1. Upstart Holdings
According to Upstart, 80% of Americans have never defaulted on a loan, yet only 48% qualify for the lowest interest rates banks offer. Management believes that discrepancy stems from a lack of data. Specifically, many of the most sophisticated credit models incorporate just 30 variables, so lenders determine loan eligibility and set interest rates based on very little information. In turn, some applicants are mistakenly approved, and others are mistakenly rejected, which means the creditworthy borrowers end up subsidizing those who default.
To solve that problem, Upstart leans on big data and AI. Its platform captures over 1,600 data points per applicant then measures those variables against past repayment events to quantify risk. To that end, internal studies have shown that Upstart's AI can cut loss rates by 75% while keeping approval rates constant, or it can boost approval rates by 173% while keeping loss rates constant. Either way you measure it, that's good news for lenders.
Not surprisingly, Upstart has seen strong demand. Since its IPO in December 2020, the number of banks and credit unions using its technology has tripled. Likewise, revenue skyrocketed 250% to $228 million in the most recent quarter, and net income jumped 200% to $29.1 million. Upstart is well-positioned to maintain that momentum.
In October 2021, the company launched Upstart Auto Retail, an e-commerce platform for car dealerships. It allows consumers to find vehicles, check out online, and access AI-powered auto loans. So far, seven banks and 291 dealerships have adopted the product.
Currently, management puts its market opportunity at $753 billion, a figure that includes all personal loans and auto loans originated in the U.S. over a 12-month period. To put that number in context, Upstart powered $8.9 billion in loans over the past year -- less than 2% of its addressable market. Moreover, Upstart can enter other markets in the future, such as student loans and mortgages. That's why this growth stock looks like a smart way to invest in AI.
2. Riskified
Riskified is a fintech company that specializes in e-commerce fraud prevention. Legacy risk-management platforms tend to be costly and inaccurate, so valid transactions are frequently rejected, and fraudulent ones are often approved. Lost revenue due to false declines hit $443 billion in the U.S. in 2021, and fraud-related losses are expected to reach $25 billion by 2024.
To fix those problems, Riskified leans on big data and AI. Compared to legacy solutions, its platform integrates more deeply with its clients' infrastructure, gathering data across any system that tracks transactions or website interactions. The company then uses AI to analyze those variables and quantify the risk of fraud, allowing it to automate the approval or denial process with 99.8% accuracy.
Across its clientele -- which range from large enterprises like Wayfair to small businesses powered by Shopify -- Riskified boosts sales by approving more transactions than legacy solutions, and it reduces expenses by blocking illegitimate charges. On average, Riskified's 10-largest merchants have seen revenue rise by 8%, and fraud-related expenses fall by 39%. That's a compelling value proposition.
Financially, Riskified's performance has been solid on the top line. Gross merchandise volume rose 28% to $20.9 billion in the third quarter, and revenue jumped 26% to $52.5 million. However, the company's gross margin fell seven percentage points to 46%, due primarily to a sharp increase in chargeback expenses. To clarify, Riskified eats the cost of any fraudulent transactions that slip by its AI, categorizing those costs as chargeback expenses. In other words, Riskified's falling gross margin could be a sign that its AI models aren't working as intended.
However, management provided an alternative explanation, calling attention to several new merchants in new industries. Put another way, Riskified's AI models faltered because the company lacks sufficient data in those markets, so the problem should resolve itself in time. Investors should watch this situation closely. If the company's gross margin is still falling a few quarters down the road, it might be time to sell.
Alternatively, if Riskified's AI models improve with more data and gross profit growth accelerates, this $1 billion company could easily grow tenfold (or even a hundredfold) in the long run.
Comments
Post a Comment