Upstart, a company that uses artificial intelligence to help approve loans, reported $4.2 billion in loans during the second quarter of 2026. That's a 50% increase from the same period a year earlier, according to results released August 4 on its investor relations site. The company approved 558,014 individual loans, also up 50%. Source: Upstart IR

The fact that both the total dollar amount and the number of loans grew at the same rate tells us something useful. It means the average loan size didn't change much. That's a healthier sign than if the company were simply handing out bigger loans to the same number of people.

Upstart runs an online lending marketplace for personal loans and car loans. Instead of using traditional credit scores, the platform uses machine learning — a type of artificial intelligence that learns from data — to decide who qualifies for a loan and at what interest rate. Source: Reuters

The company says its AI approach works better than the old way of judging credit. In a May 2026 investor disclosure, Upstart said its AI model correctly predicted whether borrowers would repay personal loans 87.4% of the time, beating a traditional credit model on the same test. Source: Upstart IR

Upstart scheduled its earnings call — where executives discuss results with investors — for August 4, 2026 at 1:30 PM PDT. The earnings release went out the day before, August 3.

The broader question is whether Upstart's AI system can stay accurate when it's handling 50% more loans. Think of it like a restaurant that suddenly doubles its customers: the food is only good if the kitchen keeps up quality at the faster pace. The 87.4% accuracy figure offers a reference point, but only if the company measured it the same way as before. Without newer accuracy numbers or data on how many borrowers are falling behind, that question stays open.

What matters next is whether Upstart is actually making more money from all these extra loans, and whether borrowers who got loans recently start missing payments as time passes. The confirmed facts show that loan volume grew. They don't yet tell us whether the company is more profitable or whether borrowers are paying back as expected.

A 50% jump in loans is a lot. But whether that growth is as good as it sounds depends on a few things we can't see yet: what it costs lenders to fund these loans, whether investors want to buy them, and whether the AI can still tell good borrowers from risky ones when reviewing a much larger pool of applicants.

Here's what we can confirm: Upstart approved far more loans, the average loan size held steady, and the company says its AI outperforms traditional methods. Here's what we can't confirm yet: how much money Upstart made, whether borrowers are repaying on time, and whether the business is actually more profitable. Those answers should come from the full earnings report and the company's financial statements.