Stripe Capital disbursed $3.8 billion in merchant cash advances and business loans in 2025, across 81,000 individual transactions, generating $420 million in lending revenue against a base of merchants processing payments on Stripe's platform [1]. That single-year figure, combined with disclosed volumes going back to the product's 2019 launch, puts cumulative originations past $12 billion, a total that exceeds the entire outstanding loan book of all but a small number of chartered US banks. The U.S. has approximately 4,500 FDIC-insured commercial banks and savings institutions [4]; the vast majority hold total loan portfolios well below this threshold. Stripe Capital achieved it without a banking licence, without accepting a dollar of retail deposits, and without filing a single call report with the FDIC.

The mechanics of the product matter for understanding why the growth has been rapid and why the regulatory exposure is asymmetric. In the United States, Stripe Capital loans are originated by Celtic Bank, a Utah-chartered industrial bank and FDIC member; merchant cash advances are provided by YouLend, a third-party specialist [5]. Stripe itself acts as the servicer and technology layer: it identifies eligible merchants based on payment processing history, presents pre-approved offers, and collects repayments by withholding a percentage of daily Stripe sales. The merchant sees a single, seamless product inside their Stripe dashboard; the regulated credit activity sits in Celtic Bank's books, not Stripe's. This structure, which mirrors the banking-as-a-service architecture used by dozens of fintech lenders, allows Stripe to scale a credit product without incurring bank capital requirements, without subjecting itself to Community Reinvestment Act obligations, and without the supervisory relationship with the Federal Reserve or OCC that a chartered bank would carry.

The data points that are confirmed are $2.4 billion in 2022 and $3.8 billion in 2025 [1][2]. The intervening years are not publicly disclosed with precision; the estimated series above uses linear interpolation between confirmed anchor points and is presented as an approximation, not a verified figure. What the confirmed bookends establish is a 58% increase in annual volume over three years at a base that was already material, and a cumulative total that passed $12 billion sometime in 2025. By comparison, the SBA's flagship 7(a) programme approved $31.1 billion across roughly 70,000 loans in fiscal year 2024 [3]: Stripe Capital, operating without any government guarantee and without any regulatory mandate to serve underserved communities, originated more than one-tenth of that volume in 2025 alone, with an average advance size in the low tens of thousands of dollars.

The business case for Stripe's merchant base is clear enough from Stripe's own published research: merchants that accepted Capital financing grew revenue 27 percentage points faster over the following year than comparable merchants that declined [6]. The mechanism is not surprising. A small e-commerce business facing a seasonal inventory purchase or a marketing spend opportunity that its working capital cannot cover will grow faster if capital is available at the moment of need. Stripe's structural advantage is that it processes the payments, so it observes the merchant's revenue in real time: the underwriting model is not a static assessment of historical financials but a continuous feed of transaction data that conventional credit bureaus cannot replicate. A bank offering a small business term loan to the same merchant would be working with tax returns and bank statements that are months old. Stripe's model is current to the prior business day.

The historical parallel that frames this most usefully is not PayPal Working Capital or Square Loans, which are operational analogues, but GE Capital, which at its peak in 2007 generated $65.6 billion in revenue and represented 57% of General Electric's operating earnings, all through a non-bank financial structure that had grown faster than regulators had thought to classify [7]. GE Capital was eventually designated a non-bank systemically important financial institution in 2013, which imposed Federal Reserve supervision and capital requirements. GE responded by dismantling the business: it sold or wound down most of GE Capital's assets between 2015 and 2018 because the compliance cost of operating at scale under bank-equivalent supervision eliminated the margin advantage that had justified the non-bank model in the first place. The architecture that produced the growth was also the architecture that could not survive the regulatory response to the growth.

Stripe is not GE Capital; $3.8 billion in annual disbursements is a different order of magnitude from a half-trillion-dollar balance sheet. The comparison is instructive not on scale but on structure: the non-bank arrangement that allows Stripe Capital to operate without deposit insurance, without CRA obligations, and without call report transparency is precisely what the FDIC's 2024 Small Business Lending Survey identified as a gap in its visibility into the small business credit market [4]. Celtic Bank, as the originating institution, files call reports. But the economic exposure, the servicing, the underwriting model, and the customer relationship all sit with Stripe, which files nothing comparable. Whether that structure remains regulatorily stable as the volumes compound is the question that the $12 billion milestone makes genuinely urgent, and it is a question the industry would be unwise to assume has already been answered in its favour.