Adam Foroughi, AppLovin

What It Covers

AppLovin founder Adam Foroughi traces the company's path from a 2012 mobile app discovery tool to a $140B advertising platform, covering the failed Chinese acquisition, a $6B stock buyback at a $3.8B market cap, the Axon AI model development, and how a team of 400 generates over $5B annually in cash flow.

Key Questions Answered

• •Contrarian Buyback Strategy: When AppLovin's stock collapsed 92% from $115 to $9 per share in 2022, Foroughi deployed $6B in buybacks — not through open market purchases, but by negotiating directly with known sellers holding roughly 50% of shares. The company was generating over $1B in EBITDA against a $3.8B market cap, a 5x cash-flow-to-market-cap ratio that made the math straightforward despite external skepticism.
• •Vertical Integration for Data Acquisition: When advertisers refused to share purchase-behavior data with AppLovin — data critical for training machine learning models — Foroughi acquired 14–15 gaming studios over five years to generate that data internally. The strategy was explicitly temporary: once third-party developers trusted the platform and shared their data, AppLovin sold all studios to TripleDot for a clean exit, refocusing entirely on the advertising platform.
• •Performance Marketing as Arbitrage: AppLovin's core product philosophy is turning advertisers into arbitrageurs. If a developer spends $1,000 and the platform guarantees measurable returns exceeding that spend within their target payback window — 30 days, 6 months, or 1 year — the only constraint on scaling becomes the advertiser's bank balance. This self-reinforcing model eliminates the need for a traditional sales force entirely.
• •Headcount as a Quality Filter: After going public, Foroughi reduced equity recipients from hundreds to roughly 100 critical contributors, then cut overall headcount by 40% in 2024 — while the business grew nearly 100% year-over-year. He personally approves every new hire, converting automatic backfill postings into a friction-heavy justification process that reduced annual hiring attempts from hundreds to tens, forcing managers to prove necessity before adding headcount.
• •Axon Model Architecture Progression: AppLovin's advertising model evolved through three distinct phases: a rules-based system, Axon One using traditional machine learning, and Axon Two using deep learning launched in April 2023. The Axon Two release triggered the stock's move from roughly $9 to a peak of $750, as the model enabled any advertiser to plug in, spend, and receive measurable returns without manual optimization — making the product effectively self-selling.
• •Regulatory Blind Spot in Cross-Border M&A: Foroughi announced a $1B investment from a partially state-owned Chinese fund at a $1.4B valuation in 2016 without understanding CFIUS review requirements. After a year-plus of failed regulatory navigation, he restructured the deal from a 70% equity stake — which triggered control concerns — to a convertible note converting to 10% equity, staying below the control threshold. A board with capital markets experience would have flagged this before announcement.