China has emerged as the world's top provider of open-source artificial intelligence models, surpassing U.S. contributions for the first time according to a joint MIT-Hugging Face study released this week. Chinese-developed models now account for 17.1% of global downloads, edging out America's 15.8% share.
Flagship systems like DeepSeek and Alibaba's Qwen are driving this growth, enabling developers worldwide to accelerate AI product development through free access and modification. At the 2025 OpenAtom Developers Conference in Beijing, Chinese Academy of Engineering academician Ni Guangnan noted: 'Open source has become the engine of global IT progress, with Chinese models like Qwen and Kimi ranking among the world's most adopted.'
The shift aligns with China's 15th Five-Year Plan (2026-2030) objectives emphasizing technological openness and international collaboration. Notably, 80% of U.S. AI startups now incorporate Chinese open-source models according to industry reports, demonstrating the ecosystem's global reach.
Analysts highlight that China's open-source communities are fostering unprecedented cross-border developer collaboration, with platforms attracting talent from over 200 countries and regions. This cooperative model is reshaping AI innovation pathways while supporting Beijing's vision of mutually beneficial technological advancement.
