NVIDIA Announces Isaac GR00T N1 — the World’s First Open Humanoid Robot Foundation Model — and Simulation Frameworks to Speed Robot Development

NVIDIA has unveiled the Isaac GR00T N1, a customizable humanoid robot model aimed at various applications in robotics. This model features a dual-system architecture designed to enhance reasoning and skills for robots, enabling them to complete tasks such as grasping and transferring objects more effectively.

In conjunction with the GR00T N1, NVIDIA has partnered with Google DeepMind and Disney Research to advance an open-source physics engine named Newton, which aims to facilitate more accurate robotic learning. The GR00T N1 makes use of this collaboration to improve robotic interaction with physical environments.

According to Jensen Huang, founder and CEO of NVIDIA, “The age of generalist robotics is here.” The GR00T N1 is offered to developers as a pretrained model, allowing them to adapt it for specific use cases with less reliance on extensive training data.

The introduction of the GR00T N1 comes at a time when industries are facing significant labor shortages, estimated at more than 50 million globally. This development reflects a push towards integrating robotics into everyday tasks across various sectors, from material handling to domestic chores.

The technologies announced also include simulation frameworks that are part of the NVIDIA Omniverse platform. These frameworks provide developers with tools to generate synthetic data, which has been shown to dramatically enhance the model's capabilities when combined with real-world data.

In a demonstration, a humanoid robot using GR00T N1 autonomously completed domestic tasks, showcasing the potential for real-world applications of the technology. Key industry participants, including Agility Robotics and Boston Dynamics, have also gained early access to the GR00T N1, marking a collaborative effort to expand humanoid robot functionalities.

Overall, NVIDIA's focus on robotics aims to evolve the developer community's approach to humanoid robots, enabling greater adaptability and advanced learning capabilities in real-world settings.