ViMi Labs, the technology development partner of Blue SI, has been awarded one of Europe’s most competitive research and innovation grants: first-ever Advanced Innovation Challenge on “Accelerating Physical AI”. The project APOLLO -Agentic Physical AI Operating System for Self-Driving Labs- aims at accelerating autonomous scientific discovery through exascale-enabled agentic AI and physical intelligence
Today, bringing a new material from the lab to the market can take 10 to 20 years. Much of that time is spent on slow, repetitive trial-and-error in the laboratory. Robots can help, but setting up a robotic lab usually takes years of specialist engineering and costs a lot. Every new task has to be programmed manually. APOLLO changes this by building an “operating system” for self-driving labs: software that lets AI agents plan experiments, train robots and run them, while scientists stay in control.
In simple terms, it works in three steps:
- Learn in a virtual lab first. Robots practise their tasks in a realistic digital copy of the lab (a “digital twin”).
- Let AI agents run the work. A team of specialised AI agents picks the best experimental method, simulates it, checks whether it is ready, and turns it into instructions a robot can follow.
- Keep humans in charge. A built-in “Guardian” layer makes sure a human approves every physical action, and that every result can be traced back to where it came from.
The result is a lab that does more than follow a fixed recipe. It can learn new skills and get better over time. And it does not depend on one robot brand. APOLLO is led by ViMi Labs (Jülich, Germany). In its first phase, APOLLO will be tested in two real robotic labs in Germany and Canada.
For ViMi and Blue SI, APOLLO is a step toward a future where AI does not just give advice. It helps carry out the work, from the lab bench to the factory floor. The disruptive advantage is to move away from fixed-recipe automation to labs that train, validate and improve their own physical skills.