On Ep. 1 of Brainfood, Simon Taylor, Head of Market Development @ Tempo is joined by David Vélez, Founder & Global CEO @ Nubank to discuss Nubank's US launch, strategy, Hispanic customers, AI banking agents and more!
On Ep. 1 of Brainfood, Simon Taylor, Head of Market Development @ Tempo is joined by David Vélez, Founder & Global CEO @ Nubank to discuss Nubank's US launch, strategy, Hispanic customers, AI banking agents and more!
Timestamps:
Sy Taylor 0:11
Welcome to Brain Food. My guest today is David Belez, who is the founder and global CEO of the one and only Newbank. Newbank has true to its name, developed a blueprint for what a bank can be in the 21st century. It's mobile-first, wildly efficient compared with its peers, and staggeringly large. It's building on its success in Latin America. It now has more than 140 million customers, more than Chase or any other domestic U.S. banking franchise with quarterly revenue of 5.9 billion. It is still a fraction of J.P. Morgan's 58 billion, but two things stand out: growth and efficiency. News revenue grew 39% year over year on an FX neutral basis. J.P. Morgan's grew 27% or 15% if you exclude some temporary stuff. New bank's efficiency ratio is like LeBron's numbers, all-time greatest. The overhead required to generate revenue was 19.5% versus 47 at J.P. Morgan. That's the kind of number that other CEOs read and weep. And of course, they do that by looking at customers that other big banks ignored first in Brazil, then Mexico, Colombia, and now in the United States, and with New Global possibly around the world. But what I find exciting and interesting is that New Bank is leaning into tokenization with stablecoins and leaning into AI with its new former foundation model. All of the companies that are going to take the growth from the next decade are not using just the fintech playbook of cloud and mobile. They're using AI and tokenization today. So I'm speaking today, the day after News American launch, and we're going to have a conversation about what is new becoming and where the keys to success lay. Please welcome my guest, David Belles. David, good to see you, man. I read somewhere that you're a big fan of Chuck Feeney. Do you want to explain who that is, and like, how's that shaped what Newbank
David Vélez 2:14
is? Wow, great question. So, I was an associate at General Atlantic, a private equity fund here in New York a while ago, I was giving this book on Chuck Finney's life, which is named "The Billionaire Who Wasn't, and I read it, and it was one of the most fascinating biographies I've ever read, and one of the coolest lives. Like I said, wow! I want to live his life. That's a really cool life because, and it was super inspiring. Basically, Chuck was an entrepreneur, the the founder of DFS Shoppers, all those stores that you see around airports around the world, and he sold it in the '80s to I think LVMH. He made a lot of money with that, but then he decided that he was going to give all that money away through the rest of his life, and that his goal was to let the last check bounce. Literally, he wanted to die with zero, but the way that the book talks about it is how he went around the world solving all these problems on an anonymous basis, and he participated in the the peace process in Ireland and built universities in the Philippines and worked in healthcare in Indonesia. So it's like it was it was an amazing read, and to me, since then that has also become kind of my goal at some point, and Andrew Carnegie has has also another similar quote, which is, "The man that dies rich dies disgraced. Wow! And he's a little brutal, but if you think about it, if you think about it, with so many opportunities and problems around the world, dying with a bunch of money in the bank-what a missed opportunity, right? I just find it crazy. So, so from that perspective, is obviously anybody can spend the money, but it's like, how can you spend it in a way that has impact and that brings people brings things a little bit better? Yeah,
Sy Taylor 3:51
because you're kind of like when it comes to impact, you're kind of operational efficiency obsessed as a culture. Like you've just launched in the U.S. and I saw this crazy stat about there was like a basis point of the company's operational efficiency that was the budget for launch in the U.S. So, like, how do you think about launching in new countries and being efficient with that?
David Vélez 4:13
Smart efficiency is one of our five values as a company. We have we're very much about culture and the thing about efficiency and us being efficient about what we do is critical to sustain our first value, which is we want customer to load fanatically, and we didn't invent this. I think Amazon sort of exemplified very well, which is you need to grow to be more efficient, so that you can pass the efficiency to the end customer of your better and better products, and then create this flywheel just accelerates with scale, right? You you grow to charge less. You don't grow to charge more. And so for us, that has always been a very important thing to do. And and yeah, I remember there are some stories on New Bank. I remember in the early days that I would get mad at people that would be taking the elevator for one floor because we were wasting energy. They could just. Simply take the stairs one floor. It's like, why are you doing that? And the
Sy Taylor 5:02
step count. Think about the step count. Exactly. You're
David Vélez 5:04
doing you're doing you're doing exercise. Why are you taking the the elevator one floor? Or or people printing in ink? I would get so mad when see people. Why are you printing? Why are you printing in ink? It's like 10x more expensive than printing in in black and white. So efficiency has always been a big part of our business model. And today I think you see that beauty and the benefit of that scale. Because, for example, just to give you a couple data points, we have 140 million customers in Latin America with 10,000 employees. Our competitors have 100 million customers with 100,000 employees and 6,000 branches. Right? We have zero branches. You won't find we're the largest private bank in Latin America, probably on Bank Latin America. You won't find a single coin in any of our offices, or or a vault, right? It's all digital, and so you can
Sy Taylor 5:45
work with cash, right? Like cash is a pretty important thing.
David Vélez 5:48
Not not really, not anymore. There's no cash anymore. We have no cash. You find no cash. These these systems, these countries have digitalized incredibly quickly, right? If you look at Brazil today, it's digital. Cash is disappearing with pigs. You obviously have debit cards and credit cards, but people are cash is disappearing, and so that just gives us this cost structure advantage that we can offer to back to the consumer via much much better products. And and you mentioned U.S. launch. As we announced the U.S. launch, we think that's that might be the biggest opportunity because while there is access to financial services in the U.S. there is there is a lot. People have access to bank accounts. People have access to credit products. Although about 100 million of Americans have restricted access, you would you wouldn't call them unbanked. You would call them underbanked. The cost of being a bank consumer in the U.S. is insanely high, and people don't realize it. It happens in ways that it's not apparent. The majority of U.S. consumers.
Sy Taylor 6:40
Paul Graham has this great term "schlep blindness, which is this idea that that you're doing these little schlepps every day that you just think is the way I have to do things. Yes, and it's not until somebody comes along and solves the schleps that you realize that that was the better product that was waiting there. Although I don't imagine there's a schlep quite as bad as the story you told in your launch yesterday, which is where you have to wait in some like prison before you can get into the branch in in Brazil when you were opening an account. That's a that's a funny story. But talk to me about some of the schleps in the U.S. that people don't realize exists.
David Vélez 7:10
Yeah, I mean it's it's interesting. So I'm not Brazilian. I'm Colombian, and it took a foreigner to come to Brazil and to see the system with new eyes to realize something's wrong here. Right. And and and when I would talk to the locals, my friends in Brazil, they would say, "What do you mean? Bank is fine. This was 2013. Like there is a lot of banking branches. Or or they would say, "Yeah, it's it's crappy. It could be better, but what is the alternative? There's nothing else. There's nowhere to run. And and it was even hard for them to clearly identify the issue. In some ways, the U.S. is today the same way, in specifically around cost, and specifically, what I think is interesting is if you look at the balance sheets of the big banks in the U.S. And by the way, this is the same thing in Europe or in Latin America, really across the world. Banks by now are making, are having 60 to 70% of their $15 trillion deposits, 60 to 70% where they're paying zero to consumers, zero. So this is $10 trillion, but they pay zero, and the average yield is 0.42% per year. In an environment where your interest rates are four and a half to 5% this is about $600 billion that, instead of going to the consumers' pockets, are staying in the back in the bank's balance sheets to pay for a very expensive bank infrastructure that consumers don't want to pay for, right? If you think about it, it's almost as if you were paying the rent of an apartment that you never use. That's where consumers do it, but it's it's hitting. It's not obvious. It's not doesn't necessarily come in terms of a fee. It is the foregone yield that you should be getting on your money, then in a more competitive environment you would see it. You would see it happening, and this is just one aspect. Then you look at the fees, close to 100 billion dollars in fees of monthly fees, weekly fees, overdraft fees, in remittances. You have banks are still charging five to 10 percent per transaction, and we see with our Latin American customers as they go. So ultimately, I think one of the biggest opportunities somebody with the business with the with the cost structure that we have. To your point about efficiency, if that was an opportunity in Latin America, it's a gigantic opportunity in the U.S. to use the business model on behalf of the consumer and pass much more of that consumer wealth back to the consumer.
Sy Taylor 9:17
And talk to me about wedge and market entry and like your ideal client here, because I imagine the diaspora is the starting point. I heard yesterday, Chris and you talking about your customers already were asking you to launch here, so you probably have a bit of a wedge. But Chime's been here for a while; they're very aggressive on cost, and Cash App is very aggressive on cost, and there is competition in the U.S. for that that kind of consumer base and cost consciousness. So, how do you win?
David Vélez 9:44
I think the first thing to consider is though you add some all of the names that you mentioned plus Nubank plus every other fintech around the world. All of us combined, all of this fintech that you have covered so well for so long, we represent less than 5% of the world's financial services. We are. Tiny. We haven't moved the needle as an industry in in 15 years of fintech existing. So that tells you a couple of things. That tells you first, there is a lot of inertia in banking, and so that will continue to happen. Second, there are ways to think about that inertia to the to to kind of remove for the next as we think about the next 10, 20 years. We think what was happening with open finance and AI. We think these are catalysts for that inertia to to start disappearing. But third, that's the competition. The competition is not a lot of the fintechs that you're mentioning. The competition is the big incumbent banks that are still paying effectively zero. They're charging all these different fees. They're paying for a bunch of banking branches that people are not really willing to use. And so yes, there are players. I think for us specifically, what differentiates us first, I think there is this this sense. There is about 100 million underserved Americans. Within that 100 million, there is a significant percentage of Hispanics. The Hispanic population of the U.S. is over 80 million Hispanics, and is now the large one of the largest demographic groups, and this is a crazy statistic, by the way. Nine out of 10 new Americans being born over the next decade are Hispanic babies. Wow! Effectively, all the demographic growth of the entire country will come from the Latin family. So this is my employer, right? So this is a very important constituent that happens to know Nubank very well. We have 140 million customers in in Latin America. They they know each other. They send money to each other. They talk to each other. So so for sure, this is a niche that is very valuable for us initially. But we think that this goes beyond because ultimately we think we're trying to build the very best mass market product in in the entire country, and so this goes beyond. This will go for the mass market American that wants to get paid to bank and doesn't want to be paying to bank.
Sy Taylor 11:47
I think also you talk about your customers being fanatic as like the first value in the company, and you know from the UK, Monzo is very near and dear to my heart. I think they they stuck the landing on that in the UK. I haven't really seen that in the U.S. market, so like, what is it that makes a customer a fanatic? Like, make that tangible because it it does sound like marketing BS, but you show it in the the cost of acquisition because that's super low for you. But but what does it take to make that?
David Vélez 12:13
If you think about it, it it might sounds a little bit like BS, but if you think in kind of first principles and and try to simplify capitalism. Literally, you simplify capitalism. Capitalism ultimately, if you truly have capitalism, which means you truly have free market and competition in a perfectly competitive environment, whoever offers the best product and service and experience wins. That's it. It means consumers have a choice, and and see if consumers have a choice and they can choose from five or six or 10 banks, they will choose whoever tries them best, gives them the best product, charge them less fees. Ultimately, is the best thing, and so that is consumer obsession. Consumer obsession is obsession to be the product that consumers will choose. And what has happened in banking globally, really, is that it hasn't been a free market, yeah, a lot of the regulation that has existed, especially here since 2008, has been really end up helping the big banks maintain an oligopolistic structure, maintaining a lot of power, and that's what we saw in Latin America, where when we began, there were five banks that own 80% of all the assets, and when you are protected from competition, then you don't really have to try for consumers that much. You almost forget that they exist, and they become a nuance. It's like, oh my God, one consumer again complaining about the subject. So, so that's what it really means. But it's hard to execute though, because I think you need a bunch of different elements. You need an element of culture. Yeah, your culture has to be optimized for that. People have to know that their mandate, number one, number two, number three priority is optimize for net promoter score is you're not optimizing for unit economics. You're not optimizing for revenue or for profit. You're optimizing for net promoter score, which is different. You make different decisions.
Sy Taylor 13:49
Yeah,
David Vélez 13:50
a lot of the time it's inconvenient from a business model perspective.
Sy Taylor 13:53
It's yeah, exactly. It's almost counterintuitive. Yeah, because you're optimizing. You get the results. Because I look at new banks like unit economics as LeBron numbers, you know, like they're all-time greats, and you're putting up these numbers of cost of acquisition, cost to serve, revenue per user, RPAK that I think make other bank CEOs look and weep. But you're doing that by not focusing on those numbers, and I think that's the cultural difference that really stands out to me.
David Vélez 14:18
I think I would say slightly different. I would say we we focus on the advantages that technology can give us to get the efficiency we discussed previously. Right. So once you already begin operating at a significantly lower cost, then you're optimizing for the long-term user experience versus the short-term profitability. Let me give you one very simple example of what this really means, because I think people sometimes example. You bring it to life. Are you a philanthropy, right? You don't want to make any money. Like, how did a couple years ago, one engineer showed up and said, "We're making more money per customer. RPAK engineer economics suddenly increased, and people obviously are happy about that. But let's go find out what happened. And it turns out that we there was a bug in the system. We had removed an email reminding people to pay on time, and we were the only bank in Latin America that were like reminding the customers to pay on time because we want them if they're gonna be late, we wanting to be mindful that they're being late and they're gonna pay an interest. So we reminded like three times, and we forgot to send those emails, and people start getting late. And so for us, because we have this culture of of obsession, we said, well, not only you have to go back and put the email back, send another email to people that overpaid and say and apologize for not reminding them to pay on time, and then give them the money back proactively.
Sy Taylor 15:31
Wow!
David Vélez 15:31
Now you would say, but again, like you don't want to make any money. It's like, no, no, no. We are a business. We want to make money. But what happens in the psychology of that customer that gets that email? Say, oh my God! This is the first time the customer. This bank is is watching for my back. Is giving me the money that that I shouldn't have paid. We have just make an investment in the next 10 to 20 years of loyalty with that consumer.
Sy Taylor 15:52
Yeah.
David Vélez 15:52
So in the long run, it's NPV positive.
Sy Taylor 15:54
Yeah.
David Vélez 15:55
In the short term, we're making less money, but in the long term, it's totally aligned with optimizing the value of the business.
Sy Taylor 16:00
I think that moving from a bad landlord to a good waiter, you know, the good waiter just shows up and fills your glass, and just things just happen. Whereas the bad landlord charges you a fee every time you leave the house, sort of thing. And it's like there's a there's a penalty for being there's
David Vélez 16:13
a cost from a consumer perspective. I was like, you you'll remember that, right? You're like ah, and you get to a point where you say, I want to get out.
Sy Taylor 16:20
Yeah, or I sort of begrudgingly accept it because I don't know any better, and then I taste something better, and it's like, oh, this is this this can be a lot better.
David Vélez 16:28
Exactly.
Sy Taylor 16:28
Talk to me a little bit, just changing tack for a second about New Global, because I was trying to watch this thing on my phone from 30,000 feet via Starlink. Incredible presentation, but I was sort of expecting that you were going to do U.S. launch and that that was going to happen, but I wasn't expecting the new Global product. So tell me what that is.
David Vélez 16:48
Sure. So New Global is a stablecoin-backed account, savings account for customers in over 35 different countries around the world, a number of European and Latin American countries, and the idea here is to enable customers from all of these different countries to start saving with high yield, start being able to spend that money. You can create a virtual debit card that you add to your wallet, and you can start sending money to any other Nuban customer or any of the 140 million customers, when I order new customers for free, real time, and how do we get to that view? Partly is us continue to execute the thesis we've been executing for 13 years, which is we think we continue to think that the future of financial services globally will be of consumer obsessed digitally native companies. So as we think about how do we go for from three countries to 100, new global is a way to to go broader, faster in a thinner layer. Yes, that's one way. But the second aspect of it is, we've been big fans and and big believers on what stablecoins and crypto and Bitcoin has enabled or will enable over the past 1015, years, and so we serve a mass market customer. We don't serve a customer that is trading, that is, you know, the high frequency trader, super sophisticated, or customer is a mass market customer. And when you look at crypto penetration in the world, is still about 1% So we are where the internet was in 1999, and at some point you go from the early adopters to a mass market, and the mass market is the mass market is what we understand. So what you see in the global is a way to use the technology of stablecoins and crypto in a under the hood in a way that for the consumer what they're seeing is just a phenomenal banking experience. By the way, it sits on with a banking partner or Signum a banking partner in Switzerland. So, what consumers really are getting is a Switzerland bank account savings on stablecoins, being able to spend without any fees and send money around the world for free real time.
Sy Taylor 18:44
I think that idea of almost like being able to go into other countries and then being able to see where the users are and they're pulling you into other markets is almost gives you visibility into okay outside of the markets we've gone deep in maybe it's I don't know somewhere in Switzerland or maybe it's somewhere else in Canada. Who knows which country it is? But you sort of get that visibility for where the pull is. I think you made a great point there. The other question I have is like, who are the other partners being under the hood? I know you said Signum. Is there anybody else there that's doing it? And this 3.5% is that something you're funding? Like, where's that coming from?
David Vélez 19:17
Yeah. So so Signum is our bank partner. We're offering Circle. You you are with USDC and Euro C. Yeah, with them, so digital dollars, digital euros. Effectively, the yield that we're paying is we're passing effectively all the yield we get back to the consumer.
Sy Taylor 19:31
Right.
David Vélez 19:32
Because I think this is part of the thesis that we overall thesis that we have globally on deposits, which is consumer should own that yield, not companies. We are just passing that. We're just providing a service, so we're passing effectively all that yield back to the consumer on the stablecoins, and they get a very, very good, very safe, effective savings product. And our partner Mastercard also on the debit side, and then we have a couple of partners like the local and bridge, enabling certain connectivity for certain rail. In specific countries,
Sy Taylor 20:01
got it. Super helpful for the finance nerds. They'll really appreciate that. So thank you, David. I know you also evolved in Tempo, which is where I spend a little bit of time. What's it been like collaborating with like the Collisons and Matt Huang on on that sort of thing?
David Vélez 20:13
Oh, it's been great. I mean, I think we're part of we're part of Tempo as well, and we're actively thinking about how do we integrate some of the technology in there. And and and and there is also a big question about what the rails of the future are going to look like. Yeah, I think everybody's coming in from different angles. So it's been great to be able to to collaborate with them and and and actively discuss what the future financial services is. I think everybody kind of kind of agrees on the end goal. Like what what the photo 1020 year from now sort of looks similar.
Sy Taylor 20:39
Yeah,
David Vélez 20:39
we like to describe it as the world of payments will look like Brazil. This is sort of the not obvious way for people, especially in developed economies, that they don't really know the future payment will be Brazil or India. Because if you are in Brazil and India today, payments are free, real time, 24/7. It coexists with debit and with credit and with and that they're not necessarily zero sum. What what who loses cash? Cash disappears. Yeah. But ultimately, you're moving bytes from one place to another one. There is no cost to moving that byte. Yeah. And people are still paying five to 10% for a transaction and waiting a week to get that. Yeah. There's
Sy Taylor 21:15
risk in those bytes, right? Like there's things that can go wrong, but a lot of that risk comes from the fact that banks often don't work bank holidays or weekends in some countries, and so that creates the risk, which creates the cost. And as soon as you go global and 24/7, which stablecoins can do, you start stripping that cost out and really attacking it. That's it's. I love that topic, and I could bore you with it forever, and I won't because there's another one I want to I want to cover off, which is new former. Right, you acquired a company called Hyperplane in I think 2024.
David Vélez 21:43
That's right,
Sy Taylor 21:44
and I fell off my chair because I was like, they're going to do something with this, but we're not going to find out what it is, and because Newbank never tells you what it is until it's ready, until it's done. You're not sort of like hype merchants. So, what is Newformer, and what does it actually do for you?
David Vélez 22:00
So we have been consumers of machine learning since we started the organization, right? So since 2013, we one of the biggest opportunities we saw in financial services were around credit, and credit is messy. Credit is capital intensive. You need a balance sheet. You need funding. So it's something that most fintechs decided not to get into, right? Because investor will give them a low multiple. We decided to fully embrace, and partly it's because it's 70% of the $10 trillion financial services profit pool. So this is the biggest price: is credit. It's not remittances. It's not deposits. It's not float. It's credit, but it's also where the biggest consumer pain point is. You have 3 billion people around the world without access to credit in Latin America. About 250 million people do not have access to credit. So, when we saw interest rates in Brazil at 1,200% we said there's got to be a better way to do this, right? By using better data, analytics, machine learning, and so from the beginning, we we were very focused with try to develop this credit capability as a core competence. And then, with the advent of AI, we started understanding what other approaches might be to this credit decisioning. We've always used more initially linear regression models or gradient boost models. And when we start hearing about LLMs and potential different approaches, we say, "Wow, there is something interesting here about the technology. We saw Hyperplane. We met the Hyperplane team. A lot of Brazilians and Americans and Indians, and and Rohan and they're the co-founders. And we say, "Wow, this this team is incredible, and we were able to partner with them to then try to basically what NoFormer is is there then, and and we're now in in V2 and V3 is create our own foundation model, the same sophistication that the large AI labs have built on language models for any more general kind of a language and request. We're doing it specifically for financial services, and initially specifically for credit for credit decisioning. So we're now in several versions. Now we published our entire research. I think in 2025 we were the first, I'd say, financial services firm in the world to publicly open. Today it's a it's a unsophisticated model of NEA lab is using over 500 terabytes of data. We are increasing our V2 is 4x has 4x more context and 4x more data that are V1. What's great about these use cases there is a lot of conversation in other companies' applications about AI. Is this what is the ROE of AI? Everybody's excited about using the tokens. What is the ROE? Here is a specific area where you can see the ROE. This this model has increased significantly the AUC, the area that recovered most of our credit models, and that translates directly into more NPV. Like we can measure the hundreds or millions of NPV being created by new generations of this model.
Sy Taylor 24:40
Because suddenly, customers that looked not creditworthy start to look credit worthy under that curve because it's more accurate and more more powerfully forecasted. Which means your 250 million goes from being oh well, I can address 100 of those to 150, which kind of is mission aligned. Are there any promising results you've seen from that? It's deployed. It's in production. It's in all. Markets like where are we at in terms of live?
David Vélez 25:02
Yeah, no. So we are. We'll have V3 by December. Each generation is four or 5x better than the last one. Again, more data, more context, more data points. We have now integrated into a lot of different decisions. What's great about the foundation model is that we don't need to do what we used to do, which was a lot of investing a lot of efforts in feature engineering.
Sy Taylor 25:21
Yeah, you
David Vélez 25:21
want to test the GPS of the smartphone in a in a model. You have to go and engineer that feature, put it in the model, and there is a there is some some some time of figuring out how do you do that test. You have all that. You you grab all the data. We have over 30,000 data points of 140 million customers over 13 year history, and you aggregate it, and then this foundation model enables, for free, the ability to create models with applications around approval of a credit. But if you approve, what should be the limit? Yeah. If it's collection, for example, do I call you at 5p.m. or I call you at 2p.m. Should I give you an offer for collections and renegotiation of 10 to 50 for fraud, for example. There is a lot of applications on fraud, transactional fraud. There is applications on anti-money laundering, applications on know your customer. Right, you have a lot of information, especially as we as we go into remittances in Mexico and Colombia and Brazil. There is a big big point. So it just gives us a huge amount of degrees of freedom of how we use our own data. And when we launch one of the things that we'll be testing, and we're excited about this, as we launch a new country, the fact that we bring 140 million data from 140 million customers in these models should allow us to grow and underwrite faster customers that we never seen before. Wow! With with transactional and behavioral information, we haven't proven that yet. No, no, we're excited. I'm
Sy Taylor 26:39
excited to see that get deployed and figured out because people who've never done feature engineering don't know how much time that takes. And I'm imagining your internal like data science and research teams are now spending more time doing different tasks rather than managing data pipelines and just kind of turning it over and trying to get marginal efficiencies. I've got this bigger thing that I can throw at lots of things. So now, how do I make the big thing better, which changes it? But I know you've also like built these personal banker agents. Do you see a world in which the customer is interfacing more with an agent like Instinct or Muse than with your own mobile app? Like, is is that too early to say, or how do you feel about that?
David Vélez 27:20
I think it is very possible that I mean you'll have a new bank agent inside News. You'll have a new bank agent working with Grok or working with ChatGPT. You have a connector, and and there is one on one end. New bank is banking your agent. On the other end, there is a specific agent that is helping you figure out what you do with your life. Historically, there's been a lot of conversation about financial advice. That's the way you solve lack of allegation, sophistication. That never really worked. People don't do financial education.
Sy Taylor 27:48
No, it doesn't work.
David Vélez 27:49
It doesn't work. It's just boring. It's it's complex. So we're excited about what what AI could really mean for effectively enabling every person around the world to grab somebody to grab their hand and helping them make the best financial. Hey, do you
Sy Taylor 28:03
want to do this? You might want to save a little bit. I could move this over here if you like. It's those nudges things that start to feel because these were always like special PFM apps, and the the budgeting nerd really liked them, but everybody else didn't use them. There's something magic about that that agent experience. Like, have you seen anything in early customer testing that stood out to you when you've been like rolling out some of these agents?
David Vélez 28:26
It's a bit counterintuitive, but here we've actually purposely being a little a little slow to truly open it up to everybody. We actually don't think AI is fully ready.
Sy Taylor 28:36
Yeah,
David Vélez 28:37
there is a you know AI is solving incredible mathematical problems for a lot of these problems, still specifically, if you're doing it with that consumer obsession, which is what we're optimizing for, we're not optimizing for saving money or efficiency. So it's not just
Sy Taylor 28:50
a call center. It's not a
David Vélez 28:52
call center optimization. A lot of people went in and put the AI on call center, and I think that was actually a huge disservice to AI because this Navier wants to deal with these chatbots,
Sy Taylor 29:01
yeah.
David Vélez 29:01
But going back to culture, if we are warning customers to love fanatically, the last thing we want to do is put our customer with this chatbot that's not going to answer any questions. So we've been very thoughtful and very very careful about where how this AI agent gets constructed. We have about 15 million customers already using it inside the app in different ways, partly in in ways that are truly useful. So there is a financial summary and financial advice that is getting to levels of NPS that are now ready for mainstream to go out. Getting there, but for example, sending money via voice, we call it picks with AI. There is a lot of renegotiation. One of the biggest pain points for customers is you are in debt, right? You you cannot pay your bills. You have a lot of debt in a lot of different banks. Get me out of debt. How do I get out of here? And so there's been a lot of great progress in in the agent helping unsophisticated customer to understand interest rates at different banks, understand all the different problems, and try to try to solve it. So we're going, we're moving, we're getting excited about where we're heading to, but it's a little earlier still if you're trying to optimize for for customers. The experience
Sy Taylor 30:08
just isn't quite there. But I I want to go inside the organization now, like agents and use of AI. I'm imagining you know you're the CEO of a bank, but you're the CEO of a tech company at the same time. A lot of banks say they're tech companies that that do banking. I don't think they are. But like, how does your job look different today than it would have done two, three years ago because of AI? And like, what are you tracking and looking at to see that the company's doing well?
David Vélez 30:34
I think what's what's really exciting about AI is that it allows us to question everything, like reinvent. It's like we talked a lot in Sanuic about how companies have new founding moments. There's not one single day that you found the company. You have refoundings, especially if your organization are willing to challenge yourself actively and maintain that change
Sy Taylor 30:52
how you do what you do.
David Vélez 30:53
Thanks for you do because I think especially a lot of successful organizations get to a point where they feel that they win.
Sy Taylor 30:57
They've won.
David Vélez 30:59
We won, and then they look around. Any change now becomes a threat, and and and you become too enamored with your own success. And then, like anybody comes in and proposes a different strategy, say no, no, no, like we got, we got it. So AI has enabled us to question everything. We literally deconstructed our business in 120 different workflows, and we have all teams rethinking how to reinvent that flow through AI, putting AI in the center. And we're seeing huge amount of improvement operationally. It's funny; a lot of the improvements, the fast improvements, actually have nothing to do with AI. It was like obvious stuff that we were doing, or obvious steps in a workflow that we could have removed even without AI, AI gives us the easy win, the free win of reinventing, and then you put AI on top of that, and it's 10 to 100x opportunity gains across the entire organization. So it has changed a lot of big priority, or especially 1215, months, has been really leading the transformation on the organization as a CEO, it has also enabled us to become much more horizontal in the organization. I have now 20 direct reports, and each direct report is has a much more horizontal team. We went from about 10 layers to five layers. Wow! Much more horizontal. That's faster speed of execution, and that's kind of more connected, more ingrained into the reality of the business.
Sy Taylor 32:21
I think that thing of like we question every workflow is something that I have not seen large financial institutions do. They're putting AI onto their existing workflows to try and make those go faster, but they might be the wrong ones, and there might be steps you could remove.
David Vélez 32:33
You're doing the faster horrors, which is you might not need that, right?
Sy Taylor 32:36
Yeah, completely not. So I'm interested as well in your day to day. What are you tracking outside of AI to know that the company's successful? Like, what are the anecdotes? What are the data points? What do you care about day to day? Is it customer feedback? Is it a number? Like, what's top of mind for you?
David Vélez 32:52
Yeah. So, still today, even after the announcement of US and Global today, still the biggest opportunity we have as an organization, or the most tangible opportunity we have as an organization is Brazil, and then Mexico and Colombia. So the opportunity is still even in Brazil, where we have now close to 65% of the adult population of the country. It's like we at some point we're going to run over Brazilians, right? Yeah, we're still getting close to a million customers every month in Brazil. We'll need to find a way for Brazilians to have more babies and to see demographics grow, because otherwise we're going to run out of Brazil. But we only have 7% market share that is growing, and and a lot of the focus, my focus is, are we continue to execute extremely well in Brazil? We now need several segments. We have options for for small businesses, for under 18, for what we call super core and high income. We are launching several payment solutions. We're building what's the beginning of a really interesting payment ecosystem with something called New Pay, which is our BNPL and doing on-off payments inside Brazil. So there is a lot of innovation. There is AI on the product. So a lot of what I'm looking at is: Are we executing on Brazil? What is the NPS? What is the RPAK or average revenue per active customer slope? We're about $17 per customer. Incumbent banks are 45. We don't charge all the fees that they charge, so maybe we won't get to 45. We'll get to 40 or 35, but that still means we could more than double average revenue per active customers over the next few years, and so that means that since we maintain the same cost to serve, we could just double the business in Brazil alone.
Sy Taylor 34:25
Yeah, with product expansion, and then you think about geographies after that.
David Vélez 34:28
Exactly, with product expansion and even maturation of cohorts, what you see is simply by time growing by and customers going from two years to three years, they'll just consume more of our products. And then Mexico is early days. We're now the largest digital bank in the country, but we only have 16 million customers there. We have 2% of the profit pool. Huge opportunity of Mexico to become another Brazil. Colombia is also going in the same direction, and then the international expansion. So a lot of the focus is is the core, and then obviously starting to track what are the big levers. For growth over the next 10 to 20 years, which is what we announced this week.
Sy Taylor 35:03
I'm so fascinated to watch New Global. As somebody sitting in Europe, I'm hoping you guys will come join me in the UK at some point. That would be fun. We'll definitely try. Please do give it a shot. So I have this hypothesis, and I'm doing a lot of interviews and writing at the moment about like the fintech era was cloud and mobile. My crazy statement for the day is like fintech is dead. The reason I say that is because cloud and mobile now are table stakes. Like I think you have to be good at those to get them, and that stablecoins, tokenization, and especially AI are going to be the thing that drives into the next decade. So, do you agree with me that like the fintech era is now somewhat table stakes? Like you have to do that to be competitive at all, and the position is moving, or are you still getting like a ton of value out of just being cloud and mobile and branchless? I guess that's still a big competitive advantage in some spaces, but not others.
David Vélez 35:51
I agree with you in some in some ways, and disagree with you in some ways. And a reason why I agree with you in some ways is that the opportunity of fully disrupting as an early stage startup one of the big banks is less obvious today. Yeah, the advantage. Maybe there'll be an AI advantage, but AI will require a huge amount of scale and data. So it's hard to see how a lot of the data and scale will benefit the team that is just starting from scratch because you a lot of these proprietary data. Imagine one of the AI labs had a monopoly of all the data that they use their frontier model on.
Sy Taylor 36:27
Yeah, it
David Vélez 36:27
would be very hard for anybody to compete. That's a little bit of what companies because finance data is regulated
Sy Taylor 36:33
and safe and secure, and so it's much
David Vélez 36:35
harder to get. So in some ways, I do think I agree with you. The reason I disagree with you is like if you just can't stay back and you see an entire financial services industry,
Sy Taylor 36:43
yeah,
David Vélez 36:44
95% of the financial services consumer banking revenues still flow from banking banks from from banking branches. Yeah, still owned by incumbent banks is what we were discussing earlier, right? So it's almost as if 95% of all the media was consumed by blockbusters. Still today,
Sy Taylor 37:00
there's a lot of room to
David Vélez 37:01
36, and banks are growing branches. Yeah. So blockbusters still opening branches.
Sy Taylor 37:07
It's the only way they know how to
David Vélez 37:08
right. And and you know, listen, like there are certain reasons why you would do that. There are certain niches. There are certain consumer segments. If you have $5 million you want to invest, you want to see a human. If you want to get a mortgage, probably as well. If you are a small business, then it's a $10 million loan. You probably as well. But for the 90% of consumers, you don't really you don't need to go into the bug buster to get a DVD. You just cannot go online.
Sy Taylor 37:31
Yeah.
David Vélez 37:32
And so that dynamic of being able to compete with the industry that owns 95% of all the profit pool in the world that still is very much in coming back. That are still they're not decreasing necessarily efficiency. They're increasing branches still. That still hasn't played out, and that's that's still just mobile and cloud.
Sy Taylor 37:49
Yeah.
David Vélez 37:49
So you got room
Sy Taylor 37:50
to run, but the opportunities elsewhere. I think that's probably fair, and I think I actually agree with you. But I like a lightning statement anyway. It's going to got to have some fun. Talk to me about ambition. Where are you setting the boundaries here? Do you want it? Because I have never seen a truly global consumer brand. Like you get regional ones, you get transatlantic. You know, there was I think in the two early 2000s there were a lot of like universal banks that were trying to go global, but they all sort of retrenched after a little while. Are you trying to go around the world here? You're looking at Asia one day, Middle East. Like if you were to say, you know, in 20 years' time, is that part of the ambition, or was the limitation here, if there is one? The
David Vélez 38:27
problems that we're trying to solve are not specific to any geo. So that's why I think maybe if you go back even to the universal banks 2030, years ago, you you solve brands trying to execute that because the problems, as we like to describe them, is these problems around access, cost, and intelligence, and this is global in nature. The problem is that financial services, as you very well know, is is very localized and it's very protected from regulatory perspective. And so, going back to Netflix, you saw with a Netflix business model, you can open up 50 countries at once. The website is there, and people, and you have access to customers. Or in some areas like Amazon or Uber, you can small small teams and you pop up new countries. If I open up our website to customers in a specific in a lot of countries today and start offering banking services to them, we're gonna have local licenses.
Sy Taylor 39:14
Yeah,
David Vélez 39:14
I mean travel. This is very regulated, so that's why banking, that trend of global firms offering products have been much more delayed than what you see in our industries. You saw it in crypto. There are some large crypto firms that have done that, but mostly unregulated or outside the regulation, and they're not not trying to do things like credit, specifically in the local markets. But I do think the trend is in that direction, and so technology and AI and digital make it easier, reduces the barriers, because we don't need to put a bunch of banking branches in every corner. In 2010, if you were city, you need a banking branch. You need a billion dollars of capex.
Sy Taylor 39:53
Yeah,
David Vélez 39:54
you don't need mainframes in every country anymore. You have cloud, so that's huge. You have. Consumers are using smartphone everywhere around the world, and you can use a lot of that data for underwriting, for example, or getting to the customer. So, in a way, technology has decreased the locality and the isolation of each country. Makes it possible. It's still not easy. It's still complicated, and you require a lot of thoughtful systems to be able to execute a multi-country strategy to be able to solve those three problems on a global nature, so we're gonna execute. We, I think, the announcement for us here is yes, we're gonna give it a shot, starting with the U.S. and new global. This is a multi-decade opportunity.
Sy Taylor 40:30
Yeah,
David Vélez 40:30
this won't happen overnight. We think we have a lot of capabilities that we built over 13 years, waiting to 140 million customers. That consumer obsession, the the the credit, the the way we use AI to give it a shot, and ultimately, it's a market that is still dominated by big incumbent banks that we we think we can we can just bite, take take a little bite of that of that market.
Sy Taylor 40:48
Just a little bite would be nice. Talk to me about what's the hardest strategic question you're wrestling with right now.
David Vélez 40:55
There's a couple one. First one is expansion, right? Is a geographic expansion in one axis versus segment expansion, or in another axis. So you go broader, or you go deeper in specific markets. Because since, as I mentioned, we still have a lot of growth in Brazil.
Sy Taylor 41:09
Yeah.
David Vélez 41:10
So that's that's a big strategic question and execution question. How much can you do until you bring enough soldiers to lose the war? Right. Yeah. Companies start executing so much, you become really, really stretched and become start hurting your execution on the core. So that's one question. The second question has always been, and especially in test 12 months, is AI. How does AI changes things? Yeah. From an internal perspective, what can be reinvented? We talk about credit. We talk about systems, but from a consumer perspective, also is is why one one we we think that one of the reasons why, specifically in the USAI, has gotten so bad reputation, partly is because it hasn't been readier to touch consumers yet. In some ways, and it hasn't been super obvious how useful it can be
Sy Taylor 41:51
better Google.
David Vélez 41:52
Exactly, and so we've well, we are very forward in some areas here. We've been very a little slower and a little bit more careful. Yeah, even though it's a marriage organization, we're taking just much more risk on the AI side. So that managing that speed and and knowing where to accelerate and where to decelerate and where to slow, where to go fast, is something that we are actively thinking now as a team.
Sy Taylor 42:14
I've heard a few people describe New Bank as one of the world's best technology companies that just happens to be a bank. I know you taught yourself Clojure at one point. I think I read somewhere. Like, what what does it mean to be a technically literate CEO? Like, how does that help you do your job? Yeah,
David Vélez 42:32
I did start learning Clojure. I I would confess, I did not become too fluent in in in in Clojure, but AI gave me the opportunity to then get back into coding again, abstracting clojure, and and get back to code, and push even push a little bit of the teams. Right, we one of the one of the experiments, one of the things that we talked for a while is a non custodial crypto app, for example. And I was able to code one in an afternoon and show the materials like, guys, we're talking about this. Here you go, I did this.
Sy Taylor 42:59
Yeah, but showing up, I think, sometimes with a prototype just moves a conversation forward so much. It's it's really impressive that that can happen.
David Vélez 43:07
So that gives superpowers. Now, I think it's important for the CEO and for every executive to be spending today, especially right now, spending time using the AI tools because they're transformative.
Sy Taylor 43:15
And how much of a week is that at the moment?
David Vélez 43:18
Because some weeks is one hour, some weeks is I do get a morning and into a morning to focus. I need five six hours. Some weeks I get a little bit more blood, but I don't think it's the best use of my time to to be doing that necessarily.
Sy Taylor 43:32
Yeah,
David Vélez 43:33
I'm not gonna be the best programmer, even with AI inside. But you
Sy Taylor 43:35
have to be literate, right?
David Vélez 43:36
But I have to be literal. I need to understand the pace of how the technology is moving, and there is also a little bit right now in social media. I was like, "Oh, I'm spending 90% of my time coding as a CEO's. Like, no, that's not your comparative advantage. No. So, so, so you have to you have to be literate. You have to know what you're talking about. Technology is moving very fast, so you have to be connected. But at the same time, there is my comparative advantage is doing something else. So, last question for you: If if we come back and we do this interview in three years' time, what are your hopes for your US launch, and where do you think we we stand on that? I think we're going to be very focused on finding that product market fit and those customers that really love the product fanatically. And those could be in a number of different niches, segments, demographic. We're going to be much more focused on on making those first 1000 customers or 10 10,000 customers incredibly excited at their customers versus necessarily what is the growth curve looks like how many customers do we have the way we've expanded historically we don't go into a new market and go in with a big bang and invest hundreds of millions of dollars in marketing we we like to try to grow things a little organically listen to customer iterate. We're okay if things take a little bit longer, just so that we can find the right unit economics and the right product. So in a three years, three year time, we absolutely should have known if we found that that that those that that unit economics, we found that great consumer love in certain niches, and then we're preparing to then invest a bunch of money to capture a big. Chunk of the market is also not going to be cheap to capture. It's going to require capital by no means to get to a segment of the of the U.S. market will be will be for free. Will take a significant amount of money, but we want to do it in a way where we have actually then conviction about the NPV or the ROE of that dollar that gets invested.
Sy Taylor 45:17
Product market fit plus having conviction on unit economics is a great combo, David. I have enjoyed this so much as a fan of all you're doing. Thank you very much. Great interview. Thank you. Thank you.
David Vélez 45:26
Thank you.