Nebius recently announced the launch of the 'Physical AI Living Lab', a specialized program designed to accelerate the growth of robotics startups in the UK and Europe.
This initiative is significant because it directly tackles the two largest obstacles early-stage robotics companies face: access to sophisticated simulation tools and the massive computing power needed to run them. The program provides a complete toolkit, built on NVIDIA's powerful Physical AI stack, including world models like Cosmos 3, and runs on Nebius's high-performance cloud infrastructure in the UK. This allows founders to move from virtual testing to real-world deployment much more quickly.
The timing of this launch is driven by three key converging factors. First is the maturation of AI technology. NVIDIA recently released Cosmos 3, an open-source world model, and the Physical AI Data Factory blueprint. These tools make it possible for startups to generate vast amounts of realistic, physics-consistent synthetic data for training their robots, a process that was once prohibitively expensive and complex.
Second, there are strong policy tailwinds. The EU AI Act is creating a clear regulatory landscape, with key rules for general-purpose AI models coming into force in August 2026. Simultaneously, the UK government is actively promoting its own computing power through its 'Compute Roadmap' and AI Research Resource (AIRR). This creates a stable environment where startups know the rules they need to follow and have access to local, powerful computing resources, which is crucial for data residency and low-latency training.
Finally, there is clear market demand. The success of UK-based companies like Wayve, which has raised substantial funding, demonstrates a strong appetite for the very simulation-heavy infrastructure that the Living Lab provides. In essence, the lab acts as a bridge, connecting cutting-edge tools, supportive government policies, and tangible market needs. It provides a mentored, compliance-aware on-ramp for the next generation of robotics innovators.
- Physical AI: Artificial intelligence designed to understand and interact with the physical world through robotics and other embodied systems.
- World Model: An AI model that learns an internal representation of how the world works, allowing it to simulate future events and understand cause and effect.
- Sovereign Compute: A nation's independent capability to build and operate its own large-scale computing infrastructure for AI and other strategic purposes.