Industry

Industry — Understanding the Playing Field

1. Industry in One Page

Latin American retail banking is a deposit-funded consumer-credit business wrapped in a payments and fee layer. Five incumbents in Brazil, Mexico and Colombia have historically held 68%–80% of all loans and deposits in each market and earned high spreads because branch-heavy distribution kept new entrants out and tens of millions of adults underserved. Smartphone penetration, instant-payment rails (Pix, CoDi, Bre-B) and Open Finance reset the entry cost: a mobile-first balance sheet can underwrite, fund and service customers at a fraction of the legacy cost. The retail financial-services revenue pool across Brazil, Mexico and Colombia was $227.9 billion in 2025 (interest income + fees, net of funding), growing 7% reported and 11% on an FX-neutral basis. This is not a payments app: interest income on credit-card and loan balances funded by sticky retail deposits is where 80%+ of the money is made. Payments and fees buy engagement and data; spread on the lending book is the profit engine.

Takeaway: profits stack across the value chain, but credit underwriting on top of cheap deposits is where banks earn their spread; everything else either feeds that engine with data or monetizes it with fees.

2. How This Industry Makes Money

Retail banks earn a net interest margin — the gap between what they charge on credit-card balances and loans (often 30%–400% APR in Brazil) and what they pay for funding (deposits, wholesale debt, capital). Out of that gap come expected credit losses (provisions for loans that won't be repaid), transactional costs (interchange paid to networks, fraud), and operating expenses (branches, staff, technology, marketing). What's left is pre-tax income. A second leg — fee and commission income — is generated by interchange on debit/prepaid spend, brokerage and asset management, insurance distribution, and increasingly subscription-style products. For Nu specifically, interest income was 85% of 2025 revenue and fees 15%; that ratio is broadly representative of digitally-led retail banks that lend.

Two structural cost lines dominate any retail bank: funding cost (about 29% of Nu's 2025 revenue) and credit losses (about 27%). Combined, they consume roughly half of every revenue dollar before a single branch or engineer is paid. This is why "cost of funding" and "cost of risk" are the two metrics that distinguish a great bank from an average one — saving a few hundred basis points on either flows almost entirely to pre-tax income.

Bargaining power sits in three places:

• Depositors: in markets with high policy rates (Brazil Selic averaged 14.3% in 2025), depositors can demand close to the interbank rate. A bank that brands itself as the customer's "primary relationship" (rather than a one-time deposit) earns funding below 100% of the interbank rate. Nu cites 81% of CDI in December 2025.
• Card networks and merchants: interchange is set by Mastercard/Visa and paid by merchants. Banks distribute the cards and receive the bulk of interchange. Caps or reforms here directly hit fee income.
• Regulators: they set capital requirements, interchange rules, and — in Brazil since 2023 — caps on revolving credit-card interest. Each can compress a layer of the margin stack overnight.

Capital intensity is high in the regulatory sense (banks must hold equity against risk-weighted assets) but low in the physical sense for a digital model — branches and back-office headcount are the legacy fixed cost the new entrants are demolishing.

3. Demand, Supply, and the Cycle

Retail-banking demand is shaped by three structural forces and one cyclical one. Structural: smartphone penetration (GSMA forecasts 93% in Latin America by 2030), formal-credit expansion among populations that have been historically unbanked, and migration of payment volume from cash to instant electronic rails. Cyclical: the policy-rate cycle. When central-bank rates rise, banks' funding costs rise immediately while loan repricing lags, compressing spread; rising rates also pull deposits out of low-yield savings into higher-yield products, raising the cost of funding. When rates fall, the dynamic reverses and credit demand picks up.

Sources: 20-F citations of World Bank, BIS, IMF, BCB, Banxico, ABECS data; rates and unbanked figures vary by source date.

The under-penetration is striking: Mexico's credit-card penetration of 11.3% sits four times below Brazil's, and household debt at 17% of GDP is roughly a quarter of the developed-market norm. That gap is the demand backdrop for a generation of credit growth. But it isn't free money — high policy rates raise both opportunity cost for depositors and credit-loss potential for the lender.

Where the cycle hits first in a retail bank is interest expense on deposits (immediate, on every dollar of funding) and expected credit loss provisions (immediate, forward-looking under IFRS 9). Loan-book yield repricing comes later. That is why digital-bank quarterly results in Brazil swing on the trajectory of the Selic and on shifts in NPL formation in unsecured consumer credit; both data points are released monthly by the central bank.

4. Competitive Structure

Brazilian, Mexican and Colombian retail banking is one of the most concentrated banking markets in the world. The Herfindahl–Hirschman Index (a 0–10,000 measure of concentration where higher = more concentrated) sits at 889 in Brazil, 993 in Mexico and 1,262 in Colombia, versus 328 in the US, 373 in Germany and 488 in France. The top five incumbents in each market hold between 68% and 80% of all loans and deposits. That concentration is exactly why disruption has been possible — it created persistent excess profit and customer dissatisfaction that a low-cost digital entrant could attack.

The structure is oligopolistic on the incumbent side and fragmented on the challenger side. Disruption has come from three distinct types of entrants — pure digital banks, payments-led fintechs that built deposit and credit franchises on top of merchant relationships, and e-commerce platforms that bolted financial services onto their user base. Each subtype has a different cost structure and a different right to win.

Two facts about the structure deserve emphasis. First, digital banks' share of total outstanding loans to Brazilian individuals went from below 1% in 2018 to above 8% by September 2025 (and above 14% if mortgages and earmarked rural credit are excluded). That is a structural shift, not a cycle. Second, the incumbent block is not monolithic — Itaú reported a 2025 net interest margin of 3.6%, well below its 2024 4.4%, and incumbent ROEs have come under pressure as digital competitors have peeled off the highest-spread, most-profitable consumer segments first. Concentration is high, but it is also actively contested.

5. Regulation, Technology, and Rules of the Game

Latin American regulators are unusually proactive in trying to engineer competition into a concentrated industry. The clearest cases — Brazil's Pix instant-payment system (October 2020) and Open Finance (rolled out from 2021) — were explicit central-bank decisions to lower switching costs and force interoperability. Pix reached 179.7 million unique users and $6.3 trillion of transaction volume in 2025, a 29% reported increase over 2024 (34% FX-neutral) and the dominant payment rail across consumer and SME use cases. Mexico has copied the playbook with CoDi (2019) and DiMo (2023); Colombia launched Bre-B as its low-value instant-payments system. The economic effect is consistent — cash and cheque float migrate to electronic balances held in mobile-first apps, and the incumbents' interchange and float advantages erode.

Technology is the second rule-changer. The 2010s shift from on-premise core-banking systems to cloud-native ones eliminated the cost moat that incumbent IT systems used to provide. Nu's NuCore platform — proprietary, cloud-based, built in Clojure — illustrates what the marginal entrant can now do. The 2024 acquisition of Hyperplane added an AI layer for personalization that incumbents will struggle to replicate without rebuilding their data foundations. AI- and ML-driven underwriting compounds: more customers generate more data, which sharpens credit decisions, which lowers cost of risk, which funds lower customer-facing prices, which acquires more customers.

6. The Metrics Professionals Watch

Retail banking has its own vocabulary. The metrics below explain value creation in this industry better than headline EPS — beginner analysts should ground a model on these before tracing them up to net income.

The most useful skeptical check on any LatAm digital-bank narrative is to triangulate three of these: (1) cost of funding as % of policy rate, (2) NPL formation in the unsecured book, and (3) ARPAC against cost-to-serve. A bank that is widening its ARPAC–cost spread while keeping cost of funding well below 100% of the interbank rate and NPLs stable is creating real value. A bank that is growing customers fast but with deteriorating funding cost or unsecured-loan NPLs is buying revenue.

7. Where Nu Holdings Ltd. Fits

Nu is a digital-native challenger that has reached scale — large enough by customers to be the #1 private financial institution in Brazil (113 million customers, 62% of adults), the leading new-credit-card issuer in Mexico, and a meaningful entrant in Colombia. It sits on the digital-bank rung of the industry ladder defined above, but with a balance sheet (deposits $41.9B, credit-card receivables $18.3B, loans $9.4B at FY2025) that increasingly looks like an incumbent's by composition, while operating costs look like a tech company's. The US OCC conditional national-bank charter granted on January 29, 2026 marks an intended step outside Latin America, into the US digital-banking market.

8. What to Watch First

A short checklist of observable signals that will tell a reader whether the industry backdrop is improving or deteriorating for Nu, in roughly the order they move:

1. Brazil Selic and cost of funding as % of CDI. Monthly BCB releases. A widening gap between Nu's funding cost and CDI (currently 81%) compresses NIM directly. A narrowing gap is the single best operational signal.
2. NPL 15–90 and NPL 90+ in unsecured Brazilian consumer credit. BCB monthly statistics plus Nu's own quarterly disclosure (3.9% / 6.6% as of December 2025). A meaningful uptick in the 15–90 bucket is the leading indicator of credit-loss provisions one to two quarters ahead.
3. Pix transaction volume and Open Finance call volumes. Both expand the addressable deposit pool and lower switching costs. Quarterly BCB reports.
4. Brazilian and Mexican incumbent bank ROEs and NIMs. Itaú, Bradesco, BBVA México disclosures every quarter. Persistent ROE/NIM compression at incumbents signals share-shift continuing.
5. Digital banks' share of total Brazilian consumer loans. Reported by BCB; currently around 8%, with the trajectory mattering more than the level.
6. Brazilian regulatory developments on revolving credit-card cap, RWA methodology, and Mexican banking license issuance for Nu. Each can reshape spread or balance-sheet capacity in a single notice.
7. Updates on Nu's OCC pre-opening conditions and FDIC insurance application for Nubank, N.A. The US charter is the only catalyst with multi-year TAM implications outside Latin America.

If those seven move favorably together, the industry backdrop supports Nu and the rest of the report can be read with a tailwind. If three or more move against the company in the same quarter — particularly a Selic spike combined with rising NPLs — the industry is working against the franchise, not with it.