Tencent Holdings Limited (700) Earnings Call Transcript & Summary

March 19, 2025

Hong Kong Stock Exchange HK Communication Services Interactive Media and Services earnings 75 min

Earnings Call Speaker Segments

Wendy Huang

executive
#1

Hello, and good evening. Thank you for being with us. Welcome to the Tencent Holdings Limited 2024 Fourth Quarter and Annual Results Announcement Webinar. I'm Wendy Huang from Tencent's IR team. Right now, everyone attending is in a listen-only mode. Once management finishes its presentation, we'll hold a question-and-answer session. <<operator_script>> Please also note that this webinar today is being recorded. Before the presentation begins, we'd like to remind you 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. Information regarding general market conditions comes from a variety of sources outside of Tencent. This presentation also includes some unaudited non-IFRS financial measures, which should be considered alongside, but not as a replacement 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 see our disclosure documents in the IR section of our website. Now let me introduce the management team joining the webinar tonight. Our Chairman and CEO, Pony Ma, will begin with a short overview; President, Martin Lau, will give an update on our AI initiatives; Chief Strategy Officer, James Mitchell, will provide 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

Now I'll hand things over to Pony.

Pony Ma

executive
#3

Thank you, Wendy. Good evening, and thanks to all of you for being here. During 2024, we bolstered the long-term franchise value of our core services. At Weixin, we are enhancing transaction features through the rollout of Mini Shops and improvements to Weixin Search. Time spent by users on Video Accounts climbed swiftly, driven by better documentation algorithms. Our evergreen game line-up rose to 14 titles in 2024 from 12 in 2023, and we cultivated new games showing evergreen potential. Three drama series from Tencent Video placed within the industry's top five for 2024, and Tencent Music extended its sector leadership, reaching 121 million subscribers. Marketing services revenue outpaced the industry as we upgraded our advertising technology platform and grew user traffic. Within fintech, we enhanced risk controls and lowered payment funding costs, reinforcing the overall fintech franchise and its profitability. Regarding AI, we pushed HunYuan's capabilities forward, applied AI to internal use cases, and laid groundwork for breakout growth in consumer AI use cases. On the financial side, our revenue growth rate strengthened across 2024, reaching double-digit growth in the fourth quarter. Growth and operating profit rose faster than revenues as we moved toward high-quality revenue streams. We remain focused (ph) on returning value to shareholders, distributing HKD32 billion in cash dividends and buying back HKD112 billion of our shares over the year.

Pony Ma

executive
#4

Now, turning to our financial figures for Q4. Revenue came in at RMB172 billion, marking an 11% increase compared with a year earlier. Gross profit reached RMB91 billion, which is 17% higher year-on-year. Non-IFRS operating profit stood at RMB59 billion, rising 21% versus the same period last year, while non-IFRS net profit attributable to equity holders amounted to RMB55 billion, a 30% gain year-on-year. At this point, I'll pass things over to Martin.

Martin Lau

analyst
#5

Thank you, Pony, and a good evening and good morning to everyone. During my part of the presentation, I'll discuss what we're doing to invest in AI for growth, and I'll give an overview of our AI efforts and the way we're putting money into AI both as something that multiplies growth across our current businesses and as a fresh engine of growth. This covers the headway we've made on our in-house HunYuan Foundation Model, our multimodal approach aimed at giving users the finest AI experience, bringing AI into the services we offer to enterprises, how AI is helping us unlock growth potential within existing businesses, and lastly, the increased investment we're making in AI for the future. Let me turn to the first topic. Our work in AI actually goes back to 2016, the year we set up our AI lab. Starting in the early part of 2023, we've been putting substantial resources into our proprietary HunYuan Foundation Model, which provides a key technological base for both our consumer and enterprise businesses and over time will act as a growth multiplier for us. Thanks to our HunYuan investments, we can build end-to-end foundation model capabilities spanning infrastructure, algorithm, training, alignment and data management, and we can also customize solutions to match the varying requirements of internal and external use cases.

Martin Lau

analyst
#6

When it comes to milestones, we were an early adopter of new techniques in Core LLM, such as the mixture of expert architecture in March 2024, the Heterogeneous MoE-based HunYuan Turbo in September 2024, and the Hybrid-Mamba-Transformer MoE-based Turbo S model in February 2025. Conventional transformer models run into limitations with context length, but our current Turbo S LLM boosts efficiency in processing long sequences by means of Mamba optimization. We also put out our HunYuan T1 deep thinking model in February 2025, which ranks among China's first long chain of thought models and delivers performance comparable to top tier models. Beyond LLMs, we've launched a multi-model HunYuan Foundation Model whose capabilities span image, video, and 3D generation. HunYuan's image generation models earned the highest score from FlagEval in December of last year. When it comes to video generation, our model stands out in video output quality and took first place on Hugging Face in December of last year. And our 3D Generation model was the industry's first open-source model supporting text and image to 3D generation. On top of that, we actively give back to the open-source community and have open-sourced a series of advanced models in the HunYuan family for 3D generation, video generation, large language, and image generation.

Martin Lau

analyst
#7

A number of these models have become very popular with developers around the world. Moving on to our consumer-facing AI products: we follow a multimodal approach so that users get the best possible AI experience, which lets us draw on every model available to address differing user needs. This matters to us since each AI model is tuned for distinct capabilities, performance metrics and use cases, and combining multiple models tends to handle complex tasks better than any single model can. What we've learned in software businesses like online games also shows that being both a developer and an operator creates synergies. Through investing in our own foundation models, we can fully tap our proprietary data to tailor solutions that fit customized internal and customer needs. At the same time, using external models lets us gain from innovations throughout the industry. On the product side, our AI-native application Yuanbao offers access to multiple models, among them Chain of Thought reasoning models like HunYuan T1 and DeepSeek-R1, plus the fast thinking model HunYuan Turbo S, with the option to integrate web search results. Search results in Yuanbao can directly reach high-quality proprietary content from the Tencent ecosystem, including official accounts and video accounts.

Martin Lau

analyst
#8

With HunYuan's multimodal abilities, Yuanbao handles prompts containing images, voice, and documents, not just text. Leading models are accessible in a stable, uncapped way through our cloud infrastructure. Between February and March, Yuanbao's DAU rose 20-fold, making it the third largest AI native mobile app in China measured by DAU. Beyond that, we've begun trialing AI capabilities inside Weixin to improve the user experience—for instance in search, language input, and content generation—and more AI features will be added to Weixin as we go forward. Turning now to the enterprise side: we've been speeding up AI integration across our cloud business, spanning infrastructure, platform, and software-as-a-service offerings. With our infrastructure-as-a-service AI solutions, enterprise customers get high-performance AI training and inference at scale, while developers can reach and roll out mainstream foundation models. On the platform-as-a-service front, our TI Platform flexibly handles model fine-tuning and inference needs. We'll deliver strong solutions that help enterprise customers customize AI assistants with their own proprietary data, and help developers create mini programs and mobile apps via natural language prompts. AI powered tools are increasingly benefiting our SaaS products. In Tencent Meeting, real-time transcription plus meeting summarization features have drawn considerable uptake, with monthly active users of these AI functions doubling year-on-year to 15 million.

Martin Lau

analyst
#9

Tencent Docs likewise boosted how productive users could be when generating and working with content. During 2024, revenue from our AI cloud roughly doubled compared with a year earlier. We have devoted more GPUs to internal needs — at first for ad tech (ph) and for training foundation models, and more lately for AI inference tied to Yuanbao and Weixin — which has cut how many GPUs we make available to outside customers and therefore held back growth in our cloud services revenue. When it comes to external workloads, we have directed the GPUs we can spare toward clients and use cases that carry high value. Starting in the fourth quarter of 2024, we have increased our GPU purchases, and as these GPUs get deployed, we anticipate that revenue growth across our cloud services overall will speed up. Turning now to our established businesses. We think our AI investment has been delivering positive returns already, and I will walk you through three illustrations of how AI strengthens our current products and businesses while producing return. On advertising, we have upgraded our ad system with neural network AI capabilities since 2015. Since 2020 we have rebuilt the ad tech platform on large model capabilities, which allows analysis of long-sequence user behavior across multiple properties, and that has led to greater user engagement and higher click through rates. Since 2023, we have kept layering in large language model capabilities to make approvals of ad content more efficient, to better grasp merchandise categories and users' commercial intent so ad targeting is more precise, and to give advertisers generative AI tools that streamline the ad create process, drawing on AI powered ad targeting capabilities along with generative AI ad creative solutions.

Martin Lau

analyst
#10

Our Marketing Services segment is already clearly benefiting from AI integration, with revenue up 20% in 2024 despite a challenging macro environment. In games, we've been using machine learning technology in our PvP titles since 2017. We've harnessed AI in games to refine the matchmaking experience, enhance game balance, and provide AI coaching for newcomers, which supports our evergreen games strategy. Our games business is now bringing in large language model capabilities to boost efficiency in 3D content production and to power in-game chatbots. For our video and music services, we're using AI to raise productivity and improve animation, live action video, and music content creation. Our content recommendation algorithms run on AI and have proven effective at boosting content discovery. Through these initiatives, we're better able to unlock the potential of our great content platforms. And lastly, as AI's capabilities and benefits have grown clearer, we've ramped up our AI investments to serve our internal business needs, train foundation models, and meet the searching demand for inference that we're seeing from our users. To concentrate our resources around this critically important AI effort, we've reorganized our AI teams to sharpen focus on both rapid product innovation and deep model research. In line with our stepped-up execution momentum and decision-making velocity, we raised annual CapEx more than three-fold to $10.7 billion in 2024, equal to roughly 12% of our revenue, with a notable increase in the fourth quarter of the year, as we purchased more GPUs for both inference needs and our cloud services.

Martin Lau

analyst
#11

Our plan is to raise capital expenditures again in 2025, and we expect CapEx to represent a low-teens percentage of revenue. On the R&D side, we'll keep putting money into our own models and speed up work on AI applications across each of our business groups. We're putting marketing dollars to work as well, to build user awareness and drive uptake of new AI products like Yuanbao. Over time, we think these investments will deliver solid economic returns, while we also have the capacity and the intention to keep giving capital back to shareholders, and we intend to repurchase no less than HKD80 billion of our stock in 2025. With that, I'll hand things over to James.

James Mitchell

executive
#12

Thanks, Martin, and greetings to all of you. Total revenue for the fourth quarter came back to double-digit expansion, rising 11% compared with a year earlier. VAS accounted for 46% of our revenue, and inside that, the social network subsegment made up 17%, domestic games 20%, and international games 9%. Marketing Services contributed 20% of our revenue, while FinTech and Business Services accounted for 33%. Gross profit climbed 17% year-on-year, helped by bigger contributions from high margin revenue streams including Domestic Games, Video Accounts advertising and Weixin Search advertising, together with cost efficiency from cloud services. Looking at segments, VAS gross profit rose 19% year-on-year and made up 49% of total gross profit. Marketing Services gross profit was up 19% year-on-year and contributed 22%, while FinTech and Business Services gross profit grew 11% and contributed 29%. Within Value added Services, segment revenue reached RMB79 billion, an increase of 14% year-on-year. Social Networks revenue rose 6%, driven by higher revenue from app based game item sales, music subscriptions and Mini Games platform service fees. Music subscription revenue grew 18% year-on-year as subscribers reached 121 million. Tencent Music strengthened cooperation with labels and artists and introduced super VIP privileges, including AI powered audio effect matching and collectible artists cards. Long-form video subscription revenue was up 3% year-on-year, with subscribers reaching 113 million.

James Mitchell

executive
#13

Our self-commissioned drama series Love Game in Eastern Fantasy ranked as the most watched drama across every online platform in China during November. Revenue from Domestic Games rose 23% year-on-year, coming off a low base quarter, fueled by evergreen titles — Honor of Kings, Peacekeeper Elite, VALORANT — plus contributions from recently launched games, DnF Mobile and Delta Force. International Games revenue climbed 15% year-on-year, or 16% on a constant currency basis, on robust performances from Brawl Stars and PUBG MOBILE and the early access release of Path of Exile 2. Turning to communications and social networks, Mini Shops — our platform for indexed and standardized merchandise — is progressively adding features to stimulate new and repeat transactions. As an example, we introduced a gifting feature, which leverages Weixin social interactions and people's desire to gift each other fun or trendy items. This gifting feature has the benefit that gift recipients must fill in their delivery addresses, which builds out our delivery location graph, making future shopping transactions more convenient for consumers and future delivery more efficient for merchants. Weixin Search continued to grow queries and revenue at rapid rates. We integrated Tencent HunYuan and DeepSeek large language model capabilities to enhance the relevance and quality of Weixin Search results, and Tencent's own model powered results now cover over 90% of question based searches.

James Mitchell

executive
#14

On the Domestic Games side, a number of our evergreen titles saw DAU and monetization gains during Q4 thanks to IP collaborations and high-profile events. Gross receipts for Honor of Kings rose by a double-digit percentage year-on-year, driven by higher DAU and popular outfits inspired by the anime series Detective Conan. VALORANT's gross receipts more than doubled, lifted by tie-ins with the world championship winning team Edward Gaming and with Riot's animated series Arcane Season 2. Fight of the Golden Spatula posted a double-digit percentage increase in gross receipts on Arcane-themed champions. We're looking to cultivate more evergreen titles. Take Delta Force, which we released recently: it brought in over RMB1 billion of gross receipts across PC and mobile in China in the fourth quarter, and our pipeline features highly anticipated games including the Hidden Ones, Light of Motiram, GODDESS OF VICTORY: NEW HOPE and VALORANT Mobile. A few weeks ago, during this 2025 Spring Festival period, each of our five highest grossing games raised its DAU compared with the 2024 Spring Festival period, which shows the health of the game industry and the vitality of our evergreen titles. That DAU growth came from gameplay and social activity initiatives like Peacekeeper Elite's Tang Dynasty-themed map and an Honor of Kings' Social Space with fireworks and selfie spots.

James Mitchell

executive
#15

Looking at our International Games, Brawl Stars — from Supercell over in Finland — stood as the third biggest mobile game by DAU across the industry outside China during 2024. Gross receipts rose multiple times year-on-year in Q4, helped along by the Angels vs. Demons season plus the redesigned battle pass. Path of Exile 2, coming from our subsidiary Grinding Gear Games in New Zealand, is a fresh action RPG built for PC and console that offers an in-depth character development system. After its early access launch in December, the title held the number one spot among premium games by revenue on Steam for six weeks. Warframe, from Digital Extremes — our subsidiary in Canada — put out a major update called Warframe 1999 during the fourth quarter, which lifted DAU and pushed the game to its best fourth quarter gross receipts ever across its 11 year history. Across the whole of 2024, Warframe's gross receipts climbed more than 30% year-on-year, hitting a life to date record as well. Turning to Marketing Services, revenue was up 17% to RMB35 billion in the fourth quarter, driven by stronger user engagement and the continuing AI upgrade of our advertising platform. More specifically, we improved our AI models so they grasp users' interests more holistically and understand how users respond to ads, which lets our system serve more relevant ad recommendations.

James Mitchell

executive
#16

Marketing Services revenue rose across the majority of industries. When we break it down by inventory, Video Accounts Marketing Services revenue was up more than 60% year-on-year, driven by stronger user engagement, the AI improvements we discussed earlier, and a rise in consumer transactions inside Video Accounts, which produced additional close loop e-commerce ads. Mini Programs Marketing Service revenue climbed quickly year-on-year, while Weixin Search revenue more than doubled year-on-year, helped by a greater number of commercial queries along with AI optimized ad placements and ad formats that lifted click-through rates. Turning to FinTech and Business Services, segment revenue reached RMB56 billion in the fourth quarter, an increase of 3% year-on-year. FinTech Services revenue returned to low-single digit year-on-year growth during the quarter, supported by better commercial payment volumes and higher revenues from wealth management and consumer loan services. In commercial payments, revenue improved to largely stable year-on-year, and transaction counts grew faster year-on-year in the fourth quarter compared with the third quarter. Business Services revenue posted modest year-on-year growth in the fourth quarter, helped by higher cloud services revenue and greater technology service fees stemming from rising e-commerce transaction volumes. However, as Martin noted, directing more GPUs toward Tencent's own needs temporarily held back cloud services revenue growth in the quarter. Business Services gross margin improved year-on-year thanks to greater efficiency.

James Mitchell

executive
#17

On WeCom, revenue was up more than 2x compared with a year earlier, as enterprises show growing readiness to pay for advanced communication features. And Tencent Meeting posted revenue growth of over 40% versus the prior year, helped by greater adoption among enterprises. With that, I'll hand things over to John.

John Lo

executive
#18

Thank you, James. Hello, everyone. In Q4 2024, our total revenue came in at RMB172.4 billion, which is an 11% increase year-on-year. Gross profit reached RMB90.7 billion, rising 17% year-on-year. Operating profit stood at RMB51.5 billion, growing 24% year-on-year. Interest income was RMB3.9 billion, roughly flat compared with a year earlier. Finance costs totaled RMB2.5 billion, a 29% decline year-on-year, reflecting the exchange — a foreign exchange gain this quarter versus losses during the same period last year. Share profit of associates and JV was RMB9.3 billion, compared with RMB2.4 billion in last year's same quarter. Under IFRS, share profit was RMB7.7 billion, an increase of RMB4.5 billion from the same quarter a year ago, driven by associate company-specific factors such as business growth, new content releases, and improved operating efficiencies. Income tax expense rose 22% year-on-year to RMB11.8 billion, mainly as a result of operating profit growth. On a non-IFRS basis, diluted EPS was RMB5.91, up 33% year-on-year, ahead of non-IFRS net profit growth thanks to a lower share count from our share buybacks. Turning to non-IFRS financial figures for quarter four: operating profit was RMB59.5 billion, up 21% year-on-year. Net profit attributable to equity holders came to RMB55.3 billion, up 30% year-on-year. The gap in Y-on-Y growth rates between operating profit and net profit stemmed mainly from a higher non-IFRS share of profit from associates and JV, which rose 72% to RMB7.7 billion this quarter.

John Lo

executive
#19

Now let's turn to gross margin. In the fourth quarter, gross margin overall came in at 53%, which is a rise of 3 percentage points compared with a year earlier, and the same pattern held by segment. For value-added services, gross margin reached 56%, 2 percentage points higher year-on-year, driven by a greater share of high-margin domestic games revenue, better margins in music subscription revenue, and a smaller contribution from lower-margin live streaming revenue. Marketing Services posted a gross margin of 58%, an increase of 1 percentage point versus a year ago, thanks to growth in the high-margin Video Accounts and Weixin Search revenues. At FinTech and Business Services, gross margin was 47%, 3 percentage points above the prior year, reflecting a larger share of high-margin revenues together with enhanced cost efficiency in our cloud services. Turning to fourth quarter operating expenses: selling and marketing expenses amounted to RMB10.3 billion, a decline of 6% against the elevated base in the same quarter of last year. Selling and marketing expenses accounted for 6% of revenues, versus 7% a year earlier. Total G&A expenses stood at RMB31.4 billion, an increase of 16% year-on-year, largely attributable to higher R&D expenses, which climbed 21% year-on-year to RMB19.8 billion, primarily because of greater staff costs and GPU service depreciation tied to our AI initiatives. Excluding R&D, G&A rose 8% year-on-year to RMB11.6 billion.

John Lo

executive
#20

By the end of the quarter, we had roughly 111,000 employees, which is a 5% increase year-on-year, or 2% quarter-on-quarter. In the fourth quarter of 2024, the non-IFRS operating margin came in at 34%, 3 percentage points higher year-on-year, broadly consistent with the growth in gross margin. Now I'll go over a few key metrics on cash flow and the balance sheet. Operating CapEx for the fourth quarter was RMB34.9 billion, up 421% year-on-year, which reflects more investment in GPUs and servers as we scale our AI capabilities. Non-operating CapEx reached RMB1.7 billion, an increase of 103% year-on-year, mainly because of data center construction that's in progress. Total CapEx came to RMB36.6 billion, up 386% year-on-year. Free cash flow stood at RMB4.5 billion, down 87% year-on-year, largely as a result of higher CapEx spending on GPUs, servers, and data centers. Sequentially, free cash flow declined 92%, owing to timing differences in the settlement of certain accrued expenses, greater CapEx spending on GPUs and servers, together with seasonally lower gross receipts from games. Our net cash position was RMB76.8 billion, down 20% quarter-on-quarter, which reflects cash outflows tied to CapEx and share repurchases. Turning to capital returns, during the full year of 2024 we bought back 307 million shares for total consideration of HKD112 billion, which more than doubled both the number of shares repurchased and the total consideration compared with last year.

John Lo

executive
#21

When calculating our 2024 diluted EPS, the weighted average count of shares came in 2% lower than a year earlier. Assuming shareholders give their approval at the 2025 AGM that's coming up, we're putting forward a yearly dividend of HKD4.5 per share — that works out to a 32% rise over the prior year. Shareholders will receive this dividend on May 30, 2025. Thank you.

Wendy Huang

executive
#22

Thanks, John. At this point we'll take questions from the line. <<operator_script>> Our first caller is Kenneth Fong of UBS. Kenneth, your line is open.

Kenneth Fong

analyst
#23

Hello there, and good evening to the management team. Congratulations on the robust quarter, and I appreciate your taking my questions. I have two things to ask. The first concerns how AI CapEx affects the financials. As we ramp up CapEx spending on AI, additional depreciation and R&D expenses will inevitably put pressure on our margin. Over the last several years, we've recorded a substantial margin gain thanks to our emphasis on high-quality growth. So looking ahead, how do we strike a balance between growth and improving profitability? My second question is about Mini Shops. During Q3, we introduced the Weixin Mini Shops along with the Blue Packet gifting function, both of which users have welcomed warmly. So could management walk us through the strategy and the key initiatives for e-commerce this year? Thank you.

James Mitchell

executive
#24

Hi, Kenneth. Thanks for asking. There's a saying that the only certainties in life are death and taxes. As a company executive, I'd say we'd be failing in our duty if we claimed lower margins were unavoidable — and we definitely don't think that's true. On the subject of higher research and development spending on AI, basically in every single year of Tencent's history we have been raising our R&D outlay across a range of different projects. So in our view, R&D spending isn't really a pressure on margins as such. CapEx is a more subtle matter, because we did raise CapEx to a new, higher steady state in Q4 of last year, and in time that additional CapEx will translate into additional depreciation across the coming several years. Still, it's worth examining precisely where that CapEx is being directed, so we can tell whether the depreciation turns into margin pressure or not. The most immediate use of the CapEx is GPUs to support our ad tech business and, to a smaller degree, our games businesses. And as you can see in our results, and as you can hear from what Martin discussed, that CapEx in fact produces good margins and high returns.

James Mitchell

executive
#25

The second area where CapEx went was GPUs used to train large language models. At one point last year, people believed each new generation of large language model would need an order of magnitude more GPUs — that belief came to an end with the breakthroughs DeepSeek showed. These days the industry, and we inside it, are achieving far greater productivity in large language model training from the GPUs we already have, without having to add extra GPUs at the pace people previously anticipated. Third, some CapEx goes to our cloud business, where we purchase GPU servers, rent them to customers, and earn a return. It might not be the highest-return business in our portfolio, but even so, the return is positive. It covers what the GPUs cost and therefore the associated depreciation. Then finally, where I think there could be short-term pressure is CapEx for 2C inference. That is one more source of cost pressure. Still, we consider it a manageable cost pressure, since that CapEx forms a subset of total CapEx. And we are also optimistic that in time that 2C inference activity we are generating — just like earlier activity across different Tencent platforms — will be monetized via a mix of advertising revenue and value added services.

James Mitchell

executive
#26

All in all, we get that there are questions on your end about the increase in CapEx and how that ultimately shows up in profitability over time, but we're genuinely pretty optimistic about our ability to keep growing the business while safeguarding margins. Thank you.

Martin Lau

analyst
#27

I'd like to make just one additional point on top of the very thorough response James gave. In inference for consumer-facing products, there are actually plenty of avenues through which we can lower unit cost by technical means — through software and through better algorithms. So I think that's another factor worth keeping in mind. Now, regarding the Mini Shop, I'd say first that it's a very long-term initiative for us. Any particular initiative is just one of many things we can do to build up this ecosystem over the long run. Second, I'd like to remind the audience about how the Mini Shop is positioned. It's really a unified platform connecting all the components of our Weixin ecosystem. And there's a standard, indexed merchandise information data structure, so that merchandise information can flow freely across the different components of the Weixin ecosystem. The purpose is actually for our consumers to be able to find quality products and merchants. And if you look at the different components of the Weixin ecosystem, there is social infrastructure, there is content, there is search, there are Mini Programs, there's a transaction platform, and there's also WeCom.

Martin Lau

analyst
#28

There are many pieces inside the unified Weixin ecosystem that Mini Shop genuinely wants you to work within. If you look at gifting, that is just one single feature belonging to the social side of that ecosystem. So it is merely one among a great many capabilities we are able to introduce so as to truly take advantage of the whole Weixin ecosystem. Based on what we have heard at the outset, as you noted, it has in fact been received rather warmly. We observe plenty of users employing this capability to send presents to friends over the Chinese New Year stretch. That gifting likewise amplifies the word-of-mouth impact of high-quality items, since you will only give away products you personally found to be genuinely good. As for those who get the products, which somewhat extends the original purchase, they would type their address into the delivery address field, and that in turn helps us build up an infrastructure of a delivery graph, and that would in fact assist us in finishing transactions more easily for the many people who have already entered their address. And typically it would additionally prompt the recipients of the gifts to occasionally pass them along to their own friends.

Martin Lau

analyst
#29

All in all, I'd say the word-of-mouth impact has been quite strong, and merchants have given us positive feedback. That said, at present it's just a single piece of the broader Asian ecosystem. So our view was that as time goes on, we'll keep developing this ecosystem and platform patiently, treating the whole thing as a marathon, not a sprint. And we believed we can actually travel quite far and for a long time down this road. Also, if you examine Mini Shop's GMV, it kept expanding at a really rapid rate during Q4 of last year.

Wendy Huang

executive
#30

Thanks, Martin. We’ll move on to the next…

Kenneth Fong

analyst
#31

Much appreciated.

Wendy Huang

executive
#32

Thanks, Kenneth. Now we'll take a question from Alicia Yap at Citigroup.

Alicia Yap

analyst
#33

Thank you. Good evening, management, and thanks for taking my questions, plus congratulations on the strong results. Two questions. The first one concerns the enterprise-facing service you mentioned — I'm just wondering whether management could illustrate a bit the demand and adoption rate you've observed for your AI (ph) service over the past two months, and how you expect demand growth to unfold in the coming quarters. And on the SaaS products, besides Tencent Meeting and Tencent Docs that you mentioned, could management also share some of the software solutions we plan to roll out that could potentially help with cross-selling and upselling to our cloud customers. The second question is about the consumer-facing application. I think in addition to Yuanbao, which overseas has achieved a very strong breakout over the last couple of months, there's also this EMA (ph) Copilot, and Weixin also has the enhanced search features. So I'm just wondering how all these products will eventually evolve over time, and whether Yuanbao will be positioned as the AI gateway that consolidates all the search and discovery entrance and will be complementary to the WeChat super app. Thank you.

Martin Lau

analyst
#34

Alright. When it comes to demand for AI services, it's genuinely very strong, but as I mentioned before and James touched on as well, we're actually facing supply constraints. Part of why you're seeing such a large jump in CapEx for the fourth quarter is that we have a number of rush orders for GPUs, for inference as well as for our cloud service. And we'd only be able to capture that big increase in AI service demand once we actually get these GPUs installed in the data center, which takes some time. So I'd say we probably haven't really captured much of that during the first quarter, but over time, we'll capture a fair amount of it as the GPUs arrive and get installed. And on the different software solutions, you mentioned Tencent Meeting, Tencent Docs, and WeCom is another very strong product — in fact, it's actually our biggest SaaS product by revenue, and it has grown a lot, with revenue doubling year-on-year in the previous quarter. Beyond that, we also see security PaaS software and audio and video PaaS software, including real-time communications as well as live broadcasting software, that we're actually selling to our cloud customers.

Martin Lau

analyst
#35

A good number of those can also be enhanced with AI, delivering added value for our customers. Then, when it comes to the consumer-facing application side, yes, we do actually have a wide range of different consumer-facing applications, and you should expect additional ones down the road. My view is that AI is genuinely at a very early stage, so discussing what the ultimate state might look like is really difficult. That said, I would offer, first, that every product will keep evolving into something very useful and even more powerful for users. So Yuanbao can serve as a very strong AI native assistant, and the EMA (ph) Copilot could be your personal library as well as a collaborative library for team collaborations. And Weixin can have a great many different features coming, right? Beyond those products, I think our other offerings will feature AI experiences, including QQ, including browser and other products. So I believe we will see increasingly more AI — consumer AI-facing products — and at the same time, every one of the products will keep evolving. If you look at Yuanbao, I think it is indeed consolidating a lot of different functionalities, but it will not be the sole gateway.

Martin Lau

analyst
#36

Every one of our products would genuinely seek out distinctive use cases where they can harness AI to deliver a great user experience for their users. Yet at the same time, the various products we offer are also able to collaborate, so that the proper pathway exists for our AI products to expand their own user base. So I believe that will keep on evolving, and that would benefit us as we create a whole host of AI solutions and consumer applications.

Alicia Yap

analyst
#37

Thanks, Martin.

Wendy Huang

executive
#38

Thanks. We’ll go ahead and take the next question from Shi Jialong at Nomura.

Jialong Shi

analyst
#39

Thanks, Wendy. Thanks to the management team, and good evening. Once more, congratulations on a really solid quarter. My question is actually a follow-up to the previous one. We've seen Tencent Yuanbao display quite strong growth momentum since this year. So could management go into more detail on the strategies for growing the user base further in this highly competitive market? Yuanbao combines an AI chatbot with an AI search provider. So I'm just wondering which of those two parts Tencent is most excited about? And related to that, how large a share of the entire search market will AI search eventually account for in the future when it comes to search queries, and how do you plan to monetize AI search going forward? Thank you.

Martin Lau

analyst
#40

Okay, so with Yuanbao, at present it functions as a chatbot plus search. Over time, though, my expectation is that it will expand into a full-fledged AI assistant offering a wide range of capabilities that serve various kinds of users. That means it would span from learners — students who are trying to study — and would also encompass all sorts of other people, including knowledge workers aiming to get their tasks done. We would also address deep research, which lets users conduct very thorough investigation into various subjects. So I believe it will end up with a great many different applications. The distinctive edge for Yuanbao, clearly, comes down to innovation — it's about continually layering on features and capabilities to meet what users need. That said, I'd say Yuanbao holds a particular advantage through its ability to tap Tencent's content ecosystem, particularly around official accounts and video accounts, which represent information sources of very high quality. At the same time, we believe our multi-model approach genuinely helps users reach the strongest model and also gives them the benefit of combining models to handle their complicated requirements. And going forward, as I hinted earlier, a lot of our various high-DAU products will begin incorporating different AI features and capabilities, and some of those will have a pathway and access into Yuanbao, so our range of products would truly support one another.

Martin Lau

analyst
#41

So I'd say those are the areas where we can keep delivering on an ongoing basis in a way that's competitive. That said, we also hold some distinctive advantages of our own.

James Mitchell

executive
#42

Regarding your broader question about AI prompts compared with traditional search, opinions will differ from person to person, and it will take time for things to unfold. Yet at a high level, when we consider how web search historically absorbed the web directory, and when we consider how we ourselves behave with AI prompts compared with traditional search, I believe there's a possibility that AI search will absorb traditional search. After all, the web directory, traditional search, and AI prompts all serve as mechanisms for reaching the Internet's knowledge graph. That said, among these, AI prompts introduce new technology, new efficiency, and also new transactional capabilities via agentic AI that traditional search could not offer. So as for how the AI prompt will be monetized, time will tell. Still, I think we can already observe in the Western world that the first monetization comes through subscription models, and then over time performance advertising will follow. In China, I think it will begin with performance advertising, and then value-added services will follow. Thank you.

Jialong Shi

analyst
#43

Thanks.

Wendy Huang

executive
#44

Thanks. Our next question comes from Ronald Keung at Goldman Sachs.

Ronald Keung

analyst
#45

Thanks, Pony, Martin, James, John and Wendy. So I have two questions. The first is about <<inaudible>> advertising: given the healthy 17% growth we saw, should we treat this fourth quarter exit rate as a proxy, let's say, for 2025 this year? And given the discussion of the AI-powered enhancements and what we've seen from Meta — that machine learning and advantage plus shopping — all of these AI could drive growth acceleration for some of the global peers. So could there be a possibility of even acceleration in this marketing services growth line as we head into this year with more AI applications of that? Separately on games, how should we view the growth outlook this year for Domestic and International, especially the higher base effect for Domestic by the second half of this year? And with AI, what are the implications to this business? You talked about longevity, but also competition and cost structures in this games industry. Thank you.

James Mitchell

executive
#46

Let me start with a quick look at advertising and game trends, and I'm not certain whether Martin will cover game AI—we'll figure that out in due course. On advertising, we're quite happy with the fourth-quarter growth rate, which clearly beat the industry. There weren't any special tailwinds worth flagging. Growth was fairly organic and quite broad-based across pretty much every industry we monitor. We believe that's because, in pretty much every industry we monitor, the AI enhancements we're deploying are delivering better return on investment for advertisers than they had before and than what's available elsewhere. You compared us with some global peers, which I think is the right thing to do—with the caveat that some of those global peers tend to shift to a sort of fully loaded ad load much earlier in their products' evolution. Take short video: by contrast, we tend to raise the ad load gradually and incrementally for newer products like short video, and that's still the case. So I think there's a difference in how quickly we choose to race down the runway versus some of our global peers.

James Mitchell

executive
#47

But on the whole, provided the macro picture holds broadly steady, we'd say we're pretty comfortable with the advertising side of our business. Turning to gaming, your question is really about a tougher comparison in the back half of the year. I suppose that's the peculiar burden of this sort of sector — succeed, and a year down the road everyone starts fretting over the base effect. Still, there are a few things we can actually point to in our games business right now. First, we closed out last year with deferred revenue — a big chunk of which is generated by gaming — sitting at a high-teens percentage increase versus a year earlier. That deferred revenue will therefore come through into reported revenue across the first half, and also across the back half of next year — 2025 — plus a bit in 2026. Second, something we highlighted during the prepared remarks: how users behave around Chinese New Year gives us a rather useful signal on whether our games are, broadly speaking, gaining or losing popularity. So the fact that all five of our top-grossing titles posted year-on-year growth in daily active users during this year's Chinese New Year, compared with the same period a year ago, counts as an encouraging leading indicator.

James Mitchell

executive
#48

Both of those are, in a sense, observable facts. There are also subjective opinions. From a subjective standpoint, we think a couple of our games are well along the path toward graduating into evergreen status. We named one of them: Delta Force. That represents a genuinely significant opportunity tailwind. Second, as we discussed during the prepared remarks, we have quite a few new titles in development that excite us. And third, we genuinely think games gain from AI technology improvements, both directly and potentially indirectly. On the direct side, game developers employ AI to help them produce more content faster and serve more users more effectively. On the indirect side — and this may be a story spanning multiple decades rather than the second half of this year — we think that as humanity adopts AI more widely, people empowered by AI will have more time and also more desire for high agency activities. And one of the finest outlets for them to express themselves in a high agency rather than a passive manner is interactive entertainment, meaning games.

Martin Lau

analyst
#49

And I think there's just one additional point worth making here, which is that when we consider the competitive dynamics, right, we actually believed AI would make evergreen games even more evergreen. And we're already seeing, sort of, how AI can assist us in executing and amplifying our evergreen strategy. And part of that sits within production, right — you can now create great content in a shorter span of time, so you're able to keep updating the games with a higher frequency of high-quality content. And with the PvE experience, when the bots are smarter, right, you actually, sort of, make the game more thrilling and more akin to PvP. And within PvP, a lot of the matching and balancing and coaching of new users can actually, sort of, be accomplished in a much better way once you apply AI. So all of this would actually help games that are already popular and large scale become even more popular and more appealing to users.

Ronald Keung

analyst
#50

Understood. Thank you.

Wendy Huang

executive
#51

And our next question comes from Alex Yao at JP Morgan.

Alex Yao

analyst
#52

Thanks, management, for taking my question, and congratulations on a strong quarter. First, on commercial payments: it's encouraging that your commercial payment revenue went from negative growth in Q3 to flat year-on-year in Q4. Could you share what you've observed about how commercial payment activity has trended so far in the first quarter of 2025? My second question follows up on AI and CapEx. You guided to a CapEx-to-revenue ratio in the low-teens for 2025, similar to the ratio for '24. So basically, this guidance implies a significant slowdown in CapEx growth. Can you walk us through the rationale behind this CapEx-to-revenue ratio? Is it because you foresee slower demand growth for GenAI, or because the big step up in 2024 will be sufficient to address GenAI demand in 2025? Thank you.

Martin Lau

analyst
#53

On the commercial payments side, what we're seeing is that transaction volumes are actually rising even more, yet pricing pressure on the ASP has carried on. So on balance, from a value standpoint, it remains roughly flat. Our read on this is that consumer willingness to spend is genuinely returning. On the supply side, though, there's still considerable pressure on pricing. So ideally this is a positive signal that we're nearing the end of a difficult market, correct? And as consumer demand keeps getting better, suppliers will gradually become less aggressive on price, and in time that should turn into value growth. Still, that's something we'll need to watch as we move ahead.

James Mitchell

executive
#54

Now, regarding your question on CapEx revenue — yes, that's right, because the increase we made in late 2024 ought to be enough to handle GenAI and other requirements in 2025 at roughly this new normal run rate. There's a time lag for us between placing orders for the GPU servers and getting them fully deployed in data centers. So during Q4, and for part of Q1, that was the situation we were in. But as Martin mentioned when talking about exiting the first quarter, we are deploying those GPUs, and we capture the benefit from them both for our internal inference needs for Yuanbao and for our external Tencent cloud needs — renting to clients and producing direct revenue. And I think if you step back and consider the bigger picture, there was a stretch last year when people questioned whether our CapEx was large enough compared with our China peers, compared with our global peers. And now, among the listed companies, I believe we had the biggest CapEx of any China tech company in Q4. So we sit at the forefront among our China peers. Broadly, China tech companies are spending less on CapEx as a percentage of revenue than some of their Western counterparts.

James Mitchell

executive
#55

However, for a while now, we've held the view that this stems from Chinese firms generally putting priority on efficiency and utilization — making efficient use of GPU servers. And that doesn't necessarily hurt the ultimate effectiveness of the technology being developed. And I think DeepSeek's success truly kind of epitomized and cemented, showed that reality. Thank you.

Martin Lau

analyst
#56

The one thing I’d like to add, though, is that this situation is extremely fluid, correct? So what we’re providing is really, in a sense, our expectation. But honestly, right, that expectation could shift. If all of a sudden there’s, say, a spike in demand, then we can certainly, say, raise our order for GPUs. So I think we’ll be, say, quite flexible and dynamic when it comes to responding to market dynamics.

Wendy Huang

executive
#57

Thanks. The next question comes from John Choi at Daiwa.

John Choi

analyst
#58

Thanks for taking my question. First, I'd like to follow up on advertising. Could management give us a bit more detail? I believe during the prepared remarks you mentioned that we've seen growth across most industries — so which verticals benefited the most? And on the advertiser side, we noticed that video accounts, mini programs, and Weixin search all rose fairly significantly. So how has advertiser behavior changed, and how has that helped our advertising revenue growth? My second question concerns capital allocation. I think management said that our share buyback this year will be lower than last year, while dividends, on the other hand, will go up. So what's our priority when it comes to allocating capital? And at the same time, how should we think about M&A investments, given that we're seeing very strong demand in the AI industry at this stage? Thank you.

James Mitchell

executive
#59

Absolutely. Regarding your question about advertising, several categories showed year-on-year growth, including e-commerce, financial, fast moving consumer goods, games, local service, education, and healthcare — quite a long list spanning the board. As we roll out these AI enhancements, we generally start with Video Accounts, and from there they gradually spread to other Weixin properties, to Tencent properties, and to our ad network, so the advantages end up being felt more and more widely as time goes on. That covers the advertising question.

Martin Lau

analyst
#60

When you think about how we allocate capital, the main guiding idea is that we aim to put money to work in ways that produce returns for the company and for shareholders, right? And that plays out in different areas at different times. At certain points we put a great deal into the ecosystem and into our ecosystem partners, building up a very sizable investment portfolio. After that, we began giving cash back to shareholders via dividends and share buybacks. And today there's AI, and we believe AI holds a great deal of potential for generating returns down the road. So at this point we essentially want to shape an investment approach where we're putting money into our future while also delivering return in the current period via dividend and buybacks, right? So that's the main guiding idea, I'd say. And when you consider whether we're able to do that, I'd tell you we have very strong financial capacity for it, since our operating cash flow is very strong. We likewise hold a very sizable and valuable investment portfolio, and a big portion of that is in fact quite liquid. So we have sufficient financial resources to put money into the future, in AI and, should the need arise, right, into certain investment activities.

Martin Lau

analyst
#61

Our investment portfolio basically funds itself, and at the same time it gives shareholders a return for the current period, right? So when we look at this year — we've already said we will pay a cash dividend of HKD41 billion, and we'd bring our share buyback down a bit from last year; we said we'd repurchase 100 billion, and this year we said we'd repurchase 80 billion, right? So basically you can think of it as a 20 billion flexibility buffer that lets us sit through the year and see whether we need to put more money into AI. But if you look across the two years, last year we said the dividend was 32 billion and the share buyback was 100 billion, which is a 132 billion target return. And this year our target return is 80 plus 41, and 121. But if you add last year's overspending on our share repurchase of 15, right, then this year's number is actually sort of similar to last year's target. So I think we're trying to strike a very good balance between investment for the current period — return, as well as future investment. But we have a very strong set of financial resources to do that.

Wendy Huang

executive
#62

Thanks, and we’ll go next to Robin Zhu at Bernstein. Robin, you’ll have to unmute yourself. Perhaps we’ll move on to Thomas Chong at Jefferies instead.

Thomas Chong

analyst
#63

Hello, and good evening. Thank you to management for fielding my question, and congratulations on such a strong set of results. My question returns to the earlier one about spending as well as higher-margin business and monetization. Given that we have pursued a high quality growth strategy over the past couple of years, and we are also looking into a margin expansion story in which operating profit growth outpaces revenue growth, but we also point out that as time goes on, the delta between the two will narrow down — so, I think, connecting all the dots together and putting all the puzzles together, how should we think about the margin expansion story in 2025 and the next few years? Thank you.

James Mitchell

executive
#64

Let me try to address that. So in some sense, a major factor behind the operating margin leverage we've seen over the last two years, especially, is that we've benefited from leverage on gross margin. And the reason gross profit has grown faster than revenue is that, while our base business, broadly speaking, carries a blended gross margin of roughly 50%, a number of newer revenue streams account for most of our revenue growth, and those come in at gross margins of 70%, 80%. So as that mix has shifted, it has been lifting gross margin and causing gross profit to grow more quickly than revenue. Now, naturally, over time there is a base effect—or, more precisely, a kind of asymptotic effect. Once the blended gross margin moves nearer to the incremental gross margin, then unavoidably the pace at which the blended gross margin improves slows down. So I believe that holds over the longer term. That said, when we look at our high-quality revenue streams—like the Video Accounts ads, like the search ads, like some of the value added financial services, like the e-commerce transaction fees, like our own games—then generally those keep growing more quickly than our overall revenue and keep contributing margins substantially above our blended margin.

James Mitchell

executive
#65

So we think that gross profit leverage will keep working in our favor, and because of that, operating profit leverage relative to revenue will too, on a kind of gradual, asymptotic basis.

Martin Lau

analyst
#66

Naturally, spending on AI would in fact run somewhat counter to that. So you simply need to, in a sense, strike a new balance between the two, kind of…

James Mitchell

executive
#67

Yeah.

Martin Lau

analyst
#68

I believe it's worth just clarifying the complete picture.

Wendy Huang

executive
#69

Thanks. Since we’re short on time, the final question will come from Liao Yuan at CITIC Securities.

Yuan Liao

analyst
#70

Thanks, Wendy. And thanks to management for taking my questions. Congratulations on the strong quarter. My question is about AI and AI agents. We've seen a lot of companies release their own large function models in AI agents. I just want to know how our management views the key competitive factors going forward in large language models, and what our competitive advantages are. Thank you.

Martin Lau

analyst
#71

So what was — I missed the final piece of that. What was the..

Yuan Liao

analyst
#72

Right. What gives you an edge, given that the market has plenty of AI agent offerings? So going forward, how do we boost user engagement by adding further features and functions?

Martin Lau

analyst
#73

I mean, there are many kinds of AI agents, right? Essentially, an AI agent is a model that draws on the model's capabilities and then connects to various software tools so that a complex task can get done, right? That's a fairly general notion, and AI agents can exist. You can have AI agents that live inside different apps on a standalone basis. The way we see it is that there will be a multitude of AI agents. But for us, I think we'd be able to create standalone AI agents by drawing on models of high quality and also by drawing on the fact that we have a lot of consumers across our various software platforms, such as our browser and, over time, Yuanbao. But at the same time, right, even inside Weixin and inside QQ, we can have AI agents. And those AI agents can actually draw on the ecosystem inside the apps and deliver really great service to our users by completing complex tasks. Take Weixin, for example. Weixin has a lot of users, a very long daily time spent per user, as well as a high frequency of users opening the app.

Martin Lau

analyst
#74

That's one benefit. The second benefit is that, if you look at what goes on inside Weixin, it's genuinely very, very diversified, correct? It isn't merely entertainment, it isn't merely transactions, it's actually social communication plus content. And plenty of people do their work inside Weixin, plenty of people do their learning inside Weixin, and plenty of transactions pass through Weixin, and there's a multitude of Mini Programs that in fact let all sorts of different activities be carried out, right? So if you look at the Mini Program ecosystem, we can easily build an agent based on the model that actually can connect to a lot of the different Mini Programs and get activities and complex tasks completed for our users. So, I think those are all very distinctive advantages that we have. Now, of course, these are experiences we want to build very carefully. We want to build very patiently so that it would deliver the right experience to the users with a lot of attention to their data security, to their sense of comfort and security, right? So these are the kind of things we need to pay attention to when we build these products.

Martin Lau

analyst
#75

But as time goes on, I think those represent really good opportunities for us.

Wendy Huang

executive
#76

Thanks, Martin. That brings our webinar to a close. We appreciate everyone being with us on today's webinar. If you'd like to review our press release along with other financial information, you can find it in the IR section of our company website at www.tencent.com. A replay of this webinar will be posted shortly as well. Thanks again, and we'll see you 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.