Neura Robotics, E.ON and German energy company enviaM have announced a partnership to develop physical AI applications for electricity networks, with initial real-world trials planned for 2027.
The collaboration will focus on using intelligent robots to inspect and monitor energy infrastructure, including substations, and support workers performing physically demanding, repetitive or potentially hazardous tasks.
Mitnetz Strom, a subsidiary of enviaM and one of eastern Germany’s largest electricity distribution network operators, will serve as the pilot partner.
David Reger, founder and CEO of Neura Robotics, says: “Physical AI and cognitive robotics will fundamentally reshape the energy sector. The key is to combine technological progress with the practical expertise of the people who operate our energy infrastructure every day.”
Training robots for real-world grid operations
The partners will develop robotic capabilities at TUM RoboGym powered by Neura, a 2,300-square-meter robotics training facility opened in October 2026 at Munich Airport in collaboration with the Technical University of Munich.
The facility combines human demonstrations, real-world robotic data and simulation to teach robots new skills.
Energy infrastructure specialists will demonstrate tasks using motion capture systems, sensor-equipped suits and teleoperation. Robots will then practice these activities, generating additional training data before their capabilities are tested in operational environments.
The training process will use Neura’s Neuron AI multimodal foundation model, which combines vision, touch, audio, language and robot-state information.
The company’s Neuraverse platform will connect the training environment with real-world applications, allowing robotic skills to be refined and potentially reused across compatible systems.
Humanoid robots for existing infrastructure
One potential application involves Neura’s 4NE1 humanoid robot, designed to navigate environments originally built for human workers, including stairs, doors and control cabinets.
This could allow robots to undertake inspection and maintenance-support activities without requiring extensive modifications to existing facilities.
Stephan Lowis, CEO of the enviaM Group, says: “Humanoid robots can provide targeted support, particularly for physically demanding tasks or those with specific safety requirements.”
The partners will initially identify suitable applications and develop the necessary robotic skills before moving to operational testing.
Mitnetz Strom expects to gain its first practical experience with the technology during 2027.
The companies say their longer-term objective is to establish a scalable approach to physical AI in critical energy infrastructure, rather than developing isolated robotic demonstrations.
