NVIDIA Releases Isaac ROS 5.0 With Agentic Tools for Robotics
NVIDIA unveiled Isaac ROS 5.0 at ROSCon in Toronto. The release adds ROS Lyrical and Ubuntu 24.04 support, agentic workflows, and a FoundationPose inference library that estimates and tracks object poses up to 5.5x faster.
NVIDIA has released Isaac ROS 5.0, a collection of GPU-accelerated packages for developing robots on the open ROS framework. The release was announced at ROSCon in Toronto, Canada.
The update adds agentic workflows, support for ROS Lyrical and Ubuntu 24.04, and tools for building, customizing and deploying robotics applications.
Agentic tools for developers
Isaac ROS 5.0 includes reusable skills for setup and manipulation. Agent-ready documentation is designed to help AI agents understand Isaac ROS tools and workflows and turn developer intent into working applications.
A new FoundationStereo fine-tuning skill lets an agent adapt a stereo perception model to a developer’s cameras, environment and robotics application. The FoundationPose inference library for estimating and tracking object position and orientation operates up to 5.5x faster.
Pick and place is also available as a standalone agent-ready skill, combining object detection, depth estimation and pose output.
ROS Lyrical compatibility
NVIDIA worked with the Open Source Robotics Alliance to contribute a standard data-handling interface to ROS Lyrical. It is intended to help robotics software work across different computing hardware, including GPUs.
The interface is available to the ROS community, with CUDA providing a working example of GPU acceleration. NVIDIA says Isaac ROS serves nearly 1.3 million ROS users.
Companies expand the ecosystem
AgenticROS, an open source project sponsored by RealSense, connects Isaac ROS with NVIDIA Nemotron open models and NVIDIA NemoClaw blueprints so agents can interact with ROS-based robots.
NVIDIA also cites Intrinsic’s Open Machine Tending Solution, Seeed Studio’s reBot Arm, and Isaac ROS deployments by Magna, Flexiv, Mentee Robotics, Universal Robots, ROBOTIS, FieldAI and Noble Machines.
Ekumen says it uses isaac_ros_cumotion on a GPU to map a collision-free path for a warehouse arm in approximately 2–5 milliseconds.
From Jetson Orin Nano to Jetson Thor
Isaac ROS 5.0 supports scalable computing from NVIDIA Jetson Orin Nano to high-performance Jetson Thor devices. NVIDIA positions Jetson as a platform for running the physical AI stack on robots, combining ROS, accelerated perception and navigation, AI models and application logic.
The release is available now.
What we know
- Isaac ROS 5.0 adds agentic workflows for building robotics applications.
- The release supports ROS Lyrical and Ubuntu 24.04.
- FoundationPose tracks object position and orientation up to 5.5x faster.
- Isaac ROS 5.0 runs across platforms from Jetson Orin Nano to Jetson Thor.
- Ekumen maps collision-free paths for a warehouse arm in about 2–5 milliseconds.
Trust: Primary source
- The story cites a primary statement. It establishes what its author said, but does not guarantee every claim is true.
- Sources: NVIDIA Newsroom.
COMMUNITY
Discussion
Sign in to join the discussion.
Loading comments…