One97 Communications Limited (PAYTM.NS) Q2 FY2026 Earnings Call Transcript & Summary

Thank you for joining, all, and welcome to Paytm's earnings call to discuss our financial results for the quarter ending September 30, 2025. We will start our call with Q&A after the introduction to the management. From Paytm's management, we have with us Mr. Vijay Shekhar Sharma, Founder and CEO; Mr. Madhur Deora, President and Group CFO; and Mr. Anuj Mittal, SVP, Investor Relations. This time, we have made a small modification to our call format. As you can see, our senior management is live on video. We hope it makes the conversation a bit more engaging and interactive. A few standard announcements before we begin. The information to be presented and discussed here should not be recorded, reproduced or distributed in any manner. Some statements made today may be forward-looking in nature. Actual events may differ materially from those anticipated in such forward-looking statements. Finally, this earnings call is scheduled for 45 minutes. A replay of this earnings call and transcript will be made available on the company's website subsequently. We will start our Q&A now. [Operator Instructions]

Well, good morning, guys, and thank you so much for joining us and me and Madhur are here in Noida office. And I thought that we will -- we aren't able to see many of you in different meetings. So my idea was that how about we get to see each other and talk about it. So when we invite you to do the Q&A, you can switch on the video or you don't need to switch on the video. It's just a choice. But as many of you would like to do it continuously forward, we will continue on this practice. The intention was that we get to at least meet and talk about it. So as you would have seen our videos that played in the beginning of this call are all about AI and that's what it is. You've seen the numbers and we are very clearly focused on making sure that AI is part of the business model as much possible. Up till now, we've been putting it into the cost side, efficiency side, while there will be some optimization, but not material enough, what we are trying to do is that we are trying to build it on the product side and feature side. And then after that, there will be AI products, which are not possible beforehand. For example, like this playing on this address book that you spend analysis that you saw couldn't have been done earlier because you can't write a script of a song for every individual. That can only be done when the Gen AI model is there. The good thing I want to tell you is that the voice model is made by us. So when you heard this song, which you heard this lyrics, this is made by us. It is not -- it is a retrained model open source model, retrained and created in-house by our team. Now going forward, what I'm seeing is there's a tremendous amount of AI stack that we're going to see. So foundationally, inference and then use cases will show up. Obviously, bottom line of AI impact that we are able to save the cost is what we're seeing, continuously we'll see. But I'm super excited that what in front of us we have in the form of AI, where we will be able to expand in the infrastructure and the use cases in a very dramatic different way. And this is my way of intending to tell what Paytm's future is headed towards. This will be the future that we will expand on. What we have built in financial services is showing that we can do it replicable globally. So product and technology that we have built in India, they can easily be replicated globally. We've written a note around it. I'm sure some of you have a question, and Madhur and me are here to answer those questions. But at the same point in time, my intention is to tell that our financial services business stack is getting clearer and clearer what we do. Merchant payment, credit is a very strong stack. We are adding the -- as you can see, money, stock brokerage and various other elements around it. That will be our focus. And then in due course, hopefully, we are able to make insurance also as a part of this stack. So this stack now in next 3 years will start to show -- we'll try finding which countries can go in a different working model. So that will be the future investment of growth. So the point is the future growth of Paytm in revenue and bottom line is going to come from India's expansion of financial services, replication of this product technology elsewhere and AI stack, which is completely from infrastructure to the use case. So it's phenomenal. I'm super excited what is lying in front of us and what we have come back from. It is truly acknowledgment of a great team here and I truly can't say thanks to every one of us internally in the team and every one of you who stayed and helped us support life. So here it is, question answers, we can start and let's go with it.

[Operator Instructions] The first question is from Mr. Pranav Kshatriya from Emkay Global.

Okay. So people -- bring 2 people together and then we want to talk about who can talk about it. So somebody can -- I mean, Pranav is here. So I'll just head to the spot line first of all. Yes, Pranav, you can -- and this is [indiscernible] that we will get the next person in the Q&A. [Indiscernible].

First question is on postpaid. That has come back. How do you think it is scaling up? Your release mentioned that you started with a select set of customers. I would like to know if you are looking to add more bank partners and how is the initial experience with it? And what is the road map for it? Because the more I read about roughly 5 crore unique credit card users in India, in a longer term, if you have to take that number to 15 crores, we will possibly need something like a postpaid or a RuPay credit card. So what exactly are the thoughts on that? How should we see this going forward? That's my first question. So I'll take one by one.

All right. And obviously, as you understand, our intention has been always democratizing financial services, democratizing credit card equal to postpaid because of the mobile UPI expansion that we see in the country. As you can see, we were able to partner with the bank. It is rather limited on various banks, technology and other business model limitations, not on our limitations. We are open and continuing to work with as many possible banks. That said, the current bank has a lot of runway of potential disbursement. So we will continue to -- while we continue to find out and add more banks, we already have enough number of customers scope and possibility of the current partner itself. So the numbers have been very encouraging. I just want to tell you that quickly, we have reached, I would say, the average spend, which probably took 1.5 years. So it's very, very encouraging.

Okay. Interesting. Second question, there has been a good growth on net payment margin on a sequential basis. Can you give some color on what portion of this growth is coming from the subscription side versus the payment processing margin side? Because if I just assume the normal ARPU, then possibly your payment processing margins have jumped significantly on this quarter. And what exactly is driving that?

Basically, we started to focus on credit instruments for the merchants and EMI, as you can guess, because the festive season was phenomenal. I mean, we're killing an EMI disbursement on the counter. And we obviously have partnered with as many EMI issuers. So that is why that became a extra chunk, which may or may not come next quarter. But at the same point of time, I can tell you that more EMI product that we do is making us more net payment margin. Obviously, as you are aware, our Soundbox and devices continue to grow. And I would rather say linearly or a little more ramp up than before, but I would point out to the EMI capability of our team that was able to do -- I mean, we are nearly half of the big guy in the town and literally imagine, I mean, without even much noise or conversation.

I think just to address the other parts of your question, Pranav, I think your analysis is right that the slightly higher jump in terms of percentages come from improvement in payment process margin. And as Vijay said, overall, the mix sort of improving from a margin standpoint, if you will, of credit instruments and so on helps and also just general price discipline both in the industry and also as we look at how we want to do business. So that discipline has also helped.

So this discipline you're talking about is on the subscription side, right?

No, I meant on payment processing margin pricing to merchants.

Okay. Interesting. My last question would be on the indirect cost. Good to see further cut in the indirect cost. How long do you think there is runway or we should expect some point soon this cost should trend up?

I think we have mentioned in the release that for the rest of the year, we expect this to be written out. And maybe in the next quarter or the quarter after that, we'll give a sense of what we expect for the following year, but we have seen a very significant improvement in non-salespeople, in marketing expenses, even though we are investing in consumer growth as well as in software expenses line. We do think that in our other indirect expenses line, there is some opportunity to reduce costs. So we may invest more on salespeople while we save money on other indirect expenses. So that might be the trend over the next couple of quarters and then we'll have a better sense of next year as we finish our planning for the following year.

Firstly congrats on good set of numbers. Two questions from me. First is on the marketing revenues. Can you share what's the breakup looking like today? And if there's any particular line within that that's driving the slowdown? And where do you think there's a floor for this line? That's one. The second is, again, on the BNPL, coming back to the BNPL question. One, what's the unit economics looking like versus the fantastic product you had earlier? And also, do you have the same level of flexibility on pricing as what you had with the original product with this new avatar of BNPL?

Pranav, our flexibility or our role remained various service charge and fees, which were there. And rest everything else belongs to the bank or the book owner any which way. And that flexibility stays with us. So there is no one less flexibility over the other. It rather because of it is running on UPI over credit card, it starts to get acquired on more places, which gives us more revenue margin beyond just acquiring. So it is net positive because the fees and charges are same or more possible variables. At the same point of time, we have more acquiring beyond just this.

And just to add to that, on overall unit economics, we expect this to be within 20 bps, give or take, of what we had earlier. Obviously, part of the input there is credit quality and that has been very, very, very early signs, but that is looking very good. So that's on BNPL or we call it Paytm Postpaid, for Pay Next Month. So we don't say later, we say Pay Next Month. Marketing services revenue, the largest components are advertising and travel. And to your question of -- on travel, there's been some industry headwind on at least on a Q-on-Q basis. We are seeing market share growth there. Overall, we have put in writing that we think that we are at bottom or near bottom and we should expect this to grow. We have also sort of given the context of that we are putting more -- less upsell properties on Paytm app. And if you have seen, I should call out that the app is getting great reviews. It is simplified to basically two folds and it's getting great reviews for its simplicity, its focus on payments and also being super thoughtful about where the upsell properties are. But I believe that simplification journey is more or less done. So now we shouldn't have sort of less upsell properties going forward compared to where we are now.

Understood. So basically, if you get the NTUs up, the monetization is kind of rework of that is done. So it should translate into a direct revenue.

MTUs up and targeting better. I don't want to show everything to everyone clearly.

Understood. Just one follow-up on the Pay Next Month. Now the flexibility I was talking about was in the postpaid product, you had flexibility to charge 0%, 1%, 2% or 3% convenience fee to consumers depending on your risk assessment. Does that still remain in the new product?

We basically do not make any option for paying later. It is a charged card equivalent. It is not that you have to extend it. So incidentally, Pranav, there are 2 kind of Buy-Now-Pay-Later. One is where you have an obligation to pay at the month end. Another is where you have this classic lambda model of 3, 6, 12 months or any other number that you want to. We do not rotate. Why we do not rotate is because we want to have it as a payment product, not as a credit product. The idea is that it leverages our payment capabilities and expands on that instead of trying to leverage it on a credit book ownership basis. It does have a credit charge. It does have a credit cost involved into it, but it is not a EMI product. So it does not compete with equivalent of EMI for iPhone. It rather is a consumption credit. So what we target are people who need up to INR 50,000, INR 60,000 in a month to match the expenditure and you are expected to pay month end. So this rotates a very small fee, a few bps on a 12x basis, making it a return instead of trying to have a larger interest cost. So it is a fee-based product, not an interest-based product. So there is a element of NIM into it, but it is not loan. If somebody wants to take a loan, that's a separate product altogether. And EMI is a separate product. Subsequent to this, if we want to build, we can build. This creates a product differentiation in the market, creates differentiation, as you know, on counter, there are too many people who are offering EMIs. It does not clutter and cloud there. It rather is every day where you cannot spend otherwise. So credit card ended up becoming an everyday spend plus EMI product. Can this become plus EMI product? Voila, you know it.

Understood. My question was more just on -- I understand there's no interest here. There's no revolvers. But I believe there is a fee that's charged -- a convenient fee that's charged. Just wondering if you have -- you can say, charge 0% for someone, 1% or 2% for someone?

It is like that. So it is a convenience fee from the spend side and people like you and me who have a high credit score, et cetera, may get 0%. So people like us who do not have a credit obligation to take, but see it as a convenience and do not have normal number of transactions that go to the bank account, availability of bank account, et cetera, et cetera systems, they can simply just use it and we earn merchant side revenue and that's it. So it does have component of 0 fee to the consumers and there is no subscription in any which ways.

That was Mr. Pranav Gundlapalle from Bernstein. The next question is from Sachin Dixit from JM Financial.

My first question was on the working capital side, right? So generally, when we see in, let's say, fiscal year '25 or in the previous half, there was roughly a INR 120 crore dip from operating cash flow before working capital to after working capital. This quarter, that number has been almost INR 550 crores. I understand there might be that INR 190 crores of write-off also parked there, but still it has almost tripled. Can you explain what is driving the sharp dip in working capital conversion -- sorry, operating cash flow conversion?

On the specific number, I'll come back to you, but I think we should be a little bit careful in looking at our working capital on this basis because our working capital closure for the quarter actually depends largely on whether it's a weekend or a weekday or any other kind of holiday because we settle to our merchants every day. And in certain instruments, which are not vast majority of our business, but in certain instruments, for example, credit cards, we get money on T+1 working day basis. So we do have fluctuations in working capital on a week-by-week basis rather than quarter-on-quarter basis. The specific bridge that you mentioned, we can explain that to you offline, if you don't mind, Sachin.

Sure. Sure. Secondly, on the payment processing margin improvement, right, Pranav obviously asked a question. How sustainable do you believe that improvement is? Because I understand there might be some EMI component in there which might have been driven by the festive sales, which were preponed this quarter. So is it sustainable at these levels? Or we should see it fluctuate and remain in that 3 to 4 bps range [indiscernible]?

As an overall trend, we think this is sustainable. The credit card mix and EMI mix on a quarter-on-quarter basis could vary. But as an overall trend that we are seeing adjusted for seasonality and those sorts of things, we are seeing improvement relatively consistently over the last few quarters. So obviously, what we look at is instrument-wise and merchant type-wise, are we seeing improvements and that is the case.

Sachin, thing that we now have is a privilege of onboarding online merchants. Online merchants eventually are high MDR plus high sales margin, net margin. So that will also start to add to the bottom line. There is a little bit of number of customers whom we have started onboarding since last quarter. And I think this is not just sustainable. Internally, just like you saw on the cost side, we will see how we can continue better.

Understood. So that's quite structural in nature in that case. Just one final question, Vijay, on the AI piece that you mentioned, clearly shows that you are quite focused as a company on putting in AI into multiple things, all your sort of business segment notes also mentioned AI in some form or the other. So I just wanted to understand, is AI going to be a cost driver for you, like you will save on cost? Or you also see a separate monetization opportunity? For example, this AI Soundbox, right, are you charging more for it? So in that sense, whether it's just cost or also a big revenue driver for you going ahead?

Big revenue driver. I think cost cut is anyways, which we were doing and AI just added acceleration or a deeper opportunity to optimize. AI is a revenue line item. AI brings newer service, newer business, phenomenally more number of things that we can do. And the good thing I want to share here it is that you will be able to see that we have large number of merchants whether you look at small merchant, which is, let's say, a small shop where we are putting a Soundbox or a large merchant, let's say, you can talk about large online e-commerce companies. Something or other, we will have for each of them to acquire. So our merchant base will get re-cross-sell able for AI-led infrastructure, product agents, et cetera, et cetera. And right now, we are running some pilots, et cetera, for us as internal customers. And I'm very excited that we do believe that 1 year forward, there should be in a commerce cloud line item, the cloud element could be the AI element that will start to go back once again. So it is a revenue line item. I'm personally completely about revenue line item and creating product and services, which otherwise couldn't have got created. So whether it is a small shop, it is always challenging to do it for a small guy, as you know, whether they'll pay or not and it is easier to do it for a larger guy because they have a understanding of the system. And I think we have a product mix that's across small merchants, medium-sized merchant, large segment.

So and I should just maybe add, Sachin, that from a financial planning standpoint, translating what Vijay just mentioned, what -- the starting point is that AI gives us better efficiencies and better insights in addition to new revenue-generating products, of course, that Vijay talked about. So now how does those additional high-quality insights convert into better cost or better revenue? That's the question. So for example, if we are able to use better insights to reduce the credit costs for our partners, then that translates into higher collection revenue. And if you are able to get much more efficient enough to do a lot more collection effort because like the price of doing a collection effort has gone down. So it does translate -- so the starting point is actually not whether it's going to give me more revenue or give me more cost, but how do you use those insights and efficiencies to drive your businesses better and then depending on the specific case, it will give you more revenue or less cost even in existing products.

The next question is from Mr. Rahul Jain from Dolat Capital.

Firstly, Vijay, congrats on achieving peak revenue per MTU on a TTM basis that you used to do in CY '23. And this has happened despite the consumer side of it not playing out. So basically, the merchant side of LTV has improved meaningfully. The MTU count is down and that's why the revenue is not at its peak. So how you play on the consumer side of it now because merchant side of it is kind of proven we have an advantage and we've been consistently innovating on that side. Consumer side, what I see as a new thing is probably the gold where I have seen a significant number of unique user. If I have to see penetration, it looks a pretty high number to me, even if I have to assume lifetime customer that we might have. The number what I saw was upwards of 50 million on the app, I think. And then there is a stock MTF, there is a Postpaid. So anything that would create a thought process on the consumer MTU increase and also consumer ARPU increase.

I think it is clear that when we got an opportunity to work only on merchant, we just did it and you were seeing the post facto effect of it. And I still want to add that there is a huge amount of bottom line or revenue generation possibility with the merchant, like I quoted AI as one of the line item there. When it comes to consumer, I think consumer AI is only the features, which can be commercially possible as a independent product versus adding the product in wealth or, let's say, credit disbursements, et cetera, those kind of elements will show up. We clear about it that we don't have large MTU, very large MTU. We're lucky to have a large MTU. Over the period, the customers who stayed with us, I would say they are higher quality customers. They were able to see the resilience and appreciate and they have stayed on the platform longer term. So I would say monetizing those customers is a primary attention. And the best love that we could give to them was no annoyance of advertising, et cetera. So cleaning up the ad so that they feel much more comfortable and cared. We launched loyalty program in the gold coins -- gold points rather which can convert into gold which can become gold coin or whatsoever. So the intention is to make it products so simple and so lovingly acknowledging the customers who've stayed with us that we want them to become the brand ambassador to attract more number of customers. You will see UPI market share when NPCI announces this month. We do believe that there should be an impact on it. Surprisingly, our marketing spends have not grown, but our UPI market share should have grown. Let's see what results come out. Now what the intention I'm trying to say is that we rather want to make most of the customer that we have and that is also the monetizable customers. I mean, I was talking to Deepinder also the same thing and he talked about the same thing. The number of customers that India will have are not so many which will be monetizeable, but the customers that we all are lucky to have because we started tens of years back are the customers who will drive the maximum monetization. So we have started to look at maximum monetization by removing the advertising, et cetera. And it is so much of personalization that we are showing up there. Gold came out as the solution for people who are not maturing towards mutual fund or trading. So as you can very well understand the tonality we've been talking since last couple of quarters that credit disbursement is one line item. Wealth, wealth, wealth, wealth is nothing but a gold, mutual fund, stock trading, F&O, all these buckets. Now these buckets are next line item for us. So it works out really good for us. I do believe -- I mean, I can't probably give the number of what we did on Diwali month because we've not quoted it. But at the same point of time, I can say the number is so phenomenal that we were like, oh wow, we have a very big customer base who loves this property and product and we are taking the feedback and continuously expanding on that. Let me give a hint of it. If some number we were talking about being target of the March month, we not only crossed that, but many times over in the Diwali month. So that kind of good performance in the gold. The intention is that credit and wealth is a bucket where the gold mutual fund distribution or stock trading and F&O, MTF, all this bucket comes. This is exactly monetization plan that we will do, my friend.

Can I just add one thing, another framework to maybe suggest, so I look at it on CAC to LTV basis always. So the trends are that our CAC is down. You can see that in marketing costs, you can see that in cash back. Retention is up, which is probably not a surprise given the product improvements that the team has made. And we're also getting some benefit of these high repeat use cases, right? Like savings is a good high repeat use case, especially if people are doing monthly -- sorry, daily, weekly SIPs. So we -- our retention is up overall and especially for the users who adopt these use cases. So we are seeing the CAC to LTV improvement and I'm quite optimistic going forward because we are talking today when we still feel like we're at the bottom of the credit cycle, which has certainly impacted our personal loan and credit card distribution businesses. So those things we have reasons to think that, hey, those things should be better in a more normalized situation. And we are seeing good growth in areas, for example, wealth and we do acknowledge that in certain areas like advertising, travel, we should be doing more. So the CAC to LTV journey is very clear. We have seen improvements in that metric and we'll continue to see improvements in that.

Yes. Just if I can ask one more. First of all, I should say that gold coin was a very good hook and very timely given the traction point. Now coming to my question, Madhur, it's commendable the way the margin is improving. One of the aspects that you highlight was that there was a improvement in the payment processing. So is it more like we are able to charge more because we could be doing a VAS kind of a thing like EMI and all for the merchant, which is driving better take rate or lower processing cost, whatever? And since Vijay also said that we are seeing a good traction in Diwali month, so is there also a near-term margin aspiration that we are aiming at because we have a medium-term outlook, but something which we could look from 12-month perspective -- 12-month forward perspective?

So we have significant improvements in how we think about pricing and merchant-level profitability. So that helps a bunch to make sure that we are tracking this very closely and that framework keeps improving. I think generally speaking, we are flexible in terms of how do we get the right margin from a merchant, right? So an enterprise merchant may want to pay us a little bit higher margin on credit card or may want to give us more EMI volume or may want to pay -- may be okay to pay us a technology fee or something like that. So we're generally quite flexible and that's really for the team and that's an advantage that we are flexible like that. Our teams can be very responsive to where the merchant -- where the merchant preference is. End of the day, the teams and the camps and the sales folks figure out a way to make sure the merchant is profitable and that's really the very healthy way of running the business. On payment processing margins, like I mentioned earlier, we saw sort of a bottoming out from a mix standpoint. If you recall 2, 3 years ago, we used to talk about, hey, UPI is growing faster than non-UPI, et cetera, et cetera. I think a lot of that is behind us. And now what we're seeing is slight improvements in mix and slight improvements in margin by instrument. And overall, that just keeps giving us some goodness on a quarter-on-quarter basis.

I also want to tell that thanks to the online merchant onboarding, there is a lot of chatter about offline, on counter EMI. But I'm sure you understand that smartphone and many other gadgets are now sold majority or significant large number through online and that is percentage that is continuously growing. So we see that there will be even higher opportunity for us now that we have online merchant onboarding and additional product that we are launching in online. So you can call it VAS, you can call it more margin products that we are able to launch for online and offline both. And now formally, we can do omnichannel. I mean, you can imagine that a single company will be able to see their single dashboard, et cetera, et cetera. So we are very happy about it. And regulator has also made both licenses and one license together. So it's phenomenally good for payments. I mean, there is no other company that plays omnichannel and more features than us in the market. There are many individual people. Somebody does QR, somebody does copy the Soundbox and somebody does EDC and somebody does online.

The next question is from Mr. Piran Engineer from CLSA.

Congrats on the quarter. Sorry, I can't be on the video. So I just had a couple of clarifications first on previous question. Now in this Postpaid, you said you make 20 bps margin.

No, sorry, I did not say that. I said it will be within 20 bps, give or take, which it could be lower or higher compared to what we made earlier. That is our early estimate. The question earlier was, will it be similar to what you used to make earlier? I said I expect that to be within 20 bps of what we used to make earlier.

Okay. Okay. Understood. So earlier you used to make 70, 80 bps, right, if I remember correctly?

Broadly correct.

Piran, net answer, we will have more values of adding revenue line items now because it is a bank and it is on UPI. And we will have more opportunity of onboarding new merchants. So it's a rather bigger revenue and profit than before. That is what we [ estimate ].

Okay. Okay. And in this, how much MDR does the merchant pay?

It's a UPI MDR. UPI has a concept of interchange, which means that the issuer side will get something like 1.1%, 1.2%, depending on category of merchants. Everything around it comes out about merchant paying 1%. I mean, there is 1.3%, 1.4% category also and 0.9% also because some categories have less or more and this is decided by NPCI. At the same point of time, now acquiring side companies can add any margin on top of it and sell it, so we can sell to the merchant ahead of this margin. So it is -- for an acquiring business, it is COGS, C-O-G-S, and then you add a margin. So can we sell it for 1.5%, 1.9%? Depending on who and what commercial is a choice of acquiring business, which is also accrued into our acquiring side of business. In exchange, 1% that goes back is what the revenue of the issuing side is, along with the convenience fees, et cetera, that the issuer can charge.

Got it. So it's like a credit card where the issuer is also earning that convenience fee of 1% to 3%...

It's called [indiscernible].

Interchange plus various fees is issuer side.

Okay. Got it. Got it. Okay. That's pretty clear. Secondly, on this EMI spend, I know a lot of people have asked on the payment margin. So when you say EMI spends, you mean like the subvention model that Bajaj Finance does? Or do you just simply mean credit card swipes at your online merchants or POS machines that are deployed of?

So Piran, our goal is to aggregate financial institutions which could be credit card issuance or NBFCs or banks issuing EMI. So these are the financial institution. Then second part is brands. You can call the classic white good brands, LG, Samsung, Apple, et cetera, these people and then merchant side. So this is a triangulated business ecosystem where these 2 players combinedly want to reach out to the reach in the market and the reach in the market is unparalleled by Paytm. Our ability to integrate with brands is incredible and our aggregation has already been successful. So we are now like -- we are cornering actually, we are taking away the market share, chipping off the market share from the erstwhile existing players and more or less that in online it is more even harder work that -- easier work for us and we will be able to capture because this brand subventioning is the USP. Although, the brands -- so 3 kind of EMIs work in the market. Straightforward customer base; number two, brand subventions; number three, financial institution also subventions. And in one product like Apple, by the way, all 3 could be subventioning. So these 3, all products exist with us and that is why we are saying that this is a phenomenal good product. We did it last quarter. I mean, we've been doing it and we started to focus on festivities once our online onboarding started. So we've been able to do good. In fact, we will be able to do even more further product enhancements on this is what we are learning.

Okay. Okay. Understood. And just moving on to a couple of questions at my end. Firstly, on margin trade finance, this is done on your balance sheet or with a partner? And how big is that book?

The margin trade finance facility is being done by Paytm Money, which is a regulated equity broker, 100% subsidiary of ours and they do that on equity broker balance sheet which is to say effectively on a consolidated basis, you can think of it as our balance sheet, but it's Paytm Money stand-alone balance sheet.

Understood. And how big is that book today?

We haven't shared that number. We believe at this point, it's commercially sensitive, but it is growing very nicely.

Got it. And just lastly on adding headcount to your distribution, your sales employees, on the ground sales employees. Now typically we were adding maybe 1,000 or 2,000 a quarter and this quarter, it's been 5,000. What's the sort of thought process behind hiring so many?

We want to dominate in the merchant ecosystem. I would acquire a large number of enterprise onboarding now offline or online and we will do even more aggressive in offline small merchant. I mean 5,000, I wish we could have done like 2x of this. We're making money. We are investing that money back. I mean we want to dominate on merchant ecosystem like nobody's business.

But Vijay, we already have almost 5 crore merchants.

No, no, no online merchant enterprise, we have huge numbers, but there is a dramatic more because maybe you know it or not, since 2022, we were not allowed to onboard any merchant whatsoever in online. And then omnichannel, et cetera products didn't show up in the market. Right now our product market fit is phenomenal. I mean, I'm going to say that. And then cross-sell products that we have discovered is -- I mean, why do you want to even think anything else?

And finally, the merchant payment for the mid- and long-tail segment and merchant lending distribution opportunity, as you can tell, you have tracked us closely for several years, that is working really, really well. So you obviously want to continue to invest in making sure that you have -- continue with the largest merchant base, highly engaged, well serviced and then do lending on top of it.

[Operator Instructions] The next question is from Mr. Sachin Salgaonkar from BofA.

I have 2 questions. First question, Vijay, I would like to understand a little bit more on this entire AI opportunity. And I'm saying that because AI is more a generic word, which is used by multiple companies who would love to understand Paytm strategy in terms of how you guys are looking to optimize AI, both from a revenue perspective as well as room for EBITDA margin to improve. And the related question is what you mentioned in your opening remarks that if you're going into international countries in terms of the software, what you guys are offering from an AI perspective, should we consider this more on the lines of PayPay or PayPay was a unique example and this is something different? So would love to get a bit more clarity on that. And second question is more a generic question on the revenue opportunity. Clearly, you guys are now back into Postpaid. You guys are doubling down on merchants. And I did see comments on equity brokerage, what you have mentioned into your shareholder letter. So we add all of it, should we see the growth accelerating from the current 24%, 25% to around 30% and above in the medium term?

I like these. All 3 questions are very favorite for me. First of all, AI, what do you mean by AI here? Is it a buzzword or what is the revenue line item that we could do? So I'm going to tangibly take 2 product line items example. One is where the small shop takes an agent, which is served using our sound -- AI Soundbox or various other AI devices that we will launch and these agents are made for small merchants. Typically, large companies get third-party companies like Snowflake or let's say, Accenture, I'm sure they will be Infosys or Wipro also in that category that will implement AI for businesses where they will bring all enterprise data and make this interface available for a large business. Now we are going to do it for small businesses in the country and elsewhere over the period as I've said. Idea is that a small business can have a chief operating officer, chief finance officer, chief marketing officer practically there in the shop in a AI component that we're talking about. Obviously, I'm not going to judge their devices, smartphones, whether they can run models, et cetera, or not. So that is why we are building these devices, which are specifically made and manufactured by us, ourselves. Now these elements of AI effectively are practically for you, it is agents for small businesses. So Paytm is selling agents for small business and charging subscription and inference fees once the subscription usage grows ahead of it. For example, like, as you know, you're going to talk about it that how do I grow my revenue? Okay, so where is my revenue compared to somebody else nearby and where is my revenue showing up versus last month? And all these kind of questions, we believe that there is a fair amount of usage that will show up in a subscription. And if you want to reuse even for the next set of things, for example, like give me an ad, give me a marketing plan, give me -- run my ad on the social media, run my ad on this, everything can be done on this. And now those kind of elements will mean that you're using more inference. We will price inference. So right now we want to see what kind of different use cases people do it. One interesting thing, Sachin, I want to tell you, when we showed the demo and we learned was translation. So you have a shop where you speak, let's say Hindi [Foreign Language] and the buyer is in let's say English or let's say Spanish or Mexican, whatever language, I'm not saying that we will only get English-speaking tourists in the India. So you could say [Foreign Language]. So now the Soundbox was able to say, hello, this is INR 45,000. It's probably the best price I can do and this is absolutely the best. But this is our last price. Now you have an assistant who can speak the language of your customer. I mean, how phenomenal it is. It is -- I'm so elated that the use cases and features that we are seeing are way ahead of just speaking a Soundbox. I've been saying it, Soundbox will sound like a feature phone some days later. I mean, this will be like a smartphone apps as well. So there is -- this is an agent revenue and there's an upside of revenue. Now you go to the larger shops, like larger businesses, let's say mid-sized retailer or even let's say Zomato, Swiggy, or Flipkart or online merchants. There are certain products, I'm not talking them right now or offerings that are meant for them. So we are piloting internally for us as a large company. And I'm trying to figure out how to price them and structure them for those guys. So those other products will also show up. It will be an infrastructure layer. It could be very well thin layer. And that is what we will talk next about after some quarters. So AI is a revenue line item. I hope that clarifies. And secondly, you talked about international. I'm going to project something so that it is readable. This is in a earning release. As you can see this, it is an earning release document. We have 2 plans, Paytm has built proven payment acceptance and hardware, software and services stack and we will give it to a partner who will run in that country. This is like Paytm which is exactly what you were saying. I mean, although in PayPay case, we did not provide hardware yet, but we will in turn continue to look at more number of things, software, hardware, service. The full stack is available to a partner who locally in a revenue share model or we can get some strategic shareholdings, sweat equity, et cetera. And sometimes we can chip in some money also if needed, but not large capital is being allotted to this business. But rather it is simply about that we have a technology we want to monetize. This is partner of model. Some countries, some profit pool, especially where we believe there is a large proven profit pool. For example, like, some place the MDR interchange gap is so large or someplace where our product market fit seems so perfectly comfortable that we could take our own self-operated, that is Paytm-operated business. So mostly it will be partner-operated, but for the complete disclosure and understanding of the model, we will call it model #2, model #1. One model, partner-operated; second model, Paytm-operated. And we believe that this will further monetize 3 year forward. The bucket between 3 to 5 years will be starting to build from some of these. Our initial experiments, which will be very low cost, will show up. And I mean, I remember when we did a $1 million investment, it was like, what are you doing? And why? The intention was to learn can we do it in such a dramatically different geography? And our results have been phenomenal. No, you're not doubling down much -- so farther away. Our approach will be only 3, 4 -- 3 hour, 5 hour, 8 hours, that kind of flying distance between the places. Let's see how it will show up.

And this is mainly emerging markets, like, for example, [indiscernible]?

No, no. Well, emerging -- okay, let me say frontier, emerging, developed. That's a good way to say it. Yes. Maybe some of the markets are emerging because whole Southeast Asia is called emerging. So sorry, yes, that is true. I'm going to say that we will definitely go in developed market also. For example, like we went in Japan, we made money, we made product. And my personal ambition is that editorial choice award that you have a product that in a developed economy can work. I never want to be identified as cheap third-world country technology. We are first world technology world. And as Indian, I want to prove it in my lifetime that an Indian company can make this and it'll go places. And so that is why developed country. But between us, as shareholders and management, we will earn our creds. We will earn our stripes by attempting to showcase that we can do it and then we do it that far away. So short answer, it will be more or less India-like market, but it will definitely go further ahead later in the developed markets.

It's not a filter that we've put because we think our product is globally relevant. The filter that we put is what we have shared in the release which is where should we be partner-operated business and where should we be Paytm-operated business.

And third question was, Sachin...

Revenue growth basically, acceleration of that.

The revenue growth for festive season [Foreign Language] acceleration on line items which are bottom line-generating. I mean, up till now we are bottom line-generating focus. We have started to become revenue-generating focus, so you'll start seeing it.

Got it. One small follow-up, Vijay. Clearly, when we make investments into AI, there is an extra incremental cost associated with compute cost. Is that something should be materially high or we should not be worried in grand scheme of things on that?

Sachin, I can tell you very candidly. One thing I've learned and I think I should give prudence to many of listed founder peers and many other, their board member, shareholders who have told me, whenever we are going to do any capital investment, capital investment is an investment that goes beyond 20 million, 50 million onwards, we will be very much sizing it up. We will very much proactively announcing it in advance, et cetera. So right now, what you are hearing, there is no material investment plan here.

The next question is from Mr. Ankur Rudra from JPMorgan.

Good quarter. On the AI side, you gave us a lot of examples on services and agents that you'll be using through AI. But at this point, Vijay, do you have the confidence that this will be a unique monetization opportunity as opposed to a feature which perhaps becomes more of a moat for your existing products?

Yes, distribution is king. The product technology that we can make our distribution is only with us. It's a proprietary distribution.

So you will be sort of upselling more products as opposed to trying to differentiate?

Differentiation is because we have this offering and in this offering, we are upselling, which is revenue on it. So I did not get how these 2 conflicts that you will be upselling instead of differentiating. I did not get it.

What I meant is that -- what I meant is you'll change your price points for these products. It'll be priced slightly higher than what you currently have.

Yes, these are all option. I mean, it's -- you've seen it, you get a top-up option, top-up option, top-up option. So these are differentiated features and the price clemently different. So both things happen. We cross-sell, we earn extra. And the opportunity of cross-selling means that only few companies can do it actually, it requires compounding work incidentally.

Got it. Just maybe moving to ...

I can tell you that the -- I mean, there was a question asked that would you like to patent this or whatnot when I launch -- trying to do this? And my answer was what -- why? I mean, I want more people to walk into this so that the more customers market get created, people understand and so on. So we would, I mean, copy everything that we've done is anyways somebody else does after some time. So that's not a problem. It's a race towards continuously expanding. And lately, you've seen, we have started to worry about monetization. Earlier we were focused on innovation, which was good. Right now we are focused on monetization of innovation equally. So right now this is a proprietary distribution.

I'm guessing it's early days. We'll probably take a couple of years to see results of this come through.

Definitely.

Just -- go ahead.

No, I have a different question, but why don't you go on?

No, go ahead. Go ahead. I mean, a couple of years is a very large time line. Madhur probably would have said that. I mean, I'm not going to say a time line. So while answering that question, I will load it question back to the -- so does that mean -- so that's why I didn't answer. I said whatever.

Yes. Sorry, Ankur, why don't you go ahead?

My question was in financial services, we did see a nice acceleration this time. Could you maybe elaborate? I know you've stopped sharing the loan disbursal amounts explicitly, but from a color -- incremental qualitative color perspective, could you highlight how much of that came from maybe the launch of -- relaunch of Postpaid versus perhaps acceleration of ML or from the consumer side, you saw any kind of acceleration?

Yes, Postpaid was launched in September. So the numbers -- financial numbers currently are insignificant. That was not a driver. Merchant loan continues with nice trajectory of quarter-on-quarter growth on volume terms. In revenue terms, we are seeing better collection performance, which is increasing take rate. Personal loan, we have given commentary that due to headwinds, we're not seeing growth there, but we are seeing slightly better take rates and economics, but again, not a major driver. And we are seeing as financial services, we are seeing growth in other areas like Paytm Money, wealth and overall wealth, gold as Vijay mentioned. So a number of positive signs and other non-lending financial services also. Just a quick housekeeping thing. We're sort of almost out of time, but we do want to make sure we take all the questions we can. So if we could have Jayant, then Vijit, then Prateek, and to the extent possible, if you could limit yourself to 1 or 2 questions, please. We just want to get through as many people as we can.

Congratulations, Vijay and Madhur. Great set of results again. So I'll have only 1 question. Clearly, the GMV mix seems to be working in our favor and the margins are showing. This is more from GMV growth perspective. We do see UPI P2M growth is now down to the 20s. Cards, in general, I think system-level growth is in that 15 handle plus minus. So from here while we are adding new merchants I see on the Soundbox and overall, the UPI -- these would be largely UPI, right, right now. So if your UPI system itself is growing at 20, do we see a scenario where we now start thinking about our GMV growth coming to the early 20s handle over the next -- forget a quarter or 2, but maybe let's -- next 2 years, 3 years? Of course, this means better margins. But from a growth perspective, is that a fair assumption?

Our UPI growth will come from -- our payment growth will come from which element? The 20% component which is non-UPI component. Is that what you said?

UPI P2M, yes. So that is growing at now 20s, right? That is sharply -- I mean, I think it has grown so strong in the last few years. So if you take the total merchant payments digital, that is itself now growing in 20s what used to be, let's say 30%, 40% led by UPI P2M. We are gaining market share. So we are growing at 25%, 26%, but -- and our growth I'm presuming is coming from adding new merchants, which means we are at the...

We have sale merchants also.

So Jayant, the -- I think maybe just to simplify the variables, what we are seeing is we are seeing market share growth with improving economics both on merchant side and consumer side. And that's what we want to try. I think it is a little bit hard to sort of predict what UPI GMV growth will be. Of course, we are a driver of that, especially on the merchant side with how much investment we make and are making in expanding the merchant acceptance network, so we are a driver of that. But in a sense, I don't know if we're best placed to agree and guess what exactly UPI growth numbers would be. I think our focus is gain market share sensibly and keep improving unit economics. And of course, in the merchant side in particular, we have done a fantastic job of upsell of distribution of financial services to that very low penetration rate. So continue to drive that. So that's simply the business model. We do think that if UPI is growing, let's say, 20%, our GMV growth should be higher than that and that's what we'll continue to drive.

Yes. If I could squeeze one more. Credit card on UPI is clearly helping on the margins as well, right? So is there any metric that you're tracking on our market share in that particular product and how are we faring compared to our overall market share and is there any investments you can do to improve that?

I like it, I like, Jayant. We probably have more than some X than UPI market share. So our credit card, UPI credit card market share is better for us than our generic UPI market share. Generic UPI market share is actually skewed towards P2P transactions because they are always larger volume, per ticket is very large. So the UPI market share is a game of P2P and it does have some money because banks pay each other and then [indiscernible] get some money out of it. So there is no secret about it that is a critical line item. But we've been able to capture more than our market share of UPI GMV market share and order of magnitude more than that.

And on the merchant acceptance side of that, are we...

Anyways, I mean, internally, the good thing I can tell you is that we do foresee and that's the statement I'm making, we do foresee that quarter or month is not very farther when our subscription revenue get toppled by our MDR revenues.

Next question is from Vijit Jain from Citi.

My first question, with this, now that you've combined the offline and the online payment aggregation business, I think you alluded to this, Vijay, when you talked about now you can go omnichannel, which you couldn't do for the last 3 years. Is that what this merger sort of allows you to do, or is there anything else that you could do post that? And relatedly, I think I'm just wondering, is EMI the other opportunity you're referring to here? Is EMI equally an opportunity on online payments as well? Yes, that's my first question.

Okay. So first of all, you're right. Omni is a basic understanding of the companies that have online, offline. So Flipkart also has an offline business because customer goes, receives goods in-person and that point of time, it is actually an omni company. Flipkart is a more omni company than people may think about it because of cash on delivery being large numbers and that means the payment is a critical player. So all these businesses we couldn't have onboarded and just because this offline entity was different company, legal entity could have not onboarded. You know what I'm trying to say. So we were lucky that we kept working on online business for so long and finally I mean, we were able to start this. Yes, not just this element, but there are other VAS elements like spend analytics, like settlement, like payout. All those things start to show up as a new value plan items. So per payment, we make more bps or more money than we will make standalone in offline or standalone in online. Actually standalone online offline business will disappear. It's like saying laptop and device will disappear. People will start to move towards assuming that everything is same. It's a cloud kind of concept. So yes, that is there and the EMI is just an intention that I'm trying to say that the success that we want to show is that we literally worked upon with this kind of combination in a category called EMI and then voila, we just captured large market. Now online EMI is by the way bigger market in categories like phone, as you know. More phones sell online versus offline. So there are even more kind of product and solutions that we are able to build. So answer, yes, not just one thing called EMI, but many other things will show up and we've been able to do a good job of adding product or service features for payment and our merchants love this and some things are unique to us and probably the market will come up to them in a couple of quarters or years.

Good thing, Vijay. My second question and this is my last question, on the devices front right now, I mean, solid growth again Q-o-Q and I see your CapEx and your D&A expenses related have been trending down as you've guided earlier as well. So I think the question I'm trying to answer here is the D&A schedule aside, are these devices remaining active for longer and much longer and how -- what is the real time line on which you take these devices out of circulation actually because they are either obsolete or they just die?

Yes. So good question. And as you would have remembered that we continued telling last quarters that our new gross adds are bigger than the net adds because if we pick up the device, the merchant is no more active if that is the reason and so on. So now I'm happy to say that has completed -- nearly completion, that job is nearly completion. Next couple of months, we would be sorted out on that, which effectively means the device in the market is active and being used. And those devices active in the market being used mean more revenue and more NPM. So what happens to the device? Well, there is always some component. I mean, we try to recycle as much more. So that's an advantage. The only cost we do not recycle, if at all, will be the cost of logistics is more than that. So we created hubs in different, different cities now. Earlier, it was based in our headquarters and near that, Noida, the factory, et cetera, but we have now created hubs to refurbish in multiple other cities.

And just to clarify, vast majority of devices that we pick up are refurbishable. So...

Yes, we will not pick up a device that will not be refurbishable. That's also...

So it is not a device that we are picking up obsolete devices. We are picking up devices from merchants who went inactive and reactivation efforts kind of journey was done. And then at the end of that, there are different sort of frameworks that we use, at the end of that, we've decided that it's better to pick up the device rather than to continue to try to reactivate that merchant. And so generally speaking, you are getting a device which is in reasonably good order. It will require some work, maybe a battery replacement or something slightly more than that and then it's as good as new.

Got it. So Madhur, then because you've had -- you've been in the devices business for, I think, more than 4 years now. So I think what I was also trying to understand is, are those the devices that are getting refurbished for battery issues? Meaning I'm just trying to wonder if these devices just with refurbishment, their life is much longer, is that how I should think about it?

Refurbishment definitely increases the life of the device. It saves us -- I mean, an average cost of refurbishment, including reverse pickup cost, will be somewhere to the order of 25% to 30% of the cost of a new device. So it reduces CapEx meaningfully. We also think that at the moment, we can continue with or even scale up a little bit the pace at which we pick up and refurbish. So both the pickup capacity as well as the refurb capacity, we see a long runway there because of the scale of device deployment, which is 13 million. And the great thing is our business leaders are super sensitive to how positively this impacts their P&L. So they have made the investments required to make sure that we can both pick up and refurbish at scale.

The next question is from Mr. Prateek Poddar from Bandhan Mutual Fund.

Sir, could you just talk a bit about the scalability of postpaid products in terms of whitelisting of consumers, penetration over there and adding new partners on the supply side so that this can be scaled even faster? And in your view, time line for this product to reach the previous peaks or which we reached, let's say when we had this earlier avatar of postpaid Paytm perspective?

Interesting. Numbers are very early, tens of thousands, so the opportunity is way more fold. I mean, the good thing I can say is that we experienced that this country we had signed up around 10 million, 12 million customers and surprisingly, most of those customers are on our platform. So we can try talking about large -- I mean, can we target couple of million? Answer is yes. Spend per customer is already high because the product is mature and understanding of the product exists. Then the understanding of multiple banks and technology is becoming better because there is a anchor issuer now as a partner with us and we have anchor issuers, so more expected. So I do believe that it is going to become resizable, hopefully. It obviously depends on banks because it is a purely bank product and we continue to help banks pursue them to expand the technology and we are happy that banks have -- partners have expanded the technology. So you expect us to grow both consumer and bank partners.

And the time lines for this would be let's say in the next 12 months and I'm not asking for answers, but we will see multiple number of suppliers or bank partners over here?

[Foreign Language] We do not need more partners than one in 1 year, the number of customers and the book of opportunity because it is a -- if you notice, it is a very high throughput churn, so it does not have a residual book which is the advantage. So even a small...

No, but supplier will get their own demand, right? That's what happened in ML also.

Yes, you got it. So because of high rotation not requiring large residual book makes it even more advantageous versus people with large books. So the good advantage here is cost of capital should be lower, not the amount of capital should be higher.

Understood. And in that case, a pure bank will have a lower cost of capital than SMB, right? Today we have tied up with an SMB.

[Foreign Language] give or take 20 bps [Foreign Language] don't bother about margins. They are not suppressed or pressed.

Let me ask you the other way. The capital commitment provided by the partner, is it large enough for you to reach the old peaks or get towards there, that's not a constraint?

That is the question. Perfectly said. Yes.

The second and the last question is on the DLG disclosures and when I see the outstanding book ex the lender 1, which has moved away from DLG to a non-DLG, I see a substantial rise on a quarter-on-quarter basis, it's like 80% plus. Just wanted to check how should I think about this and the growth from those partners looks -- I mean that ramp up is phenomenal, right?

Yes. I think one thing to point out and you already know this, Prateek, but just to contextualize the question, that number is an AUM number. So even if you continue to the same amount in the early days, that number will go up very meaningfully. But it is the case that excluding [ partner one ], which is our largest partner overall in the loan distribution and is also -- and continues to be the largest partner, right? And continues to be a partner that is growing AUM with us, but also -- but without DLGs. So the reason that partner's AUM under DLG has come down is not because the AUM has gone down, but because the DLG has gone away. So the new book is getting booked effectively without DLG. And on other partners, we are seeing very, very good growth both in merchant side and personal loan side. Personal loan, obviously, in the context of the headwinds. But we're seeing very good growth in non-lender one, if you will, both on ML and PL. So -- and I would say particularly in ML, where I mean, just to give you a sense, non-lender one has more than doubled in the last 12 months. So it's a fantastic job by the team that while lender one is growing, they have also been able to grow, as I said, more than double non-lender one. And that's a mix, obviously, as you might imagine, of other lenders growing very fast and also new lenders coming in. So it is -- overall, this is a much higher percentage of non-lender one today than it used to be.

Okay. And sorry, just going back to Postpaid, [indiscernible], the postpaid product. Look, capital is not a constraint. We have the consumers. What does it take and when do you feel confident to get to that J curve, right? Today is early days. But when does this get to -- is it months, is it years? How should I think about it? Because the capital is not a constraint, right? If capital was a constraint, I wouldn't have asked you the question. You just said capital is not a constraint here.

[Indiscernible] last 18 months. Slow and steady wins the race. When you race it fast, you may just get out of the race.

And I think we can give you a little bit more next quarter because like I mentioned, we launched in September. We have October numbers, obviously, the operating metrics which was beyond the period that we are talking about on this call. By the time I think we have the call sometime in January for December results, I think we'll be able to point to specific point, data points and then you'll get a better sense of where this is heading. But to be clear, the idea, as Vijay is saying, is not can we reach the previous number in 6 months? We're trying to do it as sensibly as we can. Of course, there are a ton of learnings from before, all the things that work really well and some of the things which maybe we don't want to do it identically as before. So we have all of that learning, all of that understanding of consumer and merchant behavior for this product. So we're using all of that. So logically, our ramp up should be faster than before. But we're not trying to say, hey, why can't we get the same number as soon as possible?

[Foreign Language].

Have seen the compounding benefit of running a business really sensibly, particularly in merchant loans, where we have always talked about we could do a lot more in the next 6 months if we started to get a bit aggressive. But the idea is let's keep the penetration low, let's reap all the benefit that comes with the fact that the partner is seeing fantastic credit quality and existing and new partners want to do more business with us and that just builds a really strong business long term.

With that, we come to an end of this call. A replay of this earnings call and the transcript will be made available on the company website subsequently.

And if I may just add, we ran out of time. There might be other questions. Please feel free to reach out to [email protected] or -- and we will be delighted to answer -- give any clarifications there.

And I really want the feedback of the video call. If you guys like it, prefer it over audio or do you like the old way of -- traditional way of doing it? We thought that we don't get to see you one-on-one so many times or in different kind of conversations opportunities. So this is the way and I was asking Madhur that how much of...

I liked it if you were looking for my feedback. I quite liked it.

I also loved it actually that we are able to see and thank you for earlier, Pranav, Pranav square, I may call them that you switched on the video. So we will encourage more people to come on the video, but it is very awkward. Awkward hours can happen for many people in different place and moment. But we love to make it more interactive and conversation and I mean, this gives us opportunity to showcase something, present something. So now that [Foreign Language] Thank you.

Thank you, so.

Thank you all for joining.

Have a good day.

You may now disconnect your lines.