Nu Holdings (NYSE: NU) The Toll Booth of Latin American Banking – Now On Sale

The setup in a sentence: the best banking fintech on the planet just printed its first-ever $5bn revenue quarter, crossed 135 million customers, earns a 29% return on equity – and the stock trades roughly a third below where it started the year. I have wanted to write up Nubank for a long time. Fintech sits right in my circle of competence, and Nu Holdings is, in my view, the single most impressive franchise the sector has produced.

Here is the short version of the thesis. Nu is a digital-only bank that has effectively become the largest financial institution in Brazil, the third largest in Mexico, and is now scaling in Colombia – while generating real GAAP profits, not the adjusted-EBITDA fiction so common in growth land. It acquires customers for a few dollars through word of mouth, monetises them more every quarter, and runs on a cost base that traditional Latin American banks cannot dream of matching. That is a classic compounding flywheel: more customers × more revenue per customer × a scalable platform. When those three multiply together, earnings can grow for years.

The opportunity exists because the market is fixated on one quarter of rising credit provisions and a self-declared “investment year.” My Fair-PE framework says the franchise is worth roughly 24× a conservative 2028 earnings number, which points to a fair value near $37 and an annualised return of about 43% over the next three years from today’s ~$12.80. I am not buying a cheap bank; I am buying a world-class compounder at a fair-to-cheap price. This is not investment advice – I simply want to lay out the case as I see it.

1. Product, Business Model, Brand & Moat

What the company does

Scroll through the Nubank app and it feels a little like an N26 – strikingly modern, purple, clean. But the comparison flatters N26 more than Nu. Nubank started in 2013 with a single no-fee purple credit card and has since stacked on a full financial life: bank accounts, debit, lending, investments, insurance, a marketplace, even crypto and SME banking. The vision is explicit – own the customer’s entire financial life so the basket of products (and the revenue per customer) keeps growing.

How they make money

Nu earns the way a bank earns, but with a tech company’s cost structure. The two engines are net interest income (lending – credit cards and personal loans – funded by very cheap deposits) and fee/interchange income. The decisive number is ARPAC, average revenue per active customer, which now sits around $16 and has expanded sequentially every single quarter since the company began disclosing it. Crucially, monthly cost-to-serve has stayed near one dollar. That spread between rising ARPAC and a flat cost-to-serve is the whole machine.

Brand

This is where Nu separates from the European neobanks. For most of my N26 friends, N26 is the cheap second account; for tens of millions of Brazilians, Nubank is the primary bank. Net Promoter Scores are world-class, roughly half of customers open the app every single day, and that love is what drives customer acquisition costs down to a handful of dollars. They essentially grow by referral – the brand does the marketing. A January 2026 multi-year partnership with the Mercedes-AMG PETRONAS F1 team is the kind of move that only makes sense once you are playing for global mind-share.

Moat

The moat is the combination, not any single piece: a low-cost digital model that legacy banks with branch networks structurally cannot copy; a deposit franchise that funds lending cheaply; a proprietary AI credit stack (internally branded “nuFormer”) now running real-time credit decisioning in Brazil and Mexico; and the data advantage that comes from 135 million customers. New-versus-old is the oldest play in the book – better cost base, better product, a deposit base the incumbents cannot match. Replacing Nu would mean rebuilding all of that at once.

2. The Market

Latin America is one of the most attractive banking markets on earth precisely because the incumbents have been so bad for so long. Accounts were expensive, fees were punishing, and a meaningful share of the population was underbanked. That is a very different competitive backdrop from Germany, where plenty of free online banks already exist. Nu turned that gap into a land grab: it now serves over 115 million customers in Brazil (the largest private financial institution in the country), more than 15 million in Mexico (the third largest), and is accelerating in Colombia. Those three countries alone represent around 60% of Latin American GDP – and there is a long runway in the rest of the region and other emerging markets where the same playbook should travel.

Two nuances matter. First, Nu is not just banking the very poor; its customer income distribution increasingly resembles the traditional banks, which means a much larger wallet to capture. Second – and this is the honest caveat – a bank is cyclical. Brazilian interest rates have been very high; as they normalise, that should ultimately help loan demand and stabilise credit costs, but the timing is a macro variable I do not control. As for political intervention, LatAm carries the usual emerging-market risk, yet the region is comparatively commodity-rich and arguably less exposed to some of the macro cross-currents weighing on other parts of the world right now. Net-net: enormous, underpenetrated, structurally growing – with a cyclical wrapper I respect rather than ignore.

3. Culture & Management

Nu is founder-led, which is exactly what I look for. David Vélez co-founded the company and still runs it as CEO, and the culture reads as genuinely long-term and product-obsessed rather than quarter-to-quarter. The clearest tell is how they talk about credit: management states plainly that the goal is not to minimise non-performing loans at a point in time but to optimise for resilient net present value. That is the language of operators who think in lifetime value, not headline metrics – and it is rare.

Capital discipline backs up the talk. The efficiency ratio – the share of revenue eaten by costs – has collapsed from roughly 78% in late 2021 to a record-low 17.6% in Q1 2026, even as they invest heavily in AI and international expansion. They are explicitly rebuilding the bank around AI rather than bolting it on; engineering throughput is reportedly up over 50% year on year. A founder who compounds the business while driving the cost base structurally lower is the profile I want to own.

4. Financials

The numbers are what moved this from a watchlist name to a position. Growth never broke after the post-pandemic normalisation that hollowed out so many growth stocks – revenue and earnings have kept compounding, and the company is profitable on a GAAP basis, which materially lowers the downside in my eyes.

Margins tell the quality story. Q1 2026 net interest income hit a record $3.25bn with a net interest margin of 21.1%; ROE printed 29%; the efficiency ratio dropped below 18%. The headline that spooked the market was a $1.79bn credit-loss allowance, up about a third quarter on quarter, which compressed the risk-adjusted NIM to roughly 9.5%. Management was clear that this reflects seasonality, portfolio growth and mix – the credit book grew about 40% year on year to $37.2bn – rather than a break in underwriting. Late-stage (90-plus-day) NPLs actually eased to 6.5%, below the 7% peak of late 2024. In other words, the “scary” number is the cost side of a deliberately faster-growing, still-healthy loan book.

5. Valuation – The Fair-PE Model

I value compounders the same way every time: estimate a defensible earnings number a few years out, apply a Fair PE that reflects quality, and check the implied annual return. The beauty of Nu is that, despite being a hyper-grower, it can be valued on profitability – it actually earns money. On my quality scorecard (product, moat, market, management, financials) this lands very close to a 9 out of 10. For a franchise of that calibre, a Fair PE of 24× is reasonable – it sits below the trailing multiple the market happily paid not long ago, yet well above a generic bank, which is appropriate given the growth and returns.

Consensus 2028 EPS sits near $1.49. I think estimates are, if anything, slightly conservative given ARPAC expansion and the Mexico/US optionality, so my base case uses $1.56 in 2028e EPS × a Fair PE of 24 ≈ $37 fair value. Against today’s ~$12.80, that is roughly a triple over three years, or about 43% annualised. The table below frames the range.

Read that carefully: even the bear column does not lose money. That asymmetry – a world-class business where the conservative case still compounds at a healthy clip – is exactly the kind of setup the Fair-PE framework is built to find.

6. Risks – Why the Opportunity Exists

If the business is this good, why is the stock down? Because the market is paying for the next quarter, not the next decade. Three things collided into the recent sell-off:

• Credit provisions. The $1.79bn loan-loss allowance and the resulting hit to risk-adjusted margin read as a “profit miss” to headline-driven traders, even though it accompanies a fast-growing, still-performing book.

• The “investment year” framing. The CFO openly called 2026 an investment year – return-to-office costs, AI and GPU spend, and international build-out are an 80–100bp drag on the efficiency ratio. The Street dislikes near-term margin give-back, even when it funds future growth.

• Sentiment and flows. LatAm, the US-dollar-funded growth complex, and fintech broadly have been out of fashion; insider Form 144 filings and a couple of analyst target trims (BofA, UBS) added fuel. The stock is down sharply from its 52-week high near $19.

The genuine fundamental risks I am underwriting: a real deterioration in Brazilian or Mexican consumer credit (if NPLs run ahead of what the AI models predict, provisions bite into profit); execution in Mexico, where a multi-billion-dollar bet must convert savers into profitable borrowers; FX and emerging-market macro; and US regulatory hurdles on the new charter. This is a tech-enabled bank, not a frictionless software business – it carries a cyclical credit cycle, and I size the position accordingly.

7. Latest News & the Q1 2026 Call

The Q1 2026 print on 14 May was, underneath the share-price reaction, a record quarter: revenue over $5bn for the first time, net income of $871m (up 41% FX-neutral), 135m+ customers, 29% ROE, and an efficiency ratio of 17.6%. Mexico reached break-even and became the third-largest financial institution in that market – management called it the same earnings-generating formula that built Brazil, now hitting its inflection point. The stock initially dropped nearly 9% after hours on the provision number before trimming losses during the call.

The bigger strategic news landed in January 2026: Nu received conditional approval from the US Office of the Comptroller of the Currency to form a de novo national bank, Nubank, N.A., based in McLean, Virginia, with Cristina Junqueira as US CEO and Roberto Campos Neto as chairman. It still needs FDIC and Federal Reserve sign-off and must capitalise and open within regulatory timelines, with a launch targeted around 2027 – initially aimed at cross-border remittances and the 60-million-plus US Hispanic population. On the chatter side, sentiment on Twitter/X and Reddit has swung between “premium valuation finally cracking” and “gift at these prices” – which is usually what a good entry point sounds like.

One housekeeping correction worth making, since I have seen it repeated: Warren Buffett’s Berkshire Hathaway is no longer a shareholder. Berkshire built a position around the 2021 IPO but fully exited between Q3 2024 and Q1 2025. The “Buffett owns it below today’s price” line is out of date – the thesis has to stand on the business, not on a famous coattail.

8. Conclusion & Investment Case

Nu Holdings is a world-class franchise – there are not many of these – trading at a fair-to-cheap price because the market is discounting one investment year. I am not betting on a heroic re-rating from an expensive base; I am buying quality at roughly 24× a conservative forward number, with a base case of about 43% annualised over three years and a bear case that still compounds.

The catalysts that can close the gap between price and value:

• Brazil rate normalisation: as high Brazilian rates eventually ease, loan demand and credit-cost stability should improve the earnings trajectory.

• US expansion: the OCC-approved national bank charter opens the world’s largest consumer-banking market, targeting remittances and the US Hispanic segment from ~2027.

• Mexico & Colombia scaling: Mexico just hit break-even at 15m customers; turning that base profitable is a major earnings lever.

• ARPAC expansion: every additional product (insurance, investments, lending, marketplace) lifts revenue per customer against a near-flat cost-to-serve.

• Margin proof points: continued sub-20% efficiency once the 2026 investment year laps would force the Street to re-rate the multiple.

For me this earns a first position – a way to compound alongside a founder-led franchise while also adding a slice of Latin American exposure, a region with comparatively few high-quality listed tech names. The valuation is fundamentally fair today; the upside is that I get to grow with the business. As always, I would rather be roughly right about a great company than precisely right about a mediocre one.

Risk Disclaimer & Conflict of Interest

This article reflects my personal opinion and is provided for informational and educational purposes only. It is expressly NOT investment advice, a recommendation, or an invitation to buy or sell any security. Equity investing carries substantial risk, including the total loss of capital; past performance and any forward-looking estimates (including the Fair-PE valuation above) are no guarantee of future results and may prove materially wrong. Always do your own research and consult a licensed advisor before making investment decisions.

Conflict of interest disclosure: Nu Holdings (NU) is held in the portfolio of the cost-efficient Haas Invest4 Innovation Fund (invest4.net), and I and/or the fund may hold a position in the security discussed at the time of publication. I may buy or sell at any time without notice. This is a genuine conflict of interest, which I disclose transparently. Figures are current as of May 2026 and based on company filings, the Q1 2026 earnings call, and public sources.