Frontier AI Innovation Summit: A Joint Conference

On October 7–8, 2026, the German AI community and European partners will gather at Messe Stuttgart for the Frontier AI Innovation Summit, a joint event combining the All Hands Meeting of the German AI Excellence Centers and the ELIAS Alliance Annual Meeting. Designed to bridge top-tier research, entrepreneurship, and societal impact, this flagship event highlights how cutting-edge artificial intelligence strengthens Europe’s technological sovereignty. Across two dynamic days featuring keynotes, panel discussions, hands-on workshops, interactive poster sessions, and high-level networking, attendees will gain direct insights into transformative AI projects from leading research hubs across Germany and Europe.

As one of Germany’s six AI centers, the Lamarr Institute will be presenting two of its focused impact AI projects that demonstrate the power of translating foundational research into real-world applications. The first featured project, Soofi, presented by Mehdi Ali, Lamarr Lead Scientist Foundation Models, represents a sovereign European alternative to non-European foundation models. Funded by the German Federal Ministry for Economic Affairs and Energy under the IPCEI-CIS initiative, Soofi focuses on industrial applications where control, transparency, and efficiency are paramount. With Lamarr researchers contributing substantially to its pre-training, the newly introduced Soofi S—a Mixture-of-Experts hybrid Mamba Transformer trained on roughly 27 trillion tokens—serves as a primary example of how Lamarr’s foundational research translates directly into usable AI infrastructure.

Complementing this infrastructure focus, Lamarr also highlights Physics-LLM, an interdisciplinary project funded under the BMFTR’s ErUM-Data program that unites researchers from eleven institutions, including DESY, Forschungszentrum Jülich, TUM, and the Leibniz Institute for Astrophysics Potsdam. Coordinated by Dr. Tim Ruhe alongside Prof. Dr. Lucie Flek, Lamarr Area Chair NLP, and Prof. Dr. Dr. Wolfgang Rhode, Lamarr Area Chair Physics, the initiative combines state-of-the-art Large Language Models with complex physics research. By delivering an LLM-enhanced toolkit for scientific data management, Physics-LLM transforms traditional Big Data into actionable “Smart Data” under FAIR data principles, streamlining scientific workflows and advancing data-driven discovery across the natural sciences.