Alibaba
Marketplace take-rate (customer management revenue on Taobao/Tmall GMV) + cloud/AI infrastructure and model services (Alibaba Cloud, Qwen) + international commerce (AIDC), logistics (Cainiao) and local/instant-commerce services; monetization via merchant ad/commission + cloud usage.
Earnings, margins, COGS & capex
FY2026 revenue grew only 3% reported (11% like-for-like after divesting Sun Art and Intime), but the story is a deliberate margin sacrifice: income from operations fell 64% to a 5% margin (RMB50.2B) and the March-2026 quarter swung to a small operating loss of RMB848M. FY GAAP net income was RMB102.1B (~US$14.8B), down only 19% because mark-to-market equity gains cushioned it - the March-quarter GAAP net income actually rose 96% YoY to RMB23.5B on those gains. The cleaner read is non-GAAP net income, which strips the gains: it fell 62% to RMB60.7B (~US$8.8B) for the year and collapsed to near-zero (RMB86M) in the March quarter, while free cash flow turned to a ~RMB46.6B (~-US$6.8B) outflow. The offset is Cloud Intelligence Group reaccelerating to +34% for the year (+38% in the March quarter, external cloud +40%) with an ~US$5.2B annualized AI-product run rate and 11 straight quarters of triple-digit AI-product growth. Alibaba remains net-cash ~US$38B and is spending into a RMB380B (~US$53B) three-year AI/cloud buildout it now expects to overshoot.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~64¢ is cost of goods and ~31¢ operating expense, leaving ~5¢ of operating profit (~61¢ net).
Revenue trend
Margins
down sharply from 14% in FY2025 (income from operations -64%)
swung from RMB28.5B income (~12% margin) a year earlier
-56% YoY on quick-commerce + AI + user-experience investment
improving; cloud profitable and scaling with revenue
down hard on instant-commerce (Taobao Instant/Ele.me) subsidy war
depressed by investment cycle + operating-profit collapse; near-zero (RMB86M) in the March quarter
COGS structure
Cost of revenue rising faster than revenue - driven by cloud/AI infrastructure build (GPUs, data centers, energy) and by subsidies/fulfillment in quick/instant commerce; this is the primary driver of the gross- and operating-margin compression rather than a demand problem.
Capex
FY2026 capital expenditure RMB126.1B (~US$18.3B), up ~50% from RMB84.3B in FY2025, almost entirely AI + cloud infrastructure. Part of the announced RMB380B (~US$53B) three-year (FY2026-FY2028) AI/cloud infrastructure commitment - more than Alibaba's prior decade of AI/cloud spend combined - which management now signals it will likely exceed given surging data-center demand.
Latest earnings
Mixed/below on profit - revenue in line (+3%/+11% LFL) but operating result swung to a small loss (-RMB848M), non-GAAP net income fell ~100% to RMB86M and free cash flow was deeply negative (-RMB17.3B for the quarter); cloud beat on growth (+38%, external +40%). The December-2025 quarter had already shown a sharp profit decline on AI + quick-commerce spend.
No formal revenue guidance; management reaffirmed the RMB380B three-year AI/cloud capex commitment and signaled likely overshoot; framed FY2026-27 as a peak-investment period with margin recovery expected as instant-commerce losses narrow and cloud/AI scales. AGI framed by CEO Eddie Wu as a strategic, once-in-a-generation priority.
- AI-related product revenue
- 11th consecutive quarter of triple-digit YoY growth (RMB8,971M in the March quarter); ~US$5.2B annualized run rate
- Cloud Intelligence revenue growth (Mar-26 qtr)
- +38% YoY (external customers +40%)
- Non-GAAP net income (Mar-26 qtr)
- RMB86M (~US$12M), -100% YoY - the operating-profit trough
- Net cash position
- ~US$38B (2026-03-31)
- FY2026 free cash flow
- -RMB46.6B (~-US$6.8B)
Growth drivers
- Cloud Intelligence Group + Qwen AI models — the reacceleration engine (34-38% growth, external cloud +40%, ~$5.2B AI run rate, external AI-product demand)
- Customer management revenue (CMR) on Taobao/Tmall — FY +7% like-for-like (Mar-26 quarter +8% LFL); reported FY CMR +5% (quarter +1%) after a new contra-revenue business-development program, monetization via ads/commission/tech-service fee
- AI monetization — inference/API revenue (Model-as-a-Service), enterprise cloud migration, and AI woven into commerce (Qwen in Taobao/Tmall search & shopping)
- AIDC (AliExpress, Lazada, Trendyol, Miravia) international commerce, with EBITA loss narrowing sharply toward breakeven (FY EBITA -RMB2.1B, -86% loss narrowing)
- Instant/quick commerce (Taobao Instant Commerce, Ele.me) driving order frequency - currently a cost center in a price war, though unit economics and average order value improving
Bull & bear
BABA is a cheap, net-cash way to own China's leading AI-cloud franchise at the moment its growth reaccelerates, with the profit drag being a self-funded, temporary investment choice rather than structural decay.
- Cloud reaccelerating to 34-38% (external +40%) with an ~US$5.2B AI run rate and 11 quarters of triple-digit AI-product growth - the segment most likely to drive a re-rating, and it's already profitable (adjusted EBITA +35%).
- Valuation is undemanding: EV of only ~US$193B (market cap ~$231B less ~$38B net cash) on ~$148B revenue (~1.3x EV/sales) understates a leading cloud + a still-dominant commerce marketplace + the Ant/portfolio stakes.
- The margin collapse is discretionary spend (AI capex + instant-commerce subsidies), not demand loss - CMR still grew ~8% LFL in the March quarter (FY +7% LFL); as subsidies normalize and cloud scales, operating margin should recover off a trough.
- Fortress balance sheet (~$38B net cash) funds the RMB380B AI buildout while continuing buybacks/dividends - optionality most global peers can't match at this multiple.
- Qwen's traction (open-weight adoption, Apple Intelligence China) gives Alibaba genuine frontier-model credibility, a scarce asset that could compound cloud demand.
Alibaba is spending record capex into negative free cash flow to defend two businesses under structural attack, with geopolitical and compute-access risks that cap how much of the cloud upside it can actually capture.
- Group operating profit fell 64% and the March quarter turned to an operating loss with non-GAAP net income near-zero (RMB86M); free cash flow swung to a ~RMB46.6B outflow - the investment cycle could run longer and deeper than guided if the price war persists.
- Core China commerce keeps ceding share to PDD/Temu and Douyin; reported CMR was distorted by a new contra-revenue business-development program (FY CMR +5%/+7% LFL) and leans on ads + the take-rate/tech-service fee more than underlying GMV strength.
- US export controls on advanced GPUs structurally constrain the AI/cloud ambition - Alibaba may not be able to buy the compute its plan assumes, and domestic-chip substitutes lag the frontier.
- The instant-commerce/food-delivery war with Meituan and JD is a margin sink with no clear end and questionable strategic payoff.
- ADR/VIE structure plus US-China decoupling (delisting tail risk, tariffs) keep a persistent discount and cap the multiple regardless of operational execution.
What it is worth
Sum-of-the-parts / EV-to-sales cross-check. Market cap ~US$231B less ~US$38B net cash -> EV ~US$193B on ~US$148B FY2026 revenue (~1.3x EV/sales). Segments: a fast-growing, profitable Cloud/AI franchise that on standalone SaaS/cloud multiples could approach or exceed the group's entire EV; a still-dominant but decelerating China commerce cash engine; AIDC nearing breakeven; plus the Ant (~33%) and portfolio stakes. Reported P/E is distorted by the depressed FY2026 non-GAAP earnings trough.
~$75-95/ADS
investment cycle runs long with poor ROIC, continued commerce share loss, and a geopolitical/export-control or delisting shock keep the multiple compressed.
~$120-150/ADS
cloud grows but margins recover only gradually, commerce holds share modestly; persistent conglomerate/geopolitics discount.
~$180-220/ADS
cloud sustains 30%+ growth and re-rates on AI, commerce stabilizes, instant-commerce losses fade, margins recover; SOTP gap closes.
The multiple embeds a large China/geopolitics/ADR discount and skepticism that the AI capex earns its cost of capital. Re-rating hinges on (a) cloud sustaining >30% growth with improving ROIC, (b) instant-commerce losses normalizing so group operating margin recovers off trough, and (c) no delisting/geopolitical shock. Not financial advice; mainland-listed peers named for context only, not as buy/own calls.
SWOT
Strengths
- Leading China public-cloud provider with a full AI stack, including the widely adopted open-weight Qwen model family
- Dominant e-commerce marketplace franchise (Taobao/Tmall) with deep merchant relationships and CMR monetization
- Strong balance sheet: ~US$38B net cash funds the AI buildout, buybacks, and dividends simultaneously
- Diversified flywheel — commerce demand feeds cloud, cloud/AI improves commerce; plus logistics (Cainiao) and Ant affiliate
Weaknesses
- Near-zero group operating profit (Mar-26 quarter operating loss — non-GAAP net income ~RMB86M) and negative free cash flow during the investment peak
- Persistent share loss in core China commerce to PDD/Temu and Douyin e-commerce
- Instant-commerce/food-delivery price war (vs Meituan, JD) burning cash in the China E-commerce segment (EBITA -44%)
- Conglomerate complexity and ADR/VIE structure make the equity hard to value and carry a governance/geopolitical discount
Opportunities
- Monetize the AI capex cycle — enterprise cloud migration + inference/API (MaaS) revenue as China AI adoption accelerates
- Qwen as a de-facto China open-model standard (e.g. Apple Intelligence China partner) -> developer ecosystem lock-in
- AIDC international commerce reaching breakeven and scaling in emerging markets
- Margin recovery + potential re-rating as instant-commerce losses normalize and cloud mix rises
Threats
- US advanced-GPU export controls constraining compute for the AI/cloud ambition
- China consumer-demand softness and a prolonged domestic price/subsidy war
- US-China geopolitics: ADR delisting tail risk (HFCAA/PCAOB), tariffs, tech decoupling
- Competitive intensity from PDD/Temu, Douyin, Meituan, JD in commerce and Huawei/Tencent in cloud
Moats, dependencies & bottlenecks
Moats
Largest China public cloud; open-weight Qwen models widely adopted (incl. Apple Intelligence China). Compute access under export controls is the key durability risk.
Strong but eroding Deep merchant + consumer two-sided network and CMR monetization, but share is leaking to PDD/Temu and Douyin.
Cainiao logistics network and ~33% Ant Group stake extend the flywheel; Cainiao now reported within the reorganized segment structure.
Consumer + enterprise data scale improves ad targeting, recommendation, and AI training feedstock; 88VIP members surpassed 62M.
Dependencies
Supply / regulatory US export controls restrict access to top-end Nvidia (NVDA) GPUs; the entire RMB380B AI plan hinges on securing compute, pushing reliance on domestic chips (T-Head/Cambricon/Huawei) that trail the frontier.
Core commerce + local services are levered to China's uneven consumption recovery.
Post-2021 crackdown eased, but platform, data, and antitrust rules remain a swing factor.
HFCAA/PCAOB delisting tail risk, tariffs, and VIE structure create a standing discount for US holders.
~33% stake; Ant's performance and any eventual IPO materially affect BABA's other-income and portfolio value.
Advantages
- Leading China cloud with credible frontier open-weight models (Qwen) - a genuine AI franchise, not just infrastructure
- ~US$38B net cash lets Alibaba out-invest most rivals while still buying back stock
- Dual commerce + cloud/AI flywheel with proprietary data scale
- Global reach via AIDC (AliExpress, Lazada, Trendyol) approaching profitability
- Undemanding valuation relative to global cloud/AI peers
Weaknesses
- Near-zero group operating profit and negative FCF at the investment peak
- Structural share loss in core China commerce
- Cash-burning instant-commerce/food-delivery war with no clear end
- Conglomerate opacity + ADR/VIE structure depress the multiple
- Compute-access ceiling from export controls
Bottlenecks
- Access to advanced GPUs/compute under US export controls - the binding constraint on the AI/cloud ambition
- Operating-margin compression from simultaneous AI capex and instant-commerce subsidies
- China consumer-demand softness limiting core-commerce GMV growth
- Competitive intensity in core China commerce (PDD/Temu, Douyin) capping take-rate power
- Negative free cash flow during the peak-investment window
Top signals & trends
Top signals
The clearest evidence the AI investment is converting to revenue; watch quarter-over-quarter durability.
Investment cycle is deeper than many expected; the March quarter marks the profit trough and its recovery timing is the key debate.
Bullish on ambition/demand read; bearish on near-term FCF and ROIC uncertainty.
External validation of model quality; potential ecosystem/demand catalyst.
Signals an intensifying price war with Meituan/JD; watch for subsidy discipline and improving unit economics.
Capital return continues even through the investment peak.
Trends
Hyperscaler-style buildout compresses near-term margins/FCF but is the source of cloud reacceleration; ROIC is the open question.
Alibaba, Meituan and JD subsidizing orders; near-term margin sink across the sector.
Enterprise migration + inference demand; Alibaba is the share leader positioned to capture it.
Constrains compute and keeps an ADR delisting/geopolitics discount on the stock.
Caps core-commerce GMV; policy stimulus is the swing factor.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
AI GPU supplier - access constrained by US export controls, the key bottleneck for the AI/cloud buildout.
Foundry for AI/custom silicon (incl. Alibaba's own T-Head designs), subject to export restrictions.
Shanghai-listed (688256.SS) domestic AI-chip supplier - a substitute as advanced foreign GPUs are restricted.
In-house silicon (Yitian CPU, proprietary inference chips) reducing dependence on foreign chips.
Taobao/Tmall sellers paying CMR (ads/commission/tech-service fee) - the core monetization base.
Enterprises & developers (Alibaba Cloud / Qwen) Cloud compute + AI-model/API (Model Studio / MaaS) customers; the fastest-growing revenue pool.
Uses Qwen to power Apple Intelligence in China - a marquee cloud/AI customer-partner.
~1B+ annual active consumers across Taobao/Tmall, Ele.me, Amap; 88VIP members surpassed 62M.
Pinduoduo (China) + Temu (global) - the primary share-taker in value e-commerce and cross-border.
1P-led China e-commerce + logistics; now also a combatant in the instant-commerce/food-delivery war.
Global e-commerce + AWS cloud; competes with AIDC internationally and with Alibaba Cloud outside China.
Azure + OpenAI set the global cloud-AI benchmark; competes for multinational cloud workloads.
WeChat ecosystem, Tencent Cloud, and gaming; rival in China cloud and consumer attention (HK:700).
Private. Douyin/TikTok e-commerce is a fast-growing threat to Taobao/Tmall GMV and ad budgets.
Dominant China local services/food delivery; direct opponent in the instant-commerce war (HK:3690).
AI Cloud (Ernie) competitor in China; smaller cloud share.
Shopee competes with AIDC/Lazada in Southeast Asia.