The AI Race Is About WAY MORE than LLMs, and Now It's PHYSICAL

Why It Matters

If China accelerates in physical AI and robotics while the US concentrates on compute-heavy LLMs, global leadership in applied AI capabilities and economic competitiveness could shift toward more embodied, real-world AI winners. That has implications for industrial strategy, defense, and where private and public investment should be focused.

Summary

The speaker warns that focusing solely on generative AI and large language models is short-sighted, arguing the AI race is multimodal and increasingly physical. Notably, US private firms outspend Chinese firms on compute by 12-to-1, yet China invests about 42% more in robotics — a gap expected to widen. The discussion highlights “physical AI,” systems that perceive and interact with the real world rather than executing repetitive factory tasks, and suggests China may hold hardware and emerging software advantages in this domain. The shift toward robotics reframes competitive dynamics beyond cloud compute and LLMs.

Original Description

The most important AI story isn't about chatbots. It's about machines that can think, move, and act in the world, and China has a head start. Alice Han and James Kynge discuss, on China Decode.