Jun 5, 2025 · 1h 27m · 20vc
Is DPI The Only Thing That Matters? with Sam Lessin, Jason Lemkin & Rory O’Driscoll · 20VC with Harry Stebbings
Rory O'Driscoll · 34m spoken Jason Lemkin · 24m spoken Sam Lessin · 13m spoken Harry Stebbings · 6m spoken
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, host Harry Stebbings and a panel of venture capitalists debate the realities of venture liquidity, the hollowing out of mid-sized funds, and the massive structural and economic shifts caused by the rapid rise of generative AI.
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 8% of the talking time here. How this is scored →
Harry as informed peer 3.3 Guest teaching 4.3 Guest disagreement 4.3 Harry pushing back 2.7
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 36:37 Jason confronts Sam over 'wasted life' remark
Jason aggressively rejects Sam's premise that non-generational companies do not matter, accusing Sam of implying that Box founder Aaron Levy and Jason wasted 16 years of their lives.
Hardest push from Harry ▶ 13:17 Harry challenges Rory on Insight's fund strategyHarry directly challenges Rory's defense of mid-sized funds by raising Insight's strategy of holding hundreds of positions and generating $400M on Hinge Health out of a $6B fund.
Biggest teaching moment ▶ 50:38 Jason educates panel on collapsing token costsJason forcefully lays out Mary Meeker's data on a 99.7% drop in token costs, schooling B2B executives and VPs of engineering who claim AI integration is too expensive.
Harry holds his own ▶ 28:13 Harry counters early selling thesis with RevolutHarry demonstrates domain knowledge by countering Sam and Jason's early-exit logic, citing Revolut's potential rise to $150B and showing how early sellers missed $125B in gains.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| The DPI vs. TVPI Debate: Chamath's Vanity Metric Critique | 2 | 5 | 6 | 1 | Harry sets up the prompt regarding Chamath's tweet on TVPI versus DPI. Rory forcefully dismisses Chamath's view as trite, leading to a sharp debate with Sam on whether markups offer any useful signal. | |
| The Hollowed-Out Middle of Venture Capital | 5 | 4 | 4 | 5 | Harry prompts the panel on SVB's hollowed-out middle fund report and explicitly challenges Rory using Insight's fund construction and Hinge Health returns. Rory defends mid-sized fund deployment mechanics. | |
| Exit Strategies: When to Sell and Secondary Exits | 3 | 5 | 6 | 3 | Sam advocates for selling early via secondaries while Rory vigorously defends mid-tier non-Google wins, emphasizing that cash checks of $80 million are meaningful returns. Harry re-engages Rory on power law assumptions. | |
| Pricing Discipline, Anthropic, and Late-Stage Access | 4 | 4 | 4 | 4 | Harry counters the concept of early exits by pointing out Revolut's potential trajectory to $100B+, where selling at $25B would leave $125B on the table. Rory and Sam detail late-stage access vs price discipline. | |
| The "Mattering" Debate: Are Most Companies and VCs Just Cogs? | 3 | 3 | 8 | 3 | Sam argues that most companies and VCs do not matter in the grand scheme, directly insulting Box and Chime. Jason reactively confronts Sam for implying founders and operators have wasted their lives. | |
| Mary Meeker's AI Report: Unprecedented Adoption & CapEx Hogs | 5 | 4 | 3 | 4 | Harry cites specific revenue figures for Anthropic and OpenAI to challenge concerns over hyperscaler CapEx spend. Rory contextualizes the spend against Google's historic financials. | |
| Mary Meeker's AI Report: China's AI Models & The End of Monopolies | 2 | 5 | 5 | 2 | Jason details China's low-cost AI models and the 99.7% collapse in token costs, forcefully stating that VPs of engineering claiming AI is too expensive should be fired. | |
| The Rise of Agentic Interfaces and MCP: How the Next Generation Will Abstract Away the CRM | 3 | 5 | 4 | 3 | Jason explains how Model Context Protocol (MCP) and agentic interfaces threaten traditional SaaS systems of record using spa software as an example. Harry connects this to Satya Nadella's database abstraction thesis. | |
| Recent M&A, IPO Trends, and the Rise of AI Infrastructure (Snowflake, Databricks, Clickhou | 3 | 5 | 2 | 1 | Harry introduces recent M&A and filing activity. Rory and Jason break down Circle's net interest margin model and Snowflake's acquisition of Crunchy Data. | |
| YC Valuations, the Dilution Dilemma, and the Fund-Returner Math | 4 | 6 | 4 | 2 | Harry asks for tactical advice on navigating YC seed rounds offering low ownership at high valuations. Jason and Rory explain the mathematical friction low ownership causes for traditional fund sizing. | |
| The Reopening of the IPO Window and Thoma Bravo's Record Fund | 2 | 3 | 2 | 1 | Harry prompts discussion on reopening IPO windows and record buyout fundraises. Rory explains how price realism clears illiquid private markets. | |
| Kalshi Quickfire: Will Meta Release a Closed-Source AI Model This Year? | 4 | 2 | 3 | 3 | In the Kalshi prediction round, Harry actively debates device distribution and initiates a direct wager with Rory on OpenAI hardware sales numbers. |
Statements from this episode (54)
O'Driscoll: Hyperscalers turned cash-efficient businesses into CapEx-heavy drains
“The hyperscalers have taken very good, cash-efficient businesses, and they would break Mr. Buffett's heart because they've turned them into CapEx hearts.”
O'Driscoll: Public market investors are just mean VCs on steroids
“Public company investors are just mean VCs on steroids.”
O'Driscoll: Every company's market share is up for grabs during platform shifts
“Every company's market share is up for grabs when there's a platform shift.”
Lemkin: Thoma Bravo generated $30 billion in distributions last year
“Tom O'Bravo had, had thirty billion dollars in distributions last year. Thirty billion dollars. I think that exceeds most guests, right? Thirty billion in distributions, not paper markups, but distributions.”
Lessin: Venture fund paper markups are irrelevant and often negative signals
“I know full well, I know this is an investor. That the marks that you made up or the marks of SoftBank listed something somewhere are completely irrelevant. In fact, the negative signal in a lot of cases. So like the marks are irrelevant and stupid.”
O'Driscoll: Early TVPI at 3-4 years correlates with final venture fund returns
“I'm willing to bet if you got some kind of machine learning algorithm and looked at all the funds that for three or four years in had a TVPI of two X and all the funds that four years in had a TVPI of .8 X and then correlated that to the ultimate outcomes, I b…”
Lessin: $1 billion venture capital funds operate in a death zone
“I don't know how you make multiples on a billion dollars consistently, right? In VC. I think once you're doing 10, you're playing a completely different game of asset gathering and asset deployment where you no longer have the same Goals. So I personally think…”
O'Driscoll: Megafund conglomerates make it harder for mid-sized VCs to win
“I do believe the impact of the conglomerates has made everything harder, including it's harder for us to make money.”
Lemkin: Modern Series A rounds are $20M to $30M
“Every founder wants to raise 20 or thirty million in their A now. No one wants to raise an eight, the eight million dollar series A is now three safe notes, right?”
Lessin: Most capital deployed into Series A rounds is completely wasted
“Most of the money at Series A is completely wasted, right?”
Lessin: Seed investors can exit via secondaries unlike Series A leads
“I really strongly believe private to private is an important future, and I love being early and small and first because I actually can sell into the private markets in a way that you can't if you've written a 30, forty million dollar check into the Series A.”
Lessin: Venture allocators must never sell their top-performing assets
“The number one rule as a fund out an allocator is you cannot sell the things that matter.”
Lessin: VCs take longer to master selling than buying
“As an early stage investor, you get hundreds of shots to buy. You get really good at buying. You have so fewer shots at learning to sell that it actually takes much longer to learn to sell well. Right. But it's still as important as learning to buy well.”
O'Driscoll: Non-generational Series A investments can still generate meaningful returns
“The overextrapolation of the only three deals make money, so therefore everything else at series A is burning money is just, I mean, it's true in that you prefer to be in Google, but it's not the only way to make money.”
Lemkin: Menlo Ventures will 10x to 15x its investment in Chime
“They said they did the Series B, right? So they did the Series B at 203 hundred. Maybe we can figure it out on the fly. They're going to, this will be a nine. There'll be some dilution. They'll 10 or 15 X their money, right? That'll be a fun returner or more t…”
Lemkin: Anthropic revenue jumped from $1B to $3B in five months
“Because Anthropic is one to three billion in revenue in five months, right?”
O'Driscoll: Pricing Discipline Is Most Critical Near Company Exit
“I actually think price discipline Is even more important the closer you get to an exit.”
Stebbings: Revolut will likely reach a $100B to $150B valuation
“I think everyone is realizing now that Revolut will quite likely be at least a hundred billion dollar company and likely a hundred and fifty billion.”
Lemkin: VCs should sell Revolut and Chime private shares immediately
“I would probably sell Revolut and Chime right now.”
Lessin: Customer acquisition cost only goes up as companies scale
“CAC only goes up, right? Like people just consistently get this wrong, right? And so I think that. Almost always up at scale, right? It's like, this is how the world works. And like, you know, the kind of argument that you're going to get good at something and…”
Lessin: Chime is not an important or paradigm-shifting company
“Is Chime an important company? I just don't see it, right? Like, is it, you know, it's good company? Like, they provided some banking products to unbanked people or underbanked people. I get it. It's not bad. They did a good job. They built a good app. But lik…”
O'Driscoll: Only one or two startups per decade reach $100B value
“Most companies fail, some companies compound to a billion in value, a few each year compound to five billion, one every year to 10, and one or two every decade to a hundred billion.”
Lemkin: ChatGPT reached 800M users in 17 months
“ChatGPT is the fastest, even though Sam Lesson said he's not sure it's important, it's the fastest Gain of users in, in the history of the world. Zero to eight hundred million in 17 months. Okay. That's, it took Netflix 15 times longer. It took TikTok five tim…”
Lemkin: Big six tech companies spent an unprecedented $212B on CapEx
“The second Mary Meeker one was showing that the infrastructure spend is unprecedented in the history of internet, right? That the big six spend two hundred twelve billion on CapEx.”
O'Driscoll: AI CapEx spending will trigger a market scare within 2-3 years
“When you're spending half a trillion dollars in advance of revenue, at some point in the next two or three years, there'll be at least one scary moment. Right. And that could be actually an interesting opportunity to invest, but it's hard to believe it's up an…”
O'Driscoll: OpenAI misses next year's revenue target by 40% if growth matches Google
“If all OpenAI is, Is just as good as Google, then it's going to miss its number next year by about 40%.”
O'Driscoll: ChatGPT will be the Apple of AI while alternatives act as Android
“I think ChatGPT will be the Apple-level quality product, but there'll be a whole bunch of Android-type quality products that are out there that will just keep pricing honest if they're commercially available and, you know, universally.”
O'Driscoll: US companies no longer hold a monopoly on LLM development
“But I think the undeniable fact is we ain't a monopoly anymore. It's not two or three companies monopolistically producing LLMs. There are four or five companies the other side of the pond cranking them out.”
O'Driscoll: Cost of AI Intelligence Will Drop Log Orders of Magnitude Annually
“Information knowledge, intelligence, for lack of a better word, is just going to get cheaper and cheaper, faster and faster. There's going to be, there's just intelligence is available at a price that's declining, you know, by log or log orders of magnitude ev…”
Lemkin: The 'AI slow roll' is the biggest threat to B2B SaaS
“I call it the AI slow roll, and I think it's the number one thing killing B to B companies is the AI slow roll.”
O'Driscoll: Tech companies without top-down AI pivots will fail within two years
“And I agree if you don't do that, and if you don't do it well, you're just going to be slow toast over the next two years.”
Lemkin: The next generation will not use files or traditional software interfaces
“I think what I do know is that this generation will not use the Internet remotely like we do. They will not use lead contacts and opportunities. They will not use files. Files are disappearing overnight. Kids do not know what a file is. It does not matter. And…”
O'Driscoll: AI systems of work will displace legacy systems of record
“I think in front of every system of record, there's going to be some system of work that does it, and it'll sit on top of the system of record for now, and if it does a good job, it will gradually displace, make irrelevant, or replace the system of record if t…”
O'Driscoll: Value accretion will shift from Salesforce to AI execution tools
“I think, as I've always said, I think Salesforce instances will be there at scale long after I'm dead, but the value accretion will all go to the technology that's helping me do my job, not the technology that's count keeping score on how well my job's been do…”
Lemkin: Model Context Protocol is an existential threat to SaaS applications
“I think that it's early, but it's, I think MCP is an existential threat to almost every SaaS app.”
Lemkin: Databricks acquired Postgres database startup Neon for $1B
“Databricks bought its own Postgres database company Neon for a billion.”
O'Driscoll: VCs excel at 10-year thematic bets but fail at quarterly calls
“Venture guys make really good thematic bets about the next 10 years and are pretty horrible at assessing how the market's going to react to the next quarter's data. Hedge fund guys are the exact opposite.”
O'Driscoll: Public board members predict earnings stock moves with at best 50% accuracy
“You can go around the boardroom and ask people, here's what we said, here's what we're going to announce tomorrow, will the stock go up or down? At best, 50% accuracy.”
Lemkin: AI startups represent 70% of the current Y Combinator batch
“They're trying to get everyone to 50 to 60 that are AI, that just have AI, which is 70% of the batch.”
Lemkin: Seed Deals Require 3x Step-Up Next Round to Justify Risk
“For seed investing to really make sense, the next round should be three X to justify the risk. And for A or B, it's gotta be two X, or you should just wait.”
Lemkin: Ownership in YC Companies Is Less Than Half Normal Target
“All I can tell you, my little analysis of YC is I've ended up owning about a little less than half of what I would have.”
O'Driscoll: Evaluating $60M YC AI seed rounds will take eight years
“It'll take 10 years to discover if the eight billion dollar funds work and deliver an acceptable return. It'll take eight years to figure out, can you pay 60 pre for an AI Y Combinator startup and make money on average?”
Lemkin: Raising 3x Capital at Same Dilution De-Risks Founders
“Three times as much money for the same dilution de-risks your life as a founder, right?”
Rory O'Driscoll: Tech IPO window has reopened and normal service resumed
“No, honestly, I think the most salient fact is that they're happening, that the window is open, that normal IPOs, I mean, I think normal IPOs are taking place, yet normal service has been resumed.”
Jason Lemkin: Thoma Bravo raised $35B fund for B2B acquisitions
“I know we glossed over it in the beginning, but I do think Tom O'Bravo raising a record fund this week is a good sign too. That's thirty-five billion dollars to buy B to B companies, almost entirely B to B, right?”
Rory O'Driscoll: Price clears all markets for tech IPOs
“Something a friend of mine used to say years ago, and it's always struck me, it's just so true, price clears all markets. In other words, if you're, you know, you can get deal, it's, what we're seeing it now is, you can't get your deal done at 25, which is the…”
Lemkin: Initial OpenAI/Jony Ive device will be screenless, but line will include screens
“I think that the first thing that comes out will be no, but ultimately it will be all of it. It'll be, yes, it'll have audio, it'll have screen. It'll be everything.”
O'Driscoll: OpenAI and Jony Ive device will sell under 5M units initially
“Will the device sell more than five million units in the first full year of shipping? And I would take the under.”
O'Driscoll: Meta will not release a closed-source AI model this year
“Other than that, I see no reason to change the bet they're currently on, which has been the open source bet. So no, in the absence of DC intervention is my call.”
Lemkin: Meta will release a closed-source AI model to stay competitive
“And if releasing a fully closed source model brings, makes them more competitive, I think they will. I think it's more than 35%.”
Lemkin: Elon Musk has a 50% chance of stepping down as Tesla CEO
“Can he in the next 18 months find someone better than him for the job? I say actually more than 32%. Yeah, exactly. I'll go 50%.”
Lemkin: Marc Benioff has hired three co-CEOs attempting to exit Salesforce
“Mark Benioff's tried four or five times to get out too. He's hired three other co-CEOs, right? It just hasn't worked.”
O'Driscoll: Gwynne Shotwell is one of the most amazing managers ever
“Gwen Shotwell appears to be one of the most amazing managers ever down there at SpaceX. Seems to be. I mean, you know, sending rockets up and most of the time having them work and building an amazing business. That's just been stunning management.”