全球指数

BABA-SW(9988.HK)1QFY27 results review:Cloud business accelerates with clear returns on capital expenditure

中信建投证券股份有限公司2026-09-04
Key takeaway
We remain bullish on the company for three reasons:
1) Strong cloud revenue growth: Revenue from AI cloud and computing services grew 45% YoY this quarter, while the EBITA margin reached double digits. Management expects growth to continue accelerating and margins to improve QoQ. The payback period for AI computing investments is approximately three years and is expected to shorten to 2.5 years, providing strong visibility into investment returns.
2) AI application losses have peaked: As model training and marketing efficiency for the Qwen App improve, losses will continue to narrow in the coming quarters, reducing the incremental drag of AI investment on the group’s profit.
3) E-commerce profit remains resilient: Unit economics for Taobao Instant Commerce improved QoQ, narrowing overall losses. AliExpress recorded an operating profit. Stable profit from traditional e-commerce provides cash flow support for AI investment.
Key areas to watch: 1) the unveiling of the next-generation model at the Apsara Conference in September; 2) capital expenditure and free cash flow; 3) progress toward the year-end MaaS ARR target of RMB30bn; 4) paid conversion for the Qwen App and commercialization progress for Qwen Office; and 5) the launch of Apple Intelligence in mainland China in September.
Event
Alibaba released its 1QFY27 results. During the reporting period, the company recorded revenue of RMB268.953bn, up 9% YoY; adjusted EBITA of RMB27.329bn, down 30% YoY; and Non-GAAP net profit of RMB20.715bn, down 38% YoY.
Quick Take
1. Business update: Strong cloud business growth, with adjusted EBITA exceeding company guidance Revenue was in line with consensus, while adjusted EBITA beat company guidance. In 1QFY27, the company recorded revenue of RMB268.953bn, up 9% YoY. Adjusted EBITA was RMB27.329bn, down 30% YoY but above the company’s guidance of RMB26bn, mainly driven by strong cloud business growth. Non-GAAP net profit was RMB20.715bn, down 38% YoY.
Net profit was below the Bloomberg consensus estimate, mainly due to one-off expenses. 1) The company recognized a goodwill impairment charge of RMB4.458bn this quarter, related to all other segment businesses. 2) The company recognized a provision related to a EUR550mn fine under the EU Digital Services Act.
Affected by the above factors and operating investments.
Starting this fiscal quarter, the company implemented a new segment reporting structure to strengthen synergies across its e-commerce platforms and full-stack AI capabilit ies. 1) Alibaba E-commerce Group: formed through the integration of China E-commerce Group, Alibaba International Digital Commerce Group, Freshippo and certain businesses under Cainiao, with revenue disclosed under four categories: China e-commerce, China quick commerce, international e-commerce and global wholesale; 2) AI cloud and computing services: formed through the integration of Cloud Intelligence Group and T-Head; 3) AI labs and applications: comprising the AI model labs, Qwen 2C Business Unit and Qwen Office, which were previously included in All Others; 4) All Others: mainly comprising Ali Health, Digital Media and Entertainme nt Group,
Amap and other businesses.
1) Revenue from AI cloud and computing services accelerated, with a significant margin improvement.
Segment revenue was RMB48.437bn, up 45% YoY. Growth in both total revenue and external customer revenue accelerated to 45%, the highest level in 22 quarters, according to the earnings call. Revenue from AI-related products was RMB12.376bn, accounting for 35% of external cloud revenue. Adjusted EBITA was RMB5.628bn, up 133% YoY, with a margin of 11.6%, up 4.4 pcts QoQ. According to Omdia, Alibaba Cloud ranked first in China’s AI cloud market with a 38.1% market share; T-Head’s Zhenwu chips have been commercially de ploye d by more than 650 external customers across over 20 industries.
2) E-commerce Group remained broadly stable, while losses from quick commerce narrowed significantly. Segment revenue was RMB205.862bn, up 4% YoY; adjusted EBITA was RMB39.749bn, down 1% YoY. China e-commerce revenue was RMB110.900bn, down 8% YoY, including customer management revenue of RMB82.547bn, down 7% YoY; excluding the impact of accounting for the new marketing development program as a reduction in revenue, customer management revenue grew 1% YoY on a like-for-like basis. China quick commerce revenue was RMB53.295bn, up 45% YoY, driven by Freshippo and Taobao Instant Commerce;
Taobao
Instant Commerce maintained its market share while delivering continued QoQ improvement in unit economics.
International e-commerce revenue was RMB27.761bn, down 1% YoY, while AliExpress recorded an operating profit. Global wholesale revenue was RMB13.906bn, up 7% YoY.
3) Investment in AI labs and applications increased, with losses widening YoY. Segment revenue was RMB3.338bn, up 16% YoY; adjusted EBITA loss was RMB13.861bn, versus a loss of RMB3.224bn in the same period last year, mainly due to increased investment in AI capabilities and higher inference costs for the Qwen App. Qwen 3.8-Max was released with open weights in August and has 2.4 trillion total parameters; since the launch of the Qwen App, 250mn users have used its agent features for AI-powered shopping; Qwen Office is deeply integrated with the Alibaba Cloud and DingTalk ecosystems.
2. Capital expenditure and cash flow: Continued growth in AI infrastructure investment Capital expenditure increased significantly YoY, while net free cash outflow widened. Capital expenditure was RMB67.678bn this quarter, up 75% YoY, mainly due to continued investment in AI infrastructure, including fluctuations in hardware delivery schedules, increased CPU procurement driven by demand for AI agents, and higher prices for chip components. Net free cash outflow was RMB44.670bn, versus RMB18.815bn in the same period last year. As of the end of June, the company had RMB474.505bn in cash and other liquid investments; according to the earnings call, net cash was approximately USD30.7bn, or approximately USD46.5bn excluding debt with maturities of more than five years. Share buyback announced. The company bought back 13.4mn ordinary shares, equivalent to 1.7mn ADSs, for a total of USD162mn. According to the earnings call, the company will continue to maximize long-term shareholder returns through disciplined capital allocation, including investment in AI and cloud businesses, share buybacks, and dividend payments.
3. Earnings call: Clear path to returns on capital expenditure, with AI lab losses expected to peak
1) Clear path to returns on capital expenditure
More than half of the three-year RMB380bn investment plan has been completed, with an active pace of investment maintained for the full year. According to the earnings call, cumulative investment under the RMB380bn plan reached RMB190bn as of the end of the June quarter, in line with expectations; the relatively high RMB67.7bn investment this quarter was mainly due to fluctuations in hardware delivery cycles and should not be extrapolated linearly on a quarterly basis, but the company will maintain an active investment approach this year. Full-stack AI platform services operate under an asset-heavy business model, and upfront capital expenditure is a prerequisite for business growth.
The payback period for AI computing power investment is approximately three years and is expected to shorten further. According to the earnings call, based on the current average gross margin of AI products, capital expenditure can be recouped within three years, with the payback period expected to decline to 2.5 years or even close to two years in the future. The A100 chips procured in 2020 remain at nearly full utilization, indicating that the actual useful life of AI computing assets is far longer than their theoretical depreciation period.
The company estimates that, assuming a three-year payback period, it can generate positive cash flow by keeping revenue growth within 33%. If the payback period is shortened to less than 2.5 years, the company can maintain positive cash flow while pursuing revenue growth of over 40%.
The three paths to improving investment returns include: 1) increasing the gross margin of AI products, which drove a 4.4pcts increase in the cloud segment’s margin this quarter; 2) increasing the proportion of T-Head’s proprietary chips used as substitutes in data centers; and 3) improving cash flow efficiency by jointly building AI computing centers with partners and using prepaid contracts and other arrangements.
2) High certainty of achieving the RMB30bn MaaS ARR target by year-end
MaaS ARR exceeded RMB16bn in August, with high certainty of achieving the RMB30bn target by year-end. According to the earnings call, ARR from model and application services, including MaaS, exceeded RMB16bn in August. Annualized revenue from AI-related products reached RMB49.5bn, or approximately USD7.3bn, and is expected to approach USD10bn next quarter. Management expects cloud revenue growth to continue accelerating over the next several quarters and the EBITA margin to improve sequentially. It is highly confident in achieving the target of USD100bn in external cloud revenue by 2030 and believes there is strong visibility for the AI cloud business to reach a margin of over 20%. Within the MaaS business, the gross margins from hosting third-party open-source models and proprietary models are similar, and a thriving open-source ecosystem benefits the cloud platform.
3) Losses from AI labs and applications are expected to have bottomed out
Losses from AI labs and applications narrowed significantly QoQ and are expected to continue improving in subsequent quarters. According to the earnings call, the segment recorded a loss of RMB13.9bn this quarter. The loss narrowed significantly QoQ due to lower marketing expenses for the Qwen App. As model training and the marketing efficiency of the Qwen App improve, the company expects losses to continue narrowing over the next several quarters.
4) E-commerce business provides earnings support as instant retail substantially reduces losses The e-commerce business provides earnings support, while instant retail is expected to achieve overall profitability in FY2029. According to the earnings call, the company aims to maintain stable profit from its traditional e-commerce business while driving continued improvement in the margin of instant retail. Taobao Instant Commerce continues to improve its unit economics QoQ while maintaining market share. Instant retail gross merchandise value from non-food categories is expected to exceed that from food and beverage categories in the next fiscal year. The instant retail segment is expected to achieve overall profitability in FY2029 and may contribute 30% of the platform’s total gross merchandise value over the long term. AliExpress recorded an operating profit this quarter; 50,000 paying merchants have used its AI agent for cross-border merchants since launch.
5) T-Head is key to improving the cloud business's gross margin
The large-scale commercial deployment of T-Head's proprietary chips is a key driver of gross margin improvement for the cloud business. According to the earnings call, shipments of T-Head's previous-generation chips have exceeded 500,000 units. The Zhenwu M890 supernode instance, which is fully domestically produced, was launched on Alibaba Cloud in August and has entered large-scale sales. Sales will continue to ramp up in 2H26. It can efficiently run inference for large models with more than 2trn parameters. Kimi K3 and Qwen 3.8-Max are already providing external MaaS services through this supernode. Tape-out of the second-generation chip is expected to begin in 2H26, with computing power and interconnect bandwidth capable of supporting large-scale model training. T-Head's chips support both training and inference workloads and have served hundreds of medium-sized and large customers.
Investment recommendation:
We remain bullish on the company for three reasons:
1) Strong cloud revenue growth: Revenue from AI cloud and computing services grew 45% YoY this quarter, while the EBITA margin reached double digits. Management expects growth to continue accelerating and margins to improve QoQ. The payback period for AI computing investments is approximately three years and is expected to shorten to 2.5 years, providing strong visibility into investment returns.
2) AI application losses have peaked: As model training and marketing efficiency for the Qwen App improve, losses will continue to narrow in the coming quarters, reducing the incremental drag of AI investment on the group’s profit.
3) E-commerce profit remains resilient: Unit economics for Taobao Instant Commerce improved QoQ, narrowing overall losses. AliExpress recorded an operating profit. Stable profit from traditional e -commerce provides cash flow support for AI investment.
Key areas to watch: 1) the unveiling of the next-generation model at the Apsara Conference in September; 2) capital expenditure and free cash flow; 3) progress toward the year-end MaaS ARR target of RMB30bn; 4) paid conversion for the Qwen App and commercialization progress for Qwen Office; and 5) the launch of Apple Intelligence in mainland China in September.
We forecast revenue of RMB1,121.2bn/RMB1,233.0bn/RMB1,358.7bn for FY2027-FY2029, respectively, up 9.53%/9.97%/10.20% YoY. Adjusted net profit is forecast at RMB108.3bn/RMB141.1bn/RMB184.6bn, respectively, up 78.57%/30.24%/30.84% YoY. Based on the closing price on August 21, 2026, the corresponding P/E multiples are 18.70x/14.35x/10.97x.
Risks:
1) Cloud revenue growth may fall short of expectations: If the increase in AI computing capacity or industry competition exceeds expectations, cloud revenue growth and margin improvement may slow;
2) Returns on AI investment may fall short of expectations: Persistently high capital expenditure, slower gross margin improvement, and slower substitution by proprietary chips would weigh on profit and free cash flow;
3) AI lab losses may narrow more slowly than expected: Investment in model training and promotion of the Qwen app will continue, and losses may improve more slowly than guided;
4) Competition in e-commerce and instant retail may intensify: Market share and unit economics may come under pressure from renewed subsidy competition across the industry;
5) Regulatory and geopolitical risks: Changes in platform regulation, cross-border tax policies, and chip export controls may affect the company's business and costs.

免责声明:以上内容仅供您参考和学习使用,任何投资建议均不作为您的投资依据;您需自主做出决策,自行承担风险和损失。九方智投提醒您,市场有风险,投资需谨慎。

推荐阅读

暂无数据

公司动态

    暂无数据

盘面综述

    暂无数据

IPO动态

    暂无数据

港股涨幅榜
  • 港股通
  • 红筹股
  • 国企股
  • 科技股
  • 名称/代码
  • 最新价
  • 涨跌幅

暂无数据

扫码关注

九方智投公众号

扫码关注

九方智投公众号