Read, Watch or Listen

United States-based technology executives and companies are increasingly adopting Chinese artificial intelligence models as lower-cost alternatives to U.S. systems, underscoring a shift in the global AI market toward price-sensitive, commercially focused tools. Mozilla’s chief technology officer, Raffi Krikorian, said he moved his day-to-day workflow to Chinese startup Moonshot’s Kimi K3 model within days of its July release, describing it as faster and more efficient than established American options. Before switching to Kimi K3, he used another Chinese system, Z.ai’s GLM-5.2, for routine office functions such as managing documents, scheduling meetings, and organizing email. Market data from OpenRouter show that the five most popular AI models over a recent month were all developed in China, reflecting broad international demand. United States companies, including cryptocurrency exchange Coinbase and North Carolina-based digital legacy platform LilyList, are testing or shifting workloads to Chinese platforms such as DeepSeek to reduce operating costs for tasks like lead generation and sales research. Executives say most commercial needs do not require expensive frontier-grade American models, and that Chinese systems provide sufficient performance at a fraction of the price, a gap magnified by per-token pricing as usage scales and as agentic AI applications trigger long, multi-step workflows. Industry analysts note that Chinese startups including DeepSeek, Zhipu and Alibaba have released new models this year, while a Goldman Sachs report in July said Chinese AI systems are approaching a threshold for broad enterprise adoption, even as experts stress that they still trail U.S. leaders in overall capabilities and face questions over financial sustainability and alleged misuse of proprietary U.S. technology, which Chinese authorities deny.

Prepared by Jonathan Pierce and reviewed by editorial team.