Meitu Inc (MEITU) has released its financial results for the first half of 2026, revealing total revenue of RMB 2.21 billion from continuing operations, a year-on-year increase of 22.1%. Adjusted net profit attributable to parent company equity holders reached RMB 650 million, up 39.5% compared to the same period last year. Alongside this growth, the company has seen its user base and monetization capabilities improve in tandem, creating a dual-engine model where daily-life applications form a stable foundation and AI productivity tools accelerate at a rapid clip.

In addition, Meitu has introduced an AI compute point-based payment model, which removes the revenue ceiling traditionally imposed by subscription services and redefines the underlying monetization logic. During the earnings presentation, Meitu’s Chairman and CEO, Wu Xinhong, expressed strong optimism about the AI application layer’s growth prospects, noting that the company is entering a highly critical window of opportunity. According to the company, the ARPPU for its productivity applications is already 50% higher than that of its daily-life applications, and that gap continues to widen.

Breaking through the monetization ceiling with subscriptions plus AI compute points

After years of product iterations, Meitu’s revenue structure has shifted markedly. In the first half of 2026, revenue from its imaging and design products climbed 30.9% year-on-year to RMB 1.77 billion, with its revenue share rising to 79.9% from 74.5% in the same period of the prior year. Among these, daily-life applications such as Meitu Xiuxiu and BeautyCam, which underpin the company’s core business, contributed approximately 82% of imaging and design product revenue, serving as a steady source of cash flow.

Meanwhile, productivity applications have opened up new growth avenues and expanded ARPPU potential, with revenue jumping 40.1% to roughly RMB 323 million. In the first half of 2026, Meitu’s products including Meitu Design Studio, Kaipai, and Vmake Labs have begun to demonstrate scalable commercial value. As of the end of June, monthly active users for Meitu’s productivity applications hit 33 million, up 43.5% year-on-year and reaching an all-time high. The annual recurring revenue run-rate for these productivity tools rose to RMB 620 million, while paid subscribers grew 29.8% to 2.35 million.

During the first half of 2026, as agent capabilities were gradually integrated across its product suite, Meitu began extending toward a usage-based AI compute point consumption model, which removes the spending limits inherent in previous subscription tiers. Wu Xinhong disclosed that in the first and second quarters of 2026, total user spending on AI compute points rose by more than 46% quarter-over-quarter, accumulating to a roughly 113% increase over the six-month period. The company notes that Meitu’s productivity application ARPPU is now 50% higher than its daily-life applications, with the divergence steadily growing.

This disparity, according to Meitu, highlights a practical shortfall of the legacy subscription model: the tiered pricing structure placed a ceiling on how much users could spend, and even when they were willing to pay more, the product lacked a mechanism to capture that additional value. Meitu’s Chief Financial Officer, Yan Jinliang, explained that with the introduction of agent capabilities, that restriction has been lifted—so as users tackle more complex or frequent tasks, their AI compute point consumption rises in line with the actual workload.

High-value users spend tenfold the average ARPPU

Meitu’s Chief Product Officer, Chen Jianyi, highlighted that within Meitu Design Studio, there are already several thousand high-ARPPU users whose annualized spending is roughly ten times the company’s overall average. Moreover, the number of such users has grown approximately six-fold since December of last year. Using Portrait AI Agent product Picchi as an example, Chen noted that when personal image becomes a production asset, users’ needs and willingness to pay for imaging products undergo a fundamental shift. Meitu Xiuxiu and BeautyCam cater primarily to everyday life scenarios, centered on the goal of “making me look better.” Picchi, by contrast, targets streamers, influencers, and models who monetize their personal appearance—for them, appealing images are not just consumption but a means of production, so their propensity to pay is naturally higher than that of typical C-end consumers.

The battle between large models and the application layer? Rivals emerge from startups

Since the rise of large language models, a persistent concern surrounding AI application-layer companies has been whether they risk being overwhelmed as general-purpose model capabilities continue to advance. Meitu’s management offered a definitive stance during the earnings call. Wu Xinhong stated, “We are very bullish on the opportunities in the AI application layer. For Meitu, this is a highly significant window.” In his view, Meitu’s primary competition comes from startups, because as the industry transitions from the 1.0 PC era to 2.0 mobile applications, extends into 3.0 products such as Wink and Kaipai, and moves toward 4.0 AI-native applications, most products no longer require real-world footage to be imported—content is directly generated and created. “In the 4.0 stage, we see that many highly competitive companies globally are almost all startups. They are not constrained by existing product frameworks, user interfaces, or established market playbooks, and they redefine products and marketing in a more native way,” Wu said.

The financial report indicates that in the first half of 2026, more than 96% of generated outputs came from Meitu’s self-developed model layer, while third-party API costs remained at a mid-single-digit percentage of imaging and design product revenue. This suggests that Meitu has not simply layered on external large models, but rather built controllability over both cost and quality through proprietary vertical models.

Chen Jianyi added a product-level perspective, explaining that Meitu now evaluates market opportunities through “creator density.” The leading indicator for assessing a market’s potential, he said, is not population size but the structure of its creators—places with a large base of leisure creators and a substantial number aspiring to professionalize and monetize content represent Meitu’s opportunity. “Kaipai’s growth in the domestic market stems from users in sectors like healthcare, insurance, real estate, and food service who understand their businesses well but may lack expertise in topic selection, scripting, or video production. Kaipai allows them to generate near-professional-grade content without needing to become professional creators,” Chen explained.

During the earnings call, Wu Xinhong outlined Meitu’s long-term strategic direction: beyond deepening its AI creation applications, the company plans to expand further into AI content distribution and AI asset libraries, enabling content to be created, circulated, and accumulated within the Meitu ecosystem. This, he said, would build a complete content ecosystem chain in the AI era. “Going forward, on the one hand, we will continue to deepen our existing products and vertical scenarios; on the other, we will proactively seek out and rapidly validate new growth directions through a more efficient innovation mechanism.”