Meitu Inc (01357): JPMorgan starts Meitu at Neutral as AI expands demand but erodes feature scarcity

Overview

This initiation report assesses whether Meitu can convert the expanding AI visual-creation market into durable revenue and shareholder value. JPMorgan sees a stronger earnings base, broad consumer distribution and credible AI-product optionality, but concludes that common visual features are becoming commoditized faster than evidence of incremental Productivity monetization is emerging.

Core views

JPMorgan frames Meitu’s opportunity through an “opportunity × scarcity × value capture” lens. AI reduces the time, skill and marginal cost required to generate, enhance, edit and create visual content, which enlarges both consumer usage and the professional workload addressable by software. Merchants, brands and creators can produce more product images, campaign variants, localized marketing, videos and social assets at lower cost. Agents could further multiply workloads by coordinating generation, editing, resizing, translation and video production across multiple outputs. The report therefore argues that Meitu’s opportunity depends increasingly on its share of underlying visual work, not only on direct user attention. The same AI transition weakens the scarcity of individual editing functions. Foundation models and general-purpose platforms can reproduce more image beautification, background replacement, generation and modification capabilities. JPMorgan highlights WeChat Xiaowei’s grey-tested in-chat image-processing function as an example of how previously specialized tools can migrate into a high-frequency platform. This can alter both competition and distribution: users may complete basic visual tasks within messaging, social, productivity or agent interfaces rather than opening a specialist application. Meitu’s historical advantages in consumer distribution, visual-product expertise and user-aesthetic knowledge remain relevant, but the report argues that differentiation must move from individual features toward taste, vertical workflow know-how and persistent user or commercial context. The report identifies Productivity as the principal route to incremental AI monetization. In 1H26, Productivity MAU reached 33mn, up 43.5% year on year; paid users rose about 30% to 2.4mn; and ARR reached about Rmb620mn. Productivity ARPU was about 1.5 times that of Leisure, while DesignKit had several thousand users spending at an annualized rate of roughly 10 times Meitu’s blended ARPPU. These indicators are positive, but JPMorgan does not see a material step-up in paid-user or ARR progression attributable to the latest AI product cycle. It expects Productivity, overseas monetization and greater AI usage to contribute an increasing share of growth through 2028E, but requires evidence that AI is creating a new earnings curve rather than extending the existing subscription trajectory. Meitu is testing several product and distribution architectures to capture higher-value work. MVLAND targets music-video creation, Picchi focuses on professional portrait retouching and creator-specific aesthetics, and DesignKit addresses e-commerce design and Agent Teams. AI credits could add usage-based spending above subscription fees, allowing revenue to rise with task volume and complexity even if seats or MAU grow more slowly. Meitu is also experimenting with Skills/MCP-style distribution through WorkBuddy and with WPS integration, allowing specialist capabilities to be called inside third-party workflows. JPMorgan considers this strategically sensible because AI distribution remains fluid, but notes the trade-off: a Meitu-owned application offers more control over customer acquisition, pricing, context and monetization, whereas third-party distribution may broaden workload access while transferring some economics and customer ownership to the platform controlling routing. JPMorgan believes durable scarcity will depend primarily on Meitu’s ability to deliver outputs that users find consistently useful, understand specific commercial workflows and retain reusable customer context. Taste and aesthetic understanding could support output acceptance, retention and willingness to pay; vertical expertise can address brand consistency, platform formats, campaign variants and deployment-ready outputs; and persistent context can incorporate identities, product and brand libraries, virtual characters, music and historical creative assets. Proprietary models offer greater control over quality, cost and iteration, but the report places more long-term importance on surrounding workflow knowledge and context than on model ownership. Evidence of this scarcity would include better output acceptance, paid-user retention, repeat use of stored assets, deeper workflows and sustained professional willingness to pay as generic models improve. The current financial base provides support for AI experimentation. In 1H26, revenue increased 22.1% year on year to Rmb2.2bn, gross profit rose 15.9% to Rmb1.6bn, and adjusted attributable net profit increased 39.5% to Rmb652mn. Sales and marketing expense increased 12.7% to Rmb330mn, below revenue growth, indicating operating leverage. However, gross margin declined 3.8 percentage points to 71.5% as AI-native products increased inference and cloud costs; compute and cloud costs were approximately Rmb130mn, with more than half directly associated with inference. JPMorgan therefore tracks the sequence from usage to revenue, gross profit and cash flow: AI-credit consumption measures demand, Productivity ARPU and professional spending indicate willingness to pay, incremental gross profit measures retained value after AI costs, and CFO and FCF provide the final financial cross-check. Liquidity gives Meitu room to fund the transition. At 1H26, the company had approximately Rmb4.8bn of cash and other liquid financial resources. The report sees limited refinancing risk from Alibaba’s US$250mn, 1% convertible bond, which has an initial HK$6.00 conversion price and matures in 2028, because principal is materially below current liquidity. The more important per-share issue is the trade-off between conversion-related dilution and cash redemption, alongside buybacks, dividends and stock-based compensation. Sustained CFO and FCF growth together with stronger Productivity monetization and AI gross profit would demonstrate that experimentation is improving financial quality. JPMorgan forecasts 2026-28E revenue growth at a 14% CAGR and adjusted attributable net profit of Rmb1.4bn in 2028E. Its Dec-27 HK$4.50 target price is based on 15x 2028E P/E, using 2028E because it should better capture contributions from Productivity, international business and AI usage as well as related inference costs and operating leverage. The 15x multiple is broadly in line with mature application peers and reflects continued earnings growth and a capital-light model, offset by lower visibility on the durability of AI-related competitive advantage. A rerating would require clearer evidence that distribution, proprietary workflows and user engagement are producing differentiated AI revenue growth and attractive economics.

Analysis framework

JPMorgan first assesses how lower AI intelligence costs expand visual-creation workloads while increasing competitive supply. It then evaluates Meitu’s subscription base, Productivity metrics, product experiments and distribution models to determine whether the company can retain a share of the new workload. The analysis tests financial conversion through user retention, ARPU, AI-credit usage, gross profit, CFO and FCF, before applying a 15x 2028E P/E multiple to derive the target price.

Methodology notes

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  AI-created visual workload × Meitu’s share of workflow × monetization × unit economics

  The report traces how lower AI costs increase upstream intelligence use and downstream visual-content production, then asks how much of that workload Meitu can capture and monetize.

• Competition & strategyEconomic Moat and Competitive Advantage

  Feature scarcity versus workflow scarcity

  JPMorgan evaluates whether Meitu’s taste, vertical workflows and persistent customer context can remain differentiated as individual AI editing features become broadly available.

• Valuation methodsP/E and PEG Valuation

  15x 2028E P/E valuation

  The target price applies a 15x earnings multiple to 2028E earnings, reflecting mature-growth characteristics and uncertainty over the durability of AI-related differentiation.

• Corporate Fundamentals and FinanceFree cash flow analysis

  Usage-to-revenue-to-gross-profit-to-cash-flow progression

  The report uses incremental gross profit, operating cash flow and free cash flow to test whether higher AI usage converts into retained economic value after inference, cloud and other costs.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Meitu Inc (01357.HK)

  Primary covered company; an AI visual-creation application provider whose earnings base and distribution support experimentation, while feature commoditization challenges long-term value capture.

  Strengths

  Large consumer distribution, established paying franchise, visual-product expertise, growing Productivity business, international growth and substantial liquidity.

  Weaknesses

  Limited evidence of a clear AI-driven Productivity monetization inflection; declining gross margin from inference and cloud costs; lower visibility on durable feature-level differentiation.

  Comparison

  The 15x 2028E P/E target multiple is broadly in line with mature application peers.

  Risks

  Faster commoditization, interface loss to larger platforms, weak AI revenue conversion or unit economics, and potential Alibaba convertible-bond dilution.

Key data

• Rating and target priceNeutral; HK$4.50 Dec-27 targetTarget is based on 15x 2028E P/E.
• 1H26 revenueRmb2.2bnUp 22.1% year on year.
• 1H26 adjusted attributable net profitRmb652mnUp 39.5% year on year.
• 1H26 gross margin71.5%Down 3.8 percentage points year on year as AI products increased inference and cloud costs.
• Productivity paid usersc.2.4mnUp about 30% year on year in 1H26.
• Productivity ARRc.Rmb620mnProductivity ARPU was around 1.5x Leisure.
• International revenue48% of group revenueUp 41.7% year on year in 1H26.
• 2026-28E revenue CAGR14%JPMorgan forecast.
• 2028E adjusted attributable net profitRmb1.4bnJPMorgan forecast.
• Liquidityc.Rmb4.8bnCash and other liquid financial resources at 1H26.

Impact & implications

The report sees Meitu as financially capable of pursuing several AI monetization routes, but its valuation will increasingly depend on whether it captures profitable, differentiated visual workflows rather than simply offering widely replicable AI features. Stronger Productivity ARR, professional spending, AI-credit economics, external workflow adoption, incremental gross profit and FCF would support a more constructive view.

Risks

• Visual creation capabilities could commoditize faster than Meitu develops durable workflow, taste and context-based differentiation, pressuring paid-user growth, retention and ARPPU.
• General-purpose platforms and AI assistants could capture the customer interface, reducing Meitu’s control over acquisition, pricing and user relationships even if it supplies underlying capabilities.
• AI engagement may not convert into sufficient revenue or incremental gross profit if inference, product-development and customer-acquisition costs rise faster than monetization.
• Alibaba’s US$250mn convertible bond could dilute shareholders if converted, while redemption would consume cash.
• Higher customer-acquisition costs and weaker subscription growth could weigh on results.

What to watch

• Acceleration in Productivity ARR and paid-user growth.
• Professional-user spending through subscriptions and AI credits.
• Recurring revenue development at MVLAND, Picchi and other AI-native products.
• Retention, repeat use of persistent customer assets and deeper commercial workflow adoption.
• External call volume, repeat adoption, pricing and revenue share from WorkBuddy, WPS and other third-party workflows.
• Whether monetized AI workload, Productivity revenue and incremental gross profit accelerate together.
• CFO and FCF growth as AI becomes a larger component of revenue.