Skip to content
Nikhil KamathNikhil Kamath

If UPI Is Free, Who Pays for It? | PhonePe, Razorpay & NPCI | WTF is Payments?

Access the full data and references used in this episode below. Please open this alongside the episode while watching. https://www.canva.com/design/DAHVQSlqQZQ/Nfv8N2FMGz22OAcanl0LFg/edit We scan a QR code, enter a PIN, and move on. I wanted to understand everything that happens in between. If a payment feels instant and costs us nothing, who is doing the work, who is taking the risk, and who is paying for it? I sat down with Sameer Nigam of PhonePe, Harshil Mathur of Razorpay, and Sohini Rajola of NPCI to follow a payment through the system. We draw out the journey between an app, the banks and NPCI, get into why the confirmation on our phone and the settlement between banks are different things, and compare that with what happens when we use a credit card. From rewards and merchant fees to fraud, failed transactions and the cost of handling cash, there is a business behind every way we choose to pay. The conversation also gets into MDR, how the ecosystem funds itself, and what it takes to keep investing in wider access. Then we look ahead: cross-border payments, business payments, wearable rings, digital tipping, and AI agents that might eventually shop and pay for us. The guests disagree on how quickly that future arrives, especially in India. If you were starting a payments business today, which problem would you try to solve? Watch the full conversation and tell me below. Timestamps: 00:00 Introduction 02:23 Sohini Rajola: Her journey and NPCI's role 09:40 Banks, fintechs and digital banking 14:11 Harshil Mathur: The problem that led to Razorpay 20:36 Sameer Nigam: His journey before PhonePe 24:08 Why young Indians want to leave 30:28 UPI's early builders and the cost of overregulation 43:16 Flipkart's payment failures and the birth of PhonePe 52:02 How a UPI payment actually works 59:06 If UPI is free, who pays for it? 1:05:26 Instant payments vs bank settlement 1:08:24 Stablecoins, CBDCs and the future of money 1:14:52 Wealth inequality, inheritance and capital gains 1:17:11 The zero-MDR debate: Funding the payments ecosystem 1:21:31 Should UPI apps have market-share caps? 1:33:08 Can new banks and startups compete in India? 1:41:18 How card and merchant payments work 1:49:39 Credit-card rewards and RuPay on UPI 1:54:12 How Razorpay improved online payments 2:02:08 The B2B payments opportunity 2:05:36 Building payments for medical tourism 2:09:54 Why digital tipping is still unsolved 2:16:01 What payment data reveals about India's consumers 2:21:22 Trust and India's next wave of online shoppers 2:26:07 GST, consumption and property taxes 2:31:38 Affordable food, quality and accountability 2:37:55 Cash vs digital: Measuring India's spending 2:42:20 B2B payments, merchant data and access to credit 2:44:31 Stablecoins, cross-border transfers and capital controls 2:51:59 Will AI agents change how we shop and pay? 2:57:57 Is AI in a bubble? 3:03:01 AI costs, productivity and human work 3:08:59 What should young founders build for AI commerce? 3:16:01 Who benefits from AI? Business and investment opportunities 3:24:31 Meet the builders: Seven, Dodo Payments and PayGlocal 3:26:46 How cross-border payments and merchant-of-record models work 3:34:55 Stablecoins and CBDCs: Speed, trust and regulation 3:39:20 The founders' view on competition and market share 3:41:31 Seven's payment rings: Cashback, revenue and adoption 3:47:20 Medical tourism and other cross-border opportunities 3:51:07 Tap-to-pay, tipping and the future of wearables 3:53:48 Cheaper AI and the pace of agentic commerce 3:56:47 Fraud, trust and safeguards for AI payments 3:59:18 Final advice for young fintech founders Sohini Rajola- Executive Director Growth at the National Payments Corporation of India. Linkedin:https://www.linkedin.com/in/sohinirajola/ Sameer Nigam- Founder and CEO of PhonePE Linkedin: https://www.linkedin.com/in/sanigam/ Harshil Mathur: Co-founder and CEO of Razorpay Linkedin: https://www.linkedin.com/in/harshilmathur/ Also featuring perspectives from Vijay Khubchandani, Founder & CEO, Seven LinkedIn: https://www.linkedin.com/in/vijaykhubchandani/ Prachi Dharani, Co-Founder & CEO, PayGlocal LinkedIn: https://www.linkedin.com/in/prachi-dharani-94530a9/ Yogesh Lokhande, Co-Founder, PayGlocal LinkedIn: https://www.linkedin.com/in/yogesh-lokhande-b760522/ Ayush Agarwal, Co-Founder & CPTO, Dodo Payments LinkedIn: https://www.linkedin.com/in/aagarwal1012/ Nikhil Kamath-Co-founder of Zerodha and Gruhas Host of 'WTF is' & 'People By WTF' Podcast Twitter: https://x.com/nikhilkamathcio/ Instagram: https://www.instagram.com/nikhilkamathcio/ LinkedIn: https://www.linkedin.com/in/nikhilkamathcio?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=ios_app Facebook: https://www.facebook.com/nikhilkamathcio/ Watch 'WTF Online' on Spotify https://tinyurl.com/4tjua4th #WTFiswithNikhilKamath #PeopleByWTF #WTFOnline

Nikhil KamathhostSohini RajolaguestHarshil MathurguestVijay KhubchandanicameoAyush AgarwalcameoPrachi DharanicameoYogesh Lokhandecameo
Oct 30, 20264h 0mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 2:23

    Introduction

    1. NK

      [upbeat music] The entire history of money has been one move, repeating. Money keeps getting lighter. It started as grain and cattle, something you grew or herded or could physically run out of. Then it became gold coins stamped by kings, so heavy with meaning that entire empires ran on the weight of it. Gold became paper, a promise, instead of the metal itself. Paper became plastic, and in India, money became a sound. The next move for money might be the strangest one yet, something I've asked so many people, but nobody knows clearly. If you were to ask an AI data center person, he would probably say, "Money is moving to energy." I don't know which way this will play out. Maybe there is a business to be built around betting what that change might be. Grain, gold, paper, plastic, stablecoin, things on the blockchain, and I'm sure it'll change tomorrow as well. Today, I'm talking to people who built the last version of this story, the people who turned money into a sound. Sohini is the executive director of growth at NPCI, the organization behind UPI. Harshil Mathur is the co-founder of Razorpay. Sameer Nigam is the founder and CEO of PhonePe. We'll get into their journeys, the wins, the losses, the years nobody was watching, and where the next move in this history gets made. Because if you want to understand how India spends or how to build the next big thing in this country, there's no one better place to teach you than the three people who will be sitting across from me today. Should we start by drawing a d- diagram of the different kind of payments and map out how it began? [upbeat music]

  2. 2:23 – 9:40

    Sohini Rajola: Her journey and NPCI's role

    1. NK

      Okay, thank you guys for coming and doing this. So let's begin with introductions and, uh, who would like to go first?

    2. SR

      Okay.

    3. NK

      Okay.

    4. SR

      I can go first. Uh, so we'll start from the very beginning. So I was born in Rajasthan, uh, went to this lovely school in Jaipur, school and college both actually initially in Jaipur. Um, and then went, did my MBA at this exclusive college for defense personnel and their dependents only, Symbiosis in Pune. Um, joined a bank, Axis Bank, spent, spent about 18 years doing lots of interesting non-core bank kind of things, so all of payments, all of digital, all new initiatives. So, uh, you know, I think that was kind of the grounding. Then went and ... Oh, before that, actually, in college, uh, um, met my boyfriend and then now husband, my husband for more than 25 years now. Um, then went to, uh-

    5. NK

      What's his name?

    6. SR

      Sitesh. Yeah.

    7. NK

      Is he a regulator?

    8. SR

      No. [laughing]

    9. NK

      [laughing]

    10. SR

      More the moderator [laughing] than the regulator, I think. But yeah. Uh, then, um, I worked in, uh, Western Union, the money transfer company, that took me across, uh, the shores. Spent some time, uh, learning about a lot of different countries and the way businesses are run there. And, uh, almost like full circle, came back last year, um, to come and lead growth for NPCI, which actually can be a little interesting because how do you grow a company which is already doing, like, roughly a billion transactions a day? So I'm still figuring that one out. Uh-

    11. NK

      Is it a government posting or is it, like, private sector?

    12. SR

      You know, that's a very interesting question.

    13. NK

      Sh-

    14. SR

      Almost everybody asked me that question when I was joining as well, to say, "So are you going to be working for the government?" My father wanted to know if I'll get a white ambassador-

    15. NK

      [laughing]

    16. SR

      ... with the, with the Ashok thumb on it or not. Unfortunately, no. Um, so it, it, this was formed by RBI and IBA.

    17. NK

      Mm.

    18. SR

      Uh, so it's set up as a not-for-profit.

    19. NK

      Mm.

    20. SR

      Um, and yes, there is an element of us working b- because it's a digital public infrastructure, so we work for, we work with closely with the government and the regulator.

    21. NK

      It's a Section 8 company, right?

    22. SR

      Yes, it is. It's set up as a not-for-profit.

    23. NK

      Who funds it?

    24. SR

      Uh, so it was set up, uh, with an initial funding of 100 crores.

    25. NK

      By?

    26. SR

      Uh, by 10 banks.

    27. NK

      Mm-hmm.

    28. SR

      Um, and, uh, now actually after that, it has generated enough surplus because it's a, you know, it's not profit and it's not returned back to the shareholders, but all of the growth and the investments in the tech and the infrastructure and everything has been kind of self-funded over the-

    29. NK

      Do Section 8 companies have shareholders?

    30. SR

      Yes.

  3. 9:40 – 14:11

    Banks, fintechs and digital banking

    1. NK

      did the Axis Bank, uh, experience teach you?

    2. SR

      Oh, a, a lot of things, actually. So-

    3. NK

      Are banks becoming irrelevant? I've thought about this. To a certain extent, regulation is harder for them to compete-

    4. SR

      Mm

    5. NK

      ... on a fair keel. Like, if I'm trying to borrow money on the back of, say-

    6. SR

      Mm

    7. NK

      ... private equity-

    8. SR

      Mm-hmm

    9. NK

      ... a bank is not allowed to lend. I have to go to a NBFC.

    10. SR

      Mm-hmm.

    11. NK

      Like that, in many departments-

    12. SR

      Mm

    13. NK

      ... banks are over-regulated or regulated so much more than a fintech player-

    14. SR

      Mm

    15. NK

      ... that that less regulation itself becomes a moat for fintech players.

    16. SR

      There are many ways to look at that, Nikhil.

    17. NK

      Mm.

    18. SR

      Because see, why are ... Uh, see, this is, you're taking one part of it in isolation.

    19. NK

      Yeah.

    20. SR

      Because if you look at it, banks have to be regulated because the NBFC is not taking deposits from the customers.

    21. NK

      Correct.

    22. SR

      The fact that you're taking the public money-

    23. NK

      Mm

    24. SR

      ... and then you're doing XYZ with it, you have to be regulated.

    25. NK

      Mm.

    26. SR

      So, you know, yes, this, you can say that's the moat for the fintechs, but then there are other moats for the banks, and banks have got to leverage and work on their moats to be able to differ-

    27. NK

      Just cost of capital, no? Do they have any other moat?

    28. SR

      No, uh, there's cost of capital. There's also, uh-

    29. NK

      And even in that, like, what, 2, 3% lower?

    30. SR

      It's not just about 2, 3% lower.

  4. 14:11 – 20:36

    Harshil Mathur: The problem that led to Razorpay

    1. NK

      Okay. Harshil, would you like to go next?

    2. SP

      Sure, sure. And by the way, I didn't know you are from Jaipur. I'm from Jaipur as well.

    3. SR

      Wow, where?

    4. SP

      Yeah. I grew up in Malviya Nagar.

    5. SR

      Okay.

    6. SP

      Schooling in Seedling.

    7. SR

      Okay. So I went to MGV.

    8. SP

      Oh, nice, yeah. I mean, we spoke so many times, we didn't know we are both from Jaipur.

    9. SR

      [laughs] Okay.

    10. SP

      I was born in Jaipur. Uh, grew up, uh, in an average middle-class family in Jaipur. Dad was a, you know, worked in a bank. Um, so-

    11. NK

      Manager

    12. SP

      ... bank, he was a clerk, actually. He became a manager by the time he retired, uh, in SBI.

    13. SR

      Mm.

    14. SP

      And then, uh, so went to college in IIT Roorkee. Uh, was always-

    15. NK

      Same my dad. Became a manager around when he retired. Canara Bank.

    16. SP

      Yeah, close to two years before he retired and so-

    17. NK

      Yeah, Canara Bank.

    18. SP

      Yeah, so, because you typically get posted out of, of the city-

    19. NK

      Yeah

    20. SP

      ... when you become a manager, so he awarded it.

    21. NK

      We get transferred every three, four years, right?

    22. SP

      Yeah, so that's why he awarded till retirement-

    23. NK

      Ah

    24. SP

      ... uh, till my schooling was done and all of that. So, so yeah, went to IIT Roorkee. Was always very deeply interested in computers from, like, school days. Used to code stuff on the side, but unfortunately didn't make the rank in IIT, so got into mechanical. Uh, didn't know what to do with mechanical. I have zero interest in it.

    25. NK

      You must have got, like, ninety-nine point two percent instead of five, right?

    26. SR

      [laughs]

    27. SP

      Yeah, I mean, rank.

    28. NK

      Did not make the rank. [laughs]

    29. SP

      Rank, so I got eighteen hundred rank. You need to get under a thousand to get computer science. Uh, so took mechanical, had zero interest in it.

    30. NK

      Mm.

  5. 20:36 – 24:08

    Sameer Nigam: His journey before PhonePe

    1. NK

      You tell.

    2. SP

      What should I say? I am, uh-

    3. NK

      Whatever you want to say. Mm.

    4. SP

      I was born in Bombay, um, into a naval family. My dad was in the navy.

    5. NK

      Mm.

    6. SP

      A small town guy from UP, went to IIT. Naval architect, so I was born here. Um, I started my schooling in Germany.

    7. NK

      Will you have coffee?

    8. SP

      Of course.

    9. NK

      Black?

    10. SP

      Yes, sir.

    11. NK

      Coffee? Black is good. Black?

    12. SP

      Coffee, yeah.

    13. NK

      Black?

    14. SP

      Coffee is always welcome.

    15. NK

      Ah. So Germany, yes.

    16. SP

      Ah. Um, so that was interesting. It was, like, a four-year-old kid. Learned German before I learned English.

    17. NK

      You still know it?

    18. SP

      Uh, very poorly. I understand more than I can speak.

    19. NK

      Right.

    20. SP

      Um, but the coolest part was, I think at age seven, when they completed India's first submarine, I got ... We all got, families got invited. We sat in a submarine at age seven. I'll never forget that in my life.

    21. NK

      Wow.

    22. SP

      It was the only, only thing I remember, uh, except German, besides German chocolates. Those are the two things I took away.

    23. NK

      Mm.

    24. SP

      Moved back, um, to Delhi. Spent about six years there, and then at 13 I moved back. As a teenager, I moved back to Bombay.

    25. NK

      Mm.

    26. SP

      And it's the best damn place and time to move back. Uh, early 90s was fun times in Bombay.

    27. NK

      Yeah.

    28. SP

      Partied a lot.

    29. NK

      Yeah.

    30. SP

      Met my co-founder now, Rahul, in undergrad. Uh... [laughs]

  6. 24:08 – 30:28

    Why young Indians want to leave

    1. NK

      people went abroad, did MBA, a lot came back.

    2. SP

      Mm.

    3. NK

      Why are they not coming back now?

    4. SP

      I think by seven people, the first crop of people, post liberalization who'd gone, had seen enough, and there was a view that India's digital moment will come. I think right now the macros are not in favor. I'm talking to a lot of young kids. My son's turning 18.

    5. NK

      They don't wanna come back now?

    6. HM

      I don't think that the peop- middle class aspirins don't wanna be here.

    7. NK

      Hmm.

    8. HM

      And I think it's sad. And, um, I mean, on a serious matter, I think, I think it bothers me because I came back-

    9. NK

      Hmm

    10. HM

      ... um, I've had a great run. You've had a great run. But I think-

    11. NK

      When you say macros are bad, what do you mean?

    12. HM

      I think there's a population of people that has grown up in the Insta world, in the, in the YouTube world that has, that has seen what's possible out there, and they're seeing governments and markets opening up, and they're seeing governments not sort of getting in the way of just unleashing animal instincts. And I think we, we are at an inflection point. We are in a weird inflection point. I feel like India's still living in the past, and I think we have 300, 400 million people who want to just look ahead. They don't, they're not really married to the past. When I talk to my kids, I mean, these are Gen Zs. My daughter's 13.

    13. NK

      Mm-hmm.

    14. HM

      Um, you see what's happening around socially right now with unrest. They, they don't really care if we are 5,000 years old, 3,000 years old. They wanna move in traffic. They wanna get to some place. They wanna move ahead. And they should, damn it. Like, I think they should.

    15. NK

      Even we think like that, no?

    16. HM

      Yeah, they should. That's what I'm saying. So I'm, I'm rebellious enough, but it frustrates me. I've seen better, um, but I made a conscious choice. I remind myself every day. I made a conscious choice to move back. I think a lot of people who are seeing things, um, are not making that choice today, and it's, it's a little sad.

    17. NK

      Yeah.

    18. HM

      Well, I think

    19. NK

      But But

    20. HM

      I think

    21. NK

      You think it'll change?

    22. HM

      I think , um, if we, if we change our relationship societally with government, I think it'll change.

    23. NK

      Say again?

    24. HM

      If we change our expectation and relationship with soci- uh, with governments, and I think for India, we are maybe 30, 40, 50 years away. Uh, there's, there's the World Bank of India. There's 500 million people who depend on the government machinery to keep things moving, and there's 500 million of us who are now saying, "You're slowing us down." So India is, India is this contrast of realities. Where are we at as a country? We are 75 years old. We are a young democracy, and in many ways it's like one of the most amazing social experiments on the planet. We are borrowing a lot of pride or glory from the past to feel energized. I mean, especially with th- this government, but in general, how do you feel good if you're coming out of 700, 800 years of servitude and poverty? You have to look as far back as you can, or if you don't have that history, then you just look forward.

    25. NK

      Hmm.

    26. HM

      The US never had a history to worry about. They only look forward. They have n- they, they can't talk about the past. George Washington

    27. NK

      Hmm

    28. HM

      ... say a very interesting thing. For a civilization to do well, it's important that we believe that we are special.

    29. NK

      Yeah. Yeah, yeah.

    30. HM

      Americans inherently believe that they deserve more because they are special-

  7. 30:28 – 43:16

    UPI's early builders and the cost of overregulation

    1. HM

      UPI is a great example. I was there on ground zero-

    2. SR

      Mm-hmm

    3. HM

      ... and I promise you this, for all the tom-tomming we do-

    4. SR

      Hmm

    5. HM

      There was maybe 20 people responsible for it. I'm not talking about enabling NPCI, I'm talking about building UPI- Hmm, hmm ... and a lot of them are sitting in Bangalore. Hmm. And it's the paradigms that shaped UPI. Hmm. Is going mobile first. Hmm. It's about building tech that's interoperable. It's about building APIs that, that people can build on. Those are technology choices. Yeah. There was no brouhaha, there was no chest-thumping, none of that. Yeah, yeah. It followed so quickly that today, and you know this and I know this- Yeah ... let's be frank, right? You can't take a step forward without taking at least .75 back- Yeah ... every single time. Hmm. Because we have this incredible habit as a, as a country- Hmm ... that we, we almost invite regulation and government into places- [laughs] ... that has no goddamn business into. They have no... Like, if I think about AI, the number of committees that different departments of different regulators, we are regulated above, as you- Yeah ... across regulators, we have government bodies across the board. Yeah, yeah. I just keep going into AI reg meetings and I'm like- Yeah ... "You have no idea what you're talking about." We have barely any LLMs of our own- With the best of intent ... but we have enough and more- No, but with, with the best of intent- ... governing frameworks ... but no one's willing to call out- Hundreds of governing frameworks ... but no one's willing to call out just how-

    6. NK

      What is Microsoft gonna do for us and-

    7. HM

      ... how dangerously backwards it is to not understand what you're trying to regulate. Hmm. And that's why I'm saying I think it'll take 40, 50 years because one generation will just revolt. When I see our kids, they don't care. But you want us, want it to take 40, 50 years? I don't want it to. I'm, I'm extremely noisy and I'm extremely deliberately noisy about these things. My dad was an impatient man. Hmm. My mom was an entrepreneur in the '80s, right? So you're talking about being in banks. Hmm. She could not open a bank account for her first business unless her spouse or her father co-signed. This was India in the '80s. Yeah, yeah. That lasted into her 90s, into her second venture.

    8. NK

      Really?

    9. HM

      Oh, yes, sir. What is-

    10. NK

      Yeah?

    11. HM

      You needed a male, male co-sponsor for a bank account.

    12. NK

      Yeah.

    13. HM

      Um, they were, they were stun-

    14. NK

      Introducer.

    15. HM

      Yeah. You, yeah, that was the word.

    16. NK

      Introducer

    17. HM

      To open a bank account someone had to introduce you-

    18. NK

      Introduce you to the bank

    19. HM

      ... and that's why the-

    20. NK

      I like that the pseudo regulators agree, agreeing with the overregulation.

    21. HM

      No, it's not even-

    22. NK

      We are agreeing with the overregulation [laughs]

    23. HM

      I object to, I object to both the terms, the pseudo and the regulator, Nikhil. [laughs]

    24. NK

      [laughs]

    25. HM

      No, I agree with Sameer, it's not even regulator coming in. A lot of times I've seen people say, "Hey, this should be regulated."

    26. NK

      Yeah.

    27. HM

      "This should be..." Like, why do you want to bring government into every single thing?

    28. NK

      Why do you think people say that?

    29. HM

      I think there's just inherent, um, as some people put it, like trust-

    30. NK

      Because, because we've been conditioned to believe-

  8. 43:16 – 52:02

    Flipkart's payment failures and the birth of PhonePe

    1. NK

      Aage ka batao. You were at Flipkart. Your company was bought by Flipkart.

    2. HM

      Huh. Then, uh, four years into it, we saw this rocket ship. Four years into it, um, there was this very interesting decision to try and build India's Black Friday moment.

    3. NK

      Mm.

    4. HM

      And, uh, my co-founder and I, I think this is why we work together on payments now, um, he was heading the supply chain. The order management udhar thi, checkout and payments mere saath tha.

    5. NK

      Mm.

    6. HM

      We spoke to all the banks and we said we want to do this massive sale. Everyone's gonna be squatting on their apps at 8:00 AM, and we'll open the floodgates. So we spoke to HDFC-

    7. SR

      Mm-hmm

    8. HM

      ... we spoke to Axis, ICICI, uh, sorry, uh, SBI. And we asked them, "Can you, like, 100X the capacity?" And they said, "Why?" And we said, "We are doing this crazy sale, and we are 100X-ing our capacity." I used to walk around with a placard every, every evening saying 25X. All servers, everything had to be scaled up. Um, we had the sale, the first Big Billion Day sale.

    9. NK

      At Flipkart?

    10. HM

      At Flipkart. And this, in those days it was a billion day sale.

    11. NK

      This is Sachin, Sachin was-

    12. HM

      This is Sachin, Binny, Mukesh was there at that time.

    13. SR

      Sujeet.

    14. HM

      Sujeet was there at that time. Um, Mukesh had just joined. Sujeet was there. Um, and-

    15. NK

      They're all same area-

    16. HM

      Yeah

    17. NK

      ... in a weird way. Sachin was in the same apartment. Sujeet is one of my best friends.

    18. HM

      So Kalyan had, was there-

    19. SR

      This is, this is that-

    20. HM

      ... and he was-

    21. NK

      Flipkart founders

    22. SR

      ... this is that place called Adarsh in Bangalore.

    23. HM

      No, that's now-

    24. SR

      In fact-

    25. NK

      Sujeet, Sujeet lives in Adarsh

    26. HM

      ... I live in Ejipura

    27. SR

      That's what everyone says

    28. HM

      They were in Raheja.

    29. SR

      Okay.

    30. HM

      At that time I was Kotla.

  9. 52:02 – 59:06

    How a UPI payment actually works

    1. NK

      Should we start by doi- drawing a d- diagram of the different kind of payments and map out how it began?

    2. SR

      Wahan se shuru karein?

    3. HM

      Sure.

    4. NK

      Okay. So I'm gonna have a bite and then I will organize my papers by then. There are 20 pages of questions I have. So before UPI, we spoke a little bit about the origin of UPI. Before UPI, how did payments work? Let's take the Visa and Master example, and let's take the bank-to-bank example. Who would like to take up the Visa and Master example?

    5. HM

      Sure.

    6. NK

      I, I know the top-line numbers that Visa and Master charge typically 2% to facilitate a transfer.

    7. HM

      They don't charge that much.

    8. NK

      Huh. The issuer bank makes 1.4. Visa and Master make a margin of 0.2 or two rupees.

    9. HM

      Slightly less.

    10. NK

      And the acquirer bank makes 3.5 rupees or 0.35.

    11. HM

      Broadly in that range.

    12. NK

      In that-

    13. HM

      That's for credit card, just to be clear.

    14. NK

      Yeah.

    15. HM

      Debit was 65 basis points.

    16. SR

      Debit is much lower.

    17. HM

      Yeah.

    18. NK

      Visa and, Visa and MasterCard making two rupees for a transaction of 1,000 ends up making $40, $60 billion of profit-

    19. HM

      Yeah

    20. NK

      ... across the world.

    21. HM

      Scale.

    22. NK

      Yeah?

    23. SR

      Scale.

    24. HM

      Scale. Low margin, high scale.

    25. SR

      Yeah.

    26. NK

      Okay, we can come back to Visa and Master next. Let's start with how this works.

    27. SR

      Okay. [laughs]

    28. NK

      How does UPI work?

    29. SR

      Okay.

    30. HM

      No, but I think the core principle is the same, whether you have Visa or MasterCard or not.

  10. 59:06 – 1:05:26

    If UPI is free, who pays for it?

    1. SR

      So there is a cost here. You had asked the cost question.

    2. NK

      Yeah.

    3. SR

      There's a cost for every entity, even your bank, Canara Bank.

    4. NK

      What is it? Is it like a data cost for your 10 API calls that, which is the main cost?

    5. SR

      For PhonePe, GPay, Paytm, the apps, it's the cost of the banks, because now earlier you were walking into a bank withdrawing 5,000 rupees, so that was one transaction.

    6. NK

      Mm-hmm.

    7. SR

      Now you're making 50 such transactions, so there's the core banking. For every transaction, the bank sends you a SMS. The, so, so there's the bank, the PSPs, the apps, and the, if there is a merchant transaction, which we haven't gotten into-

    8. NK

      Mm-hmm

    9. SR

      ... the acquiring party's, uh, costs. So that is the entire cost.

    10. NK

      Even before we go there, now my money has transferred from Nikhil to Rushank, my Canara Bank to HDFC.

    11. SR

      Yeah.

    12. NK

      How much has it costed?

    13. HM

      Everybody put together?

    14. SR

      Everybody-

    15. NK

      Yeah

    16. SR

      ... put together for the transaction, we had ... I think there was some calculation that was done.

    17. HM

      Point, 0.25%, broadly.

    18. NK

      25 paise.

    19. SR

      Yeah.

    20. HM

      No, it's not 25 paise. I think, I think if you net it out at an ecosystem level, it's about 25 bps.

    21. NK

      Which is 25 paise.

    22. HM

      Not rupees.

    23. NK

      On 100 rupees.

    24. SP

      On 100 rupees.

    25. SR

      On 100, yeah, yeah. He said 100 rupees, so.

    26. HM

      So this is where B2B and P2P dynamics are different.

    27. NK

      Mm.

    28. HM

      Retail dynamics are different.

    29. SP

      P2P-

    30. HM

      Because the switching fees and all the other fees come in.

  11. 1:05:26 – 1:08:24

    Instant payments vs bank settlement

    1. NK

      And because these transactions are instantaneous, there's no concept of float anywhere. Nobody gets to keep the money for a while before the receiver gets it.

    2. SR

      Yeah, no, this is unlike, like, what you have in, say, China, et cetera-

    3. NK

      Mm

    4. SR

      ... where the Alipay, WeChat, which are basically-

    5. NK

      They have a float part, right?

    6. SR

      There's, those are wallet-based.

    7. SP

      Those are closed loop.

    8. NK

      Yeah.

    9. SR

      This is a real-time instant payment system, so money moves from ... The float is already either in your bank account or the other person's bank account, so they are in any case meant to, banks are meant to have the float, so none of the other intermediaries, uh, monetize this.

    10. SP

      See, earlier the bank who sent the money was making that float, now the bank who's received it is going to make the float.

    11. SR

      Yeah.

    12. SP

      So the float is made either way, but-

    13. HM

      One of the, one-

    14. SP

      ... the system is not making the float

    15. HM

      ... one of the coolest things about the NPCI architecture or UPI architecture is the real-time payment just basically is a accounting and debit or credit-

    16. NK

      Mm

    17. HM

      ... that's happening, whereas they do the settlements-

    18. SR

      Yeah

    19. HM

      ... out of, out of band. What that-

    20. NK

      Really?

    21. HM

      Yeah, because it-

    22. NK

      When does that happen? After how much time?

    23. HM

      That pool stays with UPI or with NPCI, and then they settle, what, eight times a day now?

    24. SR

      We settle 10 times a day.

    25. HM

      10 times a day. So it's a massive, I would imagine, a massive pool sitting there for the settlements. They're able to move that quickly because it's much, much more just about ledger management than money moving.

    26. NK

      When you say it sits at UPI, it sits at NPCI?

    27. HM

      NPCI.

    28. SR

      So it's in the RTGS account in RBI, so you have to move it from bank A to bank B, and you do that settlement, do that settlement file posting.

    29. HM

      So you could have 100 million transactions in two hours in peak hour, but ultimately the cross settlements would be like 800 settlements, 400 on either-

    30. SR

      Yeah

  12. 1:08:24 – 1:14:52

    Stablecoins, CBDCs and the future of money

    1. NK

      ... , in the end, we'll go to it. We'll go to what Stripe is doing, what's happening with stablecoins. Is there opportunity in cross-border payment? Somewhere somebody will build. I was interviewing this, uh, Coinbase founder, Brian Armstrong.

    2. SP

      Brian, yeah.

    3. NK

      He made a very strong case for stablecoins. India may dollar because-

    4. SR

      Mm

    5. NK

      ... I don't think the government will allow it. But what about the use case for a stablecoin with something else backing it?

    6. SP

      Yeah, especially for cross-border, I think there's a lot of problems to solve.

    7. NK

      Yeah. See, nobody wants to hold rupee as paper because of how quickly it's depreciating. But if it could be based on a commodity that India holds, maybe there's an opportunity there. But we'll get into-

    8. SR

      Yeah

    9. NK

      ... things, things like that people will want to build.

    10. SP

      Yeah.

    11. SR

      Mm.

    12. NK

      But the one thing I realized with Brian is in that entire crypto world, the only thing that's left is Bitcoin and stablecoin.

    13. SP

      The value of the stablecoin. Uh, stablecoin takes away the speculation part, no. So fair- The problem with central banks across the world is they're so in love with the concept of their CBDCs, which is a much more complex problem. I think in 10 years you'll end up with peering of UPI with a lot of networks and picks and all these other networks. Real-time networks peering will move much faster than CBDCs.

    14. SR

      CBDCs-

    15. SP

      Because nobody wants to rely on somebody else's CBDC.

    16. NK

      It's work, no?

    17. SP

      No.

    18. NK

      Nobody's gonna run on somebody else's CBDC, so we're still creating disjointed outcomes.

    19. SP

      Huh. No, and then-

    20. NK

      Commonly, yeah

    21. SP

      ... egos and price come in, right?

    22. NK

      Yes.

    23. SP

      Mera CBDC bada, uska CBDC bada.

    24. NK

      But what if government

    25. SP

      No, you're talking about-

    26. NK

      So I'm talking about for cross-border.

    27. SP

      For cross-border.

    28. NK

      I'm talking internally for the country.

    29. SP

      Internally, UPI solves it. Why do you need CBDC?

    30. NK

      UPI, it will solve it.

  13. 1:14:52 – 1:17:11

    Wealth inequality, inheritance and capital gains

    1. HM

      Yes. Can I play provocateur here?

    2. SR

      Mm-hmm.

    3. HM

      Um, so at heart, I like old-school American capitalism, which is-

    4. NK

      Hmm

    5. HM

      ... grounded in the concept that regardless of birth, enterprise in one lifetime, you can actually just become anybody you want to be.

    6. NK

      Yeah.

    7. HM

      I think capital markets are extremely distorted right now globally, in that money makes money way faster than labor makes money or enterprise makes money, and that's a global trend, which is why income disparity is out of whack. So role of government, ironically, in the democracies is in need much more than ever. I feel that.

    8. SR

      Yeah.

    9. HM

      That said, um-

    10. NK

      But that is working in some places.

    11. HM

      In some places.

    12. NK

      Yeah.

    13. HM

      Now-

    14. NK

      There's something called the Gini coefficient, which measures-

    15. SR

      Yeah

    16. NK

      ... income inequality. And in places like Australia, some Nordic countries-

    17. HM

      Nordic countries, yeah

    18. NK

      ... it is coming down. It's coming down by interventions around tax largely. Uh, some places have high property tax.

    19. SR

      Yeah.

    20. NK

      Some people have, some places have inheritance tax-

    21. SR

      Inheritance tax. Yes

    22. NK

      ... that we should totally have.

    23. HM

      Wealth tax.

    24. SR

      Yeah.

    25. NK

      I don't get-

    26. HM

      I'm down with progressive tax regime

    27. NK

      If Sameer earns money in his lifetime-

    28. HM

      Yeah

    29. NK

      ... he should be allowed to do what he wants.

    30. SR

      Mm-hmm.

  14. 1:17:11 – 1:21:31

    The zero-MDR debate: Funding the payments ecosystem

    1. SR

      ... 0 Yeah

    2. HM

      No, and I'll tell you what the underlying cost is. Again, going back to societal investment, right? Forget whether MDR zero was required or not. The minute you take money out of that sector, 100 PayPaps have come and gone in the last six years. Some of the biggest names in India, the Tatas-

    3. SP

      The Ambani's, everyone enters, everyone looks at the space and says, "There's money here, they leave in two years." There's always a new kid on the block. Depending on your funding, you'll either go on IPL or you'll go digital on YouTube, waste a lot of money and realize- Paisa toh ban nahi raha. Paisa nahi ban raha. Unless you're invested in solving payments for payment's sake for the long run, with no clear monetization, you disincentivize the industry, the very same private sector that makes UPI happen, you disincentivize everyone actively.

    4. SR

      But I think that's why, that's why the con- the, the whole question about is it time to bring MDR back-

    5. SP

      Correct

    6. SR

      ... has happened, right? Like-

    7. SP

      I agree. No, I'm, I'm very grateful that the government is looking at bringing it back. Don't get me wrong, I'm very appreciative of their effort, and it's not politically easy, as we are saying.

    8. SR

      Yeah, yeah.

    9. SP

      My, my only submission is, I think that the introduction of zero MDR, like many other policies over the last, this thing so many years, not, not in this government, across India's history, requires more research and trust of the private sector to give them fair feedback. If you want to grow the market, ask the market participants what the challenges are.

    10. SR

      So I'll tell you what's important for us right now, Sameer, when, as and when MDR gets introduced. I think the big thing for us is, yes, in a way, yourselves, acquirers, issuers, everybody will hopefully have some way to-

    11. SP

      We are large acquirers, yeah

    12. SR

      ... you know, recoup what you guys have been investing in. But the key thing is, are you going to then invest it? And who-

    13. SP

      Let's say if you are right and he starts profit optimization, somebody will come and disrupt him.

    14. SR

      I'm saying the question that, uh, it's amongst the things to say, "Okay, fine," like now if the ecosystem has some way to recoup some of the investment-

    15. SP

      But the only one holding it right now-

    16. SR

      Where does, how does that double down your investment?

    17. SP

      MDR has not held me back. I don't care about these things because, again, like NPCI, we are very, very clear. You come into our buildings, every kid will tell you, we will go to a billion. We just launched for feature phones, you know this. So not everything has to be about money, and it isn't.

    18. SR

      Yeah, that's, that's the-

    19. SP

      But money helps.

    20. SR

      That's the key-

    21. SP

      This is, this is not critiquing you.

    22. SR

      Yeah.

  15. 1:21:31 – 1:33:08

    Should UPI apps have market-share caps?

    1. SP

      But, you know, there was this author, Ayn Rand. I know. She said monopolies and duopolies are generally not created in a free market, capitalistic societies. If you go from thirty percent to a hundred percent, where's the equilibrium? I'll tell you the equilibrium. Hmm. It goes to barriers to entry. Hmm. Including capital as a barrier to entry. Hmm. Let me rephrase the question. Sure. Indigo now has majority market share in airlines. Yes. If they continue to keep growing- Yes ... ninety percent, ninety-five percent, at that scale, it's very hard for anybody to come in and compete with them. Perfect. The problem with this is if tomorrow something goes wrong with Indigo and there are no airlines in the country- Yeah ... then what happens? So perfect- Sorry, I, I, I, I, sorry. It's even, it's even worse than what happens in Indigo. Indigo has economies of scale. At a certain scale, their cost is lower, so they can operate far more better. In internet world, it's not just economies of scale, there's a network effect, which becomes even harder and harder to match. Hmm. So the network effect doesn't play out in Indigo because I can still book a different airline. I don't lose because Indigo is the largest airline. Actually, the switching cost is about three seconds. But in, but in consumer, you lose because I am not f- participating in the same ecosystem that my friends are on. You're getting this completely wrong. No, but it is, it is the case today also. There is-- If that were true, PhonePe's users would only send money to PhonePe users, and GPay's users would send money only to GPay users. But if I, if I send Pay, PhonePe user sending money to a PhonePe user, my transaction commission rates are higher. No, they're not. It is, it, it is. No. I have seen that. NPCI makes all transactions route through them. Yes. Yes? Yeah.

    2. SR

      They try.

    3. SP

      Once again, no, but it is. So that's one. But more importantly, taking the Indus, uh, Indigo analogy, with Indigo, the sectors in which they apply, if they're the only airline flying, let's say from Bombay to Indore. Hmm. If they decide to shut down, that sector shut down. Hmm. If they decide to jack up the prices by three X, it becomes unaffordable. In a regulated payments market, we are neither controlling the participants, that is exclusively NPCI, all banks and all TPAPs, nor are we controlling the end routing, nor do we create a barrier to entry, and we definitely have no say on pricing. So where does choice end, and where does, uh, this thing, monopolistic practice start? Let's define that boundary, because by that measure or by that argument, every third person should be asked to move off WhatsApp. There's network effect. Every third person should use Telegram, every fourth person should use Signal. That is actually defying gravity in a perfectly competitive market.

    4. SR

      No, but that's the point Harshil is making.

    5. SP

      But-

    6. SR

      Harshil is making exactly that point, Sameer. He's saying it starts to-

    7. SP

      In a perfectly competitive market, people should be allowed to choose. Agreed. But, uh, s- but the digital payment structure, we take it as that it is critical for the country to continue to operate in a way that- Yeah ... there, there's no single entity they should decide. I'm agreeing with you, maybe thirty percent is an right place. You can't, you can't- I'm not saying it is at thirty percent- You can't just... No, no ... but you will have to stop at certain scale. I'm sorry. Their monopoly on operating systems was transferring into other verticals. Yeah. And, and the anti-- They never said stop distributing Windows. They said stop distributing Internet Explorer. They, yes. So in my case, if you said you became the biggest lender because you will not pay me, you can take that argument. So, so is it okay, is it okay if you can continue to be, continue to grow, but you can't cross-sell other things? Why? You're applying that same logic, right? No. You, you have a certain advantage of cross-selling insurance and lending that other players will not have. And that's my point. That's actually a great example. But you can't cross-sell in that. That's ex- a great example. I've been in insurance for six years. This assumption that because you have a large platform, you suddenly become large there. My insurance revenue is probably a twentieth, maybe less than a PolicyBazaar, whose distribution base theoretically is one fiftieth of mine. That ar- ACCOA is bigger. Every single guy in the country- Yeah, but that can change these things. The, you'll, you'll wait for that- No, but it's not- You will, you'll wait for that ecosystem to change in that direction. No, because a laptop or a desktop would ship with one OS, you can't change the OS. Payment apps can be transferred from one app to the other overnight, and that's why I'm saying again, a Navi picks up a share, a Cred picks up a share. There has been fragmentation, twenty percent- Still, still small

    8. HM

      So still small is not my fault or anybody's fault. It's somebody's choice

    9. SP

      I'm not, I was saying 80% is the right cap, but I'm saying there is definitely a cap.

    10. HM

      Why is there a cap?

    11. SP

      There's a cap. You can't have-

    12. HM

      Why?

    13. SP

      ... 90% of the market with-

    14. HM

      Why does it apply to any market?

    15. SP

      It does apply to every market. I'm not saying it won't apply to us.

    16. HM

      No.

    17. SP

      It will apply to any market.

    18. HM

      It should not apply, because-

    19. SP

      Every market

    20. HM

      ... that means that the merchant coming in 20 years later cannot use Razorpay as the best PG if you are. There's no market-

    21. SP

      Stopping, stopping onboarding is not the only way to stop it, but at some point the regulators have to worry about monopolism in any market.

    22. HM

      Worrying, worrying applies to NPCI. They're 80% of the market.

    23. SP

      Yeah, and that's why regulator has a lot of control in the way-

    24. HM

      Pricing

    25. SP

      ... so some way you apply that.

    26. HM

      I have no problem with regulation. I'm talking about market caps. Those are artificial.

    27. SP

      You can't have 70%, 80% market share.

    28. HM

      I can have 80% market share. You can tell me you have to follow these 20 things. Risk mitigation-

    29. SP

      To ensure you, to ensure you create space for others

    30. HM

      ... risk mitigation and market manipulation are not the same. Risk mitigation and market manipula- this is my, my point of fear of big. We'll never become big. We are nothing of nothing. We talk about NPCI and UPI being amazing. Ali and WeChat separately process more transactions than NPCI today.

  16. 1:33:08 – 1:41:18

    Can new banks and startups compete in India?

    1. HM

      The other thing when I think about this is at one point, lots of new people were coming into everything. I felt, or I feel, that it's easier for a regulator to regulate fewer larger p- players than it is to regulate hundreds of small players. That's why we don't have new banks. That's why we don't have so many other things that we should have. I think it's important to reduce the capital and regulatory burden required for new people to join many industries. That has to somehow be encouraged. I don't know how that happens. Like, if a young guy in Jaipur wants to start a bank, it's impossible. But-

    2. SP

      But you know there are already about 1,300 banks in India.

    3. HM

      Yeah, but so-

    4. SP

      For a population of?

    5. HM

      For a population of 1.4 billion.

    6. SP

      Still less banking density by US or Europe.

    7. HM

      Yeah.

    8. SP

      Nothing, nothing.

    9. HM

      I'm saying-

    10. SP

      Much lesser than US, yeah

    11. HM

      ... much lesser. There is a-

    12. SP

      A lot of those 1,300 are defunct also

    13. HM

      ... No, no, no, these ... I, I'm talking about 1,300 which are there.

    14. SP

      There's 35 universal banks.

    15. HM

      Universal banks, no. You're talking about-

    16. SP

      There are 1,200 odd cooperative banks

    17. HM

      ... cooperative banks.

    18. SP

      But those are not allowed to do business.

    19. HM

      Yeah, cooperative banks.

    20. SP

      You have to become a shareholder to do business.

    21. HM

      But, but you need, uh, Harshil, for the-

    22. SP

      Yeah

    23. HM

      ... you tell me, for ... No, no, no. You have to. No, you have to go, you have to go 30, 40, 50 kilometers outside Bombay, no? You have to ... We have to go into the interiors.

    24. SP

      No, no, I agree.

    25. HM

      Waha pe nahi hai, cooperative bank hi hai. We have to get the cooperative bank onto UPI also.

    26. SP

      But see, all the banking issues have come from the cooperative banks only. Even though their pie of the overall business is low-

    27. HM

      Yeah

    28. SP

      ... all the scams are there.

    29. HM

      The scams ... So I'm not ... See, the re-

    30. SP

      Barrier to en- ... So that is a barrier-to-entry industry. That's a, that's a taxicab medallion industry, right?

  17. 1:41:18 – 1:49:39

    How card and merchant payments work

    1. NK

      P2M [laughs] Harshil will do the P2M. Yeah. Okay.

    2. HM

      Yeah. So P2M in India typically starts as a ... Let's take card. It's easier.

    3. NK

      Hmm.

    4. HM

      It starts with a merchant wanting to accept money, because most card transactions are pull. Do you understand pull and push?

    5. NK

      No, explain it.

    6. HM

      So UPI is a push to Harshil.

    7. NK

      I know nothing.

    8. HM

      I initiated a transaction to send money to somebody. That's a push.

    9. NK

      Okay. So Harshil sending Nikhil-

    10. HM

      No, so that a P2P transaction, the way Sameer demonstrated, was a push transaction-

    11. NK

      Hmm

    12. HM

      ... where I said I want to send money to you.

    13. NK

      Yeah.

    14. HM

      A pull transaction is that I want to get money from you.

    15. NK

      Okay.

    16. HM

      So most P2M transactions typically start as pull. In India, even P2P push-

    17. NK

      Razorpay wants money from Nikhil's account.

    18. HM

      Yeah. So a merchant wants some money from Nikhil's account.

    19. NK

      100 rupees.

    20. HM

      Uh, yes. Le- let's start with, let's take an example of Amazon, right? An Amazon wants money from Nikhil's account.

    21. NK

      Okay. Amazon wants 100 rupees from Nikhil.

    22. HM

      Nikhil's account, right. So Amazon-

    23. NK

      Because I have shopped on Amazon.

    24. HM

      Yes. So Amazon sends a request to a payment gateway, right?

    25. NK

      Okay.

    26. HM

      Let's say Razorpay, whatever, right? So a payment gateway. Payment gateway has a partnership with an acquiring bank, and then, um, that is sending a request to an acquiring bank, the bank that the payment gateway has partnered with to do transactions.

    27. NK

      Who is your partner?

    28. HM

      We have multiple, but let's take an example of Axis Bank.

    29. NK

      Okay.

    30. HM

      Axis Bank is a partner bank, right?

  18. 1:49:39 – 1:54:12

    Credit-card rewards and RuPay on UPI

    1. NK

      We looked at data around this, and it says that Gen Z-

    2. SP

      Mm

    3. NK

      ... and people under the age of 30 have two credit cards on average-

    4. SP

      Mm

    5. NK

      ... in urban locations.

    6. SP

      Yeah.

    7. NK

      So even though we are seeing crazy high UPI numbers, everybody still has credit cards.

    8. SP

      Yeah. The premium segment has a certain u- use case for credit card. A, the access to credit, and B, the reward ecosystem that they get access to. Yeah.

    9. HM

      But actually, credit to NPCI

    10. SR

      Majority of those cards under 30 are actually RuPay cards today-

    11. SP

      Oh, RuPay

    12. SR

      ... because they know they can access the whole UPI network-

    13. SP

      And it works on RuPay itself. Yeah

    14. SR

      So RuPay cards, uh, two things. Uh, one is you can actually add that to your app, any, whatever, if you use PhonePe, GPay, whichever one, and then you never really need to take your card out because just you can use it just to scan QR-

    15. NK

      But can't I do that with Visa?

    16. SR

      No. Visa, MasterCard doesn't, uh, cards don't work on UPI QR codes.

    17. NK

      They don't work on UPI.

    18. SR

      RuPay rides on the UPI rail. So you can ride on the UPI network. So today, now you are doing a UPI transaction, but also getting your rewards. You're getting all the other benefits that you get. You're getting your credit.

    19. NK

      Mm-hmm.

    20. SR

      So that's the, uh, advantage. In fact, today you can, uh, you know, the, the best experience I think that you can have is you have a RuPay card, you link it on your app, and then you activate bio auth on it. So today, when you do a UPI transaction, you typically put a PIN, right? Now, with the bio auth, you just scan and click on the bio auth, it sees your face, and the transaction is through. And approval rates, I'm sure you would have seen that bio auth approval rates are-

    21. SP

      Significantly higher

    22. SR

      ... significantly higher. Much smoother transaction.

    23. SP

      Yeah.

    24. NK

      What if I get a banking license?

    25. SR

      Mm-hmm.

    26. NK

      I'm not SBI, I'm Bank XYZ.

    27. SR

      Mm-hmm.

    28. NK

      And Bank XYZ tells Harshil-

    29. SR

      Mm-hmm

    30. NK

      ... I won't charge the one point four five percent, I'll do it for free because I want to acquire a client.

  19. 1:54:12 – 2:02:08

    How Razorpay improved online payments

    1. NK

      Right. And before you, there was what, Billdesk? What were the other-

    2. SP

      They're still there.

    3. NK

      Oh.

    4. SP

      They're still there. The Billdesk, uh, like-

    5. NK

      PayU.

    6. SR

      PayU

    7. NK

      Even they're going to IPO, no?

    8. SP

      Billdesk has not gone for IPO.

    9. NK

      No?

    10. SR

      Yeah.

    11. NK

      Okay. What, what did you do differently that you were able to get more market share?

    12. SP

      Yeah, so the foc- so I think in the online pay... I'll just explain the online payments world. India took a, let's go slightly macro. US is one of the highest success rates in the world, but has highest fraud rates because US doesn't have two-factor authentication.

    13. NK

      Right.

    14. SP

      So our fraud rates are quite high, on an average about 100 to 200 basis points, but a very high success rate, your transactions typically close.

    15. NK

      Mm-hmm.

    16. SP

      And the acquirers have to focus a lot on just controlling the fraud side.

    17. NK

      Mm-hmm.

    18. SP

      In India, we removed consumer-side fraud by adding two-factor authentication. So customer-side fraud reduced drastically, where I can't take your card number and do anything unless you give me an OTP. Now, there's still social engineering and all that happens, but at scale fraud that happens in US, where I have millions of cards and I can debit millions of cards together-

    19. NK

      Mm-hmm

    20. SP

      ... it can't happen in India. You have to still have to do social engineering, it takes effort. What happened as outcome of that is because of us at introducing this friction called two-factor, we reduced the fraud rate significantly, but we dropped the success rate significantly. So our failure rates went up. The data that I got Razorpay brought in is how can I work on d- in solving a lot of problems in this chain to reduce the fraud rate signi- to reduce the failure rate significantly? The more failure rate I can reduce, the better downstream impact to the merchant, more business online merchant is doing. And in online merchant, it matters a lot more because in offline, you walk into a store, your card fails, you pay through another card, you pay through UPI. The guy will not typically let you walk out without closing a sale. In online, the sale typically drops or the customer moves to COD. So the impact on, there is directly impact on either top line or bottom line because my payment failed.

    21. NK

      Mm-hmm.

    22. SP

      Particularly in these new age customers that we target. Now, in the government, you'll typically still not lose a sale. If I have to pay my electricity bill, I'll pay if it fails five times. But if I'm buying on Amazon or I'm buying on Zomato, transaction fails twice, I'll go to Swiggy, I'll order COD. I'm not gonna wait and try three times. So that is what we really focus on, how can I improve success rate significantly with a lot of interventions we did across this value chain to improve success rate. And that is my core. The second big area was how can I allow small merchants to onboard? Most large acquirers that, like, at least when we started, the problem that I faced that when I wanted to start a small business on the side, getting approval, getting through a payment gateway itself was a fairly painful and complicated, at least a month-long process. How can I bring it down while ensuring that the fraud doesn't happen on the merchant side? And that required a lot of work to ensure that I can enable a small merchant who has no history and still ensure that I'm on the hook if he runs away.

    23. NK

      Right.

    24. SP

      That was a fairly hard problem to solve for, that I can onboard them without-

    25. NK

      Can I ask you, what does Paytm do in all this? Where does it sit?

    26. SP

      Paytm sits on the consumer side. It also sits on the merchant side. They have both. They also have offline, uh, and-

    27. NK

      Wouldn't it make sense for you to sit on the consumer side and for you to sit on this side, like together? Like if you both-

    28. SP

      And we all sit on every side.

    29. NK

      Yeah, yeah.

    30. SP

      They have-

  20. 2:02:08 – 2:05:36

    The B2B payments opportunity

    1. NK

      Can I also ask you, like we spoke about Paytm and Other Billdesk, other people in India who are in this ecosystem. Globally, who's doing what which is interesting, that there might be an opportunity around? You can talk about WIX, you can talk about SEPA, you can talk about FedNow, RTP.

    2. HM

      Who is-

    3. NK

      What is China d- doing differently? Just the payment ecosystems.

    4. HM

      I think, I think-

    5. SP

      In India or in China?

    6. HM

      ... the single biggest-

    7. SP

      In India, massive.

    8. SR

      India

    9. SP

      B2B is a massive problem, like completely uncharted.

    10. HM

      B2B is where most of this money is still in cash.

    11. SP

      And most of B2B is still not formal.

    12. HM

      Supply chains are just not solved.

    13. NK

      So what would I build? I'm starting a business in B2B.

    14. HM

      Yeah.

    15. NK

      What do I build?

    16. SP

      B2B, the complex part is-

    17. HM

      You build an app for small merchants to be able to, to onboard their, their everyday suppliers. An average kirana shop, or not kirana shop, but a small-sized retail store, has 70 to 80 distributors that give them stuff every week. Those people need to be paid out, and those payments are anywhere between 1,500 rupees to about 50,000 rupees.

    18. NK

      Why can't he just do UPI?

    19. HM

      It doesn't work. That ch- that channel is not used to accepting it that way, and there's other needs, by the way. It goes back to needs. Like he's building for the online segment. In the offline world, they need to onboard these people. They need to have their payment identities linked. They need to have, in some cases, recurring payments. In some cases, you have Khata-

    20. SP

      I'll add, I'll add to it. The payment is a small part of the problem. The reason it doesn't happen is there's a workflow around B2B payments.

    21. HM

      Yes.

    22. SP

      So typically, when I'm making a payment to you in a B2B environment, I need to first get an invoice. I need to verify it. I need to reconcile it. I need to then check how much is due. I need to cons- consider that is, am I deducting the right amount of TDS? Is the right amount of GST being paid? I need to later do a GST filing against it, so it needs to be completely matching. There's a lot of workflow around a payment, and that takes a lot of breaks-

    23. HM

      But that's the opportunity

    24. SP

      ... makes it a lot more inefficient. So it's not as simple as-

    25. HM

      That's what 40 million people don't have

    26. SP

      Yeah, so it's not as simple as picking up the phone and paying money to you. All of that ecosystem around it, and generally it's not, like, in a single-person store, it might still be the same person doing it. In a larger B2B organization, even a 15-people organization, it might be two or three people doing different parts of this.

    27. NK

      Can you simplify for me one level more? What do I exactly build?

    28. SP

      A way to build this en- solve this entire workflow. Let me give an example of how a large company does it, like, say, a Reliance or a Tata. They are able to do it efficiently. They're working at far more scale. They all use typically an SAP or something to bring all of it together, so that in one workflow I can do this transition and the data can be shared, reconciled, automated out. Now, they're extremely expensive products that a small guy cannot. So if I were to look at that and see how can I, how can I make a small guy, guy's life as simple as it is for the large conglomerate-

    29. NK

      Mm-hmm

    30. SP

      ... at far lower cost, I think that's a solvable problem.

  21. 2:05:36 – 2:09:54

    Building payments for medical tourism

    1. SR

      finance app."

    2. NK

      Sohini, what would you start?

    3. SR

      Yeah.

    4. NK

      Now you've been on the regulator side. You've been a, not regulator, you've been in a Section 8 company.

    5. SR

      Yeah, yeah.

    6. NK

      If you had to go back to the private space-

    7. SR

      Mm

    8. NK

      ... and build from everything you've learned-

    9. SR

      Yeah

    10. NK

      ... what would you build?

    11. SR

      So, um-

    12. NK

      You remember you have to compete with these guys.

    13. SR

      Yeah, I know. Um, so I'd, I'd go back to... Uh, so this was something that I had thought I would build actually, um, uh, in, while I was still doing my cross-border money transfer role. Again, this about creating an ecosystem for ... India has a huge medical tourism, right? But-

    14. NK

      Inbound.

    15. SR

      Inbound. Inbound medical tourism. Um, and again, the payment space around it, and I, I figured that out with the, in fact, some of the cases where a lot of transactions were coming in and they were getting declined because our fraud and risk models were saying, "Why are these people coming in and taking so much money?" And, and turned out these were actually people who were living in and around hospitals in India. Many of them were coming actually from Middle East, and there was no other way. They don't have a bank account here. There's no way for them to... And they were just, they thought something would last for three months, but it lasted six months, and now they don't have a way to actually get money. Um, and that entire ecosystem, whether it is the wellness or the absolute, the cure, I would build, I, I, I had one time actually thought that I should build a more elegant solution which can be connecting these real-time payment systems for a more seamless cross-border money transfer specific to the use cases so that that risk part can be taken care of.

    16. NK

      What would you build that on top of?

    17. SR

      Uh, as in the payment system layer? Uh-

    18. NK

      Stablecoins have no off-ramp within India

    19. SR

      ... it can... No, this is not, it, you can't go with a stablecoin.

    20. NK

      And that won't happen. I'm not for it.

    21. SR

      Yeah, yeah. It wouldn't be a stablecoin. Uh, it will have to still at the... So you have to build a payment rail, and on both sides, there's got to be the, the s- the, you know, the, what do you call it? Where the money sits will have to be banks, uh-

    22. NK

      But they're coming from everywhere. They're not coming from one place, right?

    23. SR

      Yes. Yes. Which is why, which is why one would really need to build a, a very nice well-thought-out product to be able to think about what are the different kind of rules. And it's not everywhere. Like, if you look at India receives, I recall I'd done, received remittances from 198 different countries/territories. Medical tourism we are not getting from 198, so it'll still be a much smaller number. So it is possible.

    24. NK

      You are saying Middle East largely?

    25. SR

      Largely Middle East. There is, uh, there's-

    26. SP

      Individual markets you can still solve for.

    27. SR

      Yeah.

    28. SP

      It might be hard to solve for 109 markets-

    29. SR

      Yeah

    30. SP

      ... but two or three you ... And that's how you start. Like you pick up-

  22. 2:09:54 – 2:16:01

    Why digital tipping is still unsolved

    1. HM

      You were asking about apps.

    2. NK

      Yeah.

    3. HM

      One thing that I think is ripe for the picking, and some kid will solve it, I don't think any of us have the imagination, um, tipping. Just being able to tip in the UPI era is a major problem-

    4. SR

      Mm

    5. HM

      ... because right now, whether it is mobile-to-mobile or it is a virtual private address, you feel awkward asking, and the other party feels awkward ask- I mean, you feel awkward asking people for their payment identity, and they feel awkward asking, and on the other hand, it's just the normal friction. Earlier, if you'd settle cash, let's say your, uh, restaurant bill was 930 bucks, you would leave the 70 on the table, and that's the tip. Or if you had a driver.

    6. NK

      Yeah.

    7. HM

      So now we end up asking people-

    8. NK

      Giving

    9. HM

      ... "Bhaiya, number batao"

    10. NK

      Oh.

    11. HM

      And I think it's, it's a pretty weird sort of, uh, way of doing it. A lot of times people just don't do it because of the headache of it. Uh, and I think people are quite comfortable tipping. I'd, I'd just gone down to the Bay a couple of weeks ago, and there it's so different because in the ... When they give you the POS device, the default is 18% and 20, 22%.

    12. SR

      Right.

    13. NK

      Yeah, it has to be a add-on.

    14. SR

      That's why you end up saying-

    15. NK

      So that's the US.

    16. HM

      No, but I'm saying that it's fitted into the POS here.

    17. NK

      US, if you don't give tip, they'll ask you what's wrong with you.

    18. HM

      Yeah, they get mad at you.

    19. SR

      Yeah.

    20. HM

      What works, waha pe woh aap se ... As he said, it's a, it's a push. So they're saying, "I want this," but you can add a tip to it. In UPI, there's no add-on. So the problem is once the payment request has been created, you can't add on a tip, and that's a fundamental difference in architecture, but it's a very limiting aspect as far as tipping goes. Tipping on-

    21. NK

      So you build it on top of a existing UPI app as a service?

    22. HM

      So in cards, in cards, when the merchant actually raises the amount, you can add a tip, so it's added on, and then you process it.

    23. NK

      Right.

    24. HM

      It's a, it's a two-part processing. I tried that with my card here because of 2FA. The first part goes through, the second gets rejected-

    25. NK

      It can't happen

    26. HM

      ... because it's a separate transaction.

    27. SR

      Yeah, you got to do a separate-

    28. NK

      Yeah, in Swiss you can try it because there's no two-factor. In India, you can't do that.

    29. HM

      So here it gets rejected.

    30. NK

      So it has to be done before the PIN is entered. Uh, so the before the PIN is entered, you have to upcharge.

  23. 2:16:01 – 2:21:22

    What payment data reveals about India's consumers

    1. HM

      is large.

    2. NK

      A lot of the people watching this-

    3. HM

      Mm-hmm

    4. NK

      ... are also students of economics-

    5. SR

      Mm

    6. NK

      ... and trying to figure out what's happening with the economy. You have data on consumers-

    7. SR

      Mm

    8. NK

      ... in a way that nobody has. Share something, no?

    9. HM

      I mean, okay. So I think, I think two, two things that I've seen over the last decade that have just still continue, uh, continue to blow my mind. I think living in the metros all our lives, we assume that the money of India is in the metros. Everything I've seen on the UPI data, and correct me if I'm wrong, breaks that myth, shatters that myth. If you see money movement, at least the digital part of it, um, tier two and beyond has been contributing 65, 70% forever.

    10. NK

      Of consumer behavior?

    11. HM

      Consumer payments. And the crazy part in there is both the user adoption as well as the money movement, and they're not just on the receiving side. Because [laughs] Because here when I talk to people, right, naturally I say, "Money must be going from Bombay to the village." No, money is going from the village to the village, and to many different places. So, so the, um, money in India actually runs deep, and we are seeing ... We didn't see that in e-commerce because e-commerce was lagging digital payments. We didn't see that in ticketing. So can I tell you that in, in online, it is not that deep. Yeah. And it's very shallow. No, so online is- And online, it's still not penetrated as much. So that's a ... So I think- But there's a five-year lag because online will follow digital payments adoption. Yeah, and online, we have not solved for a lot of problems. Like UPI has still replaced cash in day-to-day life. Online e-commerce hasn't replaced people's day-to-day shopping downstream beyond metros. And it hasn't-

    12. NK

      Is, is the pace of adoption high?

    13. HM

      It is growing, but it's still very tiny. Like, like I can put 15-

    14. NK

      Like, can you put a percentage to it?

    15. HM

      I would if ... Uh, 70% of all online e-commerce purchases are done by 15 million users, one five, 50 million households.

    16. NK

      Five zero or one five?

    17. HM

      One five.

    18. SR

      One five.

    19. NK

      That's nothing.

    20. HM

      It's that narrow.

    21. SR

      Nothing.

    22. HM

      But 70% online spends, e-commerce spends are there, and that's true for most players.

    23. NK

      That's like Netflix audience.

    24. SR

      Yeah.

    25. HM

      70% by value or- By value, by value.

    26. NK

      By value.

    27. HM

      By dollar value. By number of transactions, it's a lot, but it still spreads out. But by dollar value, it's very tiny. But it's not the-

    28. SR

      That's not important

    29. HM

      ... picture of economy that so it's-

    30. NK

      Like whenever I see economics, it says that, say, in the US, whatever their GDP is, 30 trillion, whatever number they claim, they say that a majority of that consumption, 60, 70%, is QSR, like quick service restaurants.

  24. 2:21:22 – 2:26:07

    Trust and India's next wave of online shoppers

    1. HM

      Order four-

    2. SP

      I'll say one more reason the e-commerce penetration being narrow is, like, also the behavior that we have to build for penetrating downstream. We haven't built that on apps, right?

    3. NK

      Right.

    4. SP

      So for example, in, like, the e-commerce apps are built more like supermarkets, right? Where everything is available, all information as well. I have to pick and decide myself.

    5. NK

      Mm.

    6. SP

      So downstream in India, that behavior doesn't exist. Like, supermarkets haven't really penetrated. So, like, almost 40, 50% of travel in India is purchased through, uh, these travel agents. Now, when you go to these travel agents, what do they do? They open MakeMyTrip only and buy. But I can do it myself, but people don't want to do it because I'm confused. I'm overwhelmed with the interface that all information is there, but I don't know. First of all, it's not in my language. Second, I don't know if it is solving exactly what I want. I might mischoose. So that there's still a lack of trust because it's not replicated the behavior I'm used to.

    7. NK

      Mm.

    8. SP

      So I go to the travel agent, I ask him 20 questions. "I want this flight to this place. I want this time. I want this time. Are you sure this is right?" "Yes, this is right." I want the comfort that this g- person gives me. I don't trust myself to make the purchase on MakeMyTrip. But insurance we spoke about, like the majority, and almost 90% plus of insurance purchased through brokers in India.

    9. HM

      No, but-

    10. SP

      Offline brokers.

    11. NK

      Mutual funds also

    12. SP

      Because, because of, because of this, I want to talk to somebody who give ... And not all of the time that somebody's fully incentivized to help me. But I still trust that guy more than an open platform where I have to pick and choose myself. I might get confused. I might not ans- I might mistakenly choose something else, and I might lose my money. I'm not trusting that interface.

    13. NK

      Also, if I want to change or cancel tomorrow, I know this guy I can catch to give.

    14. SP

      I can catch somebody. I have the trust with him. I don't know his-

    15. HM

      Commodity categories like motor insurance, for example-

    16. SP

      Mm

    17. HM

      ... which is also mandatory, has moved largely online.

    18. SP

      Billing has moved. Like, like-

    19. HM

      Right

    20. SP

      ... people paying their government electricity bills have moved because there we-

    21. HM

      So things where you know you have to do it and there's a price

    22. SP

      ... and it is high frequency.

    23. HM

      Yeah.

    24. SP

      I've built that trust. I... There's repeatability. But I'm booking a flight ticket maybe once a year.

    25. HM

      And discretionary-

    26. SP

      So I don't want to m- make a mistake in it.

    27. NK

      No, in tier two, tier three towns, people are still-

    28. HM

      Mm

    29. NK

      ... wanting to go and get the whole assisted, uh-

    30. SP

      But see, if you were in a massive delta-

  25. 2:26:07 – 2:31:38

    GST, consumption and property taxes

    1. SP

      Online spends. Online spends.

    2. HM

      Online Hindi It was quite awesome. So you had to isolate out the impact of Diwali and Dhanteras and all of those things which you normally also do for your assessment. Um, direct correlation, GST rate goes up, activity, retail activity went up. I mean, uh, goes down, retail activity goes up. And if GST went down by two percent, you were seeing a five and a half, six percent increase in consumption. GST went down in some categories by five percent. There was a 13% increase in-

    3. NK

      Wow

    4. HM

      ... uh, buying. GST went up in one category, and I forget which one, by 4 or 5%, something became sin good or whatever. That one came down by about 6%. And where GST was flat for that entire shopping season, retail behavior-wise, it was down minus 0.2%. So the control set was actually that.

    5. NK

      Oh.

    6. HM

      All categories where GST was unchanged, marginal year-over-year change. But wherever the GST changed, it is that, like, directly attributable category by category. So we ran through all the categories.

    7. NK

      So in a way, you're saying if GST was not there-

    8. HM

      Yeah

    9. NK

      ... consumption would pay more taxes than having GST?

    10. HM

      Oh, the, I think they know that by now. And they know-

    11. NK

      But then why do they have GST?

    12. HM

      Hmm?

    13. NK

      Then why have GST?

    14. HM

      Because you need taxes.

    15. NK

      No, but they would get the tax other ways, no?

    16. HM

      How otherwise?

    17. NK

      Like if consump-consumption goes up-

    18. SP

      No, but how would you make it?

    19. HM

      It's easier for India to at least-

    20. SP

      Majority of taxation revenue is indirect taxes. Majority of revenue are from indirect taxes

    21. HM

      It's actually easier to charge GST in India or TDS in India than to charge income tax.

    22. SP

      Income tax is very low penetration, so majority is indirect.

    23. NK

      3%, right?

    24. SP

      Yeah. In fact, the, there's the other side argument to be made, why do you charge income tax at all?

    25. NK

      Yeah.

    26. HM

      Remove it.

    27. SP

      Move everything to indirect.

    28. HM

      It's better off doing that.

    29. SP

      It's better off doing that.

    30. NK

      But then the consumer-

Episode duration: 4:00:05

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode hBvInWQWpZ0

Get more out of YouTube videos.

Read along on your phone, search it beside the video, or ask your favorite AI app what was said.

iOSAndroidClaudeChromeChatGPTComing soon