Nvidia has announced that its open-source, pre-trained AI foundation model to power humanoid robots is now publicly available. The “generalist” model is called the Nvidia Isaac GR00T N1 or Groot N1 for short. It was introduced at the chip giant’s widely watched annual GTC conference in San Jose, California, US on Tuesday, March 18. In his GTC 2025 keynote address, Nvidia CEO Jensen Huang proclaimed that “the age of generalist robotics is here” “With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI,” he added. The AI foundational model was trained on existing datasets as well as synthetically generated data. Nvidia further said that it will be publicly releasing the simulation frameworks and blueprints for generating synthetic training data. The rise of generative AI has also spurred innovation in humanoid robotics with mass-produced, general-purpose robots that move similar to humans emerging as a realistic goal in the near future. While companies such as X1 and Figure are working on refining and scaling their humanoid prototypes, bigger players like Nvidia and Google have made bids to provide the underlying AI technology to power them. Recently, Google DeepMind launched Gemini Robotics, its suite of AI models designed to supply robots with accuracy and dexterity for performing complex tasks. According to a video posted by Nvidia on its YouTube channel, Groo N1 has a “dual-system architecture” inspired by human cognition for “thinking fast and slow”. The trillion-dollar company claimed that Groot N1’s slow-thinking system will enable humanoid robots with perception and reasoning skills to decide the right actions to take. The model’s fast-thinking system, on the other hand, will help to quickly translate the planned actions into robotic movements that include manipulating objects over multiple steps, as per Nvidia. Groot N1 is the end result of Nvidia’s Project Groot, which was announced by the company at last year’s GTC conference.
