The AI Monetization Inflection Arrives as the Profit Troughs: Upgrading Alibaba to Outperform

Results vs. Consensus

On the headline profit metrics this was the ugliest quarter in years, and it barely mattered. Revenue landed roughly in line, cloud accelerated, China e-commerce beat, and every profit line cratered under the peak of the AI + Cloud and quick-commerce investment cycle. The difference from our September read is that management paired the trough with specific, checkable monetization targets, and the market chose to underwrite the payoff.

Year-over-Year Comparison (Q4 FY2026 vs. Q4 FY2025)

Full-Year FY2026 Context

Full-year revenue crossed approximately RMB1.02 trillion (up ~2.7% reported, roughly +11% like-for-like excluding disposals), the highest in Alibaba's history, but FY2026 will be remembered as the year the company chose to sacrifice near-term profit for AI leadership, culminating in a March-quarter operating loss. It is the pivot year: the fiscal year cloud revenue growth reaccelerated into the 30s-to-40 percent and the fiscal year free cash flow turned deeply negative.

Segment Performance

The reporting structure is unchanged from the prior two quarters: China E-commerce Group, AIDC, Cloud Intelligence Group, and All others. The story this quarter is the growing divergence between an accelerating, margin-improving cloud franchise and a deepening technology-investment loss in All others.

Cloud Intelligence Group — the inflection quarter

Cloud revenue grew 38% to RMB41.6B with external-customer revenue accelerating to +40% (from +29% two quarters ago). AI-related product revenue reached RMB9B, a RMB36B annualized run-rate and 30% of external cloud revenue, and posted a 11th consecutive quarter of triple-digit growth. Crucially, management disclosed model & application services (MaaS/Model Studio) ARR for the first time, above RMB8B today, and framed it as a distinct, higher-margin revenue engine on top of infrastructure. Segment adjusted EBITA margin held at 9.1%.

"The AI and cloud commercialization inflection point has arrived ... we expect that in about 1 year, AI-related product revenue will cross the 50% threshold, becoming the primary engine driving the Cloud business's revenue growth ... external revenue growth is expected to continue accelerating beyond its current 40% rate over the coming quarters." — Eddie Wu, CEO

Assessment: This is the disclosure that changes the story. In September the cloud narrative was "demand is strong, trust us"; now it is a run-rate, an ARR trajectory, a mix target and an acceleration commitment. A software-like MaaS layer growing on top of an accelerating IaaS business, with margins that management expects to expand over the next one to two quarters, is exactly the earnings engine the bull case needed. It is the single biggest reason for the upgrade.

China E-commerce Group — the accounting-noisy but healthy core

Segment revenue grew 6% to ~RMB122B. Reported CMR grew just 1%, but this was almost entirely an accounting artifact: a new merchant business-development program reclassified subsidies from sales & marketing into contra-revenue, and on a like-for-like basis CMR grew 8%, a meaningful reacceleration from the base-effect slowdown management had warned about. Segment adjusted EBITA fell 40% to ~RMB24B on quick-commerce and user-experience investment, but excluding quick commerce it would have been stable year-over-year.

"Excluding loss from our quick commerce business, our Alibaba China E-commerce Group EBITA would have been stable year-over-year and will fluctuate quarter-over-quarter due to significant investment in merchant retention and user experience." — Toby Xu, CFO

Assessment: The core marketplace remains the profit and cash engine and it is not deteriorating; the +8% like-for-like CMR is the number to watch, not the accounting-depressed +1%. The reclassification muddies optics but does not change economics, and the underlying take-rate story is intact.

Quick commerce — inflecting toward profitability

Quick-commerce revenue grew 57% to ~RMB20B, with order volume 2.7x the year-ago level (non-food orders 3x) and sequential unit-economics and AOV improvement. Management reaffirmed a clear target: UE turns positive by the end of FY2027.

"From April onwards, while maintaining order volume, we've continued to drive substantial improvement in UE ... So we are confident that UE will turn positive by the end of fiscal year '27." — Jiang Fan, CEO of Alibaba E-commerce Business Group

Assessment: The trajectory management promised in September is materializing: losses narrowing, scale holding, synergy into the marketplace (Freshippo, Tmall Supermarket) visible. The FY2027 UE-positive target is now a concrete, gradeable milestone rather than an aspiration, which de-risks the largest single drag on group profitability.

AIDC and All others — the two tails

AIDC revenue grew 6% and its EBITA loss narrowed to near breakeven on AliExpress Choice efficiency, extending last quarter's swing toward profit. All others, however, saw its EBITA loss balloon to RMB21.2B as foundation-model training and consumer Qwen-app user acquisition costs landed here, a sharp step-up from the RMB3.4B loss two quarters ago.

Assessment: AIDC is a quiet positive that supports the "we can turn losses when we choose" credibility. All others is where the bill for the AI ambition is being paid, and it is large and rising; investors need to treat this segment's loss as the true cost line of the model/consumer-AI bet and watch whether it plateaus.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: The most confident and, importantly, the most specific of the four calls we have followed. Management treated the profit trough as a settled decision and spent its energy quantifying the AI payoff (run-rates, ARR targets, margin trajectory, a capex-vs-opex framework) and a two-year path to cash-flow recovery. Where prior calls leaned on demand anecdotes, this one leaned on numbers, and the tone shifted from defending the spend to selling the return.

1. "The AI commercialization inflection point has arrived"

The organizing message of the call was that Alibaba has moved past the pure-investment phase into monetization at scale, evidenced by the RMB36B AI run-rate, the 40% external cloud acceleration, and the MaaS ARR disclosure.

"Alibaba's AI has moved beyond the initial investment phase and progressed commercialization at scale ... AI + Cloud revenue growth is accelerating with improving margins." — Eddie Wu, CEO

Assessment: The claim is now backed by disclosures a skeptic can track quarter to quarter. If cloud external growth accelerates above 40% and MaaS ARR hits the RMB10B/RMB30B waypoints, the inflection is real; if not, the thesis breaks in public. That gradeability is a feature, and it is what justifies moving off the fence.

2. MaaS/Model Studio: a new high-margin ARR engine

Management put a first number on model-and-application services ARR (above RMB8B, growing more than 10x since last November) and guided it above RMB10B next quarter and RMB30B by year-end, emphasizing its structurally higher margin than infrastructure.

"We expect model and application services annualized recurring revenue, ARR, inclusive of the Model Studio platform to surpass RMB 10 billion in the June quarter and RMB 30 billion by year-end. The high margin profile of this revenue stream is becoming increasingly apparent." — Eddie Wu, CEO

Assessment: This is the highest-value disclosure of the quarter. A rapidly compounding, high-margin software layer reframes Alibaba Cloud from a low-margin IaaS utility to a two-layer business where the faster-growing layer is the more profitable one. It is the mechanism by which cloud can eventually carry group margins, not just group growth.

3. Cloud gross margin is set to expand

Management argued cloud gross margin should rise over the next one to two quarters and materially over two to three years, driven by the MaaS mix shift, reasoning-efficiency gains that lift tokens per server, over-100% server cost inflation that confers pricing power, and rising T-Head chip deployment.

"Overall in the next 2 to 3 years, we can expect to see a significantly higher gross margin for Alibaba Cloud, and we can expect to start to see that in the next 1 to 2 quarters." — Alibaba management

Assessment: A concrete, near-term, falsifiable claim. If cloud margin starts climbing next quarter, it validates the entire monetization thesis; management has effectively staked its credibility on a visible metric, which is exactly what an analyst wants after quarters of "trust us."

4. T-Head: proprietary silicon as a moat and a margin lever

Management highlighted that T-Head's proprietary GPU chips are in scaled production with over 60% of that compute already serving external customers, making Alibaba the only China cloud able to deliver self-developed chips at scale, an autonomy advantage amid a global compute shortage and a future gross-margin tailwind.

Assessment: Strategically significant and underappreciated. In-house silicon addresses the single biggest risk to the China-AI-cloud thesis, chip supply under export controls, while structurally improving unit economics. It is optionality with a clear mechanism, though the current deployment ratio is still low and the ramp will take time.

5. Capex: overshooting RMB380B, with an OpEx release valve

Asked to size the capex required, management framed a roughly 10x increase in data-center infrastructure versus 2022 by the early 2030s, said the RMB380B three-year plan will likely be exceeded, but noted some compute is being acquired via OpEx and that T-Head lets Alibaba sell or co-build capacity, providing flexibility.

"It's likely, given that kind of investment, that we will overshoot the original CapEx figure that we had stated of RMB 380 billion. But at the same time, we can acquire some compute through OpEx ... there are different ways that we can get to where we need to get." — Alibaba management

Assessment: The capex is going higher, as flagged in September, but the OpEx/T-Head framing is a more sophisticated capital-allocation answer than the market feared. It signals Alibaba will scale to demand without being purely capex-bound, which softens the free-cash-flow bear case at the margin.

6. Free cash flow: a two-year recovery path, spelled out

The CFO directly addressed the negative FCF, attributing it to AI investment and laying out why it recovers: stable Taobao/Tmall cash generation, quick-commerce losses narrowing substantially, AIDC turning profitable, and rising cloud cash flow, all against a fortress balance sheet and capital-markets access.

"Looking ahead over the next 2 years in terms of quick commerce, the losses will narrow very substantially. At the same time, AIDC will develop from making a loss to being profitable. So we see these developments over the next 2 years as being highly positive for our net cash flow." — Toby Xu, CFO

Assessment: The missing piece from September. Management moved from "the balance sheet lets us spend" to "here is the bridge back to positive cash flow." It is a two-year story and unproven, but it converts an open-ended cash-burn fear into a trackable trajectory.

7. Quick commerce: scale held, economics improving

Order volume ran 2.7x year-ago with non-food at 3x, and management reiterated UE turns positive by end of FY2027 while continuing to generate synergy into conventional e-commerce (Freshippo, Tmall Supermarket acceleration, higher CMR).

Assessment: Delivering on the September plan. The combination of held scale, improving UE and demonstrated marketplace synergy is the evidence we said we needed before upgrading; it is now on the table.

8. The CMR reclassification

Management flagged that a new merchant subsidy program shifted spend from sales & marketing to contra-revenue, depressing reported CMR to +1% versus +8% like-for-like.

"Such subsidies previously recorded as sales and marketing expenses are now recorded as a contra revenue item to CMR. Accordingly, CMR grew 1% year-over-year ... on a like-for-like basis, CMR would have grown 8% year-over-year." — Toby Xu, CFO

Assessment: A presentation change, not an economic one, but it will confuse screens and headlines for a couple of quarters. The +8% like-for-like is the real monetization signal and it is a healthy reacceleration.

9. Consumer AI and the 2B/2C balance

Qwen app fully integrated Taobao/Tmall commerce on May 7 and is embedded across Taobao, Alipay, Amap and Fliggy. Management acknowledged 2B monetizes faster today and gets priority compute, while defending continued 2C investment as a one-to-two-year payoff.

Assessment: A disciplined answer: prioritize the paying enterprise workloads now, keep the consumer option alive for later. It reassures that resources will follow ROI, though the Qwen-app spend is a real, large loss line (in All others) that must eventually justify itself.

10. Cloud margin philosophy: "margin is still secondary"

On near-term cloud margins management was explicit that share and token-consumption growth come first, with margin secondary, while still arguing structural margin expansion is coming from mix and cost dynamics.

Assessment: The right priority for a land-grab phase in a scarce-compute market, and consistent with the leadership position (35.8% China AI cloud share). The tension is that "margin secondary" plus "capex overshoot" keeps near-term profitability low; the payoff is share that compounds.

11. Capital returns held through the trough

The board approved a US$1.05/ADS FY2026 dividend (~US$2.5B), maintained through the operating-loss quarter, alongside continued buyback capacity.

Assessment: A confidence signal. Maintaining the dividend during the peak-investment, negative-FCF quarter tells shareholders management views the cash drain as temporary and self-funded by the balance sheet, consistent with the two-year recovery framing.

Guidance & Outlook

Alibaba issues no formal numeric guidance, but this call carried more specific forward targets than any in recent memory, which is itself a tell about management's confidence.

Implied trajectory: Take the targets at face value and FY2027 should show cloud growth accelerating with margins expanding, MaaS ARR tripling, quick-commerce losses narrowing toward breakeven, and group FCF beginning to recover, even as absolute capex rises. The guidance style has shifted from conservative-and-vague to specific-and-confident, which raises both the reward (if hit) and the accountability (if missed).

Street positioning: Consensus entering the print was cautious after the December-quarter miss, with the stock de-rated and expectations low. The specific ARR and cloud-acceleration targets reset the debate onto Alibaba's preferred terrain, cloud/AI, and away from near-term EPS.

Analyst Q&A Highlights

Composition of MaaS ARR and the impact of token price hikes on cloud margin

The opening question probed how much of the newly disclosed ARR comes from in-house versus third-party models, and what recent token price increases mean for MaaS and cloud margins.

Q: "How much of that ARR is driven by our in-house models like Qwen versus third-party models? And given the recent token price hikes, what would be the implications to MaaS and also our Cloud margins as a result?"
— Ronald Keung, Goldman Sachs

A: "Most of that revenue is coming from our own proprietary models ... customers' acceptance for higher prices, and we have increased per token prices, is good and the demand continues to be high and growing ... Inherently, MaaS will have higher gross margin than IaaS ... the rapid growth in this business over the next few quarters will result in a very positive impact on our overall gross profit margin."
— Alibaba management

Assessment: The answer confirmed the two things bulls needed: pricing power (customers accepting token price hikes for agentic workloads) and margin accretion (MaaS mix lifting cloud gross margin). It directly links the AI acceleration to future profitability rather than just future revenue.

ROI on AI investment and the framework for balancing spend against earnings

A central line of questioning pressed on how investors should assess return on the AI spend given the free-cash-flow and EBITA drag, and how management balances aggressive investment against earnings stability.

Q: "While our AI investments have driven impressive 40% cloud growth, they have also created significant drag on the group free cash flow as well as our EBITA. So how do investors assess the return on this investment? And what is the management's framework for balancing the aggressive AI spending versus earnings stability?"
— Kenneth Fong, UBS

A: "The negative free cash flow is primarily due to the very significant investments we've been making in AI ... looking forward to the next 2 years ... the losses [in quick commerce] will narrow very substantially ... AIDC will develop from making a loss to being profitable ... there isn't a single card on our service that is idle. So we see the ROI on this investment in the next 3- to 5-year period as being extremely clear."
— Toby Xu, CFO, and Eddie Wu, CEO

Assessment: The most important exchange of the call. Management answered the cash-flow bear case head-on with a two-year recovery bridge and the "two factories" ROI framing. It did not quantify a precise return multiple, but "no idle cards" plus a segment-by-segment path to cash recovery is a materially fuller answer than September's balance-sheet appeal.

Quick-commerce unit-economics drivers and the multi-year outlook

A question sought detail on the UE improvement (AOV, subsidy ratio, fulfillment) and whether the three-year quick-commerce view had changed.

Q: "I just wanted to get some more color about the drivers behind [UE] in terms of AOV, subsidies ratio, fulfillment ratio ... Is there any update or changes in terms of how we think about the landscape or the UE in the next 3 years?"
— Thomas Chong, Jefferies

A: "Overall order volume was 2.7x that of the same quarter last year with non-food orders at 3x ... we've continued to drive substantial improvement in UE through enhanced fulfillment logistics efficiency as well as order mix optimization. So we are confident that UE will turn positive by the end of fiscal year '27."
— Alibaba management

Assessment: The reaffirmed FY2027 UE-positive target, backed by 2.7x order growth and specific mix drivers, is the operational proof the September thesis required. It converts the largest profit drag into a milestone with a date.

MaaS competitive advantage, AI coding growth in China, and SaaS willingness-to-pay

A multi-part question asked how Alibaba's MaaS platform is differentiated from Chinese AI startups, when China's AI-coding growth catches up to the US, and whether low China SaaS willingness-to-pay caps monetization.

Q: "What are the major advantages that Alibaba has when compared to the other major AI platforms in China as well as Chinese AI start-ups? ... when do you think we'll see that kind of growth in China in terms of AI coding? ... Chinese customers are less willing than U.S. customers to pay for SaaS ... might [that mean] less potential than we see with the U.S. counterparts?"
— Jialong Shi, Nomura

A: "We are investing at a much higher scale and across a much broader range of different model types ... those AI start-ups really are partners rather than competitors ... China is already there [on coding] ... growth that we're seeing on our own Bailian platform from November ... through to May ... is higher than a 10x growth. In terms of our ARR, it's already over RMB 8 billion."
— Alibaba management

Assessment: Management positioned startups as MaaS customers rather than rivals and argued China's coding-agent demand is already inflecting, citing 10x Bailian growth. The willingness-to-pay concern was acknowledged but framed as a temporary gap that closes as models deliver real value, a reasonable but not-yet-proven claim.

2B versus 2C resource allocation

A line of questioning asked how Alibaba prioritizes compute across enterprise (2B) and consumer (2C) initiatives, and whether it would shift resources away from the Qwen app toward cloud and MaaS if enterprise keeps gaining traction.

Q: "How do we evaluate the strategic priority and resources allocation between 2B and 2C initiatives? If going forward enterprise side continues to gain more traction, will we consider gradually shifting more resources away from Qwen app to Cloud and MaaS?"
— Ellie Jiang, Macquarie

A: "At present, we see higher willingness to pay on the 2B side ... most of our infrastructure resources are therefore channeled to the 2B side. But ... AI ultimately is an invention that's there to be a helper ... I'm confident that [consumer paying] will come to be the case in China ... probably ... in the next, say, 1 to 2 years."
— Alibaba management

Assessment: Reassuring capital discipline: compute follows the ROI, which is enterprise today. The consumer bet is being kept alive as optionality on a one-to-two-year horizon rather than force-funded, which limits the risk that Qwen-app spending balloons without return.

Cloud EBITA margin outlook versus international peers

A question asked whether accelerating cloud can eventually reach the margins of international hyperscaler peers.

Q: "I'd just like to understand what your view is on EBITA margins in the cloud business over the next few quarters. Do you think as this business accelerates, we can expect to see similar margins as we see in your international peers?"
— Joyce Ju, Bank of America

A: "Our objective is to ... drive growth in token consumption and to acquire larger market share ... margin is still secondary ... the cost for us to deploy one new server this year is double what that same server would have cost a year ago ... we have a certain pricing power ... overall in the next 2 to 3 years, we can expect to see a significantly higher gross margin for Alibaba Cloud, and we can expect to start to see that in the next 1 to 2 quarters."
— Alibaba management

Assessment: A candid "share first, margin second" answer, but paired with a near-term margin-expansion commitment. The over-100% server cost inflation point is a clever pricing-power argument. The honest read: margins improve but a hyperscaler-like margin is a multi-year aspiration, not a next-year event.

Capex required and T-Head penetration/margin uplift

The closing question asked what capex level is needed to serve MaaS and long-term cloud demand, and how much margin uplift to expect as T-Head chip penetration rises.

Q: "What kind of level of CapEx investment is required in order to satisfy the demand from both MaaS and also the long-term cloud revenue? ... What is the current penetration of T-Head being deployed on AliCloud? And as this penetration increase[s], margin uplift we should expect from our in-house chip?"
— Gary Yu, Morgan Stanley

A: "We're talking about a 10x increase [in data-center infrastructure vs 2022] ... it's likely ... that we will overshoot the original CapEx figure ... of RMB 380 billion ... [T-Head] penetration ... is still relatively low ... as we deploy more and more of our own proprietary T-Head chips, the new chips will certainly contribute very, very significantly to gross margin expansion."
— Alibaba management

Assessment: Confirms capex goes higher (the FCF cost) but also that T-Head margin uplift is still almost entirely ahead of us (the future benefit). The low current T-Head penetration is both a risk (execution/ramp) and a large embedded option on cloud margins.

What They're NOT Saying

1. No absolute capex number, still. "Overshoot RMB380B" plus an OpEx release valve is more nuanced than before, but management again declined to re-anchor the capex figure, so the FCF trough depth remains hard to size precisely.
2. No group-level margin or FCF-inflection date. The two-year recovery is directional; there is no committed quarter for when group EBITA or free cash flow turns positive. The recovery is a bridge, not a timetable.
3. All others loss trajectory left open. The RMB21.2B segment loss (foundation models + Qwen app) was explained but not bounded; management did not say when or at what level this technology-investment loss plateaus.
4. MaaS margin is asserted, not quantified. "Higher than IaaS" and "expanding" are qualitative; no MaaS gross-margin figure was given, so the profit contribution of the RMB30B ARR target is unmodelable in dollars.
5. Competitive response underplayed. With quick-commerce losses narrowing, management said little about how JD.com and Meituan might react to Alibaba pulling back subsidies, the variable that most affects whether the FY2027 UE-positive target holds.
6. Little on the geopolitical overhang. The US-China regulatory backdrop (including Alibaba's presence on the US Department of Defense's list of Chinese military-linked companies, which the company is challenging) went largely unaddressed on the call despite being a live tail risk for the ADR.

Market Reaction

• Pre-print setup: BABA closed at $134.78 on May 12, down 8.1% year-to-date (from $146.58 at year-end) and up just 2.4% over the trailing twelve months, though up 5.3% over the trailing 30 days. The 52-week closing range was $103.83-$189.34, leaving the stock roughly 29% below its high and materially de-rated from the ~$160 level and ~90% YTD gain it carried into the September print. The S&P 500 was +8.1% YTD.
• Reaction (May 13 session, before-open print): Shares gapped down 2.0% to open at $132.06, traded as low as $130.33, then reversed sharply to close at $145.81, up 8.2% ($134.78 to $145.81), near the intraday high of $146.87. Volume was 40.2M shares, 4.2x the 30-day average, the heaviest of any of the four prints we have tracked.

This is the mirror image of the September reaction, and the contrast is the whole story. In September a revenue/cloud beat faded to a 2.3% loss because spending scared a crowded, richly-valued stock. Here a worse profit print (operating loss, non-GAAP EPS down 95%) produced an 8.2% rally on record volume, because the stock was cheap, positioning was washed out after the December-quarter miss, and management finally handed the market the quantified AI monetization inflection it had been waiting for. The move was a re-rating of the narrative, not a response to the reported earnings, which is precisely the setup in which we want to be long.

Street Perspective

Debate: Is the AI monetization inflection real or a well-timed narrative?

Bull view: The disclosures are specific and gradeable, RMB36B AI run-rate, cloud external +40% accelerating, MaaS ARR targeted at RMB10B then RMB30B, so this is proof, not spin; a high-margin software layer is now compounding on top of accelerating infrastructure.

Bear view: ARR targets are management projections, the profit is still negative, and China internet has a history of promising inflections that arrive slower and smaller than guided; the +8% stock move front-ran results that have not yet appeared in earnings.

Our take: The bulls have the better of it because the targets are near-term and checkable, cloud margin should visibly expand within one to two quarters and MaaS ARR should cross RMB10B next quarter. That makes this a falsifiable thesis with catalysts, which is why we are willing to underwrite it now rather than wait.

Debate: Does the cash burn undermine the story?

Bull view: The burn is investment, not distress; Taobao/Tmall throws off stable cash, quick-commerce losses are narrowing, AIDC is turning profitable, and ~US$38B net cash plus capital-markets access funds the build with room to spare.

Bear view: FCF was still a RMB17.3B outflow, capex is going higher, the All others loss ballooned to RMB21.2B, and "two years to recovery" is a long time to ask shareholders to wait with no committed inflection date.

Our take: The burn is real and the timeline is soft, but the CFO's segment-by-segment recovery bridge plus a maintained dividend through the trough is credible enough to own. We would grow concerned if the All others loss keeps expanding without a plateau, and we will track it as the key cash-flow risk.

Debate: Is the stock cheap enough to look through the profit trough?

Bull view: After a ~29% drawdown from the high and roughly flat over twelve months, the stock embeds little of the cloud/AI optionality; on a sum-of-the-parts basis (cloud + marketplace + ~US$38B net cash + stakes) the downside is well-supported and the cloud re-rating is largely unpaid-for.

Bear view: Near-term earnings are negative, so valuation rests on out-year estimates and a geopolitical-discount that may be permanent for a China ADR; "cheap" China tech can stay cheap.

Our take: The de-rating is exactly what was missing in September. Buying an inflecting cloud franchise after a drawdown, rather than after a double, is the right side of the risk/reward, and the SOTP floor limits the downside while the AI catalysts provide the upside.

Model Update Needed

Valuation impact: The upgrades to cloud growth, the new high-margin MaaS layer and a credible FCF-recovery path all push intrinsic value up, while the de-rated share price lowers the entry multiple. We continue to frame fair value on a sum-of-the-parts basis, cloud (now with a software layer deserving a higher multiple) plus the marketplace cash engine plus ~US$38B net cash plus equity stakes, and note the current price captures little of the cloud re-rating. The net effect moves our stance from balanced to favorable.

Thesis Scorecard Post-Earnings

Scoring the same standing thesis established at initiation, updated for the March quarter and the interim December-quarter print.

Overall: Thesis strengthened. The central pillar (cloud/AI) inflected with hard numbers, quick commerce delivered on its September commitments, and the two bear points that kept us at Hold, an unproven return on spend and a full valuation, both eased (return now quantified; valuation de-rated). The near-term profit and FCF remain poor, but the print reads as the trough of a deliberate cycle rather than deterioration.

Action: Upgrade to Outperform from Hold. We flagged at initiation that we would move up on two consecutive quarters of quick-commerce UE improvement with cloud holding above 30%; both conditions are now met (UE improving through the December and March quarters; cloud external at +40%), and a de-rated entry price plus concrete near-term catalysts (cloud margin expansion within one to two quarters; MaaS ARR crossing RMB10B next quarter) complete the case. Key things to grade next quarter: cloud gross-margin expansion actually appearing, MaaS ARR hitting >RMB10B, quick-commerce loss continuing to narrow, and the All others loss beginning to plateau. We would revert to Hold if cloud margin fails to inflect or the All others loss keeps expanding without a ceiling.