Video · 2025-06-20 · 1h 1m · 6 moments

Coatue’s Laffont Brothers. AI, Public & VC Mkts, Macro, US Debt, Crypto, IPO's, & more | BG2

✦ AI generated

timeline · colored by role

Claim◆

Bad venture investments in the right long-term trend can poison your judgment about that trend even when the underlying thesis was correct.

Philippe Laffont opens by warning that a failed bet on the wrong company can wrongly convince investors the whole trend (like crypto) was a mistake.

transcript

Philippe Laffont: Sometimes you make some venture bets and they don't work and then you're like [ __ ] I just invested in the wrong trend and in fact sometimes you invested in the wrong company but it is the right trend and those bad investment cloud your judgment.

Data◆

Bitcoin's total market cap is tiny relative to global net worth, so it plausibly could grow several times larger even as a rough 'company' comparison.

Philippe Laffont frames Bitcoin against global net worth (~$450-500T), equities, real estate, and gold to argue it could realistically be 2-3x larger.

transcript

Philippe Laffont: the market cap of the world, the net worth of the world is like 450 500 trillion. Equities I think are like 120 or stuff. Real estate's probably another 100 150. Then there's a value that people have in their homes. Uh gold is about 15 to 20 trillion uh above and under uh the ground. And then we're like Bitcoin at two.

supports · 1

Mechanism

By joining anonymized credit-card and email-receipt data, Coatue could numerically prove that ChatGPT subscribers reduce their Google page views over time.

Philippe Laffont describes how Coatue joined two proprietary consumer datasets to prove, not just anecdotally sense, that ChatGPT usage cannibalizes Google search page views.

transcript

Philippe Laffont: what we try to do here is as part of the data science that we do, we process probably a 100 million credit card receipts a day. So we have a very fine view of what the US consumer does. And we have another data set where we know what consumers do based on their email receipts. And the trick was to try to join those two data sets.

Prediction◆

If AI drives a 1990s-style productivity boom (2.5-3.5% annual productivity growth), US debt-to-GDP could bend down toward 80% instead of rising to 140%.

Philippe Laffont lays out Coatue's analysis that sustained AI-driven productivity gains comparable to the 1990s internet/PC era could meaningfully lower, not just stabilize, the US debt-to-GDP ratio.

transcript

Philippe Laffont: we said what would it take for actually debt to GDP to stay at 100 or maybe even bend the curve and go down to 80. And what's really surprising I think if you just show maybe the next slide or or so yeah uh I think if we move forward yeah just a little bit you'll see that if productivity for the next decade or so was about 2 and a half to three and a half% per year we could achieve substantial reductions in this key ratio of debt to GDP.

extends · 1provides context · 1

Claim

Meta paying a huge premium for only 49% of Scale AI reflects both the enormous size of the AI prize and the extreme speed of competitive escalation in AI.

Thomas Laffont explains Zuckerberg's unusual 49%-stake, full-price deal for Scale AI as rational once you weigh the trillion-dollar scale of the AI prize against Meta's own market cap.

transcript

Thomas Laffont: I tend to think it's related to two factors, right? One is the size of the prize. So I think he clearly sees that this is the biggest prize in tech in the world frankly. And so I think relative to his 15 billion to all of us is a massive number probably in the scale of the multi-trillion opportunity that he sees he might just think it's a bet I would make all day long.

Highlight slides

Bad Bets Can Poison Good Theses✦ from: Bad venture investments in the right long-term trend can poison your judgment about that trend even when the underlying thesis was correct.Company Risk vs. Trend Risk✦ from: Bad venture investments in the right long-term trend can poison your judgment about that trend even when the underlying thesis was correct.Bitcoin vs. Global Wealth✦ from: Bitcoin's total market cap is tiny relative to global net worth, so it plausibly could grow several times larger even as a rough 'company' comparison.Room to Grow: Asset Comparison✦ from: Bitcoin's total market cap is tiny relative to global net worth, so it plausibly could grow several times larger even as a rough 'company' comparison.GPU Allocation vs. Cloud Revenue Share✦ from: Mapping Nvidia GPU allocation against cloud revenue market share reveals which cloud providers are winning or losing the AI infrastructure race, including a surprising reinvention of Oracle.Amazon: Under-Indexed on GPUs✦ from: Mapping Nvidia GPU allocation against cloud revenue market share reveals which cloud providers are winning or losing the AI infrastructure race, including a surprising reinvention of Oracle.Oracle's Reinvention✦ from: Mapping Nvidia GPU allocation against cloud revenue market share reveals which cloud providers are winning or losing the AI infrastructure race, including a surprising reinvention of Oracle.AI Productivity Boom Could Bend Debt Curve✦ from: If AI drives a 1990s-style productivity boom (2.5-3.5% annual productivity growth), US debt-to-GDP could bend down toward 80% instead of rising to 140%.The Productivity Threshold✦ from: If AI drives a 1990s-style productivity boom (2.5-3.5% annual productivity growth), US debt-to-GDP could bend down toward 80% instead of rising to 140%.

Related episodes

Video · a16z Podcast

Ben Horowitz on the Global Race for Tech, Power, and Influence

→

Audio · All-In Podcast

SpaceX-Cursor Deal, SaaS Debt Bomb, New Apple CEO, SPLC Indictment, Colon Cancer Spike

→

Video · Lenny's Podcast

The rise of taste, human authenticity and judgment in an AI world | Adam Mosseri (Head of IG)

→