Physical AI starts with edge AI: models that perceive and decide locally, agents that reason and act without waiting on the cloud, and eventually robots and autonomous systems.
The keynote set out six design requirements for a distributed AI factory, from centralised AI governance to low-latency local inference, and a reference architecture that runs micro AI factories in stores and on factory floors, governed from a central AI factory. Use cases spanned real-time video analytics, visual quality inspection, predictive maintenance and a virtual store manager: an agentic assistant at the edge of every store that answers most questions locally, even offline.
Later the same day: the Neo Clouds vs Hyperscalers panel.
Photos · 6






