Chinese-built AI models are gaining traction among U.S. companies as they close the performance gap with leading American rivals while remaining significantly cheaper to use. Recent releases from Chinese firms such as DeepSeek and Z.ai are viewed as highly competitive against systems from Anthropic and OpenAI.
The share of tokens used by U.S. companies on Chinese AI models via OpenRouter has surged, reaching over 30% weekly since February 8, peaking at 46%. This is a stark increase from an average of just 11% over the past year. As U.S. companies face rising costs for advanced AI models, they are becoming more cost-conscious, shifting towards cheaper open-source and open-weight models.
For instance, AI startup Lindy transitioned entirely from Anthropic's Claude models to DeepSeek, resulting in significant cost savings. The performance of Chinese models is improving, with some, like GLM 5.2, performing nearly on par with top U.S. models at a fraction of the cost.
Analysts suggest that these developments could lead to a shift in the AI landscape, where companies may increasingly opt for Chinese models to control costs and maintain flexibility in their AI strategies