Flo Crivello's Lindy.ai switches entirely to Chinese AI models, citing significant cost savings.
U.S. technology firms are increasingly adopting cheaper Chinese AI models to manage rising operational costs associated with advanced artificial intelligence. Flo Crivello, founder of the San Francisco-based startup Lindy.ai, announced the company migrated all traffic to DeepSeek-V4, stating the change saved millions of dollars. The trend reflects a sector-wide concern over escalating expenses. While American firms like Anthropic and OpenAI lead in AI development, many businesses are finding robust, more affordable alternatives. Experts note that cost per token is driving these shifts, allowing businesses to focus on high-volume, repetitive tasks where efficiency matters most. Eugene Cheah, founder and CEO of Featherless, noted that reliable, accessible models are often preferable to the most advanced, costly options. Furthermore, reports indicate that companies are optimizing spending by utilizing models from China, such as those from MiniMax and Xiaomi. This focus on cost-controlled implementation suggests a pragmatic market adjustment aimed at ensuring American enterprise profitability in the competitive AI landscape.