What Happened to Synapse? The BaaS Collapse That Froze $219M

Synapse Financial Technologies was a banking-as-a-service middleware company that connected fintech apps to small US banks. It filed for Chapter 11 on 22 April 2024, a planned sale to TabaPay collapsed in May, and Synapse shut off its services around 11 May. Because only Synapse's ledger recorded which end user owned which dollar in the banks' pooled accounts, the banks froze the accounts, and more than 100,000 people lost access to about $219 million for months. The trustee later found that end users were owed more than the banks could locate, a gap put at $65–96 million.

No bank failed, so deposit insurance paid nothing. Synapse is the defining failure of the US partner-bank model, and it changed how fintechs are built.

What did Synapse do?

Sankaet Pathak founded Synapse in San Francisco in 2014 to give software developers an API to a bank. At the time a fintech that wanted to offer an account or a debit card had to negotiate with a bank, connect to its core system and a card processor, build compliance processes and keep its own ledger. That took a year or more. Synapse did it once and sold it as a platform. Andreessen Horowitz led its $33 million Series B in 2019 (as reported).

The structure had three layers:

The banks held end users' money mostly in pooled "for benefit of" (FBO) accounts in the name of Synapse or a fintech. The bank's records showed the big pooled balance. Only Synapse's ledger showed each person's share. Synapse also ran cash-sweep products through an affiliated broker-dealer that moved balances among several banks, and between program and operating accounts.

Synapse earned platform and per-account fees from fintechs, transaction fees, and a share of interchange and deposit income passed through from the banks. The amounts were never disclosed. The book's study of fintech, neobanks and embedded finance explains why small banks wanted this business: banks under $10 billion of assets are exempt from the Durbin cap on debit interchange, so they could earn more per swipe and share it.

Why did Synapse go bankrupt?

Synapse was a volume business with concentrated customers, and its margins were squeezed from both sides. Banks wanted more of the economics and more control as regulators tightened. Large fintechs wanted to cut out the middleman.

The largest of them did. Mercury, a business-banking app for start-ups, had launched on Evolve through Synapse in 2019 and was described as Synapse's largest customer. In 2023 Mercury moved to a direct relationship with Evolve, completed by October. Synapse and Evolve fell into disputes over records and balances. When rates rose in 2022–23, new fintech programs dried up with venture funding.

On 22 April 2024 Synapse filed for Chapter 11 with a stalking-horse sale to TabaPay for about $9.7 million. The bankruptcy estate had little else to sell, because end users' money was never Synapse's asset; it sat at the banks.

How did the accounts get frozen?

TabaPay withdrew in May 2024, each side blaming the other. Synapse had no money to keep running and cut off services to its partners around 11 May. Without the ledger, the banks could not tell who owned what in the pooled accounts, so they froze them.

Yotta, a prize-linked savings app, alone had about 85,000 users frozen (as reported). Over the following months the banks reconciled what they could and made partial payouts. The trustee's estimates of the unexplained gap settled at $65–96 million. In June 2024 she said the source of the shortfall was "not known at this time" (as reported by Banking Dive). Some users recovered a small fraction: 13,725 Yotta depositors were offered $11.8 million of $64.9 million in November 2024, as reported.

Why didn't FDIC insurance protect Synapse customers?

FDIC deposit insurance protects depositors against the failure of an insured bank. It does not protect against a non-bank intermediary losing track of money. Pass-through insurance, which extends cover to the beneficial owners of pooled accounts, works only if the records identify each owner and amount. In the Synapse case those records were the very thing in dispute.

Many apps had told users their money was "FDIC-insured" without explaining that difference. California's financial regulator fined Yotta $1 million for such claims on 15 May 2026. The trust customers placed in the apps rested on a guarantee that did not cover the risk they were actually exposed to.

What failed: technology, controls or incentives?

The book's chapter on financial plumbing argues that in finance "the ledger is the product." A software company that loses a customer's photo has a service problem. A financial company that records a balance wrongly has created or destroyed money. Good practice is to reconcile every end user's balance every day, three ways: the fintech's ledger, the bank's records of the pooled account, and the processors' records of payments in flight.

Synapse failed that test in three ways, according to the book's reading of the trustee reports and regulators' actions:

1. The ledger of record sat outside the banks. The institutions legally responsible for the deposits could not say who owned them.
2. Money moved between banks. Sweep arrangements spread balances across four institutions and between program and operating accounts, and nobody reconciled end users' balances across all of them.
3. Responsibility was split. Each party could point at another. Evolve said Synapse's ledgers were inaccurate; Synapse blamed Evolve and Mercury; Mercury called Synapse's claims meritless (as reported). Who caused the shortfall was never established.

The culture was that of a fast-growing start-up selling to other start-ups, where the speed of launching programs was the product. Regulators reached the opposite conclusion about what mattered.

What was the BaaS crackdown?

Synapse came at the end of a two-year wave of enforcement against the small banks that powered US banking as a service. Blue Ridge Bank (OCC, 2022), Cross River Bank (FDIC, published April 2023), Sutton Bank and Piermont Bank (FDIC, February 2024), Lineage Bank (FDIC, early 2024), Thread Bank (FDIC, 2024) and Evolve (Federal Reserve, June 2024) all faced orders over fintech oversight, anti-money-laundering controls or third-party risk. The agencies' June 2023 guidance on third-party risk said plainly that a bank remains fully responsible for compliance performed by its partners.

The CFPB sued Synapse in 2025 for failing to keep adequate records of consumer funds. A stipulated judgment approved in September 2025 imposed a $1 civil penalty, a device that opened the CFPB's Civil Penalty Fund to victims. As of mid-2026 the claims process had not opened, according to a California regulator's notice, as reported.

What changed after Synapse?

The industry's answer was fewer layers. Sponsor banks such as Column, Lead Bank and Cross River built their own developer platforms so that fintechs connect directly and the bank keeps the end-user ledger. Surviving middleware firms moved toward designs in which the bank holds the authoritative records. In September 2024 the FDIC proposed a rule requiring banks to keep records identifying the owners of custodial accounts with transactional features; it had not been finalized at last check.

Fintechs drew their own lessons. Mercury, which left Synapse in time, won preliminary conditional approval for its own national bank charter in April 2026. Chime, which always contracted directly with its two partner banks and used no middleware layer, agreed in September 2026 to buy one of them, Stride Bank, for $590 million in cash, pending approval. The general direction is toward owning, or at least verifying, the ledger. Our comparison of neobanks and traditional banks covers the charter decision.

Which factor did Synapse fail on?

In money-equation terms, Synapse failed on the first factor, trust, and specifically on whether the trust was deserved. End users trusted apps, the apps trusted Synapse, and the banks trusted Synapse's ledger. None of that trust rested on a reconciled record. The book's Chapter 24 on building a financial company states the lesson for founders: "we don't hold the money" is not the same as "we have no responsibility for the money." The company that owns the customer will be blamed for any gap.

It is also a case of distribution economics outrunning controls. Synapse's value was speed of launch, a third-factor advantage, but the compliance and reconciliation work that should have grown with the number of programs and accounts did not. The book's company profile of Synapse sets out the full record, including the open questions.

The takeaway

Synapse did not lose money in a market crash or a bad loan book. It lost track of other people's money, in a structure where no single party owned the whole record, and deposit insurance turned out not to cover that risk. The fix the industry adopted, putting the ledger back inside the bank and reconciling every balance daily, is a reminder that in finance the record of who owns what is the product. For how this fits into the economics of fintech and banking, see our guide to how financial companies make money and our guide to fintech unit economics. The book covers the case in depth in Chapter 12, "The Plumbing: Core Systems, Ledgers and Open Banking".