Key Moments
Adam Foroughi, Applovin CEO: Surviving a 92% Drawdown, Ads as ML 1.0 & the $50B Game Ad Market
All-In PodcastWant to know something specific about what's covered?
We've already dissected every moment. Ask and we will deliver (with timestamps).
Key Moments
AppLovin's ad platform, hidden in mobile games, now drives $50B in ad spend and outperforms Facebook for e-commerce, despite a near-total market cap collapse.
Key Insights
AppLovin's mobile gaming ad platform is associated with approximately $50 billion in annual advertising spend, with the company itself handling around $20 billion after a 60% year-over-year growth.
Advertising technology is described as 'ML 1.0', serving as a foundational implementation for many technologies now driving modern AI and large language models.
During 2022, AppLovin's market capitalization fell from $28 billion (and peaked at $40 billion) to $3.8 billion, despite achieving $1 billion in EBITDA, representing a 92% drawdown.
The company executed an aggressive stock buyback program, repurchasing approximately $6 billion of its stock and retiring 20-25% of shares, which at its peak was worth over $50 billion.
Apple's privacy changes, while initially presenting a headwind, have led to user demand for more relevant ads, as users prefer watching engaging ads over 'spam' for rewards.
AppLovin maintains an industry-leading EBITDA margin of 84%, attributed to its lean operational structure, algorithmic focus, and automation, making it difficult for competitors to erode its margins.
An advertising powerhouse hidden in plain sight
Adam Foroughi, CEO of AppLovin, discusses his company's significant, yet often understated, presence in the mobile advertising market. AppLovin operates an ad platform embedded within over 100,000 mobile games, quietly outperforming established players like Facebook for e-commerce brands. The sheer scale of the mobile gaming universe, with over a billion daily adult players, presents a massive monetization opportunity. AppLovin disclosed $11 billion in ad spend on its platform nearly two years prior to the interview, and has since grown 60% year-over-year, reaching an estimated $20 billion today. Factoring in other players, the total ad spend in this ecosystem is estimated to be around $50 billion annually, a market that rivals the size of social media advertising in its past peak.
Advertising as the genesis of modern AI
Foroughi posits that the first wave of internet advertising, particularly through platforms like Google's Adwords and AdSense, was essentially 'ML 1.0' – the initial, critical implementation of technologies that now underpin today's artificial intelligence. While large language models (LLMs) today have a greater societal and economic impact, advertising provided a highly profitable proving ground for deep learning models. Recommendation systems, though structured differently from LLMs, follow a similar developmental trajectory. Research in LLMs can often be ported to recommendation systems, and vice-versa. Many researchers in the LLM space began their careers working on advertising systems, highlighting the deep, symbiotic relationship between these fields. A key advantage of advertising as a model is its ability to immediately translate the value of a prediction – whether it's an ad impression or an engagement on a social post – into economic terms.
The evolution of ad relevance and discovery
Foroughi contrasts early 2000s advertising, which he describes as 'complete garbage' due to technological limitations, with today's highly relevant ads. Companies like Facebook recognized early on that by combining vast amounts of data with robust technology, ads could become highly personalized. This has led to a paradigm shift where ads are increasingly perceived as content, driving significant engagement. AppLovin observes this in mobile gaming, where users actively engage with in-game ads, often in the form of mini-games, because the recommendation technology has become so adept at suggesting relevant experiences. This shift from intrusive, irrelevant advertising to discovery-based marketing is crucial for brands seeking to create demand for products users didn't even know they wanted. This 'discovery' aspect, as opposed to search-based transactions, is where significant economic expansion occurs, differentiating platforms like Meta and AppLovin from search-centric businesses.
Navigating the LLM and chatbot advertising landscape
The rise of large language models and chatbots presents a new frontier for advertising. While Google's search business historically captured bottom-of-funnel advertising (consumers researching known purchases), LLMs can now close that loop entirely, posing a direct competitive threat to search ads. AppLovin, however, operates in the discovery space – creating intent for products users haven't searched for. This model, similar to Facebook's, focuses on generating economic expansion by introducing novel products and experiences. The core difference lies in whether a transaction was inevitable (search) or created by the ad (discovery). Foroughi emphasizes that discovery ads offer greater economic uplift because they tap into previously unrealized consumer desires. This distinction is vital as the advertising world adapts to AI-powered interfaces, where the focus shifts from answering queries to creating opportunities.
Debunking the 'phone listening' myth
Concerns about phones actively listening to conversations to serve targeted ads are largely unfounded, according to Foroughi. While users might experience seemingly uncanny ad relevance after discussing a product, this is more likely due to a combination of other trackable online activities – such as searches, website visits, or product lookups – rather than microphone surveillance. He explains that the technical and data infrastructure required to constantly parse microphone input in real-time for ad targeting is not currently realistic for advertising companies. However, he acknowledges that social networks can leverage user connections and relationships to influence ad experiences, where one user's search or activity might inform ads shown to their network.
Surviving a 92% market cap collapse
AppLovin's journey included a dramatic public market valuation crisis. After a $28 billion IPO in April 2021, with $600 million in EBITDA, the company's market cap surged to $40 billion before collapsing by 92% to $3.8 billion in 2022, even as it generated $1 billion in EBITDA. Foroughi attributes this to a confluence of factors: a lack of sophisticated investor demand due to the crowded IPO market and a 'goofy' company name. Facing this downturn, AppLovin shifted strategy, halting investor outreach and initiating an aggressive stock buyback program, repurchasing $6 billion of its shares, which reduced outstanding shares by 20-25%. This internal investment, turning the company into its own best investor, turned a deeply depressing period into a significant value-creation opportunity.
The 'us against the world' mentality and employee incentives
Managing internal culture during the 92% market cap drawdown was challenging, with employees facing family and friend skepticism. Foroughi fostered an 'us against the world' mentality, emphasizing resilience and a commitment to recovery. To align employee interests with the company's turnaround, AppLovin implemented a performance stock plan for key personnel, acknowledging the current hardship but incentivizing them to persevere for future upside. This approach, combined with the company's technological advancements from 'ML 1.0' to 'ML 2.0' (deep learning models), fueled rapid growth. By September 2023, with strong performance, the stock had recovered significantly, prompting Foroughi to re-engage with investors, leading to another surge in market cap.
Privacy headwinds and the user's desire for relevance
Regulatory privacy changes, particularly from Apple and the EU, have impacted ad targeting. While these rules aim to protect user privacy by grouping users into broader categories rather than precise targeting, the outcome has been mixed. Paradoxically, users have complained about receiving more 'spam' and less relevant ads, leading to increased demand for personalized content. Foroughi argues that clear regulations are beneficial for technology companies, allowing them to adapt. Consumers, especially when watching ads for rewards in games, prefer engaging, relevant content over generic 'garbage.' AppLovin's advanced deep learning networks are adept at navigating these privacy constraints while still delivering value to advertisers and users, demonstrating that privacy compliance and ad effectiveness are not mutually exclusive.
Strategic divestiture and the focus on core competency
AppLovin's brief foray into acquiring game studios was a strategic data play. To train its initial deep learning models, the company needed proprietary data, which game developers were reluctant to share. By acquiring studios, AppLovin seeded its first model with necessary data, achieving success and driving rapid growth. Once the platform demonstrated its efficacy and began attracting third-party advertisers, AppLovin divested these acquired studios, refocusing on its core ad technology business. This strategic move highlights a disciplined approach to capital allocation and a commitment to leveraging its core strengths in machine learning and advertising.
The enduring value of discovery in an agentic world
Looking ahead to a future with AI agents optimizing tasks, Foroughi believes that traditional discovery platforms will remain relevant. While agents may handle routine tasks like subscription management, the human desire for 'window shopping' and the transactional experience itself provides a dopamine hit that many consumers, not just the technologically advanced, enjoy. He likens AppLovin's audience to that of the New York Times or Yahoo – vast numbers of everyday shoppers who value the process of discovery and comparison. This emotional and experiential aspect of shopping, Foroughi argues, is something agents cannot fully replicate, ensuring a continued role for discovery-driven advertising, even as technology evolves.
Competing with giants through focus and lean operations
AppLovin's ability to compete with tech giants like Meta and Google stems from its lean operational model and intense focus on the mobile gaming ad space. Foroughi emphasizes a mindset of constant vigilance, believing they are always at risk of being 'screwed,' which drives continuous hard work. A team of subject matter experts dedicated to translating mobile gaming experiences into transactional behavior allows them to move faster than larger, more bureaucratic competitors. This laser focus and agility are critical competitive advantages. Despite operating in a domain dominated by advertising behemoths, AppLovin has achieved remarkable success by staying lean and algorithmically driven.
Industry-leading margins and defensible moats
AppLovin boasts an industry-leading EBITDA margin of 84%, a testament to its operational efficiency and algorithmic prowess. The business model is designed to provide arbitrage for advertisers: they pay AppLovin to acquire consumers, who then transact, covering the cost of acquisition and generating profit. This performance-based, scalable model minimizes leakage, as AppLovin is not the full advertiser but powers the connection between advertiser and consumer. The sustainability of these high margins is attributed to complex technologies and differentiated data, creating a moat that is difficult for competitors to breach. Just as Anthropic has built an advantage in LLMs through model power and scale, AppLovin leverages its technological innovations and data advantages to maintain its market position.
The edge of global engineering talent
Foroughi highlights the significant contribution of his engineering teams in Palo Alto, Beijing, and Singapore. He praises the humility, hard work, and sharp intellect of Chinese engineers, seeing them as among the brightest minds globally. This diverse, international talent pool is integral to AppLovin's innovation and problem-solving capabilities. Foroughi believes that working with exceptional people, even if he perceives himself as the 'dumbest person in the room,' is a key driver of his enthusiasm and dedication to showing up for work each day.
Mentioned in This Episode
●Software & Apps
●Companies
●Organizations
●Concepts
AppLovin IPO and Market Performance
Data extracted from this episode
| Year | Market Cap | EBITDA |
|---|---|---|
| 2021 (IPO) | $28 Billion | $600 Million |
| Peak (2021) | $40 Billion | N/A |
| 2022 | $3.8 Billion | $1 Billion |
| September 2023 | $80 per share (approx.) | N/A |
| Week after Sept 2023 | $150 per share (approx.) | N/A |
| Post-recovery Peak | $250 Billion | N/A |
AppLovin Stock Buyback Impact
Data extracted from this episode
| Action | Amount | Share Impact | Value at Peak |
|---|---|---|---|
| Stock Buyback | $6 Billion | Retired 20-25% of shares | Over $50 Billion |
Common Questions
AppLovin is an advertising company that helps mobile game developers monetize their games. It operates in the massive mobile gaming ecosystem, which sees over a billion people playing daily and generates around $50 billion in annual advertising spend.
Topics
Mentioned in this video
An advertising company that helps mobile game developers monetize their space, growing significantly year-over-year and operating in a $50 billion mobile gaming ecosystem.
Mentioned as a company that did a good job of making ads relevant by pairing data with technology, and whose ad business is driven by discovery moments.
Mentioned in the context of Adwords and AdSense as part of the first wave of internet advertising that sparked critical technologies, and its search business competes with LLM-based advertising.
Mentioned as having an ad product and that ChatGPT is going to make its services free, impacting the advertising landscape.
Referred to for its amazing ad business and its role in creating discovery moments for consumers, similar to AppLovin's aspiration.
Mentioned regarding privacy regulations and changes that have impacted the advertising industry, such as restricting precise user targeting on iOS.
A company that was mentioned as an example of aggressive acquisition of slower-growth businesses that may not interest traditional venture capital.
Mentioned as a property that still has a large daily user base, similar to the audience AppLovin serves.
Mentioned as one of the smartest companies in advertising with strong engineers and models, alongside Meta and Google.
Used as a historical example to illustrate how companies can absorb and integrate various business aspects, relevant to discussions on business leakage and opportunity.
Mentioned as a company that is doing well in the large language model space, used as an example to illustrate how differentiated data and innovation can create an advantage.
More from All-In Podcast
View all 456 summaries
39 minBill Gurley: Searching for Feynman
39 minJared Isaacman: A New Era for NASA and American Space Exploration
33 minNick Shirley: Exposing Government Fraud, Suing California & Taking on the Media
44 minMeta's Dina Powell McCormick: The Case for Data Centers, Backlash, AI Job Boom & Meta’s Future
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free