Adam Foroughi

1. Key Themes

From Outdated ML to Neural Architecture: The Axon 2 Transformation

AppLovin's inflection point was replacing a tree-based ML model (Axon 1) with a deep neural network using semantic embeddings — a complete re-architecture that unlocked both better performance and lower infrastructure costs. The transition was driven by a tiny team in an extremely compressed timeline.

"Before, the last generation model tried to model this by clustering those combinations into different clusters and view that this tree-based model that's basically a huge set like hundreds of thousands of if-else branches... The Axon 2 approach uses this concept nowadays people are very familiar with thanks to the language model. It's called semantic embedding. So basically we use learnable embedding tables to encode those ID-like item IDs, user IDs that otherwise would make no meaning." 00:14:33

"Once we are able to make prediction more accurate, advertisers see better returns and our business grow." 00:00:18

Radical Leanness as a Structural Advantage

AppLovin runs its entire advertising engine — now a $100B+ company — on roughly 400 people total, with ~100 on engineering. This isn't an accident; it's a deliberate cultural design rooted from the founding days and now amplified by AI.

"The company was probably ten people at the beginning and then maybe for the first three, four years, max 50 people. So we always ran really lean. And in an advertising company, if you can't make money, then you probably shouldn't be in business." 00:05:50

"About two to three years ago, we probably cut it down from 600 to 400. And we haven't really changed all that much since." 00:07:56

Engineers Sitting "On Top of" AI, Not Next to It

AppLovin's philosophy is that AI raises the boundary of what humans need to handle — and engineers must continuously move upward as that boundary rises, focusing on what AI cannot yet do: long-horizon planning, taste, and architectural judgment.

"I don't want our engineer to sit next to AI. I want our engineer to sit on top of AI. So as AI is improving, there's a boundary between AI and a human is also moving... As AI is improving, human just sits on top of it and focus on what AI is not good at." 00:00:18

"You as a human is held accountable for every decision that your AI made for you. So it is absolutely not acceptable when I come to an engineer and say, hey, why did you implement this system this way, if the engineer tells me, oh, I don't know, AI did this." 00:37:45

Performance Advertising as a Self-Reinforcing Revenue Engine

AppLovin's business model is explicitly built to not require traditional sales — advertisers measure ROI directly and scale themselves. This creates a fundamentally different (and far more capital-efficient) growth loop than typical ad platforms.

"We want an advertiser to be able to spend money on our platform, measure it, and know that they end up getting more revenue and profit from the money that they spend. If you can give them a user and they make profit on it, then they'll scale pretty large numbers. You don't have to sell that through." 00:20:47

"We get companies that scale on our platform that are small businesses to most but have very large P&Ls and they're able to do that because of a solution like ours." 00:21:37

The "Billion People" Insight: Gaming Data Is Universal Behavioral Data

The conventional wisdom was that AppLovin's gaming data couldn't transfer to e-commerce. The contrarian and ultimately correct insight was that a billion users are not just gamers — they are whole humans, and understanding them deeply enables cross-vertical prediction.

"One thing that I lose sight of is that it's a billion people on the other side, not a billion just people who are only playing games and doing nothing else. So if you understand the person, then you can actually deliver value to them outside of just a game-to-game scenario." 00:22:34

"The people who were playing mobile casual games turned out to be a lot of heads of households. It skews slightly female, but it is this sort of 30 to 50 year old age group that's the heavy mobile casual gamer." 00:23:24

No Product Management: Engineers Who Own Business Problems

AppLovin has almost no product management. Engineers understand business metrics directly, write their own specs, and architect solutions without a handoff chain. This creates faster iteration and better product intuition.

"One of the cultural philosophies we had is almost no product management and have engineers that are competent enough about business problems to just write their own architecture, write their own product to solve business problems that maybe not even a business team knew existed." 00:26:19

"You have engineers that didn't even talk to a business person sitting down together at breakfast on a napkin, writing out a solution to a problem that we never even discussed and then something that could become a big business." 00:26:46

Playing Offense During a 92% Stock Decline

When the stock tanked after IPO, AppLovin stopped doing investor relations entirely, generated cash, and became their own best investor through aggressive buybacks — rather than trying to pitch a story to skeptical public markets.

"We actually didn't do any investor relations for well over a year at the bottom because what were we going to convey to investors? We're rebuilding our technology. Stock is dirt cheap but nobody buys something that's dirt cheap... So we were able to flip it and say, we're just going to buy our own shares. We're going to become our best investor." 01:07:18

The Path to a Trillion: Cash Flow Math and Adjacency Expansion

Adam (CEO) laid out the explicit financial model for reaching a trillion-dollar valuation: they need $30B+ in annual cash flow, which requires expanding far beyond gaming and consumer e-commerce into adjacent high-scale categories, including potentially connected TV and the open web.

"We probably on an EBITDA basis, I think we're over a $7 billion run rate. And we generate somewhere around 75% cash off the EBITDA dollar... To take it from this scale up and be worth a trillion dollars, we've got to believe that we can get to $30 billion plus of cash flow a year." [00:01:02:51]

"What is compelling as we think about what's next in our business is what are the other spaces that create this form of discovery where you don't actually know what the user wants. An example... is connected TV. User doesn't know what they want. If you can start them down the flow and convert them, that's super high value." 00:59:39

Taste Over Output as the True Competitive Differentiator

Both Gio and Adam converge on "taste" — the judgment of what NOT to build — as the most underrated competitive advantage in an era where AI makes building cheap and fast.

"I would attribute the success of Axon largely to what we decide not to do, not actually what we did." 00:00:18

"I think just in this era, AI allows everyone to build almost everything you want. Taste is a very important thing and I think has been underestimated." 01:15:55

2. Contrarian Perspectives

Most Companies Will Never Successfully Transition to AI-Native — Replace, Don't Retrain

Most executives believe they can train existing teams to be AI-native. Adam (CEO) argues this is nearly impossible at scale and that the only real solution — wholesale replacement of people and culture — is impractical for most organizations.

"Technically you almost have to replace near everyone and rebuild the culture up to be able to deal with the world as it is today. But that's not really possible for most companies. And so there's no clear answer for how those companies can get there." 00:54:16

Struggling SaaS Companies Should Not Be Saved — They'll Be Taken Private by PE

Rather than believing in a recovery narrative for SaaS companies with falling multiples, Adam argues the structural endpoint for most is private equity restructuring — not a public markets comeback.

"A lot of times when companies reach that point where it's tough... they start getting accumulated by private equity or effectively plucked off the public markets taken private because the only way to recover it is a complete restructure. You fire a lot of people, you cash flow it, you lever the business up and that fits the private equity model." 01:09:13

AI Creative Will Not Flood Ad Systems With Diversity — It Will Homogenize Them

While many assume AI-generated ads will create abundance and variety, Adam (CEO) argues the opposite: AI-generated creative will converge on the same outputs, destroying the alpha that differentiated advertisers currently enjoy.

"If the AI starts writing all the ad creative, then all of it's going to eventually look the same and the user won't respond to the ad. So how do you create alpha? You have to have really smart people that understand systems, understand what they're asking from the AI to output for them." 00:38:15

Chatbot/LLM Advertising Is a Poor Fit for Discovery — Search Is Already There

While the market speculates about massive advertising opportunities inside LLMs, Adam (CEO) argues that chatbot usage patterns (~99% used as search engines) make them bottom-of-funnel by nature — and that space is already well served. AppLovin's model requires top-of-funnel discovery, which chatbots structurally don't enable.

"If 99% of the use case is search, then the ads in chatbots are probably going to look and feel a lot like search... For us, what is compelling as we think about what's next in our business is what are the other spaces that create this form of discovery where you don't actually know what the user wants." 00:58:10

Investor Relations During Distress Is Counterproductive — Go Silent and Buy Back Shares

The conventional playbook for a company with a crashing stock is to engage investors more, do more roadshows, tell the story. AppLovin did the opposite: went silent for over a year and deployed capital into buybacks instead.

"We actually didn't do any investor relations for well over a year at the bottom... Investors always, just like the private markets, will chase something that they expect is going to grow really quickly but they very rarely tend to buy something where they think it's just cheap because things trade wherever they deserve to trade." 01:07:18

3. Companies Identified

AppLovin

Mobile advertising platform and AI-powered ad tech company. The subject of the podcast — built from near-zero to a $100B+ market cap, operating with ~400 employees, ~$7B+ EBITDA run rate, and 75% cash conversion. Core product is the Axon 2 recommendation model powering game and e-commerce advertisers.

"Post-Axon 2, we recovered and now are a 100 billion dollar company." 01:05:45

Meta (Facebook/Instagram)

Social media and advertising giant. Referenced as the model AppLovin studied for pixeling advertiser websites to build user behavioral data beyond owned platform activity.

"This is what powers a lot of what Facebook was able to build over the years... they have a lot of people and you start understanding the behaviors of those people through what people are doing on advertiser properties. You pixel websites, you get data." 00:22:58

Adjust

SaaS mobile measurement/attribution company acquired by AppLovin. Mentioned as a separate, never-integrated business that was part of the broader portfolio.

"Not including those and not including Adjust, which is a software as a service company we bought, never integrated." 00:07:56

Wayfair

Large e-commerce retailer. Used as the example of a brand-sensitive advertiser for whom AI-generated creative at current quality levels would be insufficient.

"If you get a brand as big as a Wayfair and you hand them something out of the box that looks not ideal for their brand, they lose trust in you as a platform." 00:47:04

Higgsfield

AI video generation tool for influencers. Mentioned as an example of a creative AI company optimized for social/UGC short-form content — but not suited for AppLovin's longer-form brand advertising needs.

"Higgsfield is a tool for influencers. That's not the advertiser on the other side of our product." 00:47:04

Anthropic

AI company behind Claude. Referenced as a major AI spend destination, with the concern being that non-AI-native companies wastefully duplicate spend across teams with no coordination.

"Everyone's paying Anthropic and OpenAI for the same stuff. Like it's like, what are we doing here?" 00:53:47

OpenAI

AI company behind ChatGPT. Referenced alongside Anthropic as an AI infrastructure provider and as a tool Gio uses for text tasks.

"For text tasks, I like to use ChatGPT these days." 00:54:55

Brex

Corporate spend and finance platform. Sponsor of the podcast; mentioned as used by Vercel, OpenAI, Anthropic, Granola, and Deepgram.

"The companies building what's next from Vercel, OpenAI, Anthropic, Granola, and Deepgram all made the same call. They all run on Brex." 00:27:10

Helsing

European defense AI company. Co-founder Torsten mentioned as having come from the gaming space, used as a reference point for what founders from gaming go on to build.

"We had Torsten who then went on to found Helsing." 01:24:48

Anduril

Defense tech company. Palmer Luckey mentioned as another gaming-adjacent founder who went on to build a major defense company.

"We had Palmer Luckey who went on to found Anduril." 01:24:48

4. People Identified

Giovanni Ge (Gio)

CTO of AppLovin. Architect of the Axon 2 model that transformed the company. Joined November 2022 from big tech when the stock dropped 30% on his start date. Built the neural embedding-based recommendation system that powers AppLovin's advertising engine.

"He was the architect behind it... We liked Gio a lot in the interview and both of us gave him the hard pitch on joining." 00:04:00

CEO and co-founder of AppLovin. Former derivatives trader turned advertising entrepreneur. Described as intensely focused, having removed all distractions from his life to concentrate on the business. Credited as the strategic and cultural architect of the company.

"I always optimize my life to remove all distractions and focus on the one thing that I always get at, which was building a business in the category that I'm in." 01:20:22

Basil (Bazel)

Former CTO of AppLovin, still referenced as a key technical leader. Described as having personally written ~60% of the company's code before Gio joined. Exemplified the individual-contributor-first engineering culture.

"Bazel probably wrote, I don't know, 60% of the code of the company like himself. And the mentality was always individual contributors can help us succeed in this very, very competitive field." 00:06:47

Max Levchin

Co-founder of PayPal, CEO of Affirm. Referenced as a hands-on technical CEO/CTO type that the host had asked about the future of coding skills.

"I asked Max Levchin this question because he's a hands-on CTO, CEO type. He's super technical. I asked him, do you still think people need to learn how to code?" 00:36:37

Mickey Malka

Prominent venture investor (Ribbit Capital). Referenced in the context of Silicon Valley losing the ability to make beautiful things — connecting to the "taste" theme.

"I was just listening to David's interview with Mickey Malka and Mickey Malka was talking about how Silicon Valley lost the ability to make beautiful things." 01:16:05

Torsten

Co-founder of Helsing (European defense AI company). Named as an example of a gaming-space executive who went on to build a significant company in a completely different domain.

"We had Torsten who then went on to found Helsing." 01:24:48

Palmer Luckey

Founder of Anduril (defense tech) and Oculus. Named alongside Torsten as a gaming-adjacent founder who made a major pivot to defense technology.

"We had Palmer Luckey who went on to found Anduril." 01:24:48

Patrick O'Shaughnessy

Investor and podcast host (Invest Like the Best). Referenced as having conducted an earlier interview with Adam Foroughi in 2022 that explored AppLovin's ecosystem dynamics with Apple and Google.

"I was listening to this podcast you did with Patrick O'Shaughnessy back in 2022 and you were explaining how delicate but dynamic and competitive the gaming ecosystem is." 01:00:58

5. Operating Insights

Let New Hires Push Code Within the First Week — Skip the Onboarding Theater

The fastest way to retain top engineering talent and accelerate their growth is to give them real deployment access immediately. The dopamine hit of code in production with real-world feedback is more motivating than any training program, and it self-selects for people with genuine curiosity and initiative.

"Within a week, they can push a test. They start getting results. That moment is a pretty substantial dopamine hit. They get something that they developed or thought to develop quickly into production, reaching a ton of users, and they get real-world feedback loop around it." 00:33:00

Give Engineers Full Business Context, Not Just Tasks

Disconnecting engineers from business outcomes is the most common and damaging management mistake in technical organizations. When engineers understand P&L impact, they make dramatically better build decisions without needing product managers to mediate.

"Instead of assigning tasks to them and telling them to do exactly what they have to do, I showed them how their work is impacting the business. And because our engineers are very talented, once they are provided with that kind of context, they have a much better sense of what to build." 00:30:36

Separate Prototype Systems from Core Infrastructure — Different Quality Standards Apply

Maintaining code quality at scale requires explicitly distinguishing throw-away experimentation systems from core infrastructure, and applying human accountability standards (not AI excuse-making) only to the latter.

"Internally, we also have this concept of, you know, what is the system you want to prototype? And what is the core system that you have to keep absolutely clean?... For all the core infrastructure, the core system... you can use AI however you want. But at the end of the day, you as a human is held accountable for every decision that your AI made for you." 00:37:16

The Right Hiring Screen: Intelligence + Low Ego + Response to Adversity

This three-factor framework for hiring is more predictive than any skills test. Low ego enables listening, feedback absorption, and self-questioning — which are the primary drivers of compound growth in high performers. Response to adversity reveals durability.

"The type of people I really like in my team, they usually have two characters. One is intelligence. The second is having low ego... When a person has low ego, he tends to listen, he tends to empower his team, tends to help others, and more importantly, they take feedback." 01:23:21

"The third thing which I look for and always ask about is how they react to adversity. It's nothing is easy in the world. And some people grew up with easier upbringings than other people. And usually, people who know how to respond well to adversity do really well." 01:24:07

Fix Equity Compensation in Dollar Terms, Not Share Count, to Survive Stock Volatility

When a stock collapses, companies that issue fixed-percentage equity grants face catastrophic dilution spirals. Thinking in fixed-dollar compensation terms preserves equity discipline and makes the company survivable through downturns.

"We always thought about that in fixed dollar terms and we want to pay our people exceptionally well but we don't want to overpay our people when we keep a lean team. So, it's always manageable." 01:06:49

6. Overlooked Insights

In-Game Ads Are Effectively a 10-Minute TV Network With Interactive Data — an Entirely Different Medium Than Social

This was mentioned briefly but its implications are enormous for advertisers and investors. AppLovin's ad inventory is not comparable to banner ads, social feeds, or even standard video. Users engage for 60 seconds per ad, see ~10 ads per conversion funnel (equaling ~10 minutes of engaged contact time), and interact with playable mini-games generating behavioral data throughout. This makes AppLovin's inventory structurally closer to premium interactive television than to digital advertising — and means its competitive moat and pricing power are deeply underappreciated.

"If you think about average time spent on most websites that are really successful, user spends, call it 40 minutes a day on a really successful social property. If a user sees 10 ads on our platform a day, that's 10 minutes of engagement with the website effectively. That's like a mini game serving website. Tons of data is able to be extracted out of those engagements." 00:18:37

"Because I said this earlier in the pod, if you're giving them 10 minutes of content, it better be interesting. Otherwise, you're actually going to turn them off." 01:02:23

AppLovin's Core Gaming Audience (30-50 Year-Old Heads of Household) Is the Most Valuable and Most Overlooked Consumer Demographic in Digital Advertising

This was dropped in a single sentence but represents a massive strategic insight. The mobile casual gaming audience — skewing female, 30-50 years old, heads of household — is the decision-maker for most consumer spending categories: home goods, food, retail, financial products, insurance. Yet this audience is almost entirely invisible to social platforms that optimize for younger, more active scrollers. Any brand targeting this demographic and not yet on AppLovin's platform is leaving significant ROI on the table, and this demographic profile is a major structural reason why e-commerce expansion onto AppLovin works far better than analysts predicted.

"The people who were playing mobile casual games turned out to be a lot of heads of households. It skews slightly female, but it is this sort of 30 to 50 year old age group that's the heavy mobile casual gamer. And so we had this really good user." 00:23:24