Mixue Group (Trans): Per-store sales down double digits; entering a 3-yr investment phase

Dolphin Research Trans of Mixue Group's FY26 Interim (1H26) earnings call:

I. Key takeaways

1. Shareholder returns: first post-IPO dividend of approx. RMB 1 bn.

a. The payout will come from cumulative retained earnings as of 30 Jun 2026, and management stressed this is a special dividend with no commitment to future distributions.b. Future capital returns, including dividends and buybacks, will be considered flexibly based on ops, cash reserves, growth plans, and internal investment needs. Cash reserves were RMB 21.64 bn as of 30 Jun.

2. Margin guide: interim pressure persists in 2H with no timing for an inflection.

a. As core ingredient upgrades roll out, fresh milk/beans coverage expands, and in-store deliveries scale, supply-chain upgrades will further weigh on GPM in 2H. The company also reiterated 'quality up, no price hikes.'b. Management will reassess the timing of margin recovery based on the pace of raw material upgrades and supply-chain cost-down progress, and will update the market once visibility improves.

3. 1H headline financials

a. Topline: revenue RMB 15.2 bn (+2.3% YoY); GPM 30.4% (down 120 bps YoY); net profit RMB 2.32 bn (-14.7% YoY), NPM 15.2% (vs. 18.3% a year ago).b. Opex and cash: selling & distribution expenses RMB 1.12 bn (+22.9% YoY; 7.4% of revenue); admin expenses RMB 610 mn (+39.4% YoY; 4% of revenue); R&D RMB 74 mn; net op. cash inflow RMB 2.05 bn.c. Mix and one-offs: merchandise & equipment sales contributed 97.3% of revenue; franchise & related fees 2.7%. FX losses widened on HKD holdings.

4. Three drivers of GPM contraction

a. 'True-fresh' upgrade lifted ingredient and logistics costs: short-shelf-life coffee beans, fresh milk, and frozen blended juices replaced ambient inputs, raising procurement costs while supply prices to franchisees stayed flat.b. A higher mix of cold-chain inputs in overall sales structurally diluted GPM.c. Strategic price hold in coffee: despite higher bean costs and recipe upgrades, prices stayed unchanged, with ~RMB 200 mn in benefits passed to stores in 1H.

II. Earnings call details

2.1 Management remarks

1. Strategy and investment cycle

a. A new investment cycle began in 1H centered on 'safety and quality' across supply chain, store ops, and brand IP, requiring time for capability build and subsequent performance uplift.b. External context: last year's delivery wars inflated AOV/store and shifted orders online, squeezing margins. Category and brand lines between tea and coffee are blurring, intensifying competition.c. Three-year goals: fortify cold-chain to enable 'true-fresh' upgrades domestically and overseas; enhance consumer experience via store image refresh and digitalization; build global brand influence around the 'Snow King' IP.d. 2H outlook: high comps remain and competition may intensify. Management expects continued, potentially greater operational pressure.

2. Product and brand matrix

a. Mixue Bingcheng refreshed classic SKUs and upgraded traditional 'fresh-brew' coffee to on-site grinding; over 3,000 stores had fully automated coffee machines by 30 Jun. New SKUs include melon-series ice cream.b. Lucky Coffee advanced 'three fresh, one on-site' (fresh beans, milk, fruit, made-to-order), switching to 60-day beans and low-temp fresh milk nationwide, and launched fruit-coffee such as iced coconut and bayberry.c. Three-brand strategy: Mixue Bingcheng opens steadily, fills lower-tier gaps, and upgrades store image; Lucky Coffee optimizes channels and completes county-level coverage; Xianpi Fuli Jia leverages the existing supply chain to build a national network.

3. Store network and overseas

a. As of 30 Jun, ~64,000 stores across 17 countries, with nearly 60,000 in China. Mixue Bingcheng flagship stores now cover 26 domestic cities.b. Overseas, localized warehousing and distribution networks were established in nine countries, and a full cold-chain logistics upgrade began in SE Asia in 1H.c. Indonesia and Vietnam are restructuring legacy stores; store counts fell in 1H, but management said quality improved notably. New entries included Mexico, Kyrgyzstan, and Brazil.

4. Brand IP and digitalization

a. 'Snow King' IP: first manga series launched in Jun and animation season 3 aired recently. Jan–Jul Snow King merchandise GTV rose ~200% YoY, with rollouts across SE Asia and the Americas.b. Digitalization, labeled a 'relative shortcoming' by management, is a key focus area. The mini-program and membership system directly connect to consumers, informing product innovation and ops decisions.c. By end-1H, smart meal-prep machines covered 18,000+ Mixue Bingcheng stores to improve efficiency, standardization, and food-safety risk control.

2.2 Q&A

Q: Is the deliberate slowdown in domestic store openings a short-term adjustment or a long-term direction? How do you view long-term growth?

A: It is a long-term, deliberate strategic choice, not a temporary adjustment. With ~64,000 stores, any oversight is magnified by scale, so the development logic must shift to 'progress with stability.'This year, investment intensifies around 'three safeties' (food, production, information) and 'three qualities' (product, experience, content) to avoid pendulum swings of 'three steps forward, one back.' Such volatility appeared in SE Asia over the past 2–3 years and domestically since 2025.Growth spans three dimensions: deepen China with limited store growth but room to lift sales/store; expand coverage in blank trade areas, lower-tier townships, transport hubs, tourist spots, industrial parks, schools, and flagships. Deepen SE Asia (700 mn population, young, suitable climate) while steadily penetrating North/South America and Central Asia. The Snow King IP aims to be a key long-term growth driver. Management will allocate ample, even full, resources to globalization, new brand incubation, and digital plus AI transformation.

Q: Did average sales/store drop by double digits in 1H at the group level? What about the main domestic brand, and why? What's the 2H trend?

A: Confirmed: both group-level and Mixue Bingcheng's domestic avg. sales/store declined by double digits, with greater pressure in 2H.Management noted that reverse-calculating same-store from GMV/revenue per store is distorted by two factors: the absolute number of new openings, and brand mix differences. Lucky Coffee is accelerating openings and Xianpi Fuli Jia was consolidated only last year, and these early-stage brands form a larger share of net adds, depressing the reverse-calculated average.Two drivers for Mixue Bingcheng's domestic decline: base effects, as last year's delivery wars lifted store sales to unprecedented highs; and fiercer competition as tea and coffee boundaries blur. Promotional subsidies supporting new categories softened traditional price bands such as '≤RMB 10' and 'RMB 10–20.'

The same pressures persist in 2H: last year's Q3 was peak delivery-war season, i.e., the highest comp. This year's 'true-fresh' upgrade and digital initiatives require sustained investment before showing through, likely not fully reflected by year-end. Management remains cautious on a short-term rebound.

Q: As tea-coffee convergence accelerates, what is the company's differentiated edge?

A: Follow the trend, fix weaknesses, and hold the line on 'quality up, no price hikes'; see convergence as an opportunity.The company has long held dual tea and coffee DNA: it first sold coffee in winter 2008 in Zhengzhou, moving from instant to in-store fresh-brew, and recently to on-site grinding via fully automated machines. In 2017, it launched the standalone fresh-brew brand Lucky Coffee, and is upgrading ingredients and store image while innovating specialty recipes, bean origins, and brewing methods.Differentiation centers on higher quality without price increases. Costs from quality upgrades and strategic investments are not passed to consumers but absorbed through supply-chain efficiency, scale, and digital capabilities. Over 30 years, this has been crucial to success: more lemon in lemonade (from 30g to 40–45g) and better milk powder for ice cream, while prices stayed affordable. Consumer choices increase under convergence, but value-for-money remains paramount.

Q: What are the milestones and plans in the supply-chain upgrade? How long will this cycle take?

A: 1H focused on cold-chain at the ingredient level, with a roughly three-year cycle. Fresh milk and short-shelf-life beans are rolling out across Mixue Bingcheng and Lucky Coffee, with core inputs for classics like Peach Oolong, Passion Fruit Full Cup, and Jasmine Green switched.Three focus areas: shift ambient inputs to frozen/fresh-chilled; ambient pearls and coconut milk are now frozen, fresh milk has a 19-day shelf life, coffee beans shortened from 12 months to 2 months, and jasmine tea is cold-chained end-to-end. Build compound blends on frozen inputs (e.g., golden passion fruit). Shift procurement/production from large batches to small-batch, high-frequency, fast turns.Logistics upgrades span 30+ parks and tri-temp vehicles; delivery moved to 'to-store' to reduce store burden and lower minimum order from ~RMB 9,500 to RMB 6,000 per shipment. The three-year horizon reflects the scope: the prior ambient-to-frozen shift (2020–2023) covered ~10k domestic stores; this round covers ~60k, plus new short-shelf inputs, equipment, and store utility retrofits. The upgrade does not alter the high-quality, affordable positioning.

Q: Why expand self-owned capacity when consumer demand shifts quickly?

A: In-house production underpins supply-chain stability and consistency, not merely cost, combined with external sourcing.China now has six production bases. In-house capacity serves as a backstop when external suppliers face capacity swings or risks, ensuring no stockouts and stable supply to stores, while enabling end-to-end control from farms to production lines.The same logic applies to self-operated stores and fleets: owning fleets revealed where cold-chain breaks occur and where to standardize. Fresh milk remains externally sourced; a joint venture dairy farm with Junlebao is in place, and upstream raw milk processing may be explored. Years of factory know-how feeds front-end innovation, informing choices like strawberry puree vs. chunks, or golden vs. purple passion fruit. This round focuses on retrofitting existing lines, with new investment mainly in HPP and fresh milk lines, keeping capex under control.

Q: Can you break down the domestic avg. sales/store decline further? After delivery subsidies faded, did orders return to stores?

A: Confirmed: orders fell as subsidies faded and store sales remained under YoY pressure, with no specific read on order reversion.Two external shocks weighed on averages in 1H: post-delivery-war normalization as last year's platform subsidies inflated delivery orders, then receded; and a more fluid competitive landscape with blurred category and price bands, intensifying price competition and diverting traffic.Beyond externalities, the core is timing: upgrades need time to land. The 'true-fresh' strategy and on-site grinding were finalized in late-2025 to early-2026; ingredient upgrades were in pilot in 1H and only scaled in 2H, and coffee machine rollout began in Jun. Many initiatives remain in planning/testing and have yet to show up in current results. Consumers have not yet felt tangible product changes, which is a timing gap, not a directional issue.

Q: How are the pilots on ingredient upgrades and on-site grinding performing? How far along is the rollout?

A: Fresh milk and short-shelf beans now cover 6,000+ stores; on-site grinding targets ~14,000 stores this year.Fresh milk and short-shelf beans aim for nationwide coverage by 2028. A flagship upgraded classic has become a benchmark; with nationwide rollout, its Jun monthly sales ranked third across all products, behind lemonade and ice cream.Other inputs like passion fruit and jasmine green have been advancing since Jul, with broader impact expected in 2H and next year as capacity and cold-chain build-out catch up. On-site grinding focused on testing in 1H, now covering 6,000+ stores, with full coverage by 2028. Stores with machines outperform the average, with a notable pickup in morning cups. 'Mixue Coffee' is positioned as 'daily coffee' to drive higher-frequency, low-burden consumption. Marketing spend of ~RMB 100 mn is planned this year, with provenance storytelling around 'true-fresh' and education on morning-use occasions.

Q: Beyond product upgrades, what will drive store profitability in 2H?

A: Four tracks in parallel: store image upgrades, relocations, flagships, and digital.Store image: a refreshed Gen-9 format launched with only ~10% higher capex; >2,000 upgrades are planned in 2026, layered by location to balance core-area experience and lower-tier cost efficiency to avoid heavy franchisee burdens. Relocations: underperformers by daily sales over the past six months will be relocated if issues are site-driven, with subsidies provided. Over 2,000 relocations were completed in 2025, lifting sales ~27% on average.Flagships: deeply localized designs blending local history, figures, cuisine, and intangible heritage into destination stores with exclusive SKUs and merchandise; the first Tier-1 city flagship opens in Beijing in 2H. Online: deepen digital ops with more precise, time-slot/scene-based marketing; members get mini-program exclusive pricing to boost engagement and repeat.

Q: When will avg. sales/store see a meaningful improvement? Any timeline to return to growth?

A: No specific timeline; significant pressure remains in Q3 and likely 2H.'Coarse' variables like delivery-subsidy fades weigh quickly, while 'slow' variables such as ingredient upgrades, on-site grinding, brand conditioning, and experience take time. The next 2–3 years focus on product quality and experience. Once upgrades are complete, consumers will enjoy cold-chain-fresh beverages, improved store environments, and more convenient digital services, and growth should follow naturally.

Q: How is the SE Asia recalibration progressing? When might we see a positive inflection in same-store or openings?

A: Store quality, experience, and products have improved, but no timing for an inflection.Store quality follows 'open good stores, close weak ones, lift legacy stores, prevent relapses': higher admission standards improved the avg. sales of new stores vs. legacy. Legacy performance improved via activation programs, training, and relocations, while strengthening client pipelines and site reserves. New formats such as mall, island, and service-area stores are under exploration.Consumer-facing efforts focus on localized store image and lower renovation costs. Given the high share of riders in SE Asia, signage was redesigned to boost visibility at 100m/50m/10m. New renovation schemes in Indonesia and Thailand cut upfront capex. Marketing follows a 'one country, one plan' calendar. Digital efforts aim to lift delivery and member channels, negotiating lower fees and subsidies with local platforms to improve franchisee profitability.Products include importing proven domestic hits and extending their sales windows. Cold chain: Malaysia (except East Malaysia) is largely upgraded, with Vietnam and others in progress; from late Apr, taro balls and grapes rolled out in both markets. Desserts are being piloted to raise per-store productivity.

Q: What is the overseas opening pace and scale plan?

A: No blind expansion; a more proactive opening pace may resume from 2027.Store count is not the primary goal; quality comes first, meaning better avg. sales and shorter payback. Two reasons: ensure sustained franchisee returns, and recognize that SE Asia's made-to-order beverage market is still early-stage vs. China. Cultural, consumption, infrastructure, and operating conditions vary widely by country and even within countries, requiring sub-regional strategies.By market: Vietnam and Indonesia grew faster previously and now face uneven store quality, requiring time to digest stock issues. Thailand and the Philippines entered later with stores concentrated in capitals, and will prioritize new regions and sites. Post-2027, focus will be on gaps in core-city trade areas, satellite cities around cores, outer islands in island nations, and provincial regions in peninsular countries, while testing mall/service-area/gas-station/industrial-park formats and expanding into Central Asia and the Americas.

Q: With lower GPM and NPM in 1H, beyond same-store declines, what else drove it? How long to an inflection?

A: No timing given; interim GPM and NPM will face phase pressure in 2H.NPM fell ~300 bps YoY mainly because the company absorbed costs from quality upgrades and strategic investments, given franchisee profits were squeezed post-delivery wars. GPM contraction of 120 bps reflects drivers noted above, including logistics upgrades for temperature control on short-shelf inputs, vehicle/warehouse retrofits, and in-store delivery increasing time and labor at the last mile.NPM was also hit by spend on brand IP, digitalization and creative talent, overseas team expansion, endorsements lifting promo expenses, and wider FX losses on HKD holdings. As precedent, during the 2022 ambient-to-frozen pivot, costs rose and the company proactively cut prices on 69 core inputs amid COVID, compressing profits, which later recovered via internal efficiency without price hikes.

Q: How should we view the 30% GPM and 15% NPM baselines?

A: These are long-term averages distilled over 30 years, not rigid rules.Ops are generally healthy around these baselines, but they are not immutable. In 2022, GPM dipped to the low 28s, while in 2024–2025 it was above 30%. Business is cyclical and faces external uncertainties, so the company will invest or harvest as needed rather than target a straight-line trajectory. Management rejected metaphors of 'ordeal and ascension,' saying the company is an ordinary enterprise solving problems through cycles.

Q: How do you assess Lucky Coffee's 1H product upgrades and endorsement marketing? What is the actual impact on store performance?

A: Sales/store declined YoY; after switching to short-shelf beans, Americano sales rose ~20% WoW in week one.Lucky Coffee initiated structural optimizations in 1H, with benefits to store performance releasing progressively. Product cadence met expectations: nationwide switch to 60-day beans was completed in Jul; by end-1H, ~50% of stores adopted low-temp fresh milk. Fresh fruit moved from jams to HPP juices and frozen fruit; the HPP frozen watermelon series topped 1 mn cups in its first week in Jul.On brand, dual-endorser marketing broadened audience reach, reframing Lucky Coffee from 'value coffee' to 'high-quality, value national coffee.' A 3D chibi 'Old K' character delivered multiple content pieces topping 10 mn views. Channels: differentiated start fees and product mixes improved third-party delivery economics; private-domain membership and precision marketing lifted mini-program DAUs nearly 3x vs. early-year. New hero series sold 57 cups/store/day on avg., with cumulative sales topping 1 mn cups. The '4.5' new image landed in 100+ stores, driving double-digit growth.

Q: What new actions in 2H will lift ecosystem engagement and store profitability?

A: Quality messaging, new SKUs and co-brands, tiered membership, flagships, and site subsidies.Quality communications under the 'fresher Lucky Coffee' theme will visualize supply-chain provenance and processing. A new coffee bean tailored for fresh fruit coffee and more new SKUs and annual co-brands will launch, using IP momentum to seed demand.Membership will add weekly/monthly passes with tiered benefits and precise pushes via communities to increase stickiness and frequency. Offline, early/late-day product bundles (morning coffee, noon tea, afternoon ice) will activate all time slots. Flagships will pilot in prime Tier-1/2 locations; two opened this year with tasting zones, and 16 specialty drinks tested nationwide have gone live on flagship menus.Equipment subsidies and other policies will lower entry barriers for prime sites, steering top franchisees into core city trade areas. Cities like Shijiazhuang, Chengdu, and Beijing each surpassed 100 stores.

<End of text>

Risk disclosure and statement:Dolphin Research Disclaimer and General Disclosure