Several members of IMPRS-IS are part of the founding team behind KE:SAI (Kyutai ELLIS Scalable Autonomous Intelligence), a newly launched Franco-German research laboratory dedicated to advancing open science in physical artificial intelligence.
KE:SAI is a joint initiative between Kyutai and the ELLIS Institute Tübingen, with operations based in Paris and Tübingen. The new lab aims to advance open science in “physical AI”—systems that can understand and act in real-world environments—through shared infrastructure, simulation, and robotics-focused machine learning research.
At the center of the founding team are several researchers closely connected to IMPRS-IS. IMPRS-IS faculty Andreas Geiger and Bernhard Schölkopf join as founding leadership.
The initiative also includes IMPRS-IS doctoral researchers Kashyap Chitta, Bernhard Jaeger, and Daniel Dauner as part of the founding technical team. Their involvement underscores a growing trend in which doctoral researchers contribute not only to academic outputs but also to the formation of new research organizations and large-scale collaborative labs.
KE:SAI is structured as a non-profit laboratory dedicated to advancing open-source approaches to physical AI. The lab’s mission is to develop a fully open technology stack for embodied intelligence, combining scalable simulation, efficient world models, and robotic learning systems.
A central research focus will be reducing the cost and complexity of real-world robotic learning through compact representations of the physical world, enabling more efficient training in simulation and transfer to real systems. The lab’s first major milestone is expected to be an open, high-performance self-driving stack, with broader applications in robotics domains such as manufacturing and healthcare.
The initiative builds on the combined strengths of the European AI ecosystem anchored by the Max Planck Institute for Intelligent Systems, the University of Tübingen, and the broader European Laboratory for Learning and Intelligent Systems (ELLIS) network. By combining academic research with applied engineering resources, KE:SAI aims to accelerate the transition of physical AI systems from research prototypes to real-world deployment.
Initial funding for KE:SAI is provided by Kyutai, and the organization is currently recruiting researchers, engineers, and doctoral-level talent in both Tübingen and Paris.
Editor's note: This article highlights the involvement of IMPRS-IS faculty and scholars in the newly announced KE initiative. For full details about the launch, readers can consult the official Kyutai–ELLIS Tübingen press release https://kesai.eu/blog/2026-05-20-kesai-launch/
More information