Alibaba raises $10.2 billion for AI infrastructure despite profit drop

This massive capital injection signals Alibaba's commitment to competing with Alphabet and Intel in the global compute race, even at the expense of short-term profitability. For the streaming and cloud ecosystem, this aggressive expansion into European availability zones like Paris and Germany provides a technical alternative to U.S. providers, though it faces mounting regulatory hurdles. The scale of this single private placement, nearly half of the EU's total AI budget, highlights the widening gap in infrastructure investment between Chinese tech giants and regional government initiatives. Watch for whether the proposed Cloud and AI Development Act successfully restricts Alibaba's operational independence in the EU market.

Alibaba's $10.2 billion capital raise places it among the most aggressive AI infrastructure spenders globally, directly competing with hyperscalers like Alphabet and Intel for compute dominance. In May 2025, Alibaba Cloud announced plans to invest 380 billion yuan ($53 billion) over three years into AI and cloud infrastructure, a commitment that dwarfs the current placement but signals a multi-year strategy to build proprietary silicon and data center capacity at scale. The company's Zhenwu M890 chip, designed for AI inference workloads, represents Alibaba's effort to reduce dependence on Nvidia hardware amid ongoing U.S. export controls that have restricted access to advanced GPUs for Chinese firms.

The regulatory and competitive landscape surrounding Alibaba's AI ambitions has intensified on multiple fronts. In early 2025, Alibaba's cloud division reported revenue growth of 13% year-over-year, driven by AI-related demand from enterprise customers, suggesting that the infrastructure spending is already generating commercial returns. Meanwhile, the European Union's proposed Cloud and AI Development Act could impose data residency and sovereignty requirements that complicate Alibaba's expansion into European availability zones. The company has been building out data center capacity in Paris and Frankfurt, positioning Alibaba Cloud as a lower-cost alternative to AWS and Azure for European enterprises, though regulatory scrutiny of Chinese cloud providers has increased following concerns raised by EU digital sovereignty advocates.

On the technical side, Alibaba's proprietary chip development mirrors a broader industry trend among hyperscalers seeking to optimize inference costs. Alibaba's Qwen series of large language models has been released as open-weight, with the Qwen 2.5 family supporting context windows up to 128,000 tokens, competing directly with Meta's Llama and Mistral's open models for developer adoption. The open-weight strategy reduces customer lock-in and encourages third-party deployment on Alibaba Cloud infrastructure, creating a flywheel where model adoption drives compute demand. Intel, mentioned as a competitor in the compute race, has pursued a different path with its Gaudi accelerator line, though Intel's AI accelerator revenue remained below $500 million in its most recent fiscal year, highlighting the difficulty of challenging Nvidia's dominance in AI training and inference hardware.