Physical AI & Robotics

Physical AI connects physical systems like cameras, sensors, digital twins, and robots with AI models and systems to perceive, reason, and act in real environments.
Physics-based simulations are essential for training autonomous systems to interact safely between humans and machines.
AI Factories play a crucial role in orchestrating data, computation, and workflows necessary for Physical AI.

The first talk in this session explains the role of simulation in connecting the physical and digital worlds.

The second talk breaks down the cognitive stack for generalist robots, exploring how real-time constraints - like strict latency, reliability, and energy budgets - impact the computation and training of large multi-modal models.

This session provides a hype-free view of where this field actually stands and where open-world physical autonomy is realistically heading.