A German National Academy member and IEEE/AAAI Fellow explains why home robots are the hardest build in embodied AI: world models, touch, and a silent market. Full transcript included.

The day after Unitree's record IPO, founder Wang Xingxing took the WRC 2026 stage to define embodied AI's first quantifiable threshold: 80% of unfamiliar scenes, 80% of tasks.

All 100 officially selected WRC 2026 exhibits, mapped by category: humanoids, quadruped robots, dexterous hands, core components and three industry reads.

China's embodied AI boom traces a 40-year relay: state labs built the roots, startups sprinted to market, and Unitree's RMB 60.99B IPO set the valuation anchor. Five drivers, three leaders, one system.
