To Improve AI Training, NVIDIA Releases the Open-Source Isaac GR00T N1

NVIDIA Isaac GR00T N1, the first open, completely customisable foundation model for generalised humanoid reasoning and skills, is part of a portfolio of technologies introduced by NVIDIA on 18 March to accelerate the creation of humanoid robots. Other technologies include open-source physics engines like Newton, which is being developed alongside Google DeepMind and Disney Research specifically for the development of robots, and simulation frameworks and blueprints like the NVIDIA Isaac GR00T Blueprint for creating synthetic data. The GR00T N1, which is currently available, is the first of a series of fully customisable models that NVIDIA will pretrain and make available to robotics developers globally. This will hasten the transformation of industries that are facing a labour crisis that is predicted to affect over 50 million people worldwide. Nvidia CEO Jensen Huang announced in his keynote speech at GTC 2025 that the era of generalist robotics has arrived. With the Nvidia Isaac GR00T N1 and new frameworks for data collection and robot learning, robotics developers worldwide will be able to explore the next frontier in the AI era.

GR00T N1 Promotes the Humanoid Developer Community

The dual-system architecture of the GR00T N1 foundation model is based on ideas from human cognition. "System 1" is a quick-thinking action model that mimics human intuition or reactions. "System 2" is a meticulous, slow-thinking approach to decision-making. System 2 uses a vision language model to plan activities by reasoning about its surroundings and the commands it has been given. These designs are subsequently converted into precise, continuous robot motions by System 1. Both human demonstration data and a vast amount of synthetic data produced by the NVIDIA OmniverseTM platform are used to train System 1. In addition to performing multistep tasks that call for extensive context and combinations of broad skills, GR00T N1 can readily generalise across typical tasks, including gripping, manipulating objects with one or both arms, and transferring goods from one arm to another. These features are applicable to a variety of use cases, including inspection, packing, and material handling. For their particular humanoid robot or task, developers and researchers can post-train GR00T N1 using synthetic or real data.

Other Smart Features of GROOT

Huang used a post-trained strategy based on GR00T N1 to demonstrate 1X's humanoid robot automatically completing household cleaning tasks during his GTC keynote address. The autonomous capabilities of the robot are the outcome of 1X and NVIDIA's AI training partnership. The CEO of 1X Technologies, Bernt Børnich, said that learning and flexibility are key to the future of humanoids. He added that NVIDIA's GR00T N1 significantly improves robot thinking and skills as we build our own models. We completely deployed on NEO Gamma with very little post-training data, furthering our goal of building robots that are companions rather than merely tools that can help people in significant, incalculable ways. Other top humanoid developers throughout the world, like Boston Dynamics, Mentee Robotics, NEURA Robotics, and Agility Robotics, have early access to GR00T N1.