Community banks added $482 billion of deposits while stablecoins grew

Deposits at community banks grew about 26%, or $482 billion, between June 2019 and March 2026, a window a Fortune commentary says spans the entire rise of stablecoins ^[4]. It is the only observed number in the argument. That implies a June 2019 base near $1.85 trillion ^[15] and a compounded pace of roughly 3.5% a year over those six and three-quarter years ^[16].

The exposure figure works differently. The American Bankers Association, citing an April 2025 Treasury Borrowing Advisory Committee estimate, puts as much as $6.6 trillion of transactional deposits theoretically exposed to stablecoin migration ^[2]. An April 2025 Better Markets report counted roughly $2.5 trillion held collectively by banks with less than $10 billion in individual assets, a total that had barely moved in three decades while the largest banks grew ^[1]. The exposed pool is about 2.6 times what the whole small-bank cohort holds ^[17], so the estimate has to be a system-wide count. The two figures also count different things, transactional deposits on one side and the cohort's entire balance on the other.

"The fear isn't irrational, but the data so far don't support it," the Fortune piece wrote ^[7]. It pointed to work by CRA International and the Council of Economic Advisers that found no statistically significant relationship between stablecoin growth and community bank deposit outflows over the period ^[5].

The loss the piece does describe is harder to find on a balance sheet. A business leaves its cash at the community bank and buys payments, foreign exchange, merchant services and treasury management somewhere else. The outside platform collects the transaction data, the fee revenue and the daily customer contact ^[11]. Mercury says it serves more than 300,000 businesses and individuals ^[8]. A ten-person startup that builds its financial operations on a platform today is an expensive, disruptive migration in a decade ^[14].

Nacha reported 33.6 billion ACH payments worth $86.2 trillion in 2024 ^[9], an average of about $2,565 per payment ^[18]. That is the transaction size a commercial customer actually runs. The rail already clears it. The commentary's build advice is to buy or partner for the basic infrastructure and keep compliance, liquidity, lending, data and payment routing in house ^[12]. It describes stablecoins as strongest on always-available connectivity across open networks, especially cross-border, and tokenized deposits as adding faster settlement and software controls inside a familiar bank liability ^[19]. Engineers put on a proprietary chain are not working on custody and routing, and the GENIUS Act has already given banks a federal framework to partner under ^[10].

The case against this reading is about yield. The ABA is lobbying Congress to close what it calls a yield loophole in stablecoin rules ^[3], which means the no-outflow finding was measured under the rule the ABA wants changed. If a payment token pays interest and sits in the same wallet as the payment, the switch is cheaper than it was for money-market funds and brokered CDs. Those out-yielded checking accounts for decades without emptying them ^[13]. The other way I would be wrong is aggregation. A national 26% ^[4] can hide a distribution in which the banks nearest crypto payroll and remittance corridors are already losing balances, and the CRA and CEA work as described does not report that cut ^[5].