管理層發言
Hello, and welcome to today's Kaspi Second Quarter and First Half 2026 Financial Results Call. I'll now turn the call over to David Ferguson from Kaspi to begin. Please go ahead.
Thanks, Sami. Good morning, good afternoon, everyone. I'm David Ferguson from Kaspi. Welcome to our second quarter and first half 2026 financial results call. With me is Kaspi's CEO and Co-Founder, Mikheil Lomtadze; and our Deputy CEOs, Tengiz Mosidze and Yuri Didenko. As usual, Mikheil will take you through the financial and strategic highlights from the second quarter. I'll talk through the broader financials, and then we'll open the call up to Q&A. So on that note, over to you, Mikheil.
Yes. Thank you, David. Thank you for everyone joining this call. So our financial performance for the second quarter is strong. The revenue went up 16%, adjusted EBITDA 5%. Considering the performance and strong financial position of the company, the Board is recommending to increase the dividend by 18% compared to the first year dividend. Our core businesses continue to perform: marketplace GMV plus 15%, driven by e-commerce, which is our strategically most important focus, both in Kazakhstan and Türkiye; it's 28% growth on a constant currency basis. TPV, our major business in payments, still continues to grow very nicely, around 15% growth year-on-year, and average net loan portfolio continues to grow strongly around 18% year-on-year. So e-commerce is the area which we believe is extremely important for us. That's the final destination for our consumers and merchants and where we can add most of the value in terms of enabling purchases and connecting merchants and sellers.
E-commerce GMV has grown nicely around 28% on a constant currency basis year-on-year, and the take rate continues to expand. The main driver of the take rate is value-added services, which will continue scaling, and that's about delivery and advertising. What is important is how engaged the consumers remain and how frequently they transact with us. The number of purchases per consumer continues to grow in both Türkiye and Kazakhstan, and e-commerce purchases grew 33%, which is a very nice growth, and we delivered in excess of 76 million purchases in the second quarter. About 20% of our GMV is 1P, and that is mainly the e-grocery in Kazakhstan, which is the fastest-growing e-commerce vertical for us, and the 1P in Türkiye, which is electronics and historically has been a category for 1P in Türkiye. We are about 53% of our GMV in Kazakhstan and 47% in Türkiye. Again, our priority is to continue building our e-commerce capabilities and making sure that delivery quality and speed are improving.
The value-added services are monetized, and we continue working on converting the traffic and the properties that we have into purchases from our consumers or connecting merchants and consumers to each other successfully. E-commerce has been growing the take rate with the value-added services. As you can see, value-added services grew 49% on a constant currency basis and 27% in reported terms, growing faster on a constant currency basis than e-commerce revenue. We are very responsible in terms of growing those additional services and making sure that we deliver value for merchants. We also make sure that our services are highly reliable on the delivery side and highly relevant on the advertising side when we promote something to our consumers that they actually need, delivering value to both merchants and consumers through this experience. We're also approaching a new stage of our company's development.
As you know, Kaspi has done reinventions and transformations multiple times during its history. We started from financial services, then we expanded into an ecosystem of everyday services and then united all the everyday services in a single super app. Now we are approaching the stage when we want to develop a personal AI assistant, which will help with everyday tasks for our consumers and merchants. On July 1, we launched Kasper, which is the assistant for our consumers. We started with one task now, which is enabling shopping. Kasper is built on our technology, on our data, and on our consumer experience, and it is integrated in our super app. We call him Kasper, our new personal assistant. He can understand your needs, recommend the best products, engage in a conversation with you and ask clarifying questions, compare different products, give you reviews, and so on. It's enabled in voice and text, so you can either type your task or record it by voice, and then Kasper helps you to find the right product for you across 20 million products on our Kazakhstan e-commerce platform.
He can follow up with smart questions and then he can recommend the right product for you, which you can then complete in our e-commerce. We launched it on July 1 and have been scaling during the month. It is now available to all our consumers in Kazakhstan. It's still early to give detailed performance metrics, but the metrics we already have are encouraging. One out of five customers for whom the feature was available actually used it. Response time is around three seconds. This tells you that the investments we've made both in compute and in optimizing speed have paid off. Consumers get responses very quickly, which is extremely important for any AI model or assistant. Consumers are searching across pretty much the entire catalog: 22 product categories have been covered so far, which is nearly our entire catalog, showing Kasper performs tasks across a wide range of products. Eight in ten conversations, or 80%, end with a product recommendation, and 60% of those lead customers to approve a specific product.
For example, if you're looking for an adhesive remover, Kasper will guide you through the process and place the product in the shopping cart. Speed is important: consumers are 50% faster in product discovery, 50% faster at adding to cart, and 30% faster to the basket compared to regular product discovery. We believe this technology leapfrogs traditional ways of finding products — you no longer scroll extensively, read and analyze information, and tap through multiple filters. Kasper speeds up the process. He works on the task with you rather than requiring you to type in a search box. Highly relevant, highly reliable, faster — these are the most important metrics for us. Our goal is to build a trusted assistant; an assistant must be trustworthy, because if you don't trust your assistant you can easily stop using it. We treat Kasper as a personal assistant and our priority now is to build trust, meaning a highly reliable, highly relevant service that actually helps you buy the exact product you need.
This is our priority as we scale customers. It's only one month, but the initial metrics are very encouraging based on customer interactions. Here are examples of the kinds of queries that show how different this interaction is compared to traditional search. Traditionally, customers type the product name they have in mind, and the marketplace offers a selection which you read through, compare by ratings, and narrow down with filters and navigation tools. Now customers give a task: 'I want to give my granddaughter a gift for her second birthday' or 'I want an aftershave with a strong scent' or 'I want a spray that creates a cool mist next to the house.' These tasks are different and the assistant solves them. We're true believers in this technology which we have been developing behind the scenes and getting ready to scale. It is remarkable how consumers interact and how the assistant helps.
Those examples ended up in real orders. Importantly, customers care about delivery — e.g., they need items delivered within three hours — and other practical problems such as needing a product to remove double-sided tape left on a plastic window. This is not regular search; this is asking someone for practical advice. All these things are encouraging. We're scaling Kasper as we speak in our e-commerce platform. Our mission is to develop a personal assistant for everyday tasks, and e-commerce shopping is the initial task we're focused on. In the future, Kasper can help with other tasks across all our services in our super app. I'd like to give a short demo for those watching the screen. David, can we go to the demo? Kasper has a dedicated space in our e-commerce. You can type the task you want Kasper to perform, for example, 'I need a vacuum cleaner.' Kasper does a basic analysis and starts asking clarifying questions, such as what type of vacuum cleaner suits you best, and then goes into requirements.
One important factor for vacuum cleaners is the size of your apartment, so he asks for that and asks about your budget. In a couple of seconds he pulls together different vacuum cleaners available in Kaspi Shop and they are described in simple language. He gives you a list of acceptable vacuum cleaners and you can ask him to compare specific models. When you ask for a comparison, on one screen you can see the main characteristics and compare products. You can use your voice to give the task; you could ask, for example, 'Which of these vacuum cleaners is best suited for pet hair?' and he will give a selection of products that fit that criterion and explain why. If you see technical terms like HEPA filter and ask what it is, he will explain in simple language and then you select the vacuum cleaner you want and proceed to checkout. All chats are stored and personalized so you can go back. We are beginning with the shopping experience because that's where we can add the most value.
Over time, Kasper will expand to all our services in our super app, and the service is highly scalable. When built on high-quality data, it becomes highly relevant. Obviously, Kasper is scalable to our other markets and we will think about scaling to Türkiye as well. Another quick update: we have secured a banking license. We completed the acquisition of Rabobank and are building up our fintech capabilities. We will be investing around $300 million, as stated before, to provide capital for the bank. We're scaling fintech products now and building capabilities to roll them out next year. We expect no material impact this year. Fintech and financial services are where we started, and we believe combining fintech and e-commerce delivers the most value to consumers and merchants. This is a strategic priority and we will be rolling out fintech products next year for both consumers and merchants.
We're not idle. We have been working on launching a new shopping mall based on the consumer finance license we already have in Türkiye. We've launched a new shopping loan on Hepsiburada. The flow is very similar to what you have in Kasper: you select the product, choose the monthly payment that best meets your needs, and proceed through checkout seamlessly. We're piloting this and the new shopping loan was already 0.54% of GMV in June. For financial services, originating the loan is step one; being paid back is more important. For months we have been building risk management capabilities across approval, managing consumers, and the collection process — the entire loan journey. We're rolling these out and are comfortable with the metrics as we pilot the new cash loan on Hepsiburada. We strongly believe that combining consumer and merchant shopping experiences with fintech products will create substantial value for consumers and merchants and therefore for the company.
The products we will focus on are shopping loans, BNPL, merchant finance, consumer finance products, savings accounts, and so on. Expect us to launch these products in Türkiye next year; our banking license and technology and risk management efforts are coming together nicely. We're very optimistic about launching financial services and fintech products in Türkiye.
All right. So thank you, Mikheil. So just to run through the financials, starting firstly with Marketplace. Marketplace constant currency GMV growth of 15% year-on-year. As you saw, that's driven by e-commerce GMV growth of 28% with e-commerce and travel broadly flat, consistent with trends in the first quarter. Take rate increased 110 basis points to 12.1%, again driven by e-commerce and specifically advertising and delivery. The revenue and EBITDA growth of 11% and 9% are reported growth, not constant currency, so impacted by a 21% depreciation of the Turkish lira versus the Kazakh tenge. That's the first thing to keep in mind. The second thing to keep in mind: the 9% EBITDA growth, that differential versus revenue growth — that margin pressure — is the least pronounced we've seen over the last couple of years and despite the investments that we're making into Hepsiburada. Moving on to payments.
TPV growth of 13%. That's a slight moderation, reflecting a slight moderation in inflation. If inflation continues to come down, TPV growth will reduce accordingly. The take rate declines by 7 basis points. Long-run trend though, driven by changes in product mix in favor of Kaspi Pay. The result is reported revenue growth of 5% and EBITDA down 1%. The pressure on EBITDA is two things. Number one, it's the investment in Kaspi Pay, our payments product. It's tech and product development spend. And number two, to keep in mind that adjusted EBITDA excludes the interest revenue that payments generate. That interest revenue is up 15% year-on-year. That's not reflected in EBITDA, but is reflected in net income and is net income accretive. Then on to fintech. Firstly, we talked on our last call about strategically focusing on loans that generate more revenue. These are longer-duration loans. Within the loan portfolio, the mix is shifting: BNPL, short-duration, low revenue-generating loans are getting smaller in the mix, while other loans, including general-purpose merchant financing, are growing in share.
You're seeing decent loan portfolio growth of 18%. The mix is changing in favor of higher revenue-generating loans. Pricing is stable, and so the result is faster revenue growth — revenue growth above net loan portfolio growth — with revenue up 23% year-on-year. The second point is that cost of funding remains an issue, up 150 basis points year-on-year in the second quarter. However, as some of you will have seen, Kazakhstan lowered its national bank rate at the end of June. We lowered one of our product deposit rates last week, effective last Wednesday. That was our first rate cut for over two years. It applies to our three-month duration product, which is around 30% of deposits. We lowered the rate from 20% to 19%. This isn't reflected in Q2 numbers. Some of it will be reflected in Q3. It's a three-month duration product, so it will be reflected to a much greater extent in Q4 and then fully as we go into next year.
The bigger point to keep in mind is this isn't just about one rate cut. For the last several years on this call we've been talking about how high rates have been a pressure on the bottom line. If inflation continues to fall in Kazakhstan, rates will continue to come down. You can see the growth in our deposits is strong, up 21%. So naturally, we will be able to pass those rate cuts through, and that will be very beneficial to the bottom line over the next couple of years, not just one quarter. In the second quarter, EBITDA was up 6% versus revenue growth of 23%. On the risk side, cost of risk was 0.7%, up slightly from 0.6% in the second quarter of last year but flat quarter-on-quarter. We expect cost of risk to moderate slightly in the second half of the year. The NPL ratio and NPL coverage trends are consistent with what we previously discussed: the portfolio mix shifts, particularly towards merchant financing and, to a lesser extent, car loans.
These are products with a higher probability of collection. Therefore, we keep those NPLs on the balance sheet for longer. A higher probability of collection means they require less coverage, so this remains a function of changing mix. If you look at real-time credit metrics, whether it be first- or second-payment default or delinquency rates, they remain low and stable. To wrap everything up for the second quarter: reported revenue up 15%, driven by e-commerce and fintech revenue growth. Adjusted EBITDA up 5%, impacted by higher rates and investments into Kaspi Pay and net income flat, again reflecting those pressures on EBITDA. Keep in mind that regulatory changes announced last year, particularly higher national bank reserve requirements, were introduced in two phases. The first phase was last year. The second phase was in the second quarter of this year, so you see that pressure in net income and it will be in the base going forward.
Another way of putting it: this clearly illustrates where the pressure on profitability is coming from — interest rates. We've always said that; it is cyclical. It now looks at the start of the cycle to be going from a headwind to a potential tailwind. When you think about the investments we're making into Kaspi — new tech and product spend, sales and marketing, not just in Kazakhstan but also in Türkiye — you can see that from an earnings perspective or from a dividend perspective, the cash we're able to return is relatively small compared to the overall scale, but it's resilient despite these factors and these investments. On guidance, guidance is reiterated: GMV up 17% as of the first half of the year and full-year guidance remains around 20%. We expect faster trends in the second half versus the second quarter driven by the timing of promotional events and other product initiatives. TPV growth looked like 13% versus guidance of around 15%.
If inflation moderates, that will be a downward pressure, although integration with Apple Pay should see us benefit from higher overseas volumes, particularly over the summer period. As we talked about, we've moved from TPV guidance to average net loan portfolio guidance: 20% in the first half of the year, guiding for 15% for the full year. EBITDA is trending up 7% year-to-date versus guidance of around 15%. Overall, we're comfortably on track for where we expected to be at this point in the year. On that note, Sam, let's open the call to Q&A.
分析師問答
Our first question comes from Gabor Kemeny.
Can I first ask about the fintech business, please? Indeed a decline in your deposit pricing for the first time. I think you cut your deposit rate around two months after the Central Bank policy rate cut. Is this dynamic reflective of how you expect your pricing to evolve in light of the Central Bank policy rates? And can you share your thoughts on how deposit pricing may evolve in the next few quarters? Yes, my other question would be just on the combination of this very quick deposit growth in the quarter, coupled with a drop in your deposit pricing. If you can elaborate a bit further on these trends, please, which clearly left you in a better funding position than you have been for some time. And my final question would be on the marketplace dynamics. It looks like Kazakhstan was growing more quickly this time than Türkiye. Can you shed some light on how these respective markets are evolving?
Mikheil, do you want to take actually all of those questions?
Yes, sure. So thank you for your questions. In terms of the deposit base, we are always focused on acquiring customers and delivering them the best product and experience. Previous actions resulted in very strong customer and deposit inflow. In terms of our strategy for pricing going forward, our general strategy remains the same: we look at market dynamics and the relationship between how we price our products and how we acquire customers. If we believe there is room to reduce the interest rate because market dynamics allow us to do so, then we will do it. There is no magical formula behind it; the decision to reduce the rate was driven by those dynamics. What you can expect is that this specific product, the three-month savings account, which is around 30% of our deposit base, will have its impact reflected as deposits turn because the duration is three months. So you should expect the financial impact later in the year as those deposits reprice.
In terms of marketplace dynamics, we have a different strategy in each market. E-commerce is a priority in both markets. In Kazakhstan we are developing the consumer experience by focusing on specific verticals, which is driving strong e-commerce growth. We are also growing extremely fast in e-grocery, which helps with consumer engagement and marketplace profitability and growth. In Kazakhstan we are working vertical by vertical. Electronics has not fully recovered because of price changes and supply chain challenges — for example, GPU and chip price movements driven by global events — but other verticals like clothing, car spare parts, home items are growing nicely. We are happy with our progress in Kazakhstan. In Türkiye our strategy, considering we are building out consumer products and fintech, has been to focus on foundational items: consumer experience, Net Promoter Score, merchant experience, and delivery speed, which we have improved dramatically year-over-year.
Consumer frequency of transactions increased 15% in Türkiye. Our priority has not been maximizing growth at all costs. It is better to have a smaller number of customers who love us than a larger number who transact only occasionally. We want customers who love our experience because they will adopt new products we plan to launch next year, especially fintech products.
I'll just add on Türkiye. When you look at its performance in the second quarter, you should look at order growth over the first half because there was a lot of retail disruption in Türkiye in March and April of last year. That distorted the comparable quarter-on-quarter both in Q1 and Q2. If you look over a longer period, H1 orders increased just under 18%, which gives a better indication of the business's performance this year.
Our next question comes from Maksim Nekrasov.
I have a couple of questions. The first one is very simple. Basically, your first half EBITDA growth was around 7% and was already trending above the full-year guidance, while you mentioned the reduction in the deposit rates that should benefit you in the second half of the year. So simply, why was the EBITDA guidance unchanged? And how should we think about second-half growth and profitability? The second topic I wanted to ask about, maybe not surprisingly, is AI and Kasper. I know it's quite early, but can you tell us about the early benefits you've been seeing so far or any measurable impact? And also in terms of the costs and level of investment — should we expect any significant costs related to that project?
Thanks, Maksim. Maybe I'll take the first question on the guidance, and then Mikheil can take the AI-related question. You are right: we lowered the rate on the three-month deposit, which is around 30% of the deposit base, last Wednesday. It will take three months to fully reprice that. So you're looking at the benefit really starting to come through from the second part of November — effectively only one full month this year. So yes, there is some benefit this year, but it's for a relatively short period. The full benefit of that and any other rate cuts will be felt from the beginning of next year.
On Kasper, we want to pull out the slide.
Great. So in terms of Kasper, we are just one month into it. Obviously, we have been working with Kasper ourselves for much longer, but the consumer rollout across the country on our e-commerce platform has been completed. As I mentioned earlier, the initial results are encouraging: one out of five customers use Kasper when it is available, and importantly Kasper completes tasks much faster — two times faster for consumers to find a product and 30% faster to basket. Those are important metrics. The metrics we're focused on now are all about trust. Kasper needs to perform the tasks he is given because trust is the most important early-phase metric for this service. The service needs to give recommendations, guide users, and be highly reliable and relevant. In terms of investments, we made significant investments in compute capacity, including a modern data center last year that provides sufficient compute.
That is why the response time is about three seconds, which is remarkable given the breadth of our catalog. Kasper can deliver relevant recommendations across millions of SKUs quickly and provide value such as reviews and delivery times. Going forward, our comparative advantage is that we operate in a 20 million population market. Building and scaling this service in a market of that size is far more cost-efficient than doing so where you need to support a 100 million population. That allows us to develop the product and tune the models at reasonable cost. We're not talking about price per token; we are a transactional business. We will measure this functionality based on the cost to complete the task — the task being enabling a purchase. We are a transactional business: we enable consumers to buy, pay, and shop. Everything we do ultimately results in a transaction, which is a powerful model. Kasper operates on highly accurate data, which is the foundation of our competitive advantage and allows us to deploy the service at reasonable cost and scale it to other markets in the future.
Got it. If I may add another question on payments: there have been some news about the national QR system. I wonder if you can comment if you saw any changes or any impact? And how should investors think about the long-term impact on the payments business and possibly the take rate in the future?
Our take rate is trending towards the majority of payment transactions, and that is transactions through our payment system and QR, which is priced around 0.95%. That's the trend we observe. In the second quarter we introduced Apple Pay and Google Pay. Previously we felt we could build a better local experience, but we introduced Apple Pay and Google Pay because consumers frequently ask for the ability to use them when traveling. That drove additional payment volumes when our consumers travel abroad. In Kazakhstan, the impact is mainly on international volumes because domestically our mobile app experience is very strong. We have roughly 800,000 points — give or take — where you can pay with the Kaspi mobile application. Consumers in Kazakhstan are very satisfied and merchants transact seamlessly through our technical capabilities. Regarding the national QR system, the priority for the National Bank, for us, and for other market participants was to ensure the payment system is highly scalable and highly secure given current volumes.
We have worked closely with the National Bank over the last several months to help build secure payment functionality. That functionality is now available and consumers continue transacting with our merchants as they were used to. We are also getting additional volumes as other consumers transact into the payment network that is accessible for everyone. We're happy there is a wide variety of payment methods so consumers can choose the most convenient option. When traveling, they use Apple Pay; when in Kazakhstan, they typically use the mobile application.
Our next question comes from James Friedman.
In your prepared remarks, you alluded to some of the growth initiatives you're anticipating for Türkiye next year. Could you remind us what some of those growth plans are for 2027?
Thank you. Our approach is focused on foundational improvements for merchant and consumer experience. Things that will drive growth next year and that are coming through this year include increasing delivery speed. We have improved delivery speed substantially; higher delivery speed means better conversion rates because customers are happier. E-commerce growth will continue driven by consumer engagement, mobile app usage, and the ingredients that underpin them such as delivery and user experience. Fintech products are another key driver of long-term growth and we are very excited about them because of our experience and because consumers want to buy items, not loans. Being present at the point of purchase is the best place to introduce financing products. The shopping loan we introduced, now at around 0.4% of GMV, enables customers to finance purchases seamlessly. There is a range of merchant products, like merchant finance, which we have in our home market, and we will bring those to Türkiye.
We will also launch savings products because to fund growth you need deposits, and we have simple, transparent savings products that are very popular in our home market. These are the ideas that will fuel growth in 2027: consumer shopping finance, merchant finance, and savings products. The $300 million of capital invested gives us a short start because that funding supports scaling fintech products next year. This year we focused on operational integration after acquiring the bank — taking operational control and building banking systems and reporting. Risk management and mobile application improvements are already in place for the shopping mall rollout.
Great. And this is the first time that I, analyzing the company, have seen rates go in your favor. I'm just wondering how long does it take to get repriced in the market? What I mean is in terms of consumer behavior, what have you noticed historically in terms of rate changes going the other way? How durable do you think this cycle will be?
James, I look at inflation. If inflation continues to fall, national bank rates — which are high by historical standards in Kazakhstan — will continue to come down. If national bank rates come down, our deposit rates will follow over time. When rates rose, we were among the first players to increase rates. When rates fall, I would not expect us necessarily to be the first to reduce rates. The long-term dynamic will flow through. Any rate cuts this year should give increased confidence about earnings growth next year. That's the timeframe to think about. But it's not about a single cut; you want to see inflation falling and rate cuts being sustained over a meaningful period for the benefit to be durable. Historically, the rate environment has been a headwind for much of the period since early 2024, and if it reverses it will help us over multiple quarters.
We currently have no further questions. I'd like to hand back to David for some closing remarks.
All right. So Sami, thanks very much. Thank you, everyone, for your time today. Happy to follow up offline. We are in London and New York in early September post the holiday period, so happy to follow up in person. Thanks again for your time today. Keep in touch and have a good summer. Thanks, everyone. Bye-bye.
Thank you. Bye.
This concludes today's call. We thank everyone for joining. You may now disconnect your lines.