Feb 24, 2026 · 1h 26m · invest-like-the-best
Inside Dan Sundheim's Bets on Anthropic, OpenAI, and SpaceX · Invest Like The Best
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 interview on 'Invest Like The Best', D1 Capital founder Dan Sundheim shares detailed perspectives on public and private equity investing, the economics of foundational AI models, and portfolio risk management. He reflects candidly on firm turnarounds, evaluating scaled business moats, global geopolitical risks, and his personal career evolution.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Patrick holds 19.2% of the talking time here. How this is scored →
Patrick as informed peer 5.4 Guest teaching 6.1 Guest disagreement 1.6 Patrick pushing back 1.1
speaking balance: gold is Patrick, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:21:10 Pushing back on Warren Buffett's leadership view
Dan explicitly disagrees with Warren Buffett's famous doctrine that the business matters more than the leader, asserting that in 5 to 10 year investing timeframes, talent and leadership drive technological outcomes.
Hardest push from Patrick ▶ 27:00 Patrick challenging the software apocalypse narrativePatrick pushes past simple bull narratives on AI compute to question what happens when software stocks get decimated and the broader real economy has to digest disruptive technology.
Biggest teaching moment ▶ 23:45 Schooling on hyperscaler unit economics and insourcingDan reframes the consensus view that hyperscalers will perpetually win from AI, demonstrating how LLM customer concentration and GPU inference advantages will incentivize models to insource their own compute.
Patrick holds their own ▶ 13:57 Patrick's Netflix fixed-asset amortization analogyPatrick synthesizes complex capital cycles by offering the Netflix upfront asset amortization framework, which Dan immediately embraces and expands upon with the Spotify personalization dynamic.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Patrick as informed peer | Guest teaching | Guest disagreement | Patrick pushing back | Why |
|---|---|---|---|---|---|---|
| Contrasting Public and Private Equity Markets | 5 | 5 | 1 | 1 | Patrick sets the stage by asking Dan to contrast public and private markets in valuation and market psychology. Dan explains the structural difference between private investors pursuing intrinsic value versus public markets where different market participants play completely different games. | |
| AI-Driven Synergies Between Public and Private Markets | 6 | 6 | 2 | 1 | Patrick prompts Dan on how private market holdings inform public investing views. Dan breaks down the unprecedented synergies driven by AI and reflects on early pattern matching with Anthropic's Dario Amodei and Jeff Bezos's early letters. | |
| Sponsor Messages: Vanta and Ridgeline | 5 | 7 | 2 | 1 | Following sponsor breaks, Patrick asks how Dan frames the LLM business model debate. Dan delivers a comprehensive masterclass on capital intensity, scaling laws, API commoditization myths, and operational leverage. | |
| The Netflix and Spotify Framework for AI Models | 6 | 6 | 1 | 1 | Patrick proposes a Netflix analogy regarding upfront capital amortized over huge user bases. Dan validates and expands the mental model, pairing Netflix's high fixed-cost economics with Spotify's personalization moat over commodity content. | |
| Strategic Focus, End-Market Expansion, and Advertising in AI | 5 | 6 | 2 | 1 | Patrick asks about advice for AI executive leaders managing massive compute scale. Dan analyzes OpenAI's multi-market approach versus Anthropic's tight enterprise focus, noting his early push for OpenAI to implement ad-supported monetization. | |
| The Future of Cloud Hyperscalers and AI Infrastructure | 6 | 7 | 2 | 1 | Patrick brings up reports of Anthropic securing gigawatts of power and questions if hyperscalers are at risk. Dan explains why hyperscalers may see margin compression as foundation models insource compute and become better at GPU inference than traditional cloud providers. | |
| AI Disruption in Software, Moats, and Real-World Economics | 6 | 7 | 2 | 2 | Patrick probes how the real economy will absorb AI after the initial software sell-off. Dan discusses moats, identifying why systems of record like ERPs will withstand short-term disruption while software overall faces margin pressure. | |
| Navigating Career Choices and Mindsets for the Next Generation | 6 | 6 | 2 | 1 | Patrick asks about advice for younger generations entering finance and probes the GameStop drawdown period. Dan shares deeply personal reflections on holding investor dinners at the trough of their drawdown and reframing D1's risk strategy. | |
| LP Redemptions, Accountability, and Rebuilding Trust | 5 | 5 | 1 | 1 | Patrick asks about Dan's feelings toward LPs who redeemed capital during the crisis. Dan explains his philosophy of extreme accountability and why capital legitimately leaves when performance falters. | |
| Investor Psychology, Macro Productivity, and Human Purpose | 5 | 6 | 2 | 1 | Patrick asks about the psychological balance required for investing and Dan's macro outlook. Dan expresses optimism for disinflationary AI productivity but skepticism about universal basic income, arguing human wiring requires purposeful work. | |
| High-Stakes Tech Investments: Rivian versus SpaceX | 5 | 7 | 1 | 1 | Patrick asks Dan to unpack his paired bets on Rivian and SpaceX. Dan dissects why auto manufacturing remains an unforgiving, capital-intensive business versus SpaceX's unprecedented engineering feat and Starlink's expanding telecom TAM. | |
| The Art of Short Selling and Exploiting Structural Inefficiencies | 6 | 6 | 2 | 1 | Patrick asks Dan why he remains so passionate about short selling despite market headwinds. Dan outlines how passive investing, retail flows, and pod-shop time horizons have made markets less fundamentally efficient, creating abundant short opportunities. | |
| Enterprise Value and Cash Flows in Asset Management | 5 | 5 | 1 | 1 | Patrick questions whether Dan thinks about building long-term enterprise value into D1. Dan candidly explains that hedge funds are cash-flow machines without terminal value and that his sole focus is performance rather than scaling headcount. | |
| Origin Story: Value Investors Club and Orthodontics Centers Case Study | 6 | 6 | 1 | 1 | Patrick asks about Dan's early Value Investors Club post on Orthodontics Centers of America. Dan recounts how uncovering basic accounting fraud via unit economics launched his hedge fund career. | |
| Career Evolution at Viking and the Decision to Launch D1 | 5 | 5 | 1 | 1 | Patrick asks about Dan's decision to leave Viking Global after reaching CIO. Dan details his career progression across sectors and explains why managing more than half the firm's capital made launching D1 the natural next step. | |
| Art, Aesthetics, and Investing as an Art versus Science | 5 | 6 | 1 | 1 | Patrick inquires into Dan's interest in art and aesthetics. Dan explains why he views investing as more art than science and why low-cost scaled producers like SpaceX and Costco represent the most beautiful business models. | |
| Underappreciated Global Markets: Asia, Japan, Korea, and Defense | 5 | 7 | 2 | 1 | Patrick asks about underappreciated global opportunities. Dan points to defense and industrial assets in Japan and Korea, followed by a sober warning regarding US-China geopolitical risks surrounding Taiwan's semiconductor supply chain. | |
| Lessons from History and the Real Dictators Podcast | 5 | 6 | 2 | 1 | Patrick asks about takeaways from the Real Dictators podcast and CEO leadership. Dan pushes back on Warren Buffett's maxim that business quality trumps management, arguing that over 5-to-10-year horizons, great leaders are paramount. |
Statements from this episode (48)
Sundheim: Private investors share goals; public market players play different sports
“So there's fewer people competing, but they're all doing the same thing. Whereas the public market, there's tons of people competing, but they're all playing a different sport.”
Sundheim: D1 Capital Originally Invested in OpenAI at $125 Billion Valuation
“We invested originally at the hundred and twenty-five billion dollar round, so I don't think people were entirely sold on LLMs as a business model, but if you want to invest in LLMs as a business model, OpenAI was kind of the, you know, it was the one.”
Sundheim: Investing in the Second Market Player Usually Fails
“In most industries, ah, investing in the second player is not, ah, the path to glory.”
Sundheim: Written Communication Is the Best Test of CEO Clarity
“I place a lot of weight, rightly or wrongly, on clarity of thought and the ability to communicate as a CEO, like, what you want to achieve and how you're gonna achieve it, and especially in written form, because taking the time to write something down, you act…”
Sundheim: Dario Amodei Communicates With Better Clarity Than Anyone Since Jeff Bezos
“And Dara just did that better than almost any CEO I've seen since Bezos.”
Sundheim: LLM API businesses are durable and differentiated like cloud providers
“You've just seen with Cloud Code and even OpenAI's API business, like, these are durable businesses, and yes, can you switch? You can, the same way you could switch AWS or Azure but it's not worth it for a lot of businesses to do it, and there's sufficient dif…”
Sundheim: Leading AI model developers generate high gross margins
“If you look at the underlying margins of these companies, they are not the margins that you see in a commoditized industry. You know, the gross margins are quite high.”
Sundheim: The Capital-Intensive LLM Market Will Ultimately Support Only Five Competitors
“At this point, you probably have four or five LLMs that will be relevant in the long term. I don't see that changing. Not that there's not sufficient talent out there. It's just that the capital required to get into this business is too great in these companie…”
Sundheim: OpenAI and Anthropic models are not fundamentally different
“If you were to say, like, what is an important difference of Netflix versus these models is Netflix's content was differentiated. The models are more similar than they are different in that, like, you know, at any given time OpenAI may have a better model, Ant…”
Sundheim: Personalization, Not Answer Quality, Will Be the Moat for AI Assistants
“I think if you're Google or you are OpenAI, the differentiating factor will not necessarily be that Google gives you a better answer. Like if we were just like to, you know query Gemini or ChatGPT on something I don't think it's the case that we would say defi…”
Sundheim: Spotify creates pricing power on commodity music via personalization
“The music on Spotify is no different than Apple music or Amazon music, right? It's, theoretically, it's a pure commodity. What makes Spotify have pricing power? What makes it differentiated? Why would people, you know, be incredibly upset if you said, like, yo…”
Sundheim: Companies rarely succeed pursuing multiple end markets simultaneously
“I rarely have seen any company succeed trying to do, go after Multiple end markets at the same time. You know, it's usually you have an A team. An A team is focused on one thing. Your culture as a company is oriented towards either consumer or enterprise.”
Sundheim: Anthropic Pivoted to Enterprise After Failing to Gain Consumer Traction
“Anthropa took a different approach and just said, we are going to focus on enterprise. They tried consumer early on, but it became clear they didn't have traction, so then they just went all in the enterprise.”
Sundheim: D1 Invested in OpenAI Early and Pushed Them Toward Ads
“Early on, we invested in OpenAI, this is probably a year and a half ago, I said to them, like, you have to do ads.”
Sundheim: Tech firms cannot compete against ad-funded rivals without ads
“If you're going to do it ultimately one, you can't really compete against companies that are using ads if you're not. Very hard. If you're ultimately gonna do it, you might as well start earlier, because you have to build a culture around, you know it just tak…”
Sundheim: LLMs Will Generate Massive Cash and Insource Compute Within 10 Years
“At some point in the next five to 10 years, they'll be generating enormous amounts of free cash flow. When that happens, I think that, ah, they're likely to insource the compute.”
Sundheim: Neo-Clouds Will Survive Because They Outperform Hyperscalers at Running GPU Clusters
“I certainly would not make the case that they are fantastic businesses, but I don't think they're going away like people thought. Because I think they're better at Running GPU clusters than the traditional hyperscalers are.”
Sundheim: AI Capital Intensity Will Challenge AWS and Azure Profit Margins
“Over the next 10 years, I think the business, these hyperscalers, AWS, Azure, will grow fast. I think the margins, my guess, will be challenged, both because the businesses are getting a lot more capital intensive, because AI is capital intensive, more capital…”
Sundheim: AI Disruption Will Create Many Shorting Opportunities, Starting With Software
“In our letter, I said, like, there are going to be a lot of shorts, some longs because of AI. And software's the first one.”
Sundheim: AI Will Take Years to Displace Enterprise Systems of Record
“So I think I mean, systems of record are going to be difficult to displace. I think companies, while You know, it's neat to create software for small productivity enhancements. If you really want to run your entire business on something like an ERP system or a…”
Sundheim: Betting on AI scaling laws stopping is a very low-probability assumption
“You know, betting that scaling laws are gonna stop is, like, a really low probability assumption. I mean, there's just nothing to suggest that's the case. In fact, everything suggests the opposite”
Sundheim predicts AI will likely arbitrage away all human work eventually
“It is likely that at some point in the future, everything that we do is, ah, you know, arbitraged away by AI.”
Sundheim: Moated jobs are useless career advice without genuine interest
“Unless someone's really interested in something, they're not going to be good at it. So, you know, it might be the case that like being a plumber or being an electrician is like the most, you know, moated job in the world, but if you don't want to be a plumber…”
Sundheim: Discouraging kids from studying because of AI is destructive
“I saw a podcast recently with a Google researcher who left, and he said, like, oh, I don't even tell my daughter to study. It's just like, go out and have a good time. I think that's like a very destructive, like, way of going through life. Like, you should go…”
Sundheim: D1 Capital Never Came Close to Going Out of Business
“First of all, we never came close to going into business. That was just nonsense.”
Sundheim: GameStop Squeeze Marked Structural Shift in Equity Market Dynamics
“GameStop was, it, you know, it changed, it was the beginning of a change in the market structure and on the retail side, and so I knew we had to adapt to that.”
Sundheim: Fund Managers Cannot Disprove Negative Narratives in the Short Term
“So it's impossible to disprove the negative narrative in the short term. It takes a lot of time.”
Sundheim: D1 Shifted to a Lower-Risk Portfolio Strategy Following 2022 Drawdown
“Not in that the stock selection, all of that was gonna be the same, but the portfolio construction was gonna be done in a way that was much less risk-prone, and the analogy I gave was like, we're gonna hit singles and doubles.”
Sundheim: D1 Deserved Redemptions and Harbors No Ill Will Toward Departing LPs
“I don't harbor any ill will. I mean, look, I think that you know, the act of redeeming is, like, to some extent, we deserved it, right?”
Sundheim: AI scaling will drive powerful economic growth and disinflation
“It has to be the case that if you believe in scaling laws and you believe in AI, that economic growth will be very powerful. I mean, this is the ultimate productivity tool. And what productivity does is allows you to grow while having disinflation, which is li…”
Sundheim: UBI Will Fail Because Humans Are Inherently Wired to Work
“I don't think humans, as much as people like Dario, who I respect a lot, might say, like, well, we're just gonna give everybody a check and, like, everybody will just kind of live off universal basic income. I just don't think humans are wired to just collect …”
Sundheim: Automotive manufacturing is a bad, capital-intensive business to scale
“Autos are a bad business. It could be software autos, hardware autos. It is a bad business, and it's a really tough business to scale and very capital intensive.”
Sundheim: Great private tech investments take longer to prove out than failures
“The bad ones tend to, like, be More obvious. Faster. The great private tech investments, I think, are sometimes slower to prove how great they are because, like, you have these amazing founders who are just, you know, constantly making decisions which take the…”
Sundheim: SpaceX Starlink Will Drastically Undercut Other Broadband Options Within Years
“I think that the cost, they've come so far down the cost curve, I think that in a relatively short amount of time, like months, few years they are going to be dramatically cheaper than, Any other form of delivering broadband.”
Sundheim: Story-Driven Retail Trading Creates Endless Short-Selling Opportunities
“There are tons of people investing in things that are just based on stories, like because of social media and because of Robin hood. And so there's just endless amounts of shorts.”
Sundheim: Hedge Funds Are a Horrible Business Model With Zero Terminal Value
“I don't think hedge funds are a good business. Our business, like, it's horrible. It's like, it's amazing cash flows. The cash flows really well. It has no terminal value.”
Sundheim: Running a family office takes you out of the arena
“I could be a family office right now, and there's Plenty of, you know, positive things about being a family office. The drawback is like, you're not in the arena. I'm very collaborative with other investors. It's not like I'm sharp elbowed, but being out there…”
Sundheim: Most passionate investors start with a deep value bent
“I think most investors who love investing start out with a deep value bent because it's, you know, if you want to read about great investors, historically, most of them were deep value investors, Ben Graham, you know, Buffett.”
Sundheim Managed Over 55 Percent of Viking Global's Capital by 2016
“By 2016, I was managing, ah, just over half of Viking's capital, somewhere 55, something percent of Viking's capital.”
Sundheim: Having One PM Manage Over Half a Fund's Capital Hurts LPs
“It was pretty clear to me that from a business perspective, it was not in Andreas's best interest to have one person manage more than half the capital. I don't think that'd be even good for LPs.”
Sundheim: The bull case for human art grows as AI automates content
“Art is important because it is created by people, and I think the bold case in art would be, like, as everything else is, like, automated and, you know, in infinite supply because it's being created by AI, You know, art created by people reflects emotion and, …”
Sundheim: Europe Is Economically Stagnant but Easier for Fundamental Stock Pickers
“Europe is economically stagnated, so I'm not sure anyone should pay attention to it, other than if you are a pure fundamental stock picker, it's an easier market.”
Sundheim Predicts Japan Will Once Again Become a Global Military Power
“Japan's probably gonna become a military power at some point in the future again”
Sundheim: Germany, Korea, and Japan Excel in Engineering and Physical Assets
“Japan and Korea were not well positioned for the last 20 years because it was just, like, digital companies, but when it comes to, like, actually hard assets and good engineering, they, Germany, Korea, Japan have a lot of companies that have excellent physical…”
Sundheim: Disrupting Taiwan's Semiconductor Supply Chain Would Trigger a Global Depression
“If that supply chain were to get screwed up or disremediated, we would have, like, a, Incredibly bad economy. On the order of, like, depression type economy.”
Sundheim: Replicating the semiconductor supply chain requires 10-20 years
“I think some understanding of we need 10 to 20 years to replicate this supply chain.”
Sundheim: When dictators repeatedly state their goals, believe them
“Usually when dictators say they say something, and they say it, like, religiously, You should believe them.”
Sundheim: Talent and Leadership Matter More Than Business Moats in Tech Investing
“Buffett always says, like, the business is more important than the leader, because eventually, like, I kind of disagree with that. I think if you look over 30 years, sure, but over any period of time, like, businesses are just people, and if you have amazing p…”