Dan Sundheim - The Art of Public and Private Market Investing - [Invest Like the Best, EP.460]

Episode Summary

Executive Summary: Patrick O'Shaughnessy interviews Dan Sundheim, founder of D1 Capital, on public vs. private markets, AI's business models, hyperscalers, software disruption, SpaceX, and managing through GameStop. Sundheim argues AI is a once-in-a-generation productivity shift, but one that creates new concentration, capital-intensity, and geopolitical risks that investors must navigate.

Main Topics: Public vs. private markets (Priority: 5/5): Sundheim says private markets are less competitive but public markets are more distorted by short-term incentives. LLMs as a business model (Priority: 5/5): He frames LLMs as a mix of Netflix and Spotify: heavy upfront spend, high-margin distribution, and personalization. AI competition and company strategy (Priority: 4/5): He thinks only a few model providers will matter long term, and focus versus multi-market expansion is a key strategic tradeoff. Hyperscalers, chips, and infrastructure (Priority: 5/5): He expects cloud margins to face pressure as AI customers concentrate and model firms increasingly insource compute. Software disruption and market repricing (Priority: 4/5): He sees software as the first public-market sector to face real AI-driven short pressure, while system-of-record businesses remain protected for now. GameStop and firm resilience (Priority: 4/5): He recounts how the 2021 drawdown forced a less risk-prone portfolio style and a more transparent investor relationship. Geopolitics, growth, and leadership (Priority: 5/5): He is bullish on AI-driven growth but worried about Taiwan/China semiconductor risk and the societal impact of superintelligence.

Key Arguments: Private markets offer fewer competitors, but public markets are more efficient in ideas and more distorted by short-term behavior. AI makes public and private investing more synergistic because private leaders reveal where technology is heading. OpenAI and Anthropic are not commodities; margins and product specialization suggest durable LLM businesses. LLM economics resemble Netflix/Spotify: huge fixed training costs, high incremental margins, and sticky personalization. Most winners in AI will be a few large model companies because capital requirements and scale advantages are enormous. Hyperscalers may grow faster near term, but margins likely compress as AI customers become concentrated and insource compute. Software stocks are vulnerable first, but system-of-record apps like ERP/CRM should be slower to displace. GameStop forced D1 to reduce risk, hit singles and doubles, and rebuild investor trust over time. AI could drive stronger economic growth and disinflation, but also unprecedented social and labor disruption. Taiwan's semiconductor dominance creates a severe geopolitical tail risk that could trigger depression-like global damage.

Data Points: Expense reviews automated: 85% - Ramp's AI claim in the ad read before the interview. Expense review accuracy: 99% - Ramp's AI claim in the ad read before the interview. Company savings: 5% - Ramp's ad read says it saves companies 5%. Initial OpenAI investment round: $125 billion round - Sundheim says D1 originally invested in OpenAI at this valuation round. Private-company synergy at firm founding: 25% of the time - He says private investments overlapped with public ideas about a quarter of the time early on. D1 capacity managed at peak at Viking: just over half of Viking's capital - He says he managed more than 50% of Viking capital before founding D1. June investor dinners: 4 nights - He held semi-annual dinners with LPs over four nights after the drawdown. Trough of drawdown: end of May 2022 - He references this as the low point before the June 2022 LP dinners. Historical semiconductor concentration: 90-something percent - He says Taiwan produces most of the world's most advanced semiconductors. Potential telecom impact: 97, whatever percent - He says Starship could dramatically cut launch costs by this magnitude.

Pivotal Quotes: "The public markets are the most competitive in the world, even though they are less efficient than they were before." — Dan Sundheim: He compares competition and efficiency across public and private markets. "I think the real debate is these are extremely capital-intensive businesses." — Dan Sundheim: He reframes the central question for LLM economics. "We are now the start of a mission to dramatically improve our returns, improve our firm and earn back our reputation as being great investors." — Dan Sundheim: He describes the reset after the GameStop period and the new posture of D1.

Implications: The unresolved question is how fast AI adoption, geopolitics, and capital intensity reshape returns; investors should focus on moat, concentration, and resilience.