The four major AI applications in the US stock market all exceeded expectations

AppLovin's Q1 financial report once again exceeded market expectations and drove the stock price up.

So far, four companies focused on AI applications in the US stock market—Duolingo, Palantir, Tempus AI, and AppLovin—have all delivered better-than-expected performance this season. (There was also some reaction in the A-share market this morning.)

The "surprise" from a single AI application company may be coincidental, but multiple companies performing better than expected confirms that enterprises that have taken the lead in deeply integrating AI into operations across different business scenarios are already achieving quantifiable results in cost reduction and efficiency enhancement, thereby accelerating market share acquisition and iterating more user-friendly products.

1. AppLovin

The advertising bidding engine Axon 2.0 drove a 71% year-on-year increase in advertising revenue, with the gross margin of the advertising business continuing to rise. The sale of self-developed game studios and a focus on pure software monetization further amplified operational leverage.

1. Duolingo

Leveraging generative AI to automatically generate courses and exercises, the development cycle for 148 new courses has been compressed to within a year. The paid subscription Duolingo Max has driven an increase in ARPU, with both traffic and paid subscriptions growing.

1. Palantir

The enterprise-level product AIP (Artificial Intelligence Platform) has become the main engine of performance, with US commercial revenue up 71% year-on-year. The company maintains a "high growth + positive free cash flow" combination, with a Rule-of-40 index above 80%.

1. Tempus AI

Collaboration on multimodal tumor foundational models (with AstraZeneca, Pathos) has brought in $200 million in data licensing fees over three years. After acquiring Ambry Genetics and Deep 6 AI, the closed loop of detection—data—model—pharmaceutical collaboration has begun to show revenue contributions.

Common Pathways

Algorithm Efficiency: Self-developed or upgraded inference engines reduce marginal costs; AppLovin's advertising cost per million requests has decreased by about 30%.

Platform Spillover: Outputting APIs or SaaS to the B-side forms recurring revenue; both Palantir and Tempus's data licensing businesses maintain high double-digit growth.

Valuation: The market is shifting from purely PS/PE to considering the Rule-of-40 and the pace of positive FCF. The Rule of 40 is a commonly used comprehensive evaluation metric in the SaaS industry.

The calculation formula is roughly: Rule of 40 Score = Year-on-Year Revenue Growth Rate + Profitability

If the sum of the two is ≥40%, it is generally considered that "a relatively healthy balance has been achieved between growth and profitability." Companies like Palantir proactively disclose their Rule of 40 scores in investor materials.

The domestic software ecosystem differs greatly from that of the US, with a greater inclination towards the B-side in China. Therefore, let's see if there are any relevant statements in Tencent's (on the 14th) and Alibaba's (on the 15th) earnings reports next week that can indicate some positive changes in domestic applications