Home

China

World

Politics

Business

Sci-Tech

Health

Opinions

Video

Live

Radio

World Asia-Pacific Middle East Europe Americas Africa

By continuing to browse our site you agree to our use of cookies, revised Privacy Policy and Terms of Use. You can change your cookie settings through your browser.

I agree

02:32

U.S. companies are increasingly shifting workloads to Chinese AI models to cut costs. DeepSeek leads the trend, but Alibaba's Qwen has also gained significant ground, with the company claiming its Qwen3.8-Max model rivals top U.S. systems on coding and reasoning benchmarks while offering far cheaper costs per token.

Ride-hailing giant Uber has publicly acknowledged building on Chinese open models, drawing scrutiny from U.S. lawmakers who have opened inquiries into the practice over national security concerns. Coinbase has also made the switch, cutting its AI spending by nearly half after moving to Chinese models GLM 5.2 and Kimi 2.7, even as its overall AI usage increased.

Price is the driving factor across the board with analysts at Citi finding Chinese models can charge as little as 18 cents per million tokens versus roughly $4 on average for top-tier competitors based in the U.S.

For more, check out our exclusive content on CGTN Now and subscribe to our weekly newsletter, The China Report.

Open in CGTN APP for better experience

Search Trends

More US companies switch to Chinese AI models to save money and tokens