· Adam Foroughi

Adam Foroughi: AppLovin — Buybacks at the Floor, AXON AI, and the Public-Markets Playbook

When the public market crushes a profitable business to ~5x cash flow, the founder with conviction buys aggressively against the structural sellers on the cap table; durable advertising moats come from owning the data flywheel (vertical integration to seed the model, deep learning to scale it) and an A-player-only org built for AI-leverage.

public-marketsbuybacksfounder-modeai-adstalent-densityvertical-integrationcapital-allocationmachine-learning0% confidence

Why this is in the corpus

AppLovin executed one of the most successful contrarian buybacks of the decade ($6B deployed → ~$50-60B proceeds) while EBITDA was growing; demonstrates Founder Mode capital allocation, the Axon two-layer strategy (own studios to harvest training data, then sell the data-advantaged ad platform), and lean-team talent doctrine that took the company from $3.8B floor to $140B market cap in ~3 years.

Summary for skimmers

AppLovin IPO'd at $28B in 2021, crashed to $3.8B by 2022 while EBITDA grew $700M→$1B+. Adam bought back $6B (mostly off-market from known-seller PE/founder shares), borrowed against EBITDA, returned $50-60B in proceeds. Backstory: bootstrapped after VC rejections, no board until 2018, near-miss $600M acquisition offer in 2015, near-miss Chinese take-private blocked by CFIUS, KKR cleanup in 2018. Built vertical-integrated gaming studios to feed Axon ML model with advertiser data Facebook/Google wouldn't share. Axon 2 (deep learning, April 2023) inflected stock from $9 to $750. Fired 40% of team in 2024 while revenue ~doubled to concentrate equity on ~100 critical engineers; CEO personally approves every hire; no CRO, no COO; 80%+ of code is LLM-written.

Briefing

What survives the editorial filter

This page should feel like a smart colleague already listened for you and left only the operating logic worth keeping. Not everything said in the episode makes it through.

Trust signal

Direct episode extraction

Best used for

Decision-grade retrieval metadata not yet added for this episode.

Hold lightly

No explicit downgrade reason stored yet for this episode.

Principles

Durable claims that survive beyond the speaker's biography — each with explicit limits, transferability judgment, and evidence.

Principle

The CEO is responsible for putting the best available person in every seat at every moment

Tenure is not seniority — capability at the moment is.

AppLovin swapped CTOs twice on this principle (John then Basel in 2016 then Giovanni recently). Each time the previous incumbent was performing fine; the underling was simply better.

Reassess every key role for best-available person — including against the bench you already have.

“a lot of people get complacent with the leaders that they have. And those leaders, if they''re not at that moment as good for the role as someone underneath that person underneath is gonna leave and start a new business. So like you owe it to every single person and you owe it to your company to make sure the best person is in the right role at any given time.”Adam Foroughi

Principle

A-players will not work alongside B/C/D players — protect the top of the org first

Tolerating B-players is a slow-motion firing of your A-players.

AppLovin fired ~40% of staff in 2024 while revenue nearly doubled, explicitly to keep A-players from getting contaminated. The cuts were not about cost — they were about retention of the top.

Cut the bottom not for the budget, but to keep the top.

“A players don''t like to work with B Cs Ds, I mean D''s, you''re, you''re, you''re gonna just burn your A''s out. But like anything that''s worse than an A player on the team and you''re distracting your best for working with those types of people and like trying to bring ''em up to speed, it is a huge loss.”Adam Foroughi

Principle

Voice the aspirational number out loud — then back it with a plan

Aspirational targets only work if the founder believes them and shows the work.

From billion-dollar company (2014) to $10B (2016) to $100B plan at IPO to trillion-dollar comp plan today. Each escalation followed a step-change in conviction and was paired with a concrete plan.

Conviction said aloud + plan attached = team alignment.

“every level that we''d gone up, I voiced something aspirational, but I gotta believe it to voice it… we just did a print over a billion dollar valuation. This is gonna be a $10 billion business period. Like now we have to do the work to get there. And we had a plant.”Adam Foroughi

Principle

Float too low at IPO is a self-inflicted volatility tax

Small float -> small investors -> big swings.

AppLovin sold ~7-8% in IPO at $28B. The thin float plus the COVID IPO glut left the cap table without anchor investors — directly causing the 92% drawdown.

10-15% float at IPO; let blue-chip investors build a real position.

“the IPO if I recall was about, I think we sold about like seven 8% float. So whi which is really low and probably a mistake in hindsight, why you wanna sell more float if, if anything like you need float to like lower volatility… too low. A float makes you more volatile and if the valuation is lower and you can do a 10 to 15% float, you end up attracting better investors who can end up being long-term with you.”Adam Foroughi

Principle

Loyalty to backers survives the legal terms — it''s the reputation that compounds

The deal is the start of the relationship, not the end of the negotiation.

When the Chinese deal blew up via CFIUS, Adam refused to leave the investors hanging — he restructured the equity into a convertible note so they got paid back with interest, then unwound it via KKR. Cost him optionality; preserved trust.

Pay the loyalty tax even when the contract doesn''t require it.

“I''ve always felt like if anyone bets on me, especially as someone who''s like just always had a, you''re not gonna bet on me mentality and, and I''m gonna prove to you how wrong you are. I feel like if anyone bets on me, takes an investment in our business, gotta be loyal the other way.”Adam Foroughi

Principle

AI gives A-players non-linear leverage, not B-players linear leverage

AI is a multiplier, not a leveler — top talent benefits exponentially more.

Over 80% of AppLovin code is LLM-written. The org is being explicitly designed around the assumption that 1 great engineer + AI beats 50 average engineers.

Hire fewer, better, AI-fluent people.

“if you lived in a world of B players, your one X engineer might be two x more efficient, but your 10 X engineer might be a hundred x more efficient… Your best people know how to use AI better than your mediocre people. And it''s not by a mo some sort of linear function. It''s, it''s much greater than that.”Adam Foroughi

Principle

Never save cash for a rainy day if you believe in the business

Cash hoarding contradicts conviction — deploy or return.

Adam frames cash conservatism not as prudence but as a tell that management doesn''t believe in the future they''re publicly underwriting. The same disposition powered both the buyback and the willingness to lever up at the bottom.

If you believe, deploy. If you don''t believe, don''t be running the business.

“I never believed in saving cash for a rainy day. I feel like I''m a big believer in what we''re building. I believe in where we''re going. So if I believe in the future and we''re a really high cash generated business, we should always be buying back our share.”Adam Foroughi

Principle

Always hire above yourself in every function

Every hire is either a multiplier or a ceiling — pick multipliers.

RAF on BD, Giovanni on engineering (much smarter than I am), Basel on CTO — the pattern is consistent: identify the gap, hire above yourself, become their cheerleader.

Upgrade yourself with every hire.

“I was always humble enough to know that when you hire you always wanna upgrade yourself. You wanna hire better than yourself. It''s something that you''re not as good at”Adam Foroughi

Principle

Hire only people with a chip on their shoulder

Comfortable people stop pushing; chip-on-shoulder people compound.

Adam''s own chip: his family was uprooted from Iran. RAF (high-school dropout, hired at 19) and other early hires were outcasts from other places who couldn''t get a job at big tech. The hiring filter is biographical, not credential-based.

Recruit grudges, not résumés.

“one of the things that I always looked at for hiring were find people who have a chip on their shoulder, someone who has a reason to push hard. And I had that myself”Adam Foroughi

Principle

Sell to operators on performance, not to brands on relationships

Performance buyers scale themselves; brand buyers require wining and dining.

AppLovin chose the developer/performance side from day one and was told repeatedly to add a brand sales team. Adam refused. The result: a $140B advertising company with no CRO, no COO, and ~400 people in the core business.

If your customer can prove ROI in their dashboard, you don''t need a sales team.

“if our customers are able to buy on a performance basis… their business is getting better because the ads are well spent. And if you spend your money well on ads and on the other side you can monetize, you create a spread, they become effectively a company that can arbitrage our platform, then they could scale their business.”Adam Foroughi

Principle

Buy your own shares aggressively when cash-flow-to-market-cap inverts at the floor

If FCF/market-cap exceeds 15-20% and you believe in the business, buy back as much as you can.

AppLovin generated $1B+ EBITDA against a $3.8B market cap — a 20%+ FCF yield. Adam treated this as deterministic math: five times cash flow…why don''t we just buy all the shares. The market''s mispricing is the opportunity; the trade is the consequence.

When the market values your cash flow at 5x and you believe the cash flow is durable, every dollar reinvested in your own equity returns 20%+.

“The company at the bottom was worth $3.8 billion market cap and we were generating over a billion dollars of ebitda. So you think about the cash flow to to market cap, it was about one fifth that we were generating every single year. Well, in theory we could buy back 20% of the shares of the company just in the next year if we really believed in the path we were on.”Adam Foroughi

Frameworks

Reusable systems and operating models — including when they help and when they break.

Framework

Vertical-integrate temporarily to seed the data flywheel, then divest once the model can stand alone

When advertisers won''t share data with you, become an advertiser yourself.

AppLovin bought ~15 game studios (2018-2020) to feed Axon. Spun off the entire portfolio to Tripledot in ~2024 once Axon-2 was self-sustaining. Diagnostic: are you in the integrated layer because it''s strategic, or because the platform layer is finally self-sufficient?

Vertical integration is a phase, not a destination.

“we ended up with like 14, 15 different game studios across all categories of gaming that we ended up owning so that we can not only understand what the game developer needs from marketing platform and a monetization platform, but also use their data inside our model… we got to a place where the business had grown so much that every advertiser in mobile gaming was now starting to connect into our models… But once we got to that point we realized we don''t need these studios”Adam Foroughi

Framework

Buyback via known sellers, not the open market

Find the structural sellers and offer to be their exit.

Diagnostic: which holders on your cap table are mandated to sell in the next 24 months (PE fund expirations, departed-employee vest cliffs, founder secondary needs)? Approach them directly with negotiated tender offers. AppLovin retired ~50% of shares via this route over 18 months.

Pre-identify the structural sellers. Be their buyer.

“Most companies go out and say, I''m gonna buy shares from the public markets and just take shares back. But you don''t know who''s on the other side of that trade. On the other hand, we knew that we had a cap table where about 50% of the shares were gonna sell at some point over the coming years… instead of going to the public, we went to the shareholders that we knew were gonna sell and got them to agree to sell back to us over time.”Adam Foroughi

Framework

Compensation segmentation: cash for replaceable, equity for critical, no equity for everyone-else

Equity is a retention tool for critical roles only; for everyone else, pay cash.

AppLovin segmented equity to ~100 critical engineers/product after the 2022 crash. Everyone can buy shares; only the critical 100 must be made whole on equity. Diagnostic: who specifically would the business lose if they walked? Those people get equity; the rest get cash.

Don''t spray equity. Concentrate it where it actually retains.

“if you''re a critical engineer and you''re paid a million dollars a year and your $600,000 of stock based compensation goes down to a hundred grand, you''re still getting paid 500 KA year… what we did at the, the low point for the team was go, some people were just gonna put on cash compensation, everybody at the company can buy shares in the business but we don''t need to force that decision for people that have lower total compensation… We then said, who are the people that are like critical to the business?”Adam Foroughi

Framework

The Performance-Marketing Arbitrage: make the advertiser the arbitrageur, not yourself

Performance marketing = put the customer in the arbitrage seat; their bank balance becomes the only ceiling.

Diagnostic test: (1) Does $1 in spend produce >$1 in measurable revenue within the customer''s payback window? (2) Is the measurement system in your platform (not theirs)? (3) Will they keep spending without a salesperson on the phone? If yes/yes/yes, you have performance-marketing PMF.

Build the dashboard that proves ROAS; the spend follows automatically.

“if you''re a game developer today and you plug into our platform, you''re gonna spend a thousand dollars, you''re gonna know with certainty you made more than a thousand dollars on that spend… they become an arbitrage. The only constraint on them scaling in our system is the money that they have in their bank account.”Adam Foroughi

Signals

What appears to be shifting, for whom it matters, and what happens if you ignore it.

Signal

Axon 2 deep-learning model drove ~62x stock appreciation in <2 years

Deep-learning ad models are the inflection — the era of rules-based + traditional ML is ending fast.

Forward signal: ad platforms that ship deep-learning prediction with self-serve onboarding will absorb spend from platforms that still require manual optimization. Axon-2 launched April 2023 → AppLovin from $4B to $250B market cap in <24 months. Expect comparable inflections at other ad networks shipping equivalent capability through 2026-2027.

The platform that ships deep-learning self-serve performance first wins the long tail of advertisers.

“I remember seeing a graph this like Inflection Point, the business itself inflected big time in, I think it was April, 2023 when we rolled out Axon two… any advertiser can just plug in, get going, spend money, get revenue on the other side and make a spread when you have that product. And it''s very, very scalable. The business can grow a lot… we, we went in from, I think it was April, 2023 was when the stock was 9, 10, 11 bucks to, we peaked it was probably six months ago or so. It''s 750 bucks.”Adam Foroughi

Opportunities

Only included where there is a buyer, a real wedge, and a plausible revenue path — not vague idea theater.

Opportunity

Local SMB and e-commerce advertising on gaming inventory — billion-user audience, no Salesforce

Gaming inventory is grossly under-monetized because it''s only ever been sold to game advertisers.

AppLovin: 1B+ DAU on gaming inventory, US alone 150M adults, average ad watched 35 seconds. Recently extended Axon to e-commerce advertisers ("if you spin up a Shopify store selling lipstick…"). Next leg: SMB (local laundromat) and eventually enterprise. The TAM is "every business" because the audience is "every adult.”

Performance ad inventory built on gaming has nothing to do with games — it has to do with attention.

“the hope we have is that we can get the local laundromat discovered by someone playing a game because we serve a really good ad to someone who needs their clothes wash. If that happens and we get to that level of scale, this business is gonna be much, much bigger than it is today. And the opportunity for us to really impact economic growth for businesses that are small to medium size is gonna be much bigger than it''s today.”Adam Foroughi

Lessons still worth keeping

Useful takeaways that did not fully clear the bar for durable principle status.

Lesson

Fired 40% of staff in 2024 while revenue nearly doubled to concentrate talent density

Layoffs into a doubling revenue line are about org density, not cost.

2024: AppLovin ~100% YoY revenue growth, simultaneously cut ~40% of headcount. Triggered when Giovanni (new CTO) asked "why is this person here?" repeatedly. Cuts were Q-tip: surgical, role-by-role, focused on removing process bloat and B-players around A-players.

Cut into strength to consolidate the win.

“in 23 even, or I think it was about 23 or 24, even though the business was ripping, we were firing a ton of people. We let go of about 40% of the team. I think it was 24. The business grew nearly a hundred percent year over year. And we fired about 40% of the team because I wanted to distill it down to these eight players.”Adam Foroughi

Lesson

Direct-to-consumer pivot failed twice (dating, fashion) before pivoting back into advertising

Founder-market fit beats market opportunity.

2010-2011: Adam wanted to escape B2B and own consumers; built a dating app while married, a fashion app with zero fashion sense. Both failed. The third app (app-discovery, "App-11") became the ad-network kernel. Lesson: pivot toward what you''re actually good at, not what looks more fun.

Build in your circle of competence, not your fantasy.

“we, we built in 2011 we launched a dating app and a fashion app both terrible. So a married guy is building a dating app… zero fashion sense. And I was married at the time for two terrible ideas. Pre Tinder though. And the fashion app we launched right around when Pinterest launched and they were down the block from us. So there was obviously opportunity in those markets. We were not the right teams to do it.”Adam Foroughi

Lesson

Chinese take-private blocked by CFIUS — restructured as convertible note to preserve trust and exit

Restructure deals around regulatory triggers, not the other way around.

2016-2017: Adam''s ~$1.4B take-private to Chinese PE got stuck in CFIUS. He restructured the deal into a $1B convertible note (sub-10% on conversion → no control trigger), dividend-out the cash to existing shareholders as a liquidity event, then refinanced the convert with KKR in 2018. Three-way win: investors paid; team got liquidity; control preserved.

Use convertible debt to thread regulatory eyes-of-the-needle.

“I ended up finding a way to get them a deal that then we could unravel later. We ended up doing it so that instead of their billion dollars being in an equity investment for 70% where they would get control… we ended up pivoting it to a billion dollar convertible note that if on conversion we pay them back the money plus interest, they would get 10% of the business. So under 10% there''s no control… We did that deal, we dividend out a billion dollars to shareholders.”Adam Foroughi

Lesson

Walked away from a $600M acquisition while still doubling — turned $600M into $140B

At 100% YoY growth, today''s acquisition price is yesterday''s news by close.

2015: third-party offered $600M cash. Adam''s internal valuation was ~$1B at that time. He walked. Today AppLovin is ~$140B. The lesson: in a doubling business, anchor on forward value not present offers.

If your business is doubling, the right valuation is the one 12 months from now.

“got as big as an offer for $600 million cash from… I would''ve had a number I would''ve done the deal at, but it was not that number. And so I ended up walking away from it… What''s your market cap today? 140 billion roughly. It would''ve left a lot of money on the table.”Adam Foroughi

The Plays

Try these this week

Verb-first executable actions — each one tied to a stated outcome in the episode.

Acquire competing supply (game studios) to feed a data flywheel the platform needs

Outcome: Buy first-party data producers when third-party data won''t flow.

Context: AppLovin: advertisers wouldn''t share data with them (FB/Google only). They bought ~15 game studios 2018-2020 to seed Axon. Divested to Tripledot ~2024 in a single transaction once Axon-2 worked.

“in order to solve that problem early on I went out and, and we just kicked off buying some gaming studios. So actually like the business from KKR investment to IPO ended up looking wildly different than an advertising business because we ended up with like 14, 15 different game studios across all categories of gaming that we ended up owning so that we can not only understand what the game developer needs from marketing platform and a monetization platform, but also use their data inside our model.”
— Adam Foroughi

3-5 years from acquisition to divestiture per

Before you start

· Identified data gap blocking the model
· M&A budget
· Plan to divest cleanly

strategicgrowth

Run a negotiated buyback against known structural sellers at the floor

Outcome: Negotiate buybacks against known sellers instead of the open market.

Context: AppLovin: $1B EBITDA / $3.8B mkt cap. Identified ~50% of cap table as structural sellers (PE + ex-founders). Deployed $6B (operating cash + debt) over 18 months on negotiated tenders. Result: ~$50-60B proceeds value created.

“for the following 18 months, we ended up deploying somewhere around $6 billion of buybacks of our own capital and we leveraged some to buy back share in the company. And over time, that ended up creating somewhere in the neighborhood of 50, $60 billion of actual proceeds from the buyback.”
— Adam Foroughi

18 months per

Before you start

· Profitable, FCF-positive business
· Stock trading at <8x FCF
· Identified structural sellers on cap table
· Founder/CEO conviction

capital-allocationpublic

Restructure equity into a convertible note to thread regulatory blocking

Outcome: Use convertible debt + dividend to thread regulatory thresholds.

Context: AppLovin restructured a $1B Chinese take-private (70% control, CFIUS-blocked) into a $1B convertible (converts to 10%, sub-threshold). Cash dividended to existing shareholders. Note later refinanced by KKR.

“instead of their billion dollars being in an equity investment for 70% where they would get control, which is what triggers this scrutiny, we ended up pivoting it to a billion dollar convertible note that if on conversion we pay them back the money plus interest, they would get 10% of the business. So under 10% there''s no control and then they can come along for the ride. We did that deal, we dividend out a billion dollars to shareholders.”
— Adam Foroughi

Restructure: weeks. Hold-and-refinance: 1-3 years. per

Before you start

· Regulatory block on original structure
· Counterparty willing to restructure
· Securities counsel comfortable
· Refinancing path identified

financinggrowth

Founder approves every hire personally to force the org to justify each headcount

Outcome: Make every hire require the CEO''s signature; backfills evaporate.

Context: Adam took over HR personally after noticing automatic backfills. Hire rate dropped from hundreds/yr to tens/yr.

“every time we let someone go or someone quit, there was automatically just a new job placement that went up… I took it over and I said, look, convince me that you need the person. So we went from somewhere in the neighborhood of hundreds of, of new hires that we''re doing a year or at least attempting to do a year to in the tens.”
— Adam Foroughi

Apply continuously per

Before you start

· Founder bandwidth to review hires
· Org small enough that 10s of hires/yr is realistic
· Leadership aligned that talent density > headcount growth

org-designgrowth

Bootstrap on an LLC + personal cash through the pivot zone, raise only after PMF

Outcome: Bootstrap through the pivot zone via an LLC — no investors, tax-efficient burn.

Context: Adam: previously wealthy, structured AppLovin precursor as an LLC. Bootstrapped 18 months through dating, fashion, App-11 pivots. Only raised in Nov 2012 ($25M at $4M pre, convertible to common) after the ad-network hit $1M/month run rate.

“I didn''t believe in selling to investors an idea that I wasn''t confident. So I figured if, if I''m out there with this team and we''re just playing around and trying to find a fit during that point in time where I''m gonna pivot a bunch, I don''t wanna deal with investors. So we actually set it up as an LLC early on and the nice thing is you put a million dollars into an LLC, you lose a million dollars, you get back 500 grand.”
— Adam Foroughi

Bootstrap until PMF (Adam: 18 months) per

Before you start

· Founder personal liquidity
· High prior conviction
· Comfort with pass-through tax accounting

financingpre-seed

Distribute the SDK to desperate developers for upfront cash + revenue share

Outcome: Pay cash upfront for SDK distribution when the supply side is cash-starved.

Context: AppLovin 2012: AdMob was building for brands not developers. AppLovin paid ~$10K per developer for SDK placement, combined with developer-friendly performance economics. Captured supply quickly.

“these other app developers were desperate for money… these developers needed money. And so I just went and said, look, you''ve got a million users on this calendar app. We''ll pay you 10 grand today. Just put our ad platform in there. It was up to us to make it spread on it.”
— Adam Foroughi

6-12 months to capture supply per

Before you start

· Cash to fund upfront payments
· A working SDK
· Supply side cash-starved
· Competitive alternative is bad

go-to-marketseed

Decision Moments

Actual decisions, real outcomes

Specific decisions narrated in the episode with their outcomes and transferable lessons.

AppLovin stock had crashed 92% from $115 to $9 (~$3.8B market cap) while EBITDA had grown from $700M (2021) to $1B+ (2022). FCF/market-cap yield inverted to ~20%+. Investors and observers said the business was broken; the founder''s read was that the price was broken.

Did: Deployed ~$6B over 18 months on negotiated buybacks targeted at known structural sellers (PE funds nearing fund-life end, departed founders holding ~50% of cap table). Levered the company against forward EBITDA to maximize deployment. Avoided open-market signaling.Outcome: Retired ~50% of shares; the buyback generated ~$50-60B of proceeds value as the stock subsequently rerated from $3.8B floor to a peak of $250B+ market cap. One of the most successful buybacks in public-company history.

When FCF/market-cap inverts above any reinvestment return and you have conviction, the only economically coherent move is aggressive negotiated buybacks against known sellers — not open-market, not waiting.

Part of an emerging decision pattern across multiple episodes

2015: AppLovin had a $600M cash acquisition offer from a tech company while still growing ~100% YoY at ~$50M EBITDA. Adam''s rough internal valuation was already $1B and rising. No board to pressure-test the decision.

Did: Walked away from the $600M offer. Adam set an internal valuation based on forward growth and refused to anchor on the size of the cash payout to the team.Outcome: Decision preserved the optionality that led to a ~$140B market cap today — a ~$139.4B value creation difference vs. the alternative.

In a 100% YoY-growing business, today''s acquisition offer is yesterday''s news by the close. Anchor on forward fundamentals, not the present life-changing number.

Part of an emerging decision pattern across multiple episodes

2018: AppLovin needed advertiser data for the Axon ML model but third parties refused to share (FB/Google had it; AppLovin didn''t). The instinct was that without first-party data, the model would never catch up.

Did: Bought ~15 game studios across all categories of mobile gaming. Integrated their data into the Axon training pipeline. Used the studios as both training-data source and product-feedback channel. KKR (newly on the board) supported the strategy.Outcome: Axon-1 launched in 2020; Axon-2 (deep learning) launched April 2023 and triggered a ~62x stock appreciation. The studios were sold to Tripledot in a single block ~2024 once they were no longer needed.

When third-party data won''t flow, buy first-party producers — temporarily. Set the divestiture milestone at the same time as the acquisition strategy.

Part of an emerging decision pattern across multiple episodes

2024: AppLovin revenue was growing ~100% YoY but headcount had bloated to ~700-800. New CTO Giovanni kept asking "why is this person here?" Adam saw process bloat surrounding A-players and risk of A-player churn.

Did: Fired ~40% of staff while revenue was doubling. Concentrated equity grants to ~100 critical engineers/product people. Cut C-suite to just CEO/CFO/CTO/GC — no CRO, no COO. Personally took over hire approvals.Outcome: Lean ~400-person core business generating $5B+ run-rate FCF on a $140B market cap; A-players retained; AI tools driving 80%+ of code; one of the highest revenue-per-employee ratios in public markets.

Layoffs into a doubling revenue line are about talent density, not cost. The signal "we''re lean even with this opportunity" attracts and retains the right people.

Part of an emerging decision pattern across multiple episodes

Tensions surfaced

Contradictions and trade-offs the episode raises — judgment calls a thoughtful operator has to navigate.

Tension

Vertical integration vs. platform trust — owning game studios while selling them an ad platform

Vertical integration to seed data is necessary AND it costs you trust — both at once.

AppLovin owned 15 game studios while selling ad services to competing studios. SMB game developers continued because they needed any growth they could get. Larger ones froze out. Adam''s mistake by his own admission: not proactively communicating "we''re doing this for the data, here''s the timeline to exit.”

Vertical integration to seed data is necessary AND trust-destroying. Both. Hold the tension; over-communicate the exit.

“there was a moment in time where a lot of people stopped trusting us in the, in the space because our games were growing really quickly. We were feeding their data into our model… your reputation and trust can break down at any moment if you violate trust on the other side of things… But then once they sat down with me, it was like, okay, now we can overcome that and like get down to business.”Adam Foroughi

Corpus connection

Where this episode fits for retrieval

What kinds of decisions this briefing is best pulled into.

Primary decisions

• capital-allocation
• hire
• fire
• strategic-bet
• m&a
• financing
• pivot