I must admit that when I read the news about Zillow (NASDAQ:Z, NASDAQ:ZG) shutting its home-buying business, I immediately thought of Upstart Holdings (NASDAQ:UPST) stock.

Both businesses use artificial intelligence to help them make better decisions.

Unfortunately, AI did not help Zillow avoid a fiasco at its iBuyer service. 

Fortunately, the two use AI differently. As a result, Upstart remains a very attractive long-term hold for anyone looking to bet on fintech. Here’s why. 

UPST Stock Is up 1,500% Since IPO   

We’re still more than a month away from Upstart’s one-year IPO anniversary as I write this. It’s hard to believe that the AI lending platform went public at $20 less than a year ago, at the bottom of its pre-IPO pricing range. 

At the end of August, I included UPST in a list of seven AI stocks to buy. It’s up 54% in less than three months. Yet, I’ve got a funny feeling that the California-based company is only getting started. It trades at about $335 today.

But before I get to why I think Upstart could be a $1,000 stock, let me explain how Upstart is different from Zillow. 

GeekWire reported on Nov. 3 that the company shut its home-buying unit, putting 2,000 people out of work, and causing it to write down $304 million in the third quarter as a result.

While Zillow co-founder and CEO Rich Barton could have tweaked the business model to allow for mistakes made by the algorithm, Barton realized that it could only do that for so long before the markets realized its iBuying service was based on a flawed house buying program. 

“But what we can’t solve is what the model is going to tell us about how much capital we need to raise, deploy and risk in the future in order to achieve a scale that we think is necessary to offer a fair price to customers for their homes in a competitive way,” GeekWire reported Barton saying.

Translation: the real estate market is highly fluid and constantly changing.

“All the AI and machine learning in the world isn’t yet up to the task of the complexity of valuing a home in a rapidly changing market, and this move by Zillow is proof,” MoxiWorks CEO York Baur told GeekWire. “They invented computer home valuation with the Zestimate 15 years ago, and it’s still not accurate after 15 years of refinement and billions of dollars invested.”

At the end of the day, Zillow was overpaying for a lot of homes. That adds up quickly. Barton didn’t want to sink the entire business to prove he and his management team were right about Zillow Offers. 

So, he shut it down, making a tough but smart call. 

AI and UPST Stock

InvestorPlace’s Mark Hake recently discussed Upstart’s business model, suggesting that despite being considered an AI lending platform, it is a marketing company, making most of its revenue from referral and platform fees.

Only 10% is from servicing fees, which are fees paid by lenders for Upstart to manage the collection of loan payments, etc.

One way to understand the amount of risk Upstart carries on its balance sheet is to look at pg. 27 of its Q2 2021 10-Q. For the first six months of the year, it purchased $3.4 billion in loans and immediately resold them.

As a result, less than 10% of its assets are actual loans — $82.3 million in the first six months of 2021 — with 30% of these loans held for sale. 

So, as my colleague points out, Upstart holds very little of the credit risk.

Unlike Zillow, Upstart uses machine learning and AI to fine-tune the perfect borrower. The more data in, the better it can come up with a borrower profile that reduces the loan risk for lenders. 

The most significant risk for Upstart is that AI paints the wrong picture for its lender partners, and they experience higher loan defaults. However, that’s unlikely to happen because more data equals a more thorough understanding of a potential borrower’s creditworthiness. It’s a FICO score on steroids.  

Another risk, mentioned by the company on pg. 82 of its 10-Q in the risks section is that it generates fees from just one loan product — unsecured personal loans. 

However, I view that as a potential benefit to owning UPST stock. That’s because if it ever comes up with a second or third product, the revenues will grow exponentially.

The Bottom Line

In the second quarter, Upstart’s revenue increased 1,018% to $194 million. More importantly, it generated an operating profit of $36.3 million, up from an operating loss of $11.4 million. That’s good for an operating margin of 18.7%. Expect that to continue moving higher in the quarters to come.

Now, back to the Zillow/Upstart comparison. 

Zillow was using AI to figure out how much to pay for houses in different markets. But, as I said earlier, that’s a tough gig given how people’s plans and ideas can change quickly. 

Meanwhile, Upstart is using AI to refine the customer acquisition process for lenders. That’s a service rather than an investment. As such, I would call its business model “asset-light.” 

And that’s an excellent thing. 

On the date of publication, Will Ashworth did not have (either directly or indirectly) any positions in the securities mentioned in this article. The opinions expressed in this article are those of the writer, subject to the InvestorPlace.com Publishing Guidelines.

Will Ashworth has written about investments full-time since 2008. Publications where he’s appeared include InvestorPlace, The Motley Fool Canada, Investopedia, Kiplinger, and several others in both the U.S. and Canada. He particularly enjoys creating model portfolios that stand the test of time. He lives in Halifax, Nova Scotia.