At the 2026 Zhangjiang Embodied AI Supply Chain Conference, Professor Wang Tianmiao, Honorary Director of the Robotics Institute at Beihang University, delivered a comprehensive keynote on the embodied AI industry.
Why Embodied AI Is Suddenly So Hot Wang attributes the surging interest to a fundamental shift: AI is moving from the digital world into the physical world. Robots previously understood as automated equipment now possess increasingly capable 'intelligent brains' powered by large models. Meanwhile, mature supply chains from new energy vehicles, autonomous driving, and robotics provide the hardware foundation.
Capital markets are seeking larger physical-world landing scenarios — embodied AI connects to manufacturing, logistics, commercial services, special operations, and home/elderly care, representing longer industrial chains and larger real-economy spaces.
The 'Ice and Fire' Reality Wang warns that embodied AI exhibits a classic 'ice and fire' coexistence: surging investment, favorable policies, and high market expectations on one side, versus limited truly mature commercial scenarios on the other. Many companies remain in technical verification and small-scale delivery stages.
The Four Elements of Scaling Law Wang identifies four interdependent factors: large model parameters and algorithms, human data collection scale, edge chip inference speed, and physical hardware structure. Only when all four advance together can embodied AI models enter real-world tasks.
The Second Half: High Heat, High Investment, High Trial-and-Error The industry is simultaneously exploring VLA models, world models, edge chips, dexterous hands, and new transmission mechanisms. The final model will likely be a fusion of multiple capabilities.
Wang predicts 2027 and 2030 as two critical convergence milestones. Embodied AI will first enter logistics, cleaning, retail, and commercial scenarios, then industrial environments, and eventually home and elderly care.
No Single Winner Wang sees an ecosystem of full-stack robot companies, cloud AI providers, high-value component suppliers, and operational service platforms. Success depends not on concept grandeur but on clear scenarios, real customers, cost advantages, technical moats, and sustained delivery ability.
