Washington is racing to promote American artificial intelligence abroad at the same time China’s inexpensive and increasingly capable models are becoming harder to ignore.
Why this matters
Experts note that Chinese models do not need to surpass OpenAI or Anthropic in raw capability to reshape the global AI order — broad usefulness, availability and adoption can be enough.
Obstacles to U.S. AI diplomacy
Observers point to two main factors hampering U.S. efforts to export American AI:
- An erratic export controls approach that results in on-the-fly decisions about access to advanced models.
- Insufficient attention to China’s international push to distribute open-source AI models while it deploys AI at scale domestically in sectors such as manufacturing and health care.
Policy moves and industry effects
This week the State Department expanded the Pax Silica initiative, an effort to build a U.S.-led AI and chip supply-chain bloc and reduce dependence on Chinese technology. The move comes as the AI industry continues to deal with the fallout from the government’s decision to impose export controls on Anthropic’s newest models.
At the same time, reporting by Axios’ Sam Sabin highlights that Chinese AI is closing the capability gap and that Chinese models are considerably cheaper.
What experts are saying
Emily Weinstein, a former Commerce Department staffer who now works at the U.S.-China Economic and Security Review Commission, told Axios she sees “another example of the Huawei strategy in the context of open-source AI models.” Weinstein said China is often able to offer not only the models but also underlying or associated infrastructure at little or significantly lower cost.
She warned that widespread adoption of Chinese AI in the Global South could create a “Huawei model on steroids,” leaving countries dependent on Chinese infrastructure that may not be interoperable with U.S. systems.
Daniel Remler, a former State Department technology adviser now at the Center for a New American Security (CNAS), said the industry is “kind of frozen in place” after the Anthropic decision, waiting for something more coherent — a worrying sign as China moves quickly.
Saif Khan, former counselor for critical and emerging technologies to the secretary of commerce, suggested growing calls for AI sovereignty could lead much of the world, at least marginally, to prefer Chinese open-weight models over U.S. frontier models.
U.S. countermeasures and partners’ balancing acts
The Trump administration is attempting to counter this trend by rallying allies behind American AI. This week Jacob Helberg, undersecretary of state for economic affairs, announced that 35 countries signed the “Declaration on AI Opportunity” as part of Pax Silica.
Helberg also criticized digital sovereignty — the drive for domestic AI infrastructure — calling it “backward and counterproductive” and labeling it “synchronized mediocrity.”
Yet some U.S. partners are trying to balance supporting the administration’s vision with pursuing greater technological sovereignty. The European Union notably emphasized the importance of digital sovereignty after the Anthropic decision.
Omran Sharaf, the United Arab Emirates assistant foreign minister for advanced science and technology, told Axios during his Washington visit that the UAE’s goal is to achieve “strategic autonomy through international collaboration with trusted partners.” Sharaf said it is “too early to tell” what export-control decisions like the Anthropic directive will mean for that aim.
Bottom line
For the United States the next challenge is not only to stay ahead on frontier AI capabilities, but also to persuade the rest of the world to keep building on American AI platforms — even as Chinese, low-cost alternatives and partner countries’ desire for digital sovereignty reshape adoption choices.
