Good day and good evening. Thank you for standing by. Welcome to Tencent Holdings Limited 2026 1st quarter results announcement webinar. I'm Wendy Huang from Tencent IR team. At this time, all participants are in a listen-only mode. After management's presentation, there will be a question and answer session. For participants who dial in by phone, if you wish to ask a question, please press 5 on your telephone to raise your hand. If you are accessing from the Tencent Meeting or Group Meeting application, please click the Raise Hand button at the bottom left. Please be advised that today's webinar is being recorded. Before we start the presentation, we would like to remind you that it includes forward-looking statements, which are underlined by a number of risks and uncertainties, and it may not be realized in the future for various reasons.
Information about general market conditions is coming from a variety of sources outside of Tencent. This presentation also contains some unaudited non-IFRS financial measures that should be considered in addition to, but not as a substitute for measures of the group's financial performance prepared in accordance with IFRS. For a detailed discussion of risk factors and non-IFRS measures, please refer to our disclosure documents on the IR section of our website. Now let me introduce the management team on the webinar tonight. Our Chairman and CEO Ma Huateng will kick off with a short overview. President Martin Lau and Chief Strategy Officer James Mitchell will provide a business review. Chief Financial Officer John Lo will conclude with financial discussion before we open the floor for questions. I will now pass it to Pony.
Thank you, Wendy. Good evening. Thank you everyone for joining us. We start 2026 by making significant initial progress on our new AI products, as well as continue to utilize AI to grow our existing core business. The Hunyuan 3.0 preview model built by our revamped team of AI researchers on the architected AI infrastructure is a leader in its parameter size class, delivering practical utilities and cost efficiency, and has been top ranked in OpenRouter token measure since April 28th. Our productivity AI agent solutions have attained early traction, and we believe that our WorkBuddy is currently the most widely used productivity AI agent service in China. Our core businesses continue to grow their engagement, revenue and profit, providing the cash flow to fund our AI investments as well as use cases for future AI deployments.
Looking at our financial numbers for the first quarter, total revenue was CNY 196 billion, up 9% year on year. Adjusting from for the impact on revenue recognition for our value-added services from the later Spring Festival this year compared to last year, total revenue would have increased 11% year on year. Gross profit was CNY 111 billion, up 11% year on year. Non-IFRS operating profit was CNY 76 billion, up 9% year on year. Excluding new AI products, non-IFRS operating profit was CNY 84 billion, up 17% year on year. Non-IFRS net profit attributable to equity holders was CNY 68 billion, up 11% year on year.
Turning to our key services for communication and social networks, combined MAU of WeChat and Weixin grew year-on-year and quarter-on-quarter to 1.4 billion. For digital content, Tencent Video solidified its leadership in animated series with 8 animated titles ranked top 10 industry-wide in the first quarter of 2026. For games, Honor of Kings and Peacekeeper Elite achieved record height in gross receipts during the first quarter of 2026, while newly released creature collecting game, Roco Kingdom: World, achieved breakout success. Tencent Cloud's AI agent solutions achieved rapid growth and healthy retention rate. Now I will hand over to Martin for the business review.
Thank you, Pony, and good evening and good morning to everybody on the call. For the first quarter of 2026, our total revenue was up 9% year on year. As Pony mentioned, our total revenue growth was impacted by the later timing of Spring Festival this year compared to last year, which shifted to more games-related revenue recognition to future periods. Adjusting for the timing of the Spring Festival, total revenue would have been up 11% year on year on a like-for-like basis. By segment, VAS represented 49% of our total revenue, within which social networks sub-segment was 16%, domestic game sub-segment was 23%, and international games was 10%. Marketing services was 19% of total revenue, and fintech and business services was 31%.
Our gross profit was up 11% year-on-year in the 1st quarter to RMB 111 billion. VAS gross profit increased 9% year-on-year and represented 54% of our total gross profit. Marketing services gross profit increased at 19% year-on-year, contributing 19% of total gross profit. Fintech and business services gross profit increased 13% year-on-year, contributing 28% of total gross profit. A bit of update on our AI initiatives. Over the last 6 months, we have made significant progress on our Hunyuan large language model. I would say it's just getting started. We started the initiative by completely overhauling our foundation model team, centering around newly added elite AI researchers and engineers with deep expertise in large language models.
Our new team is young, energetic and cohesive, enabling us to make progress quickly in this highly dynamic AI era. In February, we re-engineered the system and process for pre-training and reinforced learning from the ground up. We re-architected the infrastructure to support robustness, scalability, and efficiency across pre-training data and reinforcement learning. On data, we expand our data set significantly and strengthened our data collection, cleansing, and synthesis capabilities with a focus on data quality. On training, we upgraded the process for pre-training and supervised fine-tuning. We scaled up to reinforcement learning. For evaluation, we're moving away from chasing public benchmarks that can be gamed. Instead, we evaluate our model through the latest exams, human tests, product feedback, and in-house tasks to see how the model actually performs in the real world. In April, we launched Hunyuan 3 preview.
When we set out to build this model, the purpose was to build a cost-efficient and solid model for diverse applications and de-risk scaling toward larger models. The core design principles behind Hunyuan 3 preview was to deliver comprehensive intelligence and cost efficiency, optimizing it for real-world deployment. We move beyond narrow expertise and towards comprehensive intelligence, such as integrating reasoning, long context understanding, instruction follow, dialogue coding, and tool use capabilities. By co-designing inference with model, we're able to reduce costs significantly so that the intelligence is economical enough to be used at scale. Hunyuan 3 preview has delivered on these expectations. The model has already become a leading reasoning model in China and has proven effective in real-world software engineering and other productivity agentic tasks.
Internally, the model has been deployed across 131 widely used internal products, including Yuanbao, QQ, and WorkBuddy, providing valuable feedback and iterative improvement via co-design process. Externally, Hunyuan 3 preview has been well received by users and developers in real applications. It has ranked 1st among all models available on OpenRouter by token usage since April 28th and continued its lead even after its free period ended on May 8th. While significant strides have been made with Hunyuan 3 preview, we view this as merely just the first step for Hunyuan large language models. The next step is to scale up to larger models. Our Hunyuan team is already working on a larger parameter model, leveraging our infrastructure and learnings from Hunyuan 3.
By aggregating bigger and better datasets and scaling more powerful reinforcement learning, we can strengthen the model's contextual understanding, enhance its agentic capabilities in areas including coding, and increase the model's general intelligence. Through co-designing and collaborating with other Tencent product teams, we're optimizing dataset selection and focusing reinforcement learning for high-value use cases. Beyond foundation models, it has become increasingly evident that agentic AI represents a breakthrough use case after AI chatbots have become popular. Agents are more valuable in uplifting productivity from initial use cases of supporting programmers in creating code, such as with our product CodeBuddy, to now catering to a wider range of workloads and occupations, such as with Clause and WorkBuddy. These breakthroughs were made possible by more powerful models and by the harness infrastructure that allows models to utilize tools and act as interfaces that enable users to manage agents effectively.
Our platform inherently has many benefits of hosting AI agents, as users can control AI agents through our communications and browsing interfaces such as Weixin, WeCom, QQ, Yuanbao, and QQ Browser in addition to third-party applications. Users can also choose which model to use from a wide range of models based on their own needs and preferences on token cost and performance. In the future, AI agents will be able to access our Mini Programs ecosystem using Mini Programs codes as AI skills. Tencent has established an early lead in agentic AI deployment evidenced by the leading DAU of our product WorkBuddy. While early in adoption cycle, CodeBuddy and WorkBuddy are already achieving strong organic growth and high retention rates among active users and paying users.
The high time spent and high-frequency interaction with AI agents among early adopters act as a virtuous feedback loop to Tencent enabling us to identify and provide complementary software and services, which in turn drives increased AI agents usage among a broader enterprise and prosumer user base. As users utilize more AI agents for more complex tasks, paying user conversion increases, resulting in rapid growth in token usage on Tencent Cloud in recent weeks. Now with that, I will pass on to James.
Thank you, Martin. Turning to business segments, Value-Added Services revenue was CNY 96 billion, up 4% year-on-year. Social Network revenue was down 2% year-on-year to CNY 32 billion, reflecting decreased reported revenue from app-based game item sales in China arising from the late Spring Festival. Long-form video subscription revenue decreased 2% year-on-year due to fewer releases of top-tier drama series. However, we cemented our leadership in animated content with 8 of our self-commissioned series ranking among the top 10 across all video platforms in China during the quarter. We believe Tencent Video possesses competitive advantages in our creating animated series, including our abilities to cross over IP from China Literature and our games into animated IP and our use of technology tools such as Unreal Engine and generative AI for storyboarding and producing the animated content.
Tencent Music subscription revenue increased 7% year-on-year, driven by growth in our current subscribers. For domestic games, gross receipts grew at a teens percentage rate year-on-year due primarily to Delta Force, Peacekeeper Elite, Honor of Kings, and Valorant Mobile. However, revenue growth of 6% year-on-year lagged the gross receipt growth as later timing of the Spring Festival shifted some revenue recognition into the second half of 2026. International game revenue increased 13% year-on-year, mainly driven by Clash Royale, Wuthering Waves, and Valorant. Moving to communications and social networks, our WeChat Minishop transaction volume sustained a rapid growth rate. To better support merchants selling branded products, we introduced incentives including preferential take rates for brands, product subsidies, and prioritized recommendations.
Branded merchants GMV more than tripled year-on-year in the first quarter, particularly in key categories such as FMCG and beauty. For consumers, frequent buyers can use our new coupon sharing feature to share discount coupons with friends via chats, benefiting from Weixin's virality. For content creators, we upgraded the matchmaking mechanism so creators can more efficiently outreach to relevant merchants in order to promote the merchant's products in the creator's video accounts and official accounts. On the AI front, we've enabled Weixin and QQ to act as communication interfaces for controlling AI agents, allowing users to orchestrate agents from mobile for complicated task execution on PC and cloud. We scaled up the number of parameters and enhanced the algorithm for video accounts content recommendation model, enabling deeper understanding of users interests to recommend more personalized and relevant content.
Total time spent on video accounts increased over 20% year-on-year. For Mini Programs, we've upgraded the developer toolkit architecture so users can better leverage AI plugins, including CodeBuddy, to create and debug Mini Programs, and over time would enable Mini Program code to evolve into skills for AI agent use within Weixin. Total query volume on Weixin Search increased over 25% year-on-year, benefiting from foundation model-powered ranking and broadening AI search coverage to include image-based queries. Moving to domestic games, during the first quarter, Honor of Kings achieved lifetime high quarterly gross receipts, benefiting from club tier outfits inspired by Chinese ink paintings and ancient Silk Road culture. For Peacekeeper Elite, peak DAU reached a lifetime high of 90 million, and gross receipts increased over 30% year-on-year.
The expanding user-generated content platform within the game, a new class-based extraction mode, and IP collaborations with partners such as Ferrari contributed to the growth in users and gross receipts. Delta Force also achieved lifetime highs in average DAU and gross receipts for the quarter, benefiting from a Cosmic Guardian skin, the Morphosis season featuring a revamped Space City map, and crossover promotions. Among newer games, Valorant Mobile has consistently ranked among top 10 mobile games industry-wide since its launch in August last year. During the first quarter, we released distinctive weapon items such as Ignite Capsule, which contributed to strong gross receipts. Roco Kingdom World, a creature collecting open world game, released on March 26.
In its first month since launch, the game has achieved over 13 million average DAUs and is sustaining high user retention rates, with its mobile version consistently ranked among the top 10 mobile games in China by gross receipts. Taking a step back from these individual game titles, AI provides increasingly helpful tools facilitating our game developers to deliver more content and enhanced experiences. Currently, AI for games is most beneficial in areas including accelerating 3D asset production and animation, enriching player experiences with intelligent in-game guides, and delivering more realistic graphics via AI rendering techniques. Among our international games, PUBG Mobile's gross receipts grew year-on-year in the first quarter, benefiting from Eastern mythology-themed outfits, brand collaborations, and anniversary events.
For League of Legends, Riot's conducted a substantial revamp of the team processes and technology in the last 2 to 3 years, which are now collectively resulting in a faster velocity of big new content releases, including the popular ARAM: Mayhem constant team fight mode, a well-received new season in For Demacia, and the Revenant Reign Exalted tier outfit for the champion Viego. As a result, League of Legends is experiencing a resurgence in DAU and gross receipts. Wuthering Waves new season, For You Who Walk in Snow, delivered a new map content and characters contributing to the game's gross receipts growth during the period. After a consolidation period, Brawl Stars DAU and gross receipts grew rapidly year-on-year, benefiting from new abilities called Buffies, an upgraded Brawl Pass, and an improved trophy system.
For marketing services, revenue grew 20% year-on-year to CNY 38 billion, with notable growth from internet services, e-commerce and games categories. Revenue growth improved from 17% year-on-year in the fourth quarter as we deepen collaboration with e-commerce platforms and as advertising demand picked up in response to increased inventory from WeChat Channels. Our automated campaign management solution, Tencent Ads AIM+, powered around 30% of total marketing services spending from advertisers with us in the quarter. We upgraded our runtime advertising recommendation models with a unified transformer-based architecture. This upgrade provides deeper understanding of user context and intent while balancing model complexity with system efficiency. By inventory, WeChat Channels ad impressions grew rapidly year-on-year, supported by increased total time spent video views and ad load. We released more inventory of rewarded ads, which deliver high click-throughs for advertisers.
Mini game and mini drama studios increased their marketing spend within Mini Programs. Specifically for mini games, we've introduced an instant play advertising format, enabling users to play the game within the ad without needing a redirection, which reduces friction in the new user acquisition process. For our mobile ad network, we've made more rewarded or incentivized ad formats available to third-party app developers, which unlocks higher price bidders and thus more revenue. Looking at Fintech and business services, segment revenue was CNY 60 billion, up 9% year-on-year. Fintech services revenue growth is primarily driven by commercial payment volume by wealth management services. For commercial payment volume, year-on-year growth was faster than in the fourth quarter of last year, benefiting from an ongoing increase in the number of transactions, but also higher value per transaction in categories such as retail and dining services.
For wealth management, average assets per user and number of users each increased year-over-year. Turning to business services, revenue in the first quarter grew 20% year-over-year, driven by increased demand and better pricing environment for our cloud services alongside rising technology service fees generated from Minishops e-commerce. For Tencent Cloud, AI-related demands contribute to increased revenue year-over-year across GPU, CPU, and storage. We upgraded Tencent Cloud's AI agent solutions with proprietary security infrastructure, skill hubs, and control interfaces, contributing to rapidly increasing usage and initial token monetization. Tencent Cloud's international business grew its revenue over 40% year-over-year as we expanded our global footprint and captured demand for our platform as a service solutions, including media processing services and TDSQL Cloud Database. Now I'll pass to John.
Thank you, James. For Q1 2026, total revenue was CNY 196.5 billion, up 9% year-on-year. Gross profit was CNY 111.3 billion, up 11% year-on-year. Operating profit was CNY 67.4 billion, up 17% year-on-year. Interest income was CNY 4 billion, up 7% year-on-year, mainly driven by growth in cash reserves. Finance costs were CNY 3 billion compared with CNY 3.9 billion in the same quarter last year, primarily due to higher Forex gains and lower average interest rates. Share profit of associates and joint venture was CNY 3.6 billion compared with CNY 4.6 billion in the same quarter last year.
On a non-IFRS basis, share profit was RMB 7.1 billion compared with RMB 7.6 billion in the same quarter last year. Income tax expense increased by 6% year on year to RMB 14.5 billion. On non-IFRS financial figures, operating profit was RMB 75.6 billion, up 9% year on year. Operating profit, excluding new AI products, was RMB 84.4 billion, up 17% year on year. Net profit attributable to equity holders was RMB 67.9 billion, up 11% year on year. Diluted EPS was RMB 7.364, up 12% year on year, outpacing non-IFRS net profit growth due to reduced share count after our share buybacks. Moving on to gross margin for Q1. Overall gross margin was 57%, up 1 percentage point year on year.
By segment, VAS gross margin was 63%, up 3 percentage points year-on-year, primarily driven by increased revenue contributions from internally developed games. Marketing services gross margin was 55%, down 0.5 percentage point year-on-year, reflecting higher cost of revenue, including AI-related equipment depreciation and associated operating costs, as we increase AI investments to deliver more relevant content recommendations in the future. Fintech and business services gross margin was 52%, up 2 percentage points year-on-year due to improved revenue mix within fintech services.
On Q1 operating expenses, selling and marketing expenses were RMB 11.3 billion, up 44% year-on-year, reflecting increased promotional efforts to support the growth of our AI-native applications and games, including new games. R&D expenses rose by 19% year-on-year to RMB 22.6 billion, primarily due to our increased investment in AI, which drove high equipment depreciation and associated operating costs as well as higher staff force. G&A, excluding R&D expenses, decreased by 24% year-on-year to RMB 11.3 billion, reflecting a high base in the prior year period, which included a one-off RMB 4 billion share-based compensation expense related to the restructuring of an existing commercial arrangement at an overseas subsidiary.
At quarter end, we had approximately 115,000 employees, up 5% year-over-year, driven by headcount additions to games and our technology platform, including AI-related headcount. Down 1% quarter-over-quarter, mainly due to lower headcount at our independently operated subsidiaries and in-house tech support subsidiaries. Q1 non-IFRS operating margin was 38.5%, stable year-over-year. Non-IFRS operating margin excluding new AI products was 43%, up 3.1 percentage points year-over-year. To conclude, I will highlight some key cash flow and balance sheet metrics. Operating CapEx was CNY 31.2 billion, up 18% year-over-year and 84% quarter-over-quarter as we accelerated investment in server infrastructure. Non-operating CapEx was CNY 0.7 billion.
Free cash flow was RMB 56.7 billion, up 20% year-on-year, driven by growth in games, gross receipts and advertising billings, partly offset by higher server infrastructure and compute spending. On a quarter-on-quarter basis, free cash flow was up by 67%, reflecting seasonally higher game gross receipts and the timing of certain seasonal accounts payable settlements, partly is offset by higher server infrastructure and compute spending. Net cash position was RMB 146.9 billion, up 37% quarter-on-quarter or roughly RMB 40 billion, mainly driven by free cash flow generation, partly offset by share repurchase of RMB 7.9 billion and net cash outflows of RMB 7 billion related to investment in other corporations. Thank you.
Thank you, John. We shall now open the floor for questions. We will take the first question from Alicia Yap from Citigroup.
Hello, can you hear me?
Yes.
Hi, good evening, management. Thanks for taking my questions. Two questions. First, since, you know, the release of Hunyuan 3 preview, I think management also mentioned, the model has been deeply integrated to a lot of the internal core products, including the Yuanbao, Ima, and WorkBuddy. Can management share some details on the performance enhancements that you have observed in these workflows since the adoption? Additionally, what is the roadmap of integrating, you know, Hunyuan 3 more broadly into WeChat workflows? Will Mini Programs enterprises will be able to leverage these agentic workflows to improve their productivity and maybe future monetization?
Second questions is, with agents increasingly potentially replacing the traditional click-throughs on the web pages and also the apps, could management share your view on the future advertising pricing and also the resulting impact on advertiser budget? What type of digital content or the web activities are likely to remain more resilient, continuing to capture more significant user engagement? Are we thinking ahead and strategically maybe positioning our ad formats to adapt to this potential shift in the ad spending? Any thoughts or insight you could share would be helpful. Thank you.
Thank you, Alicia. In terms of Hunyuan 3, you know, we have given a pretty comprehensive overview of Hunyuan 3. As you can see from the prepared remarks, it's more intelligent and it's actually very strong in terms of reasoning despite being a smaller model. At the same time, it has significant improvement vis-a-vis Hunyuan 2 on agentic capabilities. You know, with that, when we actually integrate Hunyuan 3 into the different products, the performance was actually, you know, quite, I would say quite encouraging. You know, the different products have all expressed and marked improvement in terms of the performance because they actually sort of, you know, see it from the user's end.
The total token usage is actually at least 10x compared to Hunyuan 2. That's the clear indication that, you know, Hunyuan 3 is actually well designed. At the same time, because it benefits from a co-designing process with some of the major products, for example, Yuanbao, and WorkBuddy, right. You know, that's why, you know, it's well received by the products. In terms of the integration into the Weixin workflow, I think, you know, it will be a step-by-step process. Weixin itself actually sort of, you know, have
Been always using some part of their products, Hunyuan 2, and they upgrade already to Hunyuan 3. In some cases, they use different models and they evaluate the different models and evaluate what's the best model to use for their users. As Hunyuan 3 continue to be getting better and better, then they will be adopting more. Now, in terms of Mini Programs, I would say for the enterprises, one benefit is actually that they would be using some of our Hunyuan-enabled products such as coding or such as CodeBuddy, as well as WorkBuddy.
You know, if these companies are actually using or the users are using such products, then they would actually be able to benefit from the improved performance of Hunyuan 3. At the same time, in the future, when we start integrating Mini Programs as skills, right, you know, to allow agents to, you know, have the ability to use Mini Programs as tools, that would actually sort of add more traffic to these to these Mini Program enterprises. One is internal, right? You know, when they use our own agentic products, they can improve their own productivity, the other one is external. It can actually help their Mini Programs to be used by more users and more agents going forward.
Hi, Alicia Yap. On your question on advertising, which is an interesting question, it's certainly more of a issue potentially for e-commerce companies than it is for us because users actively choose and desire to spend their time watching short videos or listening to music or consuming content or chatting with their friends, versus generally speaking, you know, when users spend time on e-commerce, it's because they're trying to find the lowest price. It's not because they, you know, necessarily enjoy that process.
To the extent that AI agents play a bigger role in the future in facilitating price comparison, then it's possible that users will spend less time on e-commerce sites and be less exposed to ads than they are today, while the AI agents can scan infinite listings and are therefore not influenced by ads the way that human beings with a finite attention span are influenced. You know, all of that said, there's been many, you know, prior iterations of price comparison services, including search engines and the big e-commerce companies have generally thrived despite the existence of those price comparison services. I think it's, you know, premature for us to sort of have a definitive view at this point on how it will affect our friends in the e-commerce industry.
We don't see it as a primary risk for Tencent. Thank you.
Thank you. We will take the next question from Kenneth Fong from UBS.
Hi. Good evening, Management. Thanks for taking my question. I have a question on the ROI on AI investment. If we look at the global peers, they have been allocating 80% to even 100% of their operating cash flow to AI CapEx, compared to us, which we invest roughly 35% in the last quarter. However, many of them have experienced decline in free cash flow ROE, as their business become more asset heavy. Could Management share or provide more quantifiable guidance on the AI-related CapEx for this year, and what KPI are being used to assess the value creation or, and return on this investment? Thank you.
Hi, Kenneth. You know, we are seeing increased demand both, you know, from internal products as well as from, you know, external users of our model for our AI-related services. We had previously guided that we'll be increasing CapEx this year versus last year and, you know, we're now more affirmative, more confident in that guidance. We and you should expect a substantial increase in CapEx, especially in the 2nd half of this year as more China-designed ASICs become available to us month by month through the year. On the KPIs that they differ product by product and business by business. You know, at a high level for our existing activities such as advertising and games, the KPIs would be more revenue and profit related.
For our new AI products, the KPIs would be more, you know, capabilities, how intelligent is our foundation model, and usage, you know, how much token consumption is happening on the WorkBuddy related. For Tencent Cloud, you know, where, until now we actually haven't had sufficient GPUs to begin to service the external demand, the KPIs will be more, you know, revenue and market share related. Thank you.
Thank you.
Thank you, Kenneth. We will take the next question from Alex Yao from JPMorgan.
Thank you, management, for taking my question. I'd like to follow up with Kenneth's question, just from slightly different perspective. When you allocate a financial resource to CapEx, and AI investments, how do you evaluate the AI infrastructure spent internally? What is the ROI framework or payback period you are underwriting these investment against, and over what time horizon? Thank you.
Well, I think we provided some thoughts on, you know, how we, you know, view the sort of, quantitative or qualitative return on investments in the previous answer. There are certain products where we're taking a very strictly financial approach. Others where we're more focused on, you know, the benefits to our franchise over time. I don't know if you want to sort of iterate on the question 'cause I may not have caught the full meaning behind the question.
Thank you, James. I mean, ultimately, the investment community need to understand who pays these AI CapEx, from what budget, and over what time period or horizon, do we expect to tap into those new commercial opportunities.
Well, I think that, you know, in the history of Tencent, we have generally sustained good returns, and you can quantify that by looking at our return on equity over the last couple of decades, which has, you know, been a consistently high return on equity. You know, we have not got there by limiting each new product, each new service, to, you know, very, near term quantitative, you know, return on investment targets. We have got there by managing the portfolio as a portfolio and by managing products over their full life cycle, not over, you know, any specific quarter or a 12-month period.
You know, there's been many products within Tencent, you know, whether it's, you know, our expansion into games, the launch of Weixin, our movement into payments, that, you know, went through lengthy incubation periods where they had no ROI, but we were confident in the franchise value creation. Over time, they had more lengthy, you know, harvesting periods where we've been able to drive, you know, very healthy returns on that, you know, sunk investment. You know, AI includes a range of sort of shorter cycle investments as well as longer cycle investments. If we buy GPUs and we deploy them into our ad tech, you know, then that's a relatively short cycle investment.
The GPUs yield better targeting, higher click-through rates and higher revenue and profit on a pretty accelerated basis. On the other hand, when we deploy GPUs into our Hunyuan foundation model, you know, that's something which we view as important for our franchise and, you know, where we're taking a longer term view. Again, you know, we don't manage each product on a quarterly basis. We manage the portfolio, and we manage the products on a full life cycle basis.
I think just to elaborate on what James is saying, right? You know, there's sort of, you know, a range of different ways to look at this, right? You know, if you look at the model training, it's basically an investment for the future, right? There's probably not gonna be very immediate return. Over time, the capability accumulates, and it actually sort of, you know, helps unlock a lot of different business opportunities. When you look at the products that we are launching, right? You know, be it Yuanbao, WorkBuddy, CodeBuddy, right? You know, there's process in which, you know, you have free services, right? Over time, you know, you may have revenue.
The business-oriented revenue can come faster than the consumer-oriented revenues. There will be sort of, you know, a return cycle on that. If you look at the business revenue or the cloud revenue in the sense of a mass or rental of compute, then there's a clear ROI, right? You know, you have a depreciation cost, and you usually put on some kind of margin, and then sort of you rent it out. That would be sort of, you know, having a clearer return. When we have advertising, as James pointed out, right? You know, we usually see very good return from those investments. I think, you know, for different computes, there's different ways to look at it.
Thank you, Alex. We will take the next question from William Packer from BNP Paribas. Will, your line is open.
Hi, hi, Madam. Many thanks for taking my questions. Firstly, with domestic gaming your biggest revenue and free cash flow driver, could you help us think through how generative AI is impacting that business today? Is it driving incremental monetization by faster content creation cycles, or is that a delayed benefit? Are you seeing cost efficiencies benefits hit the margins today, or do you need to invest the initial limiting any benefit? As a quick follow-up, thanks for the commentary on the CapEx outlook. Could you talk through any implications for the share buyback in the second half of the year? Thanks very much.
Yeah. Why don't I address those? You know, as you hypothesized, for our game business, generative AI enables us to produce more content faster. You know, that content is in some cases.
To enhance the overall player experience, in some cases it results in direct monetization, for example, if the content is a virtual outfit. You know, that's what we are doing and that's what we are seeing. We think that we're a, you know, China leader and to some extent even more so a global leader in terms of, you know, deploying that capability and achieving that benefit. You know, the objective at this point is really, you know, faster content creation and incremental revenue generation. We're not prioritizing margin expansion per se.
It's more that, you know, as we deliver the revenue uplift that we're seeing, and if we can, you know, keep headcount fairly stable, then, you know, I suppose mathematically that combination would tend to result in higher margins over time. That's, you know, sort of a happy output rather than the intention of the process. In terms of your question about capital returns, you know, we will be stepping up our investments in AI in re-response to the increased demands that we're seeing. However, we do have, you know, a very cash generative business, as you can see from the first quarter results.
We also have a very substantial investment portfolio. We're accelerating the process of liquidizing some of that investment portfolio. That would enable us to sustain buybacks going through the rest of this year. You know, at this point in time, we believe our share price is somewhat dislocated and, you know, therefore it's an opportune time for buybacks. A particularly opportune time for buybacks. Thank you.
Thanks, Nicola.
Thank you. Next question comes from the Robin from the Bernstein.
Thanks. Thanks management. Thanks for taking my question. If I could have a couple questions. One, in terms of the changes that we've made to our data pipeline and training RL infer and so on, do you feel now that, you know, having seen the release of Hunyuan 3, we're now on a more sustainable upward trajectory where, you know, we can deliver major updates, say, you know, one major update a year, several smaller ones in between, like some of the other labs, are we there or, you know, is that still more of a work in progress?
Second, just on the kind of philosophy of some of the AI product investments, I think one of the tensions in the U.S. has been, you know, I think OpenAI has been the big DAU player, whereas Anthropic has kind of gone after a small pool of high intent power users, you know, highly willing payers effectively. Just, you know, curious on, I mean, it turns out obviously has historically been the former, but just curious your thoughts on kind of the relative merits of going down one path versus the other and, you know, to what extent does the current kind of token maxing phenomenon play into that thought? Thank you.
In terms of the production pipeline, I think, you know, we have gone in pretty at length to talk about sort of how we felt it's actually making very good progress. If you look at Hunyuan 3, we have revamped the entire team, the production process, the infra and, you know, all the different major modules in producing great models, right? Including the data pipeline, as you point out, you know, the pre-training, the post-training, the RL, as well as eval. We have deliberately actually built a smaller model to basically validate all these different points, right? The combination of this, right, when it's all integrated into Hunyuan 3 preview, is that it produces a pretty competent model at its size.
We clearly see in each one of these modules, there's a lot of work that we could be doing. I think, you know, we're happy with the result. To some extent, you know, we are actually surprised by the speed at which it's done and the fact that it's actually proved to be useful, right? You know, for a long time, I think, you know, a lot of models would actually come up pretty high in terms of the benchmarks, you know, when it's actually rolled out to different products, people complain. When you put the model to the developers and the user, right, you know, people won't use it, right?
I think, you know, if you look at how this is received in the actual use cases, it's actually, you know, better than our expectation by quite a bit. With that, I think, you know, it builds a very solid foundation for us to scale the model to the next level. In terms of how we think about the different products, we felt this is actually sort of, you know, a very early stage in terms of AI diffusion, right? You know, we would see many different products coming up going forward. Initially, it was chatbot, and everybody felt, you know, chatbot is actually the, you know, king of the product, you know.
Then, suddenly you have a coding that come up and this becomes sort of even a more, you know, eye-catching and a significant use case because it's very high value, right? You know, now we're seeing sort of agentic capability proliferating, right? You know, I think, you know, that would actually allow.
AI to be diffused to different industries and you have many different agents coming up which can help you to do work, right? You know, there's gonna be new products coming up. I think that would continue to propagate. You know, I think to some extent, right, you know, In the AI world, you actually have to find the high value, use case, you know as opposed to sort of, you know, just purely focus on DAU.
The difference between the AI revolution and internet is that, you know, this is about intelligence, and intelligence is what manifests its value in sort of, you know, how, how much people are willing to pay for it. At the same time, the intelligence is not free, right? You know, in the internet world, you basically sort of, you know, have mostly, you know, existing information and then, you know, you also create some new information and content, that's a fixed cost.
Sort of the variable cost for delivering is actually very small, right? You only have to pay for bandwidth. The compute sits on people's devices, right? You know, and as a result, you know, you can almost like go for infinite scaling. In this case, right, you know, every single delivery of a DAU actually costs you quite a bit, right? As a result, right, you can't just apply the same logic as, you know, internet and apply it to AI.
I would say the ability to find high value use cases is gonna be as important, if not more important, than just sort of blindly get a lot of DAU and user time. I think, you know, that is one important distinction. We, you know, as we think about it, how to deploy our product, you know, and how to co-design our product with the models, these are the kind of new considerations that we have to put into play.
Thank you.
Thank you. Thank you. We will move to the next one, Alex Liu from Bank of America.
Oh, thanks, management for taking my questions. I have two questions. Appreciate the sharing on the agentic AI strategy. I'm especially, you know, interested in the concept that, you know, Tencent's AI agents could access the Mini Program ecosystem and use Mini Program code as the AI skills. Just wondering on that, is there any timeline we should expect for this to materialize? A follow-up would be related to the last questions. You know, given the tight supply of computing resources right now, I was wondering how is management balancing the pace of rolling out additional AI features, such as leasing agent, against the current pretty tight compute capacity on hand? Thank you.
Well, on Mini Program, I think, you know, this is something that will be coming. I think, you know, we need to figure out sort of, you know, what's the best way of presenting these and how to, you know, allow Mini Program owners to actually actively engage this. You know, there's, you know, we have a timeline. We're not, you know, gonna be able to share with you with a definitive answer because, you know, there's a lot of design that needs to come into place, right? You know, at some point in time, I think, you know, it's a concept that our ecosystem can actually help each other out, and this is one unique advantage that we have.
Over time, you know, a lot of potential ecosystem resources can be actually turned into skills for agents. Over time, agents would actually sort of have their own identity and be able to, you know, get accounts in some of the services too, right? I think, you know, that's unique advantage that we have, and we'll be providing that over time.
On your question about balancing between the various, you know, AI products we would like to develop internally, the reality is we've already made the choice and paid the price in that we have prioritized a multiplicity of internal services ahead of, you know, Tencent Cloud. You know, I think most, you know, big tech hyperscale companies with cloud businesses have one, you know, flagship internal use case where they're allocating a large number of GPUs. We have, you know, multiple flagships. We have the Hunyuan foundation model. We have agentic developments within Weixin. We have, you know, the Yuanbao support. We have the AI deployment for advertising, for games, you know, now also for the WorkBuddy and CodeBuddy use cases.
The reason why we've been able to support all of these at once is because we have not been active in leasing out GPU capacity in Tencent Cloud. Looking through the rest of this year, as the supply of China design GPUs progressively ramps up, we'll be remedying that situation, we will be making more capacity available in Tencent Cloud and consequently driving up Tencent Cloud's rate of expansion. You know, that's where the trade-off has been made, that we have been, you know, consciously late to monetize the AI opportunity through Tencent Cloud because we've been simultaneously supporting a number of AI initiatives internally. Thank you.
Thank you. We will take the next question from Charlene Liu from HSBC.
Thanks, Wendy, thank you for the opportunity. I have two questions. First is on monetization. We're seeing Doubao starting to explore subscription models for their to C users. I would like to understand how big the management thinks that the to C subscription market looks like in China. I guess subscription aside, are ads and Mini Programs or shop key monetization avenue for to C AI or there are more, how much upside do you expect from ad and Mini Programs? That's the first question. The second question is on compute power bottleneck. The U.S. players obviously have sort of moved on to working on resolving bottleneck issues or shortages in CPU and networking chips beyond GPU.
Has these issues begin to surface? Do we anticipate it to? What are the management's plans to resolve them? Thank you so much.
In terms of the to C monetization, I would say, it's actually not easy, right. If you look at global standard in the Western market when the paid service is actually very well penetrated and the living standard is actually very high. The subscription price in the Western market is multiple times of what the equivalent service in China is like, be it music service or be it video service. The paying penetration is probably in the single digit, right. When you sort of applied it to China, I think the subscription model is not gonna be that big for the China market.
You know, you use that as a standard and then sort of, you know, start to read across for China, then it will not be that much. At the same time, it's necessary, right? For the reason that I talked about earlier. It's not like internet services in which you can have very low cost of scaling your services, right? Your every single user actually sort of cost you something in terms of variable cost. I think the more important implication is that when you have to have payment to support a service, then most likely, the service is not gonna be a winner-take-all business.
It would basically sort of, you know, be supporting multiple players, you know, who would have a share of the market and, you know, each one of them would sort of, you know, have some kind of users and some share of subscriptions. Beyond that, right, you know, when we look at, e-commerce or advertising as a way to monetize, I think, you know, it's also very early, you know, for even the U.S. players where, the eCPM is actually much higher, right? You know, the, you know, the leading player has not been able to roll out very, robust, advertising model.
Y ou know, so I think, you know, it would be for the longer term, and it would be a supplement to probably a subscription model. I said, you know, in the world of AI, you know, when you apply the compute and apply the model to different use cases and different applications, you actually need to think about what is actually the high-value use case so that you have the best return on your limited compute in order to achieve the best result.
In terms of your question about the various bottlenecks between GPU, CPU networking, and so forth, you know, to recap that the reason why there's been a GPU bottleneck, you know, that's been much more pronounced in China than elsewhere is a combination of policy restrictions on certain, you know, foreign design GPUs being brought into China. The China design GPUs, you know, facing limited fab capacity within China. You know, as a result, the country has really been, you know, short of GPU or ASIC capacity. You know, that's now being addressed because the China designed ASICs are seeing, you know, more supply from fabs within China as well as more supply from fabs in neighboring countries.
You know, by contrast, we haven't faced those sort of artificial additional constraints on CPU or networking chips. You know, we've been a big buyer of, you know, CPU and networking chips for many years before, you know, GPUs became such a big presence in data centers. We have very long-term relationships with the companies that supply the CPUs and supply the networking chips. You know, on their side, while one might think that these suppliers would be, you know, sitting back and just, you know, selling at the highest possible price into the spot market, you know, that's not actually the reality. The smart suppliers are taking very conscious, you know, 3 to 5-year forward views and, you know, negotiating long-term agreements in order to give them certainty of, you know, their revenue outlook over the next 3 to 5 years.
You know, when deciding with whom to sign those long-term agreements. You know, they're looking to work with a number of partners, not just a single partner. You know, they're looking to work with partners who, you know, have been there for many years already and will be there for many years to come, and ideally with partners whose demand they believe will grow substantially over time. You know, happily, we fulfill all of those criteria. You know, we've been, you know, a big customer for, you know, the Intel and AMD and so forth for many years. We've been progressively growing our volume with them for many years. They believe we will continue to progressively grow our volume for many years to come.
You know, on the procurement side, I'd say that the challenges are more around, you know, GPU, and those challenges are now being addressed. You know, we've been able to secure a good supply of CPU and networking chips. Thank you.
Thank you. We will take the next question from Ronald Keung from Goldman Sachs.
Thank you Pony, Martin, James, John, and Wendy. Two questions. One is on the consumer AI agent side. Compared with a potential kind of Weixin agent that we've been talking about, which is at the app level, knowing that Weixin is super app. How does management view long-term potentials or potential disruptions from operating system level agents, noting agents from iOS or Android or mobile phone, or by phone makers? How would we see that potential or disruption or threat? On second, on advertising, we've seen a very good re-acceleration, noting that we are very patient on ad load versus peers, with a slight kind of edge up of ad load in the first quarter.
I'm just thinking, is there any room for any thinking or change to have a potential further revenue acceleration for ads? Reason, I would say thinking bigger advertising profits could drive more reinvestments into AI. Would love to hear your thoughts. Thank you.
I think, you know, from a operating system perspective, right, you know, you mixing a number of different things, right? You know, there is a real operating system which is iOS and Android, right? There are all, you know, and then you log in sort of, you know, other applications which try to pretend to be operating systems. I think, you know, if you are a operating system like iOS or Android, then you actually want to make sure that the ecosystem is actually well protected and well curated and, you know, given whatever that you actually allow applications to do, right? You actually sort of, you know, want to have that balance, right?
You know, you can have an agent which try to provide services to your users, but then you actually need to have the permission, right, you know, of the different applications. Otherwise, you know, as an operating system, you are essentially robbing different apps. That's not the best way of managing an operating system. I think, you know, operating system has existed for a long time and the principle of an operating system is actually it's very neutral, and it actually provides a level playing field for all the apps and in the future, all the agents to be working with the operating system.
If you say, "Oh, you know, there's another app which sort of, you know, try to become an operating system like a service and try to sort of invade other apps," I think, you know, that's a real competition and that's not something which, you know, any app would actually allow. I think, you know, the operating system itself, which should try to sort of, you know, stop that from happening as well. I think, you know, you know, if you're talking about agents which will sort of, you know, be an app trying to compete with other app, I think, you know, that's one level of issue, right?
You know, I think, you know, operating system would always try to be sort of, you know, quite impartial and try to maintain a, you know, a healthy ecosystem for everybody to be involved in order for it to be a successful ecosystem or operating system.
In terms of the advertising revenue, you know, factually our ad load on video accounts is still the lowest in the industry, at 4%-5%. There's, you know, clearly a substantial headroom for us to increase the ad load. You know, as to whether, and to what extent we'll sort of feel the need to do so, you know, we are in a position where we have multiple revenue growth drivers beyond advertising. You know, for our game business, the jump in deferred revenue in the quarter provides us with a sort of tailwind to the reported revenue over the coming 3 quarters.
For our cloud business, we've talked about how bringing on stream, you know, more GPUs, in the second half of the year should facilitate, you know, revenue growth trends for cloud. You know, for our fintech business, you know, for many quarters we've been struggling with a situation where volume growth was positive, but pricing was negative. Now, you know, volume growth remains positive and pricing's moved to a more neutral stance.
You know, we feel with that, across our portfolio of businesses, you know, there are, you know, certain positive things playing out and, you know, we'll, you know, continue to manage it as a broad portfolio and we'll, you know, continue to manage the ad load within video accounts in a way that, you know, we think, is conducive to both, you know, supporting the overall growth of our business metrics, but also to supporting, you know, growth of, you know, time spent and engagement within the video accounts product itself.
Thank you. We will take the next question from Ellie Jiang from Macquarie.
Great. Thank you so much for taking my question, and good evening management. I have two questions. Number one is, if you look at the WorkBuddy, its current traction, it does seem like we have been gaining very early leadership, especially in the productivity agent space. Within the 20% growth in business services this quarter, could you shed some lights on what percentage of this would be kind of reoccurring agentic revenue, i.e., be kind of mass or SaaS related versus the traditional cloud revenue as well as the others? Specifically, do you have any internal ARR target, you know, for these agentic workflow products by the end of fiscal 2026 and in the next kind of 2, 3 years? Thank you.
Yeah. Ellie, the upturn in, you know, sort of productivity AI is really something that's happened, you know, not in the last few quarters or even last few months, but last few weeks. You know, I think that's true, you know, globally actually. The, you know, really it's, you know, since the end of the first quarter that, you know, the agentic AI has, you know, broken through in terms of its ability to create code, in terms of its ability to, you know, make people more productive. You know, in the first quarter, the business services growth was not, you know, a function of, you know, that token consumption.
The token consumption has, you know, been more recent than the end of the first quarter. In terms of, you know, ARR targets and so on, you know, by extension, this is all changing so quickly that, you know, we could set a target today and we'd be, you know, out by an order of magnitude 12 months from now in either direction because, you know, the usage demand is so dynamic. At this point, you know, we're less focused on, you know, hitting certain dollar ARR numbers and, you know, more focused on having the right products.
The right products includes the right product at, you know, the model level where, you know, Hunyuan 3 is, you know, very good in terms of many agentic capabilities. You know, the next iteration will be substantially better later this year. You know, also at the product level in terms of, you know, CodeBuddy and increasingly WorkBuddy, you know, being the right, you know, interface for users to, you know, access and extract the most intelligence from these foundation models. Right now that's our, you know, priority. Thank you.
Thank you. We will take the last question from Thomas Chong from Jefferies.
Hi, good evening. Thanks management, for taking my questions. My question is about AI on the content side. Given that we have seen some disruptions coming from AI in the online video space, should we expect Tencent Video will also have expecting AI drama to be a blockbuster content in the coming years? How should we think about all the content cost and the business model going forward? My second question is on AI on the fintech segment. Given the risk management is substantially enhanced with AI, how should we think about the online lending and the wealth management outlook with AI? Thank you.
Again, you know, the landscape for AI has been so dynamic. You know, our own position has changed so much, you know, in the last few weeks, in the last few months, that it's difficult to speak very definitively. I think when you talk about, you know, AI disruption in content creation, you may be alluding to, you know, mini videos, or mini drama series rather than, you know, the long form drama series that is historically Tencent Video's strength. You know, on the long form side then, you know, what we're seeing is that there is a certain, you know, double-digit percentage of the market that likes animated content.
You know, now as a result of the confluence of, you know, tools such as Unreal Engine and capabilities such as generative AI for video, it's becoming increasingly feasible to, you know, create the same 3-D assets for games and for, you know, animated content, animated linear video content. You know, both become, you know, very good best in class products within their particular categories. So that's something where, you know, Tencent is a natural leader because, you know, we have our big content IP operations, we have our game business, we have our, you know, AI technology capabilities. We have, you know, particular strength in certain multi-modalities within AI.
As Pony mentioned in the opening remarks, you know, we've actually become, you know, a very clear industry leader now in the business of, you know, producing animated TV series. You know, AI enables us to do that, you know, faster, cheaper, you know, better. It also enables us to bring, you know, far more IP into linear video format that was previously sort of stuck in novel format or in game format. You know, expand the funnel, expand the audience. You know, AI for fintech, you know, financial services represents a very big part of global GDP. It represents a very big part of our revenue. It's an industry that is inherently very data heavy.
You know, it's naturally an industry that, you know, should and will lend itself over time to uplift in productivity from AI. You know, if you think about the industries that are already being uplifted by AI, such as coding, such as, you know, advertising, then, you know, financial services is actually very logical to be uplifted in the near future too, because it shares certain characteristics, you know, with coding, with advertising and so forth. You know, to give one example, you know, on the lending side then, you know, credit scoring has historically been more of an art than a science. You know, there's been this universe of data available, but only a small subset of that universe is actually, you know, fed into the model effectively.
Versus now, with, you know, transformer-based models, you can sort of take the totality of data available and, you know, see what is predictive and, you know, improve your loan extension on the basis of that uplift in predictability. You know, I think it's an area that, you know, many companies will be investing a great deal of time and energy and, you know, we're, you know, we'll be participating as well.
Thank you. We are now ending the webinar. Thank you all for joining our results call today. If you wish to check out our press release and other financial information, please visit the IR section of our company website at www.tencent.com. The replay of this webinar will also be available soon. Thank you, and see you next quarter.