NVIDIA Announces Major Collection of Open-Source Agent Tools and Skills for Physical AI

Taipei—NVIDIA GTC Taipei—June 1, 2026—NVIDIA today announced a major collection of open-source Physical AI skills and tools, enabling developers to turn complex robotics, autonomous vehicle (AV), vision AI, and industrial digital twin workflows into agent-executable tasks. This initiative aims to reduce the cost, time, and complexity of building Physical AI workflows at scale.

As AI agents transition toward orchestrating entire development tasks, Physical AI emerges as the next frontier. Available as part of the NVIDIA Agent Toolkit, these skills empower agents to leverage NVIDIA libraries, models, and frameworks to accelerate data generation, simulation, training, evaluation, and deployment pipelines in robotics, AVs, factories, and labs.

Jensen Huang, founder and CEO of NVIDIA, stated: "AI agents are revolutionizing software development, and their impact is now extending to Physical AI, transforming systems in transportation, manufacturing, healthcare, and robotics. When agents can directly use NVIDIA libraries, models, and frameworks, Physical AI development accelerates, enabling developers to build future robots, AVs, and industrial systems at an incredible pace."

The Physical AI stack is optimized by transforming libraries, models, and frameworks into agent-callable tools. This includes the NVIDIA Cosmos world foundation model for real-world reasoning, NVIDIA Omniverse libraries for simulation and digital twins, NVIDIA Isaac for robotics simulation and learning, NVIDIA Metropolis for vision AI, NVIDIA Alpamayo for autonomous driving, and the NVIDIA Jetson platform for edge AI. Developers can securely build and deploy autonomous agents using NVIDIA NemoClaw blueprints and the NVIDIA OpenShell runtime, which provides policy-based security and privacy governance.