Jul 24, 2025 · 1h 26m · news
How Do All Providers Deal with Anthropic Dependency Risk & Figma IPO Breakdown: Where Does it Price? · 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 hosts venture capitalists Jason Lemkin and Rory O'Driscoll to dissect the shifting dynamics of the AI sector, public market valuations, and the structural challenges facing modern venture capital. The discussion offers a deep dive into platform dependency risks, IPO mechanics, and the hyper-competitive fundraising landscape for both startups and seed funds.
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 13.3% of the talking time here. How this is scored →
Harry as informed peer 5.5 Guest teaching 5.2 Guest disagreement 3.4 Harry pushing back 4.4
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 53:03 Jason rejects temporal diversification
Jason bluntly rejects Harry's argument for temporal fund deployment, calling the logic a cop-out and insisting that great founders emerge constantly regardless of market vintages.
Hardest push from Harry ▶ 1:07:19 Harry refuses big fund advantage thesisHarry forcefully interrupts and refutes Rory's claim that mega-funds benefit founders, highlighting board member turnover, signaling risk, and non-performing portfolio companies.
Biggest teaching moment ▶ 35:48 Rory explains IPO pricing mechanicsRory educates the room on how investment bankers deliberately set conservative initial IPO pricing ranges to orchestrate demand and walk prices up during roadshows.
Harry holds his own ▶ 46:58 Harry counters valuation caps with Rippling dataHarry demonstrates domain expertise by immediately challenging Jason's $30M valuation limit, citing the specific $35M Rippling seed valuation and Keith Raboi's decision.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Jason's Vibe Coding Experience and the Dangers of AI Agents | 2 | 5 | 1 | 1 | Jason monologues about his hands-on experience vibe coding over the weekend, explaining how AI agents can corrupt production databases and lie when context windows degrade. Harry remains in a listening role, asking basic clarifying questions to direct Jason's story. | |
| Lovable, Cursor, and the TAM of No-Code Tools | 4 | 4 | 2 | 3 | Rory prompts Jason to compare the software engineer market of Cursor against the broader non-technical user TAM of Lovable. Harry weighs in asserting Lovable's market expansion potential, while the guests discuss seed vs growth valuation frameworks. | |
| Platform Dependency Risk & the Power of Model Providers | 6 | 6 | 4 | 6 | Harry challenges Cursor's $28B entry valuation and asks why Anthropic wouldn't cut Cursor off immediately to protect its own products. Rory pushes back on Harry's premise, explaining regulatory risks, revenue dependency, and platform mechanics. | |
| OpenAI vs. Anthropic: The Battle for Consumer and Enterprise Dominance | 5 | 5 | 3 | 4 | Harry frames the market split between OpenAI winning consumer and Anthropic winning enterprise, forcing Rory to choose where he would invest between Anthropic at $100B or OpenAI at $300B. Rory and Jason break down enterprise software market cap dynamics. | |
| OpenAI's Internal Culture & Perplexity's Standalone Value | 6 | 4 | 2 | 4 | Harry references an essay on OpenAI's work culture and discusses Perplexity's $18B valuation and search capabilities. Harry takes a firm stance predicting verticalized AI models will rival generalist models within a year. | |
| Figma's IPO Pricing and the Mechanics of Going Public | 5 | 7 | 3 | 4 | Harry questions Figma's initial IPO pricing valuation relative to private rounds and asks about secondary share cash-outs. Rory educates on investment banking tactics like low initial pricing to build demand, direct listings, and secondary pricing risks. | |
| Rob Go's Essay on Seed Funds | 7 | 5 | 4 | 7 | Harry brings up Rob Go's essay on seed funds being squeezed by YC and multi-stage firms. When Jason claims he caps seed valuations at $30M, Harry pushes back directly using Rippling's $35M seed round as a counterexample. | |
| Fund Vintages, Future Outcomes, and Temporal Diversification | 7 | 6 | 6 | 6 | Harry cites LP wisdom from Vencap and argues sitting on hands in 2021 was superior to rapid deployment. Jason forcefully dismisses Harry's temporal diversification argument as a cop-out, arguing great founders emerge every single month. | |
| Meeting Fund-Returning Founders and the "Anti-Portfolio" Regret | 5 | 6 | 3 | 3 | Harry and Jason compare how often they meet fund-returning founders. Rory reframes the dynamic around identifying market opportunities rather than evaluating founder personalities alone, while Jason recalls historical VC collusion. | |
| Multi-Stage Fund Competition and the "Wall of Capital" | 8 | 6 | 7 | 8 | Harry describes losing a competitive growth deal to a wall of multi-stage capital. When Rory claims large funds are almost universally better for entrepreneurs, Harry forcefully pushes back, citing board member churn, portfolio bloat, and signaling risk. | |
| The Future of Seed Funds: Will They Expand or Contract? | 7 | 6 | 5 | 6 | The panel debates whether seed funds will expand or contract over the next five years. Harry presses both guests to name a single seed firm they would invest in today, forcing Rory to dodge and Jason to admit hesitation, while Harry names his own choices. | |
| Case Study: Astronomer and PR/Leadership Crises | 5 | 6 | 2 | 2 | Harry presents the Astronomer PR scandal as a case study, asking if all PR is good PR. Rory analyzes board-level decision-making, explaining that replacing a founder CEO introduces high failure risk but is necessary to remove operational drag. | |
| Kalshi Quick-Fire Round | 4 | 2 | 2 | 3 | Harry conducts a quick-fire round on Kalshi prediction markets regarding tariffs, OpenAI web browsers, and Grok macOS apps. Rory interrupts Harry to guess odds early, while Harry sides with Jason on OpenAI's browser release timeline. |
Statements from this episode (31)
Jason Lemkin: Fund managers who return no capital deserve no sympathy
“This idea that, oh, woe is me. I can't raise a third fund. I've never returned any capital. Tough luck, right?”
Rory O'Driscoll: Large venture capital funds are almost entirely beneficial to founders
“Almost everything about a big fund is good for the entrepreneur.”
Rory O'Driscoll: Consensus investment opportunities offer no pricing edge
“The market for consensus is fully priced in and fully discovered.”
Lemkin: AI agents cannot be trusted with production systems or data
“Agents cannot be trusted, and everyone in the industry knows this, and I didn't get it until this weekend. You cannot trust in it, and every single person will tell you can't trust it.”
Lemkin: VCs will make significant returns investing in AI guardrails
“Path one is guardrails, right? And there's a, and VCs are going to make a lot of money on guardrails.”
Lemkin: Windsurf was dead without access to Anthropic's Claude
“Windsurf without Claude was dead. That's why he had to find a deal that night, that weekend when OpenAI died.”
Lemkin: AI platforms like Lovable will build defensible software wrappers
“And so lovable and friends are going to build these thicker and thicker wrappers because they have to do security, right? They have to contain the AI. They're torturing Claude to do things it doesn't want to do. Claude wants to lie and seek out things and shar…”
Lemkin: New thin-wrapper AI IDEs will continually emerge
“In, in six months, if these are rappers, more, a new windsurf is going to emerge. It's just an IDE on top of the same models, right?”
O'Driscoll: AI model providers will eventually squeeze out software wrappers
“At some point when things get tougher, just like Microsoft did in the nineties, the platform provider starts to grind everyone's balls. And, you know, take more of the share away. There ain't 10 versions of PowerPoint in 20, 25.”
O'Driscoll: Anthropic will not cut off Cursor from model access
“I think it'd be very stupid of Entropic to just cut off a, probably the largest customer at the knees. And I, and the one thing, they're not stupid. So I don't think they will.”
O'Driscoll: OpenAI and Anthropic are the only AI startups that will survive
“The clear three, the two people who are clearly going to be at the table as startups when this is done are OpenAI and Entropic. It, it's hard to imagine beyond that.”
O'Driscoll prefers investing in Anthropic at $100B over OpenAI at $300B
“I think I do Anthropic at a hundred just from a back short-term momentum perspective.”
Stebbings: Perplexity raised $100M at an $18B valuation due to high demand
“Perplexities raised another hundred million at an eighteen billion dollar valuation. Honestly, there was so much demand from the 15 round that they upped it to 18, okay?”
O'Driscoll: Figma's IPO involves only 6% primary share dilution
“I think they're only raising around six percent primary share dilution.”
O'Driscoll: Figma was uniquely positioned to execute a direct listing
“It's actually one of the few companies that could have done it, it's profitable, so it's not sitting there. I mean, one of the reasons you don't do a direct listing is either you need the capital, they don't, or you have this existential dread of screwing up y…”
Lemkin: Canva should pursue a direct listing instead of a traditional IPO
“If it ever IPOs, it should do a direct listing, right?”
Lemkin caps seed entry valuations at $30 million max
“Even if you beat Andreessen, and Andreessen's offering 50, and you say, listen, the best I can do is 30, and I've done this multiple times, because 30 is the highest price I pay. It is, it's on my website, so there's multiple deals I've done where the price ha…”
O'Driscoll: Seed investing is 30-35% harder than ten years ago
“His point is the combination of those two players means at best the seed game is 30, 35% harder than it was 1008, 10 years ago. He's entirely correct.”
Stebbings: Poolside investment marked up to $5B in two years
“I'm in poolside where I've got a 250 K check. In two years, it was marked to five billion dollars.”
O'Driscoll: Scale has deployed every fund over three years since 2010
“We've taken three years on every fund since 2010.”
Lemkin: VCs now meet fund-returning founders once a quarter at best
“When I started, I met everybody in the industry cause it was small, right? So it was once a month. Now I would say I'm lucky if it's once a quarter to your point, right?”
O'Driscoll: Scale passes on 7 to 10 good deals per investment
“The stunning thing is we've looked at it like you see about seven to 10 more good deals for every one deal you do, right?”
O'Driscoll: Anti-portfolio regret is the price of good deal flow
“Anti portfolio regret is the psychological price you have to pay for being in the game because it's literally the emotional tax you pay for being in good deal flow. Because if you're not seeing 10 great deals, you're probably not going to do one great deal, wh…”
Lemkin: VCs routinely colluded and shared pitch details in earlier eras
“There was no transparency when I started all VCs colluded. In my first startup, I would go to a VC pitch and I'd go to the next one and they'd already talked a hundred percent of the time.”
O'Driscoll: Obvious startup metrics cap VC best-case returns at 2x
“Once the numbers make it obvious. That something's working. Pricing just go. Yeah. Pricing is going to go to the point where, you know, you're pricing in that two X best case.”
O'Driscoll: Half of all seed funds will shut down within five years
“The answer is less. Half of everybody's gonna die.”
Lemkin: First-time managers without markups will never raise eight-figure funds again
“Raising a seed fund in 20, 20, 21 made no, I mean, literally folks that had never closed, had never had a markup, let alone an exit, were able to raise an eight figure fund. That will never happen again.”
Lemkin: One seed investment spawned two $100M+ venture funds
“I have one investment that has not IPO'd, ok? That is only, I would say, reasonably hot, but very good, ok? It is our, that investment, where I was a seed, has already spawned two nine-figure funds out of it. Two.”
O'Driscoll: Astronomer is worse off as CEO searches fail 33% of time
“I think they're worse off in the sense that they have to do a CEO search, which is always tricky and has a one in three chance of failing.”