AI & ComputeLOT 261119 · 4 min read

Alibaba's AI cloud grows 53% while AI labs burn RMB10.5 billion

AI cloud revenue up 53% to RMB61.0 billion with AI product sales in triple-digit growth for a 13th quarter, but Qwen labs lost RMB10.5 billion and free cash flow hit negative RMB34.5 billion.

By Nathan Brooks789 words

Key takeaways

  1. AI Cloud and Compute Services revenue rose 53% to RMB61.0 billion in the September 2026 quarter; AI product revenue hit its 13th straight quarter of triple-digit growth at RMB17.92 billion.
  2. AI Labs and Applications, home of Qwen, widened its adjusted EBITA loss to RMB10.50 billion from RMB3.44 billion a year earlier.
  3. Capital expenditure rose 60% to RMB50.48 billion, driving free cash flow to negative RMB34.47 billion.
  4. Group net income rose 213% to RMB64.42 billion, mainly on mark-to-market equity investment gains; non-GAAP net income rose 92% to RMB19.91 billion.
  5. T-Head's Zhenwu V900 AI chip is scheduled for commercial release in the first quarter of 2027.
Alibaba's AI cloud revenue jumps 53% in the September quarter while its AI labs lose RMB10.5 billion
Alibaba's AI cloud revenue jumps 53% in the September quarter while its AI labs lose RMB10.5 billion — Tumisu / pixabay

Alibaba's AI Cloud and Compute Services revenue rose 53% to RMB61.0 billion (US$9.09 billion) in the quarter ended Sept. 30, 2026, even as the segment housing its Qwen models lost RMB10.50 billion on an adjusted EBITA basis. The company reported the figures on Thursday, Nov. 19. The numbers describe a business scaling fast and paying for it immediately.

Revenue from external cloud customers grew 64% in the quarter, and recurring revenue from external customers, which excludes compute equipment sales, grew 67%. AI-related product revenue reached RMB17.92 billion (US$2.67 billion), its 13th consecutive quarter of triple-digit growth. Revenue from compute equipment solutions sold to external customers grew 55% to roughly RMB10 billion, according to the results presentation.

Where the AI money went

The AI Labs and Applications segment, which contains the Qwen models, the Qwen app and the QwenWork agent platform, grew revenue 51% to RMB4.44 billion. Its adjusted EBITA loss widened to RMB10.50 billion from RMB3.44 billion a year earlier. Alibaba attributed the widening to "increased investment in AI capabilities, and higher inference cost related to Qwen app."

The loss did narrow sequentially. A six-month loss of RMB24.36 billion implies a June-quarter loss of about RMB13.86 billion. For context, the segment's loss exceeded the cloud operation's RMB8.84 billion adjusted EBITA profit — the labs consume more than the cloud earns.

The cloud segment itself performed well. Its adjusted EBITA rose 121% to RMB8.84 billion, and its margin widened to 14.5% from 10.0% a year earlier. Alibaba reorganized reporting in the June quarter: AI Cloud and Compute Services now combines the former Cloud Intelligence Group with chip designer T-Head, with prior-year figures recast to match.

Capex climbs, cash walks out

Capital expenditure rose 60% to RMB50.48 billion (US$7.53 billion). Alibaba cited faster scaling of compute capacity, higher prices for a broad range of chip components, more CPU capacity for expected AI agent adoption, and purchase timing. Operating cash flow was RMB15.69 billion, leaving free cash flow at negative RMB34.47 billion, compared with negative RMB21.84 billion a year earlier. Net cash fell to RMB220.7 billion at Sept. 30 from RMB260.8 billion at March 31.

T-Head said its next-generation Zhenwu V900 AI chip is scheduled for commercial release in the first quarter of 2027. The company also recently launched Qwen4-Max, highlighted in the release, on Nov. 16.

Group profit tripled, mostly on paper

Group revenue rose 10% to RMB271.52 billion (US$40.50 billion), the second quarter of Alibaba's 2027 fiscal year. Net income rose 213% to RMB64.42 billion (US$9.61 billion), driven mainly by mark-to-market gains on equity investments; net interest and investment income rose 265% to RMB73.31 billion, partly offset by higher impairments. Non-GAAP net income, which excludes those gains, rose 92% to RMB19.91 billion (US$2.97 billion). Non-GAAP diluted earnings per ADS were RMB8.10 (US$1.21), up 86%.

Income from operations rose 443% to RMB29.12 billion, and adjusted EBITA rose 260% to RMB32.64 billion, which Alibaba credited to better results at Taobao Instant Commerce and the cloud business, partly offset by higher technology spending.

E-commerce quietly paid the bill

Alibaba E-commerce Group revenue rose 2% to RMB197.24 billion, while its adjusted EBITA rose 243% to RMB38.85 billion, mainly on improved performance at Taobao Instant Commerce, the quick-commerce delivery service. China Quick Commerce revenue rose 22% to RMB58.76 billion, driven by Freshippo. Customer management revenue fell 5% to RMB75.03 billion, though Alibaba said it would have grown 1% excluding the contra-revenue impact of a new business development program. International e-commerce revenue fell 8% to RMB25.72 billion on lower AliExpress revenue after EU customs reform took effect, partly offset by growth at Trendyol.

Alibaba repurchased 17.95 million ordinary shares, about 2.24 million ADSs, for US$221 million in the quarter.

Chief Executive Eddie Wu said: "Our full-stack AI capabilities continue to improve commercialization and delivered a strong quarter," adding that quick commerce "continued to narrow its losses substantially while maintaining market share." Chief Financial Officer Toby Xu said the company will "maintain disciplined investments in AI capabilities with a focus on return on invested capital while targeting to achieve positive free cash flow." The release offered no numerical revenue or capex guidance.

With cloud margins widening and lab losses narrowing but still enormous, the next test is whether the Zhenwu V900 and Qwen4-Max can turn RMB50 billion of quarterly capex into revenue that outruns the inference bill.

mlq.ai (Source)

Filed under

Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,105 articles

Read also

« Previous article