Dec 11, 2025 · 1h 35m · news
SpaceX Valued at $800BN & Harvey Raises $160M at an $8BN Price & Netflix Acquires Warner Brothers · 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, host Harry Stebbings sits down with veteran investors Rory O'Driscoll and Jason Lemkin to debate astronomical private market valuations across SpaceX, Harvey AI, and Naveen Rao, the strategic shifts in AI application moats, and the deep philosophical divide between founder autonomy and active boardroom risk management.
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 12.9% of the talking time here. How this is scored →
Harry as informed peer 4.6 Guest teaching 5.4 Guest disagreement 4.2 Harry pushing back 4.2
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:29:30 Jason calls Rory 'beyond condescending' on board intervention
Jason aggressively attacks Rory's suggested deal term regarding relocating Chinese staff, calling it 'beyond condescending' and urging him to take a mulligan, which Rory outright rejects.
Hardest push from Harry ▶ 55:40 Harry rejects Jason's insult regarding enterprise implementationWhen Jason tells Harry 'You don't even know what you're talking about' regarding enterprise AI moats, Harry directly fires back with 'That's complete crap... Bullshit.'
Biggest teaching moment ▶ 21:58 Rory explains monopsony buyer power in content creationRory clearly educates the room on why Hollywood creators fear Netflix acquiring Warner Bros, shifting the framing from consumer monopoly to creator monopsony leverage.
Harry holds his own ▶ 47:01 Harry breaks down growth decay math against Harvey's $8B price tagHarry demonstrates strong expertise by stepping through the multi-year revenue decay math to prove that paying $8B requires pricing in three years of future public market performance.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome and 20VC Swag Discussion | 5 | 4 | 3 | 5 | Harry challenges Jason by asking whether SpaceX's $800B secondary sale actually indicates a lack of need to IPO rather than market strength. Rory breaks down the math behind 40x run-rate multiples for a business growing at 30 percent. | |
| The 2026 IPO Market Outlook and Anthropic Bets | 3 | 5 | 2 | 3 | Rory educates the room on power-law math and historical IPO precedents like Facebook in 2012 and Alibaba. Harry steers the conversation into setting direct pricing bets on Anthropic's future public market debut. | |
| Netflix's Warner Brothers Bid and Tech Consolidation | 3 | 7 | 1 | 2 | Rory provides a comprehensive breakdown of the Netflix and Warner Bros Discovery deal dynamics, distinguishing FTC monopsony concerns from FCC broadcast regulations. Harry asks broad structural questions to guide the detailed breakdown. | |
| Tiger Global's Downsized Fund and GP Commitment | 3 | 5 | 1 | 1 | Jason and Rory analyze Tiger Global's downsized $2.2B fund and its unusually high 20 percent GP commitment. The guests explain the financial leverage and LP alignment implications while Harry prompts on GP motivations. | |
| Naveen Rao's $5B AI Seed Round and Step-Up Valuations | 6 | 4 | 2 | 4 | Harry offers sharp industry insight by highlighting that headline mega seed rounds often hide cheaper undisclosed earlier tranches. Rory validates the observation and explains how blended investor economics function in bubble environments. | |
| Harvey AI and the Return of Premium AI Valuations | 5 | 6 | 5 | 5 | Jason forcefully argues that investors must fund outlier growth rates like Harvey's $50M to $150M ARR jump without overthinking TAM traps. Harry pushes back by pointing out European competitive threats like Lagora. | |
| SaaS vs. AI Growth Rules & Paying Ahead | 7 | 4 | 6 | 7 | Harry demonstrates strong financial acumen by modeling growth decay from $150M to $300M ARR to show that an $8B price tags three years of future public valuation. Rory agrees Harry's math is spot on, while Jason dismisses the skepticism. | |
| The Threat of Foundational Model Obsolescence ("Jasper-ing") | 6 | 8 | 9 | 8 | A heated clash erupts when Jason claims deep reasoning will render current legal apps obsolete and accuses Harry of hubris. Harry forcefully pushes back, arguing enterprise workflow integrations and law firm partnerships represent durable moats. | |
| Marc Benioff's Commodity Claim & Salesforce's Agentforce | 4 | 5 | 3 | 4 | Rory reframes Marc Benioff's claim about LLM commoditization by comparing model providers to hard drive manufacturers and consolidated airlines. Harry questions how switching ease affects B2B vs consumer software stickiness. | |
| Geopolitics, Open Source, and Chinese AI Models | 4 | 5 | 4 | 4 | Jason asserts that ChatGPT lacks long-term consumer lock-in compared to utilities like Google or Netflix. Rory disagrees using his consumer hardware replacement test, while Harry brings up founder adoption of Chinese open-source models. | |
| Airwallex, Ramp, and the "Asia Discount" Controversy | 7 | 6 | 7 | 6 | Harry addresses Keith Raboi's public data security attacks on Airwallex while disclosing his own investment. Jason suggests a racist motivation behind targeting Asian-founded companies, leading Rory to defend legitimate national security scrutiny. | |
| Boardroom Strategy & The Condescending VC Debate | 4 | 6 | 9 | 3 | Combativeness peaks between guests when Rory says he would demand Airwallex relocate all Chinese personnel as a deal condition. Jason calls Rory's stance beyond condescending, but Rory firmly refuses to retract his position. | |
| Prediction Markets, Insider Trading, and Regulation | 3 | 5 | 3 | 2 | Jason sarcastically praises prediction markets like Kalshi for enabling tech engineers to monetize insider information. Rory explains the financial mechanics and predicts impending congressional regulatory crackdowns. |
Statements from this episode (35)
O'Driscoll: Private market valuations consistently fail public market contact
“Every time that private market valuations came into contact with public market valuations, private market valuations were found wanting.”
Lemkin: VCs instructing CEOs on execution is condescending
“I think it's DC condescending to tell the CEO how they're gonna work it out. It is beyond condescending. The great CEOs will figure it out.”
O'Driscoll: SpaceX Valued Over 40x Run Rate at $15B Revenue
“Doing fifteen billion dollars, growing plus or -30%. This year, that's a pretty hefty valuation. I mean, that's, you know, north of 40 times run rate for a company this year growing 30.”
O'Driscoll: Almost Every 2025 Tech IPO Was a Down Round
“Almost every IPO this year, 20, 25 was a down round on the prior private round.”
Lemkin: Founders No Longer Fear Overpriced Valuation Rounds
“The angst is gone around mispriced rounds, overpriced rounds, and good or bad, I think that advice has no audience.”
O'Driscoll: Total private venture fair market value is roughly $2.9 trillion
“The total private FMV of venture is, you know, something like 2.9 trillion.”
Lemkin: Anthropic and Databricks will IPO in late 2026
“I think Anthropic and Databricks have an incentive to IPO, as best I understand it. They have an incentive. As Rory said, times are great. And so I think it'll be the, unless, unless, unless the market crashes for some reason, I think they're going to do it wh…”
O'Driscoll: Post-April market rebound is strongest bear market recovery since 1982
“From April, when, remember we had the tariffs and the implosion then, this has been the most, the strongest bounce back from a bear market since 1982. It's been the biggest jump, right?”
Lemkin: Anthropic will go public at a $500 billion valuation
“20 X times 26 is five 20. I'm gonna guess five hundred billion IPO, right?”
O'Driscoll: Hollywood opposes Netflix acquisition due to monopsony buyer power
“It's monopsony power, right? If Netflix becomes the biggest single buyer of content, Then if you're making content, which is what Hollywood does, making, right, you have to sell to, you know, the media buyer for Netflix becomes the most important person in you…”
Lemkin: Founders need one real and one fake offer in M&A
“And if you've been, for folks, if you've gone through M&A, you really want to have two, two real offers. Back up, you want to have one real and one fake offer, like one, at least one fake offer. You can tell CorpDev you have another offer.”
O'Driscoll: Meta and Google control 70% of all global advertising
“Facebook and Google have 70% of not just all digital advertising, but all advertising.”
Lemkin: Tiger Global made only 9 deals in 2024 versus over 100 in 2021
“They did raise a prior fund about 2.2 billion, right? In 23. They only did nine deals this year versus a hundred and some odd in 2021.”
O'Driscoll: With a 20% GP commit and 20% carry, half of 2x gains come from principal
“You have 20% of your money and then you have 20% carry. So if you two extra fund, half your gains come from your capital and half your gains come from your carry.”
Lemkin: The average growth fund GP commitment is around 1%
“I think the average growth fund is about one percent.”
Stebbings: Multi-Billion Mega Seed Rounds Include Undisclosed Cheaper Tranches
“The one lesson that I've learned from a lot of these deals, whether it's Thinking Machines or this or some others, is that although it's 500 or five billion on the sticker price when it's announced, there's actually several rounds before that Sequoia and Andre…”
O'Driscoll: Single-Day Multi-Tier Valuations Signal Top-of-Market Bubble Behavior
“When professional economists You know, make lists of things that you see in bubbles. This is widely up there on. Yeah. Kind of weird stuff you see at top of markets.”
Lemkin: Stair-Stepping Valuations in Hot Deals Efficiently Filters Oversubscribed VCs
“But it is, it does work in my limited experience because even if you do three rounds at the same day at 306 109 hundred or whatever it is, when you're oversubscribed, it's just a message to the new investors. The price is 900. It might be, the price might even…”
O'Driscoll: No Legal Software Company Has Ever Been Worth Over $2 Billion
“There's no legal company, software company been worth more than two billion ever.”
Lemkin: Being Too Negative on B2B TAM Always Bites VCs
“And at my entire life in B to B as a founder or investor, every time we try to be too negative on TAM, it bites us.”
Lemkin: VCs Must Bet on Software Companies Tripling ARR to $150M
“If this growth is durable, ok, if we're seeing signs that it's durable, if it has, if it really has a 170% NRR and 98% logo retention, And it's accelerating at one 50. This is just the bet you do. You don't pull your hair out or have to have lengthy dinners ab…”
O'Driscoll: Historical SaaS growth decay rate averaged 85% YoY
“In SAS. You could use this following rule of thumb with a high degree of accuracy. The growth rate for any year was roughly 85% of the prior year's growth, right?”
O'Driscoll: Founders must verify base models cannot easily replicate AI apps
“One of the non-negotiable disciplines before you do it is you should go and try and use the core models to do the same thing. And you should try and build a functional version of that as simply as you can. And if you can get even vaguely close to it, you need …”
Lemkin: AI customer agents remain slow, setup-heavy, and prone to hallucinations
“I could, if you've used these products in the field, you can see on the one hand, they're amazing. On the other hand, they are slow. Okay. On the other hand, they are complicated to set up and ingest data. On the other hand, hallucinations are a minor issue, b…”
Lemkin: Replit updated platform to dynamically rotate LLM providers automatically
“Now with Replit, you don't pick your model anymore. It's all dynamic. It's all gone. You don't pick a lot. You don't pick high reasoning. You don't pick short reasoning. You don't pick anything. It just, and it rotates Google and others”
O'Driscoll: Product-market fit in AI is a moving target requiring instant adaptation
“Product market fit is a rolling feast at the moment. And it's a moving target. Which is why if you look at all your good apps companies, one of the things you see is when the new model comes out, they're on it that afternoon. You know, it's pizzas and late at …”
Lemkin: Salesforce Agentforce takes 30 days to deploy and outperforms rivals
“We use Asian Force ourselves. It is great. I can talk to you how it really works. It takes about 30 days to train and deploy, and then you've got to keep going and upgrade and it's great. And it works just as well as the other agents. And in one way it works b…”
O'Driscoll: High fixed costs will consolidate foundation LLM providers to three or four
“As the cost to enter gets higher, the fixed cost barriers, there's not, I don't know if there's going to be 10 of these. There might be three or four, but there's not, I doubt there'll be 10.”
O'Driscoll: Foundation LLM providers must own consumer apps to avoid API commoditization
“If at the API level, I'm a commodity, and there's me, OpenAI, and Gemini, and the swapping between us, I have got to grab the apps. And that's what you're saying. I mean, Antropic is grabbing the coding app, And obviously OpenAI is grabbing the ChatGPT consume…”
O'Driscoll: The definitive consumer product test is immediate morning-after repurchasing
“The only test you need to run on this one is this. If you lost it today, would you go out and buy another one tomorrow morning?”
O'Driscoll: AI market rankings established by 2027 will determine 10-year trajectories
“This is one of those moments in time where how you end up ranking by the end of 26, 27 probably, I think, will determine the trajectory for 10 years, and I think the variance, the high fluctuations we're seeing now will steady out.”
O'Driscoll: Fintech valuation multiples are more bounded than AI multiples
“Fintech multiples tend to be significantly more bounded than open-ended AI multiples.”
O'Driscoll: Startups will decouple engineering teams from China over regulatory risk
“So I think you're just going to see a gradual separation. You'll just examine the risks and go, why would I, you know, yeah, I can get great engineers. I can save 20, 30%. But in return, I get embroiled in arguments on Twitter, but more importantly, I probably…”
O'Driscoll: US fintechs will use anti-China messaging against foreign rivals
“Competitors are going to be the people who use it first. Because if I was an all-American fintech company, I'd have my all-American flag out there and say, these guys are a bunch of commies. They have commies looking at your data. We're, you know, our stuff is…”