NVIDIA Releases Early Developer Previews of Isaac Sim 5.0 and Isaac Lab 2.2, Enhancing Robotics Simulation and Training Capabilities

NVIDIA has announced the early developer previews of NVIDIA Isaac Sim and NVIDIA Isaac Lab, robust frameworks for robotics simulation and learning. These updates, now available on GitHub, are aimed at accelerating the development, training, and testing of AI-powered robots in realistic physics-based environments. What’s New in Isaac Sim Isaac Sim is a powerful reference application built on NVIDIA Omniverse, designed to help users create, simulate, and validate AI-driven robots. The latest release, Isaac Sim 5.0, includes several significant enhancements: Open Source Availability Isaac Sim 5.0 is now open source, allowing developers and researchers greater flexibility and control over their projects. Advanced Synthetic Data-Generation Pipelines The update introduces new extensions that significantly expand synthetic data generation capabilities, crucial for training, testing, and validating AI-powered robots. MobilityGen: This tool generates diverse, physics-based data and perception model training data, including occupancy maps, robot states, poses, velocities, and images. Grasp Data Generation Tutorial: A new workflow automates the generation of grasp candidates, simulates each grasp attempt, and records success metrics, enhancing the training and evaluation process. Omniverse Replicator Optimizations: The new writer for NVIDIA Omniverse Replicator is specifically optimized for NVIDIA Cosmos Transfer input. It simplifies the generation and export of high-quality synthetic data, supporting both standalone and Script Editor workflows, and can be seamlessly integrated into existing Isaac Sim synthetic data generation scripts. Large-Scale Physical Environment Workflows: Isaac Sim 5.0 now supports synthetic data generation for extensive physical environments. Features like Incident Simulation trigger diverse real-world scenarios, while Caption Generation automatically annotates incident data, enriching training datasets and improving scene understanding for Vision AI models. Existing Actor and Object Simulation workflows have also been refined for better synthetic data quality and efficiency. New Robot Models and Import Tools Isaac Sim 5.0 introduces enhanced robot models and streamlined import tools, making simulation setup faster, more consistent, and more reflective of real-world behavior. Improved Sensor Simulation The latest version brings significant improvements in sensor simulation, providing users with greater realism and control. This enhancement makes it easier to define and test sensor models, ensuring that robots can operate more effectively in various environments. Standardized ROS 2 Interfaces and ZMQ Bridge Isaac Sim 5.0 fully supports ROS 2 Jazzy Jalisco, offering compatibility with standardized ROS 2 simulation interfaces. An updated MoveIt 2 Tutorial helps with motion planning workflows, and a new ZeroMQ bridge facilitates communication between Isaac Sim and external systems. This bridge also serves as a reference for developing custom middleware solutions. What’s New in Isaac Lab Isaac Lab is an open-source framework focused on training and evaluating robot learning policies. The recent updates to Isaac Lab 2.2 include: Enhanced Training and Evaluation Capabilities: These updates streamline the process of training and validating robot learning policies, making it easier for developers to deploy AI-driven robots in complex tasks. Getting Started To begin exploring these new features, developers can download the early versions of Isaac Sim 5.0 and Isaac Lab 2.2 from GitHub. Detailed feature guides are available for those who want to dive deeper into the capabilities. For a broader overview, watch the NVIDIA GTC Paris keynote by NVIDIA founder and CEO Jensen Huang at VivaTech 2025. Additional insights can be found in the GTC Paris session recordings. Stay informed by subscribing to the NVIDIA Robotics newsletter and following NVIDIA Robotics on LinkedIn, Instagram, X (formerly Twitter), and Facebook. Explore comprehensive documentation and educational content on NVIDIA’s YouTube channels, and engage with the community on the NVIDIA Developer Robotics forum. To kickstart your robotics development, consider enrolling in NVIDIA’s free Robotics Fundamentals courses. NVIDIA also provides a rich library of Isaac modules and AI models to help you build advanced physical AI systems. With these tools, developers can push the boundaries of robotics and AI integration, creating more capable and reliable autonomous machines.