AI for Robotics
The NVIDIA Isaac™ robotics platform includes a full suite of GPU-accelerated innovations in AI perception, manipulation, simulation, and software. You can now take your performance to the next level by optimizing for advancements in AI—including robot learning, LLMs, generative AI, and foundation models.
Mobile Robots
NVIDIA Isaac Perceptor's unmatched precision enables 3D perception for automated material-handling equipment and service robots, letting them operate safely and effectively in dynamic environments. NVIDIA Isaac Sim™ also provides software-in-the-loop development and testing using simulated sensors.
Robotic Manipulation
NVIDIA Isaac Lab, a lightweight app built on top of Isaac Sim, delivers robot learning with integrated physics for complex grasping tasks. Start with pre-trained models and synthetic data from NVIDIA Omniverse™ Replicator for robot training. Define the robot's task, validate its training, and deploy advanced robotic manipulation and grasping with NVIDIA Isaac Manipulator to reduce cycle times for re-training when tracking new objects.
Outdoor Equipment
NVIDIA Jetson™, with its exceptional efficiency, powers autonomous tractors, mowers, and other outdoor equipment by processing onboard sensor data and running real-time AI algorithms on the edge. Developers benefit from hardware-accelerated SDKs like Isaac ROS and the Jetson Generative AI Lab, which offer open-source generative AI models. NVIDIA Isaac Sim tests outdoor scenarios and supports the creation of robot foundational models.
Humanoids & Quadrupeds
For bipedal and quadruped robots, pretrained foundational models like NVIDIA Project GR00T (Generalist Robot 00 Technology) will significantly improve robot capabilities such as perception, control, and decision-making. These models use tools such as Isaac Lab for robot learning and NVIDIA OSMO for cloud workflow orchestration. NVIDIA Metropolis also helps robots survey a work environment for large-scale inspection tasks.