Several U.S.-based technology startups and corporations have begun utilizing Chinese artificial intelligence models to reduce operating costs. Flo Crivello, founder of the San Francisco startup Lindy.ai, reported that his company migrated its entire traffic to the Chinese model DeepSeek-V4 after finding that expenses for American models exceeded corporate payroll. Crivello stated that the switch saved the company millions of dollars.
While U.S. developers such as OpenAI and Anthropic currently lead in model capabilities, Chinese firms have focused on high-performance open-source models that are free to download and adapt. Public filings and executive statements indicate that major companies, including Uber and Airbnb, have faced rising AI budgets, with Airbnb CEO Brian Chesky previously noting the company's use of Alibaba’s Qwen model due to its cost-efficiency.
Industry experts notes that while Chinese models are estimated to be six to 12 months behind U.S. leaders in raw power, their affordability makes them attractive for specific business functions. These models are accessible through international hosting platforms like GitHub and Hugging Face, allowing U.S. firms to integrate them despite ongoing geopolitical tensions between the two nations regarding technology trade.