Kyutai and the ELLIS Institute Tübingen launch KE:SAI (Kyutai ELLIS Scalable Autonomous Intelligence), a non-profit research laboratory. Its work focuses on physical AI: systems capable of perceiving, modeling, and acting in real-world environments.
Founded in Paris in 2023, Kyutai has become known for open-source generative AI work. The ELLIS Institute Tübingen anchors its activities in fundamental machine learning research and collaborates closely with the Max Planck Institute for Intelligent Systems and the University of Tübingen.
KE:SAI will open offices in both cities. The lab will boast high-performance computing infrastructures and hardware prototyping capabilities, two essential resources for research in robotics and autonomous systems.
The project places world models at the heart of its scientific approach. These representations of the real world enable training robotic systems in simulation rather than in costly physical environments. The objective: to reduce data and compute power consumption to make this learning more accessible.
The founding team brings together Andreas Geiger (CEO, Professor at the University of Tübingen), Kashyap Chitta (CTO), Bernhard Schölkopf (Scientific Director of the ELLIS Institute and of the Max Planck Institute), as well as Bernhard Jaeger and Daniel Dauner.
World models to reduce the cost of physical learning
KE:SAI places the development of a complete technology stack for autonomous driving at the forefront of its priorities. This stack will be fully open and aimed at cutting-edge performance. The laboratory plans to extend its work to other robotic domains, notably manufacturing and healthcare.
The choice of open source reflects a democratization logic. In other AI domains, sharing models and algorithms has accelerated innovation. Physical AI has not yet benefited from this dynamic. KE:SAI aims to fill this gap precisely. The European industrial ecosystem, automotive and manufacturing sectors leading the way, represents the natural outlet for these efforts.
Kyutai is funding the lab’s initial operations. It has itself raised €300 million from the Iliad Group, the CMA CGM Group, and The Eric & Wendy Schmidt Fund for Strategic Innovations at launch.
KE:SAI is currently recruiting researchers, engineers, and PhD students at its two sites.
