Siemens Runs Live Factory Test of Nvidia-Powered Humanoid Robot

Siemens is testing whether humanoid robots can handle real shop-floor routines.

The company said a humanoid robot developed by UK startup Humanoid has been tested at Siemens’ electronics factory in Erlangen, Germany, in a live industrial trial using Nvidia technology.

In a press release, Siemens said the trial is an early step toward using “physical AI” in manufacturing. The focus extends past the robot itself to the factory software, automation systems, and AI infrastructure around it.

Inside Siemens’ live factory test

Humanoid’s HMND 01 Alpha, a wheeled humanoid robot built for industrial settings, was used for logistics work inside the Erlangen plant. Its tasks included picking up, transporting, and placing containers for human operators.

The robot operated inside Siemens’ electronics factory in a live production environment. HMND 01 Alpha was built on Nvidia’s physical AI stack and introduced into plant logistics as the companies tested how humanoid systems could be used in day-to-day industrial operations.

The company presented the deployment as a manufacturing test in an active factory, with the robot performing routine logistics work on-site. Humanoid CEO and founder Artem Sokolov said the company is building robots for “not only in controlled lab settings, but also in real-world factory environments, handling meaningful industrial tasks.” 

The tote work, by the numbers

Siemens put performance targets on the logistics run and said the robot met them. HMND 01 Alpha recorded 60 tote moves per hour, stayed operational for more than eight hours, and achieved autonomous pick-and-place success rates above 90%.

Those figures give the trial its measurable side. The robot was not only moving containers through the plant but also doing so at a sustained rate over a full working stretch, with a high success rate on the handling task itself.

Three companies, three jobs, one factory test 

HMND 01 Alpha is equipped with advanced manipulation capabilities for work in human-centric factory spaces.

Siemens handled the industrial layer through Siemens Xcelerator. That included the digital twin, AI-enabled perception, integrated control and PLC-robot interfaces, fleet management, industrial communication networks, and high-performance drives needed to connect the system to the wider factory environment.

Nvidia’s contributions included Jetson Thor for edge compute, Isaac Sim for simulation, and Isaac Lab for reinforcement learning and policy training. The simulation-first approach cut prototype development from a typical 18 to 24 months to seven months.

Deepu Talla, vice president of Robotics and Edge AI at Nvidia, said, “Factories of the future demand robots that can perceive, reason, and adapt autonomously alongside human workers.” He added that the deployment “paves the way for humanoid robots meeting real production targets on a live factory floor.” 

What it takes to move from one trial to wider deployment

For Siemens, that next step depends on coordination across the factory floor. The company said a humanoid system needs real-time data exchange with production systems and other autonomous guided vehicles, synchronized workflows with machinery and human operators, and behavior that can respond to changing conditions.

It also said this kind of coordination is what turns a humanoid from an isolated machine into part of regular factory operations. Siemens called the setup a “factory-grade model” for deploying humanoids in industrial settings.

Tesla may be preparing to turn its Shanghai Gigafactory into a launchpad for humanoid robot production.