The Trump administration has revived discussions on measures that could complicate the use of Chinese AI models with open weights. This was reported by Axios citing informed sources.

No official decisions have been made yet. The White House and the US Department of Commerce have not responded to requests from the publication. According to journalists, the recent release of Kimi K3 by Chinese company Moonshot AI has spurred renewed discussions.

Authorities Considered Several Measures

In 2025, the Department of Commerce discussed including several Chinese AI companies in a list of export restrictions. American organizations typically require permission from the department to interact with these companies.

The National Security Agency and the White House's Office of the National Cyber Director also considered issuing a warning about potential threats from Chinese AI developers.

The White House examined an order that could hold American companies accountable for security violations when using such models. However, these initiatives were abandoned due to concerns about stifling industry growth.

Currently, among the potential tools being discussed are restrictions on government procurement, inclusion of developers on a watchlist, and public warnings about risks.

Dean Ball's Forecast

Dean Ball previously served as a senior advisor on AI and new technologies in the White House Office of Science and Technology Policy. He was involved in crafting the Trump administration's AI action plan.

After leaving the White House, Ball joined the American Innovation Foundation (FAI) as a senior research fellow. On July 6, he joined OpenAI, where he leads the Strategic Futures team focused on long-term policy regarding advanced AI systems. Ball remains a senior research fellow at FAI.

Following the release of Kimi K3, he described it as a high-quality model, stating that its performance cannot be solely attributed to training on responses from other systems.

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 suggest that the Trump administration at some point realizes that their best strategy here would be to create a significant regulatory risk for using Chinese open-source models," he wrote.

Among the potential tools, Ball mentioned warnings from agencies about hidden vulnerabilities. Later, he clarified that he was predicting such a scenario rather than recommending it.

White House AI advisor David Sacks deemed the use 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

He believes that leading developers of closed systems—OpenAI and Anthropic—are already forming a duopoly in AI model revenues and may be interested in having the government limit competition from open solutions.

Awaiting Publication

Moonshot AI released Kimi K3 on July 16. The model is currently available through applications 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 private infrastructure.

US agencies have not provided evidence of hidden functions or vulnerabilities directly in Kimi K3. Moonshot AI is also not included in the export restrictions list due to this model.

At the same time, China is considering its own bans on the export of advanced AI technologies. According to Reuters, the Ministry of Commerce of China has discussed potential controls on the export of training data and the downloading of model weights by foreign users with Alibaba, ByteDance, and Zhipu.

Since July 10, Alibaba has prohibited employees from using Anthropic products in work environments and has suggested switching to its own platform, Qoder. According to Reuters, this decision was made due to security concerns.

In February, Anthropic accused three Chinese AI startups—DeepSeek, Moonshot, and MiniMax—of conducting a large-scale campaign to use Claude to enhance their own models.

In June, Anthropic reiterated its claims against Alibaba and its AI division Qwen in a letter to US Senate Banking Committee Chairman Tim Scott and senior Democrat Elizabeth Warren. The developers believe this is the largest known attack of its kind, aimed at accelerating the development of Chinese models.