The US is preparing to restrict American companies’ use of Chinese-made AI models. An unnamed State Department spokesperson stated that the use of these models raises serious concerns, saying they are “designed to advance Beijing’s narratives, censor dissent and reflect CCP ideology and values.”
US companies are increasingly opting for Chinese AI models due to lower costs and comparable quality. Coinbase CEO Brian Armstrong revealed the company’s use of two Chinese-made AIs: GLM 5.2 by Z.Ai and Kimi 2.7 by Moonshot. The CEO of startup Lindy also indicated a switch to Chinese AI, selecting DeepSeek to reduce costs.
A report from Nikkei Asia noted a surge of American companies, including Airbnb and Uber, adopting Chinese AI models. This surge coincided with Anthropic’s suspension of its AI models, Mythos 5 and Claude Fable 5, at the government’s request. It remains unclear whether the US can impose a broad ban on AI models beyond its procurement rules.
Restrictions on open-source models could raise First Amendment issues for the US government. Additionally, companies operating overseas may face challenges in altering their AI model usage. For example, Apple uses Alibaba’s generative AI for iPhones sold in China.
Chinese authorities are reportedly holding discussions with domestic companies to limit the use of homegrown AI overseas. The Chinese Ministry of Industry and Information Technology has claimed that Anthropic’s Claude Code contains a backdoor, which it describes as a “serious threat.”
Ongoing developments regarding US-China AI model usage are expected in the coming weeks, months, and years.
