The administration of U.S. President Donald Trump has revived discussions of measures that could make it harder to use Chinese open-weight AI models, Axios reported, citing people familiar with the matter.
No official decisions have been made. The White House and the U.S. Department of Commerce did not respond to requests for comment. According to the report, the Kimi K3 release from China’s Moonshot AI gave new momentum to the discussions.
Officials weighed several measures
In 2025, the Department of Commerce discussed adding a number of Chinese AI companies to the export restrictions list. U.S. organizations typically need the department’s approval to engage with entities on the list.
The National Security Agency and the White House Office of the National Cyber Director also considered issuing a warning about potential threats from Chinese AI developers.
The White House studied an executive order that could hold U.S. companies responsible for security violations when using such models. The initiative was shelved over concerns it would slow the industry’s development.
Current options under discussion include restrictions in government procurement, adding developers to the list, and public risk advisories.
Dean Ball’s forecast
Dean Ball served as a senior policy adviser on AI and emerging technologies at the White House Office of Science and Technology Policy. He helped prepare the Trump administration’s artificial intelligence action plan.
After leaving the White House, Ball joined the Foundation for American Innovation (FAI) as a senior fellow. On July 6, he joined OpenAI to lead Strategic Futures, which focuses on long-term policy for advanced AI systems. Ball retained his status as a nonresident senior fellow at FAI.
After Kimi K3’s release, he called it a high-quality model and said its results cannot be explained solely by training on other systems’ outputs.
Some observations on Kimi:
1. It’s a very good model! I don’t think its performance can be explained away by distillation or anything like that. In agentic coding sessions, it seems pretty much on par with the best public models of Q1 2026. In my fairly limited use, it also…
— Dean W. Ball (@deanwball) July 17, 2026
“I would assume that at some point the Trump administration will realize that their best strategy here would be to create significant regulatory risk for using open-source Chinese models,” he wrote.
Among possible tools, Ball mentioned agency warnings about hidden vulnerabilities. He later clarified that he was predicting such a scenario rather than recommending it.
White House AI adviser David Sacks called the weaponization of regulatory uncertainty as a competitive tool unacceptable.
I’m not sure whether Dean Ball is confessing to a regulatory capture strategy or simply predicting this will happen (he now says the latter). Either way, the weaponization of regulatory uncertainty as a competitive tool should be completely unacceptable.
He argues there’s no… https://t.co/ZxIFabAJZW pic.twitter.com/0eLk6ltU0F
— David Sacks (@DavidSacks) July 19, 2026
In his view, leading developers of closed systems — OpenAI and Anthropic — already form a duopoly in AI model revenue and may be interested in the government limiting competition from open solutions.
Awaiting publication
Moonshot AI released Kimi K3 on July 16. The model is currently available through apps and a developer API. The company promised to publish the full set of parameters by July 27 under a modified MIT license. Until the files are released, the model cannot be deployed on users’ own infrastructure.
U.S. agencies have not presented evidence of hidden features or vulnerabilities in Kimi K3 itself. Moonshot AI has also not been added to the export restrictions list because of this model.
At the same time, China is considering its own bans on transferring advanced AI technologies abroad. According to Reuters, the Ministry of Commerce of the PRC discussed with Alibaba, ByteDance and Zhipu possible controls on exporting training data and on foreign users downloading model weights.
Starting July 10, Alibaba banned employees from using Anthropic products in work environments and suggested switching to its own Qoder platform. According to Reuters, the reason was security-related issues.
In February, Anthropic accused three Chinese AI startups — DeepSeek, Moonshot and MiniMax — of a large-scale campaign to use Claude to improve their own models.
In June, Anthropic reiterated its claims against Alibaba and its AI unit Qwen in a letter to Senate Banking Committee Chair Tim Scott and senior Democrat Elizabeth Warren. The developers said it was the largest known attack of its kind against it, tied to attempts to accelerate the development of Chinese models.
