Alibaba: March-Quarter Cloud Revenue Rose 38% While Consolidated Adjusted EBITA Fell 84%, Free Cash Flow Turned to a RMB 17.3 Billion Outflow, and USD 122.25 Sits Above Conservative Value

Alibaba does not disclose quick-commerce contribution profit, subsidy expense, delivery cost per order, rider cost, merchant take rate or orders in absolute units. Management disclosed that March-quarter order volume was 2.7 times the prior-year level and that non-food orders were three times the prior year, while saying unit economics had improved. The filing also states that quick-commerce revenue is reported net of subsidies treated as contra-revenue. These disclosures are enough to establish rapid scale and improving direction, but they do not establish actual per-order profitability. Any precise subsidy run-rate presented by an outside analyst is therefore an estimate rather than a reported fact.

Cloud’s reported growth is equally real, but its economics need more work than the revenue headline suggests. FY2026 Cloud Intelligence revenue rose 34% to RMB 158.1 billion, or USD 23.4 billion at the July 31 exchange rate, while adjusted EBITA rose 35% to RMB 14.3 billion. That produces a 9.0% adjusted EBITA margin. Cloud depreciation, equipment impairment and relevant lease costs were RMB 28.9 billion, up from RMB 15.9 billion in FY2025 and more than twice segment adjusted EBITA. Group capital expenditure rose to RMB 126.1 billion from RMB 86.0 billion. Alibaba does not publish cloud capital expenditure or segment invested capital, so an incremental cloud ROIC cannot be calculated from public information.

The reported 30% AI mix applies to external cloud revenue, whose absolute amount is not disclosed. That puts a ceiling on March-quarter AI revenue at 30% of total cloud revenue, RMB 12.5 billion, approximately USD 1.85 billion. Actual revenue must be lower because the denominator includes internal cloud revenue. Alibaba also does not disclose AI-product gross margin. The evidence supports an AI revenue run rate below USD 7.4 billion annualised; it does not support an exact number or the claim that this revenue already earns attractive returns.

The balance sheet can fund the experiment. Alibaba ended March 2026 with RMB 520.8 billion of cash and other liquid investments, about USD 77.1 billion at the July 31 exchange rate. It also owned RMB 206.8 billion of equity-method investments and RMB 125.0 billion of private investments, although these portfolios require tax, liquidity and strategic haircuts before being treated as shareholder cash. Bank borrowings were about RMB 71.8 billion, and the group also had unsecured, convertible and exchangeable notes. Liquidity risk is low. Capital-allocation risk is materially higher because operating cash generation has fallen while cloud, delivery and AI spending have accelerated.

Over five years the vertical picture is stark. Revenue increased from RMB 853.1 billion in FY2022 to RMB 1,023.7 billion in FY2026, a compound rate of only about 4.7%. Operating income rose through FY2025 before falling 64% in FY2026. Net income remained volatile because Alibaba’s investment portfolio produces gains, losses and impairments that are not tightly connected to current operations. Cumulative operating cash flow over the five fiscal years was approximately 1.86 times cumulative net income, showing strong historical conversion. In FY2026 alone the ratio fell to 0.75 times, as subsidies, working-capital movements and investment spending consumed cash.

Alibaba’s historical success came from a genuine capability: it repeatedly built platform infrastructure around fragmented Chinese merchants, converted consumer traffic into advertising revenue, and then used the resulting cash to enter adjacent markets. Taobao’s defeat of eBay in China, Alipay’s role in trust formation, the rise of Tmall and Alimama, Cainiao’s logistics network and Alibaba Cloud all grew from that capability. The record becomes weaker when Alibaba owns inventory, operates physical services, subsidises fulfilment or acquires mature businesses. The divergence between marketplace economics and operating-business economics is the central fact of Alibaba’s portfolio.

Horizontal analysis reinforces the distinction. PDD has become the low-price and algorithmic-sourcing specialist, JD.com the controlled-inventory, fulfilment and service-reliability retailer. Meituan owns the deepest local-service supply and rider-density network. Alibaba remains the broadest merchant and consumer platform, with unmatched cross-category traffic and the strongest ability to fund a response. That breadth creates options, and it also encourages management to defend too many fronts at once.

In cloud, Alibaba is China’s scale leader, but the global return benchmark is much higher than its current economics. Amazon Web Services shows what high utilisation, premium services and sustained enterprise relationships can eventually produce. Microsoft illustrates how cloud can monetise AI across existing software relationships. Baidu and Tencent provide closer Chinese technology comparisons, although neither discloses directly comparable cloud segment economics. Alibaba’s 9% cloud adjusted EBITA margin and elevated depreciation burden indicate that the segment has not yet earned hyperscaler-quality returns. The investment case requires more than revenue growth: utilisation, pricing and cash returns must follow.

The new geopolitical facts warrant a distinct discount. The Department of Defense added Alibaba to the Section 1260H list in June 2026; Alibaba sued on June 23, denying military affiliation, and obtained interim relief in early July from certain lobbying-related consequences while the case proceeds. A 1260H designation is not an OFAC asset freeze, a BIS Entity List export ban or an automatic prohibition on U.S. investors owning BABA. It can restrict defense procurement and create compliance, lobbying, counterparty and institutional-ownership consequences, with additional restrictions contingent on statutes, implementing rules and litigation.

This risk is separate from the VIE structure. BABA investors own ADSs in a Cayman Islands holding company. Regulated Chinese operating assets are partly accessed through contractual arrangements with variable-interest entities rather than direct equity ownership. Alibaba states that a significant majority of its assets, revenue and cash flows sit in directly owned subsidiaries, while its major VIEs generated about RMB 119.6 billion of third-party revenue in FY2026, roughly 12% of consolidated revenue. The VIE exposure is narrower than the shorthand “investors own nothing in China,” but the contractual and enforcement risk is real.

The U.S. Department of Justice has now confirmed the reported USD 600 million pharmaceutical settlement. Alibaba’s portion consists of USD 200 million of forfeiture and a USD 125 million criminal penalty; an Ant-related entity bears the remaining USD 275 million. The direct Alibaba payment is small beside its liquid resources, but the conduct occurred over multiple years and raises a compliance-quality question rather than a solvency question.

Reported access to Nvidia H200 accelerators remains less certain. Reuters reported conditional Chinese approval and U.S. licensing progress involving Alibaba and other firms, while Nvidia stated that only limited licences had been issued and that no related revenue had yet been recognised at its fiscal year-end. Subsequent reporting described limited shipments and possible permission for Alibaba, but Alibaba has not disclosed confirmed H200 volumes, delivery dates or purchase commitments. Its T-Head accelerators are reportedly in production at scale, yet no public data establishes their benchmark performance, deployment volume or share of Alibaba’s training and inference workloads.

Compared with the prior May 19 report, three conclusions change. Cloud evidence is stronger: external growth accelerated to 40% and AI reached a meaningful share of revenue. Cash-flow evidence is worse: the predicted investment phase arrived faster and with a deeper profit collapse than the previous report’s ranges implied. External risk is also worse because Section 1260H and the confirmed DOJ settlement were not available to the prior work. The current share price is approximately 8% below that report’s USD 133.26 reference, partly offsetting the deterioration.

The twelve-month and long-term views disagree. The next year is dominated by subsidy discipline, depreciation, capex and legal headlines. The three-to-five-year outcome can be much better if quick commerce preserves traffic without permanently resetting marketplace margins, and if cloud’s current capital build produces utilisation and pricing rather than stranded depreciation. Alibaba therefore deserves neither a pure distressed multiple nor an unqualified AI premium. It is a self-funded transition whose outcome remains measurable but unproven.

Vertical history and financial evolution

Alibaba was founded in Hangzhou in 1999 by eighteen people led by Jack Ma, a former English teacher. Its first product was an online wholesale marketplace connecting Chinese exporters and small businesses with overseas buyers. The initial problem was not simply product discovery. Small Chinese suppliers lacked international distribution, payments, trust mechanisms and affordable marketing. Alibaba’s early model aggregated fragmented supply and charged for business services rather than carrying inventory.

Joe Tsai joined the founding team from private equity and became the financial and legal architect behind a company whose commercial ambition initially ran far ahead of its institutional structure. The combination mattered. Ma supplied mission, recruiting and merchant intuition; Tsai supplied capital formation, governance and foreign-investor access. The early financing base included SoftBank, Goldman Sachs and other outside investors, and Yahoo later invested USD 1 billion plus its China operations in 2005. That capital allowed Alibaba to fight competitors without demanding immediate profitability from every product.

The early decisive turn came with Taobao in 2003. eBay had bought EachNet and approached China as an extension of a global auction platform. Alibaba made listing free, localised the user experience, integrated messaging and relied on Alipay’s escrow mechanism to overcome low consumer trust. Taobao’s free model sacrificed immediate revenue but moved liquidity and merchants to Alibaba’s platform. The lasting lesson: Alibaba could subsidise a new marketplace and monetise the traffic later through advertising and merchant tools.

Tmall separated branded retail from Taobao’s consumer-to-consumer origins. Alimama turned merchant competition for visibility into a high-margin advertising market. Singles’ Day became both a demand event and an infrastructure stress test. Cainiao, founded in 2013, coordinated logistics information and partners rather than initially replacing them with a fully owned delivery system. Alibaba Cloud transformed infrastructure built for the commerce peaks into an external computing product. These moves shared one economic pattern: Alibaba built common infrastructure and collected a toll from many merchants or enterprises rather than manufacturing the products sold.

Control shaped the listing path. Alibaba.com, the B2B subsidiary, listed in Hong Kong in 2007 and was taken private in 2012. The group listed ADSs on the NYSE on September 19, 2014 at USD 68 per ADS and raised approximately USD 25 billion, then the world’s largest IPO. The prospectus emphasised a platform connecting consumers, merchants and service providers, while preserving the Alibaba Partnership’s ability to nominate a majority of directors.

The partnership structure remains more consequential than ordinary founder ownership. Alibaba has one class of shares with one vote per share, but the partnership may nominate or, in specified circumstances, appoint a simple majority of the board. Changing the relevant articles requires a 95% vote of shareholders present. As of the FY2026 report, four of ten directors were partnership nominees and six were independent. This is not conventional dual-class equity, but outside shareholders still lack ordinary board-control rights.

Alibaba returned to Hong Kong in November 2019, raising approximately USD 11.2 billion. It later sought dual-primary status, increasing access for Asian investors and providing an alternative trading venue to the NYSE. The ADSs and Hong Kong ordinary shares remain fungible.

Alibaba’s post-IPO history divides into four useful stages.

The first was platform compounding, from 2014 through roughly 2017. China’s mobile-commerce penetration, digital payments and online advertising expanded together. Taobao and Tmall acquired users, merchants bought more traffic, and the marketplace needed relatively little incremental physical capital. Alibaba used the cash to fund cloud, logistics, digital media and local services. Revenue expanded rapidly, margins remained high by retail standards, and the market valued Alibaba primarily as a Chinese consumer-internet growth platform.

The second was ecosystem expansion, from 2017 through 2020. Alibaba increased control of Cainiao, acquired Ele.me, invested in Sun Art, built Freshippo, expanded media and entertainment, and pushed Lazada and other international businesses. Reported revenue grew faster because direct commerce and logistics were booked gross, but consolidated margins diluted. The market nevertheless gave Alibaba a high multiple because investors treated loss-making adjacencies as future monopolies and placed a high value on its stake in Ant Group.

Ant’s intended 2020 IPO marked the high-water mark. The listing was expected to be the largest in history. Chinese regulators suspended it in November 2020, causing Alibaba’s U.S. market value to fall by about USD 76 billion in the immediate reaction. In April 2021, China’s competition authority fined Alibaba RMB 18.228 billion, 4% of relevant 2019 domestic revenue, for merchant exclusivity practices. The Ant suspension and antitrust case changed the market’s view of Alibaba from a private platform able to compound beyond normal constraints into a regulated national institution whose commercial freedom had boundaries.

The third stage, from late 2020 through 2023, was regulatory compression and competitive catch-up. Alibaba’s valuation fell by more than 70% from its 2020 level as PDD gained value-conscious users, Douyin and other content platforms captured merchant advertising, JD improved marketplace penetration, and China’s consumption and property cycles weakened. The market ceased capitalising every adjacency at venture-style values. It began subtracting losses from consolidated earnings.

Alibaba responded in March 2023 with a six-group restructuring intended to make businesses independently accountable and potentially listable. Joe Tsai became chairman and co-founder Eddie Wu became chief executive in September 2023. Wu had been Alibaba’s technology director at inception, Alipay’s chief technology officer and a key architect of Alimama. The leadership change brought a founder-technologist back to the centre at the moment AI infrastructure became strategic.

The proposed Cloud Intelligence spin-off was abandoned in November 2023. Management cited uncertainty created by U.S. restrictions on advanced AI chips. This was a genuine strategic reversal. It showed that the cloud asset could not be valued independently of Alibaba’s group purchasing power, internal workloads, chip inventory and geopolitical exposure.

The fourth stage began with the “user first, AI-driven” refocus and culminated in FY2026. Alibaba disposed of Sun Art and Intime, simplified segment reporting, raised dividends and repurchased shares, while concentrating new spending on cloud, AI models and instant retail. Chinese regulators said in August 2024 that Alibaba had completed the three-year rectification period following the antitrust penalty, reducing one domestic policy overhang. The operating model nevertheless became more capital intensive because the new priorities require data centres, accelerators, delivery density and consumer subsidies.

That change shows up in the vertical financial record.

The audited financial summary confirms revenue, operating profit, net income and balance-sheet figures; the operating-cash-flow and capex series comes from Alibaba’s annual cash-flow disclosures. FY2026 figures are for the year ended March 31, 2026.

The business explanation matters more than the revenue CAGR. Marketplace monetisation matured, direct retail and logistics raised reported revenue but lowered margins, and divestitures later removed low-margin sales. FY2026 reported revenue grew 3%, while like-for-like revenue excluding disposed businesses grew about 11%. The disposal-adjusted figure better measures operating demand; the reported figure better measures the cash flow and earnings available to shareholders.

Operating profit had more than doubled between FY2022 and FY2025 before falling by RMB 90.8 billion in FY2026. Non-GAAP net income fell from RMB 158.1 billion to RMB 60.7 billion. The decline was larger than the revenue slowdown because Alibaba spent against two separate objectives: quick-commerce share and AI capability. Neither objective is represented by a single accounting line.

Cash quality was historically stronger than earnings quality. Over FY2022–FY2026, operating cash flow was approximately RMB 764.8 billion against roughly RMB 412.1 billion of net income, a ratio of about 1.86 times. Investment revaluations and impairments made net income volatile, while marketplace working capital and non-cash expenses supported cash generation. FY2026 reversed the pattern: operating cash flow of RMB 76.2 billion represented only 0.75 times net income, and was less than capital expenditure. Alibaba attributed the decline primarily to quick-commerce investment and higher cloud infrastructure expenditure.

The FY2026 cash-flow decline also contains a RMB 67.9 billion increase in prepayments, receivables, other assets and long-term licensed copyrights, partly offset by RMB 37.9 billion of higher accrued expenses and payables. Some cash pressure may reverse. The subsidy and infrastructure components will persist as long as management prioritises scale over current return.

Capital returns have become less generous as investment needs increased. Alibaba paid RMB 33.7 billion of dividends in FY2026 but repurchased only about USD 1 billion of shares, far below FY2025’s repurchase pace. The company also issued convertible and exchangeable debt, partly to fund cloud and international commerce. Buybacks remain valuable when made below intrinsic value, yet the decline in repurchase scale shows that management no longer regards excess cash as abundant relative to its ambitions.

Price history follows the narrative shifts. The 2014–2020 rise capitalised Chinese mobile consumption, platform economics and Ant Group; the immediate post-Ant decline capitalised policy risk. The 2021–2022 decline added competition, weak Chinese consumption and foreign-listing anxiety. Restructuring in 2023 produced a short-lived asset-separation re-rating before the cloud spin-off reversal, and the 2024–2025 recovery reflected buybacks, policy stabilisation and AI expectations. The ADS then reached a 52-week high of USD 192.67 before retreating to USD 122.25 on July 31, 2026.

The July 31 rise of 5.1% followed reports that Moonshot AI had used a large block of Nvidia Hopper-generation computing capacity supplied through Alibaba. Alibaba disputed descriptions that the chips were H200s. It also came during strong global cloud earnings and a broader AI trade. The evidence supports a sentiment re-rating toward scarce Chinese computing capacity, not proof that Alibaba’s cloud return on capital changed in one day.

Business model, moat, industry and competitors

Alibaba reported four segments in FY2026 after combining Taobao, Tmall, Ele.me and Fliggy into China E-commerce and moving Cainiao, Amap and media into “All others.” This restatement improves strategic visibility but limits comparability with older segment histories.

“All others” generated RMB 254.4 billion of revenue and an adjusted EBITA loss of RMB 35.7 billion. It is omitted from the narrow table because its direct retail, logistics, media, maps, healthcare, Qwen, games and enterprise-collaboration businesses do not share one economic model. Consolidated FY2026 revenue was RMB 1,023.7 billion after RMB 89.6 billion of intersegment eliminations.

China e-commerce contains three different machines. Customer-management revenue is an advertising and software toll on merchant competition for traffic. Direct retail books merchandise and logistics revenue gross, which gives it retail-like margins. Quick commerce combines platform commissions, delivery and advertising, with subsidies deducted from reported revenue. Segment margins can therefore fall even when customer-management revenue grows, because delivery and direct retail dilute the marketplace economics.

Fixed costs include software engineering, data centres, corporate infrastructure, warehouses, logistics systems and permanent delivery capacity. Variable costs include product cost in direct retail, fulfilment, payment processing, traffic acquisition, sales commissions, customer incentives and rider expenditure. Marketplace advertising has strong operating leverage; direct commerce has modest leverage. Quick commerce can show negative leverage during a land grab because more orders initially bring more subsidies, rider costs and merchant support.

Cloud has high fixed costs in data centres, networking, accelerators, software development and enterprise sales. Once utilisation rises, incremental workload can carry high contribution margins. Before utilisation rises, depreciation continues regardless of demand. Alibaba’s cloud revenue acceleration and rising adjusted EBITA are encouraging; the 82% rise in segment depreciation-related expense shows that the utilisation test is still ahead.

The first enduring moat is merchant and consumer liquidity. Taobao and Tmall offer broad supply, traffic and merchant tooling across nearly every retail category. A merchant can access search demand, live commerce, advertising, payments, logistics and customer data from one operating environment. Consumers can move between branded retail, small merchants, used goods and instant delivery. This moat survived the regulatory period and aggressive competition, though Alibaba’s former near-monopoly did not.

Monetisation infrastructure is the second moat. Alimama converts merchant bids, consumer intent and transaction history into customer-management revenue. Alibaba can improve take rate through advertising relevance and merchant software without taking inventory risk. The March-quarter like-for-like customer-management growth of 8% suggests that monetisation improved despite modest core-commerce revenue growth.

The third moat is internal demand for cloud and AI. Alibaba’s commerce, logistics, mapping, advertising and consumer applications create large workloads and data-centre utilisation. This helps justify infrastructure earlier than an independent cloud startup could. The same structure creates a disclosure problem: internal demand can support revenue and utilisation without proving external product-market fit. External-customer growth of approximately 40% in the March quarter is therefore more important than total cloud growth.

Financial capacity is the fourth moat. RMB 520.8 billion of liquid resources lets Alibaba absorb a subsidy war and infrastructure build that smaller companies cannot fund. Capital is a defensive moat, not evidence that spending will earn its cost. PDD, JD and Meituan also possess substantial financial resources, so Alibaba cannot win solely by outlasting thinly capitalised rivals.

Brand alone is a weaker moat than it was ten years ago. Chinese consumers routinely use multiple applications. Merchants allocate spending across Alibaba, PDD, JD, Douyin, Kuaishou and local-service platforms according to conversion and cost. Switching costs exist in data, ratings and operating tools, but multi-homing is normal. Alibaba’s moat rests on network scale and monetisation capability rather than exclusive loyalty.

Management is technically credible and financially experienced. Eddie Wu is a co-founder with direct experience in technology and monetisation. Joe Tsai built Alibaba’s legal and capital structure. CFO Toby Xu has served since April 2022. The team correctly narrowed strategic priorities and disposed of several low-return retail assets. It also deserves scrutiny for launching two large investment programmes at once, reducing the ability of shareholders to judge either one cleanly.

Capital-allocation history is mixed. Alibaba Cloud and Taobao were exceptional internal creations. Cainiao built strategic infrastructure. The acquisitions of Sun Art, Intime, media assets and several local-service businesses generated weaker returns. Large buybacks during depressed valuations reduced the share count, but FY2026 debt issuance and reduced repurchases reveal a renewed preference for internal investment.

China’s commerce industry is mature in user penetration and remains capable of modest nominal growth through higher online share, advertising monetisation and new fulfilment formats. Profit growth is more difficult because consumer traffic has fragmented and merchant acquisition has become competitive. PDD took share through price and social discovery. JD competes on inventory authenticity, delivery and after-sales service. Content platforms compete for product discovery. Quick commerce adds a new fulfilment layer but does not create unlimited incremental consumer spending.

Local delivery is a density business. A platform needs restaurant and retail supply, frequent orders, rider coverage, dispatch software and short travel distances. Meituan entered the current conflict with the strongest density and merchant relationships. Alibaba brought broader e-commerce traffic, Ele.me supply, Taobao membership and greater cross-category purchasing data; JD brought brand trust and fulfilment capacity, but a weaker restaurant network.

Management’s strategic rationale is understandable. Consumer shopping journeys may migrate from next-day parcels toward sub-hour delivery for groceries, electronics, medicine, meals and daily necessities. If Taobao lacks instant fulfilment, Meituan can move from food into retail and take product-search traffic from Alibaba. Quick commerce is therefore partly offensive and partly defensive.

Market structure is where the economic risk lies. Three well-capitalised platforms can maintain subsidies for longer than one. The likely end state is an oligopoly rather than a winner-take-all monopoly. Rational pricing could leave Meituan strongest in meals and local services, Alibaba strongest in integrated commerce, and JD strongest in selected high-value retail. Permanent coupons, high rider costs and low merchant take rates would transfer much of the value to consumers instead.

The commerce peer comparison illustrates why one multiple cannot describe the group.

Market data are as of July 31, 2026. Latest operating figures refer to Alibaba’s March 2026 quarter, Q4 FY2026; JD’s March 2026 quarter; and PDD’s March 2026 quarter. Peer earnings are not directly comparable because Alibaba has large investment gains, JD books first-party retail gross, and PDD consolidates Temu and Pinduoduo.

PDD became the most focused commerce operator. Its algorithmic discovery, low-price supply and Temu expansion produced higher recent growth and a lower reported P/E. Customers choose it for price and novelty. Merchants accept its demanding economics to access incremental volume. Its low multiple reflects global trade risk, competition and management’s warning that investment will pressure profitability.

JD became the service-and-fulfilment specialist. Customers choose it for delivery certainty, authentic branded goods and after-sales service. Its first-party model consumes more working capital and produces lower gross margins than Alibaba’s marketplace. JD’s quick-commerce expansion adds another loss pool to an already asset-heavy model. First-quarter 2026 revenue rose 4.9% to RMB 315.7 billion, while non-GAAP net income fell to RMB 7.4 billion from RMB 12.8 billion.

Meituan became China’s local operating system for meals, hotels, services and instant retail. Customers choose it because supply density and delivery reliability are difficult to reproduce. It reported March-quarter 2026 revenue of about RMB 91.0 billion and a net loss of roughly RMB 6.8 billion as the price war continued. Management has described current competition as unsustainable. Meituan’s losses validate the severity of Alibaba’s challenge but do not prove Alibaba is winning economically.

Alibaba’s ecological niche is the broad platform. It can connect long-tail merchants, brands, logistics, payments, advertising, cloud and AI across the largest number of categories. It is less operationally specialised than JD or Meituan and less narrowly cost-focused than PDD. This makes Alibaba strongest when value comes from network coordination and weakest when value comes from owning and executing every physical step.

International commerce now contributes scale without being a major drain. AliExpress uses Chinese supply for cross-border trade; Trendyol has strong regional positions; Lazada competes in Southeast Asia against Sea’s Shopee and TikTok Shop. FY2026 AIDC revenue rose 9% and its adjusted EBITA loss narrowed from RMB 15.1 billion to RMB 2.1 billion. The improvement appears to come from logistics optimisation and spending discipline rather than explosive growth.

Cloud competition must be assessed separately. Tencent bundles cloud with communications, payments and enterprise relationships. Baidu links AI cloud to search, autonomous driving and foundation models. Huawei combines cloud, hardware and domestic semiconductor capability. Alibaba has the broadest Chinese public-cloud scale and a large external customer base. Its Qwen model family also gives it a proprietary demand generator.

Global peers set the returns benchmark. AWS reported strong growth and far higher operating margins than Alibaba Cloud, reflecting mature utilisation and premium services. Microsoft distributes Azure AI through enterprise software relationships. Alibaba does not need to match their margins immediately, but it must show that each new RMB of infrastructure creates future cash flow rather than merely revenue. The FY2026 9% adjusted EBITA margin and RMB 28.9 billion depreciation burden leave that question open.

AI workloads, enterprise digitisation, data-sovereignty requirements, semiconductor controls and government policy drive China’s cloud cycle. Domestic providers benefit from local data rules and customer preference for Chinese infrastructure. They face a lower supply ceiling for advanced accelerators and weaker access to global enterprise software ecosystems. Export controls can raise prices, delay deployment and shorten the useful economic life of available equipment.

Alibaba belongs to several overlapping cycles: Chinese consumption, digital advertising, policy, AI infrastructure and technology iteration. Marketplace demand is tied to household confidence and merchant advertising budgets. Quick commerce is tied to competitive intensity more than macro demand. Cloud is tied to enterprise capex and AI workloads. The ADS multiple is tied to U.S.–China relations, institutional mandates and long-duration interest rates.

The July 31 U.S. ten-year Treasury yield was approximately 4.74%–4.75%. A high risk-free rate raises the return required from a Chinese VIE with geopolitical exposure and reduces the value of cash flows expected far in the future.

Current fundamentals and the live debate

Alibaba’s last four reported quarters describe a deliberate deterioration in current earnings.

In the June 2025 quarter, Q1 FY2026, quick-commerce investment was still ramping and cloud growth was strengthening. The September 2025 quarter, Q2 FY2026, brought a rise of about 60% in quick-commerce revenue, but investment reduced adjusted earnings. Management said order mix and average order value were improving.

In the December 2025 quarter, Q3 FY2026, revenue rose only 2% to RMB 284.8 billion and adjusted EBITA fell 57% to RMB 23.4 billion. Cloud revenue grew 36%, while spending on delivery, user experience and technology overwhelmed the improvement. The ADS fell more than 6% following the release because adjusted earnings missed expectations.

The March 2026 quarter, Q4 FY2026, continued the like-for-like revenue acceleration and took cloud growth to 38%, but adjusted EBITA fell another 84%. This sequence establishes that the investment phase is neither a one-quarter marketing campaign nor an accounting anomaly. It has become the primary determinant of group earnings.

Figures are for the quarter ended March 31, 2026, Q4 FY2026, and the corresponding prior-year quarter. Rounding causes minor differences.

The expense bridge supports the segment bridge. Cost of revenue rose to 65.5% of sales from 61.6%, and product development to 7.8% from 6.3%. Sales and marketing rose to 21.9% from 15.3%, an increase of roughly RMB 17 billion. Alibaba attributed the marketing increase mainly to quick commerce and Qwen user acquisition, and the product-development increase to personnel and technology infrastructure.

The FY2026 collapse was principally an expensed customer-acquisition and service-cost shock, accompanied by a separate capital-intensive cloud build. Depreciation has begun to rise, but the March-quarter profit decline cannot be explained mainly by future cloud depreciation. The largest immediate EBITA losses sit in China commerce and “All others,” where subsidies, promotions, delivery investment, Qwen distribution and technology spending are recognised now.

Quick commerce generated RMB 20.0 billion of March-quarter revenue, up 57%. Order volume was 2.7 times the prior year. Dividing the revenue by an unknown number of orders would not yield contribution profit because revenue is net of some subsidies and includes different order types, commissions and service fees. The public record therefore does not permit a reliable per-order result.

A plausible upper bound can be inferred, but it must not be mistaken for disclosure. China e-commerce EBITA fell RMB 15.7 billion year over year even though like-for-like customer-management revenue grew 8%. Some portion reflects quick commerce, merchant support and broader user experience. The entire decline cannot be assigned to delivery because management does not separate those costs. Group sales and marketing rose approximately RMB 17 billion, but that line also includes Qwen and other acquisition spending.

Management has indicated a path toward positive unit economics by the end of FY2027. Positive unit economics would be a useful first checkpoint, but its definition matters. Breakeven after direct delivery and subsidy cost can still exclude central technology, permanent sales staff, depreciation and historic customer-acquisition expenditure. A satisfactory outcome requires both order-level breakeven and recovery in China e-commerce segment EBITA.

Cloud revenue of RMB 41.6 billion implies annualised revenue near RMB 166.5 billion before seasonality. External growth of 40% suggests the acceleration is not merely transfer pricing from other Alibaba units. AI products representing 30% of external revenue establish material scale. The missing disclosures are AI gross margin, external-cloud revenue in absolute terms, capacity utilisation and segment capex.

Management has said AI-related cloud revenue could exceed half of external cloud revenue within roughly a year and has set a long-range ambition for cloud-and-AI revenue to exceed USD 100 billion. The current run rate remains far below that objective, and the capital required to reach it is substantial. Alibaba announced an intended three-year AI and cloud investment of up to RMB 380 billion and has suggested it may invest more if demand supports the return.

Qwen is both a product and a cost centre. Alibaba can monetise it through token consumption in cloud, enterprise deployment, advertising relevance, shopping assistance and consumer applications. Consumer-app downloads or monthly users do not automatically create revenue. The March-quarter “All others” loss shows that distribution is currently expensive.

The accelerator picture constrains the speed and economics of this plan. Limited H200 access could improve performance and reduce software friction for some workloads. Supply conditions, licence terms and Chinese approval remain uncertain. Alibaba’s own T-Head programme reduces complete dependency on Nvidia, but the absence of public benchmarks prevents an assumption of equal total cost of ownership.

International commerce provides the cleanest positive earnings revision. Its quarterly adjusted EBITA loss narrowed to RMB 138 million from RMB 3.6 billion. AIDC is close enough to breakeven that modest logistics efficiency or advertising monetisation could make it a contributor. Its 9% FY2026 revenue growth is not fast enough to justify a premium standalone multiple, but the reduction in cash burn raises its SOTP value.

The investment portfolio remains an earnings-quality trap. March-quarter GAAP net income was RMB 23.5 billion, while non-GAAP net income was RMB 86 million. Mark-to-market gains in equity investments, and prior-year losses on the Sun Art and Intime disposals, created the difference. GAAP EPS was therefore a poor measure of current operating power.

Today’s market narrative prices cloud acceleration, potential H200 access, Qwen monetisation, domestic policy stability and a future end to the quick-commerce war. The operating fundamentals support the first two only partially and the last one not yet. The July rally reflects a higher probability placed on scarce Chinese compute and AI demand; it does not settle the return question.

The most important bull argument is that Alibaba is using a mature marketplace’s cash and balance sheet to defend consumer traffic while building China’s largest commercial AI platform. Commerce monetisation remains healthy beneath the subsidies, AIDC losses are disappearing, and cloud external revenue is accelerating. If quick-commerce costs peak while cloud growth remains above 30%, group earnings can recover sharply because the current comparison base is depressed.

The strongest bear argument is that Alibaba has entered two businesses where scale does not ensure high returns. Delivery competition can permanently raise consumer expectations and merchant costs. AI infrastructure can depreciate faster than it monetises, particularly under semiconductor constraints. The marketplace may continue funding these projects without returning to its former margin.

Section 1260H adds a different disagreement. Bulls see a designation with no immediate ownership ban, a credible legal challenge and limited direct U.S. defense revenue. Bears see a potential first step toward wider procurement restrictions, counterparty caution and institutional exclusions. Both views can be true at different dates because the consequence depends on litigation and implementation.

The DOJ settlement is more concrete but less financially material. The payment is affordable. The bears’ point concerns the duration of prohibited transactions and internal control. The bulls’ point is that a non-prosecution framework and defined payment close a known historical matter rather than create an open-ended liability.

Market calendars widely schedule the next report for August 28, 2026, covering the June 2026 quarter, Q1 FY2027. Alibaba’s investor-relations calendar had not formally confirmed that date as of the research base date, so it should be treated as a market-calendar estimate rather than an issuer-confirmed appointment. The release will test whether cloud external growth remains above 30%, whether China e-commerce EBITA has stopped deteriorating, whether “All others” losses have peaked, and whether operating cash flow covers a larger share of capex.

Valuation, risks and tracking framework

Headline P/E is not a sufficient valuation method. At USD 122.25, Reuters’ trailing P/E is 17.82. FY2026 GAAP diluted EPS per ADS was RMB 44.00, or approximately USD 6.52 at the July 31 exchange rate, implying about 18.8 times FY2026 GAAP earnings. FY2026 non-GAAP diluted EPS was RMB 26.80, approximately USD 3.97, implying about 30.8 times non-GAAP earnings. The difference reflects investment gains and excluded costs.

Free cash flow is a harsher measure. Alibaba’s definition subtracts property and equipment and relevant intangible purchases from operating cash flow. FY2026 operating cash flow was RMB 76.2 billion and capex was RMB 126.1 billion, producing materially negative free cash flow after the company’s other defined adjustments. The current FCF yield is therefore negative.

Owner earnings require a maintenance-capex estimate. Alibaba does not disclose that split. The valuation below assumes RMB 35–45 billion of FY2026 maintenance capex, broadly anchored to FY2022–FY2024 capex before the AI build. The remaining roughly RMB 80–90 billion is treated as growth capex, mainly cloud and AI infrastructure. This is an analytical assumption, not company guidance.

On that basis, FY2026 owner earnings equal operating cash flow less maintenance capex, or approximately RMB 31–41 billion, USD 4.6–6.1 billion. The implied owner-earnings yield is roughly 1.6%–2.2%, corresponding to 46–61 times owner earnings. That is roughly 2.6 to 3.4 times the headline P/E, so the valuation scenarios rely on normalised owner earnings and SOTP rather than reported net income.

The estimate is conservative in one respect and generous in another. It includes expensed quick-commerce and Qwen investment in current operating cash flow, depressing owner earnings. It assumes much of the current capex is growth rather than maintenance, raising owner earnings. The valuation requires both assumptions to be made explicit.

A sum-of-the-parts approach fits Alibaba because commerce, cloud, international operations and financial investments have different economics. It is also easy to misuse. The following values apply discounts for execution, liquidity, tax, VIE structure, partnership control and geopolitics rather than adding quoted assets at face value.

The model uses approximately 2.30 billion ADS-equivalent shares. Values are rounded and should be read as ranges rather than point precision. The available-assets line includes only a portion of liquid resources and investments because Alibaba requires operating liquidity, some investments are illiquid, and upstreaming cash across subsidiaries and VIE-related entities is not frictionless. Inputs are anchored to the FY2026 annual report and July 31 market data.

The conservative China-commerce value assumes normalised EBITA remains well below FY2025 because instant retail permanently absorbs part of the marketplace margin. The base case assumes customer-management growth and delivery rationalisation restore China segment EBITA toward RMB 145–160 billion over several years. The optimistic case assumes quick commerce becomes a defensible, low-positive-margin extension of Taobao rather than a permanent subsidy pool.

Cloud is valued at roughly 2.8 times FY2026 revenue in the conservative case, 3.8 times in the base case and 5.1 times in the optimistic case. These are below premium global cloud multiples because Alibaba’s margin is lower, capital intensity is rising, export controls remain relevant and the asset is embedded in a Chinese VIE-linked holding company. The optimistic value requires sustained growth above 25%, improving utilisation and a path toward a mid-teens or higher cash operating margin.

International commerce’s value rises as losses disappear. “All others” receives no conservative operating value because its Qwen, logistics, retail and media assets collectively consumed RMB 35.7 billion of adjusted EBITA in FY2026. A higher value requires individual businesses to become cash generative or be sold.

This is valuation-scenario analysis within a research framework, not investment advice.

The base value is close to the prior report’s USD 120–145 range, but the internal composition has changed. Cloud deserves a higher standalone value after 38% quarterly growth. China commerce deserves a lower near-term earnings value after the subsidy shock. The 1260H designation requires a more explicit holdco discount. These effects broadly offset each other.

Historical valuation is less useful than usual because Alibaba’s earnings basis changed. The company once traded as a high-growth, asset-light marketplace, and later as a regulated holding company. Today’s trailing P/E of about 18 times appears inexpensive against global technology stocks, but the owner-earnings yield is below the ten-year Treasury yield and current free cash flow is negative. The discount is not entirely irrational.

PDD trades at a lower trailing P/E despite higher recent revenue growth, reflecting Temu risk, domestic competition and concerns about future investment. JD trades at a higher P/E while suffering delivery-related margin pressure. Alibaba’s multiple sits between a distressed commerce platform and a global AI infrastructure company. Peer comparison does not establish undervaluation because all Chinese commerce peers face policy, competition and disclosure differences.

The market’s implied base expectation appears to be that the current earnings collapse is temporary but not fully reversible. At USD 122.25, investors pay more than the conservative SOTP, less than the base central value and far less than the optimistic cloud-led case. The price assigns substantial value to cloud and cash while assuming that quick commerce destroys a meaningful portion of historical marketplace profit.

The most important expectation gaps are measurable. Cloud external growth below 25% would challenge the premium placed on AI. Cloud growth above 35% with stable or improving segment EBITA would strengthen the product-demand case. China e-commerce EBITA falling below RMB 20 billion in a coming quarter would suggest that losses have not peaked. Recovery above RMB 30 billion would support the investment-cycle thesis.

The most fragile base-case assumption is that quick-commerce losses decline without a major loss of market share. Reducing the expected commerce-margin recovery to 70% lowers base SOTP by roughly USD 35–45 billion before the holdco discount, or approximately USD 11–15 per ADS after discount. The resulting base value would fall from about USD 140 toward USD 125–129.

Current price is above the conservative value of about USD 110, so the discount to the conservative case is negative. The margin of safety against that scenario is zero.

If earnings and intrinsic value remain flat for three years, an investor receives roughly the 0.9% current dividend yield, before currency movements and taxes. That is far below the approximately 4.74% ten-year U.S. Treasury yield on July 31, 2026. There is no margin of safety at this buy price under a flat-earnings outcome.

Margin-of-safety verdict: none. Alibaba may produce attractive upside from successful investment, but that upside is payment for execution and geopolitical risk rather than a conservative valuation cushion.

The principal permanent-loss risks are specific.

A prolonged quick-commerce war is the highest-probability operating risk. Probability is medium to high and impact is high. The observable indicators are China e-commerce EBITA, sales-and-marketing expense, average order value, non-food order mix and management’s unit-economics language. The transmission path runs through persistent subsidies, lower marketplace cash flow, fewer buybacks and a reduced commerce multiple.

Cloud overbuild has medium probability and high impact. The warning signs are external growth below 20%–25%, declining adjusted EBITA margin, capex staying above RMB 100 billion after growth slows, and depreciation rising faster than cloud gross profit. The transmission path is lower owner earnings, asset impairments and a lower cloud sales multiple.

Semiconductor restriction has medium probability and high impact. Confirmed H200 deliveries, U.S. licensing conditions, Chinese import approval, T-Head adoption and cloud product pricing are the observable indicators. Restricted supply would raise unit costs, limit model training and weaken Alibaba Cloud’s service quality relative to global providers.

Section 1260H escalation has low-to-medium probability and high valuation impact. The indicators are the federal lawsuit, final judgment, Pentagon implementing rules, institutional ownership filings and any addition to OFAC or BIS lists. The immediate transmission path is counterparty caution and multiple compression; a later investment or export restriction would affect both ownership and cloud operations.

VIE or domestic policy intervention has low probability but very high impact. The indicator is a change in Chinese foreign-investment, data-security or platform rules that challenges contractual control or cash transfers. The structure has operated for decades, yet the low frequency does not eliminate the severity.

Governance and capital allocation carry medium probability and medium-to-high impact. The Alibaba Partnership’s board nomination rights, limited disclosure of initiative economics and the simultaneous funding of two large programmes reduce outside shareholders’ ability to constrain spending. The indicator is continued capex and marketing growth without published return metrics.

The confirmed DOJ settlement has low incremental financial impact and medium governance impact. A repeated failure involving prohibited goods, sanctions, data or healthcare compliance would change the assessment because it would indicate a control-system problem rather than a closed historical matter.

The maximum plausible three-year loss is approximately 45%–55%, which corresponds to an ADS price of about USD 55–67. That path requires commerce EBITA to remain near FY2026 levels, cloud growth to fall toward the high teens while depreciation rises, and the holdco discount to widen after adverse Section 1260H or semiconductor developments. The conservative SOTP could itself fall toward USD 70–80 per ADS, with the market trading below that reduced value while the stress persists.

This tracking dashboard should be read against fiscal-period labels.

The first four indicators distinguish investment from destruction. Order growth alone is insufficient. The cash-flow measure catches both subsidies and infrastructure. The legal indicator must be tracked separately from China’s VIE rules. The August 28 date is a third-party market estimate as of August 2, not a confirmed company date.

Positive catalysts include a June-quarter China-commerce EBITA rebound, cloud external growth above 35%, disclosure of positive quick-commerce unit economics, confirmed advanced-accelerator deliveries, a favourable 1260H court ruling, further disposal of low-return businesses and renewed repurchases at depressed prices.

Negative catalysts include cloud growth deceleration, another quarter of RMB 20 billion-plus “All others” losses, capex growth without cash-flow recovery, Meituan or JD restarting deeper subsidies, broader U.S. restrictions, deterioration in Chinese household demand and evidence that Qwen consumer acquisition is not producing cloud usage.

The largest research uncertainties are the absence of quick-commerce orders and contribution profit in absolute terms; the absence of cloud capex, utilisation and invested capital; the lack of an exact AI-revenue denominator and AI gross margin; unconfirmed H200 procurement; and the contingent legal consequences of Section 1260H. These are not minor modelling gaps. They are the variables that determine whether Alibaba’s current spending earns an acceptable return.

Cross-synthesis and final conclusion

Looking vertically, Alibaba has proven that it can create market infrastructure around fragmented supply and then monetise the resulting traffic. Its best businesses share that pattern. Taobao aggregated merchants; Alipay solved trust; Alimama monetised merchant competition. Cainiao coordinated logistics, and Cloud externalised computing capacity. The company’s weaker record comes from businesses that require owned inventory, physical-service execution or repeated consumer subsidies.

Its early success combined era tailwinds with genuine execution. China’s internet and consumption growth provided the field. Alibaba’s local product design, free Taobao strategy, payment architecture, merchant tooling and willingness to build infrastructure won the game. Luck was present, as in every platform outcome, but eBay and other capitalised competitors had access to the same broad opportunity and failed to localise as effectively.

Several historical advantages remain. Taobao and Tmall still hold enormous consumer and merchant liquidity. Customer-management revenue remains high margin. Alibaba still has more capital and a broader ecosystem than most challengers. Eddie Wu brings technical and monetisation experience. The company is capable of funding cloud and delivery internally.

Other advantages weakened. Consumers multi-home, and merchants buy traffic from multiple platforms. PDD changed price discovery; Meituan built a local-delivery network Alibaba did not match early enough; Douyin changed the path from content to purchase. Regulation removed the option of enforcing merchant exclusivity. Alibaba’s previous dominance cannot be restored merely by spending.

Horizontal comparison identifies Alibaba’s real advantage as optionality across commerce, cloud and logistics. PDD is more focused, JD has tighter operational control and Meituan deeper local density. Alibaba can direct one billion-plus users, merchant relationships and computing capacity toward a new product faster than a standalone challenger. The weakness is that it can also direct capital toward too many objectives without publishing enough unit economics.

The current investment cycle contains one defensive and one offensive programme. Quick commerce is largely defensive: it protects product-search traffic and prevents Meituan from expanding unchecked into retail. Cloud and AI are offensive: they aim to capture a growing enterprise and developer profit pool. Defensive spending should be judged by avoided erosion as well as direct profit. Offensive spending must eventually earn a return above the cost of capital.

The two programmes deserve different tolerance. A temporary delivery loss can be rational when it changes consumer habit and creates density. That tolerance should end when market structure becomes permanently promotional. High cloud capex can be rational when external demand and utilisation rise. The tolerance should end when depreciation and capex grow faster than external gross profit.

The March-quarter evidence indicates that Alibaba has not crossed either line conclusively. Quick-commerce volume and revenue increased rapidly, while management reported better unit economics. The China segment still lost RMB 15.7 billion of year-on-year EBITA. Cloud external revenue accelerated to 40%, but FY2026 segment depreciation-related expense rose to more than twice adjusted EBITA. Both programmes have demand evidence. Neither has sufficient return evidence.

The market may be underestimating the speed of a profit rebound if spending moderates. FY2026 created unusually easy comparisons. AIDC is close to breakeven. Customer-management revenue is growing faster on a like-for-like basis. Cloud added EBITA despite the build. A modest reduction in delivery and Qwen acquisition spending could restore tens of billions of RMB in annual profit.

The market may also be underestimating how much of the old margin has been structurally surrendered. Instant delivery adds labour and fulfilment costs that search advertising did not have. AI infrastructure has shorter asset lives than traditional marketplaces. Even a successful Alibaba may emerge with faster revenue growth, lower free-cash-flow conversion and lower return on invested capital than the company investors remember from the 2010s.

The 1260H designation deserves a valuation discount but not a sanctions valuation. No current rule automatically forces U.S. investors to sell BABA, freezes Alibaba’s assets or removes it from the NYSE. Procurement, lobbying and counterparty consequences already matter. Litigation and rule-making can widen or narrow them. Treating the designation as harmless ignores the escalation path; treating it as an existing investment ban misstates present law.

Alibaba’s VIE structure deserves a separate discount because an ADS holder’s claim differs from direct ownership of every Chinese operating licence. The risk has existed since the IPO and is partly reflected in the multiple. It is not cured by a favourable 1260H ruling. Nor would a 1260H escalation automatically invalidate the VIE contracts. Combining the two into a single “China discount” obscures their different probabilities and transmission mechanisms.

By its cash amount, the DOJ settlement is unlikely to alter valuation. The direct USD 325 million Alibaba payment represents less than one-half of one percent of cash and liquid investments. Its importance lies in compliance governance. A clean post-settlement record would allow the market to treat it as closed. A second major control failure would support a higher governance discount.

The next year depends on four metrics: China commerce EBITA, “All others” losses, cloud external growth and operating cash flow after capex. Legal developments can alter the multiple, but those four metrics determine whether enterprise value compounds.

The three-year question is whether quick commerce settles into rational oligopoly pricing and whether Alibaba Cloud converts its infrastructure into higher utilisation and margin. A successful outcome can restore owner earnings above RMB 110 billion and justify the base SOTP. A failed outcome leaves the company funding delivery and AI from balance-sheet resources while its mature marketplace grows slowly.

The five-year question is whether Alibaba becomes China’s integrated AI-commerce infrastructure or a conglomerate whose profitable marketplace subsidises lower-return adjacencies. The former deserves a higher multiple than today. The latter deserves a persistent holding-company discount even if revenue remains large.

A better investment setup would combine a lower price with evidence that the spending curve is turning. USD 80–88 per ADS would provide at least a 20% discount to the conservative value. At that price, an investor could tolerate slower cloud monetisation or a longer subsidy period. Alternatively, stronger evidence can compensate for a higher price: external cloud growth above 30%, quarterly China commerce EBITA above RMB 30 billion, declining “All others” losses and improving cash flow would justify paying within the hold range.

The present price does not pre-spend the full optimistic outcome. It does pre-spend part of the recovery. At USD 122.25, Alibaba is above conservative value, near the lower end of the acceptable-hold band and below base central value. Existing investors receive substantial upside if management succeeds. New investors receive no conservative margin of safety while they wait.

The prior report’s Watch rating was defensible at USD 133.26 before the full March-quarter deterioration and before Section 1260H. This re-research reaches a slightly different rating because the share price fell to USD 122.25, cloud evidence improved and the current quote now sits inside the fair hold zone. The improvement is valuation-driven rather than risk-driven. Fundamental uncertainty increased.

The independent conclusion is Hold: Alibaba’s core assets and cloud acceleration support present value, but negative free cash flow, undisclosed quick-commerce economics and geopolitical escalation prevent a purchase rating.

The twelve-month setup remains fragile. Earnings comparisons will look weak, depreciation will rise and legal headlines can move the multiple. The three-to-five-year setup is more attractive because Alibaba has the balance sheet, traffic and engineering capacity to turn both investment programmes into durable assets. That long-term case should be earned through disclosed cash returns rather than inferred from revenue growth.

Bull reasons:

• March-quarter cloud revenue grew 38% and external-customer revenue about 40%, showing that AI demand extends beyond internal Alibaba workloads.
• Like-for-like customer-management revenue grew approximately 8%, indicating that the core marketplace monetisation engine remains intact beneath reported segment pressure.
• AIDC’s quarterly adjusted EBITA loss narrowed from RMB 3.6 billion to RMB 0.1 billion, removing a former structural cash drain.
• RMB 520.8 billion of cash and liquid investments gives Alibaba enough capacity to fund the current cycle without near-term equity financing.
• The current price is about USD 18 below the base SOTP central value, leaving meaningful upside if commerce margins and cloud returns normalise.

Bear reasons:

• Consolidated adjusted EBITA fell 84% in the March 2026 quarter and FY2026 free cash flow turned materially negative.
• China e-commerce adjusted EBITA fell RMB 85.7 billion for FY2026 even though segment revenue grew, showing that current scale has negative incremental economics.
• Cloud depreciation-related expense reached RMB 28.9 billion in FY2026, more than twice segment adjusted EBITA, while segment capex and ROIC remain undisclosed.
• Alibaba does not disclose quick-commerce contribution profit, subsidy run-rate or absolute orders, preventing verification of management’s unit-economics claims.
• Section 1260H, VIE governance and semiconductor controls can widen the discount even if operating results improve.

Pre-mortem script one: During FY2027 and FY2028, Meituan protects local-services density and JD continues subsidising instant retail. Alibaba maintains coupons and merchant incentives to preserve order share. China e-commerce adjusted EBITA remains below RMB 100 billion annually instead of recovering toward RMB 150 billion. The market cuts the commerce value from roughly ten times normalised EBITA to seven times. Cloud remains valuable, but the loss of commerce cash flow reduces buybacks and pushes the ADS toward USD 70–80.

Pre-mortem script two: Alibaba completes much of its RMB 380 billion infrastructure programme before advanced-chip supply and external demand are secure. Cloud growth slows below 20% in FY2028, depreciation rises above RMB 40 billion and adjusted EBITA margin remains below 10%. A wider U.S. restriction follows the 1260H process, raising compliance costs and the holdco discount from 25% to 35%. Cloud’s sales multiple falls below three times while commerce remains subsidised. The combined effect can cut the ADS by approximately half.

【Company-profile scores】

• Fundamental quality: medium
• Growth: medium
• Moat: medium
• Financial soundness: strong
• Management credibility: medium
• Valuation attractiveness: medium
• Risk level: high
• Suitable investor type: long-term growth, value and event-driven investors able to tolerate China, VIE and execution risk

【Investment rating】

• Rating: Hold
• One-line thesis: Cloud growth and marketplace monetisation support value, but negative free cash flow and undisclosed delivery economics eliminate the margin of safety.
• Current-price classification: acceptable hold
• Whether to wait for a better price: yes
• Buy condition: price at or below USD 88, or a higher price accompanied by cloud external growth above 30%, quarterly China-commerce EBITA above RMB 30 billion and improving cash flow
• Opportunity cost of waiting: a favourable 1260H ruling or rapid subsidy reduction could re-rate the ADS before a lower entry appears
• Target holding horizon: 3–5 years
• Expected annualised return, conservative scenario: approximately -2% to 0%, including dividends
• Expected annualised return, base scenario: approximately 8%–11%
• Expected annualised return, optimistic scenario: approximately 17%–22%
• Max-loss risk: approximately 45%–55%, triggered by structural delivery losses, cloud overbuild and a wider geopolitical discount
• Reassessment trigger: cloud external growth below 20% for two consecutive quarters
• Reassessment trigger: China e-commerce adjusted EBITA below RMB 20 billion for two consecutive quarters
• Reassessment trigger: annual capex above RMB 130 billion while cloud growth falls below 25%
• Reassessment trigger: Section 1260H consequences expand into investor-ownership or broad export restrictions
• Reassessment trigger: cash and liquid investments fall below RMB 350 billion without a corresponding increase in owner earnings

【Ideal Buy Price】80–88 USD

The range is at least 20% below the conservative SOTP value of approximately USD 110 per ADS and compensates for uncertainty in quick-commerce losses, cloud ROIC, VIE governance and Section 1260H.

Acceptable hold price: 119–161 USD, corresponding to approximately ±15% around the USD 140 base value.

Clearly overvalued price: 209–226 USD, beginning about 10% above the USD 190–205 optimistic value range.

【Valuation Range】

• current: 122.25 (close as of 2026-07-31)
• bear (conservative · ideal buy zone): [80, 88]
• base (fair · acceptable hold zone): [119, 161]
• bull (optimistic · above the clearly-overvalued line): [209, 226]

Primary-source hierarchy used in this report comprises Alibaba’s FY2026 annual report and March-quarter earnings disclosure; SEC filings; the Department of Defense Section 1260H notice; the governing U.S. statute; Alibaba’s federal complaint; the Department of Justice settlement announcement; and official JD and PDD results. Reuters, AP, the Financial Times, The Wall Street Journal and market-data providers were used for price, market reaction, legal developments and reported accelerator access.

Other tickers mentioned

• JD.US — direct Chinese commerce, logistics and instant-retail competitor
• PDD.US — low-price domestic-commerce and Temu competitor
• 3690.HK — Meituan, Alibaba’s strongest local-delivery and instant-retail rival
• 0700.HK — Tencent, Chinese cloud and consumer-platform peer
• BIDU.US — Chinese AI-cloud peer and fellow Section 1260H designee
• AMZN.US — global commerce and cloud return-on-capital benchmark
• MSFT.US — enterprise-cloud and AI monetisation benchmark
• NVDA.US — critical accelerator supplier subject to U.S.–China export controls
• SE.US — Southeast Asian commerce competitor through Shopee
• CPNG.US — logistics-led commerce reference for fulfilment economics

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.