Oct 16, 2025 · 1h 25m · 20vc
Thinking Machines Co-Founder Joins Meta for $3.5BN, Industry Venture's $665M Acquisition · 20VC with Harry Stebbings
Rory O'Driscoll · 33m spoken Roger Ehrenberg · 21m spoken Jason Lemkin · 13m spoken Harry Stebbings · 7m 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
A panel of veteran venture capitalists and host Harry Stebbings analyze breaking tech news, detailing the multi-billion-dollar valuation dynamics of AI startups, the math of venture portfolio construction, and the shifting ethical landscapes of modern founder-investor relationships.
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 9.6% of the talking time here. How this is scored →
Harry as informed peer 4.5 Guest teaching 5.2 Guest disagreement 4.2 Harry pushing back 2.7
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 34:14 Rory bluntly rejects host's data center premise
Rory directly dismisses Harry's premise comparing office building construction to data center expansion, calling the framing clever but trivial before reframing the topic around AI scaling laws.
Hardest push from Harry ▶ 1:06:03 Harry refuses premise that early winners are predictableHarry explicitly challenges the guests by stating that anyone who believes they can identify early venture winners is wrong, supporting his stance with examples like Linear versus Clubhouse.
Biggest teaching moment ▶ 1:09:07 Rory educates host using empirical revenue metricsRory dismantles Harry's argument that early venture performance is pure noise by citing data showing that hitting two years of revenue targets increases a company's probability of achieving a 5x return from 30 percent to over 70 percent.
Harry holds his own ▶ 57:22 Harry frames Founders Fund versus Lightspeed strategiesHarry demonstrates strong industry knowledge by contrasting Founders Fund's concentrated AI strategy against the index-style diversification of Lightspeed, DST, and General Catalyst across specific deal tiers.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Co-Investing with Arthur Rock | 3 | 4 | 5 | 2 | Harry opens with lighthearted teasing about Arthur Rock, prompting Rory to jokingly call him a little punk and share stories from 2004. Rory and Jason banter about Harry's young age before Harry shares an LP anecdote regarding DPI expectations. | |
| The Economics of Goldman Sachs Acquiring Industry Ventures | 6 | 5 | 2 | 2 | Harry introduces the Goldman Sachs acquisition of Industry Ventures with specific transaction metrics like AUM and earnout figures. Rory and Roger educate the room on asset management M&A valuation multiples and distribution channels. | |
| Andrew Tulloch's Departure and the Ethics of Quitting | 3 | 4 | 7 | 2 | Harry brings up Andrew Tulloch leaving Thinking Machines for Meta, sparking a heated debate among guests. Rory sharply counters Jason and Roger's moral complaints about founder loyalty by highlighting the reality of a three point five billion dollar liquid stock offer. | |
| Transactional VC Rounds and the Single-Turn Game Dilemma | 5 | 5 | 4 | 4 | Harry asks whether high-value AI rounds represent a shift toward purely transactional VC relationships. Rory explains that multi-billion dollar offers convert multi-turn relationship games into single-turn payoff scenarios. | |
| SoftBank's $5 Billion Margin Loan and Masayoshi Son's Playbook | 4 | 4 | 2 | 2 | Harry raises SoftBank taking a five billion dollar margin loan against Arm shares to invest in OpenAI. The guests analyze Masa Son's appetite for high-leverage plays and historical survival through market crashes. | |
| The Macro Reality of AI Infrastructure and Scaling Laws | 4 | 7 | 7 | 3 | Harry asks if building more data centers than office buildings proves AI is structurally different from past bubbles. Rory explicitly reframes and dismisses Harry's office comparison as trivial, explaining the macro GDP capital requirements dictated by scaling laws. | |
| "Vibe Coding" and the Insatiable Demand for AI Tokens | 2 | 3 | 4 | 1 | Jason shares his hands-on experience vibe coding eight apps and consuming massive token volumes. Rory questions whether marginal economic returns can keep pace with continuous token consumption demand. | |
| Developer Efficiency and the Compressed Startup Lifecycle | 5 | 4 | 2 | 2 | Harry cites Lovable reaching one hundred seventy million dollars in ARR in its first year to illustrate compressed startup lifecycles. The guests discuss how AI-assisted engineering forces VCs to choose between acute uncertainty or huge pre-money valuations. | |
| Regulatory Arbitrage, Regulatory Moats, and Tech Diffusion Rates | 2 | 5 | 3 | 1 | Roger emphasizes regulatory complexity as a defensive moat for startups. Rory summarizes Aaron Levie's thesis on varying tech diffusion rates across enterprise sectors. | |
| Predictability of Kingmaking in Capital-Intensive Sectors | 6 | 7 | 7 | 4 | Harry asks if recent prediction market mega-rounds demonstrate VC kingmaking. Rory directly rejects Harry's premise, asserting that capital and brand recognition do not confer kingmaker status in consumer prediction markets. | |
| Politics, Regulation, and the Rise of Unregulated Markets | 5 | 5 | 2 | 2 | Harry highlights political connections on prediction market boards. Roger and Rory discuss regulatory arbitrage, explaining how state tax hikes push volume into unregulated offshore markets. | |
| Concentration vs. Diversification in Venture Capital | 6 | 6 | 3 | 3 | Harry contrasts Founders Fund's concentrated AI strategy with Lightspeed and DST's multi-bet indexing approach. The guests unpack portfolio variance, central limit theorem, and follow-on reserve allocation strategies. | |
| The Illusions of Early Winners and Options Theory in Venture | 7 | 8 | 7 | 7 | Harry forcefully asserts that predicting early venture winners is an illusion, citing Linear versus hype-driven misses like Clubhouse. Roger and Rory counter Harry's absolute statement by applying options pricing theory and empirical post-revenue survival metrics. | |
| Portfolio Reserve Strategy and the Ethics of Board Stewardship | 5 | 6 | 4 | 3 | Jason, Roger, and Rory discuss reserve strategy, board stewardship ethics, and cross-fund LP structures. Harry probes the opportunity cost of capital when follow-on valuations inflate sharply. |
Statements from this episode (38)
O'Driscoll: A VC's Prime Career Window Lasts Only Three Years
“There's a huge period of time wherever it says you're a little too young, and then there's this brief shining moment where you're good, and then there is the When is he going to retire moment, right? And it seems to me it's like 10 years each and three years i…”
Stebbings: LPs Question VC Motivation Even After Strong DPI Returns
“For the first three funds, you know, there's no DPI. And then you return a shitload of DPI, and then LPs go, mm, but are they really hungry anymore? Cause they've made a lot of money.”
O'Driscoll: Industry Ventures Sold at Roughly 10% of AUM
“This thing traded at roughly 10% of AUM, right?”
O'Driscoll: Carlyle and KKR Trade at Roughly 20% of AUM
“Carlisle and KKR, where they own all the economics, they roughly trade market cap, 20% of AUM.”
O'Driscoll: Institutional VC Brands Are More Monetizable Than Partner Funds
“The more your business is predicated on either a brand or some institutional thing, the more it's like a business, and the less it's like just three to five partners picking great investments, the more monetizable it is.”
O'Driscoll: Founder Loyalty Erodes When Offered Multi-Billion-Dollar Payouts
“In the face of unprecedented wealth, I'm shocked to discover that most people behave badly. The loyalty conversation erodes pretty quickly when you're, you know, when you're into the third comma.”
Ehrenberg: Big Tech Will Increasingly Buy AI Startup Assets and Talent
“Essentially it's an asset purchase. And I think that's something we'll continue to see more and more of”
Lemkin: Liquidation Preferences Do Not Matter for Seed Stage Investing
“I don't believe liquidation preferences matter. I don't believe they're a big deal, despite what they are next. But my liquidation preference has always been, and I put it in quotes, not true, knowing the founder would never quit. That's my protection as a see…”
Stebbings: VCs Writing $100M Checks Without Meeting Founders Should Be Fired
“As a VC, you should be fired if you write a hundred million dollar plus check and you don't meet the co-founders, period.”
Ehrenberg: Single-Turn VC Transactions Wildly Increase Outcome Volatility
“If you reduce everything because of the scale, right? To a single turn game, then that wildly increases the volatility of potential outcomes.”
O'Driscoll: SoftBank Holds $80B in Equity Through 90% Arm Stake
“He owns 90% of Arm still in SoftBank, and Arm is trading ninety-odd billion, so he's got eighty billion of equitable equity there.”
Ehrenberg: SoftBank Could Easily Borrow $25B Against Arm Shares
“So, but it very easily could see him back in the news with an incremental twenty billion. He could lever this. He could take twenty-five billion against the arm position easily.”
Ehrenberg: $5B Margin Loan Is a 'Low Octane' Move for Masayoshi Son
“So to me, this is like a relatively low octane masa move.”
Lemkin: AI Agents Replaced Almost All SaaStr Sales and Content Teams
“I Vibe Coded eight apps in a hundred days, and we have 12 AI agents working at Sastra now. Replaced all, almost all of our sales team and our whole content team.”
Lemkin: Demand Will Absorb Every Available AI Token Produced
“Everyone will consume every available token.”
Lemkin: Current AI Infrastructure Cannot Service Existing Market Demand
“I really do think we are not, I can see it myself. We're not remotely servicing the demand that exists today. Right? So how it gets paid for is a different question, but I think this is so different than the prior waves, where we just can't even service this d…”
O'Driscoll: AI Expansion Will Be Constrained by Economic Returns
“So if anything's going to constrain this, it's going to be economics. And my big picture model here is, you know, you've got the technical trends and then you've got, you know, the economic trends. And the real, the question is, Will it get the return? I mean,…”
Lemkin: AI Developer Efficiency Will Increase Token Consumption, Not Reduce It
“The better that gets, the more tokens you'll consume. I don't think there's this, Great efficiency coming. I think it's, I think we'll get fat. We'll just build more and more stuff faster and faster.”
Stebbings: AI Startup Lovable Reached Over $170M ARR in First Year
“Lovable had their first year anniversary today. Oh, the other day. And I thought that was insane. First year anniversary, over a hundred and seventy million an hour.”
O'Driscoll: Compressed Timelines Force Choice Between Uncertainty or $2B Valuations
“The time period from we haven't launched yet to, oh my God, it's so obvious has, as Jason said, compressed to the point where it's just, you know, That sweet spot is vanishingly small, and therefore you're left with the choice of do you invest into acute uncer…”
Ehrenberg: Regulatory and IP Complexity Protects Startups From AI Disruption
“But I do think that the issues of Legal and regulatory complexity. And again, whether, whether it's financial infrastructure or it's media rights, copyright, patent, IP, it makes it more nuanced than am I able to develop the next base model or a great platform…”
Ehrenberg: Polymarket and Kalshi Are Pure Regulatory Arbitrage Plays
“This is the purest regulatory arbitrage play of all time. Okay. If you can look at the cumulative market cap of regulated sports betting and look at how it has dropped in response to the rise of Polymarket and Calci, who are not subject to the same rules and r…”
O'Driscoll: 90% of Prediction Market Business Is Sports Betting
“This is two non-sports betting companies who are doing quote prediction markets where all we talk about is the 10% of the revenue that's political, and 90% of their business is sports betting, but we're not calling it that, and they're just killing it because …”
Lemkin: Early Y Combinator Staff Could Run Independent Side Funds
“It was like the old days when you could work at YC and have your own fund at the side. You didn't have to invest through YC.”
Ehrenberg: State Gaming Tax Hikes Drive Consumers to Offshore Markets
“And the more that, you know, Illinois jacks up rates, In state, A, those, the regulated sports, the sports books that are subject to these rates reduce investment in the state, handle those down, tax revenue goes down, and those customers that are now getting …”
Ehrenberg: Game Changers Will Concentrate 75% of Capital Into 3-5 Companies
“I'm following a similar playbook of three to five companies out of the 20 to 25 companies constituting 75% of the capital. Deployed.”
Ehrenberg: Game Changers Invested $1.5M at $10M Post-Money Valuation
“We just wrote a 1.5 check at a 10 post. So, 15% ownership in a really, really cool analytics company that, yeah, is disrupting a seriously stodgy and screwed up sector that I think is generalizability outside of that space.”
O'Driscoll: Target ARR for Tech IPOs Has Doubled to $400M
“We've moved from a world where an exit is two hundred million in ARR to a world where an exit is four hundred million in ARR at an IPO, right? So you're just doing your thing here, but way over there. At the finish line has receded another two or three years, …”
O'Driscoll: Scale VC Will Expand Target Fund Size to 25 Companies
“We typically been at our stage under 20 deals per fund, and we kind of said, you probably need to aim to close it to 25. Just given this dynamic, just nothing's changing at the stage we're at, but success is further away, right?”
Stebbings: Early Venture Winners Rarely Predict Long-Term Enterprise Value
“When I look back on, like, fun one where there's, like, a meaningful timeline to actually look back on the six years now, the best performers, your linears of the world, were not obvious early, and the early out performers did not signify enterprise value in t…”
Ehrenberg: IA Ventures Retained 17% of TTD, 13% of Wise at IPO
“And we were 17% of TTD at IPO and we were 13% of wise at IPO out of a little shitty seed fund.”
Ehrenberg: IA Ventures Held 2.2% Stake in Datadog at IPO
“Datadog's a great example of one that ended up being A fund returner, but we had 2.2% at IPO because we were in there at pre-seed and then RTP led the seed and then index led the A.”
O'Driscoll: Hitting 2-Year Revenue Targets Boosts >5x VC Return Odds to ~75%
“If you get the first two years that we underwrote in terms of revenue from the moment of our investment, if you get the first two years correctly, your probability of getting greater than a five X goes from 30% to mid seventies.”
Lemkin: Invests 8% of Fund per Seed Deal, Forcing High Hit Rate
“I'm doing eight percent of my fund into almost every deal, right? So I have to have such a high hit rate at seed.”
Ehrenberg: Strategy Uses 1-2% Entry Checks, Saving 5-7% for Winners
“For me, the one to two percent of the fund at entry, that's, that is, that's what we do. And then it's that next check that could be for five percent of the fund or seven percent of the fund.”