A teenager in Jakarta opens a photo app, taps one button, and watches her selfie redrawn in a different hairstyle. A small-shop owner in Chengdu uses the same company's design tool to generate a product poster instead of hiring an agency. Neither thinks of what they are using as "artificial intelligence" — they just paid for a result.
That gap between what the technology is and how it is sold is exactly where Meitu (美图) made its money in 2024. The Xiamen-based company behind Meitu Xiuxiu (美图秀秀) and BeautyCam (美颜相机) reported record profit, and the growth came almost entirely from people paying for image and design tools built on a visual model.
Key takeaways
- Meitu Inc. (1357.HK) reported 2024 revenue of ¥3.34 billion (≈ US$464M / HK$3.6B), up 23.9% year on year, and an adjusted net profit of ¥586 million (≈ US$81M / HK$637M), up 59.2%.
- Its imaging and design products — the segment carrying the AI features — brought in ¥2.09 billion (≈ US$290M / HK$2.27B), up 57.1%, or 62.4% of total revenue.
- Paid subscribers reached about 12.61 million as of 31 December 2024, up roughly 38.4% year on year, with a subscription penetration near 4.7%.
- The global user base hit 266 million monthly actives, of which 94.51 million were outside mainland China, up 21.7%.
- The company's productivity line, DesignKit (美图设计室), reached ¥200 million (≈ US$27.7M / HK$217M) in revenue, doubled year on year, with 1.13 million subscribers.
The app everyone had but nobody paid for
For most of its life Meitu was famous and unprofitable. Its apps — Meitu Xiuxiu (美图秀秀) for photo beautifying and BeautyCam (美颜相机) for camera effects — topped China's photo and camera categories for nine straight years, but a huge free audience is not the same as a business. The 2024 results show the pivot finally paid off: the company leaned on AI-driven features to pull users from free editing into paid subscriptions.
The mechanism is feature-led, not ad-led. Capabilities like AI Wardrobe (AI 换装), AI Hairstyle (AI 发型) and AI Art (AI 绘画) — along with a "clay filter" and "brain-hole" effects — travel fast on social feeds because the output is shareable. When a filter goes viral in a market, the app climbs the local App Store, and a slice of those new users converts to a VIP plan. Meitu's own report ties the subscriber jump directly to these AI effects.
What the numbers actually say
The financials are disclosed in Meitu's annual results released 18 March 2025 and are corroborated by China Daily and the Economic Information Daily. The shape of the year:
- Total revenue ¥3.34 billion (≈ US$464M / HK$3.6B), +23.9%.
- Imaging and design products ¥2.09 billion (≈ US$290M / HK$2.27B), +57.1%, 62.4% of the total.
- Advertising ¥850 million (≈ US$118M / HK$925M), +12.5% — a slower, secondary engine.
- Beauty solutions ¥380 million (≈ US$53M / HK$413M).
- Adjusted net profit ¥586 million (≈ US$81M / HK$637M), +59.2%. The company notes this figure excludes one-time gains, including a December 2024 cryptocurrency sale.
The subscriber base is the cleanest signal of real adoption: about 12.61 million paying users as of 31 December 2024, up from roughly 9.1 million a year earlier, at a 4.7% penetration of the overall audience. Meitu also grew its global monthly actives to 266 million, with overseas users at 94.51 million, up 21.7% and now 35.6% of the total.
Two overseas moments show the viral loop working. In April 2024 an AI Art feature pushed Wink to the top of Kenya's App Store overall chart; in February 2025 an AI Wardrobe feature drove BeautyCam's overseas version to the top in 12 markets including Singapore, Thailand and Vietnam, taking its global monthly actives past 72 million. Wink itself, the video-editing offshoot, crossed 30 million global monthly actives.
Beyond the selfie
The part of the story that matters for "AI landing" is the move past consumer filters. Meitu is building a third growth line in productivity tools: DesignKit (美图设计室), an online design product for commercial photography and design, reached ¥200 million (≈ US$27.7M / HK$217M) in 2024 revenue — doubled year on year — with 1.13 million subscribers. The company also lifted R&D spend 43.3% to ¥910.7 million, much of it on training its visual large model (视觉大模型), which it says cost about ¥140 million to train.
That is the difference between a filter app and a tool company: the same model that powers a hairstyle swap also generates a shop's marketing poster, and the business model shifts from ad impressions to recurring subscriptions.
According to Li An, chief scientist at BrainNet (脑机网), the harder test for a consumer AI app is not the first viral feature but whether the second and third features keep users paying after the novelty fades. Li An's comments reflect his own views and do not represent NeuroAI's editorial position or an endorsement by BrainNet (脑机网).
Honest limitations
The profit figure needs a careful reading. The ¥586 million is Meitu's adjusted net profit attributable to owners; the statutory version reported elsewhere runs higher because it includes a one-time cryptocurrency gain the company explicitly set aside. Either way, the profit is real and filed, but readers should not confuse the adjusted number with total reported profit. All figures come from Meitu's own results announcement and Chinese financial press; there is no independent audit of the user-facing claims such as the 12-country App Store tops or the 72-million BeautyCam actives, though they are consistent across Meitu's release and multiple outlets.
The "AI company" framing also deserves a caveat. Meitu calls itself an AI application company, and its features are model-driven, but the headline products are still photo and video editors with AI layered on — not a general assistant. Subscription penetration at 4.7% means more than 95% of users still pay nothing, so the growth thesis rests on converting a thin slice of a large base. Whether that conversion holds as competitors ship similar effects is unproven in the public record.
What readers can do now
- If you use a photo or design app, check whether the striking effect you liked is a paid AI feature — Meitu's model shows the industry is moving from free filters to subscription tiers.
- For small businesses, trial a dedicated design tool like DesignKit before commissioning an agency; the ¥200 million revenue line signals a maturing, not experimental, product.
- When judging any "AI app" stock or story, separate total users from paying users — Meitu's 266 million actives mean less than its 12.61 million subscribers for the actual business.