Upstart Grew 42%, Earned a GAAP Profit, and the Stock Still Sits Near a 52-Week Low

Upstart Holdings just turned in the kind of quarter a growth-stock owner dreams about. Revenue rose 42% to $365 million in the June quarter, fee revenue was up 45%, and the company posted GAAP net income of $16.5 million — up 195% from a year earlier — on $0.16 a share. Contribution profit hit a record $193 million, originations grew 50% to $4.2 billion, and the core personal-loan business, the engine that actually drives this model, re-accelerated 27% quarter over quarter.

The stock trades around $25, down roughly 43% this year and over 60% in the past twelve months, a few dollars above its 52-week low and barely a third of its high. A $2.4 billion-market-cap fintech in the Russell 2000 is being sold as if the growth already ended. When the numbers and the price are reading two different companies, one of them is wrong. The whole argument is which one.

The fear is real, and it's in the price

Investors didn't invent the worry. In July, after a 50%-growth quarter, originations grew just 34% year over year — a visible step down. Asked to look ahead, management held its full-year guidance at about $1.4 billion of revenue and $294 million of adjusted EBITDA, pointing to a macro gauge it calls the uncertainty market index, which had risen to 1.5%, the top of its guided range. A company that just beat refusing to raise is the classic signal that the next few months look worse, and Wall Street read it that way.

There's also a long memory at work. Upstart is still roughly 87% below where it traded five years ago, and the scars from the 2021 collapse never really healed. In February, the company stopped giving quarterly guidance, guided its adjusted EBITDA margin down a notch to 21%, and swapped in a new CEO — a combination the market treated as an information vacuum and sold accordingly. When that gets layered on top of worries that lending partners are paying thinner take rates and that consumer credit is turning soft, you get a beaten-down ticker with a plausible bear case at every step.

The earnings knock is real, but it's the wrong denominator

The valuation bears cite is easy to find: on trailing GAAP earnings the stock looks expensive — roughly 40 times earnings against a consumer-finance peer group that trades in the low teens. That number gets quoted because it scares people.

It's also not measuring the right thing. GAAP earnings per share are depressed right now by a fixable problem, not a broken one. The newer secured products — auto and home loans — still lose money, carrying a combined negative 35% contribution margin even after recovering from negative 176% a year earlier; management guides them to breakeven by the fourth quarter. When near-term earnings are held down by a ramp management has already scheduled out of the numbers, trailing P/E is a poor guide to forward earnings power.

The cleaner math is on the company's own 2026 targets. Full-year adjusted EBITDA of about $294 million against an enterprise value in the neighborhood of $4 billion works out to roughly 13.6 times forward adjusted EBITDA. Enterprise value is about 2.9 times the guided $1.4 billion of revenue, or about two times trailing sales. A business compounding revenue above 40% trading near 14 times forward EBITDA and under three times forward sales is not a growth multiple. It is a fear multiple — the market has already paid for the bear case and then some.

One number decides the read

That leaves the entire thesis hanging on whether the July step-down is structural or cyclical. The evidence that it's cyclical is specific. The uncertainty index is a macro gauge, not a sign that Upstart's underwriting broke. Core personal loans — the product that defines the business — re-accelerated 27% sequentially in the same report that spooked the market. Institutions, meanwhile, are betting the opposite way the stock is: Upstart closed its largest funding deal ever, adding up to $5 billion in committed capacity, and priced $569 million of asset-backed debt at its tightest spreads in three years. Loans it keeps on its own balance sheet fell to just 5.9% of outstanding, the lowest in nearly two years, and regulators gave conditional approval for a company-owned bank charter.

None of that means the discount is automatically wrong, and the honest caveat is that this isn't a one-off sentiment dislocation. The whole AI-lending complex has been de-rated — SoFi, a far larger fintech, is down about 34% this year and still trades near five times trailing sales, more than double Upstart's two times. If consumer credit enters a genuine downturn, both deserve lower multiples, and July's 34% growth number is exactly how such a downturn would start to show up. The difference is that Upstart's price has already front-loaded that outcome, while its reported quarters keep pointing the other way.

At roughly 13.6 times forward adjusted EBITDA against a business that just re-accelerated to 42% growth with a GAAP profit, the market has priced the collapse, not the compounding. The debate now turns on one variable — whether July was a macro pause or the first crack in the credit edge. If it was a pause, this is the kind of gap between the math and the multiple worth watching on a small-cap watchlist. If it wasn't, the discount is exactly what a fading underwriting edge deserves. The numbers pick the side; only the next few months of originations can confirm it.