Universal Robots

Teradyne's subsidiary Universal Robots offers the UR AI Trainer, a hardware-software system built with Scale AI that allows robot operators to teach robots new tasks by physically guiding them through the motions, with the system capturing synchronized motion, force, and visual data to train AI models directly on production-grade robot hardware. The system was launched at NVIDIA GTC in March 2026.

The UR AI Trainer uses a leader-follower robot setup: a human operator guides a 'leader' robot, while a 'follower' robot mirrors the motion in real time. The system automatically captures synchronized motion, force, torque, and visual data required to train Vision-Language-Action (VLA) models — AI systems that combine visual perception, language understanding, and physical action. The platform is integrated with Scale AI's recording software and supports cloud integration with NVIDIA GR00T and NVIDIA Isaac Sim for simulation-based validation. Universal Robots describes it as the 'industry's first direct lab-to-factory solution for AI model training.' The product is listed on the Universal Robots marketplace as generally available.

Teradyne's subsidiary Universal Robots integrated NVIDIA's cuMotion AI path-planning software into its collaborative robot arms, making collision-free path planning up to 50–80 times faster than previous solutions. This capability was announced in March 2024 and is available to customers through UR's PolyScope X software platform.

The integration combines the NVIDIA Jetson AGX Orin edge-compute module, the cuMotion path planner from NVIDIA's Isaac Manipulator platform, and Universal Robots' PolyScope X software. The AI system calculates collision-free trajectories by simultaneously evaluating many possible paths in parallel, and can also optimize for criteria such as speed, energy efficiency, and minimal wear. Universal Robots confirmed the capability through a company press release and it was demonstrated at NVIDIA GTC 2024. The UR AI Accelerator toolkit, launched in late 2024, provides developers broader access to NVIDIA Isaac platform libraries for building AI-powered robotics applications on UR hardware.