Trump administration weighs curbs as businesses shift to cheaper Chinese AI models

The Trump administration has been signalling that it may move to restrict cutting-edge Chinese AI models, a step that could further strengthen the dominance of US companies such as OpenAI and Anthropic. Parts of the administration have previously explored imposing de facto restrictions on foreign open-source AI models.

The soaring demand for models such as Kimi K3 has reignited these discussions. Concerns stem from the growing adoption of Chinese open-source AI models by US companies, as they are significantly cheaper while offering capabilities comparable to those of Anthropic and OpenAI. The Commerce Department's consideration last year of adding several Chinese AI labs to its Entity List, which would effectively restrict US access to their technologies without a licence, further underscores the government's efforts to curb the use of Chinese AI models.

The White House has also reportedly considered an executive order that would allow US companies to host Chinese AI models only if they could guarantee security and accept liability in the event of a breach. Last year, the National Security Agency and the White House Office of the National Cyber Director also discussed issuing an advisory highlighting potential threats posed by Chinese AI labs and discouraging US companies from using their technologies.

More from Tech

A complete ban or a push to highlight potential backdoors?

Rather than imposing an outright ban, an Axios report said the administration has discussed highlighting the potential security risks, governance concerns and possible backdoors in Chinese AI models. According to the report, this approach would encourage the development of a stronger and more innovative US open-source AI ecosystem instead of relying solely on restrictions.

Businesses shift to cheaper Chinese models

A growing number of businesses have already begun adopting Chinese AI models. According to the Financial Times, startup Lindy has reportedly dropped Anthropic's models in favour of DeepSeek V4. Beyond startups, companies including Airbnb and Siemens are experimenting with integrating AI models from Chinese firms such as Alibaba and DeepSeek into their operations to reduce AI costs.

Chinese AI models are becoming increasingly attractive because they combine advanced capabilities with significantly lower costs while remaining easily accessible through platforms such as GitHub and Hugging Face, where developers can upload, download and run open-source models. A Hugging Face study published on March 16, 2026 found that Chinese open-source models accounted for 41% of downloads, highlighting their growing adoption among developers and businesses.

For many organisations, the combination of affordability, strong performance and the ability to run models locally makes Chinese AI systems an attractive alternative to more expensive offerings. While companies continue to weigh concerns around security, data control and performance in high-stakes applications, many are increasingly focused on whether a model is cost-effective and capable enough to meet their needs, regardless of whether it originates in the US or China.