Teams running NVIDIA Isaac ROS 5.0 can deploy agentic robotics workflows to build physical AI systems across standard industrial platforms.
Announced at the ROSCon gathering in Toronto, the open-source software collection connects autonomous AI agents directly to development cycles. The software targets the nearly 1.3 million engineers who rely on the Open Robotics framework to program autonomous machinery.
“NVIDIA Isaac ROS brings NVIDIA accelerated computing, physical AI models and production-ready libraries to the nearly 1.3 million ROS users, helping developers build high-performance robotics applications using free, familiar, open-source tools,” explained Katie Washabaugh, Product Marketing Manager of Robotics Simulation at NVIDIA.
Faster FoundationPose inference and stereo calibration
The software update introduces native support for ROS Lyrical running on Ubuntu 24.04. Working alongside the Open Source Robotics Alliance, NVIDIA contributed a standard data-handling interface to the core Lyrical codebase. This interface coordinates data pipelines across varied compute hardware, using CUDA as a reference implementation for graphic processors.
Developers can assign repetitive setup and manipulation sequences directly to automated agents. Specialised agent-facing documentation translates natural technical intent into functional software packages. For computer vision teams, a new FoundationStereo calibration skill delegates camera tuning to an AI agent, adjusting sensory depth parameters to match specific industrial facilities.
Object tracking gains raw execution speed. The FoundationPose estimation library now delivers pose and orientation inferences up to 5.5 times faster through an agent-ready runtime. Engineers can extract pick-and-place operations as a modular skill that links object detection, depth calculation, and coordinate outputs.
Packaging specialist Prefix.dev aligned its Pixi package manager with the stack, letting teams synchronise CUDA and ROS environments. Telemetry provider Foxglove tied its desktop and browser tools into the release, giving technicians live diagnostics across nvblox surface meshes, rosbags, and three-dimensional topics.
Industrial deployment across Jetson edge hardware
Once compiled on workstations, the physical AI stack executes on local machines via NVIDIA Jetson edge silicon. The software runs across varied edge profiles, spanning entry-level Jetson Orin Nano boards to high-performance Jetson Thor boards.
Key industrial manufacturers have begun deploying the architecture across physical production lines. Universal Robots built Isaac ROS into its AI Accelerator kit. The integration allows robotic arms on live factory floors to compensate for misaligned parts, removing the requirement for expensive, stationary alignment fixtures.
Software group Intrinsic integrated FoundationPose into its Intrinsic Core suite for computer numerical control machine tending, allowing production arms to identify and grip raw workpieces dynamically.
Automotive supplier Magna paired the perception modules with Isaac GR00T foundation models to collect synchronised sensory data, verifying plant automation routines through hardware-in-the-loop tests in Isaac Sim.
In high-speed warehousing operations, engineering firm Ekumen benchmarked the ‘isaac_ros_cumotion’ package on local GPUs. The team generated collision-free robotic arm paths within two to five milliseconds, pairing the calculation engine with standard Nav2 packages to handle facility docking and visual localisation.
Isaac ROS 5.0 is available immediately as an open-source download without licensing fees.
Learn more about physical AI during the Physical AI Expo held in Amsterdam, London, and North America.
See also: QNX and Sift link industrial edge telemetry directly to SQL
Want to learn more about the IoT from industry leaders? Check out IoT Tech Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including AI & Big Data Expo and the Cyber Security Expo. Click here for more information.
IoT News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.
