Wheels, Legs and Chips: Mapping the Physical AI Semiconductor Opportunity

The data center remains AI’s center of gravity, but a second front is opening in the physical world, on wheels and on legs. Autonomous vehicles, robotaxis and humanoid robots all run the same loop: sense, decide and act in real time, under tight power, cost and safety limits. This convergence is changing how the semiconductor value chain is built. This talk compares the technology requirements of three Physical AI segments: autonomous passenger vehicles, robotaxi fleets and humanoid robots. For each one, we look at what defines it: compute performance and efficiency, the sensor-fusion mix (camera, radar, LiDAR), memory bandwidth, connectivity, power management and functional safety.