Report contents
At a glance
- Business model: Alibaba sells merchants distribution and enterprises metered cloud computing, and merchants, not shoppers, pay for the former through advertising and commissions. A sponsored placement adds revenue at almost no incremental cost, which is why commerce has funded group profit. Growth now comes from AI data centers and subsidized instant delivery, where each yuan of revenue consumes capital first.
- Financial quality: FY2026 net income of CNY103.6 billion converted poorly, with operating cash flow of CNY76.21 billion and free cash flow of negative CNY50.72 billion as capex reached 12.4% of revenue. Borrowings are only 13.7% of assets and Fitch affirmed an A rating in October 2026, so funding capacity is intact; the doubt sits in returns, with FY2026 ROE at 10.0%.
Business model
Business model: merchants still fund the AI buildout
Alibaba sells distribution and computing, and the two have different payers. Shoppers browse Taobao and Tmall without paying a fee; merchants do the paying, through sponsored placements and sales commissions that the company books as customer management revenue. Enterprises pay Alibaba Cloud for compute, storage and model access metered in tokens. Cross-border sellers pay commissions on the international marketplaces, restaurants and couriers settle through local services, and Cainiao bills for fulfillment. The user is often not the payer, and that gap is where the economics sit.
Call it a platform conglomerate: a mature, capital-light cash engine bolted to a young, capital-hungry one. In the fiscal 2026 annual report, published May 20, 2026, the chairman and CEO write that "vast numbers of AI agents are poised to take on an ever-greater share of work in the digital economy" and describe artificial general intelligence as a critical inflection point. That is a capital-allocation statement as much as a vision. Commerce profit is being asked to fund data centers, chips and model training whose payback depends on demand that has not yet been proven at Alibaba's scale.
Advertising carries the profit, servers carry the capital
Chinese retail advertising is the piece that matters most. At the incremental margin it behaves almost like software: a merchant bidding for a sponsored slot adds revenue with negligible added cost to the platform, which is why commerce has historically generated the bulk of group profit and nearly all of its free cash. Cloud runs the other way. Revenue is earned against servers, land, power and depreciation schedules, so each dollar of growth consumes capital before it returns any.
A shift in mix between those two changes what a percentage point of growth is worth. Group revenue growth therefore says less than it appears to, and anyone reading the top line as a health check is measuring the wrong thing. Alibaba also prices commerce through several separate levers (commissions, advertising, logistics fees) that differ by category, and it does not publish the split. Outside observers see blended customer management revenue; the take rate underneath has to be inferred. The company stopped reporting quarterly GMV, so share shifts against Pinduoduo, Douyin e-commerce and JD.com are visible only through third-party traffic estimates and through how fast advertising revenue grows relative to order volume.
The newest front, subsidized same-day delivery, sits at a third point on that spectrum: low contribution margin, dependent on order density, and expensive to defend because rivals can match a subsidy faster than they can match a merchant base.
Bargaining power follows demand aggregation
Alibaba's hold over merchants comes from access to buyers, not from the fee schedule. Sellers can list elsewhere and steer budgets toward whichever platform converts, so the ceiling on take rates is set by merchant returns rather than by Alibaba's pricing ambition. A seller whose margin after advertising disappears will cut spend or leave, and the platform loses the commission and the ad budget together. The 88VIP cohort is the counterweight: subscription-paying, high-frequency users who make Taobao costly to abandon and who give merchants a reason to keep bidding.
Cloud is a different contest. Enterprise buyers can spread workloads across several providers, and Chinese cloud pricing has been competitive, so the argument for Alibaba rests on model quality through the Qwen family and on embedding itself in customer data and workflows. That may work. So far the evidence shows demand for AI capacity growing; it does not yet show that Alibaba captures it at a price that covers the capital.
Both sides of the model share one fragility: the advertising pool funds everything else, and advertising is the first line to weaken if merchants earn less. The company's own framing, that agents will become the primary interface between people and digital services, points at the same risk from another direction. If shopping starts inside an assistant instead of a search box, the sponsored slot loses value as inventory, and the profit pool that pays for the buildout shrinks.
Two observations would settle the judgment. If customer management revenue keeps pace with order volume while merchants stay and fee waivers are rare, pricing power is intact and commerce is funding the buildout without eroding. If that revenue lags volume for several quarters and concessions reappear, merchant economics are being squeezed to buy growth. On the other side, if cloud revenue accelerates while capex as a share of cloud revenue flattens, the second engine is earning its capital; if capex keeps outrunning revenue, the commerce cash is buying capacity rather than profit.
Questions worth following
- Can Alibaba lift Chinese retail take rates without losing merchants?
Advertising and commissions from Taobao and Tmall merchants supply most group profit and cash. Whether merchants absorb higher fees or shift budgets to rivals determines if the commerce engine can keep funding cloud and delivery investment without margin erosion.
- Will Alibaba Cloud's AI workloads become durable enterprise demand?
Cloud is the second engine and the destination for most incremental capital. If AI capacity use converts into recurring enterprise contracts, group profit can diversify beyond advertising; if demand is largely training bursts or subsidized trials, the returns may not cover the buildout.
- Can Alibaba's subsidized instant retail push earn repeat orders?
Same-day delivery defends shopping frequency but carries thin contribution margins and invites matching subsidies from delivery rivals. If discounts convert into habitual orders, the spending buys a durable channel; if not, it drains the commerce profit that funds cloud.
Financial quality
Financial quality: earnings hold up, cash flow does not
Alibaba still earns a lot of money; it no longer converts much of it into cash. In FY2026, ended March 31, 2026, net income fell 20% to CNY103.6 billion, operating income fell 58% to CNY59.7 billion, and free cash flow swung from positive CNY78.2 billion to negative CNY50.7 billion as capex jumped. The balance sheet shows no strain: interest-bearing debt is 13.7% of assets, and Fitch affirmed an A rating with a stable outlook on October 14, 2026. The doubt sits in returns and cash conversion, not solvency.
| CNY bn unless stated | FY2022 | FY2023 | FY2024 | FY2025 | FY2026 | Q1 FY2027 (3 mo) | Assessment |
|---|---|---|---|---|---|---|---|
| Revenue | 853.1 | 868.7 | 941.2 | 996.4 | 1,023.7 | 268.2 | Adequate |
| Revenue growth | – | 1.8% | 8.3% | 5.9% | 2.7% | 8.3% | Adequate |
| Gross margin | 36.8% | 36.7% | 37.7% | 40.0% | 39.8% | 38.0% | Strong |
| Operating margin | 8.2% | 11.6% | 12.0% | 14.1% | 5.8% | 7.2% | Weak |
| Net margin | 7.3% | 8.4% | 8.5% | 13.1% | 10.1% | 3.9% | Adequate |
| Net income | 62.3 | 72.8 | 80.0 | 130.1 | 103.6 | 10.6 | Adequate |
| ROE | 5.7% | 6.6% | 7.2% | 12.3% | 10.0% | 1.0% | Adequate |
| ROA | 3.7% | 4.2% | 4.5% | 7.3% | 5.6% | 0.5% | Adequate |
| Liabilities/assets | 36.2% | 36.0% | 37.0% | 39.6% | 41.0% | 43.2% | Adequate |
| Interest-bearing debt/assets | 8.6% | 9.5% | 10.0% | 13.0% | 13.7% | 13.6% | Strong |
| Current ratio | 1.66 | 1.81 | 1.79 | 1.54 | 1.28 | 1.36 | Adequate |
| Operating cash flow | 142.8 | 199.8 | 184.0 | 164.8 | 76.2 | 23.0 | Weak |
| Capex (OCF less FCF) | 53.3 | 34.4 | 33.2 | 86.7 | 126.9 | 67.7 | Very intensive |
| Free cash flow | 89.5 | 165.4 | 150.8 | 78.2 | (50.7) | (44.7) | Weak |
| Capex/revenue | 6.2% | 4.0% | 3.5% | 8.7% | 12.4% | 25.2% | Very intensive |
| Operating cash flow/net income | 2.29 | 2.74 | 2.30 | 1.27 | 0.74 | 2.17 | Mixed |
| Asset turnover | 0.50 | 0.50 | 0.54 | 0.56 | 0.55 | 0.14 | Weak |
| Inventory turnover | 18.6 | 18.8 | 21.7 | 27.0 | 32.6 | n/a | Strong |
The Q1 FY2027 column covers a single quarter and is not comparable with fiscal-year flows.
| CNY bn unless stated | Q1 FY26 (Jun 2025) | Q2 FY26 (Sep 2025) | Q3 FY26 (Dec 2025) | Q4 FY26 (Mar 2026) | Q1 FY27 (Jun 2026) | Assessment | Trend |
|---|---|---|---|---|---|---|---|
| Revenue | 247.7 | 247.8 | 284.8 | 243.4 | 268.2 | Adequate | Recovering |
| Gross margin | 44.9% | 39.2% | 40.5% | 34.5% | 38.0% | Adequate | Lower year over year |
| Operating income | 35.0 | 5.4 | 20.2 | (0.9) | 19.3 | Weak | Stabilizing |
| Net income | 40.7 | 21.0 | 16.4 | 25.5 | 10.6 | Adequate | Falling |
| Total assets | 1,847 | 1,884 | 1,880 | 1,902 | 1,963 | – | Expanding |
| Total liabilities | 754 | 772 | 773 | 780 | 849 | – | Expanding |
| Net PP&E | 220.8 | 246.5 | 254.5 | 281.6 | 312.6 | – | Rising fast |
| Operating cash flow | 20.4 | 10.1 | 35.2 | 9.8 | 23.0 | Weak | Volatile |
| Capex | 38.1 | 31.5 | 28.5 | 27.8 | 67.7 | Very intensive | Accelerating |
| Free cash flow | (17.7) | (21.4) | 6.8 | (18.1) | (44.7) | Weak | Deteriorating |
Returns: a bigger asset base earning less per yuan
ROE of 10.0% in FY2026 compares with 12.3% a year earlier and has never exceeded 13% in the five years shown, and leverage is not inflating it. Total assets grew from CNY1.70 trillion in FY2022 to CNY1.96 trillion by June 2026 while revenue compounded at roughly 6% and asset turnover stayed near 0.55. Taxed operating income against reported debt, cash and equity gives a ROIC of roughly 4% for FY2026 against about 10% for FY2025; that estimate is depressed because Alibaba holds large short-term investments outside the cash line, but the direction is clear. Quick-commerce subsidies halved the operating margin to 5.8%, and AI infrastructure adds capital before it adds revenue.
The gross margin of 39.8% is high for internet retail, and the quarterly path carries more information than the annual average: 44.9% in the June 2025 quarter, 34.5% in the March 2026 quarter, 38.0% in June 2026. That is a 6.9 percentage point year-over-year decline, explained by mix shifting toward instant-commerce delivery and AI compute, both less gross-margin-rich per yuan than advertising. Part of the drag is the subsidy cycle; part is a permanent change in what Alibaba sells.
Cash: capex now outruns operating cash flow
Operating cash flow halved in FY2026. Capex rose from 3.5% of revenue in FY2024 to 12.4% in FY2026 and 25.2% in the June 2026 quarter, when free cash flow was negative CNY44.7 billion. Fitch estimates CNY220–230 billion of capex in each of FY2027 and FY2028 and expects free cash flow to stay negative until FY2029; management says AI capacity can pay back in about three years, a claim with no track record yet.
The longer record is better than the latest year. Across FY2022–FY2026, free cash flow totaled CNY433 billion against CNY449 billion of net income, so earnings did convert over the cycle, helped by an asset-light marketplace. Cash conversion above 2x in FY2022–FY2024 owed something to non-cash investment losses added back to operating cash flow, which means those ratios flatter the business; FY2026's 0.74 is the more honest reading of current operations.
Balance sheet: modest debt, thinner liquidity optics
Liabilities are 41% of assets, but most of that is merchant and supplier payables. Borrowings of CNY260 billion equal 13.7% of assets, and Fitch expects a sizeable net cash position through its forecast, helped by the HKD80 billion equity placement completed in August 2026. The current ratio fell to 1.28 from 1.79 two years earlier and the quick ratio is 1.24, both in thin-coverage territory; the squeeze comes from capex and a shift out of current investments, not from short-term borrowing. Goodwill of CNY246 billion is 13% of assets, mostly from acquisitions that have not required impairment.
What to check
- The earnings bridge. FY2026 net income of CNY103.6 billion exceeded operating income of CNY59.7 billion, and the March 2026 quarter paired a CNY0.85 billion operating loss with CNY25.5 billion of net income. Investment gains and fair-value moves account for the gap. If they recur, headline profit overstates earning power; if they do not, FY2027 net income will look far worse than FY2026 even if operations improve. The reported annual EBIT of CNY8.1 billion and EBITDA of CNY14.1 billion also fail to reconcile with operating income, which the 20-F bridge should settle.
- Depreciation and leases. Net PP&E rose from CNY220.8 billion to CNY312.6 billion in four quarters. Depreciation will follow, so the test is whether cloud revenue growth outpaces it. Fitch says it will adjust for leases if data-center leasing becomes material, which makes off-balance-sheet capacity worth tracking.
- Subsidy-dependent margins. Fitch puts e-commerce EBITA margin at 13% in FY2026 and 19% in the June 2026 quarter, against 24% before the instant-commerce fight. Watch whether order volume holds as incentives shrink.
- Disclosure consistency. Receivables appear as CNY76.2 billion at March 2025, CNY232.7 billion at September 2025 and CNY36.0 billion at March 2026, so scope changes may explain the swing. Check definitions before reading a working-capital trend into it, and compare share count with the placement and buyback pace.
Questions worth following
- Will Alibaba Cloud's AI workloads become durable enterprise demand?
Cloud is the second engine and the destination for most incremental capital. If AI capacity use converts into recurring enterprise contracts, group profit can diversify beyond advertising; if demand is largely training bursts or subsidized trials, the returns may not cover the buildout.
- Can Alibaba's subsidized instant retail push earn repeat orders?
Same-day delivery defends shopping frequency but carries thin contribution margins and invites matching subsidies from delivery rivals. If discounts convert into habitual orders, the spending buys a durable channel; if not, it drains the commerce profit that funds cloud.
- Is Alibaba's RMB 109 billion base cash flow defensible?
The model's starting cash flow exceeds the operating cash flow Alibaba generated in FY2026, before any capex. If the gap is not closed by a capex peak and margin recovery, the intrinsic value anchor is too high and today's premium is larger in substance than it looks.
Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.
Continue reading the full report
- Key risksWhat could break the investment case
- Valuation and margin of safetyWhat the current price assumes, and how much room for error it leaves
- SummaryOur overall judgement and what to watch next
Free to sign up · 20 research points · unlocking this report uses 10
20 points cover 2 full reports. No card needed.
Research reflects its generation date. Preview the first two modules.
More on Alibaba Group Holding Ltd
Alibaba Group Holding Ltd
Business, moat and key questions
Open company page