Based on 33 recent Alibaba articles on 2026-10-02 11:20 PDT

Alibaba Bets Its Future on AI Infrastructure as Costs and Risks Rise

• Alibaba is pursuing a full stack AI strategy spanning Qwen models, proprietary Zhenwu chips, cloud infrastructure, agents, and enterprise applications.
• The company plans to operate more than 20 gigawatts of global data center capacity by 2032 and open cloud regions in Türkiye, Finland, and the Netherlands.
• AI Cloud and Compute Services revenue rose 45% to approximately RMB48.4 billion, while AI related capital spending increased 75% to RMB67.7 billion in the June quarter.
• Net income fell 75% and free cash flow remained deeply negative, underscoring the considerable financial burden of Alibaba’s long term infrastructure program.
• Alibaba’s possible Spanish data center highlights the growing importance of renewable power, grid access, and local regulation in global AI expansion.
• Market and governance concerns persist, including share dilution, an insider stock sale, a pending securities lawsuit, and questions about Alibaba’s partnership with the Brooklyn Nets.

Alibaba is increasingly positioning itself as an infrastructure company for the AI economy rather than simply an e-commerce leader. At the September Apsara Conference and in subsequent disclosures, it outlined a strategy covering Qwen foundation models, the Zhenwu V900 processor, agent-oriented cloud services, and enterprise applications. The V900 is expected to enter mass production in the first quarter of 2027, while future Qwen models are planned at between 5 trillion and 10 trillion parameters. Management’s target of more than 20 gigawatts of global data center capacity by 2032 illustrates the scale of the bet .

The strongest evidence supporting the strategy is coming from cloud growth, but the economics remain unsettled. AI Cloud and Compute Services revenue increased 45% year over year to roughly RMB48.4 billion, and cloud adjusted EBITA more than doubled, according to reports. At the group level, however, capital expenditures rose 75% to RMB67.7 billion, free cash flow outflows widened to RMB44.7 billion, adjusted EBITA fell 30%, and net income declined 75%. Alibaba has also raised approximately $10.2 billion through a share placement, making dilution and the timing of returns central concerns for investors .

The international expansion shows that Alibaba’s AI ambitions depend on more than software and chips. New cloud regions in Türkiye, Finland, and the Netherlands, together with additional capacity in Germany, France, Malaysia, the United Arab Emirates, and Hong Kong, would bring computing closer to customers and support data residency requirements. Preliminary discussions with Solaria over renewable power for a potential Spanish data center show how electricity supply, pricing, and grid connections are becoming strategic constraints. Spain’s proposed rules could require new data center demand to be matched by equivalent renewable generation, while also raising questions about ownership and regulatory eligibility 3.

Geopolitical and legal uncertainties could complicate the expansion. Reports that China may permit Alibaba and ByteDance to purchase Nvidia’s RTX PRO 5500 workstation chips remain unconfirmed, but they illustrate the narrow and shifting pathways available under U.S. export controls and Chinese procurement restrictions. Alibaba also faces a pending U.S. securities class action related to alleged disclosure failures surrounding its designation on a U.S. military company list and accusations involving Anthropic’s Claude model. These issues add reputational and compliance risk at a time when Alibaba is seeking to persuade customers and investors that its AI ecosystem can scale responsibly .

The company is also trying to convert its technology into visible commercial use cases. Alibaba Cloud’s tools for model orchestration, model-as-a-service, video generation, and enterprise agents, alongside the DHL logistics collaboration, suggest a push toward practical deployment rather than research alone. The Brooklyn Nets partnership offers a smaller but illustrative showcase for real-time replay, personalized fan content, and cloud-based venue services, although its related-party character raises questions because Joe Tsai chairs Alibaba and owns the Nets. Meanwhile, the October 2 market sell-off and the reported sale of nearly $12 million of shares by Chief People Officer Fang Jiang show that enthusiasm for the AI strategy remains vulnerable to broader risk appetite and investor scrutiny .

Concluding Thought

Alibaba has assembled one of the most ambitious AI infrastructure strategies among Chinese technology companies, with potential advantages in cloud distribution, proprietary silicon, and integrated models. The immediate challenge is demonstrating that rising AI revenue can outpace the cost of data centers, chips, talent, and international expansion. Its next phase will be judged less by announcements than by utilization rates, cash flow recovery, regulatory execution, and evidence that enterprise customers are willing to pay for the full stack.