AI investment pivots to embodied and physical AI as funds track capital spending

Analyst Serenity reports that Chinese primary‑market AI financing is rapidly moving away from baseline large‑language models toward embodied intelligence, physical AI and world‑model research. While funding for basic LLMs totals about $23.6 billion, capital for embodied and physical AI reaches roughly $13.4 billion and world‑model projects are now the dominant investment consensus, even though no pure‑play companies have yet emerged. Serenity suggests the United States may follow the same pattern, with capital concentrating on firms such as Anthropic and OpenAI.

Fund manager Stephen Yiu of Blue Whale Growth Fund advises investors to mirror AI‑related capital expenditures, concentrating on semiconductor‑equipment makers and hardware producers that directly benefit from AI deployment. The fund’s top holdings include Applied Materials, BE Semiconductor Industries, SK Hynix, Lam Research, Lumentum, Nvidia and SanDisk, while it has exited most software stocks and holds only Alphabet among hyperscalers. Yiu highlights long‑term demand for memory and AI data‑center infrastructure as key growth drivers.