What happened, what we can verify, and why it matters for every fintech stack
In May 2024, people who had never heard the word “Synapse” woke up and could not access their money. They were not customers of Synapse in the way most people think about “customers.” They were customers of consumer fintech apps. Those apps, like many in the US, relied on a sponsor bank relationship. Synapse sat in the middle as the middleware layer that helped move data and money between fintech programs and partner banks. When Synapse entered bankruptcy and services began to unwind, ordinary users got caught in a failure mode that almost never shows up in glossy fintech narratives: recordkeeping and reconciliation collapse.
This post is an attempt to document the Synapse scandal using verified, primary sources and reputable reporting. It is not meant to assign blame to any single party. It is meant to understand the mechanism of failure, because the mechanism is not unique to Synapse.
A short timeline, anchored to public records
April 22, 2024: Synapse files for Chapter 11 bankruptcy protection. The CFPB later summarized this date in its enforcement page describing Synapse and the bankruptcy filing.
May 2024: Disruptions spread to end users of fintech apps tied to Synapse and partner banks. AP reported that customers were frozen out of accounts and that the situation involved a “middleman” doing bookkeeping between fintechs and banks.
Summer 2024 onward: The industry and regulators begin treating Synapse as evidence of a systemic gap in custodial account recordkeeping and pass-through deposit insurance readiness. A senior FDIC official later stated Synapse’s failure created “significant hardship for consumers” and highlighted the difficulty of reconstructing balances.
September 17, 2024: The FDIC Board approves a proposed rule to strengthen recordkeeping for deposits received through third-party arrangements. The FDIC press release frames the goal as addressing risks, protecting depositors, and promoting confidence.
September 2025: The CFPB reports it obtained a stipulated final judgment against Synapse, describing specific failures in recordkeeping and mismatches with partner-bank records, including an estimated shortfall.
What exactly is “the Synapse scandal”?
The simplest way to describe it is this: the ecosystem could not quickly and confidently answer, at scale, “who owns what money, and where is it?”
AP described Synapse’s role as a non-bank “middleman” doing the bookkeeping needed “to make sure customer accounts are credited and debited correctly.”
When that bookkeeping layer failed in an already complex partner-bank setup, multiple parties had to reconcile records under crisis conditions. That is slow by definition, especially when there are multiple programs, multiple rails, and multiple ledgers involved.
This is why the Synapse episode felt different from a typical fintech outage. It was not “systems are down, we will be back in a few hours.” It was “records disagree, so we cannot safely release funds until we reconcile.”
The numbers that matter (and where they come from)
Because the case involved bankruptcy proceedings, multiple institutions, and evolving reconciliation work, numbers vary by source and date. Here are the ones that can be tied to credible reporting and official statements.
How many people were affected?
- AP reported an estimate that “frozen accounts” at one bank were “under 200,000,” attributed to a source familiar with the scope, and noted regulators believed the impacted count was in the thousands or tens of thousands (not millions).
- Forbes reported “at least 200,000” individual fintech customers lost access to money following Synapse’s collapse.
- Yale Journal later summarized: more than 100,000 people lost access to over $265 million held across several fintech platforms.
These figures differ because they come from different snapshots and definitions (accounts, people, programs, bank-specific counts). But they all point to the same underlying fact: the impacted population was large enough to become a national policy issue.
How much money was frozen or missing?
- Bloomberg reported roughly $200 million of consumer funds were frozen during the bankruptcy, citing the court-appointed trustee.
- Fortune reported “roughly $200 million” frozen and cited an estimate of “as much as $95 million” missing, attributed to the trustee.
- CFPB stated that partner banks found their held funds were less than Synapse’s records reflected, “a shortfall of between $60 and $90 million,” and that some consumers did not receive the full amount of their balance.
- The CFPB complaint PDF includes the same $60 to $90 million estimate in the narrative of record mismatch and shortfall.
How long did people lose access?
Consumers did not have any access to their funds for weeks or months as the partnering banks reconciled their records with Synapse’s records.
That sentence is one of the most important in the entire episode. It captures the real harm: not volatility, not bad UX, but prolonged loss of access tied to reconciliation of inconsistent records.
What failed, mechanically?
There are many layers to the story, but one failure mode keeps recurring in official statements and credible reporting: ledger mismatch across parties.
The CFPB’s description is unusually direct. It alleges Synapse failed to maintain adequate records of the location of consumer funds and failed to ensure those records matched partner banks’ records, causing consumers to lose access.
At a high level, these arrangements tend to include: a sponsor bank holding funds in custodial structures; one or more fintech apps interfacing with end users; middleware and program layers tracking sub-ledgers and user-level balances; processors and networks generating settlement, dispute, and fee files; operational workflows for returns, chargebacks, reversals, and corrections.
In healthy times, the system can “work” even if it is messy, because mismatches can be resolved gradually. In stress, you need a property that many stacks do not fully have: unwind readiness, meaning the ability to quickly reconstruct beneficial ownership and balances with evidence.
That is why the FDIC’s Director Jonathan McKernan wrote that Synapse’s failure revealed it “might have been time consuming and difficult, perhaps even impossible” for the FDIC to have paid deposit insurance “had one of the partner banks failed.”
This is a key insight: Synapse was a consumer hardship event even without a bank failure. A bank failure would have added deposit insurance timelines and further complexity.
Why regulators reacted the way they did
It is tempting to treat Synapse as a one-off scandal and move on. Regulators did the opposite. They treated it as a demonstration that existing recordkeeping practices in third-party deposit arrangements were not reliably “claim-ready.”
On September 17, 2024, the FDIC approved a proposed rule to strengthen recordkeeping for deposits received through third-party, non-bank arrangements. Reuters summarized the direction: banks would need to identify beneficial owners and balances, and banks would need “unrestricted access” to the data even if a middleman becomes insolvent.
The message is not subtle. Modern fintech stacks are allowed to be modular. They are not allowed to be unverifiable.
A few uncomfortable takeaways for the industry
1) Many “banking” products are, operationally, multi-ledger systems. If multiple parties hold pieces of the truth, reconciliation is not back office work. It is the control plane. Synapse is what happens when reconciliation is forced to happen under crisis conditions, at scale, with consumers watching.
2) “We are not the bank” is not a consumer-friendly concept. AP noted that fintechs “more often than not, are not banks themselves” and rely on smaller institutions as partner banks. That model can be perfectly legitimate. But when recordkeeping fails, the consumer’s lived experience is simple: their money is unavailable.
3) Systemic risk can exist without a bank run. One of the striking parts of Synapse is that it did not require a macro shock. It required a breakdown in a critical infrastructure layer, followed by disagreement over records.
4) The biggest risk is not fraud, it is uncertainty. Fraud is a known category. Uncertainty about ownership and balances is worse because it can force all parties into defensive freezes.
Closing thought: this is not a one-off accident
Synapse should be read as an alarm, not an anomaly. The fintech industry has modularized the stack faster than it has strengthened the operational infrastructure underneath it. That infrastructure is not flashy. It is recordkeeping, reconciliation, exception management, audit trails, and “break-glass” readiness.
The Synapse scandal is a clear example of what happens when the underlying plumbing cannot reliably answer basic questions under stress. The visible symptom was frozen accounts. The root cause was deeper: systems that could not reconcile quickly enough to prove the truth.
If fintech wants to keep scaling safely, the next wave of innovation has to include the boring parts. Starting with reconciliation.
Methodology and sources: This article is based on a review of primary sources, regulator statements, and reputable reporting, prioritizing documents that can be independently verified. Sources include CFPB enforcement action page on Synapse, CFPB complaint PDF (bankruptcy court filing), FDIC press release announcing the proposed recordkeeping rule, FDIC Director statement discussing Synapse’s hardship, Reuters coverage, AP reporting, and Forbes, Bloomberg, Fortune, and Yale Journal reporting.