Disclaimer: Please note that the author (Paul Coquant) is long Nubank Holdings Ltd. While he remains confident in the thesis presented here, he may adjust this exposure at any time based on market conditions. This article represents his personal opinion and should not be viewed as investment recommendations.

Introduction: Nubank business, history and management team

Nubank is the largest digital retail bank in Latin America and, by customer count, the single largest private financial institution in Brazil. It operates a zero-branch, app-only architecture across three countries: Brazil (113 million customers), Mexico (14 million) and Colombia (4.2 million). The product stack has widened from a single no-fee Mastercard in 2014 to a full retail bank: NuAccount (digital current account integrated with Brazil’s instant-payment rail PIX), personal unsecured loans, payroll-deductible and FGTS-deductible secured loans, investments (via the Easynvest acquisition rebranded Nu Invest), insurance, NuCel (mobile telephony), and travel. Revenue is split across three economic streams: credit income (interest on a $32.7bn book plus card interchange), float income (NII on customer deposits deployed in sovereigns and the interbank market), and fee income. The combination produces a 10.5% risk-adjusted net interest margin, a 19.9% cost to income ratio (vs. ~40% for the Brazilian incumbents), FY25 net income of $2.87bn on $15.8bn of revenue, and a 30% IFRS return on equity (33% annualised in Q4’25): a cost-and-capital structure that legacy retail banks cannot replicate without rebuilding their technology stack and closing their branches.

The company was incorporated in the Cayman Islands in 2013 and launched its first product, an international Mastercard managed through a smartphone, in mid-2014 into a Brazilian credit-card market dominated by the big-five incumbents, where effective annual interest rates on revolving credit routinely exceeded 400% and customer acquisition required physical branches. Early growth was referral-driven: waiting lists rather than advertising. Leading venture capital funds financed the early development of Nubank, and Berkshire Hathaway anchored the IPO with a $500m stake at $10/share in December 2021 (subsequently trimmed and fully exited by mid-2025). The history of banking licenses is instructive: Nu Financeira (SCFI) in 2018 enabled the personal-loan book in Brazil; the Nu Invest acquisition (June 2021) added the brokerage; Mexican banking approval came in April 2025; US OCC conditional approval for a national bank charter (Helen OT / Nubank N.A.) in January 2026; and a full Brazilian banking license is expected during 2026.

The company remains founder-led and founder-controlled, with a four-person executive core. David Vélez, Chairman and CEO, is a Colombian engineer who trained at Stanford and was a Sequoia Capital partner running the firm’s Latin America practice before founding Nubank after a personal experience opening a Brazilian bank account that required him to present documents behind bulletproof glass. He holds c.17% of the economic interest and c.75% of voting rights through Class B super-voting shares and has sold only once since the IPO (a $404m trim in August 2024, cited as estate planning). Cristina Junqueira, co-founder, Chief Growth Officer and CEO of the US operation, was formerly head of credit-card products at Itaú; she now carries the dual mandate of driving LatAm customer and revenue growth and building the US franchise from the Delaware entity that received OCC conditional approval in January 2026. Guilherme Lago, CFO (ex-McKinsey, Credit Suisse), is widely credited with the disciplined deposit-funded capital strategy and for the managerial-P&L framework introduced in Q4 2025 that increases transparency on per-product economics. Eric Young, CTO, manages the engineering stack that underpins the 19.9% efficiency ratio and the real-time credit and fraud engines; his remit spans the Brazilian core, the nuFormer LLM deployed in credit decisioning and collections, and the cross-market platform that now has to scale into Mexico (banking license live since April 2025) and the US.

1. A rare consumer-centric company in the financial sector

A. Unit economics: growing revenues per customer

Most banks make money the way a supermarket does: they get as many people as possible through the door, sell them roughly the same basket of products, and take a margin on each. Nubank behaves more like a streaming service. When a new customer signs up, they cost almost nothing to acquire — the typical new user comes in through a friend’s referral rather than an advertising campaign — and in the first year they barely generate any revenue at all. But that same customer, if they stay, uses more products every year: first a no-fee card, then a savings account, then a loan, then an investment account, then insurance, and more recently a mobile phone plan and travel bookings. Each extra product adds a small monthly revenue stream. The longer the customer stays, the more they are worth. This simple dynamic — of each customer quietly becoming more valuable over time — is the single most important thing to understand about the company.

The numbers make this very concrete. Nubank publishes how much money it earns per month from each group of customers, sorted by how long they have been with the company. A customer who joined in the last twelve months generates only $0.80 per month. Someone who joined between two and three years ago is worth $7.40 per month. At five years, they are worth $14.80. At seven years, $24.70. And a customer who has been with Nubank for eight years or more now generates $30.20 per month — roughly thirty-seven times more than a first-year customer. This ladder is remarkably consistent: each year a customer stays, they add another two to four dollars of monthly revenue, because each year they typically adopt one more product. Crucially, the ladder has not broken down as the company has grown from a niche product in 2014 to a household brand in Brazil by 2025 — the most recent cohorts of customers are climbing the same steps at roughly the same pace as the earliest ones.

If the company were just signing up new customers, the average revenue per customer would stay stuck near the bottom of the ladder. What actually happens is the opposite: because millions of customers have been with Nubank for several years already, the average is being pulled upward as each vintage matures. The monthly revenue across all active customers was $11.10 at the end of 2024. One year later, at the end of 2025, it was $15.00 — a 35% increase in a single year, achieved without raising prices and without changing the product mix. The mechanism is purely that older customers now make up a larger share of the base, and they spend more. Put another way, Nubank doesn’t need to find new growth to grow — it simply needs its existing customers to keep doing what they have always done.

Translated into the company’s reported revenue, this is what you see: total revenue grew from $1.7 billion in 2021 to $4.8 billion in 2022, $8.0 billion in 2023, $11.5 billion in 2024 and $15.8 billion in 2025 — a near-tenfold rise in four years. A material part of this came from new customers, but an increasing share now comes from existing customers climbing the ladder. The mature cohorts (those at $30 per month) are still small today, but they are growing fast: every year, another annual intake of early customers reaches the top step, and today’s $30 customers were themselves at $15 only a few years ago. This is the opposite of a traditional bank, where a mature customer is the least exciting one in the book. At Nubank, the mature customer is the most valuable — and the company has 131 million of them lined up, each at a different rung on the same ladder, all still climbing.

B. Alongside strong growth in the number of clients

The second engine of the company is simply how many people use it. Nubank started with a single customer — its founder — in 2013 and signed up its first paying user in 2014. Eleven years later, at the end of 2025, 131 million people across Brazil, Mexico and Colombia held a Nubank account. To put that number in human terms, it is more than the combined populations of France and Spain, and roughly four times the population of Canada. It is also the largest customer roster of any private financial institution in Latin America, and by a wide margin — the second-largest digital bank in Brazil, Banco Inter, has around 35 million customers; the biggest traditional bank by customer count, Caixa Econômica, is a state-owned institution. No company in financial services anywhere in the world has grown a customer base this fast from a standing start.

The year-by-year pace tells the same story without exaggeration. At the end of 2021 Nubank had about 53 million customers. That rose to 74 million at the end of 2022, 94 million at the end of 2023, 114 million at the end of 2024, and 131 million at the end of 2025. In each of the last four years the company has added between 17 and 20 million net new customers — roughly one new customer every two seconds, year after year. The growth rate (in percentage terms) has naturally slowed as the base has gotten bigger, but the absolute number of people joining has stayed remarkably steady. And crucially, these are not dormant accounts created for marketing purposes — 86 of every 100 customers use their Nubank account in any given month, a level of active engagement that is unusually high for a bank and closer to what you would see at a consumer technology company like Spotify or WhatsApp.

The geography of that base explains why the growth is not yet finished. 113 million of the 131 million customers are in Brazil, where Nubank now reaches around seven out of every ten adults. This is a country the company has largely saturated, and its growth there will come mostly from each customer using more products rather than from more customers walking in the door. But Mexico — the second-largest country in Latin America by population, where Nubank launched in 2019 — has just 14 million customers so far, about 15% of Mexican adults; it grew by 70% in 2025 alone, and is roughly where Brazil was in 2020. Colombia, launched in 2020, has 4.2 million customers, around 11% of Colombian adults, and is at an even earlier stage. If Nubank reaches the same penetration in Mexico and Colombia that it has in Brazil, those two markets alone would add another 80 million customers on top of today’s base.

What makes this particularly impressive is how little Nubank has had to spend to get there. Traditional banks typically spend 5% to 10% of their revenue on marketing and advertising, on top of the cost of running thousands of branches. Nubank spent around 2% of revenue on marketing in 2025 — roughly $300 million against $15.8 billion of revenue — and runs no physical branches at all. Its customer acquisition has come overwhelmingly from existing users recommending the product to friends and family. In the early years in Brazil, the company deliberately rationed access by operating a waiting list, which created social status around receiving the purple card; later the referral mechanism simply took over. This matters financially for two reasons. First, it means each new customer costs the company almost nothing to sign up, so every dollar of revenue they eventually generate is closer to pure profit than at a traditional bank. Second, it is the clearest possible evidence that customers genuinely like the product — people recommend what works and complain about what doesn’t, and 131 million accounts is the outcome of that conversation repeated tens of millions of times over eleven years.

C. Supported by an extremely low cost to serve

The third engine is the quiet one. Nubank publishes each quarter how much it spends per month to look after a single customer — answering questions, processing transactions, running the app, watching for fraud. At the end of 2025 the figure was $0.80 per customer per month. Less than a cup of coffee, less than almost any recurring expense a household has. It has barely changed in three years — $0.80 in 2024, $0.80 in 2023, $0.90 in 2022 — even as the customer base roughly doubled and the product stack widened from a card-and-account proposition to a full financial supermarket of investments, insurance, loans, mobile telephony and travel. Set that $0.80 against the $15.00 per month the average customer now pays the company and the shape of the business is already clear: it costs almost nothing to serve, and the gap between what comes in and what goes out becomes profit.

To appreciate how unusual this is, consider what a traditional bank has to pay for. An Itaú or a Bradesco operates three to four thousand physical branches, each with rent, electricity, security and a dozen staff, and employs around 90,000 people. Their technology runs on decades-old mainframes that need armies of specialists to keep alive; they print statements, process paper forms and staff call centres in several cities. Nubank has zero branches, roughly 8,000 employees and a single modern technology stack built from scratch in the last decade. The standard measure of bank efficiency is the share of revenue eaten by running costs — lower is better. For the big Brazilian incumbents it sits around 40%. For Nubank at end-2025 it was 19.9%. Compounded across 131 million customers, the gap is enormous.

More striking still, the advantage is widening rather than shrinking. At end-2022 the same efficiency measure stood at 41% — worse than the incumbents. In three years Nubank has cut it in half, with every quarter of 2025 improving on the last. The reason is that almost everything a customer does at Nubank is handled by software: opening an account takes minutes through the app with no human involvement, most queries are answered by automated tools before reaching a person, and a proprietary AI model called nuFormer has started making credit decisions, writing collection messages and handling first-line support. When the company adds a new product — a mobile plan, a travel booking, a foreign-currency account — it is mostly added in software, without new branches, new call-centre agents or new paperwork. Each extra customer and each extra product therefore adds revenue without adding much cost.

The consequence of this is a very high operating leverage, which is the mechanism behind the striking rise in Nubank’s reported profits. Revenue grew 37% in 2025, from $11.5 billion to $15.8 billion. Operating costs grew by only 12%, from $2.5 billion to $2.8 billion. Because the top line rose three times faster than the cost base, the difference between the two — which is essentially what the company keeps — grew by roughly 46%, from just under $2 billion to nearly $2.9 billion. This pattern of costs growing slower than revenue is likely to continue for several more years, because the three engines reinforce each other: older customers generate more revenue, more customers keep joining, and the cost of serving each of them barely moves. Few companies in the financial sector, and very few of any size, combine these three properties at once — which is what makes Nubank an unusually consumer-centric business inside an industry that historically has not been built around the consumer at all.

2. Nubank faces important credit, execution and currency risks for a bank

A. Nubank faces important headwinds related to credit underwriting, credit cycles and interest rates

The most important thing to understand about any bank is that its profits depend on a bet. Every time Nubank issues a credit card or grants a personal loan, it is betting the customer will pay the money back. Most of the time they do. Sometimes they don’t. When the economy is healthy the losses are small and the interest on the good loans covers them many times over. When things turn bad — a recession, rising unemployment, a jump in living costs — the losses grow faster than the interest income, and a healthy-looking bank can swing from profitable to loss-making in two or three quarters. This is not hypothetical: it happened to Itaú and Bradesco in 2015–16, and to every Brazilian bank at some point in the last forty years. Nubank, despite its technology and its cost advantage, is in the retail banking business, and credit cycles happen to everyone.

The current picture is reassuring at first glance but has a worrying detail in it. At the end of 2025, 6.7% of loans were more than 90 days behind on payments — slightly higher than a year earlier but stable for several quarters — and Nubank holds reserves worth 2.3 times those bad loans, a ratio that looks solid. The worrying detail is in a separate pool called renegotiated loans: loans where the borrower has run into difficulty and the bank has reset the terms — extended the repayment, reduced the instalment, or restructured the interest. This pool now totals $3.2 billion, or nearly 10% of everything Nubank has lent out. Some of these loans will recover. But the share of renegotiated loans that have already fallen more than 90 days behind again has risen from 13% in late 2024 to 17% at the end of 2025, and the reserves against this pool have gone from 2.1 times bad loans to 1.8 times. In plain English: the renegotiated book is rolling into default faster than it used to, and the safety cushion against it is getting smaller.

The second risk is the interest-rate cycle. Nubank earns money in three linked ways. It charges customers a high rate to borrow. It pays customers a lower rate on their deposits. And it invests the difference in Brazilian government bonds, which currently pay very well because the Brazilian central bank rate (Selic) is at 14.75%. If Selic stays high for a long time — which it has done before — the weakest borrowers default at higher rates, and good borrowers start paying down their loans faster to avoid the interest cost. If Selic falls quickly, the government-bond income that Nubank has been earning on its deposits also falls, compressing the margin between what it pays depositors and what it earns on the float. Either direction creates pressure. Worse, competitors including Inter, C6 and Mercado Pago are starting to pay 100% of the interbank rate on their customer deposits, while Nubank currently pays 87–91%. If customers start moving money to the highest payer — and a new protocol called Pix Automático will make this friction-free from 2026 — Nubank’s cost of deposits rises and its profit margin on lending falls.

The third risk sits in the loan book’s composition. Of the $32.7 billion Nubank has lent out, two-thirds is on credit cards and a quarter is in unsecured personal loans — meaning the borrower has put up no collateral. Only 8% is secured against something tangible. Unsecured lending is the most profitable type of credit in a good economy and the most dangerous in a bad one: when a household runs into trouble, unsecured loans are the first they stop paying, because the lender has nothing to repossess. Nubank is growing this book fast — unsecured loans up 55% in 2025 against a credit-card book up 32%. Credit-loss provisions already rose from $3.2bn in 2024 to $4.2bn in 2025 (+33%) — faster than revenue growth. A Brazilian recession that pushed unemployment up by three or four percentage points would likely double that cost and wipe out most of the annual profit. This is the biggest reason to size any investment position conservatively.

B. Nubank international expansion requires great execution on many aspects

Nubank already reaches roughly seven out of every ten adults in Brazil. The big question for the next decade is whether the company can repeat the Brazilian success in Mexico, Colombia and eventually the United States. If the answer is yes, the customer base can roughly double and the revenue can compound for another ten years. If the answer is no, growth will slow sharply once Brazilian cohorts mature, and the current valuation will look expensive. Expansion abroad is therefore the single most important lever for the long-term story — and it is also the hardest thing the company has ever attempted. Running a digital bank in your home market, where the founders grew up, speak the language, know the regulator, and built relationships over a decade, is very different from doing the same thing in a foreign country.

Start with Mexico, which is the most advanced of the three foreign markets. Nubank launched there in 2019 and has signed up 14 million customers, about 15% of Mexican adults — a fast start but still a small fraction of the 90 million adult population. The Mexican subsidiary is not yet profitable: it is still investing heavily in customer acquisition and technology, with losses that show up in the consolidated accounts. A full banking license was approved by the Mexican regulator in April 2025, which unlocks the ability to take retail deposits directly — the foundation on which any profitable bank rests. But activating a banking license takes time: the Brazilian business took roughly three years from its own license date to turn profitable, and Mexico is a more competitive market with established players such as BBVA, Banorte and Citibanamex, as well as home-grown digital rivals such as Klar and Stori. Colombia, launched in 2020, has 4.2 million customers — just 11% of adults — and is at a much earlier stage. In both markets, the central question is whether the referral-driven growth that worked in Brazil works equally well when the brand is younger and the incumbent banks are reacting faster.

Then there is the United States, where Nubank received conditional approval to operate a national bank in January 2026. The US is a completely different commercial environment. Customers are not underserved — the opposite, they have access to hundreds of banks, credit-card issuers and well-funded digital competitors. Chime, Cash App, SoFi and Robinhood together serve more than 200 million US accounts and have been building the same kind of app-first experience for over a decade. The payments infrastructure that underpins the Brazilian success — the instant-payments rail Pix, which makes transfers free and immediate — does not exist in the US, where interchange fees on cards are far higher but the payments landscape is dominated by established card networks. Winning share in the US will require years of investment, a differentiated product that US customers actually need, and enormous patience. Nubank’s CEO has been clear that US expansion is a long-term project rather than a near-term profit driver.

The base rate for this kind of expansion is not encouraging. Latin American consumer champions have repeatedly struggled when they have tried to cross borders. Mercado Pago, the payments arm of Mercado Libre and the most natural analog to Nubank, has had mixed results outside Argentina and Brazil. Revolut, the European digital bank often compared to Nubank, has spent five years and hundreds of millions of dollars trying to build market share in the US with modest results, and has applied for an American banking license at the beginning of 2026. Kavak and Rappi, two other LatAm scale-ups, have had to close or shrink operations in several foreign markets. The pattern is consistent: something about the specific combination of home-market dynamics, regulatory relationships and customer trust rarely transfers cleanly. None of this means Nubank will fail abroad — Mexico in particular shows promising early metrics — but it does mean that the international expansion required for the long-term growth story to hold is an execution challenge with a meaningful probability of taking longer, costing more, or delivering less than current forecasts assume.

C. FX-sensitivity is an important concern for Nubank

There is something simple but very important that is easy to miss when looking at Nubank’s reported numbers. Nubank does virtually all of its business in Brazilian reais, Mexican pesos and Colombian pesos — but it reports its results in US dollars, because it is listed in New York. This means every line of the income statement, every item on the balance sheet, has to be converted from local currency into dollars at a constantly shifting exchange rate. When the Brazilian real is strong against the dollar, the same operating performance looks better in dollar terms. When the real is weak, the same performance looks worse. None of this has anything to do with whether Nubank is gaining customers, charging the right prices, or controlling its costs. It is a purely mechanical translation effect, and over a few years it can swamp the underlying operating story.

The Brazilian real is one of the most volatile major currencies in the world. Over the last ten years it has traded as strong as 3.8 reais per dollar and as weak as 5.7 reais per dollar — a 50% move between the two extremes. In 2024 alone, the real fell around 27% against the US dollar, hitting a record weak level by the end of that year. That single move shaved hundreds of millions of dollars off Nubank’s reported book value — not because the business lost money, but because the same pile of Brazilian reais was worth fewer US dollars at year-end. You can see the effect directly in the accounts: a negative $998 million currency-translation impact on 2024 equity, followed by a positive $667 million effect in 2025 as the real partially recovered. Over two years, translation alone moved the company’s reported dollar book value by more than $1.6 billion — comparable to an entire year of profit — without a single customer doing anything different.

The same dynamic affects the income statement. When Brazilian profits earned in reais are converted into dollars, a weaker real shrinks them; a stronger real inflates them. The company publishes an adjusted figure it calls “FX-neutral” growth, which strips out this currency effect by applying the same exchange rate to both years. The gap between reported dollar growth and FX-neutral growth has regularly been ten to twenty percentage points in a single year. For an investor trying to work out how fast the business is really growing, this is a substantial complication. The underlying Brazilian business may be growing at 40% in reais while the reported US-dollar result shows 25% — or the other way around, depending on the year. Neither of the two numbers is wrong, but they are measuring very different things.

For a long-term investor, there are two ways to think about this. First, currency volatility creates short-term noise. A quarter where the real sells off hard can make earnings look like they are flatlining even if the underlying business is accelerating; the reverse is also true, and can make a quarter look better than it really is. Second — and more importantly — persistent Brazilian real weakness matters. Brazil’s currency tends to weaken over time against the dollar because Brazilian inflation is structurally higher than US inflation: the real has lost roughly half its value against the dollar over the last fifteen years. If that pattern continues, the reported US-dollar earnings power of even a successful Brazilian operation grows more slowly than the local-currency figures would suggest. In practical terms, an investor buying Nubank shares is making two bets at once: that the company will continue to grow its local business, and that the Brazilian real will not collapse permanently against the dollar. Both bets have historically paid off more often than not, but neither is guaranteed, and the currency exposure is the largest factor that is entirely outside management’s control.

And Nubank doesn’t hedge against the currency it operates in. Hedging the real against the dollar costs roughly the interest-rate differential between the two currencies — around 10% a year at current rates — which on Nubank’s $11.3bn equity base would eat over a third of annual profits to stabilise a non-cash accounting line. Besides, anyone buying a Brazilian bank listed in New York is knowingly taking on real exposure; removing it would be paternalistic and value-destroying.

3. Nubank is an impressive business whose cash flows are likely to grow at a strong pace

A. Nubank is a profitable and growing machine whose “moat” is unlikely to be threatened by traditional banking players

Nubank today is not just a fast-growing company — it is a very profitable one. In 2025 it generated $2.87 billion of net profit on $15.8 billion of revenue, roughly $900 million more than a year earlier, and earned a 30% return on shareholders’ money — a level that fewer than one in fifty publicly listed banks anywhere in the world achieves sustainably. The really striking figure, however, is the one already mentioned in Section 1.c: the share of revenue eaten up by running costs (what investors call the cost-to-income ratio) stood at 19.9% at the end of 2025. To put this number in perspective, it helps to look at how it compares with the best banks in Europe — the part of the world that has been modernising its banking system the longest, and where the gap between online-native banks and their traditional parents is the clearest.

Consider three country examples. In Italy, Fineco Bank — widely regarded as one of the best-run online banks in Europe — has a cost-to-income ratio of around 37%, against roughly 38% for its much larger former parent, UniCredit. In France, BoursoBank (the online bank of Société Générale and by most measures the fastest-growing retail bank in the country) runs at around 50%, while its parent Société Générale sits at roughly 70% — one of the least efficient in Europe. In Spain, Bankinter, an unusually well-managed mid-sized bank that has moved most of its business online, runs at around 38%, against roughly 40% at BBVA. These best-in-class European operators — after decades of cost-cutting, IT modernisation and branch closures — have got down to the mid- to high-30s. Nubank is at half that level. Out of every dollar a Nubank customer pays the company, roughly eighty cents is available to cover credit losses, taxes and profit, compared with sixty cents at the most efficient European incumbent and barely thirty cents at the least efficient. Across 131 million customers, that difference compounds into several billion dollars of additional profit power every year.

The obvious question is whether a traditional bank could close the gap. The answer, on any reasonable timescale, is no — and the reason is structural. Traditional banks are locked into three costs Nubank simply does not carry. The first is branches: Itaú and Bradesco each operate between 3,000 and 4,000 physical locations, every one of which has a long lease, a full staff, and regulatory obligations that make closing them slow and expensive. The second is legacy technology: the core software systems most large banks run on were built in the 1970s and 1980s in languages few engineers still learn, and replacing them takes a decade and carries a meaningful risk of breaking things along the way. The third is people: European experience shows that turning a 60% cost-to-income bank into a 40% one takes 10–15 years of steady restructuring, mass redundancies and protracted union negotiations. A traditional bank trying to match Nubank’s 20% today would need to reduce its workforce by something like two-thirds — a socially and politically impossible project. This is why the efficiency gap, once opened, tends to widen rather than narrow.

The combination of rapid customer growth, maturing revenue per customer, and a cost structure that traditional banks cannot match means the profit machine is likely to keep compounding for years. Analysts forecast earnings of roughly $4.4 billion in 2026 and $6.2 billion in 2027, and a reasonable base case has 2028 profits around $8 billion — roughly three times the 2025 level. This is not a growth rate pulled from thin air: it is arithmetic arising from today’s 131 million customers spending more each year while costs barely move. For this picture to be wrong, something would have to damage one of those three engines — credit costs exploding, customer growth stopping, or costs starting to rise sharply. None of those three outcomes is impossible, but none of them is the base case. We expect that Nubank will continue to do what it has been doing since 2021, with the profit stream thickening each year.

B. Who will very likely dominate the future of retail banking alongside Revolut

Step back from the quarterly numbers and ask a bigger question: who will still be running retail banking in twenty years’ time? Retail banking is a network business — the more customers you have, the cheaper it is to serve each one, and the more products you can offer them. In every mature consumer industry, networks of this type tend to concentrate into a handful of global winners: one or two dominant players per region, surrounded by smaller specialists. Think of streaming music (Spotify and Apple Music), ride-hailing (Uber and one local champion in each region), or payments (Visa and Mastercard). Retail banking, long protected from this dynamic by national regulation and physical branches, is now catching up — and the two companies most likely to be the Spotify and Uber of digital banking are Nubank in the Americas and Revolut in Europe.

The case for these two as the future winners rests on three observations. First, scale. Nubank has 131 million clients and Revolut announced in late 2025 that it had crossed around 65 million — together, they have nearly 200 million customers. No other digital bank anywhere in the world has crossed 50 million. Chime in the US has around 22 million; Monzo in the UK around 12 million; N26 in Germany around 8 million. Scale matters because the product economics only truly work when fixed costs (engineering teams, compliance systems, fraud models, the app itself) are spread across tens of millions of accounts. Second, product breadth. Both companies have moved from a single product (credit card in Nubank’s case, FX in Revolut’s) to a full financial supermarket: savings, loans, investments, insurance, crypto, mobile telephony, travel. Customers increasingly use one app for the whole of their financial life, which is exactly the experience traditional banks have never delivered. Third, technology: both companies have built proprietary engineering stacks with artificial intelligence running the risk, support and fraud systems. The traditional banks in their home markets are still replatforming legacy software; by the time they finish, Nubank and Revolut will be several generations ahead.

The geographical split between these two is also a feature rather than a bug. Nubank operates in Brazil, Mexico and Colombia and is approved to start in the United States; Revolut operates across Europe, the UK, Australia, parts of Asia and Latin America. Their territories barely overlap. Each is building scale in markets where the other is largely absent, and each benefits from the other’s existence: Revolut’s success in Europe makes it easier for Nubank to argue that digital banks can reach tens of millions of customers outside their home country, and vice versa. Neither is a direct competitor to the other today, and given how large the global retail-banking market is — billions of people holding accounts worth trillions of dollars — there is no meaningful reason why both cannot prosper simultaneously. The historic pattern of global consumer industries suggests two or three winners is what the market supports, and the two already ahead are these two.

What this means for the long-term picture is that Nubank is unlikely to be dethroned in its home region by any of the well-known competitors. Traditional Brazilian banks cannot match its cost structure; smaller digital rivals such as Inter, C6 and Mercado Pago are all below half of Nubank’s customer base and growing more slowly in net terms; and no global digital bank has shown any interest or ability to enter LatAm successfully. The real competition — and the real question — is whether Nubank can export the model abroad fast enough to keep the long-term growth rate high. That is the theme of Section 2.2, and it is the one genuine threat to the dominance argument. But within the Americas, the likelihood that Nubank is still the dominant digital retail bank in 2035 is high. Revolut’s parallel story in Europe suggests the economics of this industry reward the first company to reach true scale, and that is exactly where Nubank sits today.

C. Nubank valuation is acceptable for such a quality business, yet the margin of safety is not large

The hardest question, as with any high-quality business, is the price. As of mid-April 2026 Nubank trades at around $14.98, giving the company a stock-market value of roughly $73 billion. The most common yardstick — the price-to-earnings ratio — puts Nubank at about 23 times last year’s profit and 16 times what analysts expect it to earn this year. By comparison, Brazilian incumbents Itaú and Bradesco trade at 8–10 times current earnings. On that simple measure Nubank costs roughly twice as much per dollar of profit as its closest local competitors — not a bargain.

A second measure is the price relative to the assets behind it. European banks typically trade at 1–1.5 times their book value; the best-run (Fineco, Bankinter) at around 2 times; Brazilian banks at 1–2 times. Nubank trades at roughly 7 times book value — the single strongest traditional argument against calling the stock cheap. The defence is that Nubank earns 30% a year on that book, meaning the book itself compounds at 30% without raising new capital. In three years today’s $11bn equity should be worth roughly $24bn, and the 7x multiple on today’s book becomes closer to 3x on the book that will exist in 2028. Still not cheap, but priced for continued excellence rather than a bubble.

The more useful way to value Nubank is on what it will earn three years from now. Taking the base-case from 3.a — revenue growth of roughly 25% a year, costs growing slower, Mexico profitable by 2027 — lands around $7.5–8bn of profit in 2028. At today’s $73bn that is 9–10x the 2028 profit — attractive for a business this profitable and this fast-growing. A fair multiple for a franchise of this quality is 14–16x, which implies a 2028 price around $22 per share, or 45% upside over three years — a high-teens annual return.

A second reference point, and arguably the most relevant one, is Revolut. In November 2025, Revolut completed a secondary share sale at a $75 billion valuation — up from $45 billion a year earlier — with buyers including Coatue, Greenoaks, Andreessen Horowitz and Nvidia. In the same year Revolut generated $6.0 billion of revenue (up 46%) and $1.7 billion of net profit (up 70%) across roughly 65 million customers. Nubank, at a near-identical $73 billion market value, generated $15.8 billion of revenue (up 37%) and $2.87 billion of net profit (up 46%) across 131 million customers. In other words, Nubank has twice Revolut’s customers, more than twice its revenue and 70% more profit, for essentially the same price tag. Per dollar of profit Revolut trades at roughly 44 times earnings against Nubank’s 23 times; per dollar of revenue, 12.5 times against 4.6 times. Revolut is growing revenue slightly faster and profit growth has been higher off a smaller base, but it does not grow twice as fast, and it is private — which means its quoted valuation is negotiated between a handful of buyers rather than tested every day by public markets. On every like-for-like comparison, Nubank is materially cheaper than Revolut — and the two businesses are as close as any two comparable companies in digital banking can reasonably be.

The uncomfortable part of all this is that the base case assumes a lot goes right. A disciplined investor does not buy at fair value; they buy at a meaningful discount to it — a cushion against the unavoidable uncertainty of the future. For a business combining growth execution risk with international expansion, that cushion should be 25–30%. At $15 against a $22 fair value, the margin of safety is only around 30% — right at the edge of what we aim for, and it shrinks to nearly zero on a probability-weighted basis (because the bear case of a $9 share price is real). A move to $12 — the kind of swing Nubank has made twice in the last eighteen months — would open a 40% discount and make this a great investment.

Conclusion: Nubank’s competitive advantages make it a great long-term quality compounder, yet entry price discipline remains important

Nubank combines three engines that very few businesses — in financial services or anywhere else — manage to run at the same time: customers who pay the company more every year they stay, a customer base still growing by fifteen to twenty million a year, and a cost structure so lean that each extra customer barely adds a cent to the running bill. The result is a business earning 30% on its shareholders’ money, growing profits at nearly 50% a year, and structurally impossible for traditional Brazilian banks to match. The risks are real — a bad credit cycle, a stumble in the Mexico and US expansion, a sharp move in the Brazilian real — but none of them is the base case. The price tells its own story: at 23 times last year’s earnings Nubank is expensive against Brazilian incumbents, materially cheaper than Revolut, and fairly valued on its likely 2028 earnings power. That leaves a margin of safety that is adequate but not generous. For the patient investor the approach is simple: own this business when the market offers it at $12, not at $15. The business will still be excellent then; the entry price will be much better.