What does it mean to win the AI race?

The article distinguishes between two national strategies for achieving AI leadership. For the United States, success is often defined by being first in advanced model capabilities and chips, accompanied by tight control over access. Treasury Secretary Scott Bessent articulated this view, stating that a Chinese lead in AI would overshadow all other concerns, highlighting the focus on technological supremacy.

China's 'AI Plus' plan, in contrast, measures success by the widespread adoption of AI across its economy, integrating it into sectors like manufacturing, energy grids, and public services. Beijing aims for its open models to provide a low-cost foundation for developing economies internationally. This approach suggests that a functional model adopted widely throughout an economy could hold more practical influence than a superior model with limited deployment.

The effectiveness of these differing strategies is reflected in recent findings. Stanford's 2026 AI Index indicated a narrow 2.7% capability gap between the leading American and Chinese models in March. However, regarding generative AI use, the United States ranked 24th globally at 28.3%. China's emphasis on open models and competitive pricing appears to be yielding broader adoption, with Chinese models accounting for 41% of Hugging Face downloads in the year to February, compared to 36.5% for American models. This wider use fosters a feedback loop, as developers adapt these models to diverse languages and industries, potentially embedding Chinese foundations into the global software stack.