01 — Executive summaryA fortress balance sheet around a falling number
MIXUE Group is the largest freshly-made drinks business in the world by store count and one of the most profitable. In the six months to 30 June 2026 it reported revenue of RMB 15.22 billion, up 2.3%, gross profit of RMB 4.63 billion, down 1.6%, and profit attributable to shareholders of RMB 2.30 billion, down 14.7%. Over the same six months its store network grew 20.7%, from 53,014 to 63,987 shops.
That gap is the report. MIXUE does not sell drinks to consumers; it sells ingredients, packaging and equipment to franchisees who do. Its revenue is therefore a function of what franchisees buy, and franchisees buy more when their own shops sell more. On the company's own disclosed figures, revenue per average store fell 17.8% year on year. The growth in the top line is coming entirely from new units, not from the units that already exist.
Against that, the balance sheet is exceptional. At 30 June 2026 the company held RMB 21.64 billion of cash, term deposits and wealth-management products, carried no interest-bearing bank borrowings, and had a liabilities-to-assets ratio of 17.7%. At HK$167.30 the equity is capitalised at HK$63.5 billion, so 37.2% of the market capitalisation is financial assets. The operating business is priced at about 7.8 times normalised operating earnings.
A cheap price on a falling number is the oldest trap in value investing, and the framework we apply here is explicit about the test: a stable industry, a durable moat, margins that hold, and a discount of 20% to 50% to intrinsic value. MIXUE passes on the moat, cash flow and solvency, and fails on industry structure and margin trend. Our base case is worth HK$196, roughly 15% above the price — a discount that is real but smaller than the framework demands. We rate the shares Neutral.
| Framework pillar | Test | MIXUE | Where |
|---|---|---|---|
| 1. Industry and macro | Mature, stable, no ruinous price war | Fail | §02 |
| 2. Business and moat | Durable cost advantage, pricing power | Pass | §03–04 |
| 3. Financial verification | Cash flow matches profit, low leverage | Pass | §09–12 |
| 4. Intrinsic value | 20–50% discount to value | Marginal | §15–16 |
| 5. Risk screening | No unresolved structural risk | Open | §19 |
| 6. Decision | Buy, hold or wait | Wait | §21 |
02 — The industryA growth market that became a price war
Value investing starts with the industry, not the company, because a good business in a bad industry is a value trap with a nicer logo. China's freshly-made drinks market has moved through its growth phase inside a decade. MIXUE's own terminal retail sales rose from RMB 22.8 billion in 2021 to RMB 58.3 billion in 2024, and the growth rate has fallen in every year since 2023.
Two structural facts matter. First, the market is not an oligopoly. It is a fragmented field of national chains, regional chains and single shops competing on price, and four listed operators now disclose enough to be compared. Second, the 2025 numbers were inflated by a delivery-platform subsidy war that lifted category order volumes and will not repeat. Management described the first half of 2026 as "passing through a very unusual phase" and named the subsidy rollback as one reason per-store sales fell.
Demand itself is reasonably rigid — a RMB 6 cup is a small, habitual purchase and the category has not lost volume — but it is intensely price-elastic, and that is the problem. The framework's own screen is blunt: it excludes industries with "brutal price wars and continuous profit dilution". On that test alone, MIXUE's industry is disqualified, and everything that follows has to overcome it.
03 — The businessA supply chain that happens to own a brand
MIXUE is usually described as a tea company. It is better understood as a manufacturer and distributor with a retail brand attached. Of FY2025 revenue of RMB 33.56 billion, RMB 32.77 billion — 97.6% — came from selling goods and equipment to franchisees. Only RMB 794 million, or 2.4%, came from franchise and related service fees.
The fee line is small but far more profitable: an 82.6% gross margin against 29.9% on goods and equipment, so it contributes 6.3% of gross profit from 2.4% of revenue. The company is explicit that it does not take a share of franchisee turnover. Franchisees own their shops, pay their own rent and wages, and are accountable for their own results. MIXUE's job is to supply them at a cost no independent operator can match.
Three brands sit on that chassis: MIXUE, the tea and ice-cream chain at RMB 2–8 a cup; Lucky Cup, a coffee chain at RMB 2–10 launched in 2017; and FULU Fresh Beer, acquired in December 2025, at RMB 6–11 a 500ml pour. MIXUE's brand accounted for more than 95% of revenue and gross profit through the prospectus period, and the other two are optionality rather than earnings.
04 — The moatWhat scale actually buys
The moat is a cost advantage, and it is genuine. MIXUE was the first company in the Chinese freshly-made drinks industry to build centralised factories, and it self-produces 100% of its core ingredients across six production bases in Henan, Hainan, Guangxi, Chongqing, Anhui and Yunnan. It built its own logistics network in 2014 and now runs 28 domestic warehouses serving 33 provincial regions and more than 300 cities. It buys raw materials at origin, at a scale no competitor matches, and it sells the output at a markup that still leaves the end product cheaper than anyone else's.
The evidence that the moat is real is that it survived the price war. Gross margin has held between 28.9% and 32.5% for five years while the industry competed on price, and the company has grown to 63,987 shops without spending on stores of its own. A business with no moat does not keep a 30% gross margin while its customers fight each other.
The important qualification is what the moat protects. It protects MIXUE's margin on the goods it sells. It does not protect the franchisee's margin on the drinks it sells, and MIXUE's growth depends on the franchisee continuing to open shops. The moat is in the supply chain; the risk is in the store.
05 — The franchiseeThe unit the whole model depends on
The flywheel is simple. A franchisee opens a shop, buys MIXUE's ingredients and machines, and MIXUE books revenue. If the franchisee makes money, more franchisees open more shops, and the supply chain gets bigger and cheaper. If the franchisee does not, the flywheel stops — and because MIXUE takes no share of turnover, the company does not feel the pain immediately. It feels it one step later, when the openings stop.
Revenue per average store, annualised, fell from RMB 598,000 in the first half of 2025 to RMB 491,600 in the first half of 2026, a decline of 17.8% and below every full year shown. Broker work on the interim report puts the fall in merchandise revenue per franchisee at 17.2%, which is the same number by a different route. The franchisee's costs — rent, wages, fit-out, the franchise fee — did not fall by 17.8%.
We cannot see franchisee profit and loss; it is not disclosed. What we can see is behaviour, and behaviour has changed. Openings in the first half fell 29.3% year on year and closures rose 8.6%. The number of franchisees still grew, from 23,404 to 29,775, so the network is not collapsing. But the marginal franchisee is more cautious than a year ago, and caution at the margin is how a supply-chain model stops compounding.
06 — The networkStill opening, less productively
At 30 June 2026 the group operated 63,987 stores across 17 countries: 59,609 in the Chinese mainland and 4,378 outside it. Only 36 were company-operated. Net additions were 4,166 in the half against 6,534 a year earlier.
The mainland network is the whole story, and it is close to complete. Third-tier cities and below are 58.0% of the mainland estate, a share that has barely moved in three years, and the company already reaches more than 300 cities across 31 provinces. The gaps that made the model work — towns with no bubble-tea shop at all — have largely been filled. Every additional shop in a county town now competes with the MIXUE that was already there.
Density is the mechanism behind the per-store decline. Management has responded by slowing openings and emphasising what it calls high-quality network development, which is the right decision and also the reason the growth rate will not return to 35%. A company that deliberately stops adding shops to protect the ones it has is telling you that the addressable market is close to served.
07 — MarginQuality is a cost before it is a benefit
The framework asks one question of margins: are they stable or rising, because that is what pricing power looks like. MIXUE's are falling, and the company says why. Gross margin has gone from 32.5% in FY2024 to 31.1% in FY2025 and 30.4% in the first half of 2026. Cost of sales grew 4.1% against revenue growth of 2.3%, and management attributes the gap to a deliberate upgrade it calls "real ingredients, fresh taste, simple recipes" — moving fruit, milk and tea from ambient to chilled and frozen, and adding fresh milk and short-shelf-life coffee beans.
This is a strategic choice, not a loss of control: the company is spending margin to defend a value proposition that a price war would otherwise erode. It may work. But the framework does not grade intentions, and the record so far is three consecutive periods of gross-margin decline and a net margin down 3.0 percentage points year on year to 15.1%. Until the company shows it can pass the cost through — by price, by mix, or by volume — the margin test is failed.
08 — Cost structureEvery line grew faster than the top line
Operating leverage cuts both ways. When revenue grew 35% a year, fixed costs were an asset. In the first half of 2026 revenue grew 2.3% and every material cost line grew faster: cost of sales 4.1%, selling and distribution 22.9%, administrative expenses 39.4%. Selling and distribution rose to 7.4% of revenue from 6.1%; administrative expenses to 4.0% from 2.9%.
The selling line reflects brand IP investment and support for store operations; the administrative line is mostly people. Headcount rose from 9,102 at the end of 2025 to 10,374 at the end of June 2026, and employee remuneration rose 11.4% year on year. Some of this is the price of the quality programme and the overseas build. Some of it looks like a cost base that was sized for a company still growing at 35%.
The framework's warning is about exactly this: "period expenses" that are rigid while revenue is not. A management that can grow costs 39% in a half when revenue grows 2.3% has not yet accepted that it is running a mature business. The question for the second half is whether the spending is a step change or a new baseline.
09 — The half-yearA cost story, not a demand story
Reading the half-year line by line is the fastest way to see what actually changed. Revenue of RMB 15.22 billion was RMB 341 million higher than a year earlier. Cost of sales was RMB 418 million higher. Gross profit therefore fell RMB 77 million even as the business got bigger.
Below the gross-profit line, other income and gains fell 19.6% to RMB 127.5 million on foreign-exchange losses; selling and distribution rose RMB 209 million; administrative expenses rose RMB 172 million. Profit before tax fell from RMB 3.47 billion to RMB 2.98 billion, and a lower tax charge — RMB 656 million against RMB 750 million — absorbed part of that. Profit attributable ended 14.7% lower at RMB 2.30 billion, and basic earnings per share fell 16.3% to RMB 6.05.
Two things are worth isolating. The fall is entirely on the cost side: gross profit and operating expenses account for RMB 458 million of the RMB 397 million decline, and demand did not collapse — revenue still grew. And the company's response has been to start returning capital. The board proposed a special dividend of RMB 2.65 per share, about RMB 1.01 billion, payable on or around 6 November 2026, which is 43.8% of first-half profit. It is the first distribution since listing.
10 — Cash flowDoes the profit convert?
The framework's core test of real profit is whether it arrives as cash. MIXUE's model is structurally good at this: franchisees pay for goods before they receive them, so trade receivables were RMB 61 million at 30 June 2026 against a half-year revenue of RMB 15.2 billion — about one day of sales. Contract liabilities, which are customer prepayments, stood at RMB 601 million.
Operating cash flow has exceeded reported profit in every year shown. But the margin has narrowed sharply: 1.35 times in FY2024, then 1.03 times in FY2025, as operating cash flow of RMB 6.04 billion barely covered profit of RMB 5.89 billion. Working capital is no longer a source of cash: inventories rose to RMB 3.83 billion and trade payables fell RMB 353 million in the half, both of which absorb cash.
Investment cash flow is heavily negative and always has been — RMB 70.2 billion went out in FY2024, much of it into the deposits that now sit on the balance sheet rather than into plant. The productive spending is the supply chain: six production bases, with the Yunnan site starting up in May 2026, and capital commitments of only RMB 205 million at the half-year. The cash is not being consumed by the capital programme.
11 — The balance sheetCash against the market cap
This is the strongest part of the case. At 30 June 2026 MIXUE held RMB 21.64 billion of cash, term deposits, restricted cash and wealth-management products, up 8.3% from the end of 2025. It had no interest-bearing bank borrowings, against RMB 28.2 million at the previous year end. Its liabilities-to-assets ratio fell to 17.7%, and total liabilities of RMB 5.81 billion are dominated by trade payables and accruals rather than debt.
Set against a market capitalisation of HK$63.5 billion, the financial assets are 37.2% of the equity value. The operating business — the factories, the warehouses, the inventory, the brand — is carried in the accounts at about HK$6.1 billion of book value and is priced by the market at HK$39.9 billion.
Two qualifications keep this from being a free lunch. The cash is onshore renminbi held by a PRC company; getting it to an offshore shareholder requires dividends, which carry withholding tax. And cash earning under 3% drags on return on equity while it sits there. A fortress balance sheet is a floor, not a strategy.
12 — ReturnsThe ROE question
The framework wants a high and stable five-year return on equity, because that is the cleanest evidence of a durable competitive advantage. MIXUE's has been high. It is no longer stable.
Reported weighted-average ROE was 34.7% in FY2023 and 34.5% in FY2024. It fell to 30.2% in FY2025, and annualising the first half of 2026 gives 17.9%. Decomposing the FY2025 fall: net margin slipped from 17.9% to 17.5%, asset turnover rose slightly, and the equity multiplier fell as the March 2025 listing enlarged the equity base. The first-half 2026 decline is simpler and worse — the margin fell three percentage points while equity kept growing.
A 17.9% annualised ROE is still a good return. It is not a 30% return, and the shares were priced for the latter until this year. The question a value investor has to answer is whether 18% is the new normal or the trough of a cycle; the company's own guidance does not settle it.
13 — OverseasThe second curve has not turned
Overseas expansion was the principal reason MIXUE carried a growth premium at listing. It is not currently working. The international network peaked at 4,895 stores at the end of 2024 and stood at 4,378 at 30 June 2026, falling in each of the last four reporting periods. It is 6.8% of the group, down from 11.5% at the end of 2023.
The reductions are deliberate: management says it closed and re-tooled shops in Indonesia and Vietnam to improve the quality of what remained. But the pattern is not confined to those two markets. The company entered Kazakhstan and the United States in 2025 and Mexico, Kyrgyzstan and Brazil in 2026, and added Mongolia and Uzbekistan in September. In Hong Kong, a market it has served since early 2024, local press reported six closures during 2026, including the flagship at Mong Kok, leaving a handful of outlets.
The reason is arithmetic. A model built on a RMB 6 cup assumes cheap rent and cheap labour. Hong Kong and Japan offer neither, and management has conceded that the store economics in those markets differ. Overseas is currently a cost and a story rather than a source of profit, and we do not capitalise it in our valuation.
14 — Management and governanceFounder-led, recently listed, newly generous
MIXUE is controlled by its founders. Zhang Hongchao is chairman and Zhang Hongfu, who ran the company as chief executive through the expansion, moved to co-chairman in March 2026; Zhang Yuan, previously executive vice-president and chief financial officer, took over as chief executive. The company has raised external equity exactly once, a RMB 2.33 billion round in December 2020 with Meituan Longzhu, Hillhouse and CPE Yuanfeng, and Hillhouse took a cornerstone position at the March 2025 listing.
There is no adverse governance record to report. The interim filing discloses no material litigation, no pledge or charge over assets, and no contingent liabilities. The structure is a PRC joint stock company with H shares listed in Hong Kong, which brings the usual minority-shareholder considerations but no company-specific irregularity in the filings.
Two things give us pause, and neither is misconduct. The record is short: this has been a public company for nineteen months, and the framework's governance tests assume a longer history of disclosed capital allocation. And the capital-allocation turn is new. The company recommended no dividend for FY2024 or FY2025, then proposed a special dividend worth 43.8% of first-half 2026 profit. That is a sensible response to a decelerating business, but it is a change of posture rather than a demonstrated policy, and it arrived in the same half in which profit fell 14.7%.
15 — ValuationEarnings power plus surplus cash
MIXUE cannot be valued on a single consolidated multiple, because 37% of the market capitalisation is financial assets unrelated to selling tea. We separate the two.
Our normalised FY2026 earnings base is RMB 5.0 billion of profit attributable to shareholders, which is the first-half run rate with the usual second-half seasonality applied and is close to where post-results broker estimates have settled. Of that, roughly RMB 350 million is interest earned on the surplus financial assets. The operating business therefore earns about RMB 4.65 billion.
| Measure | Value | Note |
|---|---|---|
| Share price, 13 October 2026 | HK$167.30 | 52-week low HK$166.00; 52-week high HK$462.72 |
| Market capitalisation | HK$63.5bn | 379.62 million shares in issue |
| Cash, deposits and wealth-management products | RMB 21.64bn | HK$62.31 per share, 37.2% of the price |
| Trailing twelve-month earnings per share | RMB 14.47 | HK$15.80 converted; 10.6× the price |
| Operating earnings, normalised | RMB 4.65bn | Excludes interest on surplus assets |
| Ex-cash price / operating earnings | 7.8× | HK$105.00 per share of operating value |
| Price / book | 2.17× | Book value RMB 70.42 per share |
| Special dividend, payable 6 November 2026 | RMB 2.65 | 1.7% yield; 43.8% of first-half profit |
Seven point eight times operating earnings is a low price for a business that self-manufactures 100% of its core inputs, converts revenue to gross profit at 30%, and needs almost no capital to grow. It is also the price the market applies when it does not believe the earnings are durable. Both are true; the second is what the framework cares about, because margin of safety comes from the stability of the number, not its size.
16 — Scenarios and margin of safetyWhat the price is already assuming
We run three cases, each built from two explicit inputs: the multiple applied to normalised operating earnings, and the value ascribed to the surplus financial assets. There is no third assumption hidden inside. In every case we add the financial assets, because they are cash and short-dated deposits rather than an operating asset.
| Scenario | Operating multiple | Operating earnings | Surplus assets | Value per share | Weight |
|---|---|---|---|---|---|
| Bear | 7.0× | RMB 4.65bn | 90% of RMB 21.64bn | HK$150 | 30% |
| Base | 10.0× | RMB 4.65bn | 100% | HK$196 | 50% |
| Bull | 13.0× | RMB 4.65bn | 100% | HK$236 | 20% |
Probability-weighted value HK$190. The bull case applies a multiple still below where comparable Chinese consumer franchises trade. Farstar estimates.
The distribution is the finding. Our base case is HK$196, 17.2% above the price. Our bear case, HK$150, is 10.4% below it and assumes operating earnings never re-rate from roughly where they are today and that a tenth of the surplus cash is consumed or lost. The upside case is 41% above the price. That is a modestly favourable skew, not a compelling one: the framework looks for a market price 20% to 50% below intrinsic value.
At HK$167.30 the shares sit 14.6% below our base case — a discount, and why we would not sell, but not the discount the framework requires for a purchase. Below roughly HK$145 the arithmetic changes.
17 — The bull caseIn one page
The cost moat is real and unbroken. MIXUE self-produces every core ingredient, owns its logistics and buys at origin at a scale nobody in the category can match. Gross margin has held between 28.9% and 32.5% for five years during a price war. No competitor has taken its position, and it sells cheaper than all of them.
The fall in per-store revenue is a base effect, not a demand failure. The first half of 2025 was inflated by delivery-platform subsidies that have since been withdrawn. Management called it an abnormal phase. As that comparison laps, per-store revenue should stabilise without any improvement in the underlying business.
You are buying cash at a discount. RMB 21.64 billion of financial assets is 37.2% of the market capitalisation, the company has no debt, and the operating business is priced at 7.8 times normalised operating earnings. A re-rating to 13 times — still below comparable Chinese consumer franchises — plus the cash gives HK$236.
The capital-return story is starting. A special dividend worth 43.8% of first-half profit, a stated willingness to consider buybacks, and a capital programme of only RMB 205 million of commitments mean the surplus cash can come back. If it does, the discount closes on its own.
18 — The bear caseIn one page
The market is saturated and the company knows it. Revenue per store fell 17.8% in the half while the network grew 20.7%. The mainland estate is 58% in third-tier cities and below, with more than 300 cities already covered. Management has deliberately slowed openings, which is an admission that the incremental shop now cannibalises an existing one.
The franchisee, not the company, is absorbing the shock. MIXUE takes no share of turnover, so it books revenue on the way in and does not see the shop's profit and loss. Franchisees do. Openings fell 29.3% in the half, which is what a franchisee base does when its own returns are thinning.
Margins are falling for a reason that does not reverse. Gross margin has declined for three consecutive reporting periods as the "real ingredients, fresh taste" upgrade raises input costs, and there is no evidence the company can pass the cost through in a market where four national chains are competing on price.
The cheap multiple is the correct multiple. A business whose profit just fell 14.7% and whose industry is in a subsidy-driven price war deserves a low multiple. The cash is real, but it sits onshore, earns under 3%, and requires a dividend — taxed on the way out — to reach an offshore holder.
19 — Risk matrixTwelve things that can go wrong
| Risk | Probability | Impact | How it would show up |
|---|---|---|---|
| Revenue per store keeps falling as the network saturates | High | High | Revenue growth stays below 5% while store growth stays above 15%; the gap becomes permanent |
| Gross margin compresses further | High | Medium | Gross margin below 30% for a full year; net margin below 14% |
| The price war intensifies and peers cut deeper | High | Medium | Discounting, delivery-subsidy matching, or a franchisee support programme |
| Franchisees stop opening new stores | Medium | High | Net additions below 2,000 in a half; closure rate above 3% |
| Surplus cash is deployed into value-destroying acquisitions | Medium | High | A large deal at a full price; goodwill rising from a low base of RMB 92 million |
| The delivery-subsidy effect proves structural, not cyclical | Medium | Medium | Per-store revenue fails to stabilise even after the comparison laps |
| Overseas operations keep consuming capital | Medium | Medium | The overseas network shrinks for a fifth consecutive period |
| The cost base grows faster than revenue again | Medium | Medium | Administrative expenses up more than 20% for a second half |
| Lucky Cup and the MIXUE brand cannibalise each other | Medium | Medium | Coffee SKUs expand into Lucky Cup's core price band; Lucky Cup store growth stalls |
| A market-wide de-rating of Chinese consumer equities | Medium | Medium | Multiple compression independent of earnings; fund-flow driven |
| Food-safety or product-quality incident | Low | High | A single high-profile event; the brand is the asset most exposed |
| Governance: founder control, short public record, dividend policy reversal | Low | Medium | The special dividend is not repeated; a related-party transaction appears |
20 — CatalystsWhat to watch, and when
The thesis is unresolved rather than wrong, so the next two reporting periods matter more than usual.
The FY2026 results — March 2027
MIXUE reports half-yearly, so the next clean read is the full-year announcement. Four numbers decide it: revenue growth, revenue per store, gross margin, and net store additions. If revenue per store stops falling, the cyclical explanation wins.
The subsidy comparison laps — first half of 2027
If the delivery-platform subsidies of the first half of 2025 are the whole explanation for the 17.8% decline, the comparable period in 2027 is where the arithmetic stops working against the company. Stabilisation there would be the single most important confirmation available.
Capital allocation — the FY2026 dividend and any buyback
The special dividend of RMB 2.65 per share, payable on or around 6 November 2026, is the first distribution since listing. Whether it is repeated, and whether a buyback follows, tells us whether the RMB 21.64 billion of financial assets is a floor for shareholders or a war chest for management.
Overseas — does the network stop shrinking?
The international estate has fallen for four consecutive periods. It does not need to grow; it needs to stop being a drag. A half with flat overseas store numbers would remove one of the two open questions in the model.
21 — ConclusionA fair price for an unresolved question
The reason MIXUE is cheap is not that the market has missed something. It is that the company's growth formula — open more shops, spread the supply chain over more volume, earn a margin on every ingredient sold — has stopped producing growth. Revenue grew 2.3% while the network grew 20.7%, so the average shop is selling 17.8% less than it did a year ago. Profit fell 14.7%, margins have declined for three consecutive periods, and openings have fallen 29% while closures rose. None of that is ambiguous.
What is ambiguous is whether it is a cyclical interruption or the beginning of a structural decline. The company argues the former, pointing to a delivery-subsidy base effect and a quality upgrade that costs margin before it earns any. The framework is sceptical on principle: it avoids industries with brutal price wars and demands stable margins before it buys. On the other side of the ledger, MIXUE has a genuine cost moat, an 18% return on equity even at the trough, RMB 21.64 billion of net financial assets and no debt, and an operating business priced at 7.8 times normalised earnings.
Our base case is HK$196, 17.2% above the price; the bear case, HK$150, is 10.4% below it. That is a favourable skew but a thin one, and the 14.6% discount to base value does not clear the 20% threshold the framework sets. We would rather be paid to wait than paid to guess.
Neutral · HK$196 fair value
A high-quality supply chain with a fortress balance sheet, at a discount that is real but smaller than the framework requires for a business whose margins are still falling. We would become constructive on two consecutive halves of stabilising revenue per store, or below HK$145.
22 — AppendixFinancial summary
| RMB millions unless stated | FY2022 | FY2023 | FY2024 | FY2025 | H1 2025 | H1 2026 |
|---|---|---|---|---|---|---|
| Revenue | 13,576 | 20,302 | 24,829 | 33,560 | 14,875 | 15,216 |
| Revenue growth, % | +31.2 | +49.6 | +22.3 | +35.2 | — | +2.3 |
| Gross profit | 3,924 | 6,152 | 8,059 | 10,452 | 4,706 | 4,630 |
| Gross margin, % | 28.9 | 30.3 | 32.5 | 31.1 | 31.6 | 30.4 |
| Profit attributable to shareholders | 2,000 | 3,190 | 4,437 | 5,887 | 2,693 | 2,296 |
| Net margin, % | 14.7 | 15.7 | 17.9 | 17.5 | 18.1 | 15.1 |
| Basic earnings per share, RMB | — | — | 12.32 | 15.65 | 7.23 | 6.05 |
| Operating cash flow | 2,400 | 3,800 | 6,009 | 6,039 | — | — |
| Return on equity, % | — | 34.7 | 34.5 | 30.2 | — | 17.9* |
| Cash, deposits and wealth-management products | — | — | 11,110 | 19,990 | — | 21,641 |
| Interest-bearing bank borrowings | — | — | — | 28 | — | — |
| Total equity | — | — | 15,061 | 24,769 | — | 27,063 |
| Stores at period end | — | 37,565 | 46,479 | 59,823 | 53,014 | 63,987 |
| of which outside the Chinese mainland | — | 4,331 | 4,895 | 4,467 | 4,733 | 4,378 |
| Franchisees | — | 16,784 | 20,976 | — | 23,404 | 29,775 |
* Annualised. FY2022 and FY2023 balance-sheet items are not disclosed and are shown as unavailable rather than estimated. FY2024 cash and equity are as restated in the FY2025 annual report. Market data as of 13 October 2026.
23 — SourcesPrimary and secondary references
| Reference | Date | Used for |
|---|---|---|
| MIXUE Group, interim results announcement for the six months ended 30 June 2026 (HKEXnews) | 27 Aug 2026 | H1 2026 revenue, gross profit, profit, expense lines, store counts, franchise movements, balance sheet, cash position, gearing, special dividend |
| MIXUE Group, 2025 annual results announcement and 2025 annual report | 24 Mar / 23 Apr 2026 | FY2025 revenue, segment revenue and margin, store counts, employees, expense ratios, cash and financial assets, ROE, management changes |
| MIXUE Group, 2024 annual results announcement | 26 Mar 2025 | FY2024 revenue, gross profit, segment detail, operating cash flow, ROE, store counts, franchisees |
| MIXUE Group, prospectus | 21 Feb 2025 | FY2022 and FY2023 revenue and profit, terminal retail sales, store counts, operating cash flow, pre-IPO shareholders |
| Guming Holdings, Auntea Jenny and Luckin Coffee interim results announcements | Aug–Sep 2026 | Peer revenue growth, store counts and same-store or per-store trends |
| Broker summaries and compilations on MIXUE following the interim results | Aug–Oct 2026 | Per-franchisee merchandise revenue decline; FY2026 profit estimate revisions; consensus range |
| Hong Kong and mainland press coverage of the Hong Kong store closures | Sep–Oct 2026 | Closure of the Mong Kok flagship; size of the remaining Hong Kong footprint |
| Company and press reports on new overseas markets | Sep 2026 | Mongolia and Uzbekistan openings; overseas country count |
| Market data providers | 13 Oct 2026 | Share price, market capitalisation, 52-week range, valuation multiples, dividend yield |
24 — DisclosureConflicts, positions and method
Positions and compensation. Farstar Capital holds no position in MIXUE Group (2097.HK) or in any company named in this report, and receives no compensation from MIXUE or from any party with an interest in the securities discussed. Coverage is selected by us. This report was not previewed by the company.
Method and rating. We apply a value-investing framework derived from Graham, Buffett and Munger: industry structure, the business and its moat, quantitative verification of the accounts, intrinsic value and margin of safety, risk screening, then a decision. Neutral means we do not expect a total return materially different from the market over twelve months; fair value is our base case rather than the probability-weighted value.
Limits of the evidence. MIXUE does not disclose franchisee profit and loss, same-store sales, or the profitability of Lucky Cup and FULU Fresh Beer separately. Revenue per store here is our own calculation from disclosed revenue and store counts using average balances: a proxy for same-store performance, not a substitute for it. The interim announcement carries no cash-flow statement, so the cash-flow analysis rests on the annual accounts, and the FY2026 earnings base used in the valuation is our estimate.
Not investment advice. This report is published for informational purposes only and is not an offer, a solicitation, or a recommendation to buy or sell any security. Figures are believed accurate as of the publication date but are not warranted, and past performance does not indicate future results. Readers should conduct their own diligence and consult a licensed adviser before acting.