According to the Consumer Financial Protection Bureau, Upstart's algorithm — a company that helps banks to underwrite loans — approves 27% more applicants than a traditional model, and twice as many "near prime" consumers with FICO scores between 620 and 660. It also yields 16% lower interest rates.
Driving the news: Upstart raised $180 million in an IPO this week, ending trading on Wednesday with a market capitalization of $2.1 billion.
The big picture: Upstart started life as a way of buying equity in people, allowing them to raise cash not by borrowing it, but by pledging a percentage of their future income.
- When that didn't work out, Upstart pivoted to lending, using the same set of data inputs that it had used to underwrite its equity investments. But it was still hard to compete with banks with a much lower cost of funds.
- So Upstart pivoted again, this time selling its underwriting software to banks.
Context: Upstart, which was co-founded by Paul Gu, a Thiel Fellow, has touched on many buzzy ideas over its existence, from income-sharing agreements to Big Data to artificial intelligence. Even its IPO comes at a buzzy time for IPOs.
- The bottom line: Each iteration has been a bit more practical and a bit lower down the stack, less of a revolutionary consumer-facing company and more of a utility for lenders.
- That's still sexy enough for investors, who bid up the stock by 45% on the opening day of trading.
