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

Overview of Invest Like the Best — EP.460: Dan Sundheim — The Art of Public and Private Market Investing

Patrick O’Shaughnessy interviews Dan Sundheim, founder and CIO of D1 Capital Partners, about running a hybrid public/private investment platform, how private markets and AI-driven private companies change the public market opportunity set, the economics and risks of large LLM businesses, lessons from crises (GameStop), and macro tail risks (notably semiconductors/Taiwan). Sundheim explains his investing framework, firm changes after 2021–22 turbulence, how he thinks about leadership and loyalty, and why he’s optimistic about productivity-driven economic growth even as he flags serious geopolitical risks.

Key takeaways

• Private vs public markets

  • Private markets are less crowded and can be less efficient in certain ways; access and relationship with founders are key constraints.
  • Public markets remain extremely competitive in aggregate — “everyone’s playing a different sport” — creating opportunities for medium-term, fundamental investors.

• AI / LLMs as businesses

  • LLM firms are extremely capital intensive; the big question is whether scaling laws produce attractive returns on that capital.
  • Competition appears likely to consolidate to 4–5 durable players; differentiation will come from personalization, data history, and distribution rather than raw answer quality.
  • Useful analogies: a mix of Netflix (large fixed up-front investment amortized over growth) and Spotify (personalization/retention as the moat).

• Strategy for AI companies

  • Focus versus breadth is a central trade-off: trying to do everything risks being mediocre everywhere; focus can create a durable enterprise position.
  • Sundheim encouraged earlier consideration of ad revenue—companies that refuse it may cede important monetization levers.

• Hyperscalers and compute

  • Short-to-medium term: hyperscalers (AWS/Azure/GCP) will grow due to AI demand, but their long-term economics may worsen as AI workloads concentrate and some AI firms insource compute.
  • New GPU-specialized cloud players (neo-clouds) can be viable for some time; chipmakers have incentives to support diverse customer bases.

• Macro & geopolitics

  • The single biggest tail risk Sundheim highlights: semiconductor concentration (advanced fabs centered in Taiwan). A Taiwan disruption could trigger depression-scale economic damage.
  • Replicating the advanced semiconductor supply chain globally is a long, multi-year project; geopolitical tensions raise material risk.

• Organizational lessons & resilience

  • After the GameStop-led drawdown, D1 shifted portfolio construction to be less tail-risked and to emphasize steady returns; Sundheim emphasizes communication with LPs and inward focus on execution.
  • Loyalty matters in selecting partners and employees, but competence is the primary bar.

Topics covered

• Differences and synergies between public equities and late-stage private investing
• Why private AI/tech companies (OpenAI, Anthropic, SpaceX, Ramp) matter for public investors
• LLM business-model economics: capital intensity, scaling laws, TAM, moats (personalization, enterprise stickiness)
• Netflix/Spotify analogies for LLM monetization and retention
• Advice for AI companies: focus, monetization (ads), enterprise vs consumer pathways
• Hyperscalers, neo-clouds, GPU compute strategy and economics
• Sector implications (software, systems of record, ERP/CRM)
• Shorting, market efficiency, and who’s transacting (passives, retail)
• Personal stories: Orthodontic Centers of America short, getting hired, Viking tenure, founding D1, GameStop crisis, Rivian vs SpaceX decisions
• Leadership, loyalty, and culture in building an investment firm
• Geopolitical tail risk centered on semiconductors/Taiwan

Notable quotes & distilled insights

• “Public markets are the most competitive in the world…there are tons of people competing, but they’re all playing a different sport.” — explains why public markets still offer structural mispricings for different play styles.
• “These are extremely capital‑intensive businesses, capital‑intensive to a degree we’ve never seen before.” — on LLMs; the risk is how returns on that capital evolve.
• “LLMs are a combination of Netflix and Spotify.” — upfront fixed-cost build (Netflix) + personalization/stickiness (Spotify).
• On the GameStop crisis: changing compensation and portfolio construction to reduce short‑term tail risk was essential — focus on hitting singles and doubles to rebuild credibility.
• “The single biggest tail risk facing the global economy right now is semiconductors/Taiwan.” — a geopolitical supply‑chain risk with systemic economic consequences.

Actionable recommendations (for investors, founders, and executives)

For investors:

• If you invest in public companies materially affected by AI, develop a view on private AI leaders and the technology path (closed-loop insight between private and public portfolios).
• Look for businesses with durable moats: low-cost producers at scale or strong personalization/data advantages.
• Be mindful of asymmetric market participants (retail, passives) — they create medium-term inefficiencies short-term traders miss.

For AI founders / execs:

• Decide early whether to focus on a single end market (build an A-team) or attempt broad expansion; focus is often safer to secure a durable position.
• Build data advantages and personalization — that’s likely to be a lasting moat.
• Consider monetization options (including ads) earlier — cultural resistance can lag economic reality.
• Plan for capital intensity: raise and spend sustainably, understand the sensitivity of your roadmap to slower revenue adoption.

For policymakers / corporate strategy:

• Accelerate strategies to diversify and replicate advanced semiconductor capacity to reduce systemic risk (but accept this is a long-term effort).
• Prepare for potential compute insourcing by large AI firms and the changing landscape of cloud economics.

Memorable stories & illustrative examples

• Orthodontic Centers of America short: Sundheim’s early Value Investors Club write-up (anonymous) identified aggressive capitalizing of expenses; the piece moved the stock and helped him land an investment job.
• GameStop (2021) drawdown: intense drawdown led D1 to change comp/portfolio construction and re‑communicate strategy to LPs (“hit singles and doubles”), a pivotal resilience moment.
• Rivian vs SpaceX: two simultaneous big private bets that illustrate different capital and scaling risks — EV manufacturing and scale challenges (Rivian) versus engineering-driven launch cost advantage (SpaceX). Starship’s success materially changed SpaceX’s TAM towards global broadband/telecom.

Risks & watchlist

• AI scaling laws: if returns on training cost decline or enterprise adoption lags, the economics of LLM companies could disappoint investors.
• Hyperscaler margins & customer concentration: increasing AI workloads may compress traditional cloud economics and encourage in‑house compute for very large AI firms.
• Geopolitical risk around Taiwan/semiconductors — potential for massive global economic disruption if advanced semiconductor supply is concentrated or contested.
• Market structure: retail & passive flows, social media narratives, and short-duration quant strategies make markets noisier and create both opportunities and traps for fundamentals-driven investors.

Guest & host — brief bios

• Dan Sundheim — Founder & CIO of D1 Capital Partners; runs a hybrid public/private strategy with large private stakes (SpaceX, OpenAI, Anthropic) and a diversified public equity portfolio. Known for deep fundamental analysis, active shorting, and hands-on founder relationships.
• Patrick O’Shaughnessy — Host of Invest Like the Best; CEO of Positive Sum, curator of Colossus publication and podcast platform.

Closing summary

This episode is a wide-ranging primer on operating a cross‑market investment platform in an era of rapid technological change. Sundheim marries macro and micro perspectives: he’s bullish on productivity-driven growth from AI but warns of capital-intensity risks and severe geopolitical tail risks (semiconductors/Taiwan). For investors and founders, the practical lessons are to (1) form informed views on private AI leaders, (2) prioritize focus and durable differentiation, (3) plan for heavy capital cycles, and (4) manage organizational risk and LP communication proactively — especially in crises.