NVIDIA has unveiled a new suite of robotics AI models and open-source simulation tools designed to accelerate development in humanoid and industrial robotics. The company is opening up its Newton Physics Engine, Isaac GR00T foundation model, and Cosmos simulation libraries to developers, aiming to provide a comprehensive ecosystem for building more intelligent, adaptable robot AI. These updates integrate GPU-accelerated simulation, open foundation models, and scalable AI infrastructure, forming a complete framework for robotics research and deployment.
Simulation is a key factor in robotics development, particularly for humanoid robots, where tasks involving balance, joint manipulation, and interaction with complex environments often exceed the capabilities of conventional physics engines. NVIDIA’s Newton Physics Engine addresses these challenges by providing GPU-accelerated, open-source simulations that allow developers to model advanced scenarios such as walking on uneven terrain, handling delicate objects, or manipulating tools with precision. Co-developed with Google DeepMind and Disney Research and managed by Linux Foundation, Newton is already being tested by institutions including ETH Zurich, Technical University of Munich, Peking University, and robotics company Lightwheel.
Alongside Newton, NVIDIA introduced Isaac GR00T N1.6, an open foundation model for robotics that integrates Cosmos Reason, a vision-language-action system for physical AI. Cosmos Reason helps robots interpret abstract or ambiguous instructions and convert them into actionable step-by-step plans while applying common sense reasoning. This model is complemented by Cosmos world foundation models, including Cosmos Predict 2.5 for longer video generation and multi-camera perspectives, and Cosmos Transfer 2.5 for synthesizing photorealistic training datasets efficiently. These tools enable humanoid robots to perform complex tasks such as operating doors, manipulating objects with two hands, and navigating dynamic environments more effectively. Companies including Franka Robotics, Techman Robot, LG Electronics, and Neura Robotics are evaluating these models for research and commercial applications.
To support the growing demands of AI-driven robotics, NVIDIA has also introduced infrastructure tailored for robotics workloads. GB200 NVL72 offers rack-scale processing with 72 Blackwell GPUs, RTX PRO Servers provide a unified platform for training, simulation, and inference, and Jetson Thor delivers edge AI capabilities directly on robots. These systems are already in use by organizations such as Figure AI, Skild AI, and Unitree, enabling multi-AI workflows in real-world environments. Adoption within the research community is also significant, with nearly half of the papers at the latest Conference on Robot Learning citing NVIDIA technologies. Leading universities, including Stanford, Carnegie Mellon, and the National University of Singapore, are leveraging these tools to advance humanoid and industrial robotics research.
By making Newton, GR00T, and Cosmos openly accessible, NVIDIA is establishing a cohesive platform for robotics innovation, bridging research, simulation, and deployment. The combination of high-fidelity physics simulation, reasoning-focused foundation models, and robust infrastructure positions the company as a key enabler for the next generation of intelligent robots in industrial, service, and research applications.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
