Siemens and Nvidia said the successful test advances their push to build factories where AI-powered machines can adapt and work with human staff
German technology company Siemens and chip giant Nvidia have tested a humanoid robot on a live factory floor.
The test in partnership with robotics company Humanoid based in the United Kingdom, marks a step towards artificial intelligence-driven production where machines and humans work alongside each other. Siemens said in an announcement that the robot, powered by Nvidia’s AI technology, was deployed at its electronics plant in Erlangen, in southwest Germany.
The robot, Humanoid’s HMND 01, performed routine logistics tasks, such as picking up, moving, and placing containers used by human workers.
The robot operated autonomously for more than eight hours and completed over 90% of its tasks, moving around 60 containers per hour during the trial, Siemens said.
The project is part of its partnership with Nvidia to develop “world’s first AI-driven, adaptive” factories.
“Factories of the future demand robots that can perceive, reason, and adapt autonomously alongside human workers," said Deepu Talla, vice president of robotics and edge AI at Nvidia.
"With Siemens providing the industrial integration backbone and Humanoid deploying NVIDIA's full physical AI stack, from simulation-first training to real-time edge inference, this deployment paves the way for humanoid robots meeting real production targets on a live factory floor."
Humanoid’s robots have previously demonstrated fast learning, walking, and dexterity in real-world environments.
By using Nvidia’s simulation and training tools, much of the robot’s development could be done virtually, reducing the need for physical testing and cutting design time from up to two years to around seven months, the companies said.
This could help address labour shortages by allowing robots to take on complex tasks that currently need human involvement, as traditional automation has struggled to handle.
The companies described the trial as a “milestone in the journey to bring physical AI from vision to industrial reality”, but did not give a timeline for when such systems could be rolled out more widely.
