China launches new AI model amid escalating US export control concerns; Arizona-made chips become focal point in technology rivalry
More than 1,300 employees and executives at OpenAI, Anthropic, Google, Meta and other AI companies signed a letter last week asking the U.S. government to help build tools to deliberately slow AI development. The same week, Chinese AI companies accelerated their efforts. On July 27, Beijing-based Moonshot AI released the full weights of Kimi K3, a 2.8-trillion-parameter model available for free download. While the system requires expensive, commercial-grade computing power to run, independent benchmarks place it within striking distance of the best American systems. Arizona has a direct stake in the dispute: some of the advanced Nvidia chips central to the controversy are manufactured at TSMC's plant in north Phoenix.
Lily McElwee, president and CEO of the Phoenix Committee on Foreign Relations and a U.S.-China relations expert, characterized the K3 release as putting nearly every unresolved question in the AI rivalry on the table simultaneously. McElwee noted that Beijing's embrace of open-weight models—systems whose parameters are published for anyone to download, modify and run instead of paying for access—aligns with a longer-standing strategy. "China has always framed itself as a provider of more freely accessible, low-cost technology to the rest of the world while the United States provides these proprietary systems," she said. The pitch targets developing countries and the Global South as a way to position American firms as profit-driven.
When asked whether the open-weight push resembles China's dumping of subsidized steel to undercut competitors, McElwee drew the parallel. "I'm not saying that Beijing is acting in good faith," she said. "I'm saying it's a marketing ploy." She pointed to Huawei and 5G as a precedent: Huawei became the world's largest supplier of telecom network equipment partly by undercutting Western rivals through state subsidies and cheap financing for developing-world customers. The U.S. blacklisted Huawei in 2019 and pressed allies to exclude its gear from their 5G networks. Despite this pressure, much of Europe, Africa and Latin America built 5G infrastructure with at least some Huawei equipment.
A Booz Allen Hamilton study published in June found that three of four Chinese models tested produced significantly more vulnerable code when users identified themselves as working for the U.S. government. McElwee framed the security concerns clearly: "Washington is rightly concerned about data vulnerabilities when it comes to U.S. customers using open-weight Chinese models, because the output of those models is generating data on U.S. citizens." She drew a contrast with existing policy: "We're banning Chinese electric vehicles and other Chinese technology products. But if we're not banning AI systems, which are increasingly used to train those products and access consumer data in the United States, then we have a security vulnerability."
Kimi K3 has faced additional allegations. Michael Kratsios, director of the White House Office of Science and Technology Policy, claimed that Moonshot trained K3 using outputs from Anthropic's Claude model and employed restricted Nvidia chips routed through Thailand. Moonshot has not publicly responded to these accusations.
The Biden administration imposed the first significant restrictions on advanced AI chip exports to China—a core grievance for Beijing, according to McElwee, because chips represent the critical bottleneck in the AI race. The Trump administration eased some controls in January, moving Nvidia's H200 from near-automatic denial to case-by-case review, though the most advanced Blackwell processors remain restricted. McElwee characterized the practical impact as minimal: "There's not a blanket restriction on it anymore. It's a case-by-case basis. But to my knowledge, there have been no cases that are approved." She noted that China is acquiring chips anyway through Southeast Asian partners, stockpiling, and recruiting the scientists who design them—areas where enforcement should focus.
Washington has also taken action on imports, with the FCC banning new imports of foreign-made humanoid robots, robot dogs and power inverters on national security grounds last week. The administration has been considering restrictions on Chinese open-weight models, though several outlets reported that the White House may exempt them from its review regime.
McElwee expressed skepticism that restrictions on open-weight models would be effective, given that they are freely downloadable by design. Instead, she argued that controls should target the tangible advantage the U.S. holds: "Let's take advantage of our strong points in this AI race, which are advanced Nvidia chips, which are produced right here in Arizona."