Robotics deep dive: US VS CHINA, THE FUNDING BATTLE FOR AI Where does the AI money go? Split it by country and by what it funds, and a clean pattern falls out. The United States pours it into language. In the first half of 2026 OpenAI and Anthropic alone raised about 217 billion dollars, 43% of all venture funding on Earth (Crunchbase). Its bet on the physical world, world models and robot foundation models, was about 6 billion in the same window (Forbes, New Market Pitch). Roughly 30 to 1, language over the physical world. China spreads the bet. Its leading model labs raised on the order of 10 billion in 2026 (BigGo), still an order of magnitude behind the US on language. But its embodied-AI and world-model sector pulled in about 6.4 billion in the first half of 2026, RMB 46 billion disclosed across 226 companies (TMTPost, TechBuzzChina). Under 2 to 1. China backs the physical world almost as hard as it backs language. And the engineer who builds it? An American AI engineer averages about 200,000 dollars a year in cash pay; a Chinese one about 80,000, roughly RMB 550,000 (Built In, Glassdoor, ERI, Chinese recruiter reports). Top talent in both countries earns several times that, and this is cash, not equity.