Full Report

Franchise and Trough

Meituan runs China's dominant local-services platform — food delivery, in-store dining, hotels and travel, and instant retail — with FY2025 revenue of ¥364.9 billion [1]. After reaching a record ¥35.8 billion profit in 2024, it swung to a ¥23.4 billion loss in 2025 as a subsidy war in food delivery erupted [2]. The franchise still grew; the profit did not survive the fight. Financials are reported in renminbi (¥); the shares trade in Hong Kong dollars (HK$).

What Meituan is

Meituan operates the app most Chinese consumers open to order a meal, book a restaurant table, buy a hotel night, or have groceries and medicine delivered in under an hour. It reports in two segments: Core Local Commerce — food delivery, in-store dining, hotel and travel, and Meituan Instashopping (instant retail) — and New Initiatives, which houses grocery-retail formats, the Meituan Select community-group-buy business, and overseas delivery (Keeta). In 2025 the platform's gross transaction value and transaction volume both grew double digits, and annual transacting users, purchase frequency and ARPU reached record highs [3].

The company makes money four ways, and the mix matters. In Core Local Commerce, FY2025 revenue of ¥260.8 billion came from delivery fees (¥96.1 billion), merchant commissions (¥99.2 billion), high-margin online marketing — advertising sold to merchants (¥51.5 billion) — and other services (¥14.1 billion) [4]. Commissions and advertising are the profit engine; delivery fees largely pass through to couriers. New Initiatives added ¥104.0 billion, almost all of it lower-margin retail sales [5].

Source: FY2025 Annual Report, MD&A — Revenues by segment [6].

Size and the five-year arc

Meituan is a large-cap platform, not a startup: revenue has doubled in four years, from ¥179.1 billion in 2021 to ¥364.9 billion in 2025 [7]. The profit line tells the more important story. The company lost money in 2021 and 2022 during an earlier phase of regulatory pressure and heavy investment, turned profitable in 2023 (¥13.9 billion), peaked in 2024 (¥35.8 billion), then fell back into a ¥23.4 billion loss in 2025 [8].

Source: FY2025 Annual Report, Financial Summary [9].

The gross margin traces the same round trip: it climbed from 23.7% in 2021 to 38.4% in 2024, then gave back ten points to 30.4% in 2025 as incentives and courier costs rose [10].

The 2025 break

The loss was not a demand problem. Revenue rose 8.1%, and Core Local Commerce revenue grew 4.2% to ¥260.8 billion [11]. It was a spending problem, driven by competition. In early 2025, JD.com entered food delivery and Alibaba escalated instant-retail subsidies, and Meituan answered in kind to protect share. Selling and marketing expense rose 60.9% to ¥102.9 billion — from 19.0% of revenue to 28.2% [12]. That ¥38.9 billion increase in marketing alone is larger than the year's total loss.

The effect on segment profitability was stark. Total segment operating profit fell from ¥45.1 billion in 2024 to a ¥17.0 billion loss in 2025 [13]. Core Local Commerce — the mature, cash-generative engine — went from a ¥52.4 billion operating profit (a 20.9% margin) to a ¥6.9 billion operating loss [14]. A business that minted a fifth of its revenue as operating profit in 2024 was, one year later, subsidizing transactions.

Sources: Chairman’s Statement [15]; MD&A — Core Local Commerce segment results [16].

Whether this is a temporary defense of a durable franchise or the start of a permanent margin reset is the central question of the investment case, and later chapters test it directly. The evidence for "temporary" is that revenue and transactions kept growing and the profit erosion sits almost entirely in incentive spend management chose to deploy [17]. The counter-fact is that two of China's best-capitalised companies have now decided local delivery is worth fighting for, and a subsidy war ends only when someone stops paying.

The balance sheet that buys time

For an investor whose first concern is survival, the balance sheet is the reassuring part. At the end of 2025 Meituan held ¥106.8 billion of cash and equivalents plus ¥60.1 billion of short-term treasury investments — ¥166.9 billion of liquid resources [18]. Against that sit reported borrowings of ¥18.8 billion (non-current) and ¥3.5 billion (current), roughly ¥22 billion in total [19]. The company is deeply net-cash even before counting longer-dated deposits and investments.

Sources: MD&A — Liquidity [20]; Consolidated Statement of Financial Position [21].

The war did cost cash. Operating cash flow turned negative — a ¥13.8 billion outflow in 2025 against a ¥57.1 billion inflow in 2024 [22]. At that rate the net-cash position funds years of losses, not quarters, which is the difference between a distressed balance sheet and a defended one. The bankruptcy risk a cautious investor screens for looks remote here; the risk that matters is how long, and how expensively, the fight runs.

Founder control and skin in the game

Meituan is founder-run, with control locked in by a weighted-voting-rights structure. Each Class A share carries 10 votes to a Class B share's one. Founder, chairman and CEO Wang Xing beneficially owns 515.9 million Class A shares — about 45.30% of the votes — and co-founder Mu Rongjun a further 5.56%, so the two founders command a majority of the vote while holding a minority of the economics [23]. Meituan has roughly 6.11 billion shares outstanding in total [24].

Their pay is negligible next to that stake. In FY2025 Wang Xing's total emoluments were ¥5.26 million — a ¥5.04 million salary and pension costs, with no bonus and no share awards — and Mu Rongjun's were ¥4.32 million [25]. The founders are compensated as owners, through the equity, not through the payroll. For an investor who prizes aligned, founder-led management, the alignment is real; the flip side is that minority holders cannot force a change of course, and the decision to wage the 2025 subsidy war was management's to make.

The stock: a fallen star

Meituan is the kind of name this market once loved and now distrusts. The shares peaked near HK$460 in February 2021; they trade around HK$80.90 today, roughly 82% below that high, with a market capitalisation near HK$500 billion and a 52-week range of HK$63.65 to HK$136.10. How completely the mood has reversed is visible inside the filing itself: the convertible bonds Meituan issued in 2021 carry a conversion price of HK$431.24 — a level the stock last saw at the peak, and the reason management has been redeeming rather than converting them [26].

Consensus expects the trough to pass. Analysts model a near-breakeven 2026 followed by a return to profit in 2027, and the mean price target sits near HK$107 against the ~HK$81 quote — a stock priced for a recovery that has not yet shown up in reported earnings.

Source: share-price levels, market data and consensus estimates as reported; convertible-bond terms per FY2025 Annual Report [27].

The question this report answers

Meituan is a dominant, still-growing local-commerce franchise, founder-controlled and carrying enough net cash that survival is not the issue — thrown, in 2025, from record profit into heavy losses by a subsidy war it chose to fight. The question this report exists to answer: is that swing a durable impairment of the franchise's economics, or a defensible trough that the balance sheet can outlast — and how much of the eventual recovery the stock already excludes? Everything that follows connects to that.

---

The Delivery War

In 2025 two of China's best-funded companies attacked Meituan's core business at once. Judged on the metrics that compound — order and GTV leadership, the profitable order mix, rider and merchant density — the franchise held. But it held only by matching the subsidies, and that is what turned a ¥52.4bn segment profit into a ¥6.9bn loss. The war revealed a real but contestable moat: a cost-and-density advantage rivals could dent with cash, not erase.

Two well-funded entrants, one incumbent

For most of the last decade Meituan's food-delivery dominance was rarely tested by an equal. That changed in early 2025. JD.com entered on-demand food delivery in February, waiving merchant commissions and pledging around ¥10bn of consumer subsidies; Alibaba escalated in response, folding Ele.me into a relaunched "Taobao Instant Commerce" and committing a subsidy program the size of ¥50bn over the following year. Meituan answered in kind rather than cede ground. Press and analyst accounts of the episode — the filings describe it only as "intensified competition" — put industry-wide daily orders at roughly 100 million entering 2025 and above 250 million at the war's peak, with promotional coffees selling for as little as ¥2 and a single Saturday drawing a reported 120 million orders across platforms.

Meituan's own report frames the year with unusual candour for a Chinese issuer: 2025 was "a year of both significant opportunities and formidable challenges," met "in the face of the intense competition" [1]. Its support-fund programs alone "assisted over 500,000 merchants in alleviating operational pressures from irrational industry competition" [2]. The word management keeps returning to is irrational — the tell of a fight it did not start and does not expect to last.

What defending the line cost

The bill lands almost entirely in one place. Core Local Commerce — food delivery, Instashopping, in-store, hotel and travel — had earned a rising operating profit for years: ¥38.7bn in 2023, ¥52.4bn in 2024, a 20.9% margin [3]. In 2025 it recorded a ¥6.9bn operating loss, a negative 2.6% margin — a ¥59bn swing in a single year [4].

Source: FY2024 Annual Report (2023–2024) [5]; FY2025 Annual Report (2025) [6].

The mechanism was spending, not demand. Core Local Commerce revenue still grew 4.2% to ¥260.8bn [7]; the damage came through incentives netted against delivery revenue and, above the segment line, a 60.9% jump in selling and marketing expense to ¥102.9bn [8]. The pain deepened through the year: the fourth quarter alone carried a ¥10.0bn Core Local Commerce operating loss, a negative 15.5% margin against a positive 19.7% a year earlier, as quarterly selling and marketing expense rose 83.4% [9] [10]. Total segment operating profit for the group fell from ¥45.1bn to a ¥17.0bn loss [11]. This is the cost side of the Franchise and Trough question, itemised: the trough was bought, quarter by quarter, to hold a position.

The rivals paid too. Press and analyst estimates put the three platforms' combined subsidy-and-marketing cost above ¥100bn across the second and third quarters, with Alibaba's instant-retail losses estimated near ¥87bn for the campaign and JD's new-business unit losing on the order of ¥47bn. No participant made money; the question the war actually settled is who came out of it with the franchise intact.

Whether the moat held

On the numbers that matter, Meituan kept the lead. Through the worst of the war it reported "record DAU and MTU in food delivery," "high retention and growing transaction frequency among core users," and "maintained leadership in both order volume and GTV" [12]. Over 800 million consumers used the platform, and app DAU rose more than 20% year over year [13]. Crucially, management located its edge in the higher average-order-value segments — the profitable orders, not the ¥2 coffees the subsidies bought.

Sources: order-share estimates are analyst/press figures (JPMorgan data as reported by Sixth Tone and CNBC), late 2025; positioning per those accounts. Meituan's leadership on order volume and GTV is per the Q3 FY2025 earnings call [14].

Read the share figures with care. The late-2025 snapshot — Meituan near 50%, Alibaba near 42%, JD around 8% by daily order count — flatters the challengers, because their totals include tens of millions of subsidy-driven, low-value instant-commerce orders inside a pool that itself had more than doubled on promotion. On a restaurant-delivery GMV basis the gap is wider: Alibaba's own claim was a 40% GMV share "under a two-player market definition," implying Meituan retained roughly 60% of the economically meaningful volume. The honest read is that Meituan's share slipped from a commanding level but its lead in profitable order value did not break. Its "absolute" advantage proved defensible, not unassailable — a distinction that decides whether 2025 was impairment or trough.

Why the moat is real: density, not slogans

The advantage rivals struggled to copy is physical, not promotional. Meituan's delivery network is, in its own long-standing words, "at the core of our food delivery and Meituan Instashopping businesses, and it has always been our competitive advantage" [15]. That network — the on-demand fleet management the company built to a global scale over a decade [16] — produces order density, and density is what lets a courier chain three deliveries in one loop instead of one. A challenger can buy an order with a subsidy; it cannot instantly buy the merchant supply, the rider liquidity, and the dispatch efficiency that make that order profitable at scale. That is why the war cost Meituan a fifth of a margin point but cost JD and Alibaba far more per order gained.

The test cuts the other way in categories where density is thinner. The second front — instant retail, where Instashopping quick-commerce revenue still grew 15.9% to about ¥28bn in the third quarter [17] — is exactly the ground Alibaba and JD chose, because grocery and general-merchandise fulfilment is younger and share is more mobile. Meituan is consolidating there rather than retreating: in February 2026 it agreed to acquire the fresh-grocery operator Dingdong for an initial US$717 million [18]. In-store services — the highest-margin corner of Core Local Commerce, and one already stress-tested by short-video platforms — held its "dominant mindshare" through the year on review depth and merchant coverage [19]. Execution is not itself a moat; the moat is the density and the two-sided supply the execution has compounded, and it showed up as a defended position rather than a slogan.

The turn, and what would change the read

By late 2025 the war was de-escalating under regulatory pressure. China's market regulator summoned all three platforms, published draft rules restricting long-running, large-scale subsidies, and by August had extracted public commitments to stop the price war. The first-quarter 2026 results are the first read on the other side: management said "industry-wide subsidies eased," competition shifted "back toward fundamentals like service quality and efficiency" — Meituan's own turf — and unit economics improved [20]. The total segment operating loss narrowed from a ¥15.3bn low in the third quarter of 2025 to ¥4.1bn, on revenue of ¥91bn [21] [22]. Management says it remains "confident in the long-term growth potential and competitiveness" of the segment [23].

The read here is that the moat is real but narrow: a scale-and-density cost advantage that survived a two-front assault by better-capitalised rivals, but only by spending its own profit to do so. The strongest fact against it is that the advantage was contestable at all — a supposedly dominant franchise had to burn a year of segment earnings to defend share, and it did lose some. What would flip the read from trough to impairment is straightforward to watch: a re-escalation of subsidies into 2026, or share and order-mix that keep eroding once promotions stop. Management itself flags a nearer risk — that second-half 2026 order growth "could turn negative year over year" on tough comparisons [24]. One quarter of rational competition is evidence, not proof. But the direction — losses a third of their peak, competition back on service and efficiency — is the direction a defensible trough would take.

---

What the Price Implies

At about HK$80.9 a share, Meituan is worth roughly ¥454 billion — and about ¥367 billion once ~¥87 billion of net cash is stripped out. That enterprise value is close to 7x the pre-war operating profit of its Core Local Commerce engine alone, and about 13x pre-war group earnings [1]. Consensus models a loss in 2026 and a partial recovery in 2027, not a clean return to 2024. The price embeds a discounted, gradual normalization — neither permanent impairment nor a full rebound. Financials are in renminbi (¥); the shares trade in Hong Kong dollars (HK$).

What the market is paying

The market capitalization is straightforward: 6,111,665,005 issued shares [2] at HK$80.9 give roughly HK$494 billion, about ¥454 billion or US$63 billion. The enterprise value depends on how much cash the balance sheet really holds net of debt — and here the headline "net cash" story needs tightening.

Meituan held ¥106.8 billion of cash and ¥60.1 billion of short-term treasury investments at end-2025, ¥166.9 billion of liquid assets in total [3]. Reported bank borrowings are small, ¥22.3 billion [4]. But a second interest-bearing line, notes payable, carries another ¥58.0 billion — ¥47.4 billion of senior notes and ¥10.7 billion of convertible bonds [5]. Counting both, interest-bearing debt is ¥80.3 billion, consistent with the company's own gearing ratio of about 53% of equity [6]. Net cash is therefore about ¥87 billion — real and substantial, but well below the ¥145 billion a cash-minus-bank-borrowings shortcut would suggest, because most of the debt sits under "notes payable," not "borrowings."

Values: market cap ¥454bn (6.11bn shares at HK$80.9), net cash ¥87bn, enterprise value ¥367bn. Source: share count and debt from FY2025 Annual Report [7][8]; price and share count as reported.

Net cash of ¥87 billion is about 19% of the market value, and it dwarfs the cash the war actually consumed: free cash flow was negative ¥27.1 billion in 2025, the worst year of the conflict [9]. On that draw the balance sheet funds roughly three more years of peak-intensity losses before the cushion is gone — and a further ¥45.6 billion investment portfolio sits outside this figure [10]. Survival is not the question the valuation turns on. Earnings power is.

The earnings the price is measured against

A trough valuation only means something against a normalized number, and Meituan's normalized earnings engine is Core Local Commerce — food delivery, in-store, hotel and travel, and instant retail. Before the war, that segment earned ¥38.7 billion in 2023 (an 18.7% margin) and ¥52.4 billion in 2024 (20.9%) [11]. In 2025 it swung to a ¥6.9 billion loss [12]. New Initiatives has been a persistent drag — a ¥20.2 billion loss in 2023, narrowing to ¥7.3 billion in 2024, then widening again to ¥10.1 billion in 2025 as overseas (Keeta) investment stepped up [13][14].

Source: FY2024 Annual Report [15] (2023) and FY2025 Annual Report [16] (2024–2025).

The point the chart makes is that the pre-war earnings power was concentrated and rising: the core segment produced ¥52 billion of operating profit growing at a mid-30s pace, while the whole New Initiatives basket cost ¥7–20 billion a year. At the group level, that netted to ¥36.8 billion of operating profit and ¥35.8 billion of net income in 2024 — ¥5.85 of basic earnings per share [17]. That ¥52 billion core figure and the ¥5.85 group EPS are the two anchors any recovery case is measured against. Whether they return is the competitive question, not an accounting one; the moat verdict there was contestable but intact.

The shape of the recovery consensus models

The sell-side does not expect 2024 to come back quickly. Consensus puts 2026 revenue at ¥402.6 billion (up 10.3%) but earnings still slightly negative, at a loss of ¥0.56 per share; 2027 revenue is seen at ¥457.3 billion (up 13.6%) with EPS recovering to ¥4.07 — real, but still below the ¥5.85 the company earned in 2024.

Revenue and EPS through 2025 are reported; 2026–2027 are consensus estimates (23–40 contributing analysts). Source: FY2025 Annual Report, Financial Summary [18] (revenue) and Consolidated Income Statement [19] (EPS); consensus analyst estimates, as of July 2026 (forecasts).

Two features of the estimate set are worth separating, because they cut in opposite directions. The trough is being marked up: the 2026 EPS estimate has improved from a ¥0.65 loss ninety days ago toward a ¥0.56 loss, and near-term revisions run heavily positive. The recovery is being marked down: the 2027 EPS estimate has slid from ¥4.89 ninety days ago to ¥4.07, with more cuts than raises in the last month. Consensus, in other words, now sees a shallower bottom but a slower climb.

Source: consensus analyst EPS estimate trend, trailing 90 days, as of July 2026.

Sentiment on the stock is constructive despite the loss year: of 38 rating firms, 28 carry buy or strong-buy ratings against 2 strong-sells, and the mean price target of about HK$106.9 sits roughly 32% above the current HK$80.9. Price targets, though, are a poll, not a valuation — the useful work is in what the current price itself requires.

What the price requires

Translating HK$80.9 into the reporting currency, each share costs about ¥74.2, of which ¥14.2 is net cash — leaving ¥60 of price attributable to the operating business. Set against the different earnings anchors, the multiples fall out as follows.

P/E on price of ¥74.2/share; ex-cash strips ¥14.2/share of net cash. 2026 is a modelled loss, so no meaningful multiple. Source: derived from FY2025 Annual Report earnings [20] and consensus estimates.

The enterprise-value lens is starker. At ¥367 billion, the operating business is valued at about 7.0x the pre-war operating profit of Core Local Commerce alone (¥52.4 billion), or about 10.0x pre-war group operating profit (¥36.8 billion) [21][22]. On that arithmetic the market is assigning effectively no value to New Initiatives — the overseas Keeta build-out and the grocery-retail formats — and little to growth beyond the 2024 level. That is the sense in which pessimism is priced: at these multiples the buyer is paying for a discounted version of the pre-war core and getting the option on everything else for free.

The read that follows is measured rather than emphatic. If Core Local Commerce re-attains something near its ¥52 billion pre-war profit and resumes growing — the base case implied by the moat holding through the war — then 7x segment EBIT and ~13x pre-war earnings, with a fifth of the market cap in cash, is inexpensive for a franchise of this quality. The evidence for that path is Q1 2026's sharp narrowing of segment losses as competition shifted back toward service and efficiency (the delivery war). The strongest fact against it is that consensus, closest to the company, is cutting its 2027 recovery estimate even as it lifts the trough — a signal that the professionals modelling this see the pre-war margin structure returning more slowly, and perhaps only partly, now that JD and Alibaba remain in the market. What would decide it is the trajectory of Core Local Commerce operating margin over the next few quarters: a return toward the high-teens would validate the cheap-trough read; a plateau in low-single-digits would confirm that the price is discounting a permanently lower normal, and 18x depressed 2027 earnings is not a bargain.

What would change the read

Three observable items move this valuation more than any target does. First, quarterly Core Local Commerce operating margin (disclosed each results announcement): the swing from ¥12.9 billion of Q4-2024 segment profit to a ¥10.0 billion Q4-2025 loss is the depth of the hole [23]; the rate at which it refills is the recovery. Second, the New Initiatives loss run-rate — if overseas investment keeps widening it, the "free option" carries a real annual cost against the core's recovery [24]. Third, the direction of consensus 2027 EPS revisions: continued cuts would mean the market is still repricing the normal downward, and the multiple is less cheap than the pre-war anchor implies.

---

Demand Engine

The recovery the price already discounts rests on one thing being true: that the demand underneath Meituan is large, under-penetrated, and still growing. Through 2023 the primary record supports that — on-demand delivery transactions grew 23.9% to 21.9 billion, and instant retail grew faster off a smaller base [1]. The catch is that Meituan stopped publishing the hard demand metrics after 2023, so the most important years of the war can only be read through management's phrase "historic highs." Financials are in renminbi (¥).

A large market, lightly penetrated

Meituan's addressable market is the digitization of Chinese local services — restaurants, groceries, pharmacies, hotels, and the errands people once did on foot. At the 2018 listing, the company sized the consumer service industry at ¥18.4 trillion in 2017, growing to a projected ¥33.1 trillion by 2023, a 10.2% compound rate, per the iResearch report it commissioned [2]. The load-bearing number is not the size but the penetration: of that ¥18.4 trillion, only ¥2,705 billion of gross transaction value had moved to consumer-service e-commerce in 2017 — roughly one-seventh online [3]. A large market that is mostly still offline is the definition of a long runway.

That runway is management's own recurring theme, not a one-off IPO claim. Five years on, the FY2023 report still described "tremendous room" for the online penetration of local services and called the sector's digital transformation "inevitable" [4]. The framing is credible because it matches the mechanism: local-service merchants are fragmented, low-margin, and slow to digitize, so the online share moves up over years, not quarters.

One caveat belongs here in plain terms. The precise market-size figures above are dated — they come from the 2018 prospectus, and a current third-party estimate of China's instant-retail or local-services market was not available in this corpus. The direction (large, under-penetrated, growing) is well supported across the filings; the exact 2025 magnitude is not something the primary record pins down.

The engine, while it was visible

Demand growth on Meituan comes from three multiplying levers: more users, higher frequency, and more categories per user. For the years the company disclosed them, all three moved the right way.

Source: FY2023 Annual Report, Operating Metrics; FY2022 Annual Report, Operating Metrics [5][6].

On-demand delivery transactions — the count Meituan defines as food delivery plus Instashopping — rose from 15.5 billion in 2021 to 17.7 billion in 2022 and 21.9 billion in 2023, with 2023 growth accelerating to 23.9% [7][8]. Underneath that, the user and frequency levers were both live.

Sources: FY2021 Annual Report, Operating Metrics [9]; FY2022 Annual Report, Operating Metrics [10]; FY2023 Annual Report [11].

Transacting users grew 35.2% in 2021 to 690.5 million, then dipped 1.8% to 677.9 million through the 2022 pandemic year [12][13]. With the user base near saturation of China's mobile-internet population, the growth shifted to frequency: the average transacting user placed 40.8 orders in 2022, up 14.1% from 35.8 the year before [14]. That is the healthier engine for a platform this large: a mature user typing in one more order a week compounds faster, and cheaper, than chasing the next marginal user. Food delivery's peak daily order volume tracked it — above 50 million in 2021 and past 60 million in 2022 [15][16].

Instant retail is the second, faster gear

The part of the demand story that is genuinely company-specific — and that most distinguishes Meituan from a mature food-delivery utility — is instant retail: groceries, medicine, flowers, electronics, and daily necessities delivered in under an hour through Meituan Instashopping and its InstaMart ("閃電倉") warehouse network. It runs on the same rider fleet as food delivery, so incremental demand lands on an asset that is already paid for.

The growth rates are a step above the platform average. Instashopping's highest daily order volume climbed from above 6.3 million in December 2021 to above 11 million in December 2022, and full-year order volume then grew "over 40%" in 2023 [17][18][19]. By 2023 the InstaMart format covered more than 200 cities and its active-merchant base grew nearly 30% in a year [20]. Management notes the users are worth more than food-delivery users — higher stickiness, stronger purchasing power, skewing younger — which matters because instant retail carries a larger basket than a single meal [21].

Source: FY2021 and FY2022 Annual Reports, Chairman’s Statement [22][23].

The strategic point is one Meituan made as early as 2021: it believes "the endgame of the retail industry is 'Everything Now'" — same-hour delivery of anything, not just food [24]. By 2024 the report was describing on-demand retail as "a new lifestyle," with InstaMart growing fastest in lower-tier cities — the least-penetrated part of the map [25]. By 2025 InstaMart and the self-operated Xiaoxiang Supermarket were labelled "important supply pillars for quick commerce" [26]. Instant retail is where the runway from the previous section actually gets consumed: food delivery in top-tier cities is a frequency game, but quick commerce still adds categories and geographies.

The demand pull shows up in the reported revenue line, even in the profit trough. Core Local Commerce revenue grew 20.9% in 2024 to ¥250.2 billion, which the company attributed to "further online penetration and strong consumer demand" — not to price [27]. The top line kept growing at a double-digit rate through 2025; what broke was margin, not demand — the distinction the Delivery War turns on.

What you can no longer see

Here is the honest limit on this chapter. Meituan disclosed a clean operating-metrics table — transacting users, transactions per user, on-demand delivery transactions — through its FY2023 report, and then stopped. The FY2024 and FY2025 annual reports carry no user count, no frequency figure, and no order count; the demand engine is described only in words. For 2025 the chairman's statement says annual transacting users, transaction frequency, and ARPU "all reached historic highs," but attaches no number to any of the three [28].

Two things follow. First, the disclosure thinned exactly when it mattered most, so an outside analyst has to take the 2024–2025 demand trajectory partly on trust. Second, whatever order growth did occur in 2025 was inflated by the subsidy war — some of it is ¥2 coffees bought because they were nearly free, not durable demand — and management has flagged that second-half 2026 order growth could turn negative against those subsidy-fattened comparisons (Delivery War). Volume in 2025 is not a clean read on structural demand.

The through-line question — whether 2025 is a durable impairment or a defensible trough — needs the demand base to be intact and still growing. The evidence through 2023 says the runway is real and instant retail is extending it; the post-2023 disclosure gap and the subsidy distortion mean that read now rests more on the mechanism (a large, under-penetrated market with a maturing frequency engine and a faster instant-retail gear) than on a fresh, verifiable count. The mechanism is sound; the confirming numbers are, for now, withheld.

---

Cash Conversion

Meituan's reported profit understates the cash the business generates. Operating cash flow exceeded net profit in every one of the last five years — by ¥9.6 billion to ¥26.6 billion a year — because the platform collects from consumers before it pays merchants and couriers, and because roughly ¥16 billion of annual charges are non-cash. Even the FY2025 loss converted to a smaller cash outflow (-¥13.8 billion) than the -¥23.4 billion accounting loss [1]. The pre-war earnings power the valuation rests on is cash-backed. Financials are in renminbi (¥); the shares trade in Hong Kong dollars.

Cash ran ahead of earnings

The clearest test of earnings quality is whether reported profit turns into cash. For Meituan it turns into more. In the two clean profit years, operating cash flow ran to 2.9 times net profit in 2023 (¥40.5 billion against ¥13.9 billion) and 1.6 times in 2024 (¥57.1 billion against ¥35.8 billion) [2]. In the loss years the same gap runs the other way and cushions the downside: the platform generated ¥11.4 billion of operating cash in 2022 while reporting a ¥6.7 billion loss, and in 2025 the operating outflow was ¥9.6 billion smaller than the loss [3]. Across 2023–2025 — spanning peak profit and the trough — cumulative operating cash flow of ¥83.8 billion is 3.2 times cumulative net profit of ¥26.3 billion.

Net profit per the consolidated income statements; operating cash flow per the consolidated statements of cash flows — FY2025 AR (2024–2025) [4] [5], FY2023 AR (2022–2023) [6] [7], FY2021 AR (2021) [8].

All figures ¥ billion. Source: consolidated statements of cash flows, FY2021, FY2023 and FY2025 Annual Reports [9].

Where the cash comes from

Two mechanisms explain the gap, and neither is aggressive accounting. The first is the working-capital cycle. Meituan holds ¥3.3 billion of trade receivables against ¥364.9 billion of revenue — receivable days of about three (3.3 in 2025, 2.9 in 2024) [10]. Consumers pay up front; the company then settles with merchants and couriers on a lag. That timing turns the platform into a net holder of other people's money. Payables to merchants (¥29.2 billion), advances from transacting users (¥12.0 billion), merchant and user deposits (¥6.8 billion) and deferred revenue (¥6.3 billion) together are an interest-free float of about ¥54 billion, up from ¥48 billion a year earlier [11] [12].

Source: FY2025 Annual Report, Consolidated Statement of Financial Position and Note 30 [13] [14].

The second mechanism is non-cash cost. The FY2025 loss carries about ¥10.2 billion of depreciation and amortisation and ¥6.0 billion of share-based compensation — roughly ¥16 billion that reduces reported profit without leaving the business [15]. Share-based compensation is a real cost to shareholders through dilution, and a skeptic is right to keep it in the earnings number; but at ¥6.0 billion, or about 1.6% of revenue, and falling from ¥7.6 billion in 2024, it is modest for a platform of this size and not a lever management is leaning on harder as profit falls [16].

The adjustments are mostly non-cash

Meituan reports an adjusted net loss of ¥18.6 billion for 2025 against a statutory loss of ¥23.4 billion, a ¥4.7 billion improvement [17]. The quality of an adjusted number is in what it removes. Here the gap comes from adding back ¥6.0 billion of share-based compensation and ¥2.1 billion of impairment provisions, partly offset by stripping out ¥2.3 billion of investment gains and ¥1.5 billion of intercompany foreign-exchange gains that had flattered the statutory result [18]. The adjustment removes non-cash and non-operating items in both directions rather than only the ones that help — the opposite of an aggressive non-GAAP presentation.

Source: FY2025 Annual Report, Reconciliation of Non-IFRS Measures [19].

Two things sit on the other side of the ledger. First, the bottom line is if anything cushioned, not flattered. The FY2025 operating loss of ¥25.0 billion already contains ¥2.4 billion of fair-value gains on Meituan's investment book and ¥3.3 billion of other net gains; a ¥1.5 billion income-tax credit then narrowed the loss further, from a ¥24.8 billion pre-tax loss to the ¥23.4 billion reported [20]. Strip the investment income out and the pure local-commerce trading loss is deeper than the headline — so the reported number is not masking a worse operating reality, it is disclosing one. Second, the balance sheet carries ¥27.8 billion of goodwill, about 8% of total assets, largely from historical acquisitions. An independent valuer tested it for impairment through the loss year and management booked none; the carrying value is unchanged year over year [21]. A held goodwill balance through a loss year is a judgement to watch, but 8% of assets is a moderate exposure, not a fragile one.

The trough drained cash too

The cash-quality story cuts both ways, and the down-cycle shows how. Free cash flow — after capital spending — fell to negative ¥27.1 billion in 2025, materially worse than the ¥13.8 billion operating outflow, because capex kept climbing even as profit collapsed: ¥6.9 billion in 2023, ¥11.0 billion in 2024, ¥13.3 billion in 2025 [22]. Free cash flow, not operating cash flow, is the honest measure of the trough, and it was the largest annual cash drain in the company's listed history.

The operating outflow itself deserves a closer look, because a large part of it was a choice. The single biggest working-capital movement in 2025 was a ¥12.9 billion increase in prepayments, deposits and other assets, and most of that is Meituan's micro-credit book: loan receivables carried at amortised cost jumped from ¥0.8 billion to ¥9.8 billion, and the total consumer-loan book roughly doubled to about ¥19.5 billion [23] [24]. Growing a lending balance sheet consumes operating cash while its funding — asset-backed securities and notes — sits in financing, where Meituan raised ¥42.2 billion in 2025 [25]. Set the ¥9 billion loan-book build aside and the core-platform operating cash bleed was closer to ¥5 billion than ¥14 billion. The classification is clean, but it means two things at once: the trough was less of an operating cash event than the headline suggests, and Meituan is scaling a fast-growing consumer-credit book — carried at loss-allowance coverage of only about 1% to 3% — into a downturn, a new risk line rather than a delivery one [26].

The float is the other reason to read good-year cash flow with care. It grows when volumes and spending grow, so it is a genuine tailwind while the platform expands and would reverse in a true contraction — which means operating cash flow in the strong years overstates the steady-state cash the business would throw off if it stopped growing. In 2025 the float did not reverse; it kept building, and the cash drain came from the loss and the lending book, not from merchants and users pulling their money back [27].

Two facts frame how comfortably the balance sheet absorbs all of this. Management's response to the drain was to stop returning cash and to lean on debt: share buybacks fell from ¥26.1 billion in 2024 to ¥0.4 billion in 2025, alongside the ¥42.2 billion debt raise [28]. And the debt itself is undemanding — gearing of about 53% of equity, roughly 55% of it maturing in three years or more, with no financial covenants on any of it [29]. The ¥87 billion net-cash cushion carried the ¥27 billion free-cash drain comfortably (What the Price Implies); the buyback-to-debt swing is a capital-allocation question in its own right.

The read: reported earnings convert to genuine cash, the adjustments are clean, and the FY2025 loss is a less severe cash event than it looks — which is what makes the balance sheet's ability to outlast the trough more than an arithmetic of the cash pile. The strongest fact against a purely benign read is that free cash flow, the honest measure of the trough, was negative ¥27 billion. What would change the read is a normalised year in which operating cash flow no longer runs well above net profit — float stops growing, capex stays elevated — because that would mean the cash-backing is a growth artefact rather than a structural feature of how the platform collects and pays.

---

Capital Allocation

Meituan has returned cash to shareholders only through buybacks, never a dividend — and the record shows steadier intent than timing. It spent about HK$28 billion repurchasing stock across 2024 at an average near HK$108, then all but stopped in 2025 as the price fell below HK$80 [1] [2]. In the same year it pivoted from returning capital to consuming it. The founders hold roughly a tenth of the equity but control about half the vote. Financials are in renminbi (¥); the shares, and every buyback, are priced in Hong Kong dollars.

Buybacks, never a dividend

Meituan has never paid a dividend. The board declined to recommend one again for 2025, as it has in every year since listing [3]. The single channel it has used to return capital is the share buyback, and it opened that channel only in 2024. Through 2022 and 2023 it repurchased nothing; in 2024 it bought back 261.4 million Class B shares for HK$28.2 billion; in 2025 it bought back 3.0 million shares for HK$392 million and cancelled them [4] [5].

Sources: FY2024 Annual Report [6] and FY2025 Annual Report [7]; current price HK$80.9 per the price history cited in What the Price Implies.

The timing is the problem. The 2024 purchases ran from January to September, and all but the first tranche were bought above HK$97 — the June and July tranches, HK$16 billion of the total, at HK$105 to HK$120, and the September tranche as high as HK$143.50. The one cheap tranche, January's HK$68–77, was the smallest of the large buys. Then in 2025, with the stock cheaper, the program did not scale up to meet it; the only purchase was a HK$392 million lot in May at HK$122–132, and nothing at all as the price sank toward HK$80.

Source: FY2024 Annual Report, Report of Directors, month-by-month repurchase table [8]; May 2025 tranche per FY2025 Annual Report [9].

This runs against what management said it was doing. On the February 2024 call, as the program began, it framed repurchases as "the preferred way to return capital right now, especially because it believes the stock is undervalued" [10]. By early 2025 the stated aim had narrowed to "continuing share repurchase programs to offset dilution from employee stock plans" [11], and a quarter later to considering repurchases "as a tool for enhancing shareholder returns, depending on market conditions" [12]. The company bought most heavily when it called the stock undervalued at HK$100-plus, and went quiet once the market handed it a lower price.

A share count that barely moved

For all the money spent, the buyback did not shrink the company. Meituan had 6,135,944,107 shares in issue at the end of 2021 and 6,111,665,005 at the end of 2025 — a reduction of about 24 million shares, or 0.4%, across four years [13] [14]. The HK$28 billion of repurchases was largely absorbed offsetting new shares issued under the employee option and award schemes rather than compounding per-share value.

Share-based compensation is the reason. It ran ¥6.0 billion in 2025 and ¥7.6 billion in 2024 [15], and the shares issuable under the schemes during 2024 alone equalled 2.37% of the weighted-average share count [16]. A buyback that keeps the count flat is doing useful work — dilution left unchecked compounds against holders — but it is defence, not return. When repurchases collapsed in 2025 while grants continued, that defence lapsed: the count is now set to drift up again unless the program resumes.

From returning capital to consuming it

2024 and 2025 were opposite years for Meituan's capital account, and the swing is the real story of this chapter. In 2024 the company was a net returner of capital: financing activities were a ¥30.4 billion outflow, dominated by the buyback. In 2025 financing flipped to a ¥21.2 billion inflow, "mainly attributable to the issuance of notes payable and proceeds from borrowings" [17].

Source: FY2025 Annual Report, Consolidated Statement of Cash Flows and MD&A; FY2023 Annual Report for prior years [18].

Four decisions turned the account around. The largest was the choice to fight the subsidy war: selling and marketing expense rose 60.9% to ¥102.9 billion in 2025, an increase of ¥38.9 billion that turned Core Local Commerce from profit to loss (The Delivery War) [19]. Alongside it, the company raised debt: in November 2025 it issued five tranches of senior notes — ¥2.08 billion and ¥5.0 billion of renminbi notes plus US$600 million, US$600 million and US$800 million of dollar notes — "primarily for refinancing of existing offshore indebtedness and other general corporate purposes" [20]. It scaled a lending balance sheet, letting the micro-credit loan book roughly double to about ¥19.5 billion — the largest single use of operating cash in the year (Cash Conversion) [21]. And in February 2026 it agreed to acquire a grocery target for an initial US$717 million, its largest disclosed acquisition in years [22].

None of this threatens the balance sheet — the roughly ¥87 billion of net cash and covenant-free, long-dated debt established in What the Price Implies and Cash Conversion absorb it comfortably. But the character of the capital account changed. A business that in 2024 was buying back its own shares became, in 2025, one funding a price war, a loan book, and an overseas push partly with borrowed money — while its own stock traded below the level at which it had been repurchasing.

Control without a majority

Who makes these calls, and on whose behalf, is set by the share structure. Meituan runs weighted voting rights: each Class A share carries ten votes, each Class B one [23]. Wang Xing, the founder, chairman and chief executive, beneficially owns 515.9 million Class A shares — about 45.3% of the vote — and co-founder Mu Rongjun a further 5.6%, so the two control roughly 51% of the vote [24]. Their combined 579.2 million shares are about 9.5% of the equity. The filing states the point plainly: the beneficiaries "do not hold a majority economic interest in the share capital of the Company" [25]. A tenth of the capital directs the whole.

Alignment and control are not the same thing, and here they pull in different directions. On alignment, the signals are good: Wang Xing takes almost no pay for the control he holds — total 2025 emoluments of ¥5.26 million, all salary, with no bonus and no new equity grant — so his stake, not his paycheque, is how he is paid [26]. The board carries a four-to-two independent majority with genuinely senior outside directors, and at the 2026 annual meeting the audit committee passed to a former KPMG China chairman. On control, the offsets are thinner: the chairman and chief executive roles are combined in one person, a stated deviation from Hong Kong's governance code that the board defends as "consistent leadership" [27], and the reserved matters on which one-share-one-vote applies are limited to constitutional questions — amending the charter, appointing auditors, removing independent directors, and winding up. Capital allocation is not among them.

There is a disclosure limit worth stating: as a Hong Kong issuer Meituan publishes no ongoing insider-transaction record, so the usual test of whether insiders are buying their own stock in the open market cannot be run here. What can be said is that the founders have not diluted themselves — the 2021 convertible bonds carry a HK$431.24 conversion price, more than five times the current quote, so they are debt that will not convert into shares [28].

Reading the record

The evidence points to a capable operator whose capital-allocation judgment is the weaker part of the case. The alignment is real and cheaply demonstrated — minimal pay, a large personal stake, no self-dilution — and the dominant 2025 decision, spending a year of Core profit to hold the franchise against two better-funded attackers, is defensible on the logic set out in The Delivery War. The strongest fact against a harsh read is the war itself: JD's February 2025 entry was not knowable when the 2024 buyback was running, and a company facing an unforeseen price war was right to preserve cash rather than keep repurchasing into it. Buying back stock while burning ¥27 billion of free cash flow would have been the greater error.

What the record does not support is the idea that management repurchases with discipline. It bought HK$28 billion of stock at an average near HK$108, most of it above HK$100 while calling the shares undervalued, and then declined to buy when the market offered the same shares below HK$80. That is the opposite order from value-accretive repurchase, and it is the clearest evidence in the file of imperfect market judgment steering a balance sheet that a tenth-of-the-equity founder controls outright.

The read would change with the next capital return. A buyback resumed at scale now — with the trough passing, the Q1 2026 recovery underway, and the stock near HK$80 — would show the 2024 timing was circumstance rather than habit, and would offset the dilution now running unchecked. Continued quiet on repurchases, with the freed cash flowing instead into the loan book, overseas expansion and acquisitions, would confirm a management that prefers building the company to compounding its shares — a preference minority holders own about a tenth of, and vote on barely at all.

---

New Initiatives

New Initiatives — grocery retail, overseas delivery under the Keeta brand, and a growing consumer-credit book — earned RMB104.0 billion of revenue in 2025 and lost RMB10.1 billion, and the market assigns the whole segment close to nothing [1] [2]. The loss widened only because Meituan chose to step up overseas investment; its domestic grocery business improved through the year. Keeta is already profitable in Hong Kong. The zero mark is defensible, but on the evidence a touch harsh. Financials are in renminbi (¥); the shares trade in Hong Kong dollars.

What the market pays

The valuation lens (What the Price Implies) leaves this segment carried at roughly nothing. Strip the net cash out of the market value and the enterprise value is close to seven times the RMB52.4 billion of operating profit that Core Local Commerce earned before the 2025 subsidy war — a multiple the core engine alone can justify [3]. What is left over — a business with RMB104.0 billion of revenue, more than a quarter of the group's RMB364.9 billion top line — is thrown in for free.

Source: FY2025 Annual Report, segment revenue and operating profit by segment [4] [5].

A zero — even a slightly negative — mark on a business losing RMB10.1 billion a year is not obviously wrong. The question worth working through is whether what sits inside the segment is a permanent cash sink the price rightly ignores, or a bounded, self-funded option that is worth more than the nothing embedded in the stock.

Narrowing losses, then a step up

The segment's own history cuts both ways. New Initiatives is the bucket where Meituan houses whatever it is trying to build next, and for four years the story was one of retreat from an expensive land grab. The operating loss ran to RMB35.9 billion in 2021 — a negative 84.5% margin, most of it poured into community group-buying — then narrowed each year: RMB28.4 billion in 2022, RMB20.2 billion in 2023, and RMB7.3 billion in 2024, by which point the margin had recovered to negative 8.3% [6] [7]. Revenue climbed the whole way — RMB59.2 billion in 2022, RMB87.3 billion in 2024 — as the segment grew into its cost base [8] [9].

Source: FY2022, FY2024 and FY2025 Annual Reports, segment revenue and operating profit by segment [10] [11] [12].

Then 2025 broke the trend: the loss widened to RMB10.1 billion, a negative 9.7% margin [13]. Revenue still grew 19.1% to RMB104.0 billion, and it did so even as Meituan wound down Meituan Select ("美团優選"), its community group-buying arm — exiting loss-making regions in June 2025, closing the chapter that had cost the segment most of its early losses [14] [15]. The reversal, in other words, was not a domestic business deteriorating. It was capital being redirected.

Why the 2025 loss widened

Management is explicit about the cause: the wider loss was "primarily driven by more investments in our overseas businesses, partially offset by our efforts in improving operating efficiency in our grocery retail businesses" [16] [17]. The quarterly path shows how back-loaded that spending was. The segment loss actually shrank through the first three quarters of 2025 — RMB2.3 billion, RMB1.9 billion, RMB1.3 billion — as domestic grocery kept improving. It then jumped to RMB4.6 billion in the fourth quarter, a negative 17.1% margin, when Keeta pushed into four new countries at once. In the first quarter of 2026 it narrowed again, to RMB2.1 billion [18] [19] [20] [21] [22].

Source: Q1–Q3 2025 Results Announcements, FY2025 Annual Report, and Q1 2026 earnings call [23] [24] [25].

That pattern matters for how the loss should be read. It is discretionary and it is bounded: management guided that the 2026 New Initiatives loss "will not exceed" the 2025 level, and the spending accelerates or slows with the pace of overseas launches rather than with any structural decline at home [26]. A loss a company can dial down at will is a different object from a loss it is fighting to contain.

Keeta, the model exported

The overseas business is where the free option has its clearest content. Keeta is Meituan's own delivery playbook — courier density, dispatch technology, merchant supply — run outside China. After launching in Riyadh in October 2024, it expanded across the major cities of Saudi Arabia through 2025 [27]. By early 2025 it was the largest food-delivery platform in Hong Kong; Meituan also committed a US$1 billion, five-year investment to enter Brazil [28]. Through the year it reached 20 Saudi cities and launched in Qatar, Kuwait, the United Arab Emirates and, in Santos and São Vicente, Brazil [29] [30] [31].

The evidence that the model travels is Hong Kong. Keeta reached its first profitable month there in October 2025 — ahead of a three-year plan — and delivered positive unit economics for the fourth quarter [32] [33]. A single profitable city is not a profitable overseas business, but it does convert the export thesis from assertion to demonstration. By the first quarter of 2026 management had shifted the emphasis from new-market entry to efficiency in the markets already open, saying it would "prioritize operational improvement over aggressive new-market expansion" [34].

Sources: FY2024 and FY2025 Annual Reports; Q1–Q3 2025 and Q1 2026 earnings calls; company news [35] [36] [37] [38] [39].

What the disclosure does not give is scale. Meituan reports no overseas revenue line, no Keeta order volume or GTV, and no split of the segment loss between overseas and domestic. An outside investor can see that Keeta is growing and that Hong Kong works; they cannot size the business or model its path to group-level profitability. That opacity is itself part of why the market pays nothing — it is hard to value what the company will not quantify.

Grocery and the fintech book

The rest of the segment is domestic. The largest piece is grocery retail, led by Xiaoxiang Supermarket ("小象超市") — a self-operated, first-party grocery model built on front distribution centres, which expanded to 55 cities by the first quarter of 2026 and which management runs toward "a sustainable low single-digit profit margin over the long run" [40]. Meituan agreed to acquire the mainland-China assets of Dingdong for US$717 million to deepen its grocery supply chain and East China coverage [41]. The segment also houses B2B food distribution (Kuailv), bike and e-moped sharing, power banks, and a consumer micro-credit book [42].

Two features temper the grocery optionality. A low-single-digit target margin, even on a business the size grocery could become, is a low-return outcome by the standard of Meituan's core marketing take rate — this is scale, not a second profit engine. And the micro-credit book roughly doubled to about RMB19.5 billion in 2025, carried at a thin loss allowance (Cash Conversion); it adds a genuine credit-risk exposure to a segment already valued as an option, and its ultimate contribution depends on reserving that has not yet been tested through a full consumer downturn.

What the free option is worth

The market's near-zero mark is defensible: this is a business that lost RMB10.1 billion in 2025, discloses too little to be modelled from the outside, and sits in a segment with a track record of funding a land grab and then retreating — community group-buying consumed most of the segment's early losses before Meituan wound it down in 2025 [43]. Read against that history, Keeta could be the next expensive experiment rather than the next franchise.

The weight of the current evidence points the other way, modestly. The 2025 loss was a deliberate, bounded investment, not a domestic business breaking down; management can and says it will hold the 2026 loss no higher; and Hong Kong shows the delivery model exports to profitability rather than merely to scale [44] [45] [46]. On that reading the segment is worth more than nothing — a real call option a disciplined buyer gets thrown in, not a reason to own the stock and not a hidden crown jewel.

The strongest fact against that read is the fourth quarter of 2025: the segment's worst loss of the year, RMB4.6 billion, arrived precisely when Keeta was spending fastest, and management would commit only to not exceeding that pace in 2026 — not to shrinking it [47] [48]. Three things would decide the read in either direction: whether Keeta reaches profitability in a market beyond Hong Kong, whether Meituan begins disclosing overseas revenue and unit economics so the option can be sized rather than trusted, and whether the 2026 segment loss holds at or below RMB10.1 billion. A loss that breaks above that line would turn the option back into a sink.

---

Scenarios and Watch Items

The report's six lenses reconcile to a genuinely two-sided question, and the sell-side's own spread frames it: FY2027 earnings estimates run from ¥1.21 to ¥6.32 a share and price targets from HK$57 to HK$139, against a HK$80.9 price. The balance sheet takes solvency off the table, so the outcome depends most on how fully Core Local Commerce margins refill after the war — an outcome now observable through a short list of quarterly markers, not a forecast one has to take on faith. Financials are in renminbi (¥); the shares trade in Hong Kong dollars (HK$).

Bull and bear on the same facts

Every earlier chapter surfaced the same feature: the load-bearing facts are agreed, and the disagreement is entirely about what they imply. Core Local Commerce swung from a ¥52.4 billion operating profit to a ¥6.9 billion loss in 2025 [1], then the group's total segment operating loss narrowed to ¥4.1 billion by the first quarter of 2026, with the Core segment loss down to about ¥2 billion [2]. A bull reads a chosen, temporary defence already reversing; a bear reads two permanently-installed, well-funded competitors and a margin that may never fully return. The same holds down the list.

Sources: segment profit/loss, FY2025 Annual Report [3]; liquidity and financing, p.35 [4]; 2024 buyback, FY2024 Annual Report [5]; 2025 buyback, FY2025 Annual Report p.72 [6]; New Initiatives and Keeta, Q3 2025 call [7]; pre-war EPS, p.212 [8]; consensus estimates as of July 2026.

None of these disputes is about the numbers. Each is about durability — whether the war was an episode or a regime change — and each resolves into something a filing or a results announcement will report over the next four quarters. That is what makes the question tractable rather than rhetorical: the decisive evidence arrives on a schedule.

Which way the trough is turning

The clearest early read comes from the quarterly path of the group's total segment operating loss, which widened into late 2025 and then closed sharply. The loss ran at ¥15.3 billion in the third quarter of 2025 [9] and ¥14.7 billion in the fourth [10], then narrowed to ¥4.1 billion in the first quarter of 2026 as industry subsidies became more rational and competition shifted back toward service and efficiency [11].

Total segment operating loss (a positive bar is a larger loss). Sources: Q3 2025 [12], Q4 2025 [13] and Q1 2026 [14] earnings calls.

One quarter of narrowing is evidence, not proof. Management itself paired the improvement with a caution that second-half 2026 order growth could turn negative year over year on tougher comparisons, and that third- and fourth-quarter delivery costs are seasonally higher [15]. A re-escalation of subsidies would reset the clock. The single first-quarter data point favours "defensible trough" over "durable impairment," but it is the direction over the next three quarters — not this one — that settles the reading, and the competitive verdict was that the moat is real but contestable.

Three scenarios

Because the scenarios are most sensitive to how far Core margins refill, the sell-side's estimate range does the scenario work without invention. Consensus FY2027 EPS sits at ¥4.07, with a low of ¥1.21 and a high of ¥6.32, and the matching price targets run HK$56.7 to HK$139.2 around a HK$106.9 mean. Those three points bracket three outcomes, each measured against the pre-war anchor of ¥5.85 in basic earnings per share (FY2024) [16].

P/E on a HK$80.9 price (~¥74 at HK$1 = ¥0.918); ex-cash strips ~¥14/share of net cash. EPS anchors and dispersion are consensus estimates as of July 2026; pre-war EPS from FY2025 Annual Report [17]; scenario labels are this report's framing of the sell-side range.

The matrix makes the asymmetry legible. In the base case — a partial recovery to earnings still one-fifth below 2024 — the shares change hands at about 18x, or under 15x stripping out cash, and the mean target implies roughly a third of upside. The bull case is not heroic: it needs 2027 earnings back near the pre-war level, at which point the price is under 12x, or below 10x ex-cash, and the high target implies about 70% upside. The bear case is the discipline: if JD and Alibaba hold Core margins permanently down and 2027 earnings land near ¥1.2, then today's price is 50–60x depressed earnings, the multiple is the wrong anchor, and the low target implies roughly 30% downside. That the market's own analysts span this range — three sell ratings against 28 buys, and a ¥1.21-to-¥6.32 EPS spread — is itself the finding: the outcome is not yet decided by the evidence, and the price sits closer to the bear end on trough earnings than to the bull end on normalised ones. The full multiple derivation is in What the Price Implies.

What to watch

Each scenario resolves through observable, falsifiable items, and each appears in a document on a known cadence. The list below is the report's consolidated watch-list; the first three are the highest-signal.

Anchors: Core and New Initiatives segment margins, FY2025 Annual Report [18]; order-growth caution and 2026 New Initiatives guidance, Q1 2026 [19] and Q4 2025 calls [20]; 2025 buyback, p.72 [21]; loan-book loss coverage, Note 22 p.311 [22]; Keeta, Q3 2025 [23].

Three of these — the quarterly Core margin, the New Initiatives loss run-rate, and the direction of 2027 revisions — carry the most valuation weight and were flagged in the valuation chapter. The other four are stewardship and optionality tests: whether the founders redeploy the cushion well, whether the fintech balance sheet is reserved adequately, and whether the overseas option is one profitable city or a template. Each has a named threshold, so a reader can mark the thesis to the evidence quarter by quarter rather than re-underwriting it wholesale.

What the balance sheet buys

What ties the scenarios together is that the downside is bounded by a real asset and the upside is carried by an option the price ignores. Net cash of about ¥87 billion — roughly 19% of the market value — plus a further investment portfolio absorbed the worst free-cash drain in the company's listed history (-¥27.1 billion in 2025) without straining, because cash and short-term treasury of ¥166.9 billion sat against ¥80 billion of interest-bearing debt at year-end [24]. That cushion does not decide the outcome; it buys the time for the outcome to arrive, and it is why the bear case damages the multiple and the wait rather than the solvency.

On the evidence to hand, the weight sits with a defensible trough: the segment loss is narrowing, the New Initiatives loss is guided flat, Keeta has proven the model exports at least once, and the price pays for only a partial recovery. The strongest fact against that read is that consensus keeps cutting its 2027 recovery estimate even as it lifts the 2026 trough — the people closest to the company are pricing a slower, shallower normalisation than a clean rebound. What would change the read in either direction is on the watch-list above, and most of it reports within four quarters: a Core margin that stays negative through 2026 would move the weight toward impairment; a return toward pre-war margins with the New Initiatives loss held flat would move it toward the bull case. The report set out to judge whether 2025 was a durable impairment or a trough the balance sheet could outlast; the balance sheet has done its part, and the next four quarterly margins do the rest.

---

The AI Bet

Meituan's research and development spend rose 23.5% to ¥26.0 billion in FY2025 — a record, and the first increase in four years after three years of falling R&D intensity [1]. Management attributes the step-up to central AI investment: the proprietary LongCat models and the Xiaomei and Xiaotuan assistants [1]. At ¥4.9 billion the increase is small next to the ¥39.0 billion subsidy-war marketing surge, but unlike that surge it looks structural — and so belongs in normalised earnings.

Source: FY2025 Annual Report, MD&A — Research and Development Expenses [1].

R&D broke a three-year discipline

For three years Meituan held its research budget flat while revenue grew, and let the ratio fall. R&D was ¥20.7 billion in 2022, up from ¥16.7 billion in 2021, holding at 9.4% of revenue [2]. It stayed at ¥21.2 billion in 2023 as the ratio dropped to 7.7%, and ¥21.1 billion in 2024 as it fell again to 6.2% [3]. The absolute number barely moved for three years; the operating leverage came from spreading it over a larger business.

FY2025 reversed that. Spend rose to ¥26.0 billion and the ratio ticked back up to 7.1%, the first increase in either measure since 2022 [1]. The pace accelerated into the fourth quarter, when R&D rose 29.7% year over year to ¥7.0 billion and reached 7.6% of revenue [4]. The consolidated income statement carries the same figures — R&D of ¥25,998 million against ¥21,054 million a year earlier [5].

Source: derived from FY2022, FY2023 and FY2025 Annual Reports, MD&A — R&D Expenses [2] [3] [1].

The intensity number carries a useful calibration. At 7.1% of revenue, FY2025 R&D is still below both 2022 (9.4%) and 2023 (7.7%) [2] [3]. What changed is not the size of the budget relative to the company — it is the direction. Three years of declining intensity ended, and a research line that had been a source of operating leverage became a source of drag.

Where the money is going

Management is explicit about the destination. Meituan launched its in-house large language model, LongCat, in early 2023, and has since built out a model family — foundational models through business applications — and open-sourced core versions covering dialogue, reasoning, omni-modal understanding, and image and video generation [6]. On top of LongCat it has released consumer-facing assistants: Xiaomei as a standalone app and Xiaotuan embedded in the Meituan app for all users, meant to turn search into natural-language requests [7].

The chairman frames this as strategy, not experiment. The FY2025 letter commits Meituan to "driving the AI transformation of the physical world," and the 2026 outlook elevates it to a "Retail + Technology" corporate strategy built on combining the LongCat models with the company's merchant data, consumption behaviour and user reviews [7] [8]. The spend reaches beyond consumer apps: an internal AI coding effort put a dedicated team in front of over 6,000 frontline engineers to raise development productivity [9], and the company continues to fund L4 autonomous delivery vehicles and a Shenzhen robotics institute as longer-dated fulfilment bets [10].

The competitive context argues that at least part of this is table stakes. Meituan's largest instant-commerce rivals are making the same commitment on a larger scale — Alibaba, whose Taobao Instant Commerce is the escalating second front in the delivery war (The Delivery War), told shareholders it is "scaling up investments in our full-stack AI capabilities" including AI infrastructure and proprietary chips [11]. A local-services platform whose largest competitors are all building AI-native assistants cannot credibly not spend here. The open question is whether Meituan's version buys more than parity — whether proprietary models fed by the platform's dispatch, merchant and review data become an efficiency and discovery advantage, or simply a cost of staying in the game.

Small next to the subsidy war — but it does not reverse

The AI budget is not what broke the 2025 income statement. Of the operating-cost increases that turned a ¥36.8 billion operating profit into a ¥25.0 billion operating loss, selling and marketing dominated, rising ¥39.0 billion year over year as Meituan defended share against JD and Alibaba [5]. The ¥4.9 billion R&D increase is the second-largest cost step-up of the year, but it is roughly an eighth the size of the marketing surge.

Source: derived from FY2025 Annual Report, Consolidated Statement of Comprehensive Income [5].

The distinction that matters for valuation is between the two. The marketing surge is cyclical: it is the cost of a subsidy war that regulators have moved to cool and that management expects to ease, and the recovery case (What the Price Implies) rests on most of it reversing. The R&D step-up carries no such expectation. Management describes AI as a permanent strategic pillar, not a defensive burst, and has told the market to expect more of it, not less [8]. When the marketing spend rolls off, the research spend most likely stays.

That has a specific consequence for the earnings anchor the valuation leans on. Meituan discloses R&D only at the group level, and the segment accounts allocate operating costs — R&D among them — into Core Local Commerce and New Initiatives rather than parking them in the corporate line: unallocated items were essentially flat in 2025, at ¥8.05 billion against ¥8.30 billion in 2024 [12]. So the ¥4.9 billion increase sits inside the segment cost bases. The pre-war Core Local Commerce operating profit of ¥52.4 billion in 2024 [13] was earned while group R&D was ¥21.1 billion. If research resets to ¥26 billion or higher and that cost stays allocated to the segment, a full restore of Core Local Commerce volumes would land at a normalised profit modestly below the ¥52.4 billion headline — by the segment's share of the step-up — unless AI efficiency offsets it. The trough reverses; part of the cost structure does not.

The read

The evidence points to an AI program that is defensive in origin and optional in payoff — a necessity Meituan could not skip, wrapped around a bet it might win. The necessity is clear: every major instant-commerce rival is funding frontier models, and a platform that mediates local-services discovery cannot cede the natural-language interface [11]. The optionality is real but unproven: LongCat fed by Meituan's dispatch, merchant and review data could lower fulfilment cost and deepen the density advantage that underwrites the moat (Demand Engine), or it could settle into a permanent 7%-plus revenue tax that buys parity and no more.

The strongest fact against treating this as a problem is that the spend is modest and self-funded. R&D intensity at 7.1% remains below where it sat in 2022 and 2023 [2], the increase is a fraction of the marketing swing, and it is being financed from a business that still converts profit to cash (Cash Conversion) rather than from new leverage. This is not a company betting the balance sheet on AI; it is a company redirecting a research line it had been shrinking. What would change the read is direction and payoff: R&D intensity climbing toward the 9%–10% range with no visible efficiency return in Core Local Commerce margins would mark a genuine reset of normalised earnings power; evidence that AI is lowering cost per order or lifting conversion would move it from cost to moat. Both are quarterly-checkable, and neither is settled yet.