On March 27, Beijing time, after the market close in Hong Kong and before the market open in the U.S., Meitu Company released its financial report for FY2025. The performance in 2H25 was lackluster, with revenue falling short of expectations and gross profit significantly missing the mark. However, expense control exceeded expectations, and operating profit was largely in line. Nevertheless, the user ecosystem, which constitutes subscription revenue, is under significant pressure, leading to an overall negative impression from Dolphin Research.

During the earnings call, management addressed the notion of 'AI devours software,' emphasizing that models serve as infrastructure while applications are the value delivery layer, with the two working in synergy rather than replacing each other. Meitu, as a vertical imaging application, stands to benefit from advancements in models. Management also outlined a strategy to move toward AI Agents and discussed the prospects for B-end productivity business and international expansion.

Key takeaways:

1. User Ecosystem:

a) MAU: The earnings call explained the decline in MAU in the second half of the year by attributing it to the maturation of domestic lifestyle products and slowing growth. However, Dolphin Research is more concerned about the MAU for productivity scenarios, which stood at only 24 million and received no response. Although B-end is still in the penetration phase in terms of user structure, growth significantly slowed from 36.83% in 2023 to 8.17% in 2025.

b) Subscriber Count and Subscription Penetration:

- In terms of growth volume, the number of subscribers increased by 2.8 million in the first half of the year but only by 1.5 million in the second half. Considering the overall decline in MAU, the in-line subscription penetration rate is partly due to the shrinking denominator, suggesting that actual conversion quality may be overstated.

- From a structural perspective, subscription penetration in lifestyle scenarios reached 5.9%, with growth remaining stable in the second half compared to the first half. In contrast, productivity scenarios saw higher annual growth in subscription penetration, but the growth rate slowed from 2% in the first half to 1.2% in the second half.

Given the company's guidance for a surge in productivity tools in the second half of 2026, there are doubts about this projection at the current juncture.

c) ARPPU: Implicit ARPPU exceeded 200 yuan, up 4.21% year-on-year, reflecting the boost from the company's globalization strategy and shift toward B-end vertical scenarios.

2. Expense Control:

- Gross profit growth slowed to 20.8%, with the gross margin remaining flat compared to the first half. The significant miss in gross profit was primarily due to two factors: a decline in the proportion of high-margin advertising business and the company's strategy of 'amortizing computing costs through subscription plans' (offering users access to advanced models at lower subscription prices multiple times), which directly led to a substantial increase in computing and API-related costs.

Despite the significant miss in gross profit, core operating profit reached 466 million yuan, up 82.6% year-on-year, largely in line with expectations. This was primarily due to Meitu's effective control of selling, general, and administrative expenses, which exceeded expectations across the board.

- From a sales expense perspective, Dolphin Research initially assumed that overseas expansion would require increased sales expenses, but the proportion of sales expenses to subscription revenue remained at 16% in the second half.

- Regarding R&D spending, the company's Model Container strategy controlled expenses related to foundational model training, resulting in a mere 3.8% increase in annual R&D spending.

Dolphin Research believes that, given the current poor performance of the user ecosystem, especially the underperformance of overseas expansion, the company may still resort to increased marketing efforts to acquire customers. Additionally, it should be noted that the savings in R&D spending come at the expense of gross profit.

3. Cash Flow and Shareholder Returns: As of FY25, Meitu had approximately 4.4 billion yuan in net cash (cash + short-term investments - interest-bearing debt), indicating robust cash flow. Regarding shareholder returns, excluding the special dividend from the one-time revenue of liquidating cryptocurrency holdings, if the company maintains a 40% cash dividend payout ratio and includes the management's disclosed 300 million HKD share buyback plan, the total shareholder returns would amount to approximately 650 million yuan. Relative to the current market capitalization of 19 billion yuan, the potential yield of 3.5% is not particularly high. Moreover, it should be noted that share buybacks may not be sustainable, and excluding them, the potential return is only 2%, providing limited support.

4. Overview of Financial Data

Dolphin Research's Viewpoint

Let's first review Meitu's diverse product matrix. For an in-depth analysis, refer to: 'Meitu: Can Vertical SaaS Survive the AI Onslaught?'

Since the release of Meitu's 1H25 financial report, its market capitalization has halved. From a market environment perspective, on one hand, geopolitical risks are driving capital toward assets with high certainty. On the other hand, the narrative of 'AI devours software' is intensifying as large model vendors transition from Chatbots to the Agent era. Is Meitu being unfairly penalized? In our initial coverage, Dolphin Research explored Meitu's defenses against the impact of large models, with the following key points:

Despite only three months having passed since our initial coverage, Dolphin Research believes that some of the company's core assumptions have been challenged due to the rapid iteration of AI interaction methods and the swift advancement of model intelligence.

A) On the consumer (C-end) side, Dolphin Research still believes that Meitu's foundational features, such as its 'editor,' provide 'autonomy and controllability' that can retain moderate and heavy users in the short term (medium to long term depending on the pace of AI development), with stable user retention rates. Since image editing operations inherently rely on subjective user judgment, which is non-standardized, interactions with large models via natural language still carry hallucinations and uncertainties.

- In terms of iteration, LLM iteration speeds have significantly accelerated (e.g., MiniMax completed its iteration from M2.5 to M2.7 in just one month). Additionally, the code generation capabilities of various models continue to improve, with zero-code development enabling ordinary users to easily build scripts and even standalone software using large models. The trend of 'everyone can create tools' to some extent weakens Meitu's advantage in functional iteration and impacts light-use scenarios. Of course, Meitu still maintains certain competitiveness by building end-to-end capabilities in deep workflows and combining aesthetic demand discovery and exploration.

- From a traffic perspective, the popularity of OpenClaw seems to indicate the traffic entry point for the next generation of Agents, with domestic tech giants rushing to build their own 'shrimp farms.' Users no longer need to open specific software to accomplish specific functions, and the competition for traffic entry points has always been fierce in the domestic internet landscape.

In fact, Meitu recognized the potential of Agent interactions and launched its standalone AI Agent product, RoboNeo, in July 2025. However, as analyzed by Dolphin Research, its position as a standalone product remains unclear.

Currently, while the company has integrated Agent capabilities into various product lines, which could boost ARPPU from a token consumption perspective, this is not sufficient. Based on market feedback on single-product operating data, RoboNeo still struggles to serve as the entry point that connects all products.

B) The popularity of the 'lobster ecosystem' and the failure to secure traffic entry points led Meitu to recently release its official AI Skill, actively integrating into the ecosystems cultivated by tech giants and providing users with callable, combinable, and reusable standardized capability modules. We believe this expands B-end business scenarios and offers some monetization prospects, but it is unlikely to significantly alter its competitive advantage.

i) Why focus on the B-end? On one hand, the most widespread C-end use case is photo editing, where, as emphasized by Dolphin Research, Meitu's core competitiveness lies in high-precision controllability. On the other hand, Skill is essentially an interface-based, modular capability that better aligns with the batch and automated productivity needs of B-end scenarios.

ii) Why is there monetization potential? Taking Meitu Design Studio as an example, Dolphin Research previously highlighted 'workflow encapsulation' and 'low pricing' as key features of the product. Skill amplifies these advantages:

- Reusable workflows mean that B-end users can build automation scripts based on them, further increasing migration costs when embedded into production processes. Skill, as a direct encapsulation of mature workflows, is attractive for scenarios with a strong need for efficiency improvements.

- The pricing model, which charges based on successful outcomes, significantly reduces trial-and-error costs for B-end users.

iii) Why is it difficult to change the competitive landscape? Integrating into the 'lobster ecosystem' does not enhance Meitu's competitiveness: The development threshold for Skill is low, and competitors or even individuals can create similar offerings. The true competitive edge still lies in Meitu's years of accumulated backend data assets. However, for new B-end businesses, Meitu is still in the accumulation phase.

In summary, the logic underlying the C-end foundation has been undermined, leading to a significant drop in valuation, but the competitive landscape remains relatively benign. The C-end business, with nearly 250 million users, remains the core focus of the company. According to UBS monthly tracking data, the product's MAU share has remained relatively stable, with Meitu maintaining its leadership in the domestic photo editing sector.

In video editing, Wink has seen explosive growth overseas, forming a differentiated competition with ByteDance's CapCut, with annual revenue growth consistently exceeding 100%.

However, the MAU share in the sector continues to shrink, and combined with the underwhelming MAU data in the second half of the year, the competition for light users from large models is inevitable. Nevertheless, the user structure has improved. Therefore, on the C-end, Dolphin Research advises against overemphasizing marginal changes in MAU, with subscription penetration (implying the appeal of features to moderate and heavy users) and ARPPU growth (implying the value contribution from overseas markets) being more critical metrics.

On the B-end, the earnings call revealed that the company will hold an Imaging Festival in June this year to launch new productivity products. Based on Dolphin Research's calculations, the current market space for the product photography and live-streaming sectors that Meitu is involved in is only about 2.5 billion RMB, a result of the company's deliberate avoidance of competition with tech giants, giving it a somewhat precarious position. Undoubtedly, the company's proactive expansion into vertical B-end scenarios to counter competition from general-purpose large models is the right choice. However, from a financial perspective, it is difficult for capital to place high expectations on a B-end business with slowing growth in both MAU and subscription penetration. The company will need to demonstrate stronger resilience in future quarterly disclosures of core data.

Financial Report-Related Charts

- END -

// Reprint Authorization

This article is an original work by Dolphin Research. Reprinting is permitted only with authorization.

// Disclaimer and General Disclosure Notice

This report is intended solely for general comprehensive data purposes, designed for general viewing and data reference by users of Dolphin Research and its affiliated entities. It does not take into account the specific investment objectives, investment product preferences, risk tolerance, financial situation, or particular needs of any individual receiving this report. Investors must consult with independent professional advisors before making investment decisions based on this report. Any person making investment decisions using or referring to the content or information mentioned in this report assumes all risks. Dolphin Research shall not be liable for any direct or indirect responsibility or loss that may arise from the use of the data contained in this report. The information and data contained in this report are based on publicly available sources and are for reference purposes only. Dolphin Research strives to ensure, but does not guarantee, the reliability, accuracy, and completeness of the relevant information and data.

The information mentioned or views expressed in this report shall not, under any jurisdiction, be considered or construed as an offer to sell or a solicitation of an offer to buy securities, nor shall it constitute advice, an inquiry, or a recommendation regarding relevant securities or related financial instruments. The information, tools, and materials contained in this report are not intended for distribution to, or use by, any person or resident in any jurisdiction where such distribution, publication, provision, or use would be contrary to applicable laws or regulations or would subject Dolphin Research and/or its subsidiaries or affiliated companies to any registration or licensing requirements in that jurisdiction.

This report merely reflects the personal views, insights, and analytical methods of the relevant creators and does not represent the stance of Dolphin Research and/or its affiliated entities.

This report is produced by Dolphin Research, and the copyright is solely owned by Dolphin Research. Without the prior written consent of Dolphin Research, no institution or individual shall (i) produce, copy, reproduce, duplicate, forward, or create any form of copies or reproductions in any manner, and/or (ii) directly or indirectly redistribute or transfer them to any other unauthorized persons. Dolphin Research reserves all related rights.