NVIDIA has announced advancements the Robot Operating System (ROS) 2 framework and physical AI robotics development at ROSCon 2025 in Singapore.
It is supporting the Open Source Robotics Alliance’s new Physical AI Special Interest Group, which focuses on real-time robot control, AI acceleration and improved robotics tools for autonomous behaviours.
NVIDIA has contributed GPU-aware abstractions directly into ROS 2 to enable efficient management of CPUs and GPUs to ensure consistent high-speed performance and future-proofing ROS for rapid hardware innovations.
The company has launched the Isaac ROS 4.0 set of ROS-compatible, GPU-accelerated libraries and AI models available on the Jetson Thor platform. This toolkit supports developers in deploying physical AI for robot manipulation and mobility with enhanced performance.
Also announced were collaborations with industry partners such as AgileX Robotics, Canonical, Ekumen Labs, Intrinsic, KABAM Robotics, Open Navigation, Robotec.ai, ROBOTIS, and Stereolabs, all leveraging NVIDIA’s Jetson modules, Isaac Sim simulation framework, and AI models to push the boundaries of robotics autonomy, simulation, security, and navigation.
At ROSCon, Open Navigation demonstrated advanced autonomous navigation using NVIDIA technologies, while Robotec.ai helped establish a new ROS simulation standard integrated with Isaac Sim.
The announcements highlight the pivotal role NVIDIA is playing in advancing robotics development by seamlessly integrating cutting-edge AI, GPU acceleration and open-source collaborations within the ROS 2 ecosystem. These let developers and industry partners create more intelligent, efficient and autonomous robotic systems on a global scale.
Entelechy Asia 