The current AI strategies of China and the US are complementary, as unlike the US, China isn't “AGI-pilled” yet, focusing on embodied AI and open source models
Thoughts on US-China Competition — The United States and China are often said to be in a “race” with one another with respect to artificial intelligence. X: @deanwball and @anton_d_leicht X: Dean W. Ball / @deanwball : Today, I write about US-China competition in AI and associated technologies as I see it. “Race” is a fairly poor analogy, in my view, for understanding the magnitude of what is underway. [image] Anton Leicht / @anton_d_leicht : I think Dean's read of Chinese strategy is right. If you take it seriously, it recontextualises some export control discussion, and complicates the ‘small yard, high fence’ doctrine. Frontier-focused chip controls make the most sense on an AGI-centric view of ‘the race’. Because [image]
Context & Ripple Effects
The argument extends coverage of Beijing steering its tech sector toward AI applications, including a policy push toward practical AI deployment, rather than treating frontier-model capability as the sole measure of progress.
It also fits a prior account of Chinese firms using open-source models to work around U.S. constraints and broaden development. The significance is that AI competition may be unfolding across different layers—models, physical deployment, and industrial capacity—rather than through one finish line.
First-order effects
- The framing elevates embodied AI and open models as distinct Chinese priorities, giving domestic developers and policymakers a rationale to judge progress by deployment and ecosystem reach as well as frontier-model performance.
- U.S. AGI-focused labs and investors face a clearer strategic contrast: their model-centric agenda is not necessarily mirrored by the Chinese approach described here.
Second-order effects
- Open release strategies can widen participation by developers and businesses that cannot depend on closed frontier systems, while making model diffusion harder to assess through a small set of leading labs.
- A deployment-led agenda increases the importance of hardware, manufacturing integration, and application channels; this complements earlier state-backed support for young Chinese AI companies rather than relying solely on flagship-model competition.
Third-order effects
- If the split persists, AI leadership will be measured across multiple interdependent capabilities—frontier models, open ecosystems, compute, and embodied deployment—weakening simple “race” narratives.
- Export-control policy may face a more complex target: restricting high-end compute can matter for frontier training while leaving open-model iteration and application deployment as separate competitive paths.
The trend: U.S.–China AI competition is becoming a contest between different technology stacks and routes to commercialization, not merely a single race toward AGI.