Alibaba Cloud Targets Over 20GW Global Data Center Capacity by 2032
At the Apsara Conference, Alibaba CEO Wu Yongming committed to scaling Alibaba Cloud's global data center capacity to over 20GW by 2032 to meet surge in AI demand. Alibaba also announced plans to train a massive frontier AI model with 5 trillion to 10 trillion parameters. This aggressive long-term expansion highlights the intense global competition among cloud providers to secure massive compute power and energy supplies for AI. It reinforces Alibaba's position as a primary infrastructure powerhouse supporting the next generation of trillion-parameter models and enterprise applications. Alibaba outlined a five-layer full-stack AI architecture covering chips, cloud infrastructure, models, services, and AI agent applications. Notably, Nvidia participated as the conference's top-tier apex partner, alongside Intel and AMD, highlighting key hardware supplier collaborations.
## BACKGROUND
Alibaba Cloud is the developer of the Qwen (Tongyi Qianwen) model family, which represents some of the most capable and widely adopted open-weight LLMs globally. Expanding cloud infrastructure to gigawatt-scale capacities has become essential for hyper-scalers, as training multi-trillion parameter models requires vast amounts of continuous electricity and dense GPU clusters.