Alibaba

Alibaba Group is China's e-commerce and cloud giant, and in AI it is best known for Qwen — the open-weight model family from Alibaba Cloud that has become the most downloaded and most-forked base model line in the world. The company has committed tens of billions of dollars to AI and cloud infrastructure, positioning Alibaba Cloud as China's leading AI platform. Qwen models anchor a large share of global open-source fine-tuning and derivative model work.

Founded by Jack Ma in 1999 as a B2B marketplace in Hangzhou, Alibaba grew into China's dominant e-commerce operator (Taobao, Tmall) and built Alibaba Cloud into Asia's largest cloud provider. AI research consolidated under the cloud division and the Tongyi lab, whose Qwen (Tongyi Qianwen) models debuted in 2023.

Qwen's strategy mirrors Meta's Llama playbook but with faster iteration and broader coverage: dense and mixture-of-experts variants at every size, coder and vision-language specialists, and permissive licensing. Qwen derivatives dominate open-model leaderboards and Hugging Face download rankings, making Alibaba arguably the most influential force in open-weight AI worldwide.

In 2025 Alibaba announced a three-year infrastructure program worth over RMB 380 billion (roughly $50 billion) for AI and cloud, among the largest committed by any Chinese company. Its consumer Qwen assistant, enterprise Model Studio platform, and investments in domestic labs — including Moonshot AI — round out a full-stack AI strategy that also hedges against US chip export controls through in-house silicon efforts.