the graveyard

Nobody dies of churn: what actually killed six neobanks

September 13, 2026 · 5 min read · ← all posts · the dataset

The standard story of a failed consumer app is that people stopped using it. Growth stalls, retention sags, the money runs out. It is a satisfying story because it puts the product at the centre and implies a lesson about building better ones.

It is also, for neobanks, mostly wrong. We keep a permanent archive of every company removed from this directory, with its last verified profile and a traced cause of death. Six have left so far. Here is what actually happened to them.

The six

NeobankLifespanCause
Juno2019–2026rail withdrew — Synapse/Evolve fallout
Fi Money2019–2026rail withdrew — partner bank ended the relationship
Kard2018–2026rail withdrew — e-money provider terminated the contract
Pomelo2021–2026acquired and folded in
Z12020–2026acquired, brand retired
Will Bank2017–2026liquidated by the central bank

Three of six died because the institution underneath them stopped carrying them. Two were bought. One was wound up by a regulator over the collapse of the conglomerate it belonged to. Zero lost their customers first. Fi Money had 3.5 million of them on the day its partner bank walked away.

The rail is the product

The pattern is not subtle once you look for it. A neobank is a brand, an app and a customer relationship sitting on top of somebody else's permission to hold money. When that permission is withdrawn — by a partner bank exiting the business, an e-money provider terminating a contract, a middleware ledger collapsing — the app stops being a bank instantly, no matter how much its customers liked it.

Which is why the licence question is the first thing on every profile here. Of the 381 neobanks we track, 132 hold their own banking licence. Another 82 are explicitly running on a partner bank's charter, and the rest sit on e-money, payment, or crypto-asset permissions of varying durability. That distribution is the industry's real risk map, and it is invisible from the app store.

The 2024 Synapse collapse remains the clearest demonstration: when the middleware between fintechs and their sponsor banks failed, roughly $95M went missing and customers across 50+ apps could not reach their own money for months. Juno's presence in our graveyard is a direct downstream consequence, two years later.

why we archive instead of deletingMost directories quietly remove what dies, which leaves a survivorship-biased view of an industry where failure is the most instructive event. A reader put it better than we would have: why a project fails is more valuable than why one succeeds. So delistings move to the graveyard with the cause attached, and ship in data.json as a structured field for anyone doing failure analysis.

What it means if you are choosing

Not "avoid anything without a licence" — that would rule out most of the interesting products, including nearly every app built for people who moved and every self-custodial wallet. The useful version is narrower: know which link in the chain you are exposed to, and what happens to your balance if that link breaks. A licensed bank fails into deposit insurance and a regulator-managed resolution. A partner-bank app fails into a queue behind a ledger reconciliation. A self-custodial wallet cannot fail that way at all, because nobody is holding anything.

Those are three genuinely different risks wearing one word. Every profile on this site names which one applies, and says so in plain language rather than in a footnote — see the money map for how to read it, or the survival stack for the layer-by-layer version.

go deeperThe graveyard — every delisting, archived with its cause · the changelog — every dataset change in public · data.json — the graveyard array, machine-readable.