Skip to content

Humanoid

US Largest Robot Training Hub: 100 Humanoids Practice Object Handling

The United States has opened its largest robot training facility, where 100 humanoid robots are simultaneously practicing object handling tasks. This …

Robot Design Net · · 2 min read
US Largest Robot Training Hub: 100 Humanoids Practice Object Handling

The United States has opened its largest robot training facility, where 100 humanoid robots are simultaneously practicing object handling tasks. This development marks a significant scale-up in embodied AI training, potentially accelerating the deployment of general-purpose humanoids in logistics and manufacturing.

The facility, reported by Chosun Ilbo, addresses the critical bottleneck of real-world data collection for humanoid robots. While simulation-to-real transfer has advanced, physical training remains essential for robust manipulation and locomotion. Competing efforts include Tesla’s Optimus and Figure AI, but this hub’s scale—100 units operating in parallel—is unprecedented. Prior training setups typically involved fewer than a dozen robots.

The training hub features a controlled environment with standardized object sets and task sequences. Each humanoid is equipped with torque-controlled actuators and multi-fingered hands, enabling precise grip adaptation. The robots practice pick-and-place, assembly, and tool use, with performance metrics tracked via overhead cameras and force sensors. The facility can run 24/7, generating thousands of training hours per day. Design choices prioritize modularity: robots can be swapped in/out for maintenance without halting the entire pipeline. The infrastructure includes reconfigurable workstations and automated object replenishment systems.

This hub signals a shift from prototype testing to production-scale learning. For manufacturers, it implies faster iteration cycles for humanoid capabilities, particularly in warehousing and assembly. Research labs may leverage the facility for collaborative benchmarking. However, the high capital investment suggests that only well-funded entities can replicate this approach, potentially widening the gap between leaders and followers in humanoid robotics.


Source: 조선일보

Robotics Humanoid AI
← Back to Robot Design Net