China AI Model API Usage Surpasses US for the First Time
China Overtakes US in AI Model API Call Volume
Data from OpenRouter, the leading global AI model aggregation platform, shows that Chinese AI large language model weekly token-based API usage has surged dramatically since February 2026, now exceeding that of American AI models.
The Data
OpenRouter data reveals a clear inflection point: Chinese LLM API call volume measured in tokens has been growing at an explosive rate, overtaking US models in weekly usage. This marks a significant milestone in the global AI landscape.
What Is Driving This
Several factors contribute to the surge:
- Cost advantage: Chinese models like DeepSeek offer comparable quality at significantly lower prices
- Developer adoption: Chinese AI companies have made aggressive pushes into international developer ecosystems
- Performance gains: Recent Chinese models have narrowed or closed the quality gap with leading US models
- API accessibility: Chinese providers have simplified onboarding and offered generous free tiers
Implications
This shift challenges the prevailing narrative that US companies dominate the AI infrastructure layer. While US firms still lead in foundation model research, Chinese companies are winning the API usage battle through aggressive pricing and ecosystem expansion.
The data suggests that the AI market is becoming genuinely competitive, with Chinese providers carving out significant market share in the rapidly growing API economy.