America's New Banks Can't Legally Take Your Deposit

Key Takeaways

• The Office of the Comptroller of the Currency (OCC) received 14 new bank charter applications in 2025, nearly as many as the previous four years combined, and the pace accelerated further in early 2026.
• Most of the crypto-industry winners hold a national trust charter, a structure that cannot accept demand deposits, cannot make loans, and carries no Federal Deposit Insurance Corporation (FDIC) coverage.
• Bank of America’s CEO puts up to $6 trillion in deposits at stake in the fight over what these new “banks” are allowed to do next, and Federal Reserve staff have already published the coefficients for what that would do to lending.

The Word “Bank” Is Doing a Lot of Work

American finance is minting new “banks” faster than at any point in years. In 2025 alone, the OCC received 14 de novo charter applications, nearly equaling the total it received in the previous four years combined. By mid-March 2026, barely two and a half months into the year, the agency had already approved four new applications and received more than seven others. On December 12, 2025, the OCC conditionally approved five national trust bank applications in a single batch: two new banks, Ripple National Trust Bank and Circle’s First National Digital Currency Bank, plus conversions for BitGo, Fidelity Digital Assets, and Paxos. The wave is still cresting: on July 10, 2026, Circle received the OCC’s final approval to open its trust bank, which will operate as Circle National Trust.

Every one of those December approvals produced some version of the same headline: crypto firm becomes a bank. Investors read it as legitimacy. Customers read it as safety. The community-bank lobby read it as a “flood” that, in its words, “demands a policy response” from the OCC.

Here is the detail the headlines skip: a national trust bank is a charter that, by design, cannot do the thing you probably think a bank does. The Bank Policy Institute (BPI), the research arm of the large-bank lobby, puts it flatly: national trust companies are “generally uninsured and restricted to the activities of a trust company,” and “cannot take demand deposits or make loans.” No deposits. No loans. No FDIC.

The newest banks in America cannot legally take your deposit. Understanding why that is, and what these companies get out of the label anyway, tells you more about where your money is actually safe in 2026 than any rate-comparison chart.

What a Charter Actually Buys

A full-service national bank charter is a trade. Banks get the right to accept deposits, make loans, and process payments, plus access to what BPI calls “the federal safety net of FDIC insurance and the Federal Reserve discount window.” In exchange they submit to capital and liquidity requirements, activity restrictions, and consumer protection rules, with examiners embedded in their business.

A national trust charter takes a thin slice of that trade. It grants fiduciary powers, custody services, and the prestige of federal supervision, without deposit-taking, lending, or insurance. For a crypto custodian or stablecoin issuer, that slice is exactly what they want: federal preemption of the state-by-state money-transmitter maze, and a national regulator’s stamp for institutional clients. When the site covered Ripple’s conditional approval in December, that logic was the whole story: custody powers and preemption, explicitly without deposit insurance or discount-window access.

Where does that leave your actual cash? It depends entirely on which of five buckets the company holding it sits in:

The top two rows are boring, and boring is the point. The bottom three rows are where the word “bank” starts doing work the law does not back up.

Why Everyone Suddenly Wants the Label

Three policy moves opened the runway in twelve months.

First, the GENIUS Act. Signed on July 18, 2025, it created the first federal regulatory regime for payment stablecoins, requiring 100% reserve backing in liquid assets like dollars and short-term Treasuries, with monthly public disclosures of reserve composition. Legitimacy for the product created demand for a federal home for its issuers.

Second, the OCC finalized a rule clarifying what national trust banks may do, effective April 1, 2026, giving fintechs a cleaner path to charters that combine trust powers with payments-adjacent activities.

Third, a May 19, 2026 executive order directed federal financial regulators to identify rules that “unduly impede” fintech firms from partnering with regulated institutions and to streamline charter application processes. The now-or-never signal could not be clearer, and the application queue shows the industry heard it.

None of this is a scandal on its face. A federal charter with a named regulator beats the offshore alternative, and it beats the fifty-state patchwork it replaces. The problem is narrower and more personal: what the label communicates to you, the person deciding where to park an emergency fund.

The Insurance That Isn’t There

In March 2026, FDIC Chairman Travis Hill removed any ambiguity about the safety net. “The FDIC is planning to propose that payment stablecoins subject to the GENIUS Act are not eligible for pass-through insurance,” he said. The agency followed through: on April 7, 2026, the FDIC board approved a proposed rule stating that “deposits held as reserves backing a payment stablecoin would not be insured to payment stablecoin holders on a pass-through basis.” The GENIUS Act itself already bars issuers from marketing stablecoins as backed by the U.S. government or federally insured.

So the regulatory architecture is at least honest: the tokens are not deposits, and issuers are barred from implying otherwise. But honesty in a statute is not the same as comprehension at the checkout screen. A dollar-pegged balance, held at a company with “national trust bank” in its name, spendable from an app with a debit card, looks and behaves like a checking account right up until the moment it does not.

What that moment looks like is not hypothetical. When fintech middleware firm Synapse went bankrupt in April 2024, tens of thousands of consumers lost access to their money, with roughly $200 million in customer funds frozen and an eventual shortfall estimated at $65 million to $96 million that pass-through insurance never covered, because no insured bank had failed. Customers of the savings app Yotta bore the worst of it: data released by the company’s CEO showed 13,725 former customers lost deposited money in the collapse, refunded just $11.8 million against $64.9 million in deposits. The ledger, not a bank, was the point of failure, and the FDIC insures banks, not ledgers.

Stablecoins have already run their own version of this stress test. On March 11, 2023, USD Coin (USDC) lost its dollar peg after Circle disclosed that $3.3 billion, about 8% of its reserves, was stuck at the collapsing Silicon Valley Bank. The token regained its peg four days later. What changed in between: the Treasury, Federal Reserve, and FDIC jointly announced that SVB “depositors will have access to all of their money,” a guarantee that covered the uninsured balances too. Sit with the irony: the flagship regulated stablecoin was rescued by the exact deposit-insurance system its holders are now formally excluded from.

The $6 Trillion Math Problem

The loudest pushback here is not coming from consumer advocates. It is coming from the incumbent banks, whose objection is conveniently self-interested and still arithmetically real.

Presenting Bank of America’s fourth-quarter 2025 results in January 2026, CEO Brian Moynihan warned that as much as $6 trillion in deposits could migrate into stablecoins under certain regulatory outcomes, above all if stablecoins are allowed to pay interest. Deposits that leave, he argued, either stop funding loans or must be replaced with costlier wholesale funding.

Federal Reserve staff have put coefficients on that mechanism. A December 2025 FEDS Note found that “for each $100 billion of net deposit drain not recycled to banks, empirical pass-throughs imply a $60 to $126 billion contraction in bank lending.” The same analysis flags transaction accounts, the money in everyday checking, as more vulnerable to stablecoin substitution than savings, particularly among younger consumers who transact digitally.

The mechanics compress into one line:

$$\Delta \text{Lending} \approx (0.6 \text{ to } 1.26) \times \Delta \text{Deposits}$$

Run Moynihan’s worst case through the Fed staff’s coefficients and the implied hit to lending lands between roughly $3.6 trillion and $7.6 trillion. That is not a prediction; the entire stablecoin market stood at about $312 billion in June 2026, and it shrank that month. But it explains the lobbying war. A dollar that leaves a checking account for a stablecoin stops funding mortgages and small-business loans and starts funding the Treasury market instead. The community banks that dominate small-town commercial lending lose that dollar first, which is why the tokenization wave keeps colliding with the same trade groups.

The Last Time “Trust” Did the Heavy Lifting

If a lightly regulated institution doing bank-like business under a reassuring name sounds familiar, it should. The parallel is uncomfortably literal, down to the word “trust.” New York’s trust companies were the fintechs of the Gilded Age, and the New York Fed’s own historians describe them in modern terms: “Like shadow banks, trusts were less regulated than banks, less liquid, more levered, and perhaps most importantly, lacked direct access to emergency loans from the Clearing House.” They held cash reserves of around 5 percent of deposits, versus roughly 25 percent at national banks, and the public could not tell the difference until October 1907, when it mattered. Knickerbocker Trust, the second-largest trust in the country, faced a run; J.P. Morgan’s examiners judged its solvency uncertain, Morgan refused to lend, and Knickerbocker closed its doors on October 22. The panic that followed convinced Congress the monetary system needed a permanent fix, and the Federal Reserve Act was signed into law in December 1913.

The 2026 rhyme is not that Circle is Knickerbocker. GENIUS-regulated reserves, all cash and short-term Treasuries, are dramatically higher quality than a 1907 trust’s book. The rhyme is structural: institutions outside the insured perimeter, doing money-like business under a bank-like name, redeemable on demand, with no lender of last resort. In 1907 the runs played out over weeks of queues outside branches. A stablecoin redeems 24 hours a day, 7 days a week, at the speed of an API call. Better collateral, faster exits.

The Honest Path Exists

Here is the counterpoint that keeps this from being a simple morality tale: nothing stops a fintech from becoming a real, insured bank. It is just expensive, slow, and binding.

Varo did it, receiving its full-service national bank charter from the OCC on July 31, 2020. SoFi did it, with the OCC conditionally approving SoFi Bank, N.A. as a full-service national bank in January 2022. Square’s bank took the Utah Industrial Loan Company (ILC) route, winning conditional FDIC approval of deposit insurance in March 2020. And the wave continues: in March 2026, digital lender Upstart applied to the OCC and FDIC to establish an insured national bank, full obligations included.

One cohort built a deposits-and-lending business, and only the heavy charter fits that. The other built a custody-and-tokens business, and the light charter fits that. Both paths are lawful uses of the charter system, and both are rational. The asymmetry is in the branding: every one of these firms gets described as “a bank” in headlines, while the obligations behind the word differ by an order of magnitude. The steelman for the trust-charter firms deserves a fair hearing too: GENIUS-mandated reserves of Treasuries arguably carry less asset risk than a fractional-reserve loan book, and above the $250,000 FDIC ceiling, an uninsured bank deposit and a well-reserved stablecoin start to look more alike than different. The difference is what stands behind them on a bad day: a bank above the ceiling still has the discount window and a resolution process; a token holder has a redemption queue.

Check the Charter, Not the Marketing

For a reader deciding where money should sit, the whole story reduces to three questions, answerable in about 90 seconds:

1. Is the company itself an insured bank? Search the FDIC’s public BankFind database for the entity’s exact legal name. An insured bank appears there; an app that merely partners with banks, and a national trust company, will not.
2. If not, whose insured bank is behind it? Legitimate apps disclose their partner banks in the fine print. That is where pass-through coverage lives, and Synapse is the case study in what happens when the recordkeeping between app and bank breaks.
3. Is it a deposit at all? A stablecoin balance is not a deposit anywhere. Under the FDIC’s proposed GENIUS rules it carries no pass-through insurance, full stop.

The 2026 charter wave will keep producing “crypto firm becomes bank” headlines. The FDIC’s stablecoin rulemaking is open now, and how it finalizes the insurance exclusion will set the fine print for the next decade of money apps. None of that fine print will appear in the app-store screenshots. The word “bank” is now a marketing asset that comes in several legally distinct grades, and the only person with an incentive to check the grade is you.