NEURA Robotics establishes NEURA Gym RWTH Aachen to train physical AI

NEURA Robotics has partnered with RWTH Aachen University to establish NEURA Gym RWTH Aachen, a facility intended to help train physical AI for cognitive robots. The company says robot developers lack the vast quantities of real-world data available to large language model developers, while simulation alone cannot capture real operating conditions such as friction, variability, and unpredictability.

The planned gyms will pair physical robot training with high-fidelity simulation. They are designed to generate multimodal datasets from real-world interactions and feed them into Neuraverse, NEURA’s open, cloud-based development platform. According to the company, partners will be able to train and validate robots for specific applications before full industrial deployment, potentially lowering adoption risk and shortening deployment time.

Metzingen-based NEURA said it is building 10 such facilities across Europe, the United States, and China, with half expected to be operating by the end of 2026. Its recent Series C financing, worth up to $1.4 billion, is supporting the network.

Germany will host two centers. The Aachen site, on RWTH Aachen’s Hightech Campus Melaten, is planned at roughly 3,000 square meters. A second facility, the TUM RoboGym at Munich Airport, is being developed with the Technical University of Munich and its Munich Institute of Robotics and Machine Intelligence. It will exceed 2,300 square meters and will test fleets of cognitive and humanoid robots.

NEURA will supply the technology infrastructure, while academic researchers and corporate partners contribute specialized expertise. Although the gyms will use shared physical AI infrastructure through Neuraverse, their focus will vary with regional and partner strengths. Technologies from ecosystem partners will be available, and industries expressing interest include manufacturing, automotive, healthcare, and industrial automation.