Adam Foroughi on AppLovin's AI-Driven Comeback

Adam Foroughi runs a company most casual tech-news readers have never heard of, and that is by design. AppLovin, the mobile advertising platform he co-founded and now leads as CEO, grew for years without venture capital funding, which let it build quietly, away from press coverage (00:55). What it actually does, in the plainest terms, is help mobile game developers make money from the more than one billion people who play casual games on their phones every day (01:12). Those players, Foroughi points out, are adults and heads of households, not idle teenagers, which is part of why the audience is so valuable to advertisers.

The scale is the first surprise. Roughly $20 billion a year now flows through AppLovin's own ad platform (01:54), inside a broader mobile gaming ad ecosystem Foroughi puts at about $50 billion annually once every competing network is counted (02:18). Foroughi himself points out that social media advertising itself was only around that same size not long ago, a reminder of how quickly this particular corner of the internet has compounded (02:18).

Advertising as ML 1.0

Foroughi's more interesting argument is historical. He thinks advertising was not a side effect of the internet economy but the original proving ground for the machine learning techniques now powering today's AI boom.

"Advertising is like ML 1.0, but really was the first implementation of all these technologies that now are driving AI today." — Adam Foroughi [03:34]

He is careful to note that a large language model, the technology behind chatbots like ChatGPT, creates far more economic and social value than an ad network ever will. But advertising, he says, was the first place deep learning (a type of machine learning that uses layered neural networks to find patterns in data) turned a profit at scale, and researchers still move back and forth between recommendation systems and language models because the underlying math is related (03:56).

That history matters because it shapes how Foroughi thinks about the threat chatbots pose to his business. He splits advertising into two categories. One is bottom-of-funnel search, where a shopper already knows what they want and is just closing the loop, which is Google's core business and the part chatbots will compete for directly (06:24). The other is discovery, where a person has no idea they want something until an ad shows it to them, which is AppLovin's and Meta's territory.

"When you show a consumer an ad for something that they had no idea existed, they didn't know they needed to buy" — Adam Foroughi [07:12]

The distinction is not just semantic. A search ad, Foroughi argues, captures a transaction that would have happened anyway even without advertising, because the consumer already intended to buy. A discovery ad manufactures a transaction that otherwise would not exist, which is why he describes it as genuine economic expansion rather than a transfer of spending from one channel to another (07:12).

He also uses the conversation to knock down a popular conspiracy theory. Foroughi flatly denies that ad networks listen through phone microphones or track precise location to target ads, arguing the data transfer required to parse audio in real time and match it to an ad is not remotely practical at the scale advertising runs (08:59). What actually creates the eerie feeling of being listened to, he says, is a combination of past searches, browsing history, and social connections, since platforms can infer shared interest between friends without ever needing a live microphone feed (09:18).

From $3.8 Billion Back to $750 a Share

The most striking part of the conversation is Foroughi's account of nearly losing the company to the public markets, not to competitors. AppLovin went public in April 2021 during the COVID IPO wave at a $28 billion market cap and briefly reached $40 billion (10:54). Then, in 2022, the stock fell almost every trading day. It bottomed at a $3.8 billion market cap in the same year the company generated $1 billion in EBITDA, a profitability metric (earnings before interest, taxes, depreciation, and amortization) that was actually growing while the stock collapsed (10:54).

Foroughi, who has a finance background, says the disconnect came down to who owned the stock, not what the business was doing.

"Your price in the markets is determined by the quality of your investors." — Adam Foroughi [11:14]

Private equity holders and early employees were selling into an IPO market so crowded that serious institutional investors never bothered researching a company with what he calls a goofy name. The result was heavy supply and almost no demand (11:14). His response was to stop chasing those investors entirely and turn the company's own cash toward itself.

"Let's start buying our own stock. Let's become our best investor." — Adam Foroughi [12:05]

Over time AppLovin spent about $6 billion buying back its own shares, retiring somewhere between 20 and 25 percent of shares outstanding (12:05). Foroughi is candid that the period was brutal internally: he recalls family members calling to ask whether he was suicidal as the stock fell 92 percent from its high, and notes that his employees, without his ownership stake or public standing, were fielding the same calls from their own families (12:47). The company's answer was a retention plan built around a performance stock structure normally reserved for CEOs, extended to key employees, paired with a blunt message: the paper wealth was gone, but a recovery would make up for it (13:11).

The recovery had a specific technical trigger. In April 2023, AppLovin switched its advertising engine from a regression model, an older, simpler statistical method for predicting outcomes, to a deep learning model (13:51). Advertiser returns improved so much that when Foroughi finally resumed talking to investors that September, with the stock around $80, it jumped to $150 in a single week as institutions who had ignored the company suddenly piled in (14:34). Over roughly two and a half years, AppLovin's stock ran from $9 a share to $750 (15:09), a swing Foroughi uses to make a broader point about public market behavior: most investors, he says, follow trends late in both directions, whether the stock is being written off or being chased.

The rest of the conversation fills in how AppLovin defends a business that now runs at an 84 percent EBITDA margin, among the highest in the industry (21:36). Part of the answer is structural: AppLovin briefly owned its own game studios specifically to generate proprietary data for its first deep learning model, then sold them once outside developers were willing to share data directly (18:17). Part of it is cultural. Foroughi argues that staying lean and narrowly focused on mobile gaming lets AppLovin out-execute giants like Google and Meta in this one lane, even though those companies have more engineers and more capital (20:34).

"If you can innovate and you have differentiated data, you can build an advantage." — Adam Foroughi [22:39]

It is a modest claim for a company that went from a near-death market valuation to a quarter-trillion-dollar peak in less than three years. But Foroughi's framing throughout is less about triumph than about paranoia as a management style: build for the next threat, not the last win, and trust the model over the noise.