Neobanks: Why Building a Bank Without Branches Is Still Hard
App only banks promised to disrupt the least loved industry in America. A decade in, the survivors are real businesses, and the graveyard explains what banking actually is.
The Pitch That Launched a Hundred Apps
A neobank is a bank experienced entirely through an app, no branches, no paper, onboarding in minutes. The pitch wrote itself, traditional banks spend fortunes on real estate and legacy systems, monetize confusion through overdraft and minimum balance fees, and ship software that feels a decade old, so a startup with none of those burdens should win on cost and delight. Hundreds launched worldwide through the late 2010s, led by Chime in the US and Revolut and N26 in Europe, plus Nubank\'s spectacular run in Brazil. The customers genuinely came, tens of millions of accounts. The profits, for most, did not, and the reasons are a compact education in what a bank actually is beneath the interface.
The Economics Under the App
Start with how most US neobanks actually earn. The majority are not banks at all, they partner with small chartered institutions, the banking as a service structure this site\'s embedded finance article maps, and their dominant revenue is the interchange covered in our card economics piece, a slice of every debit swipe, available at attractive rates only through a regulatory carve out for small partner banks. The model therefore monetizes spending, not banking, it needs customers who deposit their paycheck and swipe constantly, and it earns roughly nothing from the account balance itself unless it can lend. That is the first hard truth, a checking account is a loss leader everywhere in banking, branches or not. Incumbent banks endure it because the account anchors a lifetime of lending, mortgages, cards, the net interest margin machine our banking coverage explains. A neobank without lending owns the loss leader and not the payoff.
The neobank graveyard has one epitaph, customer acquisition was the easy part. A free account with a slick app acquires users, but users are a cost until the bank holds their primary paycheck and underwrites their credit, and both of those turned out to be trust problems, not software problems.
What Separated the Survivors
The winners, and by 2026 there are genuine ones, Nubank profitable at enormous scale across Latin America, Revolut profitable and hoarding licenses, Chime reaching the public markets, converged on the same corrections. They became lenders, cards, personal loans, overdraft style advances, because lending is where banking\'s actual margin lives, and their data advantage, seeing customers\' real cash flow, gave them underwriting the incumbents\' stale credit files could not match, the same insight powering the open banking fight this site covers. They chased primacy, the direct deposited paycheck, because a secondary spending app churns while a primary account compounds. Many pursued their own charters, accepting the regulatory burden they were founded to dodge, because renting a license caps the model, the middle layer failures of 2024 froze customer funds and taught the market which logos actually held the money. And the geography lesson, the most successful neobanks grew where incumbent banking was worst, Brazil\'s oligopoly pricing gave Nubank an opening American neobanks never had against free checking and ubiquitous branches.
The Scorecard, and the Incumbents\' Counterattack
Honest 2026 accounting, neobanks won features, forced the industry to kill or soften overdraft fees, normalized two day early paychecks and instant notifications, and banked millions the branch system priced out. They did not displace the giants, the largest US banks added customers throughout, copied the features, the app gap that launched the category has narrowed to a sliver, and kept the profitable complexity, mortgages, wealth, commercial banking, that no neobank has cracked. The category\'s meaningful US wins concentrated among lower income customers underserved by fee heavy incumbents, a real social outcome attached to a structurally thin revenue base. The remaining question is the one lending answers, whether cash flow underwriting at neobank scale survives a full credit cycle, something the category, born after 2010, has never actually experienced.
The Bottom Line
Neobanks proved distribution and experience could be rebuilt in software and that the industry\'s worst fees were a choice, but the decade\'s deeper lesson is what resisted disruption, deposits are a trust franchise, lending is the margin, and the license is the moat. The survivors became what they set out to replace, licensed, lending, regulated banks with better software. The revolution succeeded exactly to the extent it stopped being a revolution, which is a lesson worth filing far beyond fintech.