[Foreign language]
Productivity tools are quite well performing. Actually, we have realized very good results in these two products. For example, the Meitu Studio launched two years ago, is already ranking high in China. In last year, the revenue was about RMB 200 million in a single product last year. Also, KaiPai, in terms of MAU alone, has become number one in China talking head video tools. The results we have realized in China mainland, and also the experience we gained from this market, allows us to globally expand with greater confidence. Lastly, I'm going to talk about our strategy for the large model development. Since our products are based on base AI technology to enhance our products, to enhance subscription, we have a strong emphasis on AI technology, particularly the deep AI technology.
Now, our home-developed large models are actively deploying different kinds of models and also utilizing high-quality API interfaces and also different applications to enhance our capability, to enhance our products. For example, based on the MiracleVision large model, part of the Meitu Designer functions, such as live ChatGPT, when it comes to semantic understanding, we have incorporated DeepSeek capability to enhance the export of our text, the reasoning in text. Also, Moqi is also providing a lot of options for the video functions for the users to choose to increase their satisfaction. We call this a model container. This strategy will allow us to reduce unnecessary training for foundational models, and will also allow us to have greater budget to specific scenario training. This is what I'm going to talk about today. That concludes my speech. Now I'll hand over to Gary for the financials.
Thank you, Xinhong. In 2024, our overall growth grew by 23.9%, achieving RMB 3.34 billion. Our productivity and globalization strategy continues to deepen the overall revenue in the past year and the past in several reporting periods. Enjoy great growth. This is reflective of our strategy. The video and photo design products grew by 57.1% to RMB 2.09 billion. With AI-driven power, our products are continuing to increase to provide better user experience, provide better video and visual processing and creativity. In 2024, advertisement revenue grew by 12.5% to RMB 850 million, basically coming from programmatic advertisement. We will continue to strengthen this kind of operation to maximize advertisement presentation, to increase click-through rates, to more precisely target the users. In beauty industry solutions, we dropped by 33.4% to RMB 380 million. For this non-core business, we are not pursuing the revenue growth alone.
In the medium and long term, the productivity tools outside of China will be focused on increasing subscription penetration rate, and also will be core centered around the visual and photo and design products. Also from and by regions centered around our global strategy. In the past year, revenue from the international market grew by 42%, much higher than the China 17% in 2024. Revenue from the international market accounts for 32% of the total growth. That is greater than the 28% in 2023. In the international market, the video and photo design products accounted for the biggest ratio, to 89%. In gross profits, for 2024, it was RMB 2.29 billion, grown by 39%. Also the gross profit was maintained at 68.7%, 7.2 percentage points higher than 2023.
Because the video, photo, and design products are highly profitable products, they will lead to drive the overall growth of our profit margin. Now we are talking about the fees. The R&D expenses were RMB 911 million, and year-by-year growth of 43.3%. A big part of it was attributable to our visual large model training. In the visual model training, the cost was RMB 140 million. That was concentrated on the computational power. We spent a lot of money on the foundational model training. As we are finishing up the video model training, this part of the cost will be reducing. We have adopted a model container strategy. We think that the R&D costs will not grow so much this year in terms of capability and computation.
We have more R&D staff just last year that contributed to the greater R&D costs. We think these R&D staff will be effectively enhancing the company's power, competitiveness in core visual front. Last year, we have completed a merger of ZCOOL, and that will bring a RMB 20 million raise in the R&D cost. On sales costs, overall in 2024, it was RMB 480 million, a year-on-year growth of 12.9%. This growth basically came from the visual and audio products promotion, which grew by 30% year-on-year to RMB 250 million. Overall, we anticipated that in 2024, the promotion fees for these products will continue to grow, but the core growth strategy is still centered around products and also PRG strategy. Relatively speaking, promotion in the international market was greater than in China. Administrative costs grew by 32.9% to RMB 400 million.
New management fees, among the new management fees, around 45 million came from the M&A of ZCOOL. Other costs came from staff manpower and also staff-related rise. In 2024, our profit could be RMB 490 million, a year-on-year growth of 39.2%. This high-profitable video and visual design products will continue to drive our business, despite our rising R&D costs for AI products. Overall, our operational cost is lower than the growth and gross profits. This operational leverage is quite significant, is quite obvious, taking our overall profitability forward. Also apart from the main business, our non-operational business items are also bringing positive impact on the company. Particularly in December 2024, we sold all the cryptocurrencies and earned one lump sum big gains. There are also some long-term investments, to some extent offset against the gains from the sale of the cryptocurrencies.
From 2024, our net profit in IFRS perspective was CNY 810 million, year-on-year growth of 113%. Also, our dividend payment strategy was adjusted by 40%. The adjusted net profit attributable to parents will have some changes. It will continue to center around the productivity and globalization, continue to deepen AI, and also continue to drive our product forward with AI. Now that closes my sharing, followed by the Q&A session. Now we have the Q&A for investors. If you are attending on site, please raise your hand. Our staff member will hand over the microphone to you. If you are attending from your mobile phone or from your laptop, you can hit the raise hand button to ask your question. When you are unmuted, you can ask your question.
You can also type your question in the chat box so we can read out your question. The investors in the English channel can also raise your hand or type down your question. If you're attending from your phone, please hit star one before you start asking. Before your question, please let us know your name and also your organization. Thank you. Now is the on-site investor Q&A.
Fourth row, left-hand side, lady in white, please. Our staff member will hand over the microphone to you.
Hi, I'm Hua Xia Internet. Mr. Wu, first, congratulations on your great profit and also on the lot of surprises in paying subscription. Two questions. First, I understand that the company has started from last year, been putting into action the globalization and productivity tools. We have realized that since the start of last year, there has been some significant changes.
I want to ask, on these two parts, these two strategies, do you have any data to share with us? That's question one. Question two, we have seen recently the development evolution of large models, particularly the visual large models. We have seen some launches from competitors, like the 2.0 Flash, and also some visual models testing results. I want to ask your views on these rapid progress of the competitors.
Okay, now we have a question answer from Mr. Xinhong from online.
I'll just debrief regarding the productivity and globalization strategy. At present, as we can see, last year our productivity holds more or less were around user growth and profits with surprising growth, just like the beauty and Meitu beauty design. The CDO, the gross profit achieved RMB 200 million.
That's the most rapidly growing product since the birth of Meitu, and the subscription penetration rate is already 12.8%. I would say in the China mainland market, or among the early launch products in China mainland, this has the highest subscription penetration. That gives me confidence that productivity tools will be a powerful driver for growth. Also on globalization, we have a very good beginning this year. In my presentation, I talked about the BeautyCam, the AI Wardrobe. They have gone viral internationally. The BeautyCam brought us more than RMB 220 million overseas new users. Our investment in these actually was not that great because it was basically some computational power and marketing and also appropriate channel purchasing investment. I mean, compared to the new growth, our investment was not that big. The return rate was quite surprising.
I believe that this year we will have some more growth like this because we have realized the AI application will be an explosive growth point. Internally, we have set up our target, which is we would like to bring into some core capabilities. We also will work harder and deeper on these capabilities so that we will remain competitive around the globe and to remain far ahead of our competitors in the world. Okay, then coming back to whether it is photo or videos and models, we are seeing like a strengthened open source. We are seeing a lot of suppliers opening source and investment growing bigger and bigger.
In general, this is a good thing because Meitu has always been benefiting from open source ecology, the ecosystem that will allow us to benefit faster from existing sources, resources from competitors, from peer enterprises, like open source models that allow us to implement our own models quicker. We understand our positioning. We are an AI application firm rather than a large model firm. Therefore, on AI application, we believe that we are strongly advantaged, whether in terms of our current user scale or because we have accumulated through 17 years technologies, particularly in algorithm and engineering. Also our aesthetics and also our insights in users, these will allow us to grow bigger and more competitive. At the same time, we should also see from models to applications, there is a big gap.
I mean, even if there are open source models and also APIs being provided, actually developers can find it very hard to use them directly. There is a lot of engineering in between. We believe that Meitu has such an economy of scale to make that happen. We have the engineering capability to compete in the application scale scenario.
Thank you. Okay, Gary, do you have anything to add?
We will have more questions from on site.
Okay, this lady, please go ahead.
Xinhong and Gary, hello. I'm Lydia from Morgan Stanley. My first question in 2024, the profit rate and also overseas productivity tool growth, can you share with us some more information? The second question is, do you think if you think AI is increasing or reducing the threshold to this industry and also how Meitu remains competitive in the AI era? Thank you.
Okay, question one, Gary, you can go ahead.
Okay, the profit guidance, actually, we generally do not define it this way, but we are aware of. Analysis actually is a realizable figure. And on productivity growth, revenue growth, our guidance is, let me put it this way. Overall, the rapid growth momentum will not change. However, in 2025, the year 2025 may not be the biggest explosive year because the whole productivity growth, revenue growth comes from overseas productivity tools. This market, and we have actively been actively deploying in this market since last year. Also, we will maybe 2026, the explosive period at the second half of 2026. On this front, I think in 2025, you will see a lot of, just like Xinhong just said before, like KaiPai penetration rate is very high.
These early launch, early indicators showing that these productivity tools will drastically transform our life scenarios. That will allow the explosive growth in the subscribed products. If we want to translate this into income, it may not be happening in 2025. What about Xinhong, do you think anything to add?
Yeah, yeah, I do have some. I think on question two, the AI, obviously, particularly the generative AI is largely reducing the threshold for app developers. As we can see in the video and photo arena, we have seen a lot of small to micro startups. Of course, this will make the whole market very vibrant. For us, in general, there are some upsides and downsides. Let me just talk about the upside of it. Open source ecosystem is, of course, beneficial to an application firm like us.
We are more working on the model container because our home-developed large model for post will not give it up because there has been a lot of capability developed inside. We will take advantage of this home model and combine it with external open source models so that we can continuously do more testing, like the efficacy of the models so that they can serve the users better. We will also be working on the model acceleration framework so that it will increase efficiency and reduce cost. The challenging thing is because more people, there are more market entrants now. For Meitu, we have to be more focused on our core competitiveness, core capability, which is, we think, these highly potential vertical scenarios, like the commodity videos, AI special effects, and AI photos so that we can serve the e-commerce industry, like video creation.
We can provide very highly efficient workflows. From models to applications, just like I said before, it's not easy. We will strengthen our capability to transform from models to applications. There is a lot of work to do here, a lot of efforts in planning. We need to have some economy of scale so that we can remain competitive. Thank you.
Next question. On site, if you have any questions, you can raise your hand. The third row, the gentleman on the third row.
Thank you for taking my question. I'm [Shanghai Jue Shang Securities, Peng Yingyuan]. I want to know about your ongoing writing down plan for shareholders. How many shares are already there or left there? The Wens heng writing down is his own personal behavior. Now, there may be 20% shares left there.
I don't think this will affect our operations because overall, we remain focused on the sector. We have very good, strong growth drivers, and the whole team is very confident. Gary, anything to add?
Let me just add something here. Wensheng, he's exited from the position as a shareholder. He is now just an investor. He has not been involved in the company's operations and not informed of anything happening inside the company. Actually, he very rarely communicated with us. From this perspective, his trading of the shares does not give rise to any hint or information. It doesn't show anything about the company being good or not, healthy or not, because the trading of stocks is more about his own personal investment strategy. From the perspective of a company, we think in just the same as a general investor.
Okay, the onsite Q&A continues. If you have any questions, you can raise your hand. Okay. Now, we have a question from Tencent Meeting online. From Guoxing Securities, Ms. Chen Shuyuan, now we are unmuting you. You can start asking.
Xinhong and Gary, good afternoon. Chen Shuyuan from Guoxing Securities. We are talking about the monthly solutions. We've seen that there is some drop in 2024. I want to ask why. This is question one. Second question is about our model cost testing because we have seen the since launch of. See, it may have some impact on our training costs. I want to ask, like the Wensheng image and Wensheng video, is there anything changing in the cost happening? Going forward, with the decreasing cost, there's any positive side on that? Thank you.
In the beauty sector, income drop. I just talked about it in my presentation first. The beauty part, the decrease in the beauty part is actually because we are moving our capability to those divisions with higher gross profit because we're not trying to make up a bigger income. From the gross profit and also the net profit perspective, even though we're dropping in 2024, actually we're growing. This growth from the perspective of net profit may be millions of RMB. Actually, it's not such a big impact on our company overall. Going forward, we're not going to invest a lot of resources to push it forward. We will pick the right time to see what we can do about it. On the large model investment and the cost reduction, first, in the presentation, I mentioned from R&D was about RMB 140 million in 2024.
This figure, I think, will remain mostly the same in 2025. There is maybe some drop. I think this corresponds to the last analyst. From the reasoning perspective, actually, our total reasoning cost accounts for the cost of revenue. In the cost of revenue, this accounts for a very small portion of it. The cost reduction from the general profit statement is not such a big thing because overall, the whole profit statement, most important is the growth in revenue and gross market and gross profit. Whether it is DeepSeek or core or other open source models, the key is that we can base on these open source models, we will have more fine-tuning, fine-tuning more vertical scenarios and services. From the revenue side, the revenue growth actually outweighs, outperforms the cost reduction. This is the actual benefit.
Okay, Xinhong, anything to add?
Our training cost actually will continue to invest in, but it will not be a significant big growth. Is this switching, moving from investing in foundational models to investing in vertical scenarios to our core competitiveness? Just like Gary said, it will be a higher investment and revenue ratio there. We also see the general growth reduction in costs. We believe that the overall training or reasoning costs will continue to grow, continue to decrease. For Meitu, we will have significant growth in AI penetration in current products. We will further utilize generative AI to redo our core competitiveness because we have been doing AI before, but it was a computer vision. Now we are using more technology to transform ourselves. This will be beneficial for better user experience.
Yeah. Of course, the reasoning cost, a larger reduction in the reasoning cost will allow us to relieve the pressure on the user side because if the reasoning cost is high, the user may find it, may find it more expensive to subscribe to our product. We still have the opportunity to maintain our costs acceptable to the users. Thank you. Thank you, Gary and Xinhong, very clear.
Thank you. Next question.
The next question comes from an online investor, [Yang Yongyi], in text. The first question is, C consumers' training scenario. What's the strategy for C consumers in 2025? What about the overseas sales overall acceleration? Second question is on the productivity scenarios. Will we expand into new industries in 2024? Any new product plan? Now, let's have Xinhong answer this question first.
The first one is the C, right? The C end strategy for China and overseas. Whether it's China or overseas, the C end users is a basic user group of Meitu. In our newly relaunched financial reporting, we have a monthly active users of 266 million. Such a big user group will become a large pool for videos and photos. I mean, most of the demands for these users will be able to match by the Meitu products. And such a pool, user pool, will be very beneficial to our newly launched productivity tools. We'll be able to do very good PMF testing organization. I mean, from user demand insight to quick iteration, reiteration, and also then to release the power. This whole series, this whole process is based on our large user group. We'll continue to grow the users on the C end.
We also hope that we will be able to form several big segments for videos and audios and photos. By the Meitu Xiuxiu, it is basically around image. Also, Wink is basically around videos. The Meitu Studio is online design. Photo, video, online design is actually they are actually separate independent arenas. We have seen a lot of different arenas can be integrated. They are independent. In Meitu Xiuxiu, when it has become a data generator, we will need to create, transform Wink and the Meitu Studio on the productivity tools.
Sorry, I want to ask, is it about strategy or something? The productivity scenario, will you expand to any new sectors in 2025? Any new products?
Okay. Actually, productivity tools, we will not expand too much on it into too many sectors. We will be more focused on one important sector, I mean, e-commerce, to do in-depth mining development because every sector has a lot of barriers, a lot of know-hows that we are not aware of or fundamental capabilities. Meitu's proven successful products are all around e-commerce, like the Meitu Studio around the e-commerce materials and also KaiPai working on e-commerce and also Talking Head video tools around video creation. Of course, we have a lot of more examples just like the and also Xiaomi, even though it was initially doing the gaming materials production. In the future, it will go into the e-commerce material creation and casting. Also, the Meitu Cloud Streaming, in addition to the AI batch trimming, we will be moving into e-commerce sector for e-commerce model image. I mean, AI batch trimming.
We want to find a big industry where we can build up a large capability, the large workflow essential to e-commerce. Then we can put together all of these core capabilities to provide an all-in-one package for this whole industry. I want to reiterate that e-commerce is, of course, not the only one that we're working on. There will be some premium, like if you can do the e-commerce materials well, design or videos well. Likewise, we may be moving on to marketing or advertising, etc., because these industries are similar, quite similar to each other. To be simple, we will be focused on one industry and then work deeper into that.
Okay, thank you. Anything to add from Gary?
Okay, no from me.
More questions from on-site. This investor, hang on one second. Executives, hi. I'm coming from Shanghai Sheng Yu Investors, Wu Xiaochuan.
I want to ask, in the 2025 development of productivity tools overseas, we understand that the payment overseas market is better than in China, but there is a large difference because we just started to do this business in 2024. Any challenges that we come across in this process and how do we address these challenges and difficulties? Could you please share with us some of your insights? On the overseas productivity tools, I think, to what level should we achieve, like in terms of revenue or user number or payment rate, and to what level do we consider we are having a good foothold in overseas? Okay. Let me try to answer this question first and then followed by Gary. On the globalization of the productivity tools, of course, there are obviously a lot of challenges.
First, the global market, particularly the European and Western American market, our team does not have the real deep insight. Last year, we started to send, I mean, to send a lot of the senior officers in the company overseas to overseas market to be stationed there so that they will have formed a deeper understanding of their local market, whether in terms of users or enterprises' demand. This process may be slow, but it will be effective. The returning staff members show us a stronger determination for globalization. They have a multi-dimensional understanding of the overseas market. Of course, there are more challenges there. For example, will it bring a big rise in the manpower cost? We are looking for some cost-effective points overseas because not every country has such a high labor cost.
We are looking for some jurisdictions where the manpower supply was sufficient with reasonable cost to establish our overseas presence. Some of the challenges are not, how should I say, that we can't do anything about it. Like those more complicated geopolitical factors, like I mean, for all Chinese companies going overseas, this is their challenge. At present, several overseas products on the productivity tools, it is still in the early stage. I can give you an example, like Meitu Studio, the overseas version like X-Design or the KaiPai overseas WeMac. Internally, we are far from satisfaction, actually. We hope, firstly, these products will play well in regions that we're not very good at. Because I said we have a large user group. The large user group is the basis of commercialization.
We will be more looking for rapid user growth and also the further localization of the team because the best localization is localization. I mean, we need to have our products launched in the target country should be a localized product, an indigenous product so that users will find it close and indigenous to them, not strange. This sounds simple. It is quite challenging to realize it from internally. Anything to add from Gary?
Maybe at this stage, we are not using revenue as a determinant for success. I may be answering your questions in two aspects. First, I think we still need to really improve user experience and value in our products because productivity is to help them achieve more goals in their jobs, earn more money, and also improve their efficiency. If our products can achieve that goal, that means we have the user recognition.
We can set a very strong foothold. Another aspect is we have been emphasizing the overseas productivity scenario. Actually, the capacity based on 2024 may be 20x than the live scenarios. The whole live scenarios market now is in a rapidly growing state. We can only say if we can find something useful in this market, that may be very hard to measure using today's parameters. Like if you keep looking forward, more than 10% or 20% growth in the future, it may not be the case. I mean, the live scenario is our third growth curve, and the tangent should be higher than what we are doing today.
Thank you for your response. Due to time constraint, the last question, last three questions from now. First, the second row on site.
Xinhong, Gary, hello. Chu Ming here. I have a couple of questions.
First, the domestic payment rate increase, what's your view? How do you maintain the rapid growth in products? What's the potential spatial growth? Another question is the overseas payment rate. What is the overseas payment rate? Any new product development plan in overseas markets? You mean overseas what? Subscription penetration? Payment rates. What's the payment rate? Also, for different products, their development plan in different countries, regions.
I'll just say something and then follow by Gary, okay? As we can see at this stage, our average subscription penetration is not quite high, actually. Whether in terms of China or overseas, there is a big room for growth. We're not in a rush to pull to raise this penetration because that would mean we have to raise the pay bar, and that will potentially impact user experience or monthly active users count.
We will apply a very stable, steady approach, I mean, by improving productivity, product power to drive the payment and conversion. This is what we do, basically. On Wink, though it is among the video tools, it ranks second, second to Jianying. Actually, the gap is quite high from the top first because market rate, Jianying is over 60%, and Wink is just about 10%. We are seeing whether for Jianying or many of our peer competitors, they find it difficult to meet all the demands or they find it hard to become an all-rounder, a general excellent enterprise to meet all the demands. We are actually complementary to each other. Just like the AI image trimming and also the video beauty actually are complementary to Jianying. Actually, they are not doing well better than us in these two segments.
Going forward, we will continue to work on these core capabilities to invest more in these areas to be first, to be number one in the world, and then we'll maintain a rapid growth. We're expecting Wink, just like I shared before, will become a new video data generator for Meitu just like the video version of Meitu Xiuxiu because we have a lot of products to provide. On the overseas market, the subscription in overseas, I think Gary can add some more. Let me put it this way. I will try to give you some hints on how to imagine this penetration in China or the overall video photo products. First of all, I want to share a little story from inside.
In the beginning, in 2021, we started to do the Meitu Xiuxiu, and then the team believed that Meitu Xiuxiu subscription penetration is ultimately 5%, 5 points. Years passed, now it's very close to 5 points, but its tangent is very steep. It's still on high growth. This story is just to share with you something about because in China, if you want to do subscribe products, particularly in video and photo tools, there are no predecessors, including our own judgments may not be just correct. This incorrectness, we often think of it in a positive way because, for example, logically, our products are characterized by different from content-type subscribe products, which after you consume the content, you don't want to pay anymore. Just like after I watch this new video, I may want to unsubscribe before the next product.
Because our products, for our products, if you use the beauty function, like if you edit a photo, and then you may keep editing photos one after another, you keep improving your editing skills. Logically, our payment continuity to some extent may be stronger. The second angle is we see, for example, in AirBrush, its payment penetration rate can be up to 50% in the United States. Of course, this is because U.S. consumers are more willing to pay. They are more willing to pay because culturally, they started from 100% payment and keep going down. They will be higher than China. If you look at the consumer behavior in China, we are more and more comfortable with pay software or subscribe software because we started from pirate software and then we grew on that.
From this angle, I think I have shared something with the investors before. First milestone may be 10%. Just like a few years back when we were doing the Meitu Xiuxiu, our ceiling was 5% in the years past. Then maybe two years ago, when I shared with other people, maybe one or two years from now, we may be achieving 8% or 9%. From the managerial perspective, of course, we want to better serve the users. From the present stage, we do not feel close to that ceiling. We are still moving ahead and we are still growing. All we can do is share with you what we have seen so far. It is very hard to predict some certain figures.
Thank you. Due to time constraint, this is time for the last question. The last question will come from online in text. Mr. Yang Yongyi's question is, surpassing or more popular Adobe, will it be part of your strategic consideration?
Actually, we have seen Adobe a role model that we all learn from. They are actually quite respectable, highly respectable because in the past 30 years of development, they have become a great leader in the photo and video design arena. I would say even a king of that. We also see Adobe, most of their products have a higher bar because they mostly serve professional designers and photographers or medium and big enterprises. We have seen in the recent years, more and more companies, by realizing certain capabilities in Adobe and make it more simple, achieve better user growth and revenue growth, including Meitu Studio and also Canva. They do the poster design like Canva. They do the UI design.
By considering this, we see a trend, which is among popular, among the general public and also professional users. There is actually a big user group among them having a great amount for video photo designs like KOLs or small to micro enterprises or consumers. Of course, there are many terms for them. In our eyes, they are just living users, user cases. By utilizing these video and photo production tools, generate income or generate fans growth, they have a very clear demand, and also they are more willing to pay. Clearly, this productivity tools deployment will be basically serving these intermediary range of users because we think these users will generate a much bigger revenue than the general public and also professional users. We think this might have been misunderstood before, but there is a big room for growth there.
Generative AI plays a very important role in this because generative AI, before its explosive development, our goal was hard to achieve, but now the technological development evolution allows us to use to reduce the bar, reduce the threshold for certain functions of Adobe to make it more user-friendly. Thank you.
Okay, anything to add from Gary?
No from me. Thank you.
Thank you all for your questions and also the executive responses. If you have any other questions regarding the company's performance, please get in touch with our IR team. Now is the immediate Q&A. This section will not be interpreted into English. Please let us know your organization and also your name before you ask your questions.
[Foreign language]