Alibaba财报分析:AI云业务强劲增长但集团盈利承压

Alibaba财报分析:AI云业务强劲增长但集团盈利承压

Driven
·Agent created by Agent
Driven·Agent created by Agent
Faster growth in Alibaba Cloud’s AI revenue is lifting earnings expectations. What do Alibaba’s latest results and disclosures show about the scale and pace of that growth, and what would investors need to see for it to translate into sustained earnings gains?

I'll pull Alibaba's latest disclosures and quote data first, then check the AI cloud numbers against the filings.

Conclusion: The AI cloud numbers are real and accelerating, but they are not yet showing up in group earnings. Alibaba Cloud's AI revenue is growing about 45% and its cloud segment margin is rising. Group non-GAAP earnings fell 38–42% in the latest quarter because AI investment, capex and losses in the new AI Labs segment are rising faster. Sustained earnings gains will need cloud margins to hold as depreciation from a much larger capex base builds, AI Labs losses to narrow, and capex to convert into revenue at a rate that clearly beats the cost.

Data as of: Latest results are for the June 2026 quarter, released August 20, 2026. Share price context is the NYSE close on Friday, November 6, 2026 (US Eastern). I found no September-quarter release in the sources I retrieved.

Scale and pace of AI revenue

  • AI Cloud and Compute Services revenue was , up year over year against a restated . This segment combines Cloud Intelligence Group with T-Head, so the comparison uses restated figures.
  • AI-related product revenue was (US$1.8bn), the twelfth straight quarter of triple-digit growth. That annualizes to about , and it is of segment revenue.
  • Sequentially, AI product revenue rose from about RMB8.97bn in the March quarter. That March figure comes from a secondary source (), not the release.
  • Management said AI-related products reached 35% of Alibaba Cloud external revenue (). I could not confirm this figure in the release itself.
  • External Alibaba Cloud revenue growth accelerated to 45%, a 22-quarter high, per the company's release page.

Pace of margin improvement in cloud

  • AI Cloud adjusted EBITA was , up 133% from RMB2,419m.
  • The margin moved from to . The company describes it as roughly 12%.

Why group earnings fell

  • Group revenue was , up 9%.
  • Group adjusted EBITA was , down .
  • Non-GAAP net income was , down .
  • Non-GAAP diluted EPS per ADS was , down from RMB14.75.
  • The AI Labs and Applications segment posted an adjusted EBITA loss of , against a loss of RMB3,224m a year earlier. The company cites higher AI investment and inference costs for the Qwen app. That loss is larger than the cloud segment's entire adjusted EBITA.
  • China e-commerce customer management revenue fell 7% reported, or +1% like-for-like.

Capex intensity

  • Capex was , up . The company cites procurement timing, more CPU compute capacity and higher chip component prices.
  • Capex rose from of revenue to . That is about 12x the quarter's AI Cloud adjusted EBITA.
  • The company reports about US$69.9bn in cash and liquid investments ().
  • The March quarter free cash flow was negative RMB17.3bn, per the same secondary source.

Consensus expectations

Driven consensus shows average EPS estimates rising from FY2027 (March 2027) through FY2029 (, 18 analysts). However, the currency and per-share basis are not labeled, and some rows are inconsistent, such as a zero EPS for FY2031. I treat these estimates as directional only.

What investors need to see for sustained earnings gains

  1. Cloud margins holding through the depreciation wave. Capex at this level will raise depreciation. AI Cloud adjusted EBITA margin needs to stay around 12% or higher in coming quarters, not just in one strong quarter.
  2. Capex converting to revenue faster than it grows. Revenue growth in the AI segment has to keep pace with capex. If capex keeps growing 75% while AI revenue slows, returns fall.
  3. AI Labs losses narrowing. Cloud profit gains are currently absorbed by AI Labs investment and inference costs. The Qwen app's paid subscriptions, launched in August, are an early test.
  4. Group non-GAAP EPS turning up. Segment margins need to show up in group non-GAAP EPS, which has fallen for the latest quarter.
  5. Disclosure quality. The new four-segment structure and restated comparatives make year-over-year comparisons harder. Investors need consistent disclosure of AI revenue and the restated base.
  6. Chip cost and supply. The Zhenwu M890 chip is used by more than 650 external customers (). Its cost advantage matters because the company cites higher chip component prices.

Risks to the thesis

  • Management's AI revenue disclosures are a segment-based metric, not audited product revenue.
  • The one-third share figure comes from call commentary, not the release.
  • The March quarter figures come from a secondary source.
  • Quarterly capex is lumpy, so one quarter's growth rate may not be a trend.

Sources

  • Alibaba Group, June Quarter 2026 Results press release, August 20, 2026 (via FinancialContent/Business Wire): financialcontent.com
  • Alibaba Group investor page, Full-Stack AI Accelerates Monetization, August 20, 2026: home.alibabagroup.com
  • Benzinga, Alibaba Says AI Now Drives More Than One-Third of Cloud Revenue, August 20, 2026 (secondary, call summary)
  • ciw.news, Alibaba's AI cloud hits US$5B run rate, May 20, 2026 (secondary, March quarter)
  • Driven Market Data: BABA quote (NYSE close, November 6, 2026) and analyst financial estimates for BABA
  • Calculations: Driven calculation tool, from the sources above

More research on BABA