Focus Media (Trans): Beyond one-off gains, growth hinges on AI and intl expansion

Dolphin Research Trans of Focus Media FY26 interim results call

I. Key financial takeaways

1) 1H26 headline metrics

Total: revenue of RMB 5.991 bn (-1.98% YoY); attributable net profit of RMB 3.128 bn (+17.39% YoY); adj. attributable net profit (ex-non-recurring) of RMB 2.291 bn (-7.06% YoY).By segment: building media revenue RMB 5.629 bn, broadly flat YoY; cinema media revenue RMB 345 mn (-26.29% YoY).Cash: operating cash inflow RMB 4.453 bn (+30.91% YoY), mainly on stronger ad collections vs. last year; as of Jun 30, cash, bank WMPs and large CDs totaled RMB 11.807 bn, up 41.77% vs. end-2025.

2) One-offs from the disposal of ShuHe Tech (fully completed in Q1)

1H one-offs totaled RMB 705 mn, including investment income of RMB 565 mn from capital reserve transfer and RMB 140 mn from deferred tax liabilities reversal.Excluding these, 1H26 attributable net profit was RMB 2.423 bn.In 1H25, ShuHe was still under equity-method accounting and contributed RMB 346 mn of investment income; excluding both sides, 1H same-base attributable net profit rose 4.47% YoY.

On Q2 alone, ShuHe contributed RMB 216 mn investment income under the equity method in Q2 2025; same-base attributable net profit was RMB 1.31 bn.There was no such contribution in Q2 2026, and same-base quarterly attributable net profit grew 1.91% YoY.

For full-year 2025, Focus Media booked an impairment of RMB 2.15 bn on ShuHe and recognized RMB 54 mn of investment income; excluding these, 2025 attributable net profit was RMB 5.044 bn, up 8.25% YoY on a same-base basis.

3) Dividend and shareholder return policy

Interim dividend was lowered YoY, which management said reflects cadence and funding considerations, not a change in shareholder return policy.In Dec 2023 the company announced a 2024–2026 three-year shareholder return plan, committing to cash dividends of no less than 80% of annual attributable adj. net profit each year; dividends were RMB 4.766 bn in 2024 and RMB 4.91 bn in 2025, both above the commitment.Dividend frequency has increased from twice to three times a year; differences among interim, Q3 and annual payments are timing only and do not affect full-year delivery.

4) Overseas: still in the investment phase, no break-even timetable

1H overseas revenue was approx. RMB 370 mn, +25% YoY (a 66% figure also appeared in the transcript, basis unclear), and the business remains loss-making and will likely continue to incur losses for the next few years.Management guidance on investment cycles: 5–7 years for a single new market in developed countries, and possibly 3–4 years to profitability in emerging markets.The company is willing and able to absorb this loss cycle.

II. Details from the call

2.1 Executive opening remarks

The interim briefing began with the VP of Finance recapping 1H performance only (financials summarized above), with no separate business strategy presentation; all business and strategy topics were addressed in Q&A.1) Attendance and information distributionAttendees: Chairman Jiang Nanchun, CFO and Board Secretary Kong Weiwei, VP Finance Wang Jingjing, CTO Ning Yijie.

The opening covered financials only; operating topics including AI, GEO, site layout, overseas and 'Tap-to-Buy' were addressed in Q&A.2) Business points from the openingThe company completed the disposal of its associate ShuHe Tech in Q1 2026, further refocusing on building media.

Building media revenue was broadly flat YoY and remains the absolute revenue mainstay; cinema media has shrunk to RMB 345 mn.Operating cash flow improved directly due to stronger ad collections vs. last year.

2.2 Q&A

Q: From 'Doubao', 'Qianwen', 'Yuanbao' earlier this year to recent AI office agents, each tech wave has been launched at scale on Focus Media. How sustainable is this trend?

A: Office-building audiences are China’s trendsetting cohort with stronger means, energy and knowledge, and are early adopters; downstream cohorts follow.In the internet era, new models such as Didi, Ele.me and Meituan all surged first via Focus Media.Office buildings are also socially contagious: once a few people adopt, gaps emerge and a follow-on consensus forms.

'Doubao' and 'Yuanbao' are more to-C, needing initial lift among trendsetters before broad consumer propagation; while products like BIDU and 'Xiao Huanxiong' are office agents that strongly fit office-building scenarios.Commuters in subways or on the road are not yet in work mode; the true moment is squeezing into the elevator in the morning with five meetings ahead, pressure peaking, or dragging oneself out at night still reminded of unfinished tasks.Only workers are most anxious in the elevator and thus most eager for an immediate helper — pain relief at the point of pain.

Airport traffic also includes business travelers, but an airport is a transit scene, not the place of highest office need, so office buildings fit AI office platforms better than any other scene.These office agents will keep iterating and will fragment further: today most are general agents, but specialized AI for HR, finance, sales and other functions will emerge, just as e-commerce evolved from Taobao/Tmall/JD to verticals like used cars.Generalists may hold most of the market, but many will build more specialized products — Alibaba just launched an AI e-commerce tool enabling one person to consolidate global e-commerce such as Amazon into a unified platform, effectively opening a one-person multinational.

Most of these are used individually today, but paid models will form, and spend will become enterprise-grade.Enterprise procurement requires consensus among users, purchasing, finance and departments, and office buildings are the easiest place to form company-wide consensus — like choosing Lark vs. DingTalk.In collaborative office software, whoever captures the time window and first-mindshare is hard to dislodge, as the switch is group-based, not individual; office-building ads will thus have durable impact.

Q: Stripping out one-offs from ShuHe’s disposal, what is the true profitability of the core business?

A: We disposed of the associate ShuHe in Q1, which impacted last year and this year’s 1H results.On a restored basis: in 2025 we booked RMB 2.15 bn impairment on ShuHe and RMB 54 mn of investment income; excluding both, 2025 attributable net profit was RMB 5.044 bn, +8.25% YoY on a same-base.In 1H26, one-offs from the disposal totaled RMB 705 mn (RMB 565 mn investment income from capital reserve transfer and RMB 140 mn deferred tax liabilities reversal); excluding these, 1H attributable net profit was RMB 2.423 bn.

In 1H25, ShuHe was still accounted for under the equity method and contributed RMB 346 mn; excluding that factor, 1H same-base attributable net profit grew 4.47% YoY.In Q2-only terms, in Q2 2025 ShuHe contributed RMB 216 mn; excluding that, attributable net profit was RMB 1.31 bn.In Q2 2026 there was no ShuHe contribution, and same-base quarterly attributable net profit rose 1.91% YoY, indicating resilient core ad operations.

Q: Does the lower interim dividend imply weaker shareholder returns? What is the forward approach?

A: The interim dividend was lower YoY mainly due to cadence/funding optimization and does not signal a change in shareholder return policy.In Dec 2023 we announced a 2024–2026 plan committing to cash dividends of ≥80% of annual attributable adj. net profit as a baseline; actual dividends were RMB 4.766 bn in 2024 and RMB 4.91 bn in 2025, both above the baseline.Frequency has increased to three payments per year, enhancing timeliness.

Interim, Q3 and annual are implementation points only; point-in-time variations are cadence and do not affect full-year delivery.1H operating cash flow was RMB 4.453 bn (+31% YoY), with solid earnings quality and capacity to honor return commitments; continuity and stability of dividends are supported.

Q: With weaker headline retail sales YoY and divergent structure (cosmetics, telecom devices, beverages/alcohol outperform; autos, home appliances, jewelry down), how do you view structural shifts in consumption? Which sub-tracks still have ad potential? Where is the recovery inflection?

A: Consumption has clearly migrated structurally, roughly into three types.Self-indulgent/social/instant gratification: categories like beauty help ease anxiety and are relatively stable, with a stronger health trend and ongoing market growth.Efficiency upgrades: smart devices and AI products that raise work/life efficiency grow faster under high work pressure.

Quality consumption with certainty: willing to pay up, but with justified premium.Autos and home appliances are big-ticket, low-frequency and driven by purchase cycles, trade-in subsidies and real estate linkage; jewelry is also affected by gold price expectations.Overall, consumers are still spending but with more cautious decisions.

Categories with low decision costs and strong emotional value (affordable yet comforting), and those with clear functional upgrades recover faster; big-ticket, low-frequency, deferrable, or pulled-forward categories recover slower.On ad potential, leading brands are more stable: under pressure, consumers prefer familiar and trusted brands, leading brands allocate more share-grab budgets, raising CR3/CR5 concentration.New products and new-tech products also win more easily because growth comes from 'new' — new categories, new functions/tech, new scenarios, new users, new channels — which typically require Focus Media to deliver social education and mindshare.

We are specifically constructive on health management, nutrition/supplements, AI products, and NEVs; within appliances, smart/AI-enabled categories should see an upgrade wave.Enterprise procurement, especially for high-tech upgrades, will drive larger spend; later on, the robotics industry will also emerge — we already see AI toys on our network.We aim to capture each wave, each tech iteration and functional innovation.

Consumption recovery still needs stronger Gov. policy support.As a medium reaching ~400 mn people, Focus Media is exploring ways to help; we are partnering with Alibaba, Alipay and others, and may roll out actions in Q4 to support consumption, injecting demand and driving partner sales — it cannot be branding only; brands can be counter-cyclical, but both short-term and long-term must be addressed.

Q: In a pressured consumption year, global leaders like L’Oréal are increasing ad spend — why? What does that mean for Focus Media?

A: Leading brands are often century-old companies that have experienced booms and plateaus across time and regions and understand cycles.They know they must maintain sufficient share-of-voice even in slowdowns: if SOV dips, brand awareness erodes within six months and share within a year; maintaining or amplifying spend makes share-grabs easier.Many Chinese clients are long-term growers who suddenly encounter volatility and, lacking cycle experience, cut budgets; Intl brands have weathered storms and are more stable, often counter-cyclical in spend to reinforce brand value, protect share and raise concentration.

Another feature is their emphasis on branding.Many Chinese companies grew on traffic dividends; as traffic peaks and costs rise, margins are squeezed, including for Intl brands.Since last year L’Oréal has rebalanced toward branding, reallocating between traffic and brand; from 2H last year through 1H this year, returns have been strong, with no short- or long-term negatives.

Kantar’s research shows roughly 70% of cycle-resilient brands’ demand comes from brand-led directed purchases and 30% from short-term promotions/conversion; if 70% of business depends on short-term promotions and traffic conversion, profits are usually elusive.Performance ads are 'fishing' — precise audience/time to catch now; Focus Media is more about 'raising fish' — educating more consumers to create demand.Otherwise the same users get recycled and CAC rises; good companies dare to educate and create incremental demand, revealing who are truly long-term brands.

Q: With consumer products a high share of revenue, how will the company navigate cycles over the next 5–10 years in a pressured consumption environment? What is the long-term growth logic?

A: Macro cycles exist, but the battle for consumer mindshare does not.In upcycles, brands fight for incremental demand; in downcycles, they fight for share.The more oversupplied and fragmented traffic becomes, the more consumers rely on familiar brands — true brands become standards, common sense and default choices; brands like Nongfu Spring can be truly cycle-resilient.

The top priority is to build certain mindshare advantages.Long-term growth rests on four pillars.First, traffic dividends have peaked and are rotating into brand dividends; Intl leaders like Coca-Cola and L’Oréal are increasing brand spend, with strong returns in recent years and solid China results in 1H.

Many Chinese SMEs grew via channel traffic, but traffic is now pricier, conversion harder and homogeneity more severe, pushing products into sameness and price compression.Traffic solves today’s sales; brand determines how long and how premium tomorrow’s sales can be.Winners will max both brand asset building and traffic/channel operations.

Second, we cover 400 mn mainstream users across daily life/work scenes, igniting the mainstream via high-frequency reach to build social consensus; consensus triggers herd behavior, and branding is fundamentally about creating trends.Third, the market is shifting from broad-based growth to structural growth: not all sectors/companies grow together, and opportunities migrate across tracks.New tech creates new categories, old categories get upgraded/remade, services/emotion consumption rises, Chinese brands premiumize and go global — all new opportunities.

Some sectors will be pressured, but we do not bet on any single sector; we serve the ongoing emergence of new platforms and demand — most new brands rising in the past five years have worked with Focus Media.As long as China’s commerce keeps innovating and brand competition continues, long-term demand is secure.Fourth, media coverage is becoming more data-driven and refined in operations.

Our moat is not only prime locations and high-value audience coverage but also daily must-pass, high-frequency, stable scenes — not occasional landmarks like the Bund — plus social consensus and trend formation at 400 mn scale.We already deliver precision, attribution, interactivity and optimization — the most advanced in global OOH — though still behind internet leaders in ultra-precision and interactivity.Strategically, we will integrate more with big platforms to build the future ecosystem.

On precision, we are doing 'a thousand buildings, a thousand faces' and have integrated data with Tmall, JD and Douyin, and are working on Tencent; once all four are integrated, precision and attribution will improve further.On attribution, after integrating with e-commerce we are now integrating payments; via WeChat Pay and other data, we can see differences between exposed and unexposed users in offline store purchases, enabling attribution not only for e-commerce but also offline.On interactivity, 'Tap-to-Buy' has engaged 20–30 mn users cumulatively, with stronger incentives planned for 2H.

Current cash red packets, lotteries and freebies are still random; in Q4 we will introduce certain benefits with the main thread 'tap to buy big brands on Focus Media' with assured discounts.Direct interactivity enables direct attribution, whereas prior data attribution was indirect.On optimization, 'Focus Smart-Ads' already lets clients see data and adjust daily; different ads can run by region and can be triggered intelligently — e.g., when temperatures exceed 35°C, pre-stored creatives for Coca-Cola auto-trigger by city, building type or climate, with dozens to hundreds of creatives.AI will further strengthen optimization.

The goal is to evolve from a scale coverage platform into brand infrastructure that is precise, optimizable, interactive and verifiable.As for market scope, we already cover 200+ cities domestically, while Chinese brands’ globalization is accelerating, and our proven neighborhood-media model is being replicated overseas.Overseas requires long-term cultivation to cross the inflection: Korea entered in 2017 and after eight years now performs very well with ~70k sites; Japan has only just 3k+ sites and typically needs 10–20k to improve materially.

Some countries remain loss-making because they have not crossed the inflection within 1–2 years; build times are longer in developed markets.Q: As consumers use AI assistants to search, compare and even decide purchases, how will AI reshape the ad chain and impact Focus Media’s moat long-term?A: AI will certainly reshape advertising, and the future boils down to two paths: become consumers’ first choice, or become AI’s priority recommendation.

AI agents will indeed search data and judge more efficiently, and agent-to-agent conversations and transactions may emerge, but AI changes the 'how to get there,' not the 'why choose you.'Leading brands must aim to be the default in consumers’ minds.At Alibaba’s launch, the CEO directly asked an agent to 'order 40 cups of Chagee, half-sugar, delivered to a specified place' — in his mind, ordering milk tea equals ordering Chagee, a reflexive choice; similarly, when asking for coffee, the agent knows one’s preference and may order Luckin.

AI will help search, compare, filter and transact fast, but the 'why choose' is human influence at its core.AI recommendations often present five brands; being absent is fatal, and among the five, consumers still pick the most familiar, as familiarity implies safety and certainty.Our own GEO observations show strong brands receive more trust and recommendations; Chinese brands can optimize to increase exposure, but probabilities still differ greatly vs. strong brands.

As AI capabilities rise, exposure is not solely a tech-optimization advantage; strong brands also work on GEO, and when everyone does, strong brands still get stronger recommendations.Algorithmic recommendations carry costs: 'Doubao' has started charging overseas at about $70 CPM, while Google/Meta ads are around $15–20.AI firms burn traffic and are not profitable, so bids will be higher; if you cannot be first choice and rely on recommendations, costs will not be cheaper than today’s traffic — likely higher.

AI will also change content supply: it can generate vast video/copy/images and flood social media, but tools are shared and lower the bar, increasing information noise.By contrast, offline life scenes in apartments and office buildings — and their media — are scarce.As information atomizes and content stacks up, human time does not increase; algorithms excel at personalization, while brand ignition requires 'hundreds of millions in sync' — Focus Media’s value in the AI era is helping brands emerge from the noise and settle into social mindshare.

AI is not an adversary but an amplifier of operating efficiency, and it will improve our own efficiency: helping clients refine product selling points and consumer reasons-to-buy to improve accuracy and creative iteration (previously a four-week cycle, hardly optimizable); enabling intelligent, data-backed delivery to optimize quickly; strengthening content review and ops, especially for SMEs; and enhancing sales service and client insights, including effect analysis, attribution and front-end delivery adjustments.AI will raise production, analysis and matching efficiency; Focus Media will provide real-scene scale ignition; brands must build long-term trust and differentiated value.AI will eliminate mediocre, homogeneous ad content, but not ads; it will compress information gaps, but not erase brand equity.

The future scarcity is not content volume, but making brands the consumer’s priority — especially first choice — where margins are higher; Nongfu Spring is a reflexive first choice.Focus Media’s long-term moat rests on the inescapability of life/physical space — unavoidable real scenes; Shanghai has only ~5,000 office towers and ~200,000 communities, scarce and non-replicable resources.The more fragmented the online information, the scarcer offline effective space and time become.

Q: This year you made a strategic investment to control Taimei Times in the GEO track. What is the strategic logic? How will GEO and elevator media synergize?

A: In the AI era only two things matter: become consumers’ first choice or AI’s priority recommendation.The former relies mainly on Focus Media — our high-frequency, certain reach builds social consensus, creating reflexive associations such as 'cheese equals Milkground,' '0 sugar/0 fat/0 calories equals Genki Forest,' 'wet wipes equals DEEYAO.'Becoming the first choice among a cohort is the lowest-cost approach; if a brand is not the absolute first choice but one of several, it must do GEO well.

FMCG more easily forms reflexes, while complex-decision categories like mother/baby or 3C often trigger search after ad exposure.Previously consumers searched on Xiaohongshu or Douyin for reviews; now they ask 'Doubao' after seeing Focus Media.We have clients whose ads generated significant search volume but poor conversion because consumers asked Doubao and the client neglected this step — the ad was intercepted at the final closure.

Focus Media commands a large share of many clients’ offline media budgets, so we must take responsibility for end-to-end outcomes.We aim to ensure that when consumers search in Doubao after our ads, the brand is recommended with high probability and its core values are recognized by AI’s semantic structures.If clients spend heavily with us, they need more certain outcomes.

This forms a complete chain: Focus Media triggers active search, GEO optimization enables priority recommendation, recommendation then clicks through to platforms like Douyin/Tmall for purchase — Doubao is backed by Douyin, Qianwen by Alibaba/Tmall, and shopping will increasingly flow through this chain.Focus Media is the outpost that triggers search and attention, then ensures priority recommendation and click-through, improving outcome synergy.Search volume itself is an effect indicator: whether search spikes post-campaign reflects delivery; with strong GEO and priority recommendation, client traffic rises further.

Q: What is the progress and timetable for acquiring NewMedia (Xinchao)?

A: We are proceeding in line with disclosed announcements and the relevant regulators’ requirements; all information requiring disclosure has been announced.Management is pushing ahead, and we will disclose promptly once thresholds are met.

Q: What is the company’s progress in AI marketing models? How are you empowering clients with AI? Any new cases?

A: AI must land against core business and real scenarios.We self-developed an AI chain platform that is fully embedded in core operations, becoming an indispensable productivity tool for sales/office staff; based on it we incubated vertical apps such as AI monitoring, AI ad video and AI PPT serving the whole ops team.Client-side deployment follows two paths.

First, we can replicate Focus Media’s proprietary marketing methodology and data capabilities; via AI-customized agents we generate professional brand strategies and, based on business inputs, refine product selling points and consumer reasons-to-buy, significantly upgrading our client service.We also share these capabilities with close partners; one major consumer brand reportedly adopted AI’s brand diagnosis and strategic recommendations to adjust and optimize its group-level brand strategy.Second, AI creative: our ad carrier is ultimately a video, and our video-generation capability based on strong model foundations is improving through iterative feedback loops, efficiently producing effective creatives for Focus Media.

From market feedback, adoption of Focus AI creatives is rising.Our creative center has deepened cooperation with some advertisers, embedding Focus AI production directly into clients’ full marketing chain — used not only in Focus but also in clients’ own channels.'Focus Smart-Ads' is a new AI-native system with growing experience among e-commerce-heavy clients, and will do more with 'Tap-to-Buy' going forward.

Internally we have also deployed AI-driven new models like the digital employee 'Wanshitong,' running for about four months.It connects multiple business systems as a connector, monitoring screen scores/status and vendors’ IoT SIM performance.Previously such work required field teams coordinating with vendors over days; now the digital employee automates the flow, shifting status monitoring from days to minutes and vastly improving information throughput.

Our AI focus ahead remains twofold: concentrate resources on core internal business systems and raise company-wide AI productivity via office automation, amplifying Focus Media’s product capabilities to deliver higher-quality outcomes for clients and consumers.

Q: What is the plan for site layout this year and beyond?

A: On screens, building TVs and smart screens will increase, especially in office towers; frame posters will be reduced — frame locations often overlap smart screens and are non-digital, based on 20-year-old product design, so we will streamline, lowering frame costs.Overall, screen count up, frames down.

On interactivity, last year we enabled 900k–1 mn interactive sites via cooperation with Alipay; this year we will add another 900k–1 mn, so around 80–90% of sites will have interactivity.Q: How did overseas expansion go in 1H? What is the go-forward strategy? How long will the investment phase last, and is there a break-even timetable?A: Overseas sites reached 196k in 1H, with revenue around RMB 370 mn, +25% YoY (a 66% figure also appeared; basis unclear), and the business remains loss-making, likely for the next few years.

Strategically, Chinese companies going global is a given, evolving from 'going out' to 'going in,' led by brand power — from selling goods to local production and product development, ideally building premium local brands.Chinese autos are already positioned as super-premium in some markets, selling for RMB 1 mn+, showing strong upward momentum.Focus Media positions itself as infrastructure for Chinese brands going global.

In each country the playbook mirrors China: lock core cities and cohorts, achieve one-off coverage, and reach 50–70% of local consumption power.Each new market typically requires 5–7 years; developed markets take longer due to higher labor costs and slower development cadence; emerging markets can reach profitability in 3–4 years.It resembles insurance: you need enough sites and users before sales scale.

Below ~10k sites there is little to discuss; the turning point is at 20–30k.Korea entered in 2017 and after 7–8 years now performs very well with ~70k sites; Singapore has done well after 6–7 years; Japan has only 3k+ sites.In Dubai one company spent 14 years to reach 1,100 sites, while Focus has already reached 5,000.

Competition-wise, aside from Vietnam there are essentially no absolute competitors — markets are small and unlike China’s intense startup/investment environment, local startups struggle; once scale forms, moats become strong.We are loss-making in many countries but already in leadership positions.Overseas is a long-term strategy, and precisely because it is hard and long-cycle, only true believers will persist; we are willing and able to absorb the loss cycle.

Chinese NEV leaders are clearly going global, but consumer brands are fewer — Genki Forest and Qiaqia have some sales and ads; in SE Asia we see only 5–10 familiar Chinese brands, with minimal advertising during sales.Appliance giants like Midea, Gree and Hisense went earlier.Over the coming years, as coverage improves, a full-scale branding wave for Chinese brands overseas will arrive.

As with client 'Oriental Leaf,' scaling from zero to RMB 1 bn took ten years, but RMB 1 bn to RMB 20 bn took only four; an inflection will come, so we will keep investing through the loss phase.

Q: Any representative cases for 'Tap-to-Buy'?

A: 'Tap-to-Buy' is Focus Media’s core infrastructure to close the loop from offline reach to online conversion.Ops metrics are developing steadily and healthily, and upgrades to product standardization, onboarding speed and user experience shared last quarter have gone live.In Q2–Q3 seasonal marketing, we focused on water and tea categories.

For water, we rolled out standardized full-funnel 'drive-to-store' strategies, including low-price new-customer offers and coupon tools for repeat purchases; 'Sanyuan' and 'Suntory' each achieved 1 mn new users and 200k repeats within flight, with strong outcomes.For tea, we partnered with 'Yeeherb' on summer campaigns, driving 1 mn verified orders via 'Tap-to-Buy'; for Yeeherb’s new product launch, we also collaborated with suppliers for channel reactivation and promotions, with sales far above plan.We also tested some collaborations during the World Cup, all based on standardized product capabilities.

We will invest more in 2H: together with strategic partner Alipay we will upgrade the product end-to-end, both expanding interactive coverage and enriching interactive formats.Around the main thread 'buy big brands on Focus, enjoy unique offers,' we will target high-conversion users/scenes/verticals via refined ops, offering more certain rewards to build habitual interactions, with Focus providing support on investments.We aim for this major upgrade to help revive consumption, create value for platforms and users, and unlock greater commercial benefits.

<End of text>

Risk disclosures and statements for this article:Dolphin Research Disclaimer and General Disclosure