SKYWAY SUMMITInvestments · CoverageNU is Nubank, the digital bank that became the largest private financial institution in Brazil and is now replicating that playbook across Latin America. Q1 2026 was, on the fundamentals, excellent: revenue surpassed $5B for the first time, net income of $871M, 29% ROE, 135M+ customers, and a record-low efficiency ratio below 18%. The stock fell anyway, because revenue missed one consensus line by ~2% and credit provisions rose. But the provision increase was seasonal (Q1 always runs higher) and the 90+ NPL ratio actually improved. The result is a rare setup: a genuinely elite bank (29% ROE is JPMorgan-beating profitability at fintech growth rates) trading near a ~14× forward P/E, an all-time-low multiple, with analysts at a Strong Buy consensus and a ~$19 target. This is a quality compounder the market just put on sale.
The Brazil playbook, now running in three countries
Nubank's model is the same everywhere it goes: enter an underbanked market with a free, no-fee, app-only account; acquire customers at near-zero cost through referrals and word-of-mouth; then cross-sell credit cards, loans, investments, and insurance as trust builds, lifting revenue per customer over time. The genius is the cost structure, with no branches and an AI-driven back end, Nu's efficiency ratio is below 18% (a traditional bank runs 50-60%), which means it can profitably serve customers that incumbents find uneconomic. The result is a compounding machine that gets more profitable as each customer cohort matures.
The reason this matters now is the combination of proven dominance and untapped runway. Brazil alone, where Nu is already the leader, represents a profit pool exceeding $100B in annual gross profit, of which Nu has captured roughly 7%. That single figure frames the opportunity: the company can keep growing for years just by deepening its position in a market it already leads, before counting Mexico (just inflecting) and Colombia (barely started). A future US entry under a national bank charter (targeting the Hispanic/remittance market) is optionality on top.
The two numbers that define the machine: ARPAC and efficiency
Nubank's quality shows up in two metrics moving in opposite, favorable directions: revenue per customer rising while cost-to-serve falls. That gap is the profit engine, and it has widened every quarter since the company began reporting.
ARPAC of ~$16/month sounds small, but multiply by 135M customers and it is a $20B+ annualized revenue run-rate, and it keeps climbing as customers adopt more products (cards, then loans, then investments, then insurance). Meanwhile the efficiency ratio of 17.6% is the structural moat: a traditional Latin American bank spends 50-60 cents to generate a dollar of revenue; Nu spends under 18. That cost gap is what lets Nu profitably bank the underbanked and still post a 29% ROE. As the AI transformation (its NuFormer models, AI underwriting, agentic banking) deepens, management expects the efficiency edge to widen further.
The low-cost model is the moat; AI widens it
The structural advantage is cost-to-serve. Incumbent Latin American banks (Itau, Bradesco, BBVA) carry expensive branch networks and legacy technology; Nu was built cloud-native and branchless from day one. This is not a moat that erodes easily, because the incumbents cannot simply close their branches and match Nu's cost base without abandoning their existing customers and infrastructure. Nu's customer-acquisition cost is a fraction of an incumbent's, its cost-to-serve is a fraction, and those advantages compound as the customer base grows.
Two pieces of leverage matter. Operating leverage: as ARPAC rises and the cost base stays flat (or falls with AI), incremental revenue drops to profit at a very high rate, which is exactly what drove the 29% ROE and the sub-18% efficiency ratio. Geographic leverage: the Brazil playbook is now a repeatable template. Mexico reaching break-even validates that the model travels; each new market follows the same maturation curve (acquire cheap, cross-sell, mature into profit), and Nu is running it in three countries at three different stages simultaneously. The investment is essentially buying the Brazil profit engine today and the Mexico/Colombia engines as free options.
Elite bank profitability at fintech growth rates
The numbers are the cleanest of the four names in this batch. Q1'26 revenue of $5.32B (a record, the first quarter above $5B), net income of $871M (+56% YoY), ROE of 29%, net margin of 44%. A 29% ROE is elite, the kind of profitability associated with the best mature banks in the world, except Nu is posting it while growing revenue 40%+ and adding ~4M customers a quarter. The one blemish, and the reason for the selloff, was a rise in credit provisions: the 15-90 NPL ratio rose 89bps to 5.0%. But that is a normal Q1 seasonal pattern (early-stage delinquencies always peak in Q1), and the more important 90+ NPL ratio actually improved 10bps to 6.5%, well below its Q3'24 peak. The market reacted to the seasonal number and ignored the structural one.
| $M (mgmt) | Q1'25 | Q1'26 ● | FY26E | FY27E |
|---|---|---|---|---|
| Revenue | 3,250 | 5,320 | ~22,000 | ~28,000-30,000 |
| Net income | 557 | 871 | ~4,000 | ~5,500+ |
| ROE | ~27% | 29% | ~28-30% | ~30%+ |
| Customers (M) | ~119 | 135 | ~150 | ~165-175 |
Q1'26 figures from the 14 May 2026 release (managerial basis, FX-neutral where noted). Forward columns illustrative build pending, engine will compute.
Method: an earnings-and-ROE blend. At a ~$16B+ FY26E net-income run-rate and ~4.8B shares, a 20-24× P/E (justified by 40%+ growth and a 29% ROE, still below where the stock traded for most of its life) yields ~$17-21. Cross-check: a 29% ROE sustainably supports a 4-5× price-to-book; at the current book, that brackets ~$18-20. The stock at ~$12.90 trades near a ~14× forward P/E, close to its all-time-low multiple. For a bank compounding at this ROE with this runway, sub-15× is structurally cheap. Blended fair value ~$17-20, matching the Strong-Buy analyst consensus near $19. The risk is Brazil macro (rates, FX) and credit normalization, not the franchise.
Elite ROE, paying a growth-bank multiple
The comparison frames the quality. Nu's 29% ROE exceeds JPMorgan (the gold standard of US banking at ~18%) and the best Brazilian incumbent, Itau (~21%), while growing revenue many times faster than either. Against fellow digital bank SoFi (also covered in this batch, ~9% ROE), Nu is in a different league on profitability. Yet at a ~14× forward P/E it trades at a discount to its own history and not far above the multiples assigned to the slow-growth incumbents it is taking share from. A bank that out-earns JPMorgan on ROE and grows like a fintech should not trade like a utility; that gap is the opportunity.
Strong Buy consensus; the selloff is the disagreement
The sell-side is overwhelmingly constructive: a 21-analyst Strong Buy consensus with an average target near $19, roughly 50% above the current price. Even the bearish actions are mild, UBS cut its target but kept a Buy, trimming from $18.10 to $16.90, still well above the current $12.90. The caution that exists is technical (the stock is below its key moving averages with negative momentum after the post-earnings drop) and macro (Brazil rates, LatAm FX, credit-cycle sensitivity), not fundamental. That is the textbook profile of a quality name handed to patient buyers by short-term traders reacting to a headline miss. The gap between the consensus target (~$19) and the price (~$12.90) is unusually wide for a Strong-Buy-rated profitable compounder.
Three risks, all macro or cyclical rather than structural. (1) Brazil credit cycle, the live risk. Nu is a Brazilian consumer lender at its core; a genuine deterioration in Brazilian consumer credit (beyond the Q1 seasonal bump) would raise provisions, compress ROE, and validate the bears. The Q1 NPL rise was seasonal, but this is the metric to watch every quarter. (2) Brazil macro and FX, the structural-exposure risk. High Brazilian rates, currency depreciation against the dollar (NU reports in USD), and political/regulatory shifts all flow directly into the equity. A US-listed stock with a Brazilian earnings base carries a permanent FX and country-risk layer. (3) Competition and expansion execution, the slow risk. Incumbents (Itau, BBVA) and other fintechs (MercadoLibre) are responding, and the Mexico/Colombia/US expansions consume capital and management attention. None of these threatens the franchise; all of them can pressure the multiple.
Prototype Skyway Coverage brief · facts current to 27 May 2026 (Q1'26 results, segment/geographic detail, NPL trends, analyst actions all sourced). Forward estimates and fair-value derivation illustrative pending the quant engine. Not investment advice. The brain advises, the trader decides.
Company & opportunity first; the Skyway edge framework is context, not the verdict. Prototype. Illustrative figures not yet engine-computed.