The New Frontier for Chinese Venture Capital: Physical AI and World Models

Recent analysis of China's private venture capital market reveals a significant shift in investment focus. Institutional investors are pouring substantial resources into two cutting-edge domains: Physical AI and World Models. This movement signals both the direction of technological evolution and investor confidence in the next generation of AI applications.

Mapping the Capital Distribution

The current investment landscape is segmented into several key areas:

  • Large Models/LLMs: Approximately $23.56 billion in total investment, remaining the largest sector by volume, though the focus of growth is evolving.
  • AI Infrastructure & Tech Stack: Around $15.74 billion, targeting foundational elements like computing power, semiconductors, and cloud platforms.
  • Embodied Intelligence/Physical AI: Roughly $13.36 billion, emerging as one of the fastest-growing sectors focused on AI's interaction with the physical world.
  • AIGC Applications: About $8.79 billion, with capital favoring mature use-cases and clear commercialization paths.

The Evolving Investment Thesis

Market observers suggest the early-stage funding window for pure foundational models may be largely closed. Capital is now consolidating in two distinct patterns.

Firstly, there's a pronounced concentration towards established industry leaders, a trend visible globally that may intensify. Secondly, World Models—AI systems capable of understanding and simulating complex real-world environments—are becoming a new theoretical and technical frontier, attracting capital seeking transformative innovation.

Where Capital Is Flowing Next

Ongoing AI investment continues to bolster two foundational pillars: infrastructure and the semiconductor supply chain. Simultaneously, a clear rotational signal is evident: capital is rapidly pivoting from purely software and algorithmic layers towards Physical AI and Embodied Intelligence.

This pivot underscores a growing investor interest in AI applications that require real-time interaction with the physical world—such as robotics, automation, smart manufacturing, and autonomous driving. It represents a crucial step for AI's journey from the digital realm into the tangible one.