AppLovin’s AI Engine Is Printing Profit

AppLovin (NASDAQ: APP) posted another stunning quarter, solidifying its position as one of the most profitable AI-powered software platforms in the public markets. For Q3 2025, revenue surged 68% year-over-year to $1.405 billion, while net income nearly doubled to $836 million, up 92%. Adjusted EBITDA hit $1.16 billion, translating to 79% EBITDA margins, a figure few technology companies can match.

This explosive profitability has caused a re-rating of AppLovin’s stock in recent months, driven largely by the success of its AI-based ad platform, AXON 2.0. Unlike traditional mobile ad networks, AppLovin’s AI models can process real-time behavioral data, predict user conversion probability, and automatically allocate ad spend with minimal human oversight. That high-frequency optimization is what’s turning the company into a cash machine.

AI Advertising Flywheel Is AppLovin’s Real Moat

AppLovin is no longer just an app monetization company. It is an AI-driven advertising infrastructure provider. Revenue from its software platform continues to scale faster than its legacy gaming business. The company’s model is simple: ingest data from thousands of apps, feed it into proprietary models, and redeploy optimized ad campaigns that cost less and convert more. This creates a flywheel — more data, better predictions, better results, more spend.

According to Kenji Sano, Technical Marketing Lead at Adalo, AppLovin’s edge lies in how tightly integrated its AI models are with developer and advertiser tools. He explains that the true power of AXON 2.0 is not just algorithmic efficiency, but its ability to make real-time decisions across billions of ad impressions without manual tweaking. For developers building on platforms like Adalo, the appeal is how AppLovin simplifies complex ad monetization workflows using automation rather than dashboards and spreadsheets.

Still, Gilmore notes that this advantage could narrow if competitors like Unity, Google, or Meta successfully integrate similar AI-based optimization tools at scale. As he puts it, the barrier isn’t building AI models — it’s training them on the right volume and diversity of behavioral data.

Shareholder Returns and Financial Strength

This quarter, AppLovin generated over $1.05 billion in free cash flow, essentially converting all operating profits into available cash. The company spent $571 million on share repurchases, and its board authorized an additional $3.2 billion in buybacks, bringing remaining authorization to $3.3 billion.

With just 339 million Class A and B shares outstanding, this buyback pace could significantly reduce share count over time, amplifying EPS growth even if revenue growth moderates. Guidance for Q4 2025 points to revenue between $1.57–$1.60 billion and adjusted EBITDA between $1.29–$1.32 billion, implying 82–83% margins — meaning profitability is accelerating into year-end rather than cooling off.

Can 80%+ EBITDA Margins Be Sustained?

The biggest question investors are asking now is simple: are these margins permanent or temporary? AppLovin’s extraordinary profitability is driven by its AI infrastructure, low incremental costs, and high software operating leverage. But the model also carries risks. Digital ad markets are cyclical, gaming spend is unpredictable, and developers may demand better revenue-sharing terms if AppLovin’s take rate grows too large.

Jason Gilmore warns that high-margin AI infrastructure models attract regulatory attention and competitive pressure. Apple’s privacy changes (ATT), Google’s evolving Android policies, and increased scrutiny of user tracking models could disrupt data collection — the fuel of AXON’s machine learning system.

Additionally, if economic conditions weaken and ad budgets shrink, AppLovin’s revenue could flatten while fixed AI infrastructure costs remain high. The company will need to prove that its AI efficiency can weather macro slowdowns better than traditional ad platforms like Meta or Alphabet.