The Tübingen-based startup is one of only ten across Europe chosen for the Next Frontier AI Challenge to build the future of embodied intelligence.
TÜBINGEN — July 2026 — Ontic Labs, a pioneering AI startup co-founded by Wieland Brendel—Principal Investigator at the ELLIS Institute Tübingen and Research Group Leader at the Tübingen AI Center—has achieved a major breakthrough. The company has been selected for the first round of the highly competitive Next Frontier AI Challenge by Germany's Federal Agency for Disruptive Innovation (SPRIND).
As one of just ten exceptional teams chosen from across Europe, Ontic Labs will receive €3 million in initial funding to design and develop next-generation frontier AI technologies.
Key Highlights
-
The Venture: Ontic Labs, co-founded by the ELLIS Institute's Wieland Brendel together with Omid Taheri, Mikel Zhobro, Gerard Pons-Moll, and Georg Martius — uniting researchers and engineers from the Max Planck Institute for Intelligent Systems, the University of Tübingen, and the ELLIS Institute Tübingen.
-
The Funding: up to €3 million in equity-free Stage-1 capital from SPRIND, within a €125 million program spanning three stages.
-
The Mission: foundation models for physical intelligence — world models that learn how the world works from large-scale video, so any robot can put that understanding to work.
- The Goal: over the next 24 months, the field narrows from ten teams to up to three labs positioned to become globally competitive European frontier AI companies. Ontic Labs intends to be one of them.
Moving Beyond Pixels: AI That Understands Physics
Under the scientific co-leadership of Wieland Brendel, Ontic Labs is tackling one of the most important missing pieces in modern AI: understanding of the physical world. Today's generative models excel at predicting words and generating images, but they lack a grasp of objects, physics, materials, and cause and effect.
Ontic Labs builds foundation models for physical intelligence: world models that learn how the world works from large-scale video. Instead of being programmed for narrow tasks, robots equipped with these models can draw on that understanding to navigate, manipulate, and adapt in complex, unpredictable real-world environments. The ambition: learn the physics of the world once — and enable every robot to perform every task in every scene.
"Today's foundation models are impressive, but they pattern-match the world rather than understand it. What's missing isn't scale — it's understanding," says Wieland Brendel. "With SPRIND's backing, we're building world models that learn physics directly from video. If we succeed, robots will no longer need to be programmed for every task; they will understand the world they act in."
Empowering the Tübingen Ecosystem
The selection of Ontic Labs highlights the deep-tech pipeline emerging from Tübingen's research ecosystem — the ELLIS Institute Tübingen, the Tübingen AI Center, the University of Tübingen, and the Max Planck Institute for Intelligent Systems. By carrying frontier machine-learning research into a company built in Tübingen, for Europe, Brendel and his co-founders reinforce the region's standing as one of Europe's leading hubs for AI innovation.
More Information