Alibaba Profit Falls 75% As AI Spending Hits Nearly $10B

Alibaba’s latest result captures the uncomfortable math of the AI race: the business is growing, AI demand is real, but the bill for building the infrastructure is getting very large.

Alibaba announced results for the quarter ended June 30, 2026, with revenue rising 9 percent year over year to RMB268.95 billion, or about $39.64 billion. Net income, however, fell 75 percent to RMB10.44 billion, while net income attributable to ordinary shareholders was RMB10.54 billion.

The pressure came from several directions, but AI spending is the part investors will watch most closely. Alibaba said capital expenditure reached RMB67.68 billion, or about $9.98 billion, up 75 percent from the same period last year. Free cash flow was an outflow of RMB44.67 billion, largely because of increased cloud infrastructure expenditure.

At the same time, the AI business is clearly not just a future story. Alibaba’s AI Cloud and Compute Services revenue rose 45 percent to RMB48.44 billion, driven by public cloud growth and demand for AI-related products. AI-related product revenue reached RMB12.38 billion and delivered its twelfth consecutive quarter of triple-digit year-over-year growth.

Alibaba also reorganized how it reports the business, combining cloud and T-Head Semiconductor into AI Cloud and Compute Services while grouping model labs, Qwen Consumer Business and QwenWork into AI Labs and Applications. That gives investors a clearer view of the company’s full-stack AI strategy.

The numbers show why Alibaba remains one of China’s most important AI companies. It has cloud, chips, Qwen models, enterprise workflow tools, e-commerce data and distribution across Taobao, Tmall, DingTalk and other services. Few companies can connect AI infrastructure, models and consumer or enterprise use cases at that scale.

The problem is that scale is expensive. Alibaba’s AI Labs and Applications segment reported a large adjusted EBITA loss, partly due to increased AI investment and higher inference costs related to the Qwen app. That is the part of AI investors often underestimate: once users arrive, serving them can be costly.

This fits the broader China AI story we have been tracking. China is pushing cheaper and more open AI models, and Alibaba’s Qwen family is central to that effort. We have argued that China may win the AI race if the US does not adapt, partly because Chinese companies are turning AI into infrastructure, not only demos.

For investors, the question is whether Alibaba’s AI spending becomes operating leverage or a permanent drag. If AI cloud demand keeps accelerating and proprietary chips reduce long-term costs, today’s capex could look strategic. If inference costs keep rising faster than monetization, the market will become less patient.

Alibaba’s quarter therefore tells a wider story. The AI boom is not free growth. It is a capital-intensive infrastructure cycle. Companies that win may build huge new businesses, but they will first have to prove that nearly $10 billion of quarterly AI spending can turn into durable profit.