The financial highlights of KRAFTON: Q3 revenue was KRW 719.3 billion, once again a new quarterly record, as it maintains its growth trajectory for three consecutive quarters. Underpinned by PUBG IP's top-line growth, Q3 operating profit reported KRW 324.4 billion. In terms of the revenue breakdown, Q3 PC revenue was KRW 274.3 billion, posting a record high quarterly revenue since PUBG PC release. Lamborghini collaboration: Third such collaboration on the PUBG PC platform offered wide-ranging options for exterior and interior customization and was met with great user traction, which drove sizable growth of PUBG PC revenue by 126% year-on-year and 43% quarter-on-quarter. Also, through Taego World Update and resurfacing of SLB mode following last year, a variety of in-game contents were served, driving concurrent user count to 890,000 in September, which supported healthy traffic. Q3 mobile revenue was down 15% quarter-on-quarter, but up 38% year-on-year to KRW 425.4 billion.
PUBG Mobile introduced a variety of concept modes like the Atlantis theme, driving good user feedback, while high-end costume skin in collaboration with car brands supported mobile revenue growth. BGMI rewrote quarterly revenue record, underpinned by regional expansion strategy and localized content. Console revenue in Q3 was KRW 11.8 billion, up 35% quarter-on-quarter, but down 2% year-on-year. All in all, revenue in Q3 once again hit quarterly record, posting KRW 2 trillion on a cumulative basis, which is more than the previous year's annual revenue. Next is breakdown of operating expense and operating profit. On the back of continuing new hires and new game development, and following the integration of Tango Studio's development team, labor cost was up 3% quarter-on-quarter and 30% year-on-year, reaching KRW 133.1 billion.
Platform fee and cost of revenue was impacted by record high quarterly revenue of PUBG PC and BGMI, which are serviced directly by KRAFTON, increasing 30% quarter-on-quarter and 69% year-on-year to KRW 101.1 billion. Commissions paid saw an increase of 5% quarter-on-quarter and 24% year-on-year to KRW 80.4 billion, on the back of growing outsourcing cost for new game development and live servicing. Marketing expense was up 41% quarter-on-quarter and 66% year-on-year, reporting KRW 29.9 billion, on increase in marketing spend following Gamescom and campaigns for BGMI's regional expansion. Despite rise in share price, changes in standards for cost recognition drove stock-based compensation down 59% quarter-on-quarter to KRW 14.5 billion. All in all, Q3 operating profit reported KRW 324.4 billion, up 71% year-on-year. Adjusted EBITDA for third quarter after SBC increased 85% year-on-year, reporting KRW 365.4 billion.
In terms of non-operating accounts, weak dollar drove foreign exchange-related loss, leading to non-operating loss of KRW 127.7 billion in Q3. Thus Q3 net profit was down 65% year-on-year to KRW 121.4 billion, while cumulative basis net profit for fiscal year 2024 was up 34% year-on-year, coming in at KRW 811.4 billion. This brings me to the end of the presentation. Thank you.
Jules, we would now like to have the Q&A.
[Foreign language]
Thank you. I have two questions I would like to ask. I think everyone can all agree the strong IP power that PUBG wields, and thanks to that you've been able to post good earnings results, and the growth has been very stellar. So I would like to ask whether you think that the earnings that you're getting from PUBG is actually structural, or we may also ask the question whether it could just be cyclical. I would like to understand, therefore, the management's view, and if you consider the growth of PUBG as something being structural, what are the levers that's going to continuously drive the growth of this game title as we move into the next year? Second question is, during the presentation you've mentioned that you are resorting to the use of small language models as opposed to LLMs.
I would like to understand as to the reason why you have opted to use SLM rather than LLM, and also is these initiatives, is it just your in-house effort, or do you also have plans to expand your partnership with any outside parties? And previously you've also in other earnings calls you've talked about Virtual Friend. I would like to understand as to how that is different from CPC, which you introduced today. Quality of such the user experience. And I think the performance that we were able to drive over the past year is the very testament to the fact that there were these structural levers behind the growth.
[Foreign language]
So regarding the performance outlook for PUBG IP next year, we are at this point maintaining this traffic growth or uptrend or at a little elevated level. And as we have mentioned previously, basically on a rolling basis, we have a full-year annual service plan in place. And based upon that plan and in light of the content pipeline that is upcoming, for instance, different modes, map updates, and collaborations, we also believe that we will be able to continue to trigger growth going forward. And in so doing, we will also be able to further scale up PUBG IP franchise, and we believe that this will form the basis for us to continuously generate good performance.
[Foreign language]
I'm Kang-wook Lee. I will respond to the second question. The first one actually related to why we are developing SLM as opposed to LLM. Basically, we are designing SLM and the size is small enough so that it could actually be installed in the gamer's client. It could be integrated to the gamer's client, and it could operate based off of that client device, which means that there's going to be significant cost savings that we can gain from using SLM as opposed to LLM. If you're running an LLM, that means that it has to run at a server, and the user would have to make use of this large language model through the API, which then means that while the gamer is playing the game, there's going to be continuous cost that's going to be incurred.
But if you're using SLM on the client device itself, then there's a significant amount of cost that can be saved. And also because you don't have to go through the network and because the SLM sits on the client device itself, basically the response time is going to be extremely fast. And also because it is not running under a server, even if there is a failure in the equipment or the network, the players are well protected in terms of the security aspect. And also because it's not running off of a server, you could also play the game offline.
[Foreign language]
You also asked about possible partnerships. There are two aspects that we can think of when it comes to partnership with outside entities. First is R&D effort in actually developing the game SLM. Second is the actual application of the developed SLM. With regards to the first point, the research and development aspect, although I cannot disclose the details at this point, we are considering partnering up with the global top-tier company in researching and developing the SLM model. The second aspect, which is application of the SLM, the use of the small language models, from a short-term perspective, we will first apply it to the game IPs within KRAFTON, the games that KRAFTON is either servicing, developing, or publishing. We will first apply it there.
But from a longer-term perspective, because we believe that we have an ample amount of competitive edge, we will also be open to open-sourcing this technology or entering into licensing arrangement.
[Foreign language]
You also asked about the difference between Virtual Friend and CPC. CPC basically is, in the definition, it's much narrower compared to Virtual Friend. I talked about two use cases regarding CPC. One of that is in inZOI, using the inZOI simulation. Unlike NPC, which is quite rigid and unpredictable, inZOI that is supported by or powered by CPC is more human-like and it is much more lively, and also under PUBG games, compared to the traditional bot, the CPCs can engage in more interactive communication with the gamer, and it is a character that is one step more evolved than what you would normally see, so basically CPC enables the in-game characters to further evolve and further evolve, so that's the difference between CPC and Virtual Friend, and also there are cases where there are Virtual Friends that are not CPCs.
What I mean by that is when there's a game, you see like YouTube streamers and broadcasters, and this is usually done by individual human beings, but it's AI that can actually do that role and replace that role. So basically in this case, it's not the game character that is being replaced. It actually is people who used to stream the game sessions that's going to be replaced. So in that case, this will not be CPC, but it will be a Virtual Friend. Going forward, we are going to focus on CPCs and further evolving the characters, the in-game characters. 다음 질문 부탁드립니다. Next question, please. 다음으로 질문하실 분은 미래에셋증권의 이미석 위원님입니다. Our next question comes from Hee-seok Lim with Mirae Asset Securities. Please ask your question.
the trailer was also revealed, so then in the future, going in a bit of an exclusive direction with Xbox, I'm first curious about that. Second, regarding new titles, you mentioned a bit various new titles, so overall, if you could provide a bit more specifics on the timeline, I would appreciate it if you could tell me.
Hello, this is Hisuk Im from Mirae Asset Securities. I have two questions. First one is on your Subnautica 2. I understand that you have a partnership with Xbox and there was a release of a trailer as well. Is this relationship exclusive? That's the first question. And the second question is regarding a little more detail on the timeline for your new title lineup.
release on two platforms, PC and console, in 2025. For console, we plan to launch on Xbox first. So, it is not exclusive to Xbox. However, the reason we selected Xbox as our partner for early access is that the purpose is greater to complete this game together with these users who are using Game Pass. Also, because we have our strategy to expand genres and platforms, we are approaching PC early access and Xbox with the so-called concept of early access.
Responding to your question on Subnautica 2, we are at this point planning on early access release in year 2025 for both PC and console platform as well. And for console, we are initially going to partner up with Xbox in distributing this game title, but it is not an exclusive arrangement. The reason why we chose Xbox early access is because our goal is to complete the level of this game title together with the Game Pass users. And because we are in a stage where we want to expand the game genre as well as the platform that we serve these games, we've decided to go with early access PC platform as well as Xbox early access.
[Foreign language]
Regarding more details on the timeline for new game titles, basically we are planning to release InZOI as of March 28 of 2025. This release is going to be followed by Dark and Darker Mobile, Subnautica 2, and Project ARC. And for the Dinkum, we will focus on scaling up the original title and servicing it through the new platform[Foreign language] . Next question, please. [Foreign language] . Our next question comes from In-chul Yu with Citi Securities. Please ask your question.
[Foreign language]
I'm just wondering whether you have any one-off revenue that you have booked. That's the first part of the first question, and in Q3, it seems you have successfully controlled your expense items, but since there are a multiple number of new games that are upcoming, like to gain your insight on what your projections are for marketing and labor costs going forward. Second question relates to the launch of Cookie Run game in India and also, you know, what is your release schedule like, the marketing related planning. Also, if you could talk about your long-term publishing strategy that you have for the Indian market, that would be helpful.
[Foreign language]
Regarding this revenue, the so-called bonus-like revenue, as a CFO, I would be very happy if somebody gave me a bonus-like revenue. I can tell you that this quarter there was no revenue incurred based off of a certain contractual arrangement. All of the growth that we have recorded came all from our organic business, and we have outperformed the growth that we've seen in the second quarter, and I think that goes to show the very strong capability and the competitiveness of PUBG IP.
[Foreign language]
Responding to your question about our marketing and labor costs for the Q4, we will continuously carry out marketing activities regarding the games that we are servicing live, and also next week we are going to the G-STAR event, so I think that for this fourth quarter, the marketing expense is going to be quite similar to the level that we've seen for the third quarter. In terms of labor cost, when we first announced our long-term vision of scale up the creative at the beginning of the year, we shared with you our objective of powering mid to longer-term growth, and in order for us to drive that growth, it's very important that we invest into talent, the development talent that is, so that we could continuously create, explore, and develop the creative that we require, so we will continuously make talent-related investment.
So I believe that with respect to the labor spending, the trajectory will be quite similar to what we have seen recently.
[Foreign language]
You also asked about Cookie Run for the Indian market. We are at this point developing characters as well as the game content that would work well against the Indian user base. Pre-order will open mid-November, and our objective is to release it before the end of the year. So Cookie Run, as you know, is not a very difficult game to play. It is very intuitive, and this game can be supported even with a very low-end device. So we are looking forward to positive responses from the Indian user base. In India, as you know, we have a significant base of BGMI user base, but now on top of that, by bringing in Cookie Run and in the genre of Battle Royale, we are looking forward to building up a new base of a gamer base.
[Foreign language]
Since during this quarter's earnings presentation, we included KRAFTON's AI-related various different endeavors and initiatives. So we were actually looking forward to a lot of questions on the topic of AI, but there weren't that many questions on AI. So if I may just also emphasize the position that we are taking as a company is that if you look back three years ago when play to earn was a big hype at the time, you would all remember the decision, the strategic decision that KRAFTON had made at that point in time. We are not going to be swayed by hypes that emerge in the market or in the industry, and also we will take on the same approach for AI as well.
As you know, if you look at the games that are out in the market or either in the pipeline of all the players in the game industry, KRAFTON is the only company that actually is bringing the AI use case and incorporating that into the development of its game InZOI. So although yes, we have a very stellar short-term business performance and achievements that we have made, but we are not going to be complacent at that, you know, at that level, and we're not going to fall under the pitfall of making a wrong decision from a longer-term perspective so that we could eventually achieve a stepwise growth in the future.
[Foreign language]
Last message that I want to share with you before ending is that you may remember for the first earnings presentation of 2024, the presentation, the voice that you've heard for, you know, when the CEO was making the presentation was actually KRAFTON's TTS. And yes, of course it was CEO Kim who responded to questions during the Q&A, but the opening presentation, the voice was generated by a TTS technology within KRAFTON. And same for today as well. The presentation that you've heard today for KRAFTON's earnings presentation was actually rendered by the TTS technology. As you can see, we are making use of AI technology in the way that we do our business, we conduct our business. So we ask for your continued interest in our company.