Tencent (Minutes): Emphasizes AI, but Lowers Full-Year Capex Guidance

The following are the minutes organized by Dolphin Research$TENCENT(00700.HK) for the FY25Q3 earnings call. For earnings interpretation, please refer to 《Tencent: Not Playing Heartbeat with Meta, Just Want Stable Happiness》

I. Review of Core Financial Information

1. Capital Expenditure:

Operating capital expenditure: RMB 12 billion, down 18% year-on-year, mainly due to changes in the supply chain.

Non-operating capital expenditure: RMB 1 billion, down 59% year-on-year.

2. Free Cash Flow: RMB 58.5 billion, basically flat year-on-year (growth in operating cash flow was offset by other factors). It increased by 36% quarter-on-quarter, mainly due to increased game revenue collection.

3. Net Cash Position: RMB 102.5 billion, up 37% quarter-on-quarter.

4. Stock Repurchase: This quarter's repurchase amount was RMB 19.2 billion.

II. Detailed Content of the Earnings Call

2.1 Key Information from Executive Statements

1. AI Strategy and Progress:

a. Strategic investments have shown results:

- Business empowerment: Enhancing precision advertising and game user engagement.

- Efficiency improvement: Increasing efficiency in coding, game, and video production.

b. Hunyuan large model continues to lead, upgrading model team and architecture:

- Enhanced complex reasoning capabilities, especially in coding, mathematics, and science.

- Hunyuan Image: Ranked first in the global leaderboard of text-to-image models in LMSYS Chatbot Arena.

- Hunyuan 3D: Ranked first in 3D generation models on Hugging Face.

c. AI application planning: Will continue to invest in promoting the popularity of Yuanbao. Generative AI capabilities will be developed within WeChat, expected to gain more market attention.

2. Core Business Highlights:

a. Communication and Social Networks: The combined monthly active accounts (MAU) of WeChat and WeChat reached 1.4 billion, achieving growth both year-on-year and quarter-on-quarter.

b. Digital Content: Tencent Music's paid users and revenue both achieved growth, consolidating its leadership in the music streaming field.

c. Gaming Business:

- "Delta Action" has become one of the top three games in the Chinese market based on revenue.

- "VALORANT" successfully expanded from PC to mobile.

3. Value-Added Services (VAS):

a. Total revenue of RMB 96 billion, up 16% year-on-year.

b. Social network revenue of RMB 32 billion, up 5% year-on-year.

- Growth drivers: Video account live streaming, music subscriptions, and mini-game platform service fees.

- Music subscriptions: Revenue grew 17% year-on-year, with paid users increasing by 6% year-on-year to 126 million.

- Long video subscriptions: Revenue decreased by 3% year-on-year, with paid users decreasing by 2% year-on-year to 114 million. The main reason is the series "Love's Ambition" (which was launched at the end of the quarter but has become one of the highest-rated series since the beginning of the year).

- Domestic game revenue: Increased by 15% year-on-year, mainly driven by "Delta Action," "Honor of Kings," and "VALORANT."

- International game revenue: Increased by 43% year-on-year (42% at constant exchange rates). The exceptionally high growth rate is partly due to one-time prepaid revenue recognition for "Dying Light: The Beast" and the consolidation of newly acquired game studio financial statements.

4. Communication and Social Networks:

a. Mini Program E-commerce: GMV continues to grow rapidly. Using foundational models (large models) to analyze user content consumption habits to understand their interests, optimize product recommendations, thereby improving sales conversion rates.

b. New Features:

- Added "Gift" feature on WeChat order and shopping cart pages, utilizing WeChat's social relationship chain.

- Upgraded "Search by Image" feature, allowing users to scan and identify real-world objects and shop directly in mini-program stores.

5. AI Features in WeChat:

a. Yuanbao feature integration: Introduced Yuanbao feature in video account and public account comment sections, which can automatically summarize content and encourage users to ask follow-up questions, receiving positive user feedback. Using content generated by Yuanbao to enrich WeChat "Look" information flow. Through the Yuanbao App, helping users explore news-related topics more conveniently.

6. Gaming Business:

a. Domestic Games

- "Honor of Kings": Revenue achieved year-on-year growth. Significant synergy with IPs "Lord of the Mysteries" and "Fox Spirit Matchmaker" from Yuewen Group. During the October anniversary event, daily active users (DAU) reached 139 million. The event featured heroes and minion skins inspired by Shang and Zhou dynasty bronze artifacts.

- "Delta Action": Ranked among the top three in the industry based on revenue this quarter. DAU exceeded 30 million in September, with PC DAU exceeding 10 million.

- "VALORANT":

PC benefited from new season content, anniversary celebrations, and global esports events, achieving record highs in DAU and revenue in September.

Mobile "VALORANT Mobile" was released on August 19, becoming the most successful new mobile game in the Chinese market this year (based on first-month DAU and revenue). The successful release of the mobile version doubled the total monthly active users (MAU) of "VALORANT" in October compared to July.

b. International Games:

- "Clash Royale": Supercell released a new auto chess mode Merged Tactics and expanded the trophy road from 9000 to 10000 trophies, significantly enhancing player engagement. September's monthly active users, daily active users, and revenue all reached record highs. Third-quarter revenue grew over 400% year-on-year.

- "PUBG Mobile": Third-quarter revenue achieved year-on-year growth. Growth was driven by Egyptian-themed skins, innovative "execution" emotes (with sound effects), dual gliders, and collaborations with "Transformers" and Lotus cars.

- "Dying Light: The Beast": Released by Polish subsidiary Techland, received very positive user reviews on the Steam platform. Due to the one-time prepaid revenue recognition model for game copy sales, this game contributed to the "exceptionally high growth" in international game revenue this quarter.

7. Marketing Services/Advertising:

a. Revenue of RMB 36 billion, up 21% year-on-year.

b. Growth drivers:

- Broad coverage of advertisers: Advertising spending increased across all major industries.

- Increase in ad exposure: Mainly from traffic growth in video accounts, mini-programs, and WeChat search, enhanced user engagement, and increased ad load rate.

- Average eCPM improvement due to upgraded advertising foundational models with more parameters, capturing more closed-loop marketing demand.

c. AI marketing solutions: Launched automated advertising campaign solution "AI Marketing+". Advertisers can achieve automation in placement, bidding, and placement optimization, thereby enhancing marketing return on investment (ROI).

d. Performance of various ad placements:

- Video accounts: Increased user engagement (DAU and per-user usage time) drove growth in ad exposure. Advertisers increasingly use marketing tools to drive traffic to short videos, live streams, and stores.

- Mini-programs: Increased activity and usage time attracted advertising from short dramas and mini-games.

- WeChat search: Increased commercial query volume and click-through rate led to significant revenue growth. Enhanced search ad relevance through upgraded large language model capabilities and optimized sponsored results.

8. Fintech and Enterprise Services: Total revenue of RMB 58 billion, up 10% year-on-year.

a. Fintech services: Revenue achieved high single-digit percentage growth.

- Growth drivers mainly from commercial payment services and consumer credit services. Offline payment volume improved in the third quarter (especially in retail and transportation), while online payment volume maintained strong growth, with overall commercial payment transaction volume growing faster year-on-year than in the second quarter.

- Consumer credit: Non-performing loan (NPL) rate remains one of the lowest in the industry and improved year-on-year, reflecting prudent risk management practices.

b. Enterprise services: Despite supply chain constraints in GPU procurement, revenue achieved double-digit percentage growth year-on-year.

- Growth drivers: Increased cloud service revenue and technical service fee growth driven by mini-program e-commerce transaction volume growth.

- Cloud product performance: Significant year-on-year revenue growth in storage and data management products (such as object storage COS, data warehouse TCHouse, distributed database TDSQL), mainly due to increased demand from customers including leading automotive and internet companies.

- Enterprise WeChat: Launched AI summary feature, which can generate project summaries and provide suggestions based on user emails and conversation content, enhancing project collaboration efficiency.

2.2 Q&A

Q: The international gaming business has shown strong revenue performance for several consecutive quarters. What has the company specifically done right to achieve such excellent results? How does the company view the future growth trend of the international gaming business? Can the company share more thoughts on the international gaming strategy? For example, will it continue to invest in high-quality overseas game studios, or will it focus more on bringing self-developed games to the global market?

A: Regarding the growth rate and strategy of the international gaming business, the growth rate reported this quarter (43% year-on-year) is indeed far above its normal trend. This is mainly due to two special factors: First, we consolidated the financial statements of recently acquired game studios; second, we adopted a one-time prepaid revenue recognition model for the buyout game "Dying Light: The Beast." Therefore, looking ahead to the fourth quarter, we expect the growth rate of the international gaming business to slow down and return to its normal potential growth level. As for the future internationalization strategy, we will continue to pursue multiple approaches, including seeking acquisitions and investments in excellent overseas game studios, strengthening cooperation with overseas developers, and striving to successfully bring more domestically developed games to the global market.

Q: The capital expenditure (CapEx) in this quarter's financial report is approximately RMB 13 billion, but the cash payment for CapEx reached RMB 20 billion. How should we understand the difference between these two numbers? Does the company have any new updates on the full-year capital expenditure guidance?

A: This is mainly due to the time difference between the accrual (accrual basis) and actual cash payment for server-related expenditures. Specifically, the credit term we pay to server suppliers is usually 60 days, which leads to a temporary mismatch between the two in specific quarters.

As for the capital expenditure outlook for 2025, we are updating the guidance. Looking back at 2024, our total capital expenditure accounted for about 12% of revenue. For 2025, our previous guidance was that total capital expenditure would account for "low teens" of revenue. Now we expect that although the actual expenditure amount in 2025 will be higher than in 2024, its proportion of total revenue will be lower than the previous guidance range.

Q: The company mentioned upgrading the "Hunyuan" team and infrastructure. Can you elaborate on the specific situation? What new features or performance can be expected from the upgraded version? Does the management have any updated views on the newly released "Yuanbao" App? How will it complement and enhance the various AI capabilities embedded in the WeChat ecosystem over the past few months?

A: In the AI field, we are upgrading the "Hunyuan" large model from two aspects. On the team side, we are actively recruiting more top research talents to supplement our existing strong engineering team. On the infrastructure side, we are comprehensively optimizing hardware and software facilities to support larger-scale, more efficient data preparation, model pre-training, and cross-domain reinforcement learning.

As for the combination of the "Yuanbao" App and WeChat, the two are complementary and mutually reinforcing. We have integrated the capabilities of "Yuanbao" into multiple scenarios within WeChat, such as in the comment sections of video accounts and public accounts, where users can use "Yuanbao" to summarize content with one click, which not only enhances the user experience but also stimulates more interesting interactions. We also use content generated by "Yuanbao" to enrich the information flow of WeChat Look. By showcasing its capabilities in high-frequency scenarios within WeChat, "Yuanbao" not only serves WeChat users but also successfully attracts more people to learn about and eventually become users of its standalone app, forming a virtuous cycle.

Q: Can the newly launched automated advertising solution "AI Marketing+" (AIM+) better serve small and medium-sized advertisers? Does the company expect this solution to increase advertiser adoption and return on investment (ROI), thereby supporting accelerated growth in advertising revenue in the coming quarters?

A: Our newly launched automated advertising solution "AI Marketing+" (AIM+) aims to help all types of advertisers improve efficiency. It automatically matches ad inventory and user profiles through algorithms, discovering opportunities with better results than manual placement, thereby enhancing ad performance. As you mentioned, small and medium-sized enterprises, due to the lack of complex historical processes, are currently the most active and first to adopt this solution. However, we also observe that more and more large advertisers are starting to adopt AIM+, which is similar to the development path of Meta's Advantage+ automated advertising solution in overseas markets. We believe this solution is valuable to all advertisers.

Q: The company mentioned that the promotion of the "Yuanbao" App and the implementation of various generative AI functions within WeChat rely on the capabilities of the underlying foundational model. However, the management's latest comments indicate that the growth rate of capital expenditure (CapEx) is relatively flat. Is there a risk that the company's investment in AI is not aggressive enough, potentially leading to competitors with stronger model capabilities or more aggressive capital expenditure strategies gaining an advantage in the future AI application market?

A: Currently, we believe that the existing GPU resources are sufficient to meet the needs of all our internal businesses (including "Hunyuan" model training), and there is no shortage of computing power. Our capital expenditure strategy is carefully considered, although there are limitations on GPU supply for external cloud service customers, internal use is sufficient. In terms of model capabilities, we are continuously iterating the "Hunyuan" large model quickly by recruiting top talents and optimizing infrastructure. We are satisfied with the progress made so far and believe that the next-generation model will demonstrate significant performance improvements. We believe that there is currently no model in the domestic market with a decisive leading advantage, and everyone is in a very competitive state, each with their strengths. We are not lagging behind, and as model capabilities improve, the user activity of "Yuanbao" is also steadily increasing.

Q: To what extent has the company internally started adopting AI to improve efficiency? When can we expect to see cost savings from AI applications (such as in sales and marketing expenses, R&D expenses, etc.), which could offset some of the company's new investments in "Hunyuan" large model and game advertising?

A: At this stage, the main impact of AI is reflected in revenue and gross profit improvement, such as the efficiency improvement in the advertising business, which has led to good growth. On the operating cost side, our AI investment is indeed part of the growth in R&D expenses. However, the effect driven by AI is more about "efficiency improvement" rather than "cost reduction". A few years ago, we already conducted large-scale organizational optimization, and the existing team is already very streamlined and efficient. Therefore, the application of AI mainly empowers our team to do more and more valuable things, rather than being used by some other companies primarily for cost and personnel reduction.

Q: The company mentioned developing generative AI capabilities within WeChat. Can the management share how these AI capabilities will create value for WeChat users? Particularly interested in the concept of "AI commerce," hoping to understand the management's specific views and plans on this.

A: Regarding generative AI (Agentic AI) and AI commerce within WeChat. Our ultimate vision (blueprint) is to create a "Weixin Agent" that can deeply understand users' social habits, interests, and intentions, and utilize the powerful communication, content (public accounts, video accounts), services (mini-programs), and commercial (e-commerce, payment) capabilities within the WeChat ecosystem to proactively complete various tasks for users. This is an ideal personal assistant form.

However, we are still in the very early stages of development. We are advancing in parallel through several approaches: First, gradually integrating the capabilities of "Yuanbao" (such as content summarization) into various scenarios within WeChat for testing and iteration; second, using AI to enhance WeChat's search function to more efficiently meet users' information acquisition and analysis needs; third, we have started developing vertical AI agents, which may be launched one by one in the future and eventually integrated to achieve the ultimate vision of "Weixin Agent."

As for AI commerce (Agentic Commerce), it is naturally part of our planning. Currently, we are focusing on building the e-commerce ecosystem itself, such as the healthy and rapid growth of GMV in Mini Shops. As the capabilities of vertical AI agents mature, we will certainly launch e-commerce services with AI capabilities in the future, but this will be achieved later in the overall development path.

Q: Is it correct to understand the company's guidance on capital expenditure for 2025: the total amount will be higher than in 2024, but the proportion of revenue will be lower than the previously given "low teens" range? Does this adjustment in guidance reflect changes in which factors? Has the supply situation of AI chips changed? Has the company's AI investment strategy changed? Has the company adjusted its expectations for future Token (model usage) consumption?

A: It is correct to understand it this way. We expect the absolute amount of capital expenditure in 2025 to be higher than in 2024, but the proportion of revenue will be lower than our previous guidance. This is not because our AI strategy or expectations for future model usage have changed, but mainly because the supply of AI chips is better than we expected, and the acquisition cost has been optimized.

Q: Can you interpret the potential impact of improved relations with Apple on the company's business, particularly mini-program games and domestic video games?

A: First of all, we have always maintained a very good cooperative relationship with Apple, collaborating in many different areas. We have been communicating with each other to make the mini-program game ecosystem more prosperous and vibrant. We are excited about the progress made so far. I think there may be an official announcement at an appropriate time in the future, and I suggest everyone pay attention then.

Q: The advertising business revenue accelerated growth again this quarter, showing strong performance. Can the management paint a picture of the short-term growth prospects of this business in 2026? When looking ahead, what new structural or cyclical factors should we consider?

A: We believe it will largely continue the current trend. From the demand side, although overall consumption in China remains cautious, it is moderately recovering, providing macro support for advertisers' willingness to spend. From the supply side, we will continue to increase the application of AI technology, especially promoting the previously mentioned "AI Marketing+" (AIM+) automated advertising placement solution to continuously improve advertising efficiency and effectiveness.

Q: In the short to medium term, how should we expect the proportion of R&D expenses to total revenue to change? We have seen very good gross profit margin (GPM) optimization across various business segments. Considering potential growth items such as future AI investments and depreciation costs, how will these gross profit margin improvements ultimately translate into the group's overall operating profit margin (OPM)? Overall, in the medium term, how will these positive (gross profit margin improvement) and negative (investment increase) factors interact, and what impact will they ultimately have on net profit margin?

A: First, everyone sees the continuous improvement in gross profit margin (GPM) across our business segments, but this is mainly not due to simple cost "optimization" measures. Indeed, in specific businesses like cloud services, we have taken multiple measures over the past two years to improve profitability, which directly led to gross profit margin improvement. But for most businesses, the improvement in gross profit margin is more a result of revenue structure optimization—we have repeatedly emphasized that we are shifting our business focus to higher quality, higher profit margin revenue sources, and this positive business portfolio change is the main driver of gross profit margin improvement.

Secondly, regarding the relationship between gross profit margin and operating profit margin (OPM), I suggest everyone not to focus too much on quarterly fluctuations. There is a dynamic change in accounting treatment here: In the early R&D stage of a product, its related costs are recorded as "R&D expenses," reflected below the gross profit; and when the product matures and is commercialized on a large scale, these costs may be transferred to "service costs," recorded above the gross profit. Therefore, the boundary between the two will move. I suggest everyone should focus more on revenue growth and operating profit growth, these two core indicators, rather than being overly concerned with the specific differences between gross profit margin and operating profit margin.

Q: Given the strong performance of global stock markets this year, can the management share its views on the company's current investment portfolio and future investment strategy and direction? Specifically, how does the company plan to deploy or recycle capital in the investment portfolio?

A: This year, global capital markets have been quite active in terms of both price and liquidity. We are utilizing this market environment to more actively recycle capital in our investment portfolio, mainly by selling some of our holdings in the public market.

Although we are also making new investments, including some emerging growth opportunities and our regular focus on gaming and digital content fields, overall, our asset disposals (divestments) have exceeded new investments (investments) by more than $1 billion since the beginning of the year. It is worth mentioning that we are also actively investing in some interesting AI startups, especially in the Chinese market, where we see a new wave of value creation and new opportunities.

Q: In the shooting game category, we seem to see signs of new and old alternation, such as "Delta Action" and "Dark Zone Breakout" performing well. How does the management view the evolution of the competitive landscape in this category? For the game "Delta Action," what specific plans does the company have (such as content updates, operational activities, etc.) to push it from its current second or third position on the bestseller list to the top? Does the management think achieving this goal is realistic?

A: First, regarding the statement of "new and old alternation," I do not completely agree in the Chinese market. Rather than "changing positions," it is more like "expanding the army" in the entire shooting game category. This year, our major shooting games, whether "Delta Action," "VALORANT," "Peace Elite," "Dark Zone Breakout," or "CrossFire Mobile," have all achieved growth in daily active users (DAU) or commercialization revenue, or both. The performance of "Delta Action" is indeed satisfactory, and "VALORANT" has performed exceptionally strong this year.

Shooting games are the number one game category globally, but not yet in China. Our current goal is to push this category to its rightful leading position in the Chinese market through efforts with a series of products such as "Delta Action," "VALORANT," and "Peace Elite."

As for how to further improve the performance of "Delta Action," our core strategy is "platformization." This game was designed with a modular architecture from the beginning, laying the foundation for platformized operations. Through platformization, we can introduce more new modes, one key direction being user-generated content (UGC)—this mode has achieved great success in "Peace Elite" and "PUBG Mobile," and we will gradually cultivate it in "Delta Action" in the future. In addition, we will continue to add PVE (player vs. environment) content, strengthen the live streaming ecosystem, and fully apply the valuable experience accumulated from successfully operating over 40 shooting games in the Chinese market over the past 17 years to "Delta Action."

Q: Can you roughly explain the relative contribution of in-app advertising (IAA) and in-app purchases (IAP) to the revenue of mini-program games on the Android platform? As reported by the media, is the current discussion with Apple limited to mini-program games, or is it a broader discussion that may involve the entire Tencent gaming business (such as traditional mobile game apps)?

A: The discussion mentioned in the media does indeed refer to mini-program games and does not involve our traditional mobile games distributed through the App Store. Regarding the revenue composition of mini-program games, at this stage, the vast majority of revenue comes from in-app purchases (IAP), rather than in-app advertising (IAA). Therefore, theoretically, if cooperation is reached, we will benefit from it.

Q: Given that the company is focusing more on the consumer credit business, will the current macroeconomic environment (such as consumer willingness, credit risk, etc.) become an important consideration affecting the revenue growth of this business? How should we assess the impact of the macro environment on the consumer credit business?

A: Our fintech business mainly includes three parts: payment, wealth management, and loans. The macro environment has the greatest impact on the payment business because the payment business is large, and its growth is closely related to the overall social consumption growth trend in China.

Currently, China's consumption growth is indeed relatively slow. The key reason behind this is that during the period when real estate prices fell, causing damage to household balance sheets, many consumers significantly increased their savings. This is different from other countries where economic downturns are driven by excessive credit and bankruptcy. Chinese consumers are not without resources but choose to save more rather than consume.

Therefore, we believe that once consumers feel they have saved enough money and gained a sense of security, while real estate prices stabilize, their willingness to consume is expected to rise, thereby unleashing consumption potential. The strong performance of the stock market this year has positively impacted household balance sheets, which is also a positive signal.

Specifically for the consumer credit business, since consumers are not overly stressed on their balance sheets but are just increasing savings, the macro environment does not have a significant impact on our consumer loan delinquency rate. More importantly, we have always been very disciplined and restrained in lending, controlling both the scale and amount of loans very strictly. We have strong data capabilities, making our credit underwriting extremely conservative and completely data-driven, so our loan delinquency rate has always been at a very low level in the industry.

Q: Looking ahead, how should we view the growth rate of the cloud business? When predicting cloud business growth, should we consider more the increase in customer spending in areas such as AI (bringing growth), or should we consider the impact of overall capital expenditure (CapEx) possibly slowing down (leading to growth deceleration)? Which of these factors will play a dominant role?

A: In the past few years, our cloud business strategy has been to focus on "quality" rather than "quantity," so revenue growth has not been fast, but gross profit has achieved very significant growth. Entering this year, we are pleased to see that cloud business revenue and gross profit are starting to grow in sync, and the business has already achieved profitability.

Currently, a major constraint on the growth of our cloud business is the supply of AI chips. When AI chips are in short supply, we prioritize meeting the needs of internal businesses (such as model training) rather than renting them out as cloud services to external customers. In other words, if AI chip supply is no longer a bottleneck, our cloud business revenue growth rate could be faster.

Q: Last quarter, the management mentioned that AI technology has significantly boosted the advertising business. Can you elaborate more this quarter on how AI specifically affects advertising conversion efficiency and pricing? If we exclude the incremental contribution from AI technology, what would be the "natural" growth rate of the advertising business? This can help us better understand the actual role of AI.

A: The strong growth of the advertising business this quarter can be roughly split into two equal parts:

1. About half of the growth (about 10 percentage points) comes from the improvement in eCPM (effective cost per thousand impressions). This is mainly due to two points: First, the optimization of our AI-driven advertising technology platform has improved the precision and conversion rate of ad placements; second, our continuously improved transaction closed-loop (such as completing purchases within video accounts) has enhanced advertisers' recognition of our platform value, making them willing to pay higher prices.

2. The other half of the growth comes from the increase in ad impressions. This reflects the improvement in user activity and usage time on our platform, as well as the addition of more ad inventory (ad load).

Q: The management mentioned that overall consumption is highly related to household spending, but also pointed out that transaction volume growth in offline scenarios such as retail and transportation is good. Can you share some more specific observations from sub-industries? In different industries, are we seeing any specific trends in transaction counts or transaction amounts that make the management more confident that the consumption improvement momentum seen in the past few quarters can continue?

A: Commercial payment transaction volume has indeed shown moderate improvement. A positive factor is that Chinese consumers have accumulated a large amount of savings in recent years, making them less anxious about real estate price fluctuations, while being more sensitive to the good performance of the stock market. These savings constitute a huge potential consumption power.

From specific trends: Online payments have maintained relatively stable growth during the past weak consumption period and the current relatively stable period. Offline payments, after being suppressed for a period, have also started to recover. Although the growth rate of online payments is still faster than offline, as offline payments improve in areas such as transportation and retail, the growth gap between the two is narrowing. This somewhat reflects the increased frequency of people's outdoor activities.

However, I must emphasize that this improvement momentum is still at a very early stage. We need to observe a few more months of data to be more confident in judging whether this has formed a sustainable trend.

Q: The management previously mentioned the "AI Marketing+" product. Are there any early quantitative data or observations on the effectiveness of this tool (such as the improvement in merchant ROI) that can be shared? The financial report mentioned "mini shops." As the transaction ecosystem within mini-programs becomes increasingly active, how does the management evaluate the future potential of this business (i.e., advertising directly related to mini-program transactions)?

A: The biggest change for advertisers with our launched automated advertising placement system (i.e., "AI Marketing+") is that it allows us as the platform to manage the bidding process on their behalf. Of course, large advertisers initially tend to be conservative, having doubts about "whether to completely hand over pricing rights to the platform."

Typically, these large advertisers will test in parallel for a period—using both automated placement and manual placement, then comparing the ROI (return on investment) of the two to verify whether the automated system can indeed bring better results.

Although we have launched this automated bidding tool relatively recently, the early results are positive. We observe that advertisers who have adopted the automated solution have indeed achieved better returns. Therefore, both the number of advertisers using "AI Marketing+" and the amount of advertising spending through this system are steadily increasing.

To evaluate the advertising potential of mini-program e-commerce, a simple and effective method is to conduct benchmarking analysis. You can refer to the "advertising revenue / GMV (gross merchandise volume)" ratio of existing mainstream e-commerce platforms in the Chinese market. Considering that the GMV of our mini-program e-commerce is growing very rapidly, you only need to apply this ratio to the GMV of our mini-program e-commerce to roughly estimate its future advertising revenue potential.

Q: Overseas, Meta's Facebook and Instagram serve different user groups through differentiated positioning. Our company's WeChat and QQ have long been similar product matrix models. How does the management view the product positioning and potential of WeChat and QQ in the domestic market? While focusing on WeChat, how can QQ or other products be used to serve and cover more different segmented user groups in the domestic market? How does the company's thinking in this regard differ from Meta's strategy?

A: There are some fundamental differences between WeChat/QQ and Facebook/Instagram.

First, the product attributes are different. Facebook and Instagram are essentially content-based social networks (Social Network), where users can consume different content based on interests, making it easier to differentiate and serve different groups. WeChat and QQ are primarily communication networks (Communication Network), where the network effect of communication platforms is stronger, naturally tending to aggregate rather than differentiate.

Second, the device environment is different. China is a highly mobile market, while other regions globally still have a large number of users using PCs. Facebook has a large number of users on the PC side, which is different from the situation in China.

Third, the age distribution of users is different. In the Meta system, Facebook's user group is relatively more mature, while Instagram is more inclined towards young people. But in our case, the situation is the opposite: QQ's users are naturally younger, while WeChat covers a wider age range.

Despite these differences, WeChat and QQ do serve different user groups and usage scenarios. Many young people, although they also use WeChat, will use QQ to enter a social space without parents, teachers, or colleagues. Therefore, our future product evolution direction is clear:

WeChat will continue to serve as an all-purpose application for the entire population, while QQ will firmly grasp its youthful characteristics and further strengthen differentiation. We will make QQ more fun, creating a significant distinction from WeChat. QQ will focus on meeting the needs of young users in specific scenarios, such as making new friends and developing interest circles. Our strategy is to ensure that WeChat and QQ better serve their respective user groups in different usage scenarios, achieving differentiated development.

<End here>

Risk Disclosure and Statement of this Article:Dolphin Research Disclaimer and General Disclosure