Nvidia has unveiled a suite of new open technologies designed to speed up the development and deployment of advanced robotics, including humanoid systems. The announcement was made alongside updates to the company’s Isaac robotics platform.
At the center of the release is the Newton Physics Engine, an open-source, GPU-accelerated simulator codeveloped with Google DeepMind and Disney Research.
Now available in Nvidia Isaac Lab and managed by the Linux Foundation, Newton is built to handle the complex balance and motion challenges of humanoid robots.
It allows developers to simulate demanding real-world scenarios such as walking on gravel, handling fragile objects, or operating in strong currents, with results that can be reliably transferred to physical machines.
Nvidia also introduced the latest version of its Isaac GR00T open foundation model, which combines vision, language and reasoning to give robots more humanlike decision-making capabilities.
The GR00T N1.6 release integrates Nvidia’s Cosmos Reason model, enabling robots to break down ambiguous instructions and adapt to unfamiliar tasks using common sense and physics-based reasoning.
Alongside these tools, Nvidia announced updates to its Cosmos world foundation models, designed to generate diverse synthetic data for training physical AI systems at scale.
The company also highlighted a new workflow in Isaac Lab for teaching robots dexterous grasping, already tested by Boston Dynamics’ Atlas robot to improve manipulation skills.
Robot makers including Agility Robotics, Figure AI, Franka Robotics, Techman Robot and Boston Dynamics are among the companies adopting Nvidia’s Isaac and Omniverse technologies.
Rev Lebaredian, vice president of Omniverse and simulation technology at Nvidia, says: “Humanoids are the next frontier of physical AI, requiring the ability to reason, adapt and act safely in an unpredictable world.
“With these latest updates, developers now have the three computers to bring robots from research into everyday life – with Isaac GR00T serving as robot’s brains, Newton simulating their body and Nvidia Omniverse as their training ground.”
