Latest / Fintech Conversations with Fexingo: Banking, Payments, and Financial Software Companies / The Neobank Profitability Puzzle Solved by Nubank
Transcript
- Lucas: There's this persistent question in fintech that I hear from investors and operators alike: do neobanks actually make money? And for years the answer was kind of a shrug. But last year, one digital bank posted over two billion dollars in net income. That bank is Nubank. Luna: Two billion is a real number. That's not just 'we're trending toward profitability' — that's actual, durable profit. Lucas: Exactly. And the reason I want to dig into Nubank specifically is that it proves the neobank model can work, but only under specific conditions. Nubank operates mostly in Brazil and Mexico — high interest rate environments, under-banked populations, and a credit card culture that's very different from the US. Luna: So the macro tailwind matters. But plenty of banks in Latin America still lose money. What's Nubank doing differently at the unit level? Lucas: Let's start with the core product. Nubank launched with a no-fee credit card — no annual fee, all app-based. That card became their wedge. By 2025, they had over 100 million customers. But here's the kicker: their cost to serve a customer is roughly 80 percent lower than a traditional Brazilian bank like Itaú. They don't have branches, they don't have legacy IT, and their cloud-native stack means they can add features without adding headcount. Luna: Eighty percent lower — that's a structural advantage. And it means they can afford to give away the card and still make money on the back end. Lucas: That's exactly it. Their revenue per active customer in 2025 was around eleven dollars a month. Most of that comes from interchange fees on card transactions and from interest on revolving credit. They also cross-sell — insurance, personal loans, investment accounts. The cross-sell ratio is about 3.5 products per customer. That's huge when your marginal cost to add a product is near zero. Luna: Compare that to a typical US neobank like Chime, which relies heavily on interchange but doesn't have the same credit product depth. Or Monzo in the UK, which only recently turned profitable on a pre-tax basis. Lucas: Right. Monzo reported its first full-year profit of about 18 million pounds for the year ending February 2026. That's progress, but it's a rounding error next to Nubank's two billion. And Monzo operates in a lower interest rate environment, with a more saturated market. Nubank's edge is partly market selection — Brazil's base rate was still above 10 percent for most of 2025. Luna: So the spread on deposits is fatter. If you're paying zero interest on demand deposits and lending at 15 percent, the math works differently than in the UK or US. Lucas: Exactly. But let's not over-attribute it to macro. Nubank's return on equity was over 20 percent in 2025. That's best-in-class globally for any bank, digital or not. They achieved that through disciplined underwriting — their non-performing loan ratio is actually lower than the Brazilian banking average. And they use a proprietary data model to score credit risk, pulling from transaction history and behavioral data. Luna: That's the machine-learning moat. The more customers they onboard, the better the model gets, and the lower the defaults. It's a virtuous loop that legacy banks can't replicate easily. Lucas: And that loop is exactly why their customer acquisition cost has been falling. In 2022, it cost them about six dollars to acquire a new customer. By 2025, it was under four dollars. Meanwhile, their average revenue per customer is climbing. So the unit economics improve every year. Luna: But Nubank's story isn't just Brazil. They expanded into Mexico in 2019, and that market is now a significant growth driver. How does the model transfer? Lucas: Mexico is similar in some ways — high unbanked population, high interest rates. But it's a different regulatory environment and a different credit bureau system. Nubank had to adapt. They launched with a debit card first in Mexico, then added credit. By the end of 2025, they had over 5 million customers in Mexico and were already profitable in that market. That's faster than any Mexican digital bank. Luna: So the playbook is: enter a market with a low-cost card product, build a data moat, then cross-sell higher-margin products. And it only works in markets with high spreads and low banking penetration. Lucas: Right. Which is why I don't think we'll see a Nubank-like outcome in Europe anytime soon. The margins just aren't there. But for emerging markets, Nubank is the blueprint. And that's what makes it worth studying. Luna: If today's episode was useful to you and you want to keep it ad-free, buy me a coffee dot com slash fexingo helps. We keep the lights on with listener support, no sponsors. Lucas: Yeah, it really does. And we're grateful for every contribution. Now, back to Nubank — I want to talk about the competitive response. Traditional banks in Brazil are fighting back. Itaú and Bradesco have launched their own digital apps. But Nubank's cost advantage is so deep that they can still undercut on fees and offer better user experience. Luna: Can the incumbents ever catch up? They have the deposit base and the brand trust, but their cost structures are anchored by branches and legacy systems. Lucas: It's hard to see them closing the gap on cost. A branch costs millions per year to operate. Nubank's entire physical footprint is a few offices and data centers. That said, regulation could be a risk — Brazil's central bank has been pushing for open banking, which might erode Nubank's data advantage over time. Luna: That's the classic fintech dilemma: you build a moat with proprietary data, then regulators mandate data sharing. It happened in the UK with Open Banking, and it's happening in Brazil. Lucas: But Nubank is already pivoting. They're building a marketplace model — offering third-party products like insurance and investments through their app. That turns open banking into an opportunity rather than a threat, because they can be the front door for financial services even if they don't originate every product. Luna: So the long-term bet is that Nubank becomes a financial super-app, like WeChat Pay but for banking. Lucas: Exactly. Their CEO David Vélez has said publicly that the goal is to be the primary financial relationship for every customer in Latin America. And if you look at engagement metrics — average daily app opens are over 8 times per month per user — they're well on their way. Luna: That's insane engagement for a banking app. Most traditional banks see single-digit logins per month. Lucas: And that engagement feeds back into the data moat. More data means better credit scoring, better cross-sell targeting, and lower churn. Their churn rate is under 2 percent annually. For context, a typical US neobank churns around 10 to 15 percent per year. Luna: So the thesis is: low churn plus falling acquisition costs plus rising revenue per user equals a compound growth machine. That's the math behind Nubank's market cap — which is around 60 billion dollars as of this spring. Lucas: Right. And that multiple — roughly 25 times 2025 earnings — is expensive by traditional bank standards. But Nubank is still growing revenue at 30 percent plus annually, and they have a total addressable market of over 600 million people in Latin America, most of whom are under-banked. Luna: So the question for investors becomes: can they sustain that growth rate as they mature? And can they replicate the model in other emerging markets? Lucas: They've already started. They launched in Colombia in 2024, and they're eyeing other markets in Latin America. But each country has its own regulatory and cultural nuances. Mexico took four years to become profitable. Colombia will likely take a similar timeline. Luna: Which means the stock's valuation is pricing in a lot of future expansion. If any of those bets stumble, the multiple could compress. Lucas: For sure. But here's what I think is underappreciated: Nubank's profitability isn't just a Brazil story — it's a proof of concept for the entire neobank thesis. If a digital bank can generate 20 percent return on equity in a high-inflation, high-interest environment, then the model is viable. The question is where. Luna: And that's the lens I think listeners should take away. Nubank isn't the template for a neobank in London or San Francisco. It's the template for any fintech targeting an emerging market with high spreads and low penetration. That's a huge opportunity, but it's not universal. Lucas: Exactly. So next time you hear someone say 'neobanks don't make money,' you can point to Nubank. But you should also point to the conditions that made it work. That's the real lesson. Luna: Well said. And I think that's a good place to leave it for today.