NVIDIA's latest robotics toolkit now lets AI agents collaborate with human developers to build and deploy robots faster, marking a significant shift in how physical AI applications are created. Isaac ROS 5.0, launched at the ROSCon conference in Toronto, integrates artificial intelligence agents directly into the robotics development workflow, automating repetitive tasks and simplifying navigation through complex codebases.

What Makes This Release Different for Robot Developers?

The new version supports ROS Lyrical and Ubuntu 24.04, offering developers a clear path to use the newest ROS platform while continuing to accelerate demanding robotics workloads with NVIDIA's accelerated computing. NVIDIA collaborated with the Open Source Robotics Alliance to integrate a standard data-handling interface into ROS Lyrical, which improves the efficiency of robotics software across diverse computing hardware, including graphics processing units (GPUs).

Isaac ROS 5.0 introduces several agent-ready features designed to speed up development. These include new NVIDIA Isaac skills for setup and manipulation that offer reusable workflows both developers and AI agents can use to accomplish robotics development tasks. Agent-ready documentation simplifies the process for AI agents to comprehend Isaac ROS tools and workflows, converting developer intentions into functional applications more rapidly.

How to Leverage AI Agent Skills in Your Robotics Projects

• FoundationStereo Fine-Tuning: An AI agent can help tailor a stereo perception model to your specific cameras, environment, and robotics application, enabling more precise perception for a given sensor configuration with greater ease.
• FoundationPose Inference: A foundational model for object pose estimation and tracking now includes an agent-ready inference library that boosts robots' ability to perceive and track object positions and orientations up to 5.5 times faster than previous approaches.
• Pick and Place Workflows: The common workflow of pick and place, which integrates detection, depth estimation, and pose output, is now available as a standalone, agent-ready skill, offering robot developers increased flexibility beyond the Isaac ROS ecosystem.

The robotics ecosystem is already adopting this agentic approach for development workflows. AgenticROS, an open-source initiative sponsored by 3D perception technology company RealSense, links Isaac ROS with NVIDIA Nemotron open models and NVIDIA NemoClaw blueprints, enabling AI agents to interact with ROS-based robots. This integration demonstrates how the toolkit is being deployed in real-world scenarios.

Which Companies Are Already Using This Technology?

Several organizations have begun integrating Isaac ROS 5.0 into their robotics platforms. Intrinsic's Open Machine Tending Solution serves as a reference application for computer numerical control machine tending, featuring built-in compatibility with NVIDIA FoundationPose for immediate object registration, tracking, and pose estimation. By using the FoundationPose perception pipeline, the solution enables robots to dynamically detect and handle parts, reducing the reliance on rigid, expensive physical fixtures and specialized systems integration.

Seeed Studio is integrating NVIDIA Isaac ROS with its reBot Arm, combining accelerated perception, spatial understanding, and motion planning on NVIDIA Jetson Thor. This integration offers developers a practical platform for building adaptable physical AI applications, ranging from object localization to collision-aware manipulation and autonomous pick and place.

Magna is using NVIDIA Isaac ROS as a modular, GPU-accelerated foundation for robotic perception, synchronized data collection, and NVIDIA Isaac GR00T model deployment. They are pairing this with Isaac Sim hardware-in-the-loop testing to accelerate the transition of intelligent automation from research to real-world manufacturing and mobility, while also reducing risks.

Ekumen, a Grid Dynamics Company, is employing GPU-accelerated Isaac ROS packages within existing ROS and Nav2 stacks to improve precision docking, 3D obstacle detection, visual localization, and real-time motion planning, with each application validated in Isaac Sim. Ekumen uses isaacroscumotion on a GPU to map a collision-free path for a warehouse arm in roughly 2 to 5 milliseconds, demonstrating the performance gains available through the toolkit.

The broader ROS community stands to benefit significantly from these advances. NVIDIA Isaac ROS delivers NVIDIA's accelerated computing, physical AI models, and production-ready libraries to the nearly 1.3 million ROS users, assisting developers in building high-performance robotics applications through the use of free, familiar, and open-source tools.

The applications developed by both humans and agents ultimately require deployment and execution directly on the robot. NVIDIA Jetson serves as a scalable computing platform designed to run the complete physical AI stack at the edge with real-time performance. It integrates ROS, accelerated perception and navigation, AI models, and application logic directly onto the robot, creating a unified ecosystem for robotics development and deployment.