Artificial intelligence has emerged as a major and rapidly growing expense for American businesses, prompting a growing number of companies to shift toward lower-cost Chinese AI models to manage their budgets.
Lindy.ai, a San Francisco-based startup that creates AI assistants for email and calendar management, recently made a complete migration to DeepSeek-V4, a Chinese model, after determining that Anthropic’s services represented their largest single expense—surpassing payroll, rent, and other costs combined. The company reported saving millions of dollars through the switch, with the Chinese model costing roughly one-tenth as much. Industry observers note that Chinese open-source models have come to dominate the market for freely available and adaptable AI systems, even as American companies like Anthropic, OpenAI, and Google maintain advantages in frontier AI capabilities.
The cost pressures extend beyond startups. Uber’s CEO recently disclosed that the company exhausted its annual AI budget within a single quarter, forcing operational adjustments. Major companies including Airbnb, Perplexity, and Nvidia have also incorporated Chinese models such as Alibaba’s Qwen into their operations. While some firms remain reluctant to publicly acknowledge their reliance on Chinese models due to political considerations, these systems are widely accessible through platforms including Hugging Face, GitHub, and various inference providers based outside China.
Experience levels with Chinese alternatives vary among companies. Some view the trade-off between cost savings and performance as acceptable, comparing the choice to selecting a reliable Honda over a Ferrari. Analytics from OpenRouter indicate that usage of China’s DeepSeek model increased from approximately 9 percent to nearly 20 percent since January, while adoption of models from MiniMax, Xiaomi, and Tencent has also risen. However, newer companies with demanding performance requirements, such as Comment.io, continue to prioritize quality and have opted to remain with American providers, particularly while those providers offer discounted subscription rates.
Industry analysts suggest that American AI companies may respond competitively by introducing more affordable offerings or developing open-source alternatives. However, some observers expect that as U.S. firms face increasing pressure to demonstrate profitability ahead of potential public offerings, pricing pressures could intensify, potentially driving further migration toward Chinese options.
Originally reported by NPR. Read the full story →