The hard lesson of the last Banking-as-a-Service cycle was not about solvency. It was about records. When a program breaks, what decides who bears the loss is not whether the bank was insured, but whether anyone can say whose money is whose. That distinction sounds academic until the moment it is tested, and then it becomes the only thing that matters.
A recent early-October 2026 analysis revisited the Synapse failure and the question it left unresolved: who owns the loss. The answer was uncomfortable. Pass-through deposit insurance protects depositors against the failure of a bank, but in that case the money was not lost to a failed bank. It was unaccounted for in a non-bank's records. Because the problem was a reconciliation gap rather than an insolvency, insurance did not make depositors whole. The FDIC's recordkeeping rule, written around exactly that ledger gap, still has not been finalized.
That is the part worth sitting with. In a bank-fintech program, the ledger is not a compliance formality that sits quietly behind the product. It is what determines, when something goes wrong, which end customer is owed what. If the record cannot establish beneficial ownership down to each person, no insurance policy and no contract can reconstruct it after the fact. The protection is the record itself.
This reframes what diligence should actually weigh. When a platform evaluates a banking partner, the questions that tend to dominate are speed, pricing, and breadth of features. The more durable questions sit underneath:
- Does the institution hold its own record of each beneficial owner, or does it depend on a third party to reconstruct one?
- Is reconciliation performed continuously, down to the end-user level, or only periodically?
- Are client funds segregated and independently recorded, so ownership is identifiable at any moment?
- When an intermediary fails, can the bank still identify whose money is whose without that partner's cooperation?
For platforms and the banks behind them, those answers matter more than any single feature. This is the discipline behind ACTC's regulated trust infrastructure: account segregation, independent recordkeeping, and reconciliation that keep each owner's funds identifiable, so ownership can be established when it matters, not reconstructed once it is too late.
The lesson of the last cycle is not that insurance failed. It performed exactly as designed. The gap was the record beneath it. The question is no longer whether a BaaS program is insured. It is whether, when something breaks, its records can still say whose money is whose.
.png)