NVIDIA Advances Isaac Platform for General-Purpose Robotics With New Reference Humanoid, Frameworks, Simulation Pipelines and Data
nvidia ·
NVIDIA has significantly expanded its Isaac open robotics platform with new tools, datasets, and a reference humanoid robot to accelerate the development of general-purpose robots. The company introduced Agile ONE SN, a full-stack humanoid built on NVIDIA Jetson Thor and Isaac GR00T, designed to streamline the transition from simulation to physical deployment for researchers. Key software updates include Isaac Lab 3.0 for robot learning and Isaac OS, an open-source runtime framework that ensures deterministic real-time performance. To support data generation and training, NVIDIA released the SimReady Central Library with over 2,000 assets and the Human 1M Dataset, alongside three new open-source reference workflows for DGX Station. These advancements aim to reduce setup times and provide developers with integrated paths from data collection to real-world application. Major European robotics leaders, including ABB, KUKA, and Universal Robots, are actively integrating these technologies into their systems. Additionally, NVIDIA highlighted partnerships with entities like Stanford University and ETH Zurich to advance frontier humanoid research. CEO Jensen Huang emphasized that these developments bring the industry closer to a 'ChatGPT moment' for robotics by providing comprehensive models, data, and computing systems.
nvidia