Mining: Workflow templates for AI tasks such as training, fine-tuning, inference and evaluation

Digital Energy: Models and pilots for innovative energy supply systems, as well as comprehensive domain knowledge in the energy sector

Digital Business: AI system engineering, agent-based systems and human-AI interaction for digital transformation in enterprises, as well as comprehensive domain knowledge in the financial sector

Human-Centered Engineering and Design: User-centered methods such as usability engineering and design thinking for requirements analysis, workshop formats for data collection, and impact analysis

Learning Center: Development of learning offerings collaboratively in interdisciplinary teams and based on scientific findings and best practices from didactics, psychology and design thinking

Official JAIF project website: https://jaif.ai

Half of JUPITER's funding comes from the European supercomputing initiative EuroHPC Joint Undertaking (EuroHPC JU), with an additional quarter each from the German Federal Ministry of Research, Technology and Space (BMFTR, formerly BMBF) and the Ministry of Culture and Science of North Rhine-Westphalia (MKW NRW) via the Gauss Centre for Supercomputing (GCS).