Today, and thank you for standing by. Welcome to Kingsoft Corporation 2025 Interim Results Conference call. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question and answer session. To ask a question during the session, you will need to press *11 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press *11 again. Please be advised that today's conference is being recorded. I'd like to hand the conference over to our first speaker today, Ms. Yinan Li , IR Director of Kingsoft . Please go ahead.
Thank you, Operator. Ladies and gentlemen, good evening and good morning. I would like to welcome everyone to our 2025 Interim Earnings Call. I'm Li Yinan, the IR Director of Kingsoft . I would like to start by reminding you that some information provided during the earnings call may include forward-looking statements, which may not be relied upon in the future for various reasons. These forward-looking statements are based on our information and the information from other sources, which we believe to be reliable. They refer to the other publicly disclosed documents for detailed discussion on risk factors, which may affect our business and operations. Additionally, in today's earnings call, the management will deliver prepared remarks in both Chinese and English. The third-party interpreter will provide consultative interpretation into English.
During the Q&A section, we will accept questions in both English and Chinese with alternating interpretation provided by the third-party interpreter. Onsite translation is solely to facilitate communication during the conference call. In case of any discrepancies between the original remarks and the translation, the statements made by the management will prevail. Having said that, please allow me to introduce our management team who joined us today: Mr. Zhou Tao, our Executive Director and CEO, and Ms. Li Yi, our Acting CFO. Now, I'm turning the call to Mr. Zhou.
[Foreign language]
Hello everyone, and thank you all for joining Kingsoft 's 2025 Interim Results Earnings Call. In the second quarter of 2025, we remain committed to executing our established strategies and the core business to maintain the stable. Kingsoft Office Group focused on AI collaboration and internationalization, continued to strengthen the allocation of R&D resources in related fields. As development solutions deeply aligned with users' scenarios need to stay competitively in the field of intelligence office. The online game business continued to advance content innovation and global expansion, achieving the development of flagship IPs and new genres.
[Foreign language]
In the second quarter, Kingsoft's total revenue reached CNY 2.07 billion, representing a year-on-year decrease. Revenue from the office software and services business amounted to CNY 1.36 billion, up 14% year-on-year. The online game business, affected by the high base effect in the same period last year, generated revenue of nearly CNY 1 million, representing a year-on-year decline. However, the expansion into new genres and global layouts have opened a new chapter.
[Foreign language]
In the second quarter, our office software and services business continues to grow steadily. For WPS AI products, we released the WPS AI 3.0 and launched the native office intelligent agent, WPS Lingxi. WPS AI 3.0 drives deep integration between AI and office software through bidirectional transformation, achieving native embedding of AI capabilities and deep enrollment in office workflows. WPS Lingxi is serving as the core capability module of WPS AI 3.0, integrating multiple AI functions, marking our transition from two-phased AI application to collaborative intelligent agents. We have also implemented intelligent upgrades to WPS 365 component, launching messaging, meeting, and email assistance to enhance office productivity.
[Foreign language]
For WPS individual business, we enhanced the product capability and user experience, and strengthened the refinement of operations and commercial strategy adjustments. These efforts have ensured stable domestic revenue growth. In the first half of the year, monthly active AI users reached nearly 13 million, with the completion of AI feature rollouts, as well as the launch and the promotion of new AI products. The number of AI users and monthly active AI users is expected to grow rapidly in the second half of the year. For overseas markets, we initiated the development of the new WPS International Edition to gradually migrate high-value domestic features.
[Foreign language]
For WPS 365, we've consistently improved the collaboration component functionalities and advanced the application development integration, enhancing end-to-end security capabilities. At the same time, we continuously enhance the deep into industry and the scenario, actively promoting the implementation of AI projects and engaging in cooperation with key clients. Based on industry typical solutions, we've replicated and scaled up their successes. For WPS software business, we actively participate in the bidding for office software among central and local governments and enterprises. Their effect has helped us maintain a leading market share in both floor and fixed layout document software market.
[Foreign language]
In the second quarter, our online game business continued to enrich the value of the classic JX3 series and expand into new game genres. Our flagship IP, JX3 Online, maintained an engaged player base and demonstrated resilience through consistent content updates and continued technical optimization. We launched the Fate of Sword: Zero, built on the core gameplay of the classical JX3 series IP, and introduced new innovations. The aim shooter game Snowbreak: Containment Zone continued to follow its long-term operation strategy and kept a stable user base. At the same time, we made significant progress in the new game genres and global expansion. Our self-developed sci-fi mecha game, Mecha BREAK, commenced its global public beta in July. On its launch date, it topped Steam's most played and trending games lists and earned high scores from several international authorities and media outlets like IGN.
This reflects our progress in new genres and global market expansion. Additionally, we are actively strengthening our cooperation with high-quality overseas IPs and deepening our localized operation capabilities in the domestic market. We obtained the publishing license for the social deduction game Goose Goose Duck in June, and it has already surpassed 5 million free registrations. It's set to launch in the second half of the year. We obtained the publishing license for Cats & Soup: Magic Recipe in May and began free registration in August, generating immediate excitement in the market. We will focus on refining the operations of our core products and continuously optimize the gaming experience based on player feedback.
[Foreign language]
Looking ahead, Kingsoft Office Group will continue to increase its R&D investment in AI and collaboration, and promote the implementation of its products and services across a wide range of office scenarios in various industries. The online game business will remain focused on premiere games, continuing to cultivate its IP franchise and steadily advance its global expansion. It aims to enhance the quality of long-term operations. We believe that their efforts will strengthen the group's core competitiveness and lay the foundation for achieving long-term high-quality growth. Next, I would like to invite Ms. Li Yi to introduce the finance performance for the second quarter and first half of 2025. Thank you.
Thank you, Zhou, and Yinan. Good evening and good morning, everyone. I will now discuss the second quarter and first half of 2025 operational and financial results. I'm starting from the second quarter's use of RMB as a currency. The revenue split was 59% for office software and services, and 41% for online games and others. Revenue from office software and services business increased 14% year-on-year, and 4% quarter-on-quarter to CNY 1,356 million. The year-on-year increase was mainly attributable to the growth of WPS 365, WPS individual business, and Kingsoft Office Group. The robust growth of WPS 365 business was primarily attributable to enhanced productive capabilities, high-quality service, as well as further penetration into industry and business scenarios, driven by the replication and promotion of typical solutions. The steady growth of WPS individual business was mainly driven by enriched AI-enabled features, improved user experience, and optimized operations.
Revenue from online games and others business decreased 26% year-on-year, and 8% quarter-on-quarter to CNY 952 million. The year-on-year decrease was mainly due to the high base effect of revenues gained from Snowbreak: Containment Zone and JX3 Online in the second quarter of 2024. The quarter-on-quarter decrease was mainly due to the decline of Snowbreak: Containment Zone and JX3 Online, partially offset by revenue growth from other games. Cost of revenue increased 5% year-on-year, and 8% quarter-on-quarter to CNY 454 million. The increases were primarily due to higher server and family-based costs, greater channel costs, as well as increased purchasing costs, along with the business growth of Kingsoft Office Group, partially offset by the lower channel costs of online games business. Gross profit decreased 11% year-on-year, and 3% quarter-on-quarter to CNY 1,854 million. Gross profit margin decreased by 3 percentage points year-on-year, and 2 percentage points quarter-on-quarter to 8%.
Research and development costs increased 50% year-on-year, and 3% quarter-on-quarter to CNY 854 million. The year-on-year increase was mainly driven by continued investments in AI and collaboration capabilities, as well as increased investments in new game servers. Selling and distribution expenses increased 12% year-on-year, and 25% quarter-on-quarter to CNY 424 million. The increases were mainly driven by greater marketing and promotional activities of Kingsoft Office Group, as well as increased promotional and advertising efforts in supporting pre-launch activities of new games. Administrative expenses increased 12% year-on-year, and 8% quarter-on-quarter to CNY 174 million. The year-on-year increase was mainly due to personnel-related expenses of Kingsoft Office Group. Share-based compensation costs decreased 30% year-on-year, and increased 32% quarter-on-quarter to CNY 71 million. The quarter-on-quarter increase was mainly due to the grant of awarding shares to the selected employees of certain subsidiaries of the company in these quarters.
Operating profit before share-based compensation costs decreased 48% year-on-year, and 31% quarter-on-quarter to CNY 454 million. Net FV gain for the second quarter of 2025 was CNY 343 million, compared with losses of CNY 12 million and CNY 22 million for the second quarter of 2024 and the first quarter of 2025 respectively. The gains in this quarter were mainly due to that we recognized the gain on deemed disposal of Kingsoft Cloud as a result of the dilution impact of the issue of new shares of subsidiaries in the second quarter of 2025. Share of loss of associates of CNY 170 million was recorded for the second quarter of 2025, compared with share of losses of CNY 169 million and CNY 153 million for the second quarter of 2024 and the first quarter of 2025 respectively.
Income tax expense for the second quarter of 2025 was CNY 104 million, compared with income tax expenses of CNY 48 million for the second quarter of 2024 and the first quarter of 2025 respectively. The increases were mainly due to the accrued deferred tax expense related to deemed disposal of Kingsoft Cloud in the second quarter of 2025. As a result of the reasons discussed above, profit attributable to owners of the parent was CNY 532 million for the second quarter of 2025, compared with the profit of CNY 393 million and CNY 284 million for the second quarter of 2024 and the first quarter of 2025 respectively.
Profit attributable to owners of the parent, excluding share-based compensation costs, was CNY 570 million for the second quarter of 2025, compared with the profit of CNY 441 million and CNY 360 million for the second quarter of 2024 and the first quarter of 2025 respectively. The net profit margin, excluding share-based compensation costs, was 25%, 18%, and 40% for these quarters: the second quarter of 2024 and the first quarter of 2025 respectively. As of now, in the first half of 2025, revenue increased 1% year-on-year to CNY 4,645 million. Office software and services made up 57% and increased 10% year-on-year to CNY 2,657 million. Online games and others made up 43% and decreased 10% year-on-year to CNY 1,988 million. Gross profit margin decreased by 1 percentage point year-on-year to 81%. Operating profit before share-based compensation costs decreased 29% year-on-year to CNY 1,109 million.
Share of losses of a social app of CNY 323 million and CNY 338 million were recorded for the first half of 2025 and 2024 respectively. As a result of the reasons discussed above, profit attributable to owners of the parent was CNY 860 million for the first half of 2025, compared with the profit of CNY 678 million in the same period last year. Profit attributable to owners of the parent, excluding share-based compensation costs, was CNY 886 million, compared with the profit of CNY 717 million in the prior year's period. The net profit margin, excluding share-based compensation costs, was 19% and 17% for the first half of 2025 and 2024 respectively. Our state of financial position, we had cash resources of CNY 25 billion as at 30 June 2025. Net cash from operating activities was CNY 665 million and CNY 1,374 million for the first half of 2025 and 2024 respectively.
Cash used for the capital expenditures was CNY 203 million and CNY 164 million for the first half of 2025 and 2024 respectively. That's all for the introduction of our operational and financial results. Thank you all. Now we are ready for the Q&A session. Thank you.
Thank you. We will now begin the question and answer session. As a reminder, to ask a question, please press *11 on your telephone and wait for your name to be announced. To withdraw your question, please press *11 again. Please stand by while we compile the Q&A roster. We will now take our first question from the line of Xiaodan Zhang from CICC. Please ask your question, Xiaodan.
[Foreign language]
Thanks management for taking my questions, and my first question is on the gaming business. Could management share your views on the performance of Mecha BREAK after its release in July? Also, any update on the full-year revenue guidance for the sector? The revenue growth of WPS Office business seems to reaccelerate since the second quarter. What is the outlook for the second half of this year? Thank you.
[Foreign language]
Got it, got it. Firstly, let me just answer the question regarding Mecha BREAK, as everybody paid attention to it. When we officially got into the entrance into the test, RPN now is already past one month. Based on this conclusion, it didn't realize the expectation. Although at the early stage when we launched, the population and also the good comments were quite good, after one month, generally speaking for the whole game, in a lot of the comprehensive elements, right now the key point is that after the test, in April, February, and March, actually, based on the feedback from the players, we did some modification. Generally speaking, the player told us that this game really needs to put a lot of energy and also spend a lot of money into it. This is their feedback.
That is why we, based on this feedback, did some modification and tried to avoid that. Actually, our team is overlapped. If this player told us that they're going to spend too much energy into it and too much money into it, it means that this game is worth the investment. It means that this game is good for them. They could play, like, they would like to play it, so in different perspectives. Right now, because we have this modification, this game becomes not that meaningful. We think that this is actually a very feared point with this modification. Second is that when we did the official test and also started to get into the charger operation, we think that the whole ecosystem has been changed, which we didn't expect, especially in China. The black industry is really bad, especially in the game industry.
We didn't expect it that much. Because of that, it really had a very big damage for the ecosystem in the game industry. Because it's become worse, and also we have some content when we officially get launched for the target design, finally, the actual result for the final players is going to have a very big difference. Comprehensively speaking, one side is that with this new IP, we have some improvement and we have some progress, but we also realized that in this new type categories and also the new playway, we need more time to control it, as it's quite similar to JX3 Online. Our strategy is that we need to have a very close interaction with the players to build up this IP. We need some certain time to do the adjustments and optimization.
OK, [Foreign language]
Okay, so regarding the officing and the outlook of the office, and that is going to include for the office, and based on our internal evaluation, we have some progress. Basically speaking, with all of the elements compared with the strategy last year, I'm here at the beginning of this year, a lot of like the personal business, the office business, and sub-business. From the personal business in the first half year of this year, based on this AI functions improvement, as I just mentioned, and especially in these two days, we have a lot of meetings, and together with the broad meetings, we realized that this direction is correct. As I mentioned at the beginning of this year, we hope that from the positioning, it's going to it's going to from zero to one, and step by step to realize all of them.
Currently, we are moving into 0.5 and 0.6, and probably 0.7. It's more stick to the reality, the official like officing situation. We believe that for the continuously with this year, for the goals of the membership, we have a strong confidence for that. Regarding the 365 business since last year up till this year, and we can see is that the goals, to currently based on the situation, we have strong confidence as [REER] that based on this collaboration of AI, we're going to promote strategically, and this is going to be quite successful as [REER]. We have some a lot of breakthroughs in the first half year, and we're going to continue to do it. In the second half year, we believe that this is going to be a very strong business.
For the software perspective, basically, we're going to have this domestic replace and also some innovation, creation. In the first half year, we're going to have a pretty high market share, and in the second half year, we're going to continue for this trend. At this moment, for the second half year, generally speaking, no matter for 365 or software optimization, we're going to have a very big confidence that we're going to be the maintain the same growth rate as the first half year.
All right, thank you. We will now take our next question from the line of Wenting Yu from CLSA. Please ask your question, Wenting.
[Foreign language]
I'll quickly translate the questions. Thanks management for taking my question. The first one is about JX3. Could you please share the recent trends in active user base and UP for JX3? Considering the upcoming high base effect in the second half of this year, how do you see JX3's revenue growth and the potential impact on our gross profit margin and operating profit margin? The second question is regarding the office software. WPS AI recently updated the AI agent-based features for PPT and knowledge library. How do you see the WPS AI agent differentiating from competitors like Manus and Auto GLM? To a broad sense of the software sector, which types of products are more vulnerable to the disruption by AI agents and which have a relatively deeper moat in your view? Thank you.
[Foreign language]
Okay, so firstly, I would like to answer the question regarding JX3 Online. This is actually a very good question. For JX3 Online, last year, when we promoted the mobile version, no matter for the income and also the active players, it's going to reach to a peak level historically. This year, this is actually a pretty big pressure for us. Up to now, although the JX3 Online is still pretty high, but compared with last year, it had a slightly drop. Also, we have like the season, and we also put a lot of energy in the Mecha BREAK. In the first half year, we're going to put too much energy into the Mecha BREAK. This is a slightly influence of the JX3 Online.
No matter for the product and also for the improvement, we think that to maintain JX3 Online as a pretty high level, we still have confidence on it because we have been operating for JX3 Online since 2009 up to now. It has been 15 years. We have met a lot of challenges. We have met a lot of difficulties. It all happens. From the long-term operation, it still works. We believe that if we would need to reach to such a high foundation as last year, and whether we could maintain this high level, we still have a little bit of pressure for it.
[Foreign language]
Okay, regarding the second question, this is actually a very good question. You mentioned especially the comparison with the other companies. Regarding this question, from the basic thing, for the next generation, we can see that for the big model, it has been created for two to three years up to now. How the big model is going to actually increase the working efficiency and also do the collaboration with human beings in their actual real, like typical scenario. I think this is going to get back to what exactly the AI is. For us, we're going to have a typical documentation creation that is not just a notebook. This is actually a typical documentation and also together with reading. We are kind of a platform like this. That is why this is not just a regular, very general solution for customers. This is actually typically for the office working.
We have to very clearly identify our position. Otherwise, we cannot compete with the other things because we need to compete in the same type of product. For the users, compared with the users, as you just mentioned, you have mentioned two companies for this matrix. I actually didn't think they are our competitors. For the big model, for the Kingsoft Office AI, the key things I think I would like to express myself is that I don't think these two, the other companies, are our actual competitors. Target for this job, we are in the officing in this area. For the AI, whether we could make a very good scenario and better serve our customers. My idea is that we need to have a core scenario to serve our customers, including the data. We have enough users, we have enough data.
This could help us to constantly improve and optimize the AI. Compared with the other companies, we are very happy to do that. Recently, we get the feedback internally. This is kind of a closed loop. From this perspective, we can think that our monthly active users could reach 650 million. For the daily active users, it's more than 100 million. For the PC, for the monthly active users, it could be more than 300 million. No matter for their B end or C end, for their office working scenario, we have enough customers, we have enough data, and we have enough scenarios to help us to optimize our product. For the AI from the 1.0 to 3.0 version, the monthly active users' growth has improved, what I just mentioned. This is not just a general thing.
We can see from this, the monthly active users, we have compared with the other companies, but that is way much better. We are actually in this officing field. This is kind of a very vertical characteristic. We have almost 30 million users doing it. Another thing is that compared with last year, their increase of the average users could be more than 50%. Comparatively speaking, comprehensive speaking is that we are focusing on the AI to improve the creation for the documentations. We have enough users, we have enough data to help us to create a better product and better experience platforms for our customers. That is why we have a strong confidence that at least for the officing field, compared with the other two companies you have just mentioned, they are not directly competitors for us.
Each of us has a vertical field, which is going to help us to improve better, to do better. We know that after several years' hard work, we're going to find a way that we think is quite actual, and we have enough customers, we have enough scenarios to help us to improve our affairs and constantly to create a better product for our customers.
Thank you. We will now take our next question from the line of Linlin Yang from Guangzhou Securities. Please ask your question, Linlin.
[Foreign language]
Quick follow-up question about the Mecha BREAK. Similar situation has occurred with other games during their initial launch period, but they eventually achieved great success. What is this game's current user base situation and how will the game be operated going forward? Second question is about the AI team. What's the management position of this AI team and what is its advantage? Can management give us some comments about its product progress and monetization outlook and the synergies with WPS and cloud business? Thank you.
[Foreign language]
Okay, we think that regarding Mecha BREAK, you are absolutely correct for the Kingsoft games, and from the historical, it can prove many times that it's very few, they have always had good comments, and like JX3 Online, and also for the Snowbreak: Containment Zone. At the beginning, at the initial stage when we launched, there's quite good feedback, and then we did the modification, and from a historical perspective, that is true. At the early stage, our team internally, we have a ton of communication, and after launch, because the customer, the players, are going to give us their feedback, we need to hear about their actual experience, and then they're going to get back to our team, and then we're going to modify it, improve it. From the long-term perspective, this strategy is quite good, it's correct.
This game is just launched for one month, so we have a lot of work, like do repeats of the works, to think back how we're going to improve it, and we are doing this job right now. I don't have that much things to say, but I think that the generally main direction is that we need to give some time to our team, and then we think that after a certain period of time, we're going to have a more clear picture, including how long time we're going to improve, and which part we have to improve. I think that we need to observe for a certain time, then have some better clear conclusion.
[Foreign language]
[Foreign language]
Regarding the second question, this is a quite good question. I need to have further explanation, so let me just quickly extend that. For the Kingsoft AI product, what is our core positioning? Our positioning is that we're going to help enterprises to use the AI better, as I just previously mentioned. This is actually the whole lay system, how we consider that. The first thing is that for their big module, their AI technology, this deeply influences their human digital economic system, which means that all of the enterprise's digital system based on this AI is going to have the restructure. For the AI itself, for the majority of the enterprises, it doesn't have that ability to train the big module.
For the majority of the enterprises, even in the future, we're going to have this machine module, but we believe that the machine module is going to shrink up here three to five types of the machine module. Actually, even if we have the machine module, the majority of the enterprises don't have this IT abilities for them to use the AI better. Comprehensively speaking, based on these four separate points, plus the lay systems, ecosystem like Xiaomi, we have the GB module has already been launched, and we also have the reference module, which is going to be launched at the end of this year. Based on that, including some other big module like DeepSeek and DripWiz, etc. Basically speaking, in the whole market, the target for the whole lay systems scenario is based on the Xiaomi's big module. Also, we're going to solve the machine module's issue.
The remaining is that we need a certain team to complete the engineering ability for the AI to help enterprises to use the AI better. Firstly, we need to have a machine module, and secondly, we need to have a good team at the back to help them to realize the AI usage. This is actually the background explanation. How about our progress? Let me just report that. Actually, same as the officing, we collaborated with the cloud, and we currently work together with [Food Only Bank]. We have to deliver the first scenario. Currently, for the National Bank, they need to report the need to generate the credit report in the loan. This is actually for them a quite heavy job. We have already completed and delivered. We work together with Kingsoft Cloud, Kingsoft Office, together with the Cloud AI. Up till now, we got very positive feedback.
Altogether, the full score is 100, and we reached above 85. Their feedback is that it's good quality and usable. Secondly, we collaborated with regional government for the officing area. We worked together with them, and after one year's optimization, basically speaking, right now it's almost 80, and we are still improving it. Based on this financing, based on this governance, we also have cooperated with a big law firm and worked together with like law university, and we need to have a deep cooperation with them. We need to have a legal module to use in the legal scenario, and we're going to create a deep product. This is actually the current progress. Since last time up till now, the total delivery we have realized from zero to one, including the governance, this is in the process, and also the legal scenario basically is still in the process.
My understanding is that we need to have a smaller scale that's suitable. We didn't officially promote it yet, but I believe that in the future, the AI big model is definitely going to be used for a lot of different scenarios. For the big model in the future, it's going to shrink into three to five types. The key thing is that we need to have the AI engineering ability, which means that we need to have this IT ability. This is actually a very big gateway for the majority of the enterprises. That is why we need to help them to use the AI better. We believe that this is going to be a very big market. Currently, we have selected several scenarios. We have completed from zero to one. In the future, we're going to constantly improve it to have some strategy layout.
Based on this progress from zero to one, we have already delivered two to three projects. The next stage, we need to think about how we have realized from zero to one. We need to think about how to realize from one to end. Lastly, it's actually the two things that together with the Kingsoft Office Group, Kingsoft Cloud, this collaboration in the future, we're going to collaborate with Xiaomi to realize the machine module. In the future, we're going to have a matrix of the AI to work together targeting how to help the enterprises to use the AI better. This is going to realize the whole chain of the AI assistance.
All right, thank you. I'm showing no further questions. Thank you all very much for your questions. With that, we conclude our conference call for today. Thank you for participating. You may now disconnect your line.