At the ROSCon 2025 conference in Singapore, NVIDIA announced a major step forward in its support of open-source robotics by contributing GPU-aware abstractions to ROS 2 enabling the framework to efficiently manage CPUs, integrated GPUs and discrete GPUs across robot architectures and by backing the Open Source Robotics Alliance’s (OSRA) newly formed Physical AI Special Interest Group which focuses on real-time robot control, accelerated AI processing and advanced robotics tooling. In tandem, NVIDIA open-sourced its Greenwave Monitor tool to help developers identify performance bottlenecks in robot systems.
Also Read: TwinThread Unveils Advisor, a GenAI Assistant for Operations
In addition to infrastructure contributions, NVIDIA introduced the latest release of Isaac ROS 4.0, a collection of GPU-accelerated libraries and AI models compatible with the Jetson Thor platform that enable manipulation and mobility workflows for physical AI and robotics deployment. Several ecosystem partners including AgileX Robotics, Canonical, Ekumen Labs and KABAM Robotics were cited as already leveraging these technologies for simulation and real-world systems. This suite of developments underscores NVIDIA’s commitment to making ROS 2 “the open, high-performance framework of choice for real-world robotic applications” by enabling higher-speed performance, future-proofing hardware abstraction and accelerating an open-source community push into next-generation autonomous systems.
