Trump administration revives push to ban Chinese open-weight AI models like Kimi K3 and DeepSeek

The U.S. government may be moving forward with plans to ban leading Chinese AI models, following the release of Kimi K3 — a powerful AI model developed by Chinese startup Moonshot AI — last week.

According to an Axios report published July 20, the Trump administration is reigniting its push for a ban, citing cybersecurity concerns.

Open-Weight Models Driving Adoption

Chinese models such as DeepSeek and Kimi K3 are open-weight, meaning their trained model weights are published for public download. This allows enterprises to:

• Keep data in-house by self-hosting models on private infrastructure
• Reduce inference costs significantly

These characteristics have led to rising adoption by U.S. companies.

Previous Attempts and Renewed Efforts

Citing several sources close to the administration, the Axios report states that the government had previously made a series of attempts to curb the growth and expansion of Chinese models in the U.S. over cybersecurity concerns. Critics say such moves will stifle competition and innovation while encouraging monopolies.

Key prior measures included:

• The U.S. Department of Commerce last year considered adding multiple Chinese AI labs, including DeepSeek, to its "Entity List" — a trade blacklist maintained by the Bureau of Industry and Security (BIS) that restricts foreign entities from purchasing sensitive American hardware, software, or technology.
• U.S. officials considered a joint National Security Agency/Office of the National Cyber Director advisory to discourage the use of Chinese AI models.
• Officials drafted an executive order holding U.S. companies liable for security breaches involving hosted Chinese models.

While these measures were initially paused due to internal pushback regarding market impacts, they have been revived following the release of new Chinese open-weight models.

Criticism from Industry Figures

Critics of the potential ban — such as former White House adviser Sriram Krishnan and outside White House AI adviser David Sacks — say the move would negatively impact innovation while handing a monopoly to leading U.S. AI labs OpenAI and Anthropic. The report implies these labs may have a hand in pushing for the ban.

"We are at a critical inflection point in AI policy. The leading closed labs, already a duopoly in terms of AI model revenue, want the government to eliminate their open-source competition," wrote Sacks in an X post on Sunday.

Cost Advantages Driving Adoption

Chinese AI models are increasingly used by numerous American companies due to their relatively low cost and perceived matching capabilities with domestic alternatives.

The open-weight nature of models such as DeepSeek V4 and Kimi K3 — which allows companies to download models locally and host them on private servers — drives adoption by:

• Providing data privacy
• Slashing API costs compared to closed Western alternatives

However, self-hosting shifts the cost of GPUs, electricity, maintenance, networking, and model operations to the company, making it generally most economical for organizations with substantial and sustained AI usage.

Pricing Comparison

Chinese open-weight models price their APIs well below comparable U.S. systems:

| Model | Price per Million Output Tokens | |-------|--------------------------------| | DeepSeek-V4-Pro | $0.87 | | Anthropic's Claude Fable 5 | $50.00 |

This aggressive undercutting has caused a massive surge in developer adoption. Coinbase CEO Brian Armstrong noted that the company runs models like GLM-5.2 and Kimi in production, cutting their overall AI spending nearly in half even as actual token consumption spiked.

Enforcement Challenges

Despite rising adoption, the U.S. government cites cybersecurity concerns as a reason for a ban. However, the question of whether a ban on Chinese AI models can be practically enforced arises, as blocking open-weight technology presents significant technical and regulatory hurdles.

• For individuals or small companies wanting to use DeepSeek via website or app despite a U.S. block, a VPN works fine.
• However, limited app availability and payment restrictions remain effective barriers.

For enterprises that self-host rather than use the hosted app or API, enforcement becomes harder for several reasons:

• Unlike closed-source APIs that require data to leave a company's network, open-weight models exist as downloadable files mirrored across public repositories like Hugging Face and independent torrents, making them hard to fully recall once released.
• Once an American enterprise downloads the weights, it can run the model entirely offline inside a private, air-gapped data center, limiting U.S. regulatory reach.