Alibaba Cloud is set to increase its global data centre capacity to over 20 gigawatts by 2032, as the company ramps up infrastructure spending for AI computing. This announcement was made by Alibaba Group CEO Eddie Wu at the Apsara Conference in Hangzhou. The company had previously committed to investing more than 380 billion yuan in cloud computing and AI infrastructure over three years starting in February 2025.
The expansion plans not only focus on data centre capacity but also include the development of AI accelerators, server processors, networking and storage systems, cloud services, and Qwen models. Alibaba aims to cover the entire AI technology stack, from models and semiconductors to data centres.
In the June quarter of 2026, Alibaba’s AI Cloud and Compute Services revenue saw a 45% year-over-year increase, reaching US$7.1 billion. The company’s adjusted earnings before interest, taxes, and amortisation for the cloud segment rose by 133% to US$830 million, with an adjusted EBITA margin of about 12%. Alibaba reported triple-digit growth in revenue from AI-related products for the twelfth consecutive quarter.
Alibaba has already spent a significant portion of its planned AI investment during 2026, focusing on proprietary chips and AI infrastructure. The company expects margins to improve as it deploys more processors developed by its T-Head semiconductor unit.
Alibaba Cloud faces increasing customer demand for AI services, with supply chain constraints limiting the pace of capacity expansion. To address this, Alibaba Cloud plans to bring AI supernodes into commercial operation at scale in the current quarter.
Alibaba also unveiled the Zhenwu V900, an AI processor designed by its T-Head semiconductor unit for model training and inference. The processor offers superior performance compared to its predecessor, the Zhenwu M890. The V900 is scheduled for mass production and commercial release in the first quarter of 2027.
Alibaba’s chip roadmap includes the introduction of general-purpose processors, such as the Yitian 720 and Yitian 730 CPUs in 2027. The company has already integrated T-Head processors into its cloud infrastructure, supplying computing capacity to Alibaba Cloud infrastructure and its model-as-a-service inference platform.
In addition to custom processors, Alibaba announced updates to its networking and storage systems for large AI clusters. The company’s HPN 8.0 Pro networking architecture provides high bandwidth, while its Cloud Parallel File Storage system supports high throughput for AI training workloads.
As hyperscalers like Alibaba, AWS, Microsoft, and Google develop more custom infrastructure, the focus is shifting toward AI computing and agent services. Alibaba continues to invest in larger Qwen models, with plans for future models with trillions of parameters.
Overall, Alibaba’s efforts to expand its AI infrastructure and connect it to cloud services demonstrate the company’s commitment to meeting the growing demand for AI technology in the digital age.
