Tencent Holdings Limited (700) Earnings Call Transcript & Summary

May 14, 2025

Hong Kong Stock Exchange HK Communication Services Interactive Media and Services earnings 82 min

Earnings Call Speaker Segments

Wendy Huang

executive
#1

Hello and good evening, everyone. Thanks for holding. You're tuned in to the Tencent Holdings Limited 2025 First Quarter Announcement Webinar. This is Wendy Huang from Tencent's IR team. Right now, every participant is in listen-only mode. Once management finishes its presentation, we'll move into a question-and-answer session. <<operator_script>> Also, please note that this webinar today is being recorded. Ahead of beginning the presentation, we'd like to remind everyone that it contains forward-looking statements, which are subject to a number of risks and uncertainties and might not come to pass in the future for various reasons. Details regarding general market conditions come from a range of sources outside of Tencent. This presentation also includes certain unaudited, non-IFRS financial measures, which should be viewed alongside, and not as a replacement for, measures of the group's financial performance prepared in accordance with IFRS. For a full discussion of risk factors and non-IFRS measures, please see our disclosure documents in the IR section of our website. Allow me to introduce the management team joining the webinar tonight. Our Chairman and CEO, Pony Ma, will begin with a brief overview. President, Martin Lau, and Chief Strategy Officer, James Mitchell, will then give a business review, and Chief Financial Officer, John Lo, will wrap up with a financial discussion before we open the floor for questions.

Wendy Huang

executive
#2

I'll hand things over now to Pony.

Pony Ma

executive
#3

Thank you, Wendy. Good evening, everyone. Thanks to all of you for joining us. In the first quarter of 2025, our high-quality revenue streams continued their solid growth trajectory. AI capabilities are already making tangible contributions to businesses such as performance, advertising and evergreen games. We also increased our spending on new AI opportunities such as the 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 RMB180 billion, up 13% year-on-year. Gross profit was RMB100 billion, up 20% year-on-year. Non-IFRS operating profit was RMB69 billion, up 18% year-on-year. And non-IFRS net profit attributable to equity holders was RMB61 billion, up 22% year-on-year. Now turning to our key services. For communication 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 growth receipt during the quarter, and Delta Force has become the highest ranked mobile games by DAU released in China in the last three years.

Pony Ma

executive
#4

On the cloud side, Tencent Cloud's audio and video solutions have held the number one spot by revenue in China for 7 years running. At this point I'll pass things over to Martin for a business perspective.

Martin Lau

analyst
#5

Thanks, Pony. And good evening — or good morning, depending on where everyone is. In Q1 2025, total revenue rose 13% compared with a year earlier. VAS made up 51% of overall revenue, and inside that, social network accounted for 18%, domestic games for 24%, and international games for 9%. Marketing services contributed 18% of total revenue, while fintech and business services accounted for 30% of it. Turning to gross profit: overall gross profit climbed 20% year-on-year in Q1, exceeding RMB100 billion for the first time ever for us. Broken out by segment, VAS gross profit grew 22% year-on-year to RMB55 billion, which was 55% of our total gross profit; Marketing Services gross profit rose 22% year-on-year to RMB18 billion, adding 18% of total gross profit; and Fintech and Business Services gross profit was up 16% year-on-year at RMB28 billion, contributing 27% of total gross profit. Looking across our businesses, the ones we already have keep producing operating leverage thanks to our high-quality revenue and growth model. That gives us room to plow that leverage back into the exciting AI opportunity in front of us. As we noted on last quarter's earnings call, we're ramping up AI investments through capital expenditures and operating expenses alike.

Martin Lau

analyst
#6

A portion of these GPU and AI investments is already producing revenue for us. Examples include better ad targeting, which lifts ad revenue, and improved content recommendation, which increases user time spent and therefore ad revenue. We also see it in the use of AI within evergreen games, which raises user engagement and thereby game revenue, and in the deployment of GPUs and AI across our computing infrastructure, APIs, and platform solutions, which brings in cloud revenue. As for our other GPU and AI investments, the ones that are longer-cycle by nature, there is naturally a time lag between when we make those investments and when they begin contributing significant revenue. While that lag persists, we expect the costs associated with those GPU and AI investments to offset our underlying operating leverage, which would produce a temporarily narrower gap between our revenue growth rate and our operating profit growth rate than what we delivered in recent quarters. With that said, we have confidence that our stepped-up investment in longer-cycle AI projects will generate substantial long-term value for our users, our business, and our shareholders. Now turning to our business segments. Value-added Services revenue came in at RMB92 billion, an increase of 17% year-on-year. Social Network's revenue rose 7% year-on-year to RMB33 billion, driven by higher revenue from app-based games item sales, music subscriptions, and Mini Games platform service fees.

Martin Lau

analyst
#7

Revenue from music subscriptions rose 17% versus a year earlier, driven by an 8% year-on-year increase in subscribers to RMB123 million along with higher ARPU. TME deepened its collaborations with labels and artists to bolster content offerings, and it broadened Super VIP membership perks with better speaker sound quality, AI-generated audio effects, and privileges at offline events. Long-form video subscription revenue was flat compared with a year earlier. Video subscribers were up 1% year-on-year to RMB117 million in a quarter whose content slate was relatively weaker. Guardians of Dafeng, our self-commissioned drama series, ranked as the most-watched drama across every long-form video platform in China during the quarter, which shows our capacity to produce excellent content. Domestic games revenue climbed 24% year-on-year, coming off the low base of the first quarter of 2024, mainly on growth from flagship titles, notably Honor of Kings and Peacekeeper Elite, plus contributions from recently launched games, notably DnF Mobile and Delta Force. International Games revenue was up 23% year-on-year, or 22% in constant currency terms, fueled by titles including Brawl Stars, Clash Royale and PUBG Mobile. Turning to Communications and Social Networks: we rolled out a set of upgrades to reinforce the Weixin content ecosystem. On video accounts, we simplified the live streaming setup process and improved recommendation algorithms so as to nurture and promote more original content.

Martin Lau

analyst
#8

User time overall rose sharply compared with a year earlier during the first quarter. With regard to Mini Shops, we offered merchants additional onboarding assistance through bigger traffic incentives and reduced deposit thresholds, aiming to broaden the selection of branded product listings. Mini Shops GMV kept up its swift year-on-year expansion in Q1. Turning to Official Accounts, we made content creation simpler by letting influencers and other users establish accounts and post content through mobile devices, which helps grow the creator community and spurs more user generated content. Beyond that, we are at an early stage of rolling out AI features inside Weixin. Users can now add Yuanbao as a Weixin context for seamless AI interaction within Weixin Chat, delivering context-aware responses and making content discovery easier while tapping the Weixin ecosystem and the World Wide Web. Weixin Search is now beginning to incorporate results powered by large language models, among them the fast thinking model HunYuan Turbo S along with the Chain of Thought reasoning models HunYuan T1 and DeepSeek R1. We supply AI tools enabling content creators to produce images that match the text of their official accounts articles and to generate video effects for video accounts videos using preset templates. We shortened Mini Programs development time with an AI coding assistant for building AI programs that accept natural language prompts and image inputs.

Martin Lau

analyst
#9

With that said, I'll hand things over to James.

James Mitchell

executive
#10

Thank you, Martin. Turning to our home-market titles, beginning with the flagship Evergreen lineup: Honor of Kings posted its highest quarterly gross receipts on record, driven by Year of the Snake-themed outfits and promotional campaigns aimed at specific user profiles, which lifted the share of players spending money inside the game. Peacekeeper Elite raised both average DAU and gross receipts during the first quarter, helped by growing interest in its extraction shooter mode, which complements the original battle royale mode along with a new Tang Dynasty-themed setting and items. A number of other evergreen titles from various launch years also turned in strong results. Crossfire Mobile, for instance, which came out over nine years ago, hit record-high gross receipts thanks to a better reward system and a new player versus environment mode. In China's overall mobile industry during the first quarter, Crossfire Mobile ranked third by gross receipts gains, trailing only Honor of Kings and Peacekeeper Elite. Valorant China, on the market for under two years, more than doubled its gross receipts and average DAU year-on-year in the first quarter, lifted by two new agents and fresh weapon skins. Among games played in Chinese Internet cafes, Valorant now sits second, behind only League of Legends. As for newer titles, Delta Force, which relaunched in China last September, pushed its combined PC and mobile user base to a post-launch peak of RMB12 million DAU in April.

James Mitchell

executive
#11

Delta Force now ranks as the sixth most popular mobile game across the industry by DAU, and it's the highest-DAU new game launched industry-wide in the past 3 years. On the international side, PUBG Mobile's gross receipts rose by a teens percentage year-on-year, lifted by Golden Moon-themed events and outfits plus collaborations with brands such as Bugatti and the K-Pop group Baby Monster. Gross receipts for Call of Duty Mobile climbed year-on-year, helped by top-tier operator outfits featuring halo and wings and by Secret Cash, a new engagement system in which players unlock rewards by completing tasks and events. Delta Force Outside China — its PC version, launched in December last year — has seen its user base grow steadily, hitting a record level of concurrent users on Steam in May this year. Since launching on April 21st, the mobile version of Delta Force has become the second most downloaded mobile game in international markets, and we plan to launch a console version of Delta Force in the second half of this year. Turning to marketing services, revenue rose 20% year-on-year to RMB32 billion, driven by 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.

James Mitchell

executive
#12

Now regarding ad tech, we've enhanced our advertising platform by adding improved Generative AI features, including tools for ad generation and video editing that speed up creative production, plus digital human solutions that help content creators and merchants with live streaming. Large language models are being used to strengthen how well our system comprehends merchandise as well as user interests throughout our apps, which lets us serve better ad recommendations. Breaking it down by inventory, Video Accounts marketing services revenue rose more than 60% year-on-year, driven by advertisers making greater use of our marketing tools to increase content exposure, promote live streams, and boost sales conversion. Mini Programs marketing services revenue posted notable year-on-year growth, capturing closed-loop marketing demand coming from mini dramas and mini games. Weixin Search revenue also climbed quickly year on year, helped by higher commercial query volume and click-through rate. Turning to fintech and business services, segment revenue came to RMB55 billion, an increase of 5% year-on-year. Fintech services revenue rose by a low single-digit percentage year-on-year, supported by growth in consumer loan services and wealth management services activities. Our total commercial payment volume has been uneven across recent quarters; it fell slightly year-on-year in the first quarter but showed improvement in April. To back our wealth management efforts, we brought HunYuan and DeepSeek large language model capabilities into our financial assistant chatbot.

James Mitchell

executive
#13

In the first quarter, business services revenue went back to a year-on-year growth rate in the teens, helped by higher cloud services revenue and by greater technology service fees brought in as e-commerce transaction volumes rose. Business Services gross margin increased year-on-year, reflecting improved efficiencies. More and more leading Internet platforms are adopting our audio and video solutions, the TDSQL database, and general cloud infrastructure, as they look to better handle usage surges and improve their overall operational efficiency. Within Tencent Cloud, AI-related revenue rose quickly year-on-year, driven by greater customer demand for GPUs, APIs, and platform solutions, though it was constrained by limited GPU availability. With that, I'll hand things over to John for the financial review.

John Lo

executive
#14

Thanks, James. For Q1 2025, our total revenue came in at RMB180 billion, marking a 13% increase compared to a year earlier. Gross profit reached RMB100.5 billion, which represents a 20% rise year-on-year. Operating profit stood at RMB57.6 billion, up 10% versus the prior-year period. Interest income amounted to RMB3.7 billion, a decline of 12% year-on-year, driven by lower interest yields. Finance costs totaled RMB3.9 billion, climbing 37% year-on-year, largely because of foreign exchange losses this quarter versus gains in last year's same quarter. Our share of profit from associates and JVs was RMB4.6 billion, compared with RMB2.2 billion in the equivalent quarter last year. Under non-IFRS measures, that profit share was RMB7.6 billion, an increase from RMB5.5 billion a year ago. Income tax expense fell 3% year-on-year to RMB13.7 billion, chiefly reflecting a high base tied to a withholding tax provision in the same period last year. Domestic corporate income tax expense rose year-on-year during the first quarter. Diluted EPS on a non-IFRS basis was RMB6.58, up 25% year-on-year, exceeding non-IFRS net profit growth thanks to a smaller share count following our buybacks. The weighted average share count used for Q1 '25 diluted EPS dropped 2% year-on-year. Turning to non-IFRS figures, operating profit was RMB69.3 billion, an 18% year-on-year gain.

John Lo

executive
#15

Net profit attributable to equity holders came in at RMB61.3 billion, marking a 22% increase year-on-year. The gap between the year-on-year growth rates for operating profit and net profit stemmed partly from higher associated income. Turning to gross margins. The overall gross margin stood at 56%, rising 3 percentage points year-on-year. Breaking that down by segment, VAS gross margin reached 60%, climbing 2 percentage points year-on-year, largely driven by a richer mix of high-margin domestic games revenue. Marketing services gross margin came to 56%, 1 percentage point higher year-on-year, helped along by growth in high-margin video accounts and Weixin Search revenues. Fintech and business services gross margin was 50%, an increase of 5 percentage points year-on-year, chiefly on the back of better cost efficiency in fintech and cross services. Regarding first quarter operating expenses, selling and marketing expenses totaled RMB7.9 billion, up 4% year-on-year, which reflected greater promotional efforts supporting growth of our AI native applications, partly offset by lower ad spend on new game launches. Selling and marketing expenses accounted for 4% of revenues, down from 5% in the same quarter last year. R&D expenses climbed 21% year-on-year to RMB18.9 billion, reflecting higher staff costs and increased investment backing our AI-related efforts. G&A excluding R&D expenses grew 62% year-on-year to RMB14.7 billion, primarily driven by RMB1 billion of RMB4 billion (ph) stock-based compensation expense tied 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.

John Lo

executive
#16

By the close of the quarter, our headcount stood at roughly 109,000 employees — a 4% increase compared with a year earlier, though 1% lower than the prior quarter. For Q1, our non-IFRS operating margin came in at 39%, which is 2 percentage points higher year-on-year. To wrap up, I'll walk through a few key metrics on cash flow and the balance sheet. Operating CapEx reached RMB26.4 billion, nearly 300% higher than a year ago, reflecting greater spending on GPUs and servers as we scale our AI capabilities. Non-operating CapEx was RMB1.1 billion, an 86% decline from last year's elevated base in the same quarter, which stemmed from land use rights acquisitions. Total CapEx amounted to RMB27.5 billion, a 91% rise year-on-year. Free cash flow came to RMB47.1 billion, 9% lower than a year earlier, mainly because of heavier CapEx outlays on GPUs and servers. Sequentially, free cash flow grew 9.5 times off a low base in the preceding quarter, mirroring seasonally stronger gains, gross receipts, and timing differences in settling certain accrued expenses. Our net cash position was RMB90.2 billion, 17% higher quarter-on-quarter thanks to seasonally stronger gains gross receipts, partly offset by cash outflows tied to CapEx investment activities and our share repurchases. Thank you.

Wendy Huang

executive
#17

Up first, we have Thomas Chong from Jefferies. Thomas, go ahead — your line is live.

Thomas Chong

analyst
#18

Good evening, and thanks to management for fielding my questions—congratulations on a strong set of results. First, on Agentic AI: could management speak to the outlook and to how our pricing is differentiated relative to peers in the market? Also, would management comment on our strategies across the various AI business models—advertising, transactions, GPU rental, and subscriptions? My second question concerns e-commerce: can management address the most recent organizational adjustment there, especially the latest progress on Mini Shops, and what we should expect in terms of our strategies and KPIs going forward? And how should we view the synergies with our ecosystem and the differentiation against modern live streaming platforms? Beyond that, could you also discuss our preparation for the June 18th campaign? Thank you.

Pony Ma

executive
#19

You posed quite a few questions, so I'll go deeper on the first one. The second one got a fairly brief response, I believe, since others might share that same question. Now, regarding Agentic AI — it's an extremely popular idea at the moment, and the concept is that AI can assist you in finishing highly complex tasks involving numerous steps, plus tool usage and possibly linking up with other applications. So when we consider that concept, there's a general Agentic AI that anyone can build. Basically, you build this agent, and it goes out into the world attempting to finish tasks on behalf of your user. But there's also, at the same time, an Agentic AI capable of residing inside Weixin and its distinctive ecosystem. And I'd say those constitute two separate products. As for the general Agentic AI, I believe we're building that capability into some of our AI native products, including Yuanbao and Ima, gradually, as these AIs keep evolving and their capabilities keep growing. So at the very start, these AIs respond to questions, perhaps very fast.

Pony Ma

executive
#20

So those are kind of the quick replies. Then as time goes on, they begin to incorporate the chain of thought, extended thinking, and reasoning model, and you're able to answer complex questions. And as time goes on, the capability can actually let them begin performing more complex tasks. So they start to evolve and gain agentic capability, and they'll be interacting with all the other apps and programs and external APIs to assist users. So that will keep evolving, and it isn't all that different from other Agentic AIs that our peers offer. But on the other hand, inside the Weixin ecosystem, I believe there's an opportunity for us to build a fairly unique Agentic AI that links up with the distinctive components of the Weixin ecosystem, including the social graph, including 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 kind of reaches into all sorts of information as well as transactional and operative capabilities across many different verticals of applications. So I think that would be extremely unique relative to other, more general Agentic AIs. And that's kind of a very differentiated product for us.

Pony Ma

executive
#21

On your question about AI business models, I’d say advertising is directly augmented by AI, since AI can actually help improve how well we target our ads. And when we deliver better results, that translates straight into additional advertising revenue. I think that’s a big opportunity we’re 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. 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’re actually also trying to increase our advertising revenue. That’s another component and pillar of our advertising revenue growth driver. GPU rental is directly related to cloud business, and that’s more like a reselling business mostly. And to a large extent right now, we’re putting it on a lower priority because, especially when there’s a short supply of GPUs, then GPU rental is 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, AI is for free.

Pony Ma

executive
#22

Regarding the subscription model that operates outside China, I don't believe it will become the mainstream business model for AI within China. When it comes to e-commerce, frankly, the most recent organizational adjustment is a minor one, because our Weixin e-commerce team was in fact incubated inside our Open Platform department, and as time passed it grew in size, and now we are simply making the formal change of separating the two, so that the e-commerce team has now become an independent department, though it is actually still managed by the same manager. So this is not a big change, and I don't think we should be reading too much into it.

Wendy Huang

executive
#23

Thanks. Just a quick reminder to keep it to 1 main question plus 1 follow-up. Our next question comes from Adi Yao (ph) at Macquarie. Adi, your line should be open.

Adi Yao

analyst
#24

Thank you. Thanks to management, and thanks to Wendy for taking my question. I really just wanted to check on where we currently stand with the Yuanbao integration into the Weixin ecosystem. I was wondering, management, whether you could shed some light on how we should view user behavioral trends now that this integration has been done. What more do we actually expect to see in terms of synergies across the different initiatives within the Weixin ecosystem, including things like Mini Shops, official accounts, video accounts, and so on? What are the next-step movements, so to speak? Do we expect to see some kind of combination of both private domains and public domains integrated into the Yuanbao feature itself? Thank you.

Pony Ma

executive
#25

Alicia, honestly speaking, we're still at a really early development phase right now, and I think as users engage with it, they end up using it more. We can observe people turning to it for asking questions, having some dialogues, and also for sending content over to Yuanbao so that Yuanbao can summarize it and help them analyze it. Those are the first use cases we're seeing. And as users engage and interact with Yuanbao, they're likely to keep using it more, sort of. Over time, definitely, we're going to link the Weixin ecosystem more with the Yuanbao chat companion. And I think we'll run various experiments, and perhaps in a couple of quarters there will be more for us to report on a systematic basis. Right now, it's a bit too early to summarize everything for you in a systematic way.

Analyst

analyst
#26

Thanks a lot. And if I can, a quick follow-up question. As adoption and usage potentially gain more traction on both the user side and the enterprise side, what are we actually planning to do about managing, I suppose, the balance between how fast we invest and the revenue runway, particularly while we're in that early ramp-up stage? Thank you.

Pony Ma

executive
#27

So at this point, I think our aim is to move at full speed when it comes to generating demand, and honestly, if demand ends up exceeding what our GPUs can handle, that's when we'd begin considering how to pace things. That would be a nice problem to have. Right now, we're not there yet.

Analyst

analyst
#28

Thanks a lot. Appreciate it.

Wendy Huang

executive
#29

Our next question comes from William Packer at BNP. William, your line is open. William, go ahead and unmute yourself — your line is open.

William Packer

analyst
#30

Thanks to management for taking my questions, and congratulations on the strong quarter. Domestic games had a very strong quarter. Could you help us think through what this means longer term for the growth algorithm in that segment? Should we view the strong performance as 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—changed things underpinning better structural growth? And as a follow-up, there's been a lot of noise globally recently around regulatory interventions changing the split of economics between app stores and apps. Per management commentary, is this something Tencent has focused on in recent years? Where should we think of China on this journey in the split of value between apps and app stores, especially in the video game segments? Are there further shifts here as a value driver for Tencent? Thank you.

Pony Ma

executive
#31

Thank you, William. Now, regarding your question on domestic games, in the first quarter we did benefit from an easy comparison against the first quarter of 2024. So that definitely worked in our favor during the first quarter in a manner that won't hold true in every quarter ahead. But we do feel, we do believe there's a long runway for growth in our domestic game revenue and indeed our international game revenue going forward. There are many reasons, but let me pick out three for now. First, we spoke at length this time last year about some of the changes we were making to how we envision, and therefore how we run, and therefore who runs our biggest domestic games. And as you can see, we've made those changes and they're delivering the results we hoped they would deliver, and we see them delivering more results going forward. A second driver or enabler of that long runway is the use of AI, which we believe is especially beneficial to the large competitive multiplayer games we've discussed at length and that make up the majority of our domestic game revenue. And that's so because while there are many ways we can, and are beginning to, deploy AI within games, some of the most interesting involve using AI to help coach new players, to help accompany existing players, to help prevent cheating and hacking and so on.

Pony Ma

executive
#32

Each of those matters a great deal in competitive multiplayer titles. Then third, our game teams have really worked hard to position themselves where the gaming audience, or the coming gaming audience, is heading in China — and historically, the share of gaming activity in China made up by first-person action games has been far lower than it is elsewhere in the world. Our stance has been that China would close that gap. If you look at the slide on our domestic games this quarter, you'll notice that the majority of the titles we called out lead that first-person action category inside China, since Tencent holds the top position in that category within China. Users are indeed gravitating toward those first-person action games — Peacekeeper Elite, Crossfire Mobile, Call of Duty Mobile, Valorant, and now Delta Force as well. Each of those games is expanding at a faster or slower pace. But taken together, since every one of them is growing and some — Valorant and Delta Force among them — are growing very rapidly, the category overall is posting very healthy growth. That is what supported our domestic game revenue growth this quarter, and it also gives us a long-term runway for domestic game revenue growth.

Pony Ma

executive
#33

Now, shifting to the app stores — this really is a period of transition when it comes to the economics, which in the past were unfairly penalizing creators of digital content, especially game companies, effectively giving a free ride to providers of physical products. And from where we sit, as a business that runs an app store while also being a digital content provider ourselves, the app stores are being over-rewarded. That shift has been going on in China for several years now, and digital content creators, ourselves included, are now receiving a fairer portion of what end users spend. In the Western world, that change has moved much more slowly, or hasn't occurred at all, until quite recently. But a number of legal cases are now in progress. A number of regulatory interventions are now taking place, and we do believe that although it's hard to predict what will happen this week or what will happen next week, over the longer term there will be a reset toward a more equitable relationship in which digital content providers can keep more of the value they create, rather than in effect subsidizing a system that benefits e-commerce and other companies that contribute nothing to the entire app store ecosystem.

Pony Ma

executive
#34

In any case, that's the direction we see things moving—more clearly in China, and elsewhere too, where we expect it to become clearer as time goes on. Thanks.

Wendy Huang

executive
#35

Thanks. The next question comes from Alicia Yap at Citigroup.

Alicia Yap

analyst
#36

Hello, and thanks. Good evening, management, and congratulations on the strong results. Thanks for taking my questions. First, I wanted to ask whether management can share with us, over the past 2 or 3 months, some of the notable shifts in user behavior that Tencent has been able to observe following the deeper integration of AI into various business applications. For example, in terms of how users are using Yuanbao, whether this AI in Weixin is being noticed by business partners, and how Tencent expects some of the analysis you actually observe could help the company further improve user value as well as future monetization potential. And then a follow-up on gaming: I know that in addition to some of the bigger games that have been benefiting from AI integration, will some of the smaller titles also be integrating with AI in the coming quarters, where we could see improved monetization or a revival of user engagement on some of the mid-tier games as well? Thank you.

Pony Ma

executive
#37

Alicia, I believe we addressed that question of yours before — it's likely still too soon to run a methodical examination of how user behavior is shifting. What we're doing at this point, I'd say, is building features and user experiences that take advantage of AI, and then watching to find out what sticks with users and what doesn't. So, as I mentioned, users seem to enjoy posing questions and engaging with the AI through additional follow-up questions, and when we introduce things like letting photos be analyzed, people do use them. There are many features that have been added right now, and we're beginning to observe that people either like them a great deal or don't use them much. So it's a process of discovery. And — but overall at present, we can see that users are interacting with and using our AI agents increasingly as they gradually build up familiarity and keep working out what value the agents can add for them. So the overall usage, I'd say, is in fact rising.

Martin Lau

analyst
#38

Regarding your second question, I touched briefly on William's question about how we're putting what we'd call game AI into our large competitive titles. That deployment is at a really early stage, so in our view the biggest opportunity lies in continuing along that path. At the same time, there's a somewhat distinct chance to bring more — more generative AI — into games that are more content-driven. By doing that, one might initially help the game studio itself create content faster, and then in time allow some amount of user-generated content, experiences of varying quality, and eventually dynamically created content, where the game itself spins up a new environment should players decide to venture off the preset map. But I think all of that is to be explored in the years ahead. And right now, the larger and more tangible opportunity centers on using what we'll call game artificial intelligence inside the big competitive multiplayer games. Thank you.

Wendy Huang

executive
#39

Thanks. We’ll go now to the following question, from Kenneth Fong at UBS.

Kenneth Fong

analyst
#40

Hi, good evening, management, and thanks for taking my question. My first question concerns the high-end GPU. The US recently brought in a licensing requirement for the high-end GPU. As we understand it, our priority is to use this GPU internally. First, could management share your perspective on this situation, and how would it affect, if at all, our capital expenditure, AI development, and product launches? And I have a follow-up on the macro situation. In our press release, we indicate that our consumer loans business has seen year-on-year revenue growth following a more conservative approach last year. Given the still uncertain macro environment, can management share some insight into what we have observed on the ground, both on the customer side and on the advertiser side? Thank you.

Martin Lau

analyst
#41

Okay, regarding GPUs, the situation is genuinely quite fluid at the moment. Since our previous earnings call, an H20 ban came along, and then the new BIS guidelines arrived just overnight. So things are moving quickly, and we simply have to handle it — on one hand staying entirely compliant, and on the other, working out the best path for us so our AI strategy can still be carried out. What puts us in a good spot is, first, I believe we hold a fairly solid stockpile of chips purchased earlier, and those will prove quite helpful as we execute that AI strategy. As for how those chips get allocated, clearly they will go toward applications delivering near-term returns for us — the advertising business, for instance, along with the content recommendation product, where we'd actually be putting many of these GPUs to work producing results and returns. Second, training our large language models comes next in priority, and that training does call for higher-end chips.

Martin Lau

analyst
#42

What's encouraging there is that during the last several months, we've begun shifting away from the notion—or the conviction—held by American tech firms, which they term the scaling law, that you had to keep enlarging the training cluster without pause. We now observe that even a reduced cluster size lets you reach quite strong training outcomes. There's also substantial upside available on the post-training side, which doesn't necessarily call for very big clusters. That in turn lets us examine the high-end chips we already hold and conclude we possess a sufficient quantity of them to keep training our models across several additional generations ahead. After that, the bigger requirement for GPUs centers largely on inference, particularly when you witness rising inference demand from users, and also as we shift toward chain of thoughts reasoning models, which need far more tokens in order to respond to an intricate question. Moving into Agentic AI demands a greater number of tokens still, so there's genuinely substantial need on the inference side. Yet where inference is concerned, I believe considerable effort can be undertaken by us to keep that need under control.

Martin Lau

analyst
#43

First, there's simply the matter of making better use of software optimization. I believe we still have a fair amount of headroom to continue enhancing inference efficiency. So, if inference efficiency can be doubled, that essentially translates to twice the GPU capacity. That's actually a really solid way for us to put our resources to work in order to boost inference efficiency. Then the second approach involves customizing models of varying sizes. Some applications in particular don't need models that are very large. We're able to build tailor-made models and distill models so they serve different use cases, and that can genuinely reduce GPU consumption for inference. And lastly, we potentially can take advantage of other chips — compliant chips that are available in China or that are available for us to import — along with ASICs and GPUs in certain cases for inference on smaller models. So I think there are plenty of avenues through which we can meet expanding and growing inference demand, and we just need to keep exploring these paths and devote probably more time to the software side instead of simply resorting to purchasing GPUs.

James Mitchell

executive
#44

Regarding your question about the loan facilitation business and what we're seeing on the ground — what we've noticed this year is that credit quality for loans facilitated through our platform has been improving steadily and consistently. We believe that's partly because of better borrower selection and so forth. But it's also partly a macro development: over the last several years, consumers in China in aggregate have built up unusually large savings balances, and, all else equal, you'd rather be lending into a consumer with unusually large savings than one with unusually limited savings. So we're enabling the facilitated loan book in aggregate to grow in a very measured way. Looking ahead, how far can we grow that loan book? We estimate that if we benchmark the loan book facilitated through our fintech platform against the loan book of our largest peer companies, we're currently a fraction of their size. And so, while we'll continue doing this in a very measured fashion, we do think there's a long, many year runway for us to keep growing these high-margin loan facilitation revenues.

James Mitchell

executive
#45

Thanks.

Wendy Huang

executive
#46

Much appreciated. Our following question comes from John Choi at Daiwa.

John Choi

analyst
#47

Appreciate it, and thanks for the congratulations on the strong results. I've got a quick question about the comment on reinvesting into AI. I believe earlier in the opening remarks you guys mentioned a smaller gap between revenue and operating profit. Could you elaborate a bit more on this — the magnitude and what kind of extensive period we're talking about? And apart from CapEx, what are 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 across Fire Mobile, I just want to understand, are we seeing a structural evolvement in this typical first-person genre, as you point out earlier, across the board, do we expect, kind of evergreen status for multiple games within this specific vertical? Thank you.

Martin Lau

analyst
#48

You've raised some interesting questions, and I'll do my best to address both—just nudge me if I miss one. Regarding the gap between our profit growth and revenue growth, and the narrowing of that gap—both its size and how long it lasts—we're in uncharted territory, and that applies not just to Tencent but to the entire world when it comes to deploying artificial intelligence. So my confidence in these statements isn't necessarily very high, but if you want to gauge the duration, looking at the past may be the best way to predict the future. Tencent has gone through many periods in which we built a new product up to critical mass and substantial popularity before monetizing it, and usually the length of those gaps—between investing to cultivate something and monetizing it and generating revenue—would fall in the range of one to two years. So clearly, it will hinge on what our peer companies in China do. Clearly, it will hinge on consumer habits and on advertiser habits. Still, I think that's a reasonable timeframe to keep in mind. As for magnitude, I won't go beyond what we said earlier, which is to refer to a narrowing.

Martin Lau

analyst
#49

We don’t anticipate the gap between revenue growth and operating profit growth that we saw this quarter will persist. It should narrow, but conversely, we also don’t anticipate operating leverage becoming negative. Regarding which costs besides CapEx—or really depreciation—might drive that narrowing, CapEx depreciation is far and away the biggest factor. We have some added marketing spend for Yuanbao, though not much for AI within Weixin. Then we mention that engineers skilled in AI command high pay. But that’s more of a mixed remark than a comment on total headcount. We aren’t seeing a step up in headcount. We’ll keep managing headcount tightly. Still, we note that engineers with such AI expertise are deservedly well compensated. That covers your first group of questions. On your second group of questions about how popular first-person action games are in China, I’d make two broad points. First, as I said, to a significant degree this reflects China gamer behavior converging with long-standing gamer behavior elsewhere in the world. Outside China, the first-person action genre is the dominant genre in games.

Martin Lau

analyst
#50

Think of it as the television equivalent of a drama series. It accounts for 40% to 50% of the time gamers spend and of the revenue they generate. Historically in China, that figure was 10% to 20%. But when we examine the various cohorts, the different generations of gamers, what we find is that younger Chinese players are drawn disproportionately to first-person action games, and titles in that category such as Valorant, which speaks very directly to, say, the 20 to 30-year-old demographic, are performing especially well. So that's the first observation. The second observation is that while our newer first-person action games are performing well, so too are a number of our established titles, among them Peacekeeper Elite, Call of Duty Mobile, and Crossfire Mobile, and I think there are several reasons for that, but one is that you may recall that a few years ago, the unlocking of the Battle Royale mode, which in the west led to a fortnight becoming very popular, with Call of Duty, War Zone, PUBG, all becoming popular, and that was a kind of step up that meant the overall genre expanded. In China, we saw that step up from Battle Royale, and then more recently we've seen another step up thanks to the popularity of what's called the extraction shooter.

Martin Lau

analyst
#51

As I said, the growth in usage for Peacekeeper Elite — the largest first-person action title in China — has mostly come from the extraction shooter mode. When it comes to Delta Force, the majority of users with high retention are playing extraction shooter mode. Our view, then, is that every few years new styles of gameplay emerge within first-person action, and these tend to grow the total audience rather than take players away from other titles, and that is the process taking place right now. We are therefore quite pleased with how several of our first-person action games are progressing in China.

Wendy Huang

executive
#52

Thanks, John. Jialong from Nomura will be the one I take the next question from.

Jialong Shi

analyst
#53

Good evening, and thank you, management, for taking my questions. I have a few questions about the ads business. Every quarter, we hear management cite AI tech as one of the drivers behind your ad business. So I'm just wondering how much additional growth is likely to come from deploying AI tech in your ad business? And separately, what are the latest ad load and eCPM for your video account ads? And on Weixin Search, I'm wondering what the latest search query market share is for your Weixin Search, and how much of the ad revenue is now contributed by Weixin Search? After you add the AI search feature into Weixin, do you see user search activity migrating from traditional search to AI search? Thank you.

James Mitchell

executive
#54

Alright, I'll do my best to tackle some or all of those points, but regarding the question of how much additional benefit AI can deliver in lifting our advertising revenue, that's really an industry-wide question. If I had the answer to it, that would be extremely useful for understanding Meta and Google and a wide range of companies beyond Tencent. And if I try to simplify the framework a bit, then a large portion — not the whole thing, but a large portion — of the uplift that AI is delivering to ad revenue at present can be measured in terms of the click-through rate on ads. In the past, banner ads got a click-through rate of roughly 0.1%. Feed ads got a click-through rate of roughly 1.0%. Thanks to AI, we've observed that the click-through rate on some ad inventories can rise toward 3.0%, for instance. And then the question becomes, what is the ceiling on that click-through rate? Right now, nobody knows the answer, because it turns almost philosophical — if you had full information or insight into a consumer, if you could infer what that consumer wants, or what the consumer, based on their prior behaviors, ought to want, and then serve an ultra-targeted ad to that consumer.

James Mitchell

executive
#55

So at that point, it becomes quite difficult to argue that the ceiling ought to be X percent instead of Y percent. Naturally, there are additional avenues through which AI boosts advertising revenue. AI can help us serve more appealing content to consumers, and that leads to them spending more time in the feed, which in turn means they see more ads. That said, I'd say ad click-through rate is probably the most significant one. As for the ad load in video accounts, we've held it pretty steady over the past — I believe it's six months now — and it's still roughly 3% to 4%. On the video account CPM, it's still a very strong figure, partly thanks to the high click-through rate we're getting, and partly because advertisers are competing for access to the inventory that's on offer. Regarding Weixin Search's query share, we haven't made that number public, but it's trending upward nicely. And on how users are reacting to AI search, I think — one of my colleagues might give you a more nuanced answer. But first, it's early days, and second, I think the entire notion of AI search is somewhat slippery, in the sense of where you draw the line between a consumer typing a prompt into a pure-play AI experience versus typing a query into a search engine that might be answered by the traditional algorithm or might be answered by a result powered by a large language model?

James Mitchell

executive
#56

Those two things are kind of merging, and in our view — since we're relatively new, a fresh player in search — that works to our advantage. It gives us an opening to grow our share of users and, in the end, grow our share of revenue too.

Wendy Huang

executive
#57

Thanks, and we’ll go ahead and take our next question from Alex Yao at JPMorgan.

Alex Yao

analyst
#58

Thanks, management, for fielding my question, and congratulations on a strong quarter. My first question concerns advertising performance, which sped up from 17% in Q4 to 20% in Q1. What we'd like to grasp is how much of that robust showing, or acceleration, stems from macro recovery as opposed to technology improvement as opposed to inventory release? Going by what was said earlier, I'd guess inventory release carries the smallest effect of the one. So, even if it were measured precisely, pinning down the impact is hard — could you rank the factors behind the revenue acceleration? Then a second question concerns the AI product format. You folks have been vigorously building user adoption for Yuanbao for some time now. Could you share with us the user feedback, retention, et cetera, for those users actively using Yuanbao? More importantly, on the product side, how do you view the future development of the chatbot AI application? Is it likely to be the ultimate form of generative AI? Thank you.

James Mitchell

executive
#59

Let me address the advertising question, though not the AI one. On advertising, Alex, I wouldn't put too much weight on the quarter-to-quarter swings in ad revenue growth. It has stayed inside a range, and honestly, Q1 sat at the top end of that range. And we wouldn't necessarily even want an acceleration. What matters more to us is really safeguarding a very long runway so that range can be sustained for many years ahead — not many quarters. And it may be that if we find ourselves getting too close to the top of the band, then we spend more time raising the ad load on our existing products, or on deploying ads against our newer AI products, and really concentrate on optimizing for user time spent and user experience rather than trying to — rather than trying to break out past the top of the band. But then the band would last for a shorter period than we'd like. So that's the broad general comment. As for your question and sub-question about the specific drivers behind the upward fluctuation in Q1, it's certainly not inventory, because we haven't raised inventory at all.

James Mitchell

executive
#60

Whether that's macro-driven is unclear to me. We likely have to hold off until we see what other companies report. Beyond that, when it comes to the pace of ad tech enhancements tied to AI, that isn't something we attempt to plan too rigidly quarter by quarter. So once more, to sum up, I'd simply describe this as a typical kind of swing inside a band, toward the top of that band. I wouldn't interpret anything into it, or I wouldn't interpret too much into it, and I wouldn't extrapolate it too far ahead either.

Martin Lau

analyst
#61

Traffic is also growing organically. That’s likely a pretty significant factor.

James Mitchell

executive
#62

So the reason the band sits in the teens instead of a single-digit positive or a single-digit negative range — and then moves around inside that band — comes down to traffic growth plus the ad tech gains. I'd say a certain amount of quarterly noise comes with those. But the larger picture, which Martin highlights, is that traffic is growing, especially in the inventory that advertisers want most: video accounts, Mini Programs, official accounts, and Weixin search. And around those inventories, we're putting ad tech in place — using AI to boost what advertisers get back.

Martin Lau

analyst
#63

To build on that, our view of the advertising business is that the runway is very long, and our whole effort goes into lengthening it rather than controlling the growth rate quarter by quarter. If we keep enhancing our ad tech, that lengthens the runway. If we keep growing traffic — and a great deal of that is driven by AI too — that also lengthens it, because when all these users come to our video accounts, we still have many people who arrive, watch for a very brief time, and then leave. If we can work out the right content to push to them through AI, that can raise user time by a pretty significant level. Once we have that, we don't have to raise our ad load; we can simply depend on the organic growth of traffic, and that would carry revenue growth, and we can save the new ad load for a later time so that our runway gets extended again. And at the same time, with less advertising, we can actually improve our product experience, and that's another competitive advantage that creates a virtual cycle.

Martin Lau

analyst
#64

I think that's how we view our advertising business. Over time, if we can actually create a stronger transaction ecosystem inside our Weixin ecosystem, right, so that more and more merchants are able to advertise, and once the click-through rate happens, then it leads to transactions very easily. Then basically advertisers' propensity to spend with us would rise, and the value of each click would rise. Right? So there are a lot of these systemic things we're doing in order to extend the runway of our advertising business, which we believe, if you take a high-level view, has a very long runway, and we are continuously working to extend that. On AI. I think right now, just within the past quarter, we actually grew our Yuanbao user base by a lot. And we spent a lot of time making sure that we can retain them, 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, it really — chat is the predominant one at this point in time.

Martin Lau

analyst
#65

However, as we introduce additional features, the interface and layout could look different. If a user is attempting to complete something complex, it might turn into more of a project than a chat, and that task may need to remain in place over a longer period. So we'd tailor the form factor and the experience for users to match whichever capabilities we're delivering. Still, I believe the upcoming phase for Yuanbao is this — right now we've reached scale. Our user base is sufficiently large, and we've arrived at a point where we're actually able to keep those users. The following phase involves many more features we can still introduce that would let us retain and activate users more effectively and also draw in new ones. That, then, will be the coming phase of development for our AI product. After that, the real next step is to make the most of our distinctive ecosystem — whether social, whether content, whether Mini Programs — so that we can offer something clearly differentiated from the other AI products and AI chatbots available in the market.

Martin Lau

analyst
#66

And naturally, as we carry out all of this, we’ll keep improving our Huanyuan model, which will let us make the dual model strategy even stronger. So I believe these are the sorts of things we’ll be doing.

Wendy Huang

executive
#67

Thanks. We’ll take our final question from Ronald Keung at Goldman Sachs.

Ronald Keung

analyst
#68

Thanks, Wendy. And thanks to Pony, Martin, James and John. I have two questions. The first concerns our WeChat e-commerce strategy. I'd just like to hear which KPIs matter most to us, and whether there's a plan or timeline for perhaps a shelf-based offering, as opposed to what we have today, where live stream search is the primary format. The second is about payment. Management mentioned a slight decline in payment volumes in 1Q. So is that more macro-driven, or is it a matter of payment strategy, given that you also mentioned some improvement in April? Thank you.

Martin Lau

analyst
#69

When it comes to how we approach e-commerce, I don't believe we run things through rigidly fixed KPIs. This is a project measured in years, with an extremely long runway. I realize I keep returning to that same word — runway — because in our view it matters far more that the business continues expanding over a very extended period, instead of chasing quick sprints just to hit some KPI and then feeling pressured by it. So as for how we've constructed that runway, we have several initiatives underway. Priority number one is essentially enhancing the fundamental shopping experience for consumers: we need products of higher quality, we need a return policy that works conveniently for them, we need a solid customer service process, and we need attractive pricing for these consumers. That's the first. Second, we aim to bring in more supply of quality products, more supply of strong brands, and motivated merchants into our ecosystem, so that the supply side keeps getting better. Third, we want to drive additional traffic toward merchants and give consumers more opportunities to engage with products and with those merchants.

Martin Lau

analyst
#70

The initial move here is really about driving more traffic via live broadcasts. That, I believe, is one route by which we can genuinely turn more sales, and it has actually been well-proven. So other platforms have already done this kind of thing. And I see this as more of a no-regret move on our part. Yet at the same time, as time goes on, we actually want to have something distinctive for the Weixin ecosystem — namely, we can link users with products via our many different modules inside Weixin, such as Mini Programs. We know a great many users are already purchasing products via Mini Programs. And if we can actually shift some of those transactions over to Mini Shops, since there's better infrastructure for supporting e-commerce transactions inside Mini Shops, we can actually amplify the transactions, and it will benefit users, and it will benefit the merchants. There's a social component, there's a communication component, there's a connection to moments, there's a connection to search. On the social side, it's — we said the gifting part is one functionality that we added, and there'll be more to be added and connected to official accounts.

Martin Lau

analyst
#71

There are plenty of additional linkages we could introduce. And as time goes on, those linkages would begin bringing more transactions to merchants in a quite distinctive manner. That's the sort of work we're undertaking to expand our Mini Shop and Weixin e-commerce ecosystem over the long run. And I'm glad to report that although these efforts are still in the first inning, we keep witnessing very robust growth in our Mini Shop transaction GMV.

James Mitchell

executive
#72

Regarding your question about commercial payment volume, our market share could shift based on subsidy levels, on how aggressively we accept transactions funded by credit cards, and so on. But the truth is, we don't make dramatic changes from one month to the next. So we didn't increase our competitive intensity going from the first quarter into April. And so I think the big picture is that consumer spending has been fluctuating over the last several quarters, but the latest data point is an upward fluctuation. And so it's possible that this upward fluctuation reflects overall consumer confidence and consumption activity stabilizing. And we hope that over time, that stabilization turns into a sustained upturn. But, of course, time will tell.

Martin Lau

analyst
#73

To add a bit more detail here, I’d say that in Q1 we observed transaction counts rising steadily while prices moved lower. Our working theory is that this may reflect the final stretch of competition on the supply side as demand begins to pick up. The April trend seems to support that view, though it comes before tariffs really take effect. Absent tariffs, consumption clearly appears to have hit bottom and is beginning to recover gradually. With tariffs now in the picture, we’ll need to see how the next quarter unfolds. It’s highly dynamic: there are tariffs, but how broad are they? The government is also being quite supportive through stimulus, and in my view Beijing still has substantial room to introduce additional stimulus. At the same time, there’s some uncertainty around the tariff side. So the two things to watch are tariffs and how they interact with the economy, plus the offsetting effect from stimulus.

Wendy Huang

executive
#74

Thanks. That brings us to the close of this webinar. Appreciate everyone being with us for today's results. Should you want to look at our press release along with other financial materials, you can find them in the IR section on our corporate site, www.tencent.com. A replay of this webinar is going to be posted shortly as well. Thanks again, and we'll see you in the next quarter.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Tencent Holdings Limited transcript — plus 132,000+ transcripts from 7,000+ companies, speaker segments, AI summaries and full-text search — through the EarningsCalls.dev API.

Get the API View API docs →

This call discussed

For developers and AI pipelines

Programmatic access to Tencent Holdings Limited earnings transcripts and 132,000+ others is available through the EarningsCalls.dev REST API. Plans from $39.99/month — full transcripts, speaker segments, full-text search, and the recently-added /api/v1/transcripts/recent polling endpoint for ETL pipelines.

More tools for stock research

OptionsBell: an unusual options activity scanner that tracks options flow across all US stocks and emails you when large trades appear in Tencent Holdings Limited or any ticker on your watchlist.

StockMarketScan: a stock screener with alerts: 24 pre-built screens refreshed every 5 minutes, watchlists and email notifications. Both offer API and MCP access and fair monthly pricing.