Bootstrapping a Business to $5 Billion in Free Cash Flow | AppLovin's Adam Foroughi

Key points

• **Walked away from a $600M all-cash acquisition offer in 2015.** Business was growing 100% YoY at $50M revenue. 'Between the time of seeing the number and getting to a deal, the business had grown a ton. I'm a finance person — I would have done the deal at a higher number, but not theirs. So I walked away.' At today's $140B+ market cap, would have been a $1B-vs-$140B counterfactual. Useful base rate for founders evaluating mid-stage acquisition offers.
• **The Chinese cross-border-deal mistake (2016) — and why no board nearly cost the company.** Foroughi announced a $1B / $1.4B-valuation private-equity round from Orient Hontai Capital (Chinese, turned out 'partially state-owned'). CFIUS spent >12 months pressing him on the data-platform implications. 'Persian CEO + Russian CTO selling 70% control to a Chinese state-owned PE fund — it was cold.' Got out by restructuring to a $1B convertible note (10% equity on conversion = under the control threshold) and dividending the proceeds. KKR cleaned it up in 2018 — Foroughi's first-ever board. **The lesson: no-board freedom is a real benefit (no veto on walking away from $600M) but also a real cost (no veto on entering the Chinese deal).**
• **The buyback math, precise this time.** $6B deployed over 18 months → produced ~$50-60B of value at today's market cap. Mechanism: identified the ~50% of the cap table that was going to sell anyway (legacy private equity + ex-founders) and bought directly from them rather than the open market — same playbook Foroughi described on 20VC last issue but with hard numbers attached. ~⅓ of today's $140B market cap is from this single capital-allocation move.
• **Axon 2 inflection — April 2023, stock $9 → $750 peak.** Pre-Axon-2: rules-based algo ('people who play solitaire also play poker'). Axon 1: traditional machine learning. Axon 2: deep-learning model that 'turned the advertiser into an arbitrageer' — they put in $1k, get $1k back in 30 days, then $2-3k by 6 months. **The performance-marketing thesis stated precisely:** 'Very few companies define performance marketing as making the advertiser an arbitrageer. Facebook does. Google partially. We do entirely.'
• **Studio acquisitions (2018-2025) as a data-acquisition strategy, not a games strategy.** Bought ~15 mobile gaming studios primarily to get advertiser-side data for training Axon 2 (advertisers wouldn't share with them but would with Facebook/Google). Foroughi doesn't play games himself — 'I suck at games, so I don't play them.' 5 years owned; sold the whole portfolio to Triple Dot (UK) in 2024-25 for one lump sum at a discount to enable a clean exit. Headcount on those studios was 1,500+ distributed globally; selling cleared the structural-overhead concentration.
• **The 'why is this person here' culture rebuild (2023-24).** Giovanni (current CTO) joined Nov 2022 and started questioning every position. 'I either had to address these questions or risk losing talent like him.' Result: **40-50% layoffs during a triple-digit-growth year** — Issue 03 framing. Coupled with comp-tier reset: top ~100 people got equity, everyone else moved to cash + ESPP. Foroughi took over HR personally and made every new hire approval pass through him. 'From hundreds of job postings to tens — people were spooked about convincing me they needed the role.'
• **~80%+ of code is AI-generated — 18 months ago.** 'Now much higher than that.' Direct cross-reference to Lemkin's 80-90% and Toby Lütke's >50% from the same week. **Cross-issue confirmation: ~80% AI-coding is now the operator-class default across martech, ecommerce platforms, and prediction-market exchanges.** Foroughi adds: 'Your 10x engineer paired with AI might be 100x — not linear. Your B player paired with AI is maybe 2x. The gap widens, not closes.'
• **Senra's framework distillation: 'ruthless efficiency and hyper-competence.'** Foroughi's response: 'Maybe we got lucky that VCs told us no — when VCs tell you yes, they want you to hire more people. We never had that pressure.' The contrarian read on VC: well-funded startups are encouraged to dilute their IQ by overhiring. Bootstrapped or modestly-funded ones default to leaner ops.
• **Trillion-dollar comp triggers exist.** Key-people incentive plan has compensation triggers all the way up to $1T market cap. 'If you believe in the future plan and think it could get that big, voice it to your team. They've seen that if I voice something big and keep working, we eventually get to a good place.' This is the operator-side of the All-In monopoly thesis from this week — Foroughi is actively planning for it.
• **The Raph story.** Started working for Foroughi at age 16 (high-school dropout). 18 years later still works for him, now runs marketing and growth at AppLovin, billionaire today. 'I can call him at 11pm on a Sunday and he picks up.' Chip-on-the-shoulder hiring criterion in action — 'find people who have a reason to push hard.' Foroughi's own version: family uprooted from Iran at age 4, dad had a major real-estate company they lost.
• **Float discipline at IPO and IPO mechanics insight.** Shopify-comparison: AppLovin sold only 7-8% float at IPO. 'Probably a mistake — too low a float makes you more volatile. If you push valuation too high you can't fill a 10-15% float and you end up with worse investors and worse volatility.' Direct contrast with Toby Lütke's IPO-pricing story this week — both had to learn the price-discovery mechanism the hard way.
• **e-commerce + Connected TV expansion path.** Now serving e-commerce ads to the same billion-plus mobile-gaming user base. 'The power user who plays Candy Crush is a middle-aged person with time, willing to watch a 35-second ad.' Aspiration: 'the local laundromat getting discovered by someone playing a game.' If that works on a billion-DAU base, the implied TAM expansion is what supports the $1T target. **CTV is the next adjacency — performance ads at TV scale is the 'holy grail' Foroughi flagged on 20VC.**

Notable quotes

Themes

• Senra's biographical layer on the Issue 03 AppLovin story
• The $6B buyback that became $50-60B in value — exact mechanics
• Chinese cross-border deal misstep as no-board cautionary tale
• 10x engineer × AI = 100x widens the talent gap, doesn't flatten it
• Trillion-dollar comp triggers as the operator-side monopoly play