Who actually holds your money? The four questions your app won't answer
Open two banking apps side by side and they look identical: a balance, a card, instant payments, a clean feed. What the interface never shows is the thing that matters most the day something breaks — who is actually holding your money, and what you are owed if they vanish. One app might be a chartered bank with deposit insurance. The next is a contractual claim against an e-money firm. The next depends on a sponsor bank's ledger you have never heard of. The next is a private key that is entirely your problem. Same screen, four different answers.
That gap has been the subject of a sharp discussion this week around the neobankbeat dataset, kicked off by Dr. Efi Pylarinou and pushed further in the comments. This is our attempt to answer it with the data.
The headline: most "banks" aren't
Of the 379 neobanks we track, 132 hold a full banking license — their own charter, holding your deposits, usually inside a deposit-insurance scheme. The other 247 are not banks in the sense that word implies on the app store.
The 192 in the middle are where the interface does the most hiding. 81 run on the partner-bank ("rent-a-bank") model, where a sponsor bank you never chose holds the balance and keeps the ledger. 48 are e-money institutions, where your funds are safeguarded but not insured — a different, weaker promise. 13 are payment institutions; 24 hold crypto or VASP licenses. And 55 hold nothing at all: they are self-custodial, and the money lives on-chain in a wallet only you control. You can see the split on the licence and structure pages.
The four questions
The clearest framing in the thread came from Dmytro Nasyrov, who argued that every provider should make four facts visible; Efi's reply was that they belong in front of the customer, visually. We agree. Here they are as a checklist you can run on any app before you trust it with a paycheck:
- Who legally holds the funds? A chartered bank, a sponsor bank behind the brand, an e-money issuer — or you.
- Who keeps the authoritative ledger? The record that decides whose dollars are whose. In modular banking the app, the program manager, the issuer and the ledger provider can all be four different companies.
- What protection applies if an intermediary fails? Deposit insurance, e-money safeguarding, an unsecured claim in a bankruptcy — or nothing but your own keys.
- How do you recover access during resolution? This is the gap between a covered deposit and a queue number in a filing.
Why it isn't academic
These providers don't fail like FTX, with a bang and a headline. They fail quietly: the app stops updating, support goes silent, and — as Anjli Amin put it in the thread — you learn that "you're not a depositor, you're an unsecured creditor waiting in line."
In July 2026 alone, five neobanks dropped off our active list — each broken at a different link in the chain:
- Juno (US) — its banking-as-a-service middleware (Synapse) collapsed; funds frozen.
- Kard (FR) — its e-money provider terminated the contract.
- Fi Money (IN) — its partner bank walked away; roughly 3.5 million customers redirected overnight.
- Z1 (BR) — absorbed by NG.CASH; the brand is gone.
- Pomelo (US) — acquired by Zepz; product paused indefinitely.
Synapse in 2024 wasn't an outlier. It was the rent-a-bank model working exactly as built — right up until the ledger that said whose dollars were whose stopped reconciling. Every removal above is on the public record in the changelog; the pattern is the point of why neobanks die.
Self-custody is a different animal
One correction from the thread is worth keeping, from Panagiotis Kriaris: a self-custodial wallet is not a neobank sitting on a sponsor bank, and lumping the two together flattens the risk. Self-custody removes the counterparty and intermediary risk of the rent-a-bank model — no one can freeze what no one holds — but it replaces that with a different set: private-key management, smart-contract risk, and no consumer-protection backstop. It isn't safer or riskier in the abstract. It's a different question, which is exactly why we keep custody and regulation_type as separate fields.
It's also where the market is heading. Among neobanks founded in the 2020s, 30% are self-custodial. In the 2010s cohort, it was 5%. A category that barely existed a decade ago is now nearly a third of new launches — carrying a risk model most customers have never had to think about.
Read the chain yourself
Every profile on neobankbeat already carries the two facts the interface hides: how the product is regulated (its own licence, a partner bank, e-money, a VASP registration, and so on) and whether it is custodial or self-custodial. Start with the licence breakdown, read what actually happens when a neobank shuts down, or pull the raw regulation_type and custody fields straight from data.json.
What we can't do is wait for the apps to start volunteering this. So the next thing we're building is a per-app money map — those four questions, answered visually on every profile, the way Efi argued they should be. It'll land in the changelog when it's ready. Until then, the money chain is something you have to read for yourself — and now you know where to look.
Cite it
Free to use, including commercially. Building on the licence and custody data, or think a number is wrong? The whole dataset is public — corrections welcome.
Counts reflect the open dataset as of August 2026 and change as neobanks launch, relicense and die. Regulation and custody are compiled from public registers, filings and product documentation; sources are on each profile page. Nothing here is financial advice, and "deposit insurance" coverage varies by scheme, balance and jurisdiction — confirm your own before you rely on it.