Sea Limited (SE) Q2 2025 Earnings Report, Transcript and Summary

Good morning, and good evening to all, and welcome to the Sea Limited Second Quarter 2025 Results Conference Call. [Operator Instructions]. And finally, I would like to advise all participants that this call is being recorded. Thank you. I'd now like to welcome Mr. [indiscernible] to begin the conference. Please go ahead.

Hello, everyone, and welcome to Sea's 2025 Second Quarter Earnings Conference Call. I am [indiscernible] from Sea's Investor Relations team. On this call, we may make forward-looking statements which are inherently subject to risks and uncertainties and may not be realized in the future for various reasons as stated in our press release. Also, this call includes discussion of certain non-GAAP financial measures such as adjusted EBITDA. We believe these measures can enhance our investors' understanding of the actual cash flows of our major businesses when used as a complement to our GAAP disclosures. For a discussion of the use of non-GAAP financial measures and reconciliation with the closest GAAP measures, please refer to the section on non-GAAP financial measures in our press release. I have with me the Chairman and Chief Executive Officer, Forrest Li; President, Chris Feng; and Chief Financial Officer, Tony Hou. Our management will share strategy and business updates, operating highlights and financial performance for the second quarter of 2025. This will be followed by a Q&A session in which we welcome any questions you have. With that, let me turn the call over to Forrest.

Hello, everyone, and thank you for joining today's call. The momentum from our strong start to 2025 has continued into the second quarter. All 3 of our businesses have delivered robust healthy growth, giving us greater confidence of delivering another great year. Shopee's GMV grew 25% year-on-year in the first half, and we expect this growth momentum to carry into Q3. Money's loan book continues to expand rapidly while maintaining a healthy NPL ratio. And for Garena, we now expect full year bookings to grow more than 30% year-on-year. Given the high potential of our market, and the stage of our business now, we will continue to prioritize growth. We still see huge opportunities in our markets to serve many more users and better many more lives with technology expanding both our addressable market and capturing more market share will pave the way for us to maximize our long-term profitability. At the same time, our company has reached a stage where we can pursue growth opportunities while improving profitability by being disciplined and cost efficient, we have turned all 3 businesses EBITDA positive since the second half of last year, and we are accumulating cash each quarter as we scale. We remain committed to growing profitably with a strong balance sheet that enables us to capture future opportunities. With that, let me take you through each business performance. Starting with e-commerce. Shopee has shown stellar growth performance throughout the first half of this year. After a record high Q1, we have delivered another record break in Q2 with quarter-on-quarter growth in gross order volume, GMV and revenue. This was driven by a sustained increase in our active fire and their purchase frequency, reinforcing our leadership across all our markets. We also delivered year-on-year profitability improvement across Asia and Brazil enabled by our expanding scale and improving cost efficiency. Our monetization improved in the second quarter, largely driven by strong growth in advertising revenue. We have delivery worked to both drive seller adoption and encourage existing ad paying sellers to use our latest ad tech solution. Since early last year, our dedicated ad tech team has worked hard to improve algorithms, enhance traffic allocation efficiency and deploy AI technologies to better serve our ad paying sellers and we have seen very encouraging results. During the second quarter, the number of sellers using our ad products rose by around 20%. And ad paying sellers average quarterly ad spend grew by more than 40% year-on-year. Our tech enhancements have allowed us to more effectively optimize Shopee's GMV and advertising revenue at the same time. We saw an 8% uplift in Shopee's purchase conversion rates and improved our ad take rate by almost 70 basis points this quarter year-on-year. Our operational priorities have proven to be a winning formula and they remain consistent, enhancing price competitiveness, improving service quality and strengthening our content ecosystem. In the second quarter, reinforced our price competitive value proposition with the launch of the campaign slogan, cheaper, faster at Shopee. It resonated well [indiscernible]. We saw a more than 10% year-on-year uplift in overall purchase frequency during the quarter, and the buyers continued to rank us as the most price-competitive e-commerce platform in Qualtrics survey across Asia and Brazil. On service quality, our logistics capabilities continue to be an important depreciator for us. In the second quarter, we reduced the logistics cost per order and improve delivery speed across Asia and Brazil year-on-year in both urban and rural areas. We continue to roll out new initiatives to address specific customer needs, such as instant delivery options in dense urban areas. This led some buyers receive their orders within as little as 4 hours of order placement, our fastest shipping channel yet. We piloted it in Indonesia and proved so successful that we have now rolled it out to Vietnam and Thailand as well. Another example is intelligent demand forecasting, where we preship commonly ordered products to warehouses closer to where we know buyer demand will likely come from, reducing buyer rating time when actual orders are placed. Our logistics innovation not only delight our customers by improving the service quality they receive, but they also make us more cost efficient letting us pass savings on to buyers. We have also been doing more to enhance buyer loyalty and stickiness. Our Shopee VIP membership program a paid subscription service, giving buyers exclusive benefits has shown very good momentum in Indonesia. Total GMV from VIP members there grew nearly 50% quarter-on-quarter and VIP members bought a monthly average of around 30% more after subscribing. VIP members have also shown a roughly 20% higher retention rate compared to nonmembers. Building on this success, we have expanded the program to Thailand and Vietnam. At the end of June, total VIP subscribers in this market reached 2 million. We plan to roll out the program to more markets over the rest of the year. Our content ecosystem continues to be a powerful engine of buyer engagement and conversion. Our AI tools empower shopping sellers to produce high-quality video content, helping them improve user conversion and make more money without having to invest in their own studio setup. In Southeast Asia, orders from live streaming and short-form video accounted for more than 20% of our total physical good order volume in the second quarter. Our collaboration with YouTube has also continued its strong momentum. As of June, more than 7 million YouTube videos featured Shopee product links across our Southeast Asian market, an increase of more than 60% quarter-on-quarter. Moving to Brazil. Shopee has continued to deliver exceptional growth while maintaining its positive adjusted EBITDA. Average monthly active buyers rose by over 30% year-on-year in the second quarter, much faster than industry average growth rate. Our strong growth in Brazil is built on solid fundamentals, especially logistics improvements and product category expansion. We have brought logistics cost per order down by 15% while also reducing our average delivery time by more than 2 days year-on-year. In the Greater Sao Paulo region, about winning 4 shop parcels were delivered the next day and 40% within 2 days, up from single-digit percentages in the same period last year. We added over 100 well-known brands to our platform, especially in higher-value product categories. This contributed to steady and healthy increases in buyers average basket sizes in the second quarter. This combination of improving delivery service while expanding our product selection has allowed us to both serve our existing users more effectively and expand into more user segments, such as urban and more affluent buyers. We will continue to push on this front and keep our strong momentum going in Brazil. This quarter, we celebrated Shoppe's 5-year anniversary in Brazil and I'm very proud of what our team has achieved in this relatively short time. We have become the market leader by other volumes. We continue to grow fast, and we are operating profitably. I'm especially happy with the role we have played in promoting digital entrepreneurship to over 8 million Brazilian. 30% of our active sellers that Shopee has their first experience selling online, and more than half of our active sellers, they rely on Shopee as their primary source of income. In summary, Shopee has delivered an exceptional performance in the first half of the year with 25% GMV growth year-on-year, and we are confident that this growth momentum will carry into Q3. We remain committed to delivering strong, profitable growth while reinforcing our market leadership across Asia and Brazil. Next, moving to digital financial services. Money had another very strong quarter. Both revenue and adjusted EBITDA continued to grow more than 50% year-on-year. and our loan portfolio remains healthy, thanks to our prudent risk management. In the second quarter, our loan book grew over 90% year-on-year to reach $6.9 billion driven by both our expanding user base and our wider range of products addressing more user needs. We added over 4 million first-time borrowers during the quarter and our new cohorts are scaling well with positive unit economics. At the end of the quarter, active users for our consumer and SME loan products exceeded $30 million for the first time, representing more than 45% year-on-year growth. Our loan portfolio remains healthy with the 90-day NPL ratio staying stable at 1.0%. We have delivered strong and healthy growth across multiple markets, reducing our reliance on any single market and improving our ability to weather local economic cycles. I'm happy to report that Malaysia's loan book surpassed $1 billion at the end of June. It is our third market to reach this significant milestone after Indonesia and Thailand. Brazil also delivered robust growth in the second quarter, driven by strong adoption of Spaylater and personal cash loan products. We have achieved such high growth while managing risk very well thanks to 3 unique advantages that Mani has. First, deep and seamless integration with our shopping ecosystem. Second, a very large base of users who are growing their credit track record with us over the years. Third, our increasing use of AI to improve our credit models. Together, these advantages uniquely enhance our underwriting capabilities in each market, enabling us to very effectively push for growth across our 3 credit product lines. On-Shopee Spaylater, Off-Shopee Spaylater and cash loan products. On-Shopee Spaylater continues to deliver solid growth across our markets with GMV penetration now in the mid-teens on a market blended basis. We promoted Spaylater later 1 month interest-free option at Shopee check out mimicking the benefit of credit card usage. We use the tier-based pricing to offer lower interest rates to prime user segments who have access to more credit options and are more price sensitive. We also introduced a feature that allows users to request a higher credit limit by voluntarily submitting proof of income. Such initiatives contributed to our record high monthly numbers for first time later borrowers in June. In addition, these measures has enabled us to capture more prime users with stronger repayment capacity and higher lifetime value. We see further runway to scale this product by deepening its penetration on Shopee in all our markets. Off-Shopee Spaylater is also growing healthily. In Malaysia, we recognize the significant user demand for greater payment flexibility. So we integrated Splat with Malaysia's national QR network. Do it now. enabling seamless and flexible off-line usage on many everyday purchases. Our Off-shopee Spaylater portfolio grew over 40% quarter-on-quarter and accounted for more than 20% of our SPaylater portfolio in Malaysia at the end of June. Building on this success, we have just launched a similar user experience with the Thailand National ProPay QR network as well. We have also gained good traction with personal cash loans as addressing the strong demand we have seen in our market for credit success in people steady live. We have scaled this product category, both quickly and profitably by cross-selling personnel cash loans to stay later users with good repayment trends. As a result, personal cash loans outstanding has almost doubled year-on-year as of the end of June, and a lot of headroom remains for this product in our market. In summary, Money has delivered excellent growth throughout the first half of the year, diversified its loan portfolio across markets and products and maintain high asset quality through prudent risk management. It is exciting that our credit business is still in the early stages in many of our markets, reinforcing our strong conviction in Money's long-term growth and earnings potential. Finally, moving to digital entertainment. After a flying start to the year, Garena continued its strong growth momentum into the second quarter. Bookings were up 23% and adjusted EBITDA was up 22% year-on-year. Multiple key typos delivered double-digit growth in the second quarter, including Free Fire, Arena of Valor, EA Sports FD Online and Duty Mobile. Free fire continues to be at the core of Garena's strong performance. sustaining its massive global user base of more than 100 million average daily active users. Free Fire continues to grow well even after 8 years bringing joy to more users in more markets because we always put what gamers want at the heart of our work. A great example is the high-profile launch of our new map, Solora, in celebration of 38th anniversary during the second quarter. Solar plants in Juba, Notter with exciting innovation. Our veteran gamer were drilled by the return of an iconic trend from Free Fire's earlier maps that used to be a central part of their game experience. And gamers both old and new, we're very excited by the new full map light rail feature, allowing them to more rapidly across entire Karen completely changing their gameplay strategy and pushing them to come up with new techniques to win. Since its launch on May 16 at the Free Fire World Series, response from players has been exceptional. It has already become our best-performing new map. We also capitalized on excitement around this new map by introducing a new camera mode let players capture photos and videos of their gameplay more easily, boosting social sharing and engagement within a month of releasing this feature average daily share of in-game footage grew by nearly 4x, dramatically expanding free far visibility. Building on this excitement, we extended our anniversary celebration into July with the 2 high-impact IP collaboration, Netflix's Squid Game and Naruto Shippuden Chapter 2. Initial response from gamers has been extremely positive. In summary, Garena has delivered a very strong performance in the first half of this year. We believe reefer has established itself as an evergreen franchise both sustaining its user engagement and growing its appeal in more markets globally. We are also committed to trying out new genres and new markets and testing the boundaries of future game experiences by embracing AI. Given all of this, we are raising our full year guidance for Garena and expect bookings to grow more than 30% in 2025 year-on-year. In closing, we are very happy with the strong set of results we delivered both in Q2 and the first half of 2025 overall. All 3 of our businesses have extended their track record of excellent execution robust growth and improving profitability. This gives us greater confidence about the second half of the year, and we look forward to delivering a strong 2025 and beyond. Thank you, as always, for your support. With that, I invite Tony to discuss our financials.

Thank you, Forrest, and thanks to everyone for joining the call. For overall, total GAAP revenue increased 38% year-on-year to $5.3 billion in the second quarter of 2025. This was primarily driven by GMV growth of our e-commerce business and the growth of our digital financial services business. Our total adjusted EBITDA was $829 million in the second quarter of 2025 and compared to an adjusted EBITDA of $448 million in the second quarter of 2024. On e-commerce, Shopee's gross orders grew 29% year-on-year to $3.3 billion in the second quarter of 2025, and GMV increased by 28% year-on-year to $29.8 billion in the second quarter of 2025. Our second quarter GAAP revenue of $3.8 billion included a marketplace revenue of $3.3 billion, up 34% year-on-year and GAAP product revenue of $0.5 billion. Within GAAP marketplace revenue, core marketplace revenue, mainly consisting of transaction-based fees and advertising revenues was $2.6 billion, up 46% year-on-year. Value-added services revenue, mainly consisting of revenues related to logistic services was $0.7 billion, up 3% year-on-year. E-commerce adjusted EBITDA was $228 million in the second quarter of 2025 compared to an adjusted EBITDA loss of $9 million in the second quarter of 2024. The Digital financial services GAAP revenue was up by 70% year-on-year to $883 million. Adjusted EBITDA was up by 55% year-on-year to $255 million. As of the end of June, our consumer and SME loans principal outstanding reached $6.9 billion, up over 90% year-on-year. This consists of $5.9 billion on book and $0.9 billion of book loans principal outstanding. Nonperforming loans past due by more than 90 days as a percentage of total consumer and SME loans was 1% at the end of the quarter. Digital entertainment bookings grew 23% year-on-year to $661 million. GAAP revenue was up 28% year-on-year to $559 million. The growth was primarily due to the increase in our active user base as well as the deepened paying user penetration. Digital entertainment adjusted EBITDA was $368 million, up 22% year-on-year. Returning to our consolidated numbers. We recognized a net nonoperating income of $83 million in the second quarter of 2025 compared to a net nonoperating income of $56 million in the second quarter of 2024. We had a net income tax expense of $144 million in the second quarter of 2025 compared to a net income tax expense of $61 million in the second quarter of 2024. As a result, net income was $440 million in the second quarter of 2025 as compared to a net income of $80 million in the second quarter of 2024.

Thank you, Forrest and Tony. We are now ready to open the call to questions. Operator?

[Operator Instructions] Your first question comes from the line of Pang Vitt of Goldman Sachs.

Congratulations for another solid quarters. Two questions from me, both on e-commerce. Number one, on the GMV performance. First half of the year has already been very impressive at 25% year-on-year, and you also expect momentum to carry into third quarter. With that, are you also looking to raise your full year guidance that you gave in the beginning of the year on the back of this strong performance? That's question number one. Question number two, can you share with us on the latest competitive landscape? Have you observed any changes in momentum, in particular, in Brazil, have you seen any change since your competitor reduced its free shipping threshold?

Yes. I think on the first question regarding the growth momentum as we shared in the opening, we do see the Q3 momentum will continue. I think what we are saying is that in Q3 likely to be around similar growth rate as we observed in the first half of the year as well. Obviously, if you take into consideration the full year growth will be better than what we shared before. I think that's one. Second 1 regarding the competitive environment. For Brazil, as you shared, we do see certain actions from the competitors in the past few months. On our side, our Brazil business has been growing very well even in the last months after the competitor adjustment, we didn't see any observed impact to our growth as far as we can see for now. I think the key thing, if you look at Brazil for us is, number one, to make sure that we operate in a best cost structure, especially on the logic side, even with the improved speed of delivery as far as shedding the opening that we are 2 days faster than last year. We still see the cost are coming down. And our cost, we believe that much lower than where our competitor is. And even compared with their slow shipping, our cost structure is still very competitive, and our speed is a lot faster than losing. So we believe that we are in a pretty good position on that as that's number one. Number 2 is around the price competitiveness. And I think this is publicly available. You can do your own benchmark that our pricing is still very competitive in the market if you -- across all price categories. even with the adjustment from the other side. I think number 3 is, if you look at our brand seller side, we're expanding our seller base to a better -- to a higher ticket items especially in the past few quarters, we believe that trend will continue as well and will be not impacted much by the competitor movement. And just all in all, we believe that in Brazil, we are still well positioned. Our growth trend will continue, especially with our core segment group, giving the strong focus that we have on the fundamental cost structure, our price companies and also the growth of our seller side.

Your next question comes from the line of Piyush Choudhary of HSBC.

Congratulations for a great set of numbers. I have 2 questions, both on e-commerce, actually. Shopee VIP membership program in Indonesia has seen great progress. So could you share like what's the potential for the user base in midterm as you are expanding across markets? And how should we think about the cost implication of this program on logistics and other benefits which you offer. Also, if you can talk about your new initiative on instant delivery, which countries you're targeting and cost implication of this? Second question is on the Shopee EBITDA margin. With the push on VIP membership and instant delivery, should we expect margins to remain around severe levels in second half? And what's the outlook for 2026?

On the VIP program, we do see very good traction in Indonesia as the first country started we see a very good pickup from the VIP program, especially given that we are able to, over time, making sure the user can renew with a credit card or Paylater products, which is the biggest problem for any of this program in this market before. We are still in the very early stage of rolling this out. As you can see that it's only a few months since we started this program. And the first 2, 3 months was really tied process. So I think it's too early to tell, but we see very good potential there to be a meaningful program in our market comparable to the other program you're seeing other e-commerce or retailers in the region or outside the region. And as it's a new program, we do foresee some investment in the early time to grow the acceptance, especially getting users to sign up to it, especially renewing the program to get into the habit, but we don't see this as a fundamental impact or 2 visible impact to our cost structure in the medium term or long term. For the instant delivery. Instant delivery, by itself is essentially to expand our product offerings to the user base, especially on the high-end user base. Actually, those users typically are better profitability compared to the low end user base. So we don't foresee any negative EBITDA impact from these services. Rather on the other side, we do believe this will give us the better profitability in our businesses as we grow it. But obviously, the delivery cost, for instance, of course, the actual cost is slightly higher, but typically, for those segments, people have a better acceptance for how much they're willing to pay for those faster services. Regarding the EBITDA margin, there is a slightly fluctuation on EBITDA margin, as you can see from Q4 last year to Q1 this year to Q2 this year. I think there is a lot of sanity involved in this Q1, the exceptional quarter, and which is the first year in our business history that Ramadan for Indonesia and Malaysia especially falls into Q1 and the holiday falls into Q2. So that has a big seasonal impact for the businesses impact EBITDA meaningfully. So going forward, we do expect a long-term trend of EBITDA margin will continue to improve. Although, again, there might be seasonal fluctuation quarter-to-quarter, but the general trend will stay.

Your next question comes from the line of Alicia Yap of Citigroup.

Congrats on the solid results. I have a follow-up on the competitive landscape in Brazil. I understand we talk about the may impact is limited. But how should we think about the newcomer for example, the TM and also ticktock in Brazil, if you can comment on that. And then just curious, what is the current percentage contribution from the higher ticket items currently in Brazil. And then second question is on the gaming. Given the increased booking guidance for this year, just wondering, is it mainly because of the outperformance of the Free Fire or you actually will expect a higher contribution from the new games?

I think for Brazil, cross-border for Brazil has been relatively a smaller percentage due to the tax structure. So TMO has remained to be relatively smaller players in the market so far. We will observe how this evolves over time. For Tiktok it just started. And again, the amount of orders -- the size of order is still relatively small in the market. We will continue to observe webbing quite imminent with their businesses. And Brazil, it's quite a different market compared to Asia. I think it's something we will observe that structure-wise, we don't see any fundamental change to the market structure that will impact our view on the market or change our trajectory of the growth in the market so far. But again, just to reiterate. If you look at e-commerce businesses, the fundamental business we are focusing on is make sure our cost is right. We can serve our customers well through our logistics and making sure our pricing is right for the market, then that will be a long-term mode that we can build for the market. For the second question, I mean, it depends on how -- on the high ticket items, it depends on how you define what is high ticket item. I think if you look at our mall segment of Brazil, which is probably a work the population you can look at is in the range of teams. If you -- even compared to Asian markets, there is a meaningful room we can grow from there. And in the long term, I do believe that the penetration of more businesses in Brazil should be higher than Asia, giving the dining capital and income gaps versus our Asian market.

Alicia, on your question for Garena, the recent guidance in terms of the bookings and considering the scale and the size of free fire. So the main driver will be still a Free Fire for this revised up guidance. And we feel very excited. I think as I shared in the -- in my remarks, and we see a very, very strong momentum of growth across all metrics for Free Fire-- and in Q2, we launched the new map. This is the first new map, actually, we have launched things like in the past 3 years. It has been a tremendous effort behind it, and there is like a very sizable developer team across the different countries. And this is a 2-year of their the result of the 2 year of hard work. So it's very, very well received and by the gamers, both the new gamers and all the gamers. So we do focus on the both the new gamer boarding experience and at the same time, the retention engagement with our veteran gamers. And the confidence also comes from the like in some collaboration, IP collaboration we mentioned in Q3, specifically Netflix's Squid Game and Naruto Chapter 2. So we have a very, very exciting and promising results like from these IP collaborations in Q1 as we shared last quarter, and this is kind of like the another new episode of this IP collaboration. We learned a lot, and we have done great, but we also see things we can even do better. So we fine-tune our collaborations. And so far, we have seen great results, so from the IP collaboration. So I think that has given us the confidence of the full year outlook of Garena.

Your next question comes from the line of Divya Kothiyal of Morgan Stanley.

My first question is on e-commerce. Could you talk about the upside to take rates from here on beyond advertising. Could you also maybe discuss what the seller response has been to the rising commissions in ASEAN? And how does that impact the overall competitive landscape? And my second question is on fintech, specifically for Brazil because you mentioned that, that has also started growing quite robustly in this quarter. Could you talk about your strategy of ramping up fintech in Brazil. Where is the BNPL penetration now? And how fast do you expect it to grow? And if you can maybe point out some differences between the way it should ramp up in Brazil versus how you've been able to grow the business in ASEAN.

On the take rate side, obviously, as we shared in the opening that we see that still a strong potential on the ad side through technology improvement and also a better product that we brought out to the sellers. I think beyond the commissions, as you ready pointed out, we do did some adjustments in the past quarter on the take rate on the seller side. Overall, we have been seeing resin can response from the ecosystem. I think the key thing that you will be looking at is how the pricing structure will change in the ecosystem. In response to the take rate change, which is a good barometer to ensure the health of the ecosystem. We have been seeing relatively expected response on the pricing side. And also we didn't see any particular seller system, et cetera. So I think from all the indicators, that I think we have been getting a reasonable response to come response from the ecosystem. And at the same time, we're able to invest some of the take rates a ecosystem, which is very well received by both the merchant on the seller side and on the buyer side. And on the competitive side of this, you probably observe that a relatively healthy competitive ecosystem we observed that some of the competitors also increased their take rates. In fact, some of them copy exactly the same thing as we did as well in certain markets. So we are less kind of worried from that front. On the second question regarding the Brazil digital finance part of the services, we do see Brazil as a very important market for us on the [indiscernible] finance side. we have seen very good growth on the loan book in the second quarter. Our active user for loans grow 2x year-to-year. Our outstanding in Brazil also grew more than 2x year-to-year. One of the key things we did, I think we shared in the last earnings call is we combined the personal cash loan and as palate limit 1 limit, which is different from how we operate in Asia. We also integrate more external data to our risk assessment system compared to Asia. And this is because in Brazil, there are more external data available compared to Asia in the market. And we also have better integration between our money product offering and the shopping product offerings. All this enable us to grow our loan book quite meaningfully and reduce our risk in the market. On top of that, just to share a little bit on the Brazil as well that we do acquire a license, which -- and we also get initial approval for SCFI license, which will enable us to have better funding sources in future. In fact, we have formed partnership with some external lenders already to support the lending fund in Brazil market. Yes. So in general, I think we are very optimistic about the potential upside on the digital financial services in Brazil. And I think we are doing a very early stage compared to Asia in terms of the growth trajectories the penetration of escalator Shopee side is still -- we're still around the single digit to double-digit range. If you compare it to Asia, we still have a very large room to grow and our personal loans are still very early stage. We have many other products in the top line to be rollout to the market.

Your next question comes from the line of John Choi of Daiwa Capital Markets.

I have my first question is about advertising take rate. I mean this quarter, obviously, 70 bps meaningful improvement. But can you kind of share like how much more upside that you see? I know that you've been maybe quite a bit of ad tech, and you guys all shared some metrics about ad products rose, a number of sellers that you're using is the quarter by 20% and et cetera. In terms of the users or the advertisers, how much more compression do you think you could expect. That's my first question. Second question is on Brazil right now. I think you mentioned I think the delivery has improved substantially to 20%, 40%, as you mentioned. What kind of investments do we have to do furthermore in order for it to further improve? Would that we'll be able to improve our profit [indiscernible]. I mean, the EBITDA margin along the way? Or is it going to be kind of a prudent approach on trying to balance investment also profitability in the Brazil market?

On the ad side, there are 2 main drivers for the stake grade. One is we have a better traffic allocation algorithms between the ad and organics. We had -- we are able to mix the ad slot and the opening slot in a more efficient way, which will improve how much we can is actually improved the conversion rate of the ad placement, which will improve the return on investment from the sellers, so we can essentially cater for more seller demand on that. I think that's number one. Number 2 is with the better seller-facing products we roll out GMV Max, which helps the seller to automatic allocate the app spend more efficiently and maximize their ROI. And we also simplify a lot more set up UI to more and more sellers. I think all this helps to achieve essentially the better take rate. As you can see that we increased will we increase both how many sellers are involved in the ad product. We have a 20% increase in the active ad sellers. We also have 40% rise in the ad revenue per seller. So we improving on both sides in terms of the seller and the number of sellers and revenue per seller. Going forward, we do see still a meaningful improvement on the attach rate improvement because number 1 is our products are still rolling out to more and more sellers so that just simply, they are more seller joining the new products. I think number 2 is we still see a very good potential on using, especially AI technology, improving our algorithms on how to improve our conversion better. There are a few experience we are running, which yield a pretty good result so far and some of them will be brought out in the later part of the year, which will hopefully giving us a meaningful improvement on the take rate. I think the improvement at take rate, I think it will be ongoing for a period of time. I think it will not finish this quarter. I think in the next few quarters, you will see the number as we go. On the second question on Brazil. Regarding the investment business, in general, our logistics service has been relatively less CapEx-heavy, so we don't buy land. We don't buy our own trucks even in Brazil. So the majority of the investment in terms of CapEx is, number one, our sorting machines, our automatic sorting machines in our sorting center. Number 2 is just setting up the delivery hubs across the countries. So it will not give us a big, I guess, profitability burden in general. And in fact, as we're expanding more and more SPX coverage and improving the efficiency more and more. This will help us on the EBITDA in Brazil. Actually, this trend has been ongoing for a period of time if you track our past. We have been possible in Brazil since a few quarters ago, and we are still profitable in Brazil even we grow in such high speed for the market.

Your next question comes from the line of Jiong Shao of Barclays.

Congrats first on the very strong results. I have 2 follow-ups. I wonder just on as, I think Chris just talked about some of the elements for ads. I was wondering what's the current take rate now for the ads? And I recall a few quarters ago, you talked about longer or medium- to long-term target of 4% to 5%. I just wanted checking to see if that remains the case. The second question is about AI. I think both fors and Chris talked about using AI to improve the ad tech and among other things for your internal operations. Could you expand a bit on your thoughts all side of using AI, improving your internal efficiency. The reason I'm asking is that 1 field e-commerce peers in Asia recently, for example, started to expand their cloud services to external customers. Since on those lines, think about what are some potential other businesses that can go beyond your core businesses by using AI.

For the ad take rate, our current asset, we are still well below the peers we have seen in the regions. We have probably around 2% as we are -- and as you already pointed out, I think they are quite sizable room we can grow over time. On the AI side, I think as we shared in a few occasions, we use AI at this point in time for 2 main purposes. One is to improve our current businesses across all different dimensions. For example, we talked about the ad improvement just now. We also use our general recommendations, and this improved our conversion quite a lot by understanding user intention better by understanding the buyer query better. We also spent a lot of effort on the AIDC initiatives that we can generate a lot more attractive pictures for the product descriptions. We also generate many of the video products to help the conversion of our sellers. On the customer interaction side, we -- our customer service chat board is 80% managed by AI agent. We also helping the seller to interact with the buyers through the CS chat by agent as well, not only reducing the cost for the sellers, but also improve the upselling potential for the sellers while talking to the buyers. We have many initiatives like that to -- across our businesses just to make our business better. That's the third type of category. The second type is to improve our internal operations for example, obviously, the product development side, but also many of our daily operations like, for example, if you look at the way we run our marketing campaigns, a lot of mine camping are very ultimate right now through AI tools, many of the process to process the payment AI-enabled to the agent, et cetera. I think obviously, we are actively looking at what are the potential we can -- potential business we can expand from the AI solutions. At this point in time, there is no particular concrete business yet.

Your next question comes from the line of Thomas Chong of Jefferies.

Congratulations on a very strong set of results. My question is about the gaming business. Can management comment about other than the success of Free Fire, how are other games perform like Delta Force. And on the other hand, from this AI actually benefit the gaming business in driving the time spend, the monetization contest.

Sure. Thanks, Thomas. Yes, for like -- I mean, as from our philosophy, we always want to like not only continually make Free Fire more engaging attractive games, but we also focus on like a new game, like we do see a lot of opportunities in the different general and the different markets. But like I think it's -- but I would say, like I think at this moment, even we see some promising results, and I think it's like the game like a Delta Force like 3 cities is still at an early stage. I think it's like we're still kind of like a fine tune stage like rather than saying, okay, this will be another fantastic [indiscernible] for us. And so we'll keep our like investors updated when we continually see like things that still are fine-tuning the product. Although for some of them, we see some initial traction. And we are very, very excited about the AI perspective in the game industry and personally, I believe game industry will be among the first batch of industries largely benefited by the advancements and the technologies. And so far, like we have seen a lot of kind of upside on the -- actually on the development and the production side. And say, for example, like to develop any new content new map, we need to generate a lot of original arts. And now a lot of like very, very basic arts can be generated by AI. So it's -- the quality is very, very decent in terms of the efficiency, the volumes are generated and the varieties that generated I mean, you can imagine it's much, much better than what a human can do. So this is largely improve our productivity, and it's really, really, really exciting. And like on the -- as you mentioned from the gamers like engagement perspective, like -- so there is a very clear opportunity we have seen in the use case is like we do believe, like I say, for example, Free Fire is a very, very social game. It's a design for team play. So it's like there is much, much more fun if you play with other players, and there is a much more combination of the strategy, the technique you can use then you play as a solo gamer. But we observed the Free Fire still have a very, very sizable gamers like only play solar games. I mean they're enjoyed, but I think they haven't really fully experienced the amazing part of the game and maybe because of their shy, they don't know how to reach out to other players. So we think like the AI-enabled bot it's kind of like their -- it's an AI game agent like as their teammates, as a peer for them to play the game together, kind of play a brother throws, sister throws, coach roads in the game and give them a little bit flavor of how this interaction will kind of feel and taste in the gameplay and as an encouragement for them to reach out to play as a team rather than individuals. I think that largely helped on the retention. And furthermore, I think we are very actively experiencing and trying to figure out how to kind of leverage the generative AI to let gamers and to generate the content rather than -- okay, so not owed to today's game experience a preset and how the experience will look like. And I think with the AI tools, actually, this experience can be much more immersive and much more interactive and much more individualized. I think it's -- well, it's still early stage because like we can see some success, but we want to make sure the experience can be applied in the large scale and in a very, very consistent quality. I think that's the things we have been working on. And -- but we're sure, I think, like this will happen sooner or later yes.

Thank you. This concludes our question-and-answer session. I would like to turn the conference back over to Mr. Casey Ong for any closing remarks.

Thank you all for joining today's call. We look forward to speaking to all of you again next quarter.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.