StoneCo just reported a credit portfolio that more than doubled to BRL 3.8 billion, and the company's own numbers show the risks growing almost as fast as the loans themselves.

The Brazilian fintech, which trades on the Nasdaq under the ticker STNE, is betting that it can transform from a payments processor into a full-service bank for small and mid-size business owners. The August 13 quarterly report laid out both sides of that bet with unusual clarity: credit revenues up 153%, retail deposits up 22%, and a government-backed loan program already at BRL 334 million. Also in the report: loans past due by more than 90 days nearly doubled, climbing from 4.67% of the portfolio a year ago to 8.60% today.

The business they are building

The core argument StoneCo is making to investors is that payments and credit reinforce each other. When a merchant runs their sales through Stone's terminal, the company can see their revenue in real time, which makes it easier to underwrite a working capital loan. That logic drove most of the portfolio's expansion this quarter.

The newer government-backed loan program adds a different wrinkle. Because the Brazilian government absorbs part of any default loss, StoneCo can offer loans to clients it would have turned away under normal underwriting. The trade-off is lower interest rates on those loans, but also meaningfully lower risk. The BRL 334.2 million already disbursed through that program represents a growing share of the portfolio designed to offset some of the pressure building elsewhere.

On the funding side, retail deposits climbing to BRL 10.8 billion actually helps. More deposits mean the company borrows less from the wholesale market, and its funding cost has dropped to roughly 85% of Brazil's benchmark short-term rate. That gap, even if it sounds technical, translates directly into margin.

Where it hurts

Brazil's interest rates are the context that makes every number here harder. When borrowing costs are high for a sustained stretch, the businesses that needed credit most urgently in late 2024 and early 2025 are the ones most likely to fall behind now. That is the pattern StoneCo is watching play out inside its own book.

The coverage ratio, the reserve the company holds against future loan losses compared to what is already overdue, fell from 279.9% a year ago to 203.6% today. Management said that partly reflects a deliberate shift toward better-rated and government-backed loans, and partly reflects the normal timing lag between loans going bad and write-offs hitting the books. Both explanations are plausible. Neither makes the direction of travel look comfortable.

The most concrete example of the risk is an BRL 11 million default from a large, long-standing client that filed for bankruptcy protection this quarter, which management described as unexpected. There is also a separate exposure tied to a liquidated card issuer, where the final provisioning cost depends on how a pending legal matter resolves.

The company's cost of risk, the percentage of the portfolio it sets aside to cover expected losses, rose from 20.2% to 21.5%. For a business growing this fast, a 1.3 percentage point move in that figure is not trivial.

The bigger tension

What StoneCo is navigating is a version of the same problem that faces any lender scaling quickly in a high-rate environment: growth that looks impressive in the revenue line creates hidden pressure in the credit quality line, and the two columns do not peak at the same time. Revenue from a loan booked in 2025 shows up almost immediately. The default, if it comes, often arrives 12 to 18 months later.

That timing gap is why the quarterly report can show credit revenue up 153% and 90-day delinquencies nearly doubled in the same document without either number being wrong. The question that matters for anyone watching this company is whether the loan vintages being written right now are better-structured than the ones now rolling into delinquency. Management says yes. The data to confirm that will arrive over the next several quarters.