Trend AI competition enters a second round centered on physical AI in factories

Summary

The author describes the AI race as entering a second round, shifting from a contest over stronger algorithms and large language models to one over embodying intelligence in physical space. Physical AI combines AI with sensors, actuators and robots so that systems can perceive, decide and act in the real world. Its core technologies include computer vision and multimodal perception, reinforcement-learning-based decisions, simulation-based virtual training and edge computing for real-time inference. Unlike software AI, it must operate amid the uncertainty and unpredictability of the physical world. Industry leaders now identify the manufacturing floor, rather than household service robots, as the real battleground. Autonomous manufacturing, in which machines perceive, judge and optimize processes without human intervention, is expected to become the core of industrial competitiveness.