Adam Foroughi, Applovin CEO: Surviving a 92% Drawdown, Ads as ML 1.0 & the $50B Game Ad Market

AppLovin Company Profile & Scale

• Adam Perugy (CEO) built AppLovin without early-stage VC funding to maintain operational quietness.
• The platform operates within over 100,000 mobile games, serving a daily audience of over 1 billion adults.
• Ad spend on AppLovin's proprietary platform grew from $11 billion (2022 disclosure) to approximately $20 billion annually.
• The total mobile gaming ad ecosystem is estimated at $50 billion annually, having evolved from a $50 billion social media opportunity.
• The company is projected to generate approximately $6 billion in cash for the current fiscal year.

Technological Foundations & AI Strategy

• AppLovin's business model is described as "ML 1.0," serving as a foundational implementation of deep learning technologies now driving AI.
• The company transitioned from regression models to deep learning models in April 2023, shifting focus to "ML 2.0" recommendation systems.
• The strategy evolved from driving user-to-game transitions to creating "discovery moments" that drive shopper behavior and e-commerce intent.
• Deep learning models are now used to predict future outcomes for ad placement, allowing for immediate economic value translation.

Market Evolution & Ad Effectiveness

• User behavior regarding ads has shifted from 2005 "spam" to 2026 content-like experiences where users engage with ads to earn rewards (e.g., lives in games).
• Adam distinguishes two ad models: "bottom of funnel" (search/LLM-driven intent which merely accelerates existing transactions) versus "discovery" (creating new demand for products users didn't know existed).
• AppLovin prioritizes the discovery model to generate economic expansion rather than simply competing with search-based transaction loops.

Capital Markets History & Buyback Strategy

• AppLovin went public in April 2021 with a $28 billion market cap ($600 million EBITDA) during a COVID IPO wave.
• In 2022, the stock price collapsed daily, hitting a low market cap of $3.8 billion despite $1 billion in EBITDA generated that year.
• The collapse was attributed to a lack of "blue chip" investor demand and supply pressure from early private investors exiting.
• Management responded with an aggressive internal buyback program, purchasing $6 billion of stock and retiring 20–25% of outstanding shares.
• Since the market recovery in late 2023, the stock price rose from roughly $80 to $150, reaching a market cap peak of $55 billion.
• Over a 2.5-year period, the share price moved from $9 to a peak of $750 before stabilizing.

Competitive Moat & Operational Efficiency

• The company maintains an 84% EBITDA margin, the highest in the market, driven by a performance-based arbitrage model where advertisers pay only upon transaction.
• Adam argues the business survived competition from Meta and Google by remaining lean, focusing exclusively on mobile gaming, and leveraging superior deep learning algorithms.
• The company divested its owned game studios after using them solely as data sources to train initial deep learning models; the model's success made third-party partnerships viable.
• Adam notes that "Chinese people are very humble, very hardworking, and sharp," citing Beijing and Singapore engineering offices as critical assets.

Regulatory & Privacy Landscape

• Apple's privacy changes (neutering precise targeting) initially forced less relevant ad delivery but subsequently increased user complaints about "spam."
• Adam asserts that deep learning networks have adapted to privacy regulations, allowing for effective group targeting without precise location tracking or microphone listening.
• The company rejects the notion that ads are "creepy" or that users do not want relevance; instead, users prefer relevant ads in exchange for in-game rewards.

Future Outlook & Agent Commerce

• While some users will adopt "agents" to optimize shopping, Adam believes the majority of shoppers still desire the manual "discovery" and "window shopping" dopamine experience.
• The company does not expect "agentic commerce" to replace human discovery for the typical shopper in the near term.
• The team views current public market trends as lagging indicators compared to private market trends, noting that sophisticated investors eventually catch up to the company's performance.