The real world still poses the greatest challenges for robots. A system may recognize objects flawlessly and master movement sequences perfectly in simulations, yet still fail as soon as a workpiece shifts, a surface gives way, or a person unexpectedly reaches into its work area.
While language models are trained with data from the internet, “Physical AI” has to gather its experience primarily in the real world. To this end, Neura Robotics is developing so-called gyms: training centers where robots can grasp, sort, or operate tools under conditions that are as authentic as possible. In the process, they capture images, touch, sounds, and spatial changes. The gym is intended to bridge the gap between the laboratory and the factory floor. Industrial companies and developers can recreate workflows there, train robots on them, and test their capabilities before they are deployed in production. The workflows they learn can then be transferred to other systems via the Neuraverse digital platform and, if desired, offered to other companies.
“The biggest bottleneck for Physical AI is no longer intelligence, but practical experience,” says David Reger, CEO and founder of the startup. “Robots learn by interacting with the real world, and that’s exactly why we’re building the gyms.”
Ten such centers are planned across Europe, two of which are expected to be established in Germany by the end of 2026: on the RWTH Aachen campus and at Munich Airport in cooperation with the Technical University of Munich.
The funding for these investments comes from a financing round of up to $1.4 billion, which Neura Robotics announced in June 2026. The investors include Nvidia, Amazon, Qualcomm, Bosch, Schaeffler, and the European Investment Bank. With the new capital, the company, founded in Metzingen in 2019, also plans to massively expand the series production of its industrial, service, and humanoid robots – and move closer to its goal of establishing Neura as a European counterweight to the major robotics providers from the US and China.
