Nov 13, 2025 · 1h 17m · 20vc
Michael Burry Shorts NVIDIA and Palantir & Has Defensibility Died in a World of AI? · 20VC with Harry Stebbings
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 20VC, Harry Stebbings, Jason Lemkin, and Rory O'Driscoll discuss the profound impacts of artificial intelligence on venture capital, market dynamics, and startup defensibility, analyzing everything from Sequoia's transition to Michael Burry's shorts and practical 'vibe coding' tools.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 14.9% of the talking time here. How this is scored →
Harry as informed peer 5.5 Guest teaching 5.1 Guest disagreement 3.0 Harry pushing back 3.9
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 45:27 Jason's visceral rejection of Harry's meeting volume
When Harry reveals his firm conducts 80 pitch meetings a week, Jason forcefully rejects the strategy, stating he would resign, give back his carry, and want to blow his brains out rather than take that many meetings.
Hardest push from Harry ▶ 34:36 Harry refuses Rory's premise on Series B clarityHarry directly refuses Rory's assertion that Series B investors can clearly identify category winners, arguing that market leaders remain completely uncertain due to aggressive moves from giants and new entrants.
Biggest teaching moment ▶ 8:20 Rory's masterclass on options math for shorting NvidiaRory educates the room by walking step-by-step through the precise options pricing, short-term expiration dates, strike prices, and 2x return requirements needed to bet against Nvidia at $188.
Harry holds his own ▶ 36:41 Harry names specific competitors to disassemble Rory's thesisHarry demonstrates deep market awareness by listing specific adjacent competitors like Vercel, Claude Code, Atlassian, and Salesforce to prove that venture bets face massive disruption regardless of early category leadership.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Sequoia Capital's Leadership Transition | 3 | 5 | 2 | 3 | Harry introduces Sequoia's leadership transition with standard background info and prompts the guests on why this signals Sequoia might be falling behind in AI. Rory and Jason provide nuanced insider analysis on VC fatigue and competitive pressure in the current AI environment. | |
| Venture Specialization: Walmart vs. Chanel | 6 | 4 | 3 | 3 | Harry introduces a sharp strategy framework comparing mega-platforms to boutiques ('Walmart vs. Chanel') and questions if Sequoia is caught in the middle. Rory elaborates on the 'manager of managers' operational dilemma, while Jason points out classic partnership economics friction. | |
| Analyzing Michael Burry's $1.1B Short on NVIDIA and Palantir | 5 | 7 | 2 | 4 | Harry questions how shorting Nvidia makes economic sense from an asset allocator perspective. Rory delivers a detailed breakdown of options math, put option mechanics, and short-term expiration constraints, educating the show on why betting against AI CapEx is mathematically brutal. | |
| Is the AI Wave a Bubble? Tracking Billions in ARR | 7 | 3 | 2 | 6 | Harry directly counters AI bubble skepticism by bringing concrete revenue statistics from OpenAI ($20B ARR), Anthropic ($70B projected), and Gamma ($100M ARR). Rory readily concedes Harry's point, agreeing that current demand and revenue traction are undeniably real. | |
| Case Studies: Saster's Use of Gamma and Replit Agents | 2 | 6 | 1 | 1 | Jason delivers an extended monologue sharing hands-on operational case studies of Gamma and Replit V3 agents within Sastr. Harry acts primarily as an attentive audience member while Jason details how agents transitioned from basic copilot tools to autonomous team members. | |
| The Decline of Defensibility and the Rise of AI Clones | 5 | 6 | 3 | 4 | Harry introduces vertical specialization and domain-specific data ingestion (citing Solve Intelligence in patent law) as a defense against horizontal cloning. Jason and Rory detail the extreme stress in seed investing when competitors can deploy feature-rich clones within 30 to 90 days. | |
| Seed Stage Defensibility and Pricing Dynamics | 2 | 5 | 4 | 1 | Harry steps back while Jason and Rory debate seed-stage defensibility and high entry valuations. Rory forcefully contends that seeking defensibility at the seed stage in AI is foolish because moats only accrue at scale. | |
| Identifying Winners: Seed Stage vs. Series B Risks | 8 | 5 | 6 | 9 | A heated clash occurs when Rory claims Series B investors can clearly identify market winners. Harry repeatedly rejects this premise, listing specific emerging competitors (Claude Code, Codex, Cognition, Vercel, Atlassian) to argue that categories remain chaotic and far from settled. | |
| Fund Construction and Diversification Math | 7 | 5 | 5 | 7 | Harry challenges Jason's rigid seed fund sizing math by introducing LP data and arguing that expanding market outcome sizes allow funds to take lower initial ownership. Jason maintains that writing large checks without strong ownership fails core portfolio construction math. | |
| The Value of Founder Meetings and VC Learning | 6 | 4 | 5 | 4 | Harry reveals that 20VC conducts 80 pitch meetings a week across partners, provoking a visceral, humorous rejection from Jason. Rory defends high meeting volumes as his primary method for acquiring market insights. | |
| Navigating Fundraising Processes: Best Practices for Founders | 7 | 7 | 4 | 6 | Harry shares personal frustration with founders running artificial fundraising processes when offered target pricing today. Rory and Jason break down founder negotiation dynamics, explaining why non-process processes built on informal relationship-building yield superior outcomes. | |
| The Highly Binary Startup Fundraising Environment | 7 | 4 | 2 | 3 | Jason declares the current market the most binary fundraising environment in decades. Harry strongly validates this with concrete portfolio data, citing a SaaS startup growing from $400k to $3M ARR that required 120 meetings to secure a single term sheet. | |
| Datadog's Success and Co-Attaching to AI Budgets | 5 | 4 | 1 | 2 | Harry brings up Datadog's 23% stock surge driven by $15M+ in AI-native customer spend. Rory and Jason summarize the strategic imperative of 'co-attaching to AI compute spend' for legacy infrastructure providers. | |
| Duolingo and the Value of Replacing Humans with AI | 6 | 6 | 3 | 4 | Harry asks why shareholder darling Duolingo crashed 25%. Jason and Rory educate on software value capture, framing that modern AI applications must either capture compute budgets or directly displace human labor headcount to avoid heavy market discounts. | |
| Excitement and Stress in the New Era of Software | 6 | 5 | 2 | 2 | The group discusses the blend of excitement and stress in modern software before Harry highlights Hummingbird VC's outlier performance with BillionToOne ($800M position on IPO). Rory and Jason analyze fund mechanics, MOIC versus ownership retention, and capital efficiency. |
Statements from this episode (43)
Lemkin: Early First-Month Growth Spikes Are No Longer Defensible
“You just can't take that early first month explosion as seriously as you used to. It's not as defensible.”
O'Driscoll: CEO Changes Always Signal Underlying Organizational Problems
“Whenever you have a CEO change having, it's cause something is wrong, right?”
Lemkin: Tech Executives From the Last 15 Years Are Wrong for Today
“The most folks from the last decade or 15 years are not the right people for the next decade.”
Lemkin: Legacy CROs Are No Longer Relevant for Today's Tech Roles
“I struggle to even recommend a lot of the CROs and executives I know for roles today. I just struggle to find them a job. They're not the right people anymore.”
Lemkin: The AI Era Requires Less Than Half of Today's VCs
“And so I think I could only imagine we don't even need half the VCs we have today for the AI world.”
Stebbings: VC Winners Will Be Mega-Platforms or Specialized Boutiques
“To me, I think it's like a specialization in leadership that's kind of correlated to actually the winners and losers in venture in the next 10 years, and I think the winners are Walmart, which is your mega platforms, it's your Thrive, it's your Lightspeed, it'…”
Lemkin: VC Partnerships Are Dysfunctional Due to Misaligned Carry Economics
“One of the reasons just partnerships are dysfunctional, and maybe it does, maybe we could move on, is that it's almost impossible for performance to tie to economics.”
O'Driscoll: AI CapEx Spending Will Overshoot and Cause a Downturn
“Zooming out, I think there's, you look at the AI capex band and you go, at some point this is going to overshoot. And then there's going to be a downturn.”
O'Driscoll: Palantir and AI CapEx Will Face a Major Correction Within Two Years
“I think, Michael, it's very hard to imagine a company trading north of a 110, 120 times revenues like Palantir, growing at 50 or 60% last quarter. Great. Amazing quarter. Kicking off cash. It's very hard not to imagine the next two years that doesn't have a si…”
Stebbings: Sam Altman Says OpenAI Will Hit $20B ARR This Year
“Altman says OpenAI is going to hit twenty billion ARR this year.”
Stebbings: Anthropic Projects Reaching $70B in ARR by 2028
“Anthropic projects seventy billion in ARR by 2028.”
Stebbings: Gamma Hit $100M ARR With 50 Employees
“Gamma raising a hundred million dollars at a 2.1 billion dollar valuation, having hit a hundred million in revenue with 50 people.”
Lemkin: SaaStr Requires $8M Annually in Sponsorships to Keep Lights On
“We have, you know, we have to close eight million of sponsors at Saster a year to keep the lights on.”
Lemkin: Tech Companies Will Never Return to 2020-2021 Staffing Ratios
“I think we're going to see a lot of these folks, you know, adding revenue teams, adding sales teams, adding marketers, never at the ratio of 20, 21. We'll never see those levels of staffing of humans”
Lemkin: I Launched 10 Apps in 125 Days Without an Engineer
“So I've been vibe coding for a 126 days. Feels like a lot of change. So I did this presentation, all hands at Replit, 125 days and 10 apps. I launched 10 apps in 125 days without an engineer.”
Lemkin: AI in 2026 Will Join Human Teams, Not Replace Them
“Next year is AI as part of your team, not replacing folks or layoffs. It is literally embedded in your team.”
Lemkin: Canva Built a Competitive Clone of AI Presentation Tool Gamma
“Canva actually has a Gamma clone now that's not bad.”
Lemkin: One Portfolio Startup Faced Five Clones Within 30 Days of Launch
“I can think of one investment I've made that has had five clones, like in the first 30 days, including one from a cloud leader.”
Lemkin: Bolt Outsourced Its AI Agent to Claude and Fell to Third
“So even Bolt, which was the early leader, they're now just at number three, They don't even have an AI agent anymore. They've outsourced it to Claude.”
O'Driscoll: AI Startups Reaching $200M ARR Will Achieve Lasting Defensibility
“In the sense of, I don't think you'll have these 203 hundred million dollar companies, and then someone else doing roughly the same thing come and displace them.”
O'Driscoll: Enterprise Defensibility Only Emerges at Scale Post-Consolidation
“The defensibility theorem emerges at scale, right? In other words, if you, I do believe what's true in most enterprise businesses, once you become the anointed winner, once a market coalesces and there's two or three people at that point in time, you know, It'…”
O'Driscoll: Seed-Stage Defensibility in AI Code Generation Is Absurd
“I think the idea that at the seed stage, you're going to find a defensible way to do co-gen or co-testing is absurd, right?”
O'Driscoll: Fearing Incumbent Competitors Signals Startup Market Leadership
“My experience is, when you go to the first board meeting where you suddenly realize you're scared of the big company adjacent competition, it probably means you've graduated from the baby class. You're one of the two or three winners in the startup land, and n…”
Stebbings: LP Position in 150-Deal Seed Fund Returned 7x
“I'm in one fund as an LP and it has a hundred, a 150 positions with a hundred to a 150 K checks. It's a seven X fund.”
Lemkin: Founders Should Send Detailed Pitch Decks Before Requesting Meetings
“Don't get in the door with me. Just send me a great deck and a great email. I will read it. I will slow down. I will spend time. And if it looks good, I'll take the meeting, but don't, no need to get in the door with me.”
O'Driscoll: Sharing Startup Data Serially With Select VCs Is a Mistake
“Conversely, if I was on the board of a company where the founder said, hey, I'm meeting with Joe at mega firm. They ask, can we share some data? They're not, you know, they just want to get a sense of it. I would shut that down and I would say, you share with …”
Lemkin: The Best Fundraising Processes Do Not Require Traditional Data Rooms
“The optimal way to run it is for everyone already to want to invest for real without games before you open your data room. And the super optimal way doesn't even require a data room because they already, you don't even need a data room because they already wan…”
Lemkin: Current Startup Fundraising Market Is the Most Binary in Our Lifetimes
“It's the most binary fundraising environment in our lifetimes, isn't it? You're either, you know, you're either YC, Neo, South Park Commons, got something, get funded, or who the hell's gonna find you in your pre-seed round, right? You either got, you know, yo…”
Stebbings: SaaS Startup Needed 120 Meetings for One Term Sheet Despite 10x Growth
“We had a company that went from 400 K to three million Naira, classic enterprise SaaS business, bread and butter enterprise SaaS that would have had five term sheets from five good firms, 120 meetings, one term sheet and it was ten million round on a 40 post.”
Stebbings: Datadog Stock Surged 23% Following AI Customer Revenue Growth
“Datadog absolutely freaking crushed. Stock up 23%. Fifteen million dollar plus AI native customers.”
Lemkin: B2B SaaS Companies Attached to AI Budgets Will Thrive in 2026
“If you have attached to AI budget and you're a data dog era and you're in that, you're actually going to have a great 20, 26 because we're there.”
Lemkin: OpenAI Now Purchases Software Like Traditional B2B Enterprises
“Like open AI is buying like a normal, buying like an Adobe or a Microsoft now.”
Lemkin: Legal Tech Company Clio Reached a $5B Valuation
“And the other one was Clio, which Raised it five billion in legal tech. It attached to AI in a different way. This is a company founded in 2008 that found its way in the AI era, right? It added FinTech payments, got up to three billion now at five billion.”
Lemkin: AI Startups Must Capture Compute Spend or Replace Human Headcount
“You're either getting money from compute, right? This massive spend, or you're getting money because you're using AI to replace humans. Otherwise, you're not gonna grow.”
O'Driscoll: LLM Software Matches One-on-One Human Tutoring Efficacy
“If the human today is someone in a class with 20 people, there's a whole bunch of examples that says LLM learning is equivalent to one-on-one human learning, tutoring one-on-one.”
Lemkin: AI CRM Startups Attempting to Disrupt Salesforce Are Unexciting
“There's like a 400 AI CRM startups out there all saying they're going to eat HubSpots and Salesforce lunch. That's not exciting to me as an investment.”
Lemkin: B2B Software Has Not Improved in a Decade Until AI
“This is the first time software has gotten better since the three of us met. It hasn't gotten any better since all of us met. It's the same crap.”
Lemkin: The NASDAQ Will Outperform Most Venture Capital Funds
“Just put the rest into NASDAQ, and you're gonna make more than most VC funds anyway”
O'Driscoll: AI Software Automates Actual Work Rather Than Just Updating UIs
“Salesforce was rebuild Siebel in the cloud, and it was fun, and we made a lot of money, but you're right, somewhat boring. What is really exciting now is you're not just talking about rebuild Salesforce exactly the same, but with a slightly modern UI. You are …”
Stebbings: Hummingbird Held an $800M Position in the BillionToOne IPO
“Hummingbird, the fund that does not get credit, is not talked about in the same way that many other great firms are, does their first bio deal in billion to one. They have an eight hundred million dollar position on the IPO. You want frickin' great venture ret…”
Lemkin: I Will Never Hold 18% Ownership at IPO Again
“And that's, I'm just obsessed with ownership now. It took me a long time to be obsessed with it, and now I've given up, but I don't see how I'll ever own 18% of something at IPO ever again, right?”
O'Driscoll: VC Funds Must Choose Between Max Multiple or High Ownership
“Fundamentally, the way they've been able to produce eight and 10 X funds has been in part by, I mean, keeping it at 40, 60, a hundred million dollar funds. I mean, obviously, you know, if you want, it's hard to simultaneously keep your ownership and keep your …”
O'Driscoll: Small High-MOIC Funds Are Best Allocation for Marginal LP Dollars
“Now, the interesting thing is both of our different outcomes have great outcomes for the GP, and this is where if you only have one dollar to play with, and this is the most important LP fact, if you only have one dollar to play with, then obviously you want t…”