Teradyne Robotics Unveils Wide Range of Production-Ready Physical AI Applications at Automate 2026

Teradyne Robotics, the parent company of Universal Robots and Mobile Industrial Robots, is set to showcase a suite of deployable physical AI solutions at the Automate 2026 trade show in Chicago. These applications aim to move beyond conceptual hype by offering production-ready tools for electronics manufacturing, logistics, and complex dexterous manipulation in unstructured environments. The announcement highlights a significant shift in the robotics sector toward integrated AI platforms that allow manufacturers to automate tasks previously considered too difficult or impractical for traditional robotic systems.

At the center of Teradyne’s presentation is the MiR1200 Pallet Jack, the company’s first physical AI-enabled product designed to navigate dynamic industrial settings. Jean-Pierre Hathout, President of Teradyne Robotics Group, emphasized that these solutions are currently available for purchase through a global network of system integrators, marking a transition from experimental technology to immediate factory-floor utility. By leveraging physical AI, Teradyne aims to address the challenges of unstructured environments, providing robots with the intelligence needed to handle variability in logistics and manufacturing workflows.

The foundational technology driving these advancements is PolyScope X, Universal Robots’ next-generation software platform built on a modern tech stack featuring containerized applications and native ROS 2 support. A key feature of PolyScope X is the introduction of Logic Programs, which are multi-threaded, PLC-style background processes that run independently of the main robot program. This allows for the coordination of multiple work cell components and data exchange without the need for external PLCs, effectively streamlining the hardware requirements for complex automation cells while maintaining cyber-secure and future-proof operations.

Teradyne is also highlighting the UR AI Trainer, an imitation learning platform developed in collaboration with Scale AI that uses force-aware data to train Vision-Language-Action (VLA) models. This system allows users to guide robots through tasks like smartphone packaging, with the resulting data capable of being integrated into the NVIDIA GR00T open VLA model or validated within the NVIDIA Isaac Sim simulation framework. Additionally, the company will demonstrate the Generalist GEN-1 model on UR12e robots, showcasing high-speed dexterous manipulation that signals a move toward general-purpose robotic intelligence capable of meeting the reliability thresholds required for real-world industrial deployments.