Oct 23, 2025 · 1h 34m · 20vc
Why VC Today is Worse than 2021 · 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 analytical episode of 20VC, host Harry Stebbings and venture capitalists Rory O'Driscoll O'Driscoll and Jason Lemkin dissect the modern tech investment landscape. Together, they debate the extreme capital intensity of the AI infrastructure boom, structural challenges in venture capital fund cycles, and the mathematical realities of vertical SaaS valuations.
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% of the talking time here. How this is scored →
Harry as informed peer 5.1 Guest teaching 5.9 Guest disagreement 4.0 Harry pushing back 4.1
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 13:08 Rory Rejecting Harry's Founder TAM Expansion Thesis
Rory directly tells Harry 'I think you're wrong' and forcefully rejects the 'great man theory', arguing that market size dictates startup scale rather than founder skill alone.
Hardest push from Harry ▶ 21:50 Harry Refusing Jason's COVID Market AnalogyHarry explicitly refuses Jason's comparison between AI software adoption and temporary COVID demand spikes, maintaining that AI delivers permanent productivity enhancements.
Biggest teaching moment ▶ 51:46 Rory's Lesson on Rationale vs. Rationality in Compute CapexRory reframes Harry's confusion regarding Poolside's data center, schooling him on the distinction between internal founder rationale and long-term economic rationality while detailing GPU supply bottlenecks.
Harry holds his own ▶ 4:26 Harry Defending Top VC Carry Packages Against Meta AI CompHarry pushes back on Rory's assertion that Meta AI packages beat VC compensation by citing specific multi-decade carry distributions at mega-funds like Thrive, GC, and Lightspeed.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Everett Randall's Move to Benchmark | 4 | 2 | 2 | 2 | Harry introduces Everett Randall's high-profile move to Benchmark as GP and notes Benchmark's generous backdated carry pool structure. Rory and Jason comment on VC mobility, firm culture, and how talent compensation in VC compares to tech. The tone is casual and agreeable across all participants. | |
| VC Compensation vs. AI Talent Packages and the Reality of Carry | 6 | 7 | 6 | 6 | Harry directly challenges Rory's claim that Meta AI talent packages beat VC carry, citing mega-platforms like Thrive, GC, and Lightspeed. Rory forcefully disagrees, explaining the difference between four-year liquid public stock and multi-decade illiquid VC waterfalls. Harry presses his counterargument before Rory details historical venture check timing. | |
| Revolut's Valuation and the TAM Debate | 6 | 8 | 7 | 6 | Harry argues that exceptional founders like Daniel Ek and Alex Bouaziz sequentially expand their addressable TAM over time. Rory bluntly rejects this 'great man theory', telling Harry he is wrong and explaining that market size dictates potential while founders merely capture it. Rory then schools the room on why Spotify actually succeeded through early European music licensing. | |
| Vertical AI, Exuberance, and the 'COVID Mistake' Analogy | 6 | 6 | 5 | 7 | Jason draws an analogy between current vertical AI exuberance and the 2020 COVID hype cycle where buyers briefly entered the market simultaneously. Harry firmly rejects the COVID analogy, arguing AI provides enduring productivity gains rather than a temporary spike. Jason and Rory counter by emphasizing enterprise decision window compression and hidden onboarding costs. | |
| M&A Advice, Vertical SaaS Math, and Venture Exit Realities | 5 | 5 | 4 | 4 | Jason outlines M&A rules of thumb and questions vertical SaaS venture math relative to Toast. Harry counters with portfolio evidence from Solve Intelligence in IP law where contract sizes exceed $100k. Rory supports the thesis by noting LLMs are uniquely tailored to word-centric legal workflows. | |
| OpenAI, Microsoft, and Oracle's Shift in Power Dynamics | 4 | 5 | 2 | 2 | Harry introduces OpenAI shifting compute spend to Oracle over Microsoft. Rory provides analysis on how Satya Nadella de-risked Microsoft's balance sheet while Oracle absorbed low-margin compute risk. The group agrees on Microsoft's savvy corporate positioning. | |
| Oracle's Balance Sheet and OpenAI's Risk Offloading Strategy | 6 | 6 | 2 | 4 | Harry highlights Oracle's high 4.6x debt-to-equity ratio as evidence of overextension. Rory agrees with Harry's concern and praises OpenAI for offloading massive balance sheet capex risks onto third-party infrastructure vendors while keeping the upside. | |
| Poolside's Two-Gigawatt AI Data Center and the Reality of Capital Intensity | 5 | 8 | 5 | 4 | Harry shares his early seed investment history in Poolside but admits he struggles to understand why they are building a 2GW data center. Rory corrects Harry's use of 'rationality' versus 'rationale', explaining that compute scarcity forces startups into capital-intensive infrastructure commitments. | |
| Defining the AI Bubble and Potential Bust Dynamics | 5 | 7 | 5 | 4 | Harry cites technical definitions of a market bust to argue AI is not in a bubble. Rory corrects Harry by explaining that his definition describes the aftermath rather than the bubble formation, drawing parallels to the 1990s telecom bandwidth crash. | |
| The Rise of 24/7 Inference and Investment in Nvidia | 3 | 5 | 1 | 2 | Jason describes a new B2B AI investment that utilizes continuous 24/7 inference workloads. Harry asks if this implies investors should simply concentrate capital into Nvidia, which Jason confirms. | |
| Temporal Diversification and the Condensed 18-Month Fund Cycle | 6 | 5 | 2 | 4 | Rory cautions against abandoning temporal diversification in fund deployment. Harry provides real LP foundation feedback confirming that 18-month deployment cycles have returned across the venture industry. | |
| Historical Venture Cycles and the 15-Year Bull Market | 3 | 7 | 3 | 2 | Harry asks Rory to reflect on historical venture cycles over his 30-year career. Rory delivers a detailed historical overview of underfunded multi-year periods following the 1987 and 2000 crashes, contrasting them with the current 15-year bull market. | |
| Is This the Toughest or Easiest Time to Invest in Venture Capital? | 4 | 5 | 6 | 2 | Harry prompts the guests on whether current market conditions are the tough or easy. Rory calls it as tough as ever due to high risk curves, while Jason forcefully disagrees, declaring it the easiest time ever to write checks and ride technological change. | |
| Content Moderation, Erotica, and the Privacy of LLM Chat Histories | 5 | 6 | 4 | 4 | Harry brings up OpenAI opening guardrails on erotica and asks whether the guests would share their LLM chat history publicly. Jason expresses strong privacy and moral concerns, while Rory highlights the legal content moderation traps facing AI model providers. | |
| Game: Replit's Path to a Billion ARR and Vibe Coding Disruption | 7 | 6 | 5 | 7 | In a quickfire game, Harry disagrees with Jason's prediction that Replit will reach $1B ARR next year, citing prosumer TAM boundaries and churn. Harry then presses Jason on why he hasn't put money into Replit given his extreme user advocacy, prompting Jason to commit $10M on air. | |
| Comparing Payroll Giants Deel and Rippling, and the Sweet Spot of Early-Stage Venture | 6 | 7 | 5 | 6 | Harry asks the guests to choose between owning shares in Deel or Rippling. When Rory tries to punt due to an adjacent investment in Papaya, Harry presses him to answer. Rory breaks down ADP/Workday incumbent dynamics before choosing Deel, while Jason defends Rippling. |
Statements from this episode (38)
O’Driscoll: Doubling down on mega-caps is easiest 2025 strategy
“It looks like the easiest way to make money in 2025 is to take the very biggest companies and double down one more time.”
O'Driscoll: Aggressive risk-taking distorts investor talent during bull markets
“Now, again, reminder, in a bull market, the most aggressive person will look the smartest. Just before the crash Because the more risk you've taken, the more money you've made.”
Everett Randle joins Benchmark as GP from Founders Fund
“Topic number one, Everett Randall joins Benchmark. Benchmark don't add partners very often. It's a big news announcement that he's joining. Joins Benchmark as their latest GP. He was with Kleiner and he was with Founders Fund before.”
Benchmark offers backdated carry pools to newly joined General Partners
“I also know Benchmark are extremely generous in terms of like backdated carry and being brought into this carry pool for this fund with fireworks and with McCore and with Legora and with Manus and with many others.”
O'Driscoll: VC compensation cannot match elite four-year AI engineer pay packages
“There's no one gonna, if reports of the billion dollar plus package are true, very few people are gonna make four, one billion dollars, four year vested in venture.”
Stebbings: Top mega-fund partners will make over $1B in distributions
“Well, I think if you are one of the top one to three carry participants in one of the large mega platforms, be it Andreessen, be it Thrive, be it GC, I would argue that you will have more than that in distributions in the next few decades.”
O'Driscoll: VC is get-rich-slow with decade-long dry spells in carry
“Venture, as someone said to me years ago, it's the get-rich-slow program, and there can be ten-year periods of non-payment.”
O'Driscoll: Revolut generated $3B revenue and $1B profit growing at 60%
“It's doing, I want us to last year, three billion in revenues, making a billion, growing at 60%.”
O'Driscoll: Spotify won by starting in Europe to avoid US label litigation
“All the other music startups based in the U.S. Got strangled at birth by lawyers, right, because it was all about IP property rights, and little old Spotify got going in a bunch of European countries that, let's be frank, your average Big Five record label did…”
Calacanis: Enterprise Buying Window for AI Tools Will Soon Disappear
“They're not going to be in market every year for an AI tool. This will be a window that will disappear.”
Calacanis: Enterprise AI Buyer Demand Will Drop to 5% Within 24 Months
“They got the price of the vendor right, but the soft costs, the training, the onboarding, the business process change were the highest it's ever been in their lifetimes. They're not going to do that every year. And so we may go back to five percent being in ma…”
O'Driscoll: Extrapolating Short-Term Growth Breaks Standard SaaS Investment Models
“If you take these growth rates and extrapolate them for the next four or five years, you could be catastrophic with your mental model of SAS. You know, slight deacceleration every year. You could be catastrophically wrong on growth.”
Calacanis: Founders should accept M&A offers if TAM expansion stalls
“If you get an M&A offer as a founder or an investor, and the founders make the decisions, the VCs don't make the decisions. If your TAM isn't really accelerating, take it.”
Calacanis: VCs will lose money on 80% of B2B AI investments
“VCs are gonna lose, like, 80% of their investments in AI B to B because they're ignoring these issues, and so be it.”
O'Driscoll: Microsoft leadership nailed OpenAI deal, but failed on shipping technology
“I think when all is said, and as I said before, on Microsoft's side, when all is said and done, the shareholders should award medals to Satya, The CFO and their GC and they should hire someone else to do their technology because they haven't shipped, right?”
Calacanis: OpenAI capital requirements are 100x Microsoft's initial high-end estimates
“OpenAI needed much more capital than they thought when they started. Microsoft bailed them out in buying 49% of the company. Now they need much more capital than Microsoft thought. Probably two orders of magnitude more than Microsoft's, ah, you know, high-end …”
O'Driscoll: OpenAI has successfully offloaded balance sheet risk to partners
“They've offloaded a lot of the balance sheet risk to everyone else, and all these other people seem to be happily taking it on right now”
O'Driscoll: OpenAI's corporate development team is second to none
“The OpenAI corporate development deal-making machine is second to none. They have a rootless instance for weakness and take advantage of other people.”
Stebbings: Early investment in Poolside founder's previous startup returned 50x
“I was actually one of the first investors. I don't know if you guys knew this. I invested in ISO, the founder of Poolside's business. Which pivoted into poolside. And so I got rolled in very luckily into like the first round, which is great. I'm very grateful …”
Stebbings: Productive capital will not leave AI for three years
“Like we will not have productive capital leave AI and data centers for more than three years. And so everyone that's like, oh, we're in an AI bubble, we're in an AI bubble and it's going to bust. I don't think so, because if we're in a bubble and you're antici…”
Calacanis: New B2B AI investment requires unprecedented inference volume
“I'm making an investment this week, a B to B AI investment that it's early ish. It's early, but it uses more inference than anything I've invested in yet. It uses far more.”
Calacanis: AI apps will consume 1,000x more compute via continuous inference
“Smart folks are going to figure out already how to use a thousand times more inference and compute than we're using today. Because instead of running a little one-off thing, right, or even using lovable for an hour, and then letting the servers with no load go…”
Calacanis: VCs raise funds every 18-24 months without temporal diversification
“Everyone's just raising the fund every 20, 18 to 24 months. There's no divert temporal diversification.”
Stebbings: Salesforce deployed $850M of its $1B AI fund
“Salesforce invested eight hundred and fifty million of their new billion AI fund.”
O'Driscoll: Venture returns over last 5 years massively trailed public markets
“The truth is the venture return over the last five years has been massively lower than the public market returns.”
O'Driscoll: VC pooled returns exceed small caps by 600 basis points
“The Cambridge pool, third year return says you get exactly what economic rationality would assume you get, which is around 600 basis points on a pooled return, not a median return above the small cap.”
O'Driscoll: Assessing VC market cycles requires a 30-year timeframe
“If you're in a business with 10 or 15 year cycles, you just have to internalize that you have to have a 30 year span to talk about cycles, right?”
O'Driscoll: ChatGPT ended the 2022 VC market correction within a year
“In the context of anyone playing this business from 2010 on, there has never been a period of longer than a year where there's been a substantive correction or a curtailment of capital. The only year would have been 22, twenty-three-ish, and God bless ChatGPT …”
O'Driscoll: Current VC market is as tough as it has ever been
“It was a great time in 2010 1015. It was tough in 21. Frankly, it feels tough today, as tough as it's ever been, because the good news is stuff is working, but there's a lot of variance, a huge amount of capital, and it feels a tough, you're way out there on t…”
Calacanis: Present VC market is the easiest ever for investing
“Yeah, because there's because there's so many entrepreneurs. There's so much change. Change is when you make money, right? Change is when you make money in venture. There's so much change. There's so many great entrepreneurs, and we don't have to worry about g…”
Stebbings: Erotica is the largest use case for Grok media generation
“And it is the largest use case for Grok on their image and video generation is erotic creation.”
Calacanis: AI companies built models by stealing all my copyrighted content
“Just like in the beginning to get these off the ground, we had to trample copyrights and destroy everyone's IP rights. All of my IP is stolen. Everything I've written, all my videos were taken without my consent.”
O'Driscoll: LLMs face higher regulatory liability than social platforms
“We didn't write the content. We're just trans, you know, we're just a connection mechanism. It's other people's content. And what's super clear on chat GPT is you are writing the content. So for things like, you know, advice that goes wrong, medical stuff, may…”
O'Driscoll: ChatGPT content moderation will be a hot seat for 5 years
“I think the content moderation, content decision job at ChatGPT is going to be a hot seat for the next five years, and I don't think erotica is going to be the hardest problem they face, right?”
Calacanis: AI vibe coding renders classic early-stage VC software evaluation obsolete
“If in six months, all this stuff can be vibed by anybody, right? Like then how the hell can we tell those early stage investors? We can still judge founders. Don't get me wrong. Right. But when a 19 year old founder walks into 20 VC and the product is really, …”
Calacanis: Offshore dev shops and WordPress agencies will completely disappear
“They'll all disappear. We don't need these crummy WordPress agencies and terrible offshore dev shops that never finish a project and charge you 20 or 50 grand. They're all going to be gone.”
O'Driscoll: AI coding startups hit $1B TAM by capturing labor spend
“If you think of this as a tools market, it's probably going to flatten out. If you think of this as replacing all the people using those tools to build crappy products, and you can just compress that labor spend, then you could get to, then the TAM clearly sup…”
O'Driscoll: Deel has more attractive market dynamics than Rippling
“At the margin, I'm, of those two choices, I would bet at the margin, despite a little distaste for what went on in terms of the de-espionage thing, I would say, maybe I want to say is the TAM and the, Competitive matrix is more attractive for deal.”