Beyond the Banking License: How Neobanks Are Transforming into Financial Operating Systems

Out of 368 neobanks he counts as actively operating this July, barely a third hold a full banking license. The rest are renting their right to exist from someone else's rail, and most of us never check which kind we're holding.

The number behind the number

Andreoli, who runs developer relations at Consensys and MetaMask, has spent six months mapping the field on neobankbeat.com — a project born, he writes, from the fact that "analysts charging $4,000 for PDFs didn't know" how many neobanks actually existed. The headline (368) is less interesting than the shape. He sorts them into 254 traditional challenger banks, 58 fiat-and-crypto hybrid apps, and 56 web3-native self-custody applications. That third bucket has grown fast: 30% of new banks founded in the 2020s and still operating are web3-native, up from just 4% in the previous decade's cohort. Builders have voted with their feet, whatever regulators think of the result.

The scale is also worth pausing on. Summing disclosed user counts, Andreoli lands on roughly 1.46 billion people served — not a projection, the actual numbers each company reports. WeBank alone carries over 400 million users, more than every US and European neobank combined. Nubank reports 131 million. Revolut, the European flagship, sits above 50 million. Geography matters: 817 million of those users are in Asia, a fact Western fintech media still struggles to put on the front page.

Two-thirds of the apps on your phone aren't banks

Here's the friction worth naming out loud. Most of these companies are licensed — but the license is an e-money authorization, a card-program sponsor, or a banking-as-a-service partner you've never heard of. The deposit insurance that protects you when a chartered bank fails doesn't travel with you in the same way when a BaaS intermediary, processor, or sponsor is the thing that cracks. Andreoli lines up the receipts. WaveCrest in 2018: Visa revoked a card issuer's qualification and dozens of crypto card projects died overnight. Wirecard in 2020: a €1.9 billion hole froze funds from European "banks" whose only mistake was being built on top of it. Synapse in 2024: ordinary Americans discovered that "FDIC insured" can mean very little if the ledger recording who owns what is the ledger that broke.

When a chartered bank fails, you get deposit insurance. When a new bank's infrastructure fails, you get a claim number.

What it means for the rest of us

You don't need to take a side on crypto to feel the consumer stakes. The "operating system" framing — a financial app that holds your paycheck, your card, your savings, your brokerage, maybe your identity — is already here. The only question is whether we, as users, can tell which apps own their own rails and which are tenants on someone else's floor. Andreoli's dataset is open, refreshes monthly, and deleted five entities this month alone — liquidated, absorbed, or quietly reshaped, with no press releases attached. The industry doesn't stay still long enough for any single snapshot to age well.

The trust contract is older than any app store. You hand over your salary, your direct debits, your identity verification, and in return you get a login and a push notification. That contract only holds if the thing behind the login is what it claims to be. Right now, two-thirds of the time, it isn't a bank. It's something wearing a bank's clothes, and the seam only shows when the stitching comes loose.