The crypto-neobank market is on the verge of a massive shakeout. My analysis shows that the vast majority of these services will not last more than a year and a half. The issue is not a lack of demand — recent data on a record $245 million in weekly deposits proves the opposite. The root of the problem lies in the rotten compliance architecture built into the foundation of most projects.

I have been studying compliance infrastructure since 2017 and see how the market rests on several technical layers. The failure of any one of them can instantly collapse the entire business. Take card programs: these are three-way relationships between the service, the issuer, and the payment network. Enhanced scrutiny from the network is layered on top of the issuer's requirements. As soon as any party loses trust in the system, cards stop working not gradually, but instantly.

The Binance example is telling: in July 2023, the exchange lost Visa in Europe, and two months later, Mastercard in Latin America. Another service, Ready, in June 2026 gave users outside the EEA just one hour to react when its relationship with the issuer was severed. The payment network requires continuous, not periodic, screening against sanctions lists at the individual account level. A transaction monitoring system with real alerts and a secure KYC level across all operating jurisdictions is needed. The £29 million fine for Starling Bank over a system that didn't generate a single individual sanctions alert for six months is a stark price for such a mistake.

I will separately highlight the on/off ramp layer. Each such step is first a compliance event, and only then a matter of convenience. On-ramp checks the source of funds and identity, while off-ramp requires a clean audit trail with a documented source of funds. It is on the off-ramp process that teams most often cut corners, because users don't see it. But regulators do. During a quarterly review, the banking partner takes a sample of transactions, and if the trail isn't clean for every transaction, the company finds out at the worst possible moment.

Yield Products and Compliance: A Deadly Combination

The most underestimated layer is yield products (earn). Shared vaults are perceived as a product solution, when in reality they are a legal solution. Commingling user funds in a single pool creates fiduciary risks and complicates the situation in bankruptcy. The "killer question" for such a structure is simple: show the ledger entry for a specific user's balance. If the answer requires reconstruction from the pool's accounting, then there is essentially no answer.

The best practice is to isolate each user's account from day one: a separate ledger entry and individually calculated yield. Only such a structure withstands scrutiny from a banking partner, regulator, or institutional investor. In Europe, Article 50 of the MiCA regulation prohibits accruing interest on euro-denominated stablecoins, so the legal path lies through tokenized government bonds and "RWA wrappers."

The foundation of the entire structure is the screening layer. Most teams build it reactively — when something breaks or a regulator asks a question. This is the wrong approach. Those who survive build compliance proactively: before the card, before the ramp, and before the yield product, with a single end-to-end audit trail. Disparate point solutions create gaps: the KYC system doesn't communicate with transaction monitoring, and that doesn't pass data to sanctions screening. Each such gap turns into a reconciliation problem at the worst possible time.

My professional opinion: The $245 million in deposits is not a story about cards or yield, but a story about survival. Compliance is not the last thing you build; it is the only thing that allows you to build everything else. The market has already seen this lesson with WaveCrest, Wirecard, FTX, Binance, and Ready. The question is only who will learn it this time.