The era of physical AI: From simulation and cloud-based training to on-device deployment, embodied intelligent robots are moving towards efficient implementation.

AI technology is rapidly expanding from the digital world to the physical world, propelling intelligent systems into the era of physical AI. With the widespread adoption of systems like robots and autonomous driving, user demand for human-like perception, autonomous decision-making, and precise execution capabilities continues to increase. IDC predicts that by 2026, AI models, vision systems, and edge computing will have achieved breakthroughs, and the number of application scenarios will triple. The global robotics market is expected to exceed $400 billion by 2029, with China accounting for nearly half of the market share. NVIDIA and other manufacturers are accelerating the development of physical AI through a "simulation-first" approach, cloud-based training, and device-side deployment.