May 22, 2026 · 30m · american-optimist
AI Investing Strategy for 2026 · Joe Lonsdale
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of American Optimist, host Joe Lonsdale interviews Eric Scott, Managing Partner at Overlook Capital, to explore the structural dislocation in modern venture capital created by massive fund concentration. They discuss Overlook's strategy of value investing in overlooked, high-growth tech companies and leveraging AI for operational efficiency.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Joe holds 21.9% of the talking time here. How this is scored →
Joe as informed peer 4.4 Guest teaching 2.6 Guest disagreement 1.5 Joe pushing back 1.3
speaking balance: gold is Joe, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 14:06 Rejecting the magical power law underwriting premise
Eric directly pushes back against the standard VC consensus, arguing that power law outcomes are obvious only in retrospect and rejecting the idea of pretending to underwrite skewed barbell portfolios.
Hardest push from Joe ▶ 14:31 Defending elite track records in picking power law winnersJoe challenges Eric's skepticism regarding underwriting breakout winners by citing the proven long-term track record of Peter Thiel and Founders Fund.
Biggest teaching moment ▶ 16:02 Explaining LP capital concentration and fund size inflationEric educates the audience on market mechanics, outlining how 9x fund size increases among the top 20 funds restrict their investable universe to companies that can absorb billion-dollar checks.
Joe holds their own ▶ 18:37 Demonstrating market pricing dynamics with live portfolio dataJoe demonstrates his deep operating and investing expertise by providing real-time data from an 8VC portfolio company scaling past $100M revenue facing compressed growth multiples.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Joe as informed peer | Guest teaching | Guest disagreement | Joe pushing back | Why |
|---|---|---|---|---|---|---|
| Early Career, HVF Labs & Meeting Max Levchin | 2 | 2 | 1 | 0 | Eric shares the origin story of how a cold email got him hired as Max Levchin's technical assistant at HVF Labs. The host listens warmly and prompts the narrative without friction. | |
| Evolution of Prototyping & AI's Role in Product Management | 3 | 2 | 1 | 1 | Joe notes how rapid prototyping has changed with AI, while Eric explains how AI amplifies core PM skills and reflects on starting companies with zero capital. | |
| Building Sci-Fi VC, Cynthego & Harbor Health | 3 | 2 | 0 | 0 | Eric details spinning out Sci-Fi VC and backing Cynthego, recognizing his edge was enabling deeply technical founders rather than being the CEO. | |
| Lessons in Incubation & Founder Commitment | 5 | 2 | 1 | 1 | Both host and guest align on the absolute necessity of full founder commitment during startup incubation, drawing on their joint history at 8VC and Harbor Health. | |
| Macro Shifts in Venture Capital & Power Law Dynamics | 6 | 4 | 3 | 3 | Eric challenges conventional power-law dogma by arguing it is only visible in retrospect and that top investors win via lower loss ratios, prompting Joe to defend Peter Thiel's deliberate selection track record. | |
| Portfolio Concentration Data & Market Inefficiencies | 6 | 3 | 2 | 2 | Eric describes how mega funds ignore sub-trillion dollar opportunities, while Joe validates this market dislocation with a concrete example of an 8VC portfolio company scaling to $100M revenue. | |
| Underwriting Returns, Concentrated Portfolios & AI P&L Effects | 5 | 3 | 2 | 1 | Eric breaks down the underwriting math of 3-5x targets yielding 25-30% IRR, explaining that AI should collapse operating costs on the P&L rather than replace traditional business moats. | |
| The GP Diligence Framework & Lean Firm Operations | 5 | 3 | 2 | 2 | Eric explains Overlook's diligence question to GPs and his decision to keep the partnership ultra-lean, capping fund size to prevent strategy drift. |
Statements from this episode (17)
Scott: Bill Gates's technical assistants became Microsoft's next-generation leaders
“Bill Gates had a bunch of these. They became the next generation of leaders at Microsoft.”
Scott: AI Amplifies Core Product Management Skills and Prototyping
“I think it bifurcates the skills a little bit. So the things that made a really good product manager, a good product manager then are only amplified. So deep understanding of the customer, ability to articulate your ideas to a team with varied skills. AI only …”
Scott: Office playgrounds are a deeply negative signal for company culture
“And you know, now I'm 37 and I realized that's actually like a deeply negative signal for the culture of the company.”
Scott: The individual incubator determines startup success, not process
“The person incubating the company is actually still the number one determinant of success. There's no like magic way to do this that you can write on a whiteboard and say.”
Scott: Annual VC deployment has increased tenfold over the past 10–15 years
“Roughly speaking, we're talking about a 10 X increase in the number of venture dollars that are put to work every year from, call it, 10 or 15 years ago.”
Scott: Venture power law dynamics are only true in retrospect
“The catch here is that the power law is only true in retrospect and it is very obviously true.”
Scott: Global venture investments are on track to exceed $1T in 2026
“So the, these numbers are kind of apples to oranges, but I think they're, we're on track in 20, 26 to clear a trillion dollars of venture capital. Globally.”
Scott: Top 20 VC fund sizes grew 9x due to capital concentration
“The LP dollars that are going into venture are concentrated into fewer and fewer funds. Those fund sizes have basically nine acts on average. If you just look at the top 20 funds”
Scott: Mega-funds face a narrowing universe able to absorb billion-dollar checks
“There aren't that many companies that can withstand a billion dollar infusion, which means The universe where they can deploy these dollars just becomes more and more narrow.”
Scott: Anthropic, Anduril, and SpaceX will be generation-defining companies
“Anthropic and Anduril and SpaceX. These are phenomenal companies. They will be generation defining companies, and it is a good strategy to give money to ATVC and Founders Fund and Thrive to invest in these companies.”
Scott: Private companies beyond the top 50 trade at value-investment multiples
“However, if you look just beyond sort of the top 50 private companies, you go from incredibly high multiples to extremely reasonable multiples that look almost, almost like they're very good value investments.”
Scott: Mega-funds neglect hundreds of startups growing 50% to 150% annually
“And there are hundreds of companies out there that are working on big, super important problems with great teams. They have high product market fit. They have growth rates that are 50 to 150% year over year. But they might not be a trillion dollar company. Tho…”
Scott: AI will slash enterprise operating costs without destroying structural moats
“We don't want to see AI impacting their moat. So if they have a network effect, a cost of capital advantage, a regulatory moat, Those things are not going to go away with AI. However, their operating costs should be falling off a cliff.”
Scott: VCs rarely choose their hottest startups for personal net worth bets
“We would ask them, if you had to put all of your net worth, liquid net worth into one company, which company would it be? And that company was rarely the hottest company in their portfolio, in part because they already had massive e-liquid exposure to that com…”
Scott: Overlook Capital will use AI analysts and eliminate partner tracks
“AI is our analyst. We will eventually bring on one analyst associate or principal at a time, but that is primarily to give back to the ecosystem. We will not be a firm that has sort of a Partner track.”
Scott: Now is probably the best time to start a company since 2013
“I think this is probably the best time to
Start a company or be running a company, at least since I've entered the workforce in.”
Scott: The narrative that only AI, defense, and energy matter is untrue
“What you're seeing in the headlines today is basically, Hey, AI is all that matters.
Maybe defense matters because the world is becoming a slightly more volatile place, and energy matters as a derivative of AI.
So if you're not already super exposed to those i…”