Universal Robots Unveils AI Trainer at GTC 2026, Pioneering Lab-to-Factory AI Model Training
SAN JOSE, Calif. – March 16, 2026 – Universal Robots (UR) today launched the UR AI Trainer at GTC 2026 in Silicon Valley, marking a significant advancement in robotics and artificial intelligence. Developed in collaboration with Scale AI, the AI Trainer facilitates a shift from pre-programmed robotic applications to fully AI-driven tasks, powered by data generated in AI training cells where robots learn through human imitation.
Bridging the Gap Between Lab and Factory
The UR AI Trainer is designed as the first direct lab-to-factory solution for AI model training. According to Anders Beck, VP of AI Robotics Products at Universal Robots, customers are increasingly seeking methods to collect high-fidelity, synchronized robot and vision data to train AI models on the robots they intend to deploy. “Our AI Trainer is the industry’s first direct lab-to-factory solution for AI model training,” Beck stated in a press release.
Showcasing Advanced Capabilities at GTC 2026
At GTC 2026, Universal Robots is demonstrating the AI Trainer’s capabilities alongside a robotic foundation model from Generalist AI, a preferred model partner. Two UR robots are autonomously completing a complex smartphone packaging task, a feat previously unattainable without recent advancements in Physical AI.
Addressing Challenges in AI Robotics Training
Traditional AI robotics training often faces hurdles due to fragmented hardware and low-fidelity data capture. Much of the existing training data is collected on research robots unsuitable for production environments, and many systems rely solely on visual feedback, limiting their effectiveness in delicate or contact-rich tasks. The AI Trainer directly addresses these limitations by utilizing Universal Robots’ Direct Torque Control and force feedback features, giving developers greater control over the robot’s physical interactions and enabling training on robust hardware used in over 100,000 industrial deployments.
Scale AI Partnership and Data Flywheel
The partnership with Scale AI enables a continuous feedback loop for optimizing physical AI systems. The AI Trainer allows human operators to guide UR robots through tasks in a leader-follower setup, automatically capturing high-quality multimodal data for robotics AI development. During demonstrations, the system records synchronized motion, force, and visual data, creating structured datasets for training Vision-Language-Action (VLA) models.
“Universal Robots is a leader in industrial robotics, and its global footprint offers the ideal foundation for data capture and AI deployment,” said Ben Levin, General Manager, Physical AI at Scale AI. “Together, we’ve created an integrated robotics data flywheel, allowing customers to train, deploy, and improve their AI models faster than ever before.” UR and Scale AI plan to release a large-scale industrial dataset collected on UR robots later this year.
Hands-on Experience and Simulated Training
GTC attendees can experience the AI Trainer firsthand at the UR booth, guiding two UR3e ‘leader’ robots to control two UR7e ‘follower’ robots in an advanced smartphone packaging task with haptic feedback. The system records demonstration data in real-time on Scale’s platform and allows for replay directly on the AI Trainer.
A parallel demonstration showcases the same smartphone packaging task trained virtually within NVIDIA Omniverse and Isaac Sim, utilizing a simulated bi-manual UR3e system controlled with Haply Inverse3 devices, providing physics-accurate simulation with real-time haptic feedback.
Leveraging NVIDIA Technology
Universal Robots is also exploring the use of the NVIDIA Physical AI Data Factory Blueprint to automate and scale synthetic data generation, transforming compute power into a production engine for high-quality robotic training data. Amit Goel, head of robotics and edge AI ecosystem at NVIDIA, emphasized the importance of this shift, stating, “By leveraging the NVIDIA Isaac simulation frameworks, Universal Robots is building a scalable engine for high-fidelity data capture and generation, providing the essential infrastructure to train the next generation of autonomous systems at scale.”
Generalist AI Demonstrates Real-World Performance
Generalist AI is showcasing its embodied foundation models, with two UR7e robots autonomously executing the complex smartphone packaging task, demonstrating dexterity, coordination, and contact-rich manipulation in a real-world environment. “Generalist is building embodied foundation models that deliver industry-leading dexterity and reliability,” said Pete Florence, co-founder and CEO of Generalist AI. “This demonstration on Universal Robots’ trusted industrial platform shows how physical commonsense can be translated into real-world capability, paving the way for deployment across industries at scale.”
Looking Ahead
Anders Beck will share his expertise on a panel at GTC titled “Beyond the Workcell: Scaling Robotics Workflows Across the Factory Floor” on Wednesday, March 18, at 11:00 a.m. The advancements showcased at GTC 2026 underscore Universal Robots’ position as a preferred platform for physical AI and signal a future where robots are more adaptable, intelligent, and integrated into manufacturing processes.