Tencent Holdings Ltd

HKEX:700

Q1-2025 Earnings Call

[Wendy Huang, executive]

Good day and good evening. Thank you for standing by. Welcome to Tencent Holdings Limited 2025 First Quarter Announcement Webinar. I'm Wendy Huang from Tencent IR team. [Operator Instructions] Please be advised that today's webinar is being recorded.

Before we start the presentation, we would like to remind you that it includes forward-looking statements, which are underlined by a number of risks and uncertainties and may not be realized in the future for various reasons. Information about general market conditions is coming from a variety of sources outside of Tencent. This presentation also contains some unaudited non-IFRS financial measures that should be considered in addition to, but not as a substitute for, measures of the group's financial performance prepared in accordance with IFRS. For a detailed discussion of risk factors and non-IFRS measures, please refer to our disclosure requirement -- disclosure documents on the IR section of our website.

Let me introduce the management team on the webinar tonight. Our Chairman and CEO, Pony Ma, will kick off with a short overview. President, Martin Lau; and Chief Strategy Officer, James Mitchell, will provide a business review; and Chief Financial Officer, John Lo, will conclude with financial discussion before we open the floor for questions. I will now pass it to Pony.

[Huateng Ma, executive]

Thank you, Wendy. Good evening, everyone. Thank you, everyone, for joining us. During the first quarter of 2025, our high-quality revenue streams sustained their solid growth trajectory AI capabilities already contribute tangibly to business such as the performance advertising and evergreen games. We also stepped up our spending on new AI opportunities such as Yuanbao application and AI in Weixin. We believe the operating leverage from our existing high-quality revenue streams will help absorb the additional costs associated with these AI-related investments and contribute to healthy financial performance during this investment phase. We expect this strategic AI investment will create value for users and society and generate substantial incremental returns for us over the longer term.

Looking at our financial numbers for the first quarter. Total revenue was RMB 180 billion, up 13% year-on-year. Gross profit was RMB 100 billion, up 20% year-on-year. Non-IFRS operating profit was RMB 69 billion, up 18% year-on-year. And non-IFRS net profit attributable to equity holders was RMB 61 billion, up 22% year-on-year.

Now turning to our key services. For communications and social networks, combined MAU of Weixin and WeChat grew year-on-year and quarter-on-quarter to 1.4 billion. For digital content, Tencent Video and TME both grew their subscription year-on-year. For games, several evergreens achieved a record high in gross receipt during the quarter, and Delta Force has become the highest ranked mobile games by DAU released in China in the last 3 years. For cloud, Tencent Cloud's audio and video solutions ranked first by revenue for the 7th consecutive year in China. I will now hand over to Martin for a business review.

[Chi Ping Lau, executive]

Thank you, Pony, and good evening, and good morning to everybody. For the first quarter of 2025, our total revenue was up 13% year-on-year. VAS represented 51% of our total revenue, within which Social Networks subsegment was 18%, Domestic Games subsegment was 24% and International Games was 9%. Marketing Services was 18% of total revenue, and FinTech and Business Services was 30% of our total revenue. In terms of gross profit, overall gross profit was up 20% year-on-year in the first quarter, surpassing RMB 100 billion for the first time in our history. By segment, VAS gross profit increased 22% year-on-year to RMB 55 billion, representing 55% of our total gross profit. Marketing Services gross profit increased 22% year-on-year to RMB 18 billion, contributing 18% of total gross profit. And FinTech and Business Services gross profit increased 16% year-on-year to RMB 28 billion, contributing 27% of our total gross profit.

If you look at our businesses, our existing businesses continue to generate operating leverage under our high-quality revenue and growth model. This provides us the capacity to reinvest the leverage for the exciting AI opportunity ahead of us. As we have highlighted in the prior quarter earnings call, we are stepping up investments in AI in the form of capital expenditures as well as operating expenses. Some of these GPU and AI investments already generate revenue for us, such as improved ad targeting, which boosts ad revenue; improved content recommendation, which boosts user time spent and thus ad revenue; usage of AI within evergreen games, which boosts user engagement and thus game revenue; and deployment of GPUs and AI across our computing infrastructure, APIs and platform solutions, which generates cloud revenue. For our other GPU and AI investments, which are more long cycle in nature, there's a natural time lag between making the investments and those investments starting to generate significant revenue for us. During this time lag period, we expect the costs of those GPU and AI investments to offset our underlying operating leverage, resulting in a temporary smaller gap between our revenue and operating profit growth rate than we have achieved in recent quarters. That said, we're confident that our stepped-up investment in longer-cycle AI projects will create substantial long-term value for our users, business and shareholders.

Turning to our different business segments. Value-added services revenue was RMB 92 billion, up 17% year-on-year. Social networks revenue was up 7% year-on-year to RMB 33 billion, driven by increased revenue from app-based games, item sales, music subscriptions and mini games platform service fees. Music subscription revenue increased 17% year-on-year as subscribers grew 8% year-on-year to 123 million and as ARPU improved.

TME strengthened content offerings through deepened collaborations with labels and artists and expanded Super VIP membership benefits with enhanced sound quality for speakers, AI-generated audio effect and offline event privileges. Long-form video subscription revenue was flat year-on-year. Video subscribers grew 1% year-on-year to 117 million during a quarter of weaker content lineup on a relative basis. Our self-commissioned drama series, Guardians of Dafeng was the most watched drama series across all long-form video platforms in China during the quarter, demonstrating the ability for us to create great content.

Domestic Games revenue grew by 24% year-on-year against the low base quarter of first quarter of 2024, mainly driven by growth from flagship games, notably Honour of Kings, and Peacekeeper Elite and contributions from recently released games, notably DnF Mobile and Delta Force. International games revenue increased by 23% year-on-year or 22% in constant currency terms, driven by titles, including Brawl Stars, Clash Royale and PUBG Mobile.

Moving on to communications and social network. We implemented a series of upgrades to strengthen the Weixin content ecosystem. For video accounts, we streamlined the live streaming setup process and enhanced recommendation algorithms in order to cultivate and promote more original content. Total user time spent grew rapidly year-on-year in the first quarter.

For Mini Shops, we provided more onboarding support for merchants via greater traffic incentives and lower deposit requirements in order to enrich the range of branded product listings. Mini Shops GMV continued rapid year-on-year growth in the first quarter. For official accounts, we simplified content creation by enabling influencers and other users to set up accounts and publish content via mobile devices in order to expand the creator community and encourage more user-generated content. In addition, we're in the early stages of rolling out AI features within Weixin. Users can now add Yuanbao as a Weixin context for seamless AI interaction within Weixin Chat, providing context-aware responses and facilitating content discovery while leveraging the Weixin ecosystem and the worldwide web. Weixin Search is now starting to include results powered by large language models, including the fast thinking model Hunyuan Turbo S, and the chain of thoughts reasoning models Hunyuan T1 and DeepSeek R1. We provide AI tools so that content creators can generate images matching the text of their official accounts articles and generate video effects for video accounts videos utilizing preset templates. We reduced the Mini Programs development time via an AI coding assistant for creating AI programs that supports natural language prompts and image inputs.

Now with that, I'll pass to James.

[James Mitchell, executive]

Thank you, Martin. For our Domestic Games and starting with our flagship evergreen games. Honour of Kings achieved record high quarterly gross receipts, as Year of the Snake themed outfits and promotion activities targeted to specific user profiles, increased the proportion of players who make purchases within the game. Peacekeeper Elite grew its average DAU and gross receipts in the first quarter through the increasing popularity of its extraction shooter mode complementing the original Battle Royale mode as well as new Tang Dynasty theme setting in items. Several other evergreen games of varying vintages also performed well. For example, CrossFire Mobile released more than 9 years ago, achieved record high gross receipts, benefiting from an enhanced rewards system and a new player versus environment mode. CrossFire Mobile was the third biggest mobile game by gross receipts industry-wide in China in the first quarter behind only the Honour of Kings and Peacekeeper Elite. VALORANT China released less than 2 years ago, more than doubled its gross receipts in average DAU year-on-year in the first quarter, benefiting from two new agents and new weapon skins. VALORANT has become the second most played game in Internet cafes in China behind only League of Legends. And among new games, Delta Force, which we launched in China last September, grew its PC plus mobile user base to a post-launch record level of 12 million peak DAU in April. Delta Force is now the sixth most popular mobile game by DAU industry-wide and the highest DAU new game released in the last 3 years industry-wide.

Among our International Games, PUBG Mobile grew gross receipts by a teens percentage year-on-year, benefiting from Golden Moon-themed events and outfits and collaborations with brands, including Gucci and the K-pop group BABYMONSTER. Call of Duty Mobile's gross receipts grew year-on-year, benefiting from a top-tier operator outfit featuring a Halo and Wings and from Secret Cash, a new engagement system where players unlock rewards by completing tasks and events. For Delta Force outside China, its PC version, which launched in December last year, has been steadily growing its user base, achieving a record level of concurrent users on Steam in May this year. The mobile version of Delta Force is the second most downloaded mobile game in international markets since its launch on April 21, and we intend to launch a console version of Delta Force in the second half of this year.

For Marketing Services, our revenue grew 20% year-on-year to RMB 32 billion, benefiting from higher user engagement, ongoing AI upgrades to our ad platform and a strengthening transaction ecosystem within Weixin. Our Marketing Services revenue increased year-on-year across all major advertiser industry categories. On the ad tech front, we upgraded our advertising platform with enhanced generative AI capabilities such as ad generation and video editing tools to accelerate ad creation and digital human solutions to facilitate live streaming activities for content creators and merchants. We're using large language models to deepen our systems, understanding of merchandise and of user interests across our apps and so deliver better ad recommendations.

By inventory, Video Accounts Marketing Services revenue grew over 60% year-on-year as advertisers increasingly utilized our marketing tools to boost content exposure, promote live streams to drive sales conversion. Mini Programs Marketing Services revenue increased notably year-on-year, capturing closed-loop marketing demand from mini dramas and mini games. And Weixin Search revenue grew rapidly year-on-year, benefiting from increases in commercial query volume and click-through rate.

Looking at FinTech and Business Services, segment revenue was CNY 55 billion, up 5% year-on-year. FinTech services revenue grew by a low single-digit percentage year-on-year benefiting from increases in consumer loan services and wealth management services activities. Our total commercial payment volume has been choppy in recent quarters and declined slightly year-on-year in the first quarter but improved in April. To support our wealth management initiatives, we integrated Hunyuan and DeepSeek large language model capabilities into our financial assistant chatbot. Business Services revenue returned to a teens year-on-year growth rate in the first quarter, benefiting from higher cloud services revenue as well as increased technology service fees generated from rising e-commerce transaction volumes. Business Services gross margin rose year-on-year due to improved efficiencies.

Leading Internet platforms are increasingly adopting our audio and video solutions TDSQL database and general cloud infrastructure as they seek to better manage surges in their usage and enhance their overall operational efficiency. AI-related revenue within Tencent Cloud grew quickly year-on-year, driven by increased customer demand for GPUs, APIs and platform solutions, although constrained by limited GPU availability.

And with that, I'll pass to John to discuss the financial review.

[Shek Hon Lo, executive]

Thank you, James. For the first quarter of 2025, total revenue was CNY 180 billion, up 13% year-on-year. Gross profit was CNY 100.5 billion, up 20% year-on-year. Operating profit was CNY 57.6 billion, up 10% year-on-year. Interest income was CNY 3.7 billion, down 12% year-on-year due to lower interest yields. Finance costs were CNY 3.9 billion, up 37% year-on-year primarily due to foreign exchange losses this quarter compared to gains in the same quarter last year. Share of profit of associates and JV was CNY 4.6 billion compared to RMB 2.2 billion in the same quarter last year. On a non-IFRS basis, share of profit was CNY 7.6 billion, up from CNY 5.5 billion in the same quarter last year.

Income tax expense declined by 3% year-on-year to CNY 13.7 billion, mainly due to a high base from withholding tax provision in the same quarter last year. Domestic corporate income tax expense in the first quarter increased year-on-year. On a non-IFRS basis, diluted EPS was CNY 6.58, up 25% year-on-year, outpacing non-IFRS net profit growth due to reduced share count after our share buybacks. Our weighted average number of shares for calculating Q1 '25 diluted EPS decreased by 2% year-on-year.

On non-IFRS financial figures, operating profit was CNY 69.3 billion, up 18% year-on-year. Net profit attributable to equity holders was CNY 61.3 billion, up 22% year-on-year. The difference in year-on-year growth rates between operating profit and net profit was partially due to higher associate income.

Moving on to gross margins. Overall gross margin was 56%, up 3 percentage points year-on-year and by segment. VAS gross margin was 60%, up 2 percentage points year-on-year primarily driven by a higher mix of high-margin Domestic Games revenue. Marketing Services gross margin was 56%, up 1 percentage point year-on-year, supported by growth in high-margin Video Accounts and Weixin Search revenues. FinTech and Business Services gross margin was 50%, up 5 percentage points year-on-year, mainly due to improved cost efficiency in our FinTech and cloud services.

On first quarter operating expenses, selling and marketing expenses were CNY 7.9 billion, up 4% year-on-year, reflecting higher promotional efforts to support the growth of our AI native applications, partly offset by reduced ad spend on new game launches. Selling and marketing expenses represented 4% of revenues, down from 5% in the same quarter last year. R&D expenses rose by 21% year-on-year to CNY 18.9 billion, reflecting higher staff costs and increased investment to support our AI-related efforts. G&A, excluding R&D expenses increased by 62% year-on-year to CNY 14.7 billion, primarily driven by one-off CNY 4 billion stock-based compensation expense related to the restructuring of an existing commercial arrangement at an overseas subsidiary, which was not included in the calculation of our non-IFRS operating profit and net profit.

At quarter end, we had approximately 109,000 employees, up 4% year-on-year or down 1% quarter-on-quarter. Our non-IFRS operating margin was 39% in Q1, up 2 percentage points year-on-year.

To conclude, I will highlight some key cash flow and balance sheet metrics. Operating CapEx was CNY 26.4 billion, up almost 300% year-on-year, driven by increased investments in GPUs and servers to ramp up our AI capabilities. Nonoperating CapEx was CNY 1.1 billion, down 86% from the high base in the same quarter last year due to acquisition of land use rights. Our total CapEx was CNY 27.5 billion, up 91% year-on-year. Free cash flow was CNY 47.1 billion, down 9% year-on-year, primarily due to increased CapEx spending on GPUs and servers. On a Q-on-Q basis, free cash flow was up 9.5x from a low base in the previous quarter, reflecting seasonally higher games gross receipts and timing difference in settlement of certain crude expenses.

Net cash position was CNY 90.2 billion, up 17% quarter-on-quarter due to seasonally higher games gross receipts, partially offset by cash outflows related to CapEx investment activities as well as our share repurchases. Thank you.

[Wendy Huang, executive]

Thank you, John. We shall now open the floor for questions. [Operator Instructions] So we will take the first question from Jefferies, Thomas Chong.

[Thomas Chong, analyst]

Congratulations on a solid set of results. My first question is about Agentic AI. Can management comment about the outlook and differentiation of Weixin versus peers in the market? In addition, can management also comment about our strategies across various AI business model, i.e., advertising, transactions, GPU rental and subscriptions? And my second question is about e-commerce. Can management comment about the latest organizational adjustment in e-commerce, in particular, the latest progress in Mini Shops? And what should we expect our strategies and KPI going forward? And how should we think about the synergies with our ecosystem and differentiate from other live streaming platforms? On top of that, can you also talk about our preparation for June 18 campaign?

[Chi Ping Lau, executive]

I think you asked many questions. So I'll try to answer the first one more extensively. I think the second one answered pretty just lightly because I think other people may have the same question. So on Agentic AI, it's a very hot concept, right? And the idea is actually, oh, the AI can actually help you to complete a very complicated tasks that involve many different steps as well as the use of tools and maybe in connection with other apps. So if we look at that concept, then there is a general Agentic AI, which everybody can do. Essentially, you create this agent and you go out to the world and try to complete tasks for your user. But at the same time, there's also an Agentic AI that can sit within Weixin and the unique ecosystem of Weixin. And I think those are two different products.

On the general agent AI, I think we are creating that capability within some of our AI native products such as Yuanbao and Ima, over time, as these AIs continue to evolve to increase in terms of their capability. So in the very beginning, these AIs actually answer questions very quickly. So those are the sort of quick response. And then over time, they include -- they start including the chain of thoughts, a long thinking reasoning model and you can answer complicated questions. And over time, the capability can actually allow them to start doing more complicated tasks. So they start evolving to have Agentic capability, and they will be interacting with all other apps and programs and external APIs to help the users. So that would continue to evolve.

And it's not that much different from other Agentic AIs provided by our peers. But on the other hand, right, within the Weixin ecosystem, I think there is the opportunity for us to create a pretty unique Agentic AI that connects with the unique components of the Weixin ecosystem, including the social graph, including the communications and community capability, including the content ecosystem, such as our Official Accounts and Video Accounts and all the millions of Mini Programs that exist within Weixin, which actually sort of gets into all kinds of information as well as transactional and operative capabilities across many different verticals of applications. So I think that would be extremely unique compared to other more general Agentic AIs, and that's sort of a very differentiated product for us.

In terms of your question on AI business models, I think if you look at advertising, it's directly augmented by AI because AI can actually help to improve the targeting capability of our ads. And when we deliver better results, then it translates directly into additional advertising revenue. And I think that is a big opportunity that we are already realizing in our performance ads, but there's more opportunity to develop over time.

Now I think transaction is actually very closely tied to advertising, right? When you have advertising that leads to direct transactions and then advertising value actually goes up significantly. And I think that's the way we are actually also trying to increase our advertising revenue. That's another component and pillar of our advertising revenue growth driver.

GPU rental is sort of directly related to cloud business, and that's more like a reselling business mostly. And to a large extent, right now, we are putting it on a lower priority because -- especially when there's a short supply of GPUs, right, then GPU rental is a lower priority for us. And subscriptions, I think it's not the most likely business model for AIs in China, right? Now everybody is actually providing AIs for free. So the subscription model, which exists outside of China, I think it's not going to be mainstream business model for AI in China.

In terms of e-commerce, I think the latest organizational adjustment is a small one, frankly, because our Weixin e-commerce team was actually incubated within our open platform department. And over time, it has grown in size. And now we are just making the formal change of separating the two so that the e-commerce team has now become an independent department but it's actually still managed by the same manager. So it's not a big change. And I don't think we should be reading too much into it.

[Wendy Huang, executive]

[Operator Instructions] We will take the next question from Ellie Jiang from Macquarie.

[Ellie Jiang, analyst]

I just wanted to really check on our current progress with the Yuanbao integration into Weixin ecosystem. Just wondering, management, if you can shed some light on how do we see the user behavioral trends after we've done such integration? What do we really further expect to see synergies across different initiatives within the Weixin ecosystem, including likes of Mini Shops, Official Accounts, Video Accounts and so on? What is kind of the movement for the next step? Do we kind of expect to see kind of a combination of both the private domains and the public domains to be integrated into the Yuanbao feature itself?

[Chi Ping Lau, executive]

So Ellie, to be honest, right now, this is still at a very early stage of development. And I think we can see users as they use it, right, they use it more. We can see that users are using it to ask questions, to have some dialogues and also to have content directed to the Yuanbao, so that Yuanbao can do summarization and help them to analyze. So these are the initial use cases that we're seeing. And as the users use and interact with Yuanbao, they likely to be sort of using it more. I think over time, absolutely, we are going to provide more linkage of the Weixin ecosystem to the Yuanbao chat companion and -- I think we would be doing different experiments. And maybe in a couple of quarters, there will be more for us to report on a systematic basis. At this point in time, it's a little bit too early to summarize everything in a systematic way for you.

[Ellie Jiang, analyst]

Great. And if I may, just a quick follow-up. As we see potentially more adoption and the usage picking up across both users and the enterprises side, what do we really plan to do in terms of managing, I guess, the balance between pace of investments versus the revenue runway, especially during the early ramp-up stage?

[Chi Ping Lau, executive]

Well, at this stage, I think we want to go full speed in terms of trying to generate demand. And if, frankly, there's too much demand that our GPUs can handle, then we would start thinking about sort of how do we pace it. It will be a nice problem to have. At this point in time, we're not there yet.

[Wendy Huang, executive]

We will take the next question from William Packer from BNP.

[William Packer, analyst]

Congratulations on the strong quarter. It was a very strong quarter for Domestic Games. Could you help us think through the longer-term implications for the growth algorithm in that segment? Should we think of the strong performance as a catch-up demand for the weaker 2021 to '23 period or have recent developments such as the concentration of consumer attention and spend on key global games where Tencent is overweight and faster content release cycles with AI change things underpinning better structural growth?

And as a follow-up, there's been a lot of noise globally recently around regulatory interventions, changing split of economics between app stores and apps. Per management commentary, this is something Tencent has focused on in recent years? Where should we think of China on this journey and the split of value between apps and app stores especially in the video game segment? Are there further shifts here as a value driver for Tencent?

[James Mitchell, executive]

Thank you, William. So on your question about Domestic Games, the first quarter, we did have the benefit of easy comparison against the first quarter of 2024. So that certainly played in our favor during the first quarter in a way that won't be true in every future quarter -- but we do feel we have -- we do believe we have a long runway for our domestic and indeed international game revenue growth looking forward. And there's many reasons, but just to pick on three for now. First of all, we talked extensively this time last year about some of the changes we were making to how we envisage and therefore, how we operate and therefore, who operates our biggest Domestic Games. And you can see that we have made those changes and they're bearing the fruit that we hoped they would bear and we see them bearing more fruit going forward.

A second driver or enabler of that long runway, is the utilization of AI, which we think is particularly beneficial to the big competitive multiplayer games that we've talked about extensively and that represent the majority of our domestic game revenue. And that's the case because while there's many ways that we can and we're starting to deploy AI within games some of the most interesting include using AI to help coach new players, to help accompany existing players, to help prevent cheating and hacking and so forth. And all of those are particularly important within competitive multiplayer games. And then third, our game teams have really strived to align themselves with where the game audience or the future game audience is moving toward in China.

And historically, the proportion of game activity in China that is first-person action games is much smaller than in the rest of the world. And we have taken the view that China would catch up. And if you look at the slide talking about Domestic Games for us this quarter, you can see that most of the games we highlighted are games that are the leaders within that first-person action vertical within China because Tencent is the leader in that vertical within China and the users are indeed moving towards those first-person action games such as Peacekeeper Elite, such as CrossFire Mobile, such as Call of Duty Mobile, such as VALORANT and now such as Delta Force.

And each of those games is growing at faster or slower rates. But in aggregate, because all of them are growing and some of them such as VALORANT and Delta Force are growing very quickly, the overall category is seeing very healthy growth. And that's what underpinned our domestic game revenue growth this quarter, but also provides us with a long-term runway for domestic game revenue growth.

Then moving on to the app stores. It is indeed a time of change in terms of the economics, which previously were unfairly penalizing digital content creators, in particular, game companies, giving a free ride to sort of physical product providers. And in our view, as a company that operates an app store but is also a digital content provider, we're over-rewarding the app stores. And that change has been underway in China for several years. And now the digital content creators, including ourselves, are getting a fairer share of the end user spend.

That change has been much slower or has not happened at all in the Western world until very recently, but there are now a number of legal cases underway. There are now a number of regulatory interventions. And we do believe that while it's difficult to forecast what happens this week, what happens next week, over the longer term that there will be a reset toward a more equitable relationship whereby the digital content providers are able to retain more of the value that they create as opposed to effectively subsidizing a system that benefits the e-commerce and other companies, which aren't paying anything into the whole app store ecosystem. So anyway, we think that's the direction of travel, both in China, where it's more evident but also in the rest of the world, where we think it will become more evident over time.

[Wendy Huang, executive]

We will take the next question from Alicia Yap from Citigroup.

[Alicis a Yap, analyst]

Congrats on the strong results. Number one, I wanted to ask can management share with us over the past 2, 3 months, what are some of the notable user behavior changes that Tencent is able to observe post the deeper integrations of AI into various of your business applications. For example, like in terms of like how user using Yuanbao, whether this AI in Weixin are being noticeable by the business partner and how Tencent expects some of the analysis that you actually observe could help the company in further improving the user value and also future monetization potential?

And then a follow-up on gaming is that I know in addition to some of the bigger games which has been benefiting from the AI integration, will some of the smaller title will also be integrating with AI in the coming quarters that we could see improving monetization or revival of user engagement on some of the mid-tier games as well?

[Chi Ping Lau, executive]

Well, Alicia, I think we actually covered your question earlier, which is it's probably too early to sort of conduct systematic analysis of the behavioral change of users. At this stage, I think we're trying to create functionalities and user experiences that would leverage AI and try to see what may or may not stick with the users. So as I said, right, the users sort of like to ask questions, like to interact with the AI with further follow-up questions. And when we put in various functionalities such as allowing photos to be analyzed and sort of people use it. So there are a lot of functionalities, which right now we have put in, and we're starting to get to see people like them a lot or are not using it that much. So it's a discovery process. And -- but as a whole, right, we can see users are using and interacting with our AI agents more and more as they sort of continue to develop affinity, as they continue to figure out what value add the agents can do for them. So I think the overall usage is actually increasing.

[James Mitchell, executive]

And on your second question, I talked a little bit in response to William's question about how we're deploying what should we say is a game AI into our big competitive games and -- we're at a very early stage of that deployment. And so we think the biggest opportunity for us is continuing down that road. Now at the same time, there's a sort of separate opportunity to deploy more generative AI into more content-driven games, and by doing so, perhaps, one would be able to facilitate initially faster creation of content by the game studio itself. And then over time, some degree of user-generated content, varying quality of experiences and ultimately, dynamically created content where the game itself is spinning up new environments if the players choose to venture off the preset map. But all of that is, I think, to be explored in the years ahead. And as of right now, the sort of bigger, more tangible opportunity is around utilizing what we'll call game artificial intelligence within the big competitive multiplayer games.

[Wendy Huang, executive]

We will move to the next question from Kenneth Fong from UBS.

[Kenneth Fong, analyst]

My first question is on the high-end GPU. Recently, the U.S. has introduced the licensing requirement for the high-end GPU. We understand that our priority is to use this GPU internally first. Could management share your perspective on this situation? And how would it impact, if any, our capital expenditure, AI development as well as product launches? And I have a follow-up on the macro situation. In our press release, we indicated that our consumer loans business has experienced year-on-year revenue growth after a more conservative approach last year. Considering the still uncertain macro environment, can management share some insight on what we have observed on the ground, both on the customer side as well as on the advertiser side?

[Chi Ping Lau, executive]

Well, on the GPU front, it's actually a very dynamic situation, right? So there -- since the last earnings call, we have seen an H20 ban. And then after that, there was the BIS new guidelines that just came in overnight. So it's a very dynamic situation, and we just sort of have to manage the situation, on one end, sort of in a completely compliant way, and on the other end, sort of we try to figure out the right solution for us to make sure that our AI strategy can still be executed.

So the good thing that we are in is that, number one, I think we have a pretty strong stockpile of chips that we acquired previously, and that would be very useful for us in executing our AI strategy. And if you look at the allocation of the usage of these chips, obviously, they will be used for the applications that will generate immediate return for us. So for example, in the advertising business as well as content recommendation product, right? We actually would be using a lot of these GPUs to generate results and generate return for us.

Secondly, in terms of the training of our large language models, they will be of the next priority. And the training actually requires higher-end chips. And the good thing on that front is that over the past few months, right, we start to move off the concept or the belief of the American tech companies, which they call the scaling law, which require continuous expansion of the training cluster. And now we can see even with a smaller cluster, you can actually achieve very good training results. And there's a lot of potential that we can get on the post-training side, which do not necessarily meet very large clusters. So that actually sort of help us to look at our existing inventory of high-end chips and say, we should have enough high-end chips to continue our training of models for a few more generations going forward.

And then the larger need for GPUs are actually sort of around inferences and especially sort of when you see a growth in demand for inference on the user side as well as when we move into the chain of thoughts reasoning model, it actually requires many more tokens to answer a complicated question. And if we move into Agentic AI, right, it requires even more tokens, there's actually a lot of need on the inference side. But on the inference side, there's actually a lot of work that could be done for us to manage the need. One is just sort of leveraging software optimization. I think there's still quite a bit of room for us to keep on improving the inference efficiency, right? So if you can improve inference efficiency 2x, then basically, that means the amount of GPUs get doubled in terms of capacity. So that's actually a very good way of investing our resources to improve on the inference efficiency.

And the other approach is we can customize different sizes of models, especially some applications do not require very large models, right, and we can tailor-made models and distill models so that they can be used for different use cases, and that can actually save on the inference usage of GPUs. And finally, we actually sort of can potentially make use of other chips, compliant chips available in China or available for us to be imported as well as ASICs and GPUs in some cases for smaller models inferences. So I think there are a lot of ways to which we can fulfill the expanding and growing inference needs, and we just need to sort of keep exploring these venues and spend probably more time on the software side rather than just force buying GPUs.

[James Mitchell, executive]

And on your question about our loan facilitation business and what we're observing on the ground. So what we have observed on the ground this year is that the credit quality for the loans that we facilitate through our platform has been gradually but consistently improving. And we think that's partly due to a better selection of borrowers and so forth. But it's partly also due to a more macro development, which is that over the last several years, consumers in China in aggregate have built up unusually large savings balance. All else equal, one would rather be lending into consumer with unusually large savings than the consumer with unusually limited savings. So we are enabling the facilitated loan book in aggregate to grow in a very measured way.

Looking forward, how far can we grow that loan book? We estimate that if we benchmark the loan book facilitated through our fintech platform, with the loan book of our largest peer companies, then we're currently a fraction of their size. And so while we'll continue to do this in a very measured fashion, we do think that there's a long, many year runway for us to continue growing these high-margin loan facilitation revenues.

[Wendy Huang, executive]

We will take the next question from John Choi from Daiwa.

[Hyungwook Choi, analyst]

Congratulations on the strong set of results. I want to have a quick question on the comment about reinvesting into AI. I think you guys mentioned earlier in your opening remarks, smaller gap between revenue and operating profit. Can you kind of elaborate a bit more on this, the magnitude and what kind of extensive period that we're talking about? And like apart from CapEx, what are the -- some of the expense items that we should be expecting? And I have a quick follow-up on the FPS for the domestic game business. As you said, Delta Force did very well this quarter, but also CrossFire Mobile. I just want to understand, are we seeing a structural evolvement in this typical first-person action genre as you pointed out earlier, I mean, across the board, do we expect kind of evergreen status for multiple games within this specific vertical?

[James Mitchell, executive]

Those are interesting questions. And I'll try to answer both of them, but you can prod me if I fail to answer one of them. So on the gap between our profit growth and our revenue growth and the narrowing of that gap, the magnitude, the duration, then we're at uncharted territory, not only for Tencent, but for the whole world in terms of the deployment of artificial intelligence. So I don't have necessarily a very high degree of confidence in these statements. But if you're thinking about measuring the duration, then the past may be the best guide to the future in that Tencent has been through many time periods where we have cultivated a new product toward critical mass and substantial popularity ahead of monetizing that product.

And typically, the duration of those gaps between investment to cultivate versus monetization and revenue generation would be in the sort of 1-year to 2-year time range. So obviously, it will depend on what our peer companies do in China, obviously, will depend on consumer habits, on advertiser habits. But I think that's a reasonable time frame to think about. In terms of magnitude, I won't go beyond what we said earlier, which is referring to a narrowing. So we don't expect the delta between revenue growth and operating profit growth that we experienced this quarter to continue. There will be a narrowing. But on the other hand, we don't expect our operating leverage to turn negative either.

In terms of what costs other than CapEx or really depreciation could cause that narrowing, then CapEx depreciation is, by far, the most important. We do have some incremental marketing expenses for Yuanbao, although not so much for AI within Weixin. And then we referenced the fact that engineers with expertise in AI are expensive, but that's more of a sort of mix comment rather than an aggregate headcount comment. We don't see a step-up in headcount. We'll continue to manage headcount closely, but we observe that engineers with that AI expertise are rightly well paid. So that's on your first set of questions.

On your second set of questions around the popularity of first-person action games in China. Then I think two broad comments. One is, as I mentioned, to a large extent, this represents China gamer behavior converging with long-standing gamer behavior in the rest of the world. That in the rest of the world, the first-person action genre is the sort of super genre within games. It's the equivalent of drama series on television. It represents 40% to 50% of gamer time spent and gamer revenue generation. In China, historically, it had been 10% to 20%. But what we see when we look at the different cohorts generations of gamers is that younger gamers in China are disproportionately attracted to first-person action games, and first-person action games like VALORANT that very directly appeal to, let's say, the 20- to 30-year old demographic are doing particularly well. So that's one observation.

The second observation is that while our newer first-person action games are doing well, so a number of our incumbent games, including Peacekeeper Elite, Call of Duty Mobile, CrossFire Mobile. And I think there's a number of reasons for that. But one is that you may remember a few years ago, the sort of unlock of the Battle Royale mode, which in the West resulted in Fortnite becoming very popular. Call of Duty: Warzone and PUBG all becoming popular. And that was a sort of step up, which meant that the overall genre expanded.

In China, we saw that step up from Battle Royale. And then recently, we have seen another step up through the popularity of so-called extraction shooter. And so I mentioned that the increase in usage for Peacekeeper Elite, which is the biggest first-person action game in China has been largely around the extraction shooter mode. For Delta Force, most of the high retention users play extraction shooter mode. So we believe that every few years, there are these new modalities of gameplay within first-person action that generally expand the overall audience more than they cannibalize and that's what's underway today. So we're very happy with the progress of a multiple of our first-person action games in China.

[Wendy Huang, executive]

I will take the next question from Jialong from Nomura.

[Jialong Shi, analyst]

I have a few questions for ads business. Each quarter, we hear management mention AI tech as one of the drivers for your ad business. So I just wonder how much more room of growth will likely come from the deployment of AI tech into your ad business? And separately, what is the latest ad load and eCPM for your Video Account ads? And regarding Weixin Search, so I just wonder what is the latest search query market share for your Weixin Search? And how much of the ad revenue is now contributed by Weixin Search? After you guys add the AI search feature into Weixin, do you see a migration of user search activity from traditional search to AI search?

[James Mitchell, executive]

Okay. I'll try to address some or all of those as best I can. But in terms of how much of the -- how much further AI benefits can boost our advertising revenue, it's really a sort of global industry question. If I knew the answer to that question, then it would be very helpful for understanding Meta and Google and a whole range of companies beyond Tencent. And if I try to sort of simplify the framework, then a big part, not all, but a big part of the uplift that AI is providing to advertising revenue today can be quantified in the form of the click-through rate on ads. And historically, banner ads achieved a roughly 0.1% click-through rate. Feed ads achieved roughly 1.0% click-through rate. With the benefit of AI, we have seen that the click-through rate on certain ad inventories can improve toward 3.0%, for example.

And then the question is, what's the upper limit on that click-through rate. And at this point, no one knows the answer because it almost becomes philosophical if you had complete information or insight into a consumer if you had the ability to infer what the consumer wants or the consumer given their prior behaviors should want and then deliver an ultra targeted ad to that consumer, then it's very hard to say that the upper limit should be X percent rather than Y percent. Of course, there's other ways that AI benefits advertising revenue. We can use AI to target more appealing content to the consumer, which means they spend more time in the feed, which means they then view more ads, but I think that ad click-through rate is perhaps the most important.

In terms of the Video Accounts ad load, we have kept it fairly stable for the last I think, 6 months now, and it's still around 3% to 4%. In terms of the Video Account CPM, it remains a very good number, partly because of the high click-through rate that we're achieving, partly because there is advertiser competition to have access to the inventory that's available.

In terms of the Weixin Search query share, we haven't disclosed that number, but it's on a good upward trend. And then in terms of how users are responding to AI search. I think it's -- one of my colleagues may have a more nuanced answer. But a, it's early days. And b, I think the whole concept of AI search is a little bit slippery in that where does one draw the boundary between consumers typing a prompt into a pure-play AI experience versus typing a query into a search engine that could be answered by the traditional algorithm or could be answered by a large language model powered result. The two are sort of blending together. And we think that as a relative newcomer and new entrant into the search industry, that's good for us because it's a chance for us to build user share and ultimately build revenue share as well.

[Wendy Huang, executive]

We will take the next question from Alex Yao from JPMorgan.

[Alex Yao, analyst]

Congrats on a solid quarter. So the first question is on advertising performance, which accelerated from 17% in Q4 to 20% in Q1. So we want to understand to what extent is such a strong performance or acceleration driven by macro recovery versus technology improvement versus inventory release? Based on the prior commentary, I guess, inventory release should be the smallest impact for one. So even if accurately quantify, the impact is difficult. Would you be able to give us a rank of the impact on the revenue acceleration? And then another question is on the AI product format. So you guys have been actively acquiring user adoption for Yuanbao for a while. Can you share us with the user feedback, retention, et cetera, on those users who are actively using Yuanbao. More importantly, in terms of the product, how do you see the future development of chatbot AI application? Is it likely the ultimate form of generative AI?

[James Mitchell, executive]

So why don't I answer the advertising question, but not the AI question. I think on the advertising, Alex, I wouldn't read too much into the quarter-by-quarter fluctuations in the advertising revenue growth rate. It's been within a band. And frankly, the first quarter was at the upper end of that band. And we wouldn't even necessarily desire an acceleration. For us, it's more important to really preserve a very long runway for sustaining that band for many -- not quarters, but many years to come.

And it may be that if we see ourselves being too close to the top of the band, then we take more time in terms of increasing the ad load for our existing products or in terms of deploying ads against our newer AI products, and really focus on optimizing for user time spent, user experience rather than trying to have -- rather than trying to break out at the top of the band, but then the band exists for a shorter period of time than we would like it to. So I think that's the broad general comment.

In terms of your question and sub-question about the specific drivers for the fluctuation upward in the first quarter. It's certainly not inventory because we haven't increased inventory at all. I don't know whether it's macro or not, we probably need to wait and see what other companies report. And then the cadence of ad tech improvements around AI is not something that we try to sort of overly plan on a quarter-to-quarter basis. So again, in sum, I'd just say that this was a normal sort of fluctuation within a band to the upper end of the band. And I wouldn't read anything -- but I wouldn't read too much into it, and I wouldn't over-extrapolate it forward either.

[Chi Ping Lau, executive]

There's an organic growth of traffic, too. So that probably is a pretty important driver.

[James Mitchell, executive]

Yes, the traffic growth and the ad tech improvements are why the band is in the teens rather than the single-digit positive or the single-digit negative range. And then the fluctuations within that band, I think that there's some quarterly noise associated with them. But the bigger picture, as Martin points out, is that we have traffic growth, particularly for our most desirable advertising inventories, meaning the Video Accounts, the Mini Programs, the Official Accounts, Weixin Search, and we're deploying ad tech around -- we're deploying AI to improve the return to advertisers around those inventories.

[Chi Ping Lau, executive]

Yes. Well, just to add to that, right, the way we look at our advertising business is that there is a very long runway. And all we do is try to sort of extend the runway rather than to sort of manage the growth rate on a quarter-to-quarter basis, right? If we can actually keep on improving our ad tech that extends the runway, if we can keep on increasing the traffic, right? And a lot of it is actually driven by AI as well, right? Because when we have all these users coming to our Video Accounts, we still have a lot of people who come in and then watch it for a very short period of time and then leave, right? If we can actually figure out the right content to push to them through AI, then it can actually increase on the user time by pretty significant level, right?

Then when we have that, we cannot necessarily increase our ad load, we can just rely on the natural growth of the traffic and that would sort of carry the revenue growth, and we can reserve the ad load for a later time so that our runway is extended again. And at the same time, then by having less advertising, we can actually make our product experience better, and that's another sort of competitive advantage, and that forms a virtuous cycle. So I think that's the way we think about our advertising business.

And over time, if we can actually build stronger transaction ecosystem within our Weixin ecosystem, right? So more and more merchants can actually advertise. And once the click-through rate happen, then it leads transaction very easily, then basically the propensity for advertisers to spend with us would go up and the value of each click would go up, right? So there are a lot of these systemic things that we're doing in order to extend the runway of our advertising business, which we believe, if you take a high-level view, right, there is a very long runway, and we are continuously working to extend that.

On AI, I think -- right now, we -- just within the past quarter, right, we have actually grew our Yuanbao user base by a lot. And we spent a lot of time in making sure that we can retain them with that, which I think we have proven ourselves to be able to retain the users and provide them with good enough experience so that they keep coming back and they keep using Yuanbao more. I think when you talk about the new product form factor, right, it really -- chat is the predominant 1 at this point in time. But then when we roll out more functionalities then the user interface and the format may be different, right? When you try to execute a complicated task for users, then it may become a project rather than a chat, and you may have to have that task stay there for a longer time.

So we would be customizing the form factor, the user experience according to sort of what kind of functionalities that we are providing. But I think the next stage for Yuanbao is that -- the current state is that we are at scale. We have large enough number of users, and we have got to a stage that we can actually retain these users. The second stage is that we still have a lot of functionalities that we can add, right, which will help us to retain and activate users better and also to help us to attract new users. So that will be the next stage of the development for our AI product.

And then the next thing, right, is really to leverage our unique ecosystem, be it social, be it content, be it mini programs in order for us to provide something that's very well differentiated from the other AI products and AI chatbots in the market. And of course, while we do all these, right, we will continue to upgrade our Hunyuan model so that we can make the dual-model strategy even stronger. So I think these are the kind of things that we'll be doing.

[Wendy Huang, executive]

We will take the last question from Ronald Keung from Goldman Sachs.

[Ronald Keung, analyst]

One -- two questions. One is on our WeChat e-commerce strategy. Just want to hear what are the key KPIs that we see? And is there a plan or a timeline for maybe shelf-based offering versus today, which is live stream search as the main formats? Second is on payment. I see the management noted a slight decline in the payment volumes in 1Q. So are these more macro related? Or is it a payment strategy, which -- because you've also mentioned some improvement in April?

[Chi Ping Lau, executive]

In terms of e-commerce strategy, I don't think we operate by setting prescriptive KPIs. This is a very long-term project with a very long runway. Again, right, I keep on sort of going back to the same term runway because we felt it's much more important for us to make sure that the business can keep growing for a very long time rather than sort of go for the sprint. So that you have the KPI and then sort of you need it.

So in terms of how we build the runway, we have a number of initiatives that's going on. The number one priority is really sort of improving the basic shopping experience for consumers, such as we need to have higher-quality products. We need to return policy that's convenient for them. We need to have a good customer service process. We need to have good pricing for these consumers. So that's one.

Number two is that we want to get more supply of good products, right, supply of good brands and motivated merchants into our ecosystem so that the supply side can keep improving. And the third one is really to get more traffic to the merchants and provide more chances for consumers to interact with products and with the merchants, right?

And the first step is really growing traffic through live broadcast. I think that's one way through to which we can really convert more sales, and it's actually well proven. So this is something which other platforms have already done, and I think this is more like a no regret things that we can do. But at the same time, over time, we actually want to have something unique for the Weixin ecosystem, right, which is we can connect users to products through our many different modules within the Weixin ecosystem such as Mini Programs. We know that there are a lot of users who are already buying products through Mini Programs. And if we can actually convert some of these transactions to Mini Shops, and because of this better infrastructure of supporting e-commerce transactions within Mini Shops, we can actually magnify the transactions, then it will be helpful for users, it would be helpful for the merchants. There's social component, there's communication component, there's connection to moments, there's connection to search.

Like on the social side, it's -- we said the gifting part is one functionality that we added, and there will be more to be added. And connection to Official Accounts. So there are a lot of potential connections that we can add. And over time, these connections would start delivering more transactions to the merchants in a very unique way. So that's sort of the kind of things that we're doing in order to grow our Mini Shops and Weixin e-commerce ecosystem over the long term. And I'm happy to say, even as these are the things that we're doing in the first inning, we continue to see very strong growth in our Mini Shops transaction GMV.

[James Mitchell, executive]

And on your question about commercial payment volume, then our market share could fluctuate depending on the level of subsidies, depending on how aggressive we are in accepting credit card funded transactions and so forth. But in reality, we don't make dramatic changes on a month-to-month basis. And we didn't dial up our level of competitive intensity moving from the first quarter to April. And so I think the big picture is that consumer spending has been fluctuating in the last several quarters, but the most recent data point is a fluctuation upward. And so it is possible that fluctuation upward is reflective of overall consumer confidence and consumption activity stabilizing. And we hope that over time, that stabilization becomes a sustained upturn. But of course, time will tell.

[Chi Ping Lau, executive]

Well, if I can answer it sort of slightly -- adding more color, right? I think when we look at the first quarter, we see the number of transactions continue to go up, but then the pricing actually goes down, right? And our hypothesis is that it could be a little bit of the last leg of the supply side competition as the demand side start to warm up. And I think the April trend somewhat confirms it. But that's really before the tariffs start kicking in, right? So without the tariff, I think we clearly see that the consumption pattern has bottomed out and start to slowly recover.

But now with sort of the interaction with the tariff, I think we just have to see what will be happening in the next quarter. And it's very dynamic right now. There's tariff but then sort of what's the extent of the tariff. And then I think the government actually sort of is very supportive with the stimulus. And I think the Chinese government still have a lot of room for rolling out more stimulus. And now there's sort of a little bit of a choose on the tariff side. So I think tariff and the interplay of that, the impact of it on the economy and then the counteract effect of the stimulus would be the two factors to watch.

[Wendy Huang, executive]

Thank you. We are now ending the webinar. Thank you all for joining our results today. If you wish to check out our press release and other financial information, please visit the IR section of our company website at www.tencent.com. A replay of this webinar will also soon be available. Thank you, and see you next quarter.