Aug 7, 2026 · 1h 30m · news
The AI Boom Will Create Enormous Roadkill: Who Wins & Loses? | David Frankel
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 interviews David Frankel, Managing Partner at Founder Collective, exploring early-stage venture fund economics, the AI valuation bubble, founder psychographics, and long-term liquidity strategies.
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 23.8% of the talking time here. How this is scored →
Harry as informed peer 6.0 Guest teaching 4.7 Guest disagreement 2.4 Harry pushing back 3.8
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 34:27 Momentum trading is not my business
Frankel bluntly dismisses Stebbings's premise that $1B is the new Series A, categorizing that framing as momentum trading rather than true venture investing.
Hardest push from Harry ▶ 25:45 Challenging fund size disciplineStebbings directly challenges Frankel on leaving returns on the table by refusing to raise larger opportunity vehicles alongside his core fund.
Biggest teaching moment ▶ 22:34 The risk of orphaned seed investmentsFrankel explains the structural risk when junior VCs at mega-funds move on, leaving high-priced seed founders orphaned without internal board champions.
Harry holds his own ▶ 33:50 Asserting the $1B Series A realityStebbings marshals real-time market data from Cursor, Cognition, and Mercor to substantiate why top-tier AI rounds command billion-dollar price tags.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Reflecting on an 11-Year Relationship and Early-Stage Belief | 6 | 5 | 3 | 4 | Stebbings poses a sharp thesis that $50M-$100M funds are in the dead zone of venture capital. Frankel responds by dissecting historical data across the top 500 venture outcomes to explain why patient seed investing still returns funds. | |
| Co-Investing Strategy and Capital Efficiency in Seed Rounds | 6 | 4 | 3 | 5 | Stebbings probes whether Founder Collective can realistically co-invest small checks alongside aggressive multi-stage platforms and questions value investing in non-AI sectors. Frankel clarifies their positioning as an insurance policy and highlights off-piste opportunities. | |
| True Entrepreneurs vs Founders and the CEO Growth Trajectory | 5 | 5 | 2 | 4 | Stebbings notes the professionalization of founders from top colleges and questions turning down deals based on secondary co-founders. Frankel contrasts trendy founders with resilient entrepreneurs, emphasizing the steep CEO learning curve and founder alchemy. | |
| Founder Psychographics and Venture Scaling Expectations | 7 | 4 | 3 | 5 | Stebbings questions whether traditional growth pacing like triple-triple-double-double is dead in modern venture. Frankel counters with the reality of extended fund lifecycles, citing SeatGeek's long journey and warning against abandoning steady compounders. | |
| AI Waves, Expected Market Roadkill, and Seed Differentiation | 6 | 5 | 2 | 3 | Frankel lays out historical return distributions and predicts massive roadkill in the current AI bubble, defending seed math. Stebbings prompts on whether seed investing has reached peak bubble conditions. | |
| Mega-Platform Mechanics, High Valuations, and the Orphan Risk | 6 | 6 | 4 | 5 | Stebbings plays devil's advocate, suggesting mega-platforms win if seed is commoditized and cheap capital is attractive to founders. Frankel pushes back by outlining the danger of junior VC turnover leaving portfolio companies orphaned. | |
| Fund Size Discipline, LP Alignment, and Resisting AUM Expansion | 7 | 5 | 3 | 6 | Stebbings directly challenges Frankel on why Founder Collective resists expanding AUM into larger opportunity funds when competitors do. Frankel explains GP skin-in-the-game, alignment on DPI, and personal fulfillment at the early stage. | |
| Ownership Targets, Dilution Flexibility, and Missed Deals | 6 | 5 | 3 | 4 | Stebbings shares his firm's mistakes of passing on winners over low ownership percentages. Frankel explains that he never passes on exceptional founders solely due to ownership constraints, citing his early Suno investment. | |
| Is $1 Billion the New Series A? Valuation Realities vs Momentum | 7 | 5 | 4 | 6 | Stebbings aggressively asserts that $1B valuations have become the new Series A baseline for top-tier AI companies. Frankel firmly rejects this premise, stating Stebbings is describing a momentum trading business rather than value-focused seed investing. | |
| Managing Momentum Assets, Early Value Capture, and Secondary Liquidity | 6 | 5 | 3 | 4 | Frankel and Stebbings analyze early value capture and the mechanical differences between capturing 80% of value early versus chasing follow-on momentum. Frankel acknowledges the contrast between their investment philosophies. | |
| Long-Term Applied AI Themes and Fund Return Concentration | 6 | 5 | 3 | 5 | Stebbings argues large platform funds can deliver massive aggregate dollar returns, but Frankel highlights the shift in LP base to sovereign wealth funds seeking IRR over fund multiples. Frankel notes how early applied AI bets in Fund II were made long before current hype cycles. | |
| AI Cannibalization of SaaS, SaaS-pocalypse, and Embedded Software Moats | 7 | 5 | 3 | 4 | Stebbings highlights the threat of AI cannibalization hitting software leaders before liquidity events occur. Frankel argues embedded workflows and mission-critical systems provide defensive moats against software disruption. | |
| High-Conviction Underwriting: The 10x Next Round Framework | 6 | 5 | 2 | 4 | Frankel describes Founder Collective's underwriting rule requiring confidence in a 10x return by the next round and explains why he funds founders with deep vertical domain edge. Stebbings humorously clarifies cultural definitions of nepo babies. | |
| Navigating Secondary Markets and Portfolio Liquidity Strategies | 7 | 4 | 2 | 3 | Stebbings and Frankel discuss the growing liquidity in secondary markets and the discipline of taking 20% off the table to return fund capital early, reflecting on past exits in Uber and SeatGeek. | |
| Mega-Rounds, LP Evolution, and the Financialization of Venture | 6 | 5 | 2 | 3 | Frankel details the increasing financialization of venture, noting whole-portfolio secondary transactions and LP minimums. He reaffirms his preference for hands-on founder interaction over fund financial engineering. | |
| Founder Focus, Loyalty, and Second-Time Founders | 5 | 4 | 2 | 3 | Stebbings asks whether founder focus and loyalty are declining due to side projects and angel funds. Frankel notes that moderately successful second-time founders often outperform wildly wealthy repeat founders due to sustained hunger. | |
| Wealth, Unreasonable Expectations, and Everyday Kindness | 5 | 4 | 1 | 2 | Stebbings and Frankel discuss how wealth elevates personal expectations and reduces patience, with Frankel emphasizing that kindness and human connection matter most in late-career venture. | |
| AI Platform Shifts and Competitive Disruption | 6 | 5 | 2 | 3 | Stebbings asks about the market impact of trillion-dollar AI foundation model companies. Frankel notes that historical platforms always face disruption and predicts Google's contextual search gives it strong defensive capabilities over Microsoft. | |
| AI Workforce Impact, Productivity Gains, and Vertical Expertise | 6 | 5 | 2 | 4 | Stebbings questions whether older knowledge workers can retrain as easily as AI-native youth. Frankel points out that deep vertical domain expertise and human trust remain indispensable in high-stakes service sectors. | |
| High-Stakes AI Models and Enterprise Query Valuation | 5 | 4 | 1 | 2 | Stebbings shares an anecdote about enterprise queries fetching eight-to-nine figure valuations for critical decisions. Frankel adds that high-value governmental and defense simulations will command massive willingness to pay. | |
| US vs. China Geopolitical AI Competition | 6 | 6 | 3 | 3 | Frankel gives historical context on global economic shifts and highlights China's aggressive development in open-source AI, biotech, and hardware. Stebbings and Frankel discuss regulatory divergence between the US, Europe, and China. | |
| Seed Arbitrage, Value Investing, and Venture Strategy | 6 | 5 | 2 | 3 | Stebbings asks what market condition would convince Frankel to expand fund size. Frankel explains that only clear pricing dislocations or orphaned Series A/B opportunities would justify raising larger vehicles. | |
| Long-Term Founder Journeys, AI Disruption, and Decision Frameworks | 6 | 4 | 2 | 4 | Stebbings notes the low IRR on long-duration exits like Olo's 17-year journey, while Frankel defends the satisfaction of long-term founder partnerships. They discuss Suno's rapid path from creation to consumption tool and historical misses like Klaviyo. | |
| Relationship Principles, Future Technological Frontiers, and Farewell | 4 | 3 | 0 | 1 | Frankel shares personal principles around presence and relationship building, followed by reflections on autonomous vehicles and AI-driven medical breakthroughs before a warm farewell. |
Statements from this episode (69)
Stebbings: $50M–$100M Seed Funds Will Be the Worst-Performing Vintage
“The hardest part of the market is seed in many ways, and the worst performing funds will be the 50 to a hundred million dollar funds. I say this to explain because you're too big to be collaborative, To write those hundred to 250 K checks and be a friend, but …”
Frankel: Fewer Than 100 Companies Created Since 1999 Sustainably Top $10B
“There were less than a hundred companies over the last 25 years, less than a hundred, that are sustainably over ten billion dollar companies.”
Frankel: Median of top 500 companies created in 25 years is $2.6B
“The median of the top 500 companies created in the last 25 years, the median is 2.6 billion.”
Frankel: On-piste deals are always expensive; best founders are off-piste
“If you're on piste for the last, I've been doing this for 18 years nearly, is, it was always expensive. It was always tough. But you find some of the best people off piste.”
Frankel: Founders Use Founder Collective as Insurance in Mega Seed Rounds
“I think we're being seen and I'm, I could be over extrapolating the last 20 deals that we've been involved in almost as an insurance policy. Where we're side by side, we're putting in 500 K or a million. There's been eight, nine million dollars going in, and T…”
Frankel: Heavily Funded AI Startups Are Not Capital Efficient
“There's very little evidence yet that These hot, hot AI companies that are raising huge amounts of money are capital efficient, right? They're anything but capital efficient. There's, like, the jury's out on whether that's going to work still.”
Frankel: Saying a startup is not AI is like not using the internet
“Well, look, everybody's AI, right? Like, it's almost like saying that you're not AI today is like, I'm not using the internet, right? It's like, why wouldn't you use the most contemporary tools? So everyone's AI.”
Frankel: Uncapped Convertible Notes Suck at the Seed Stage
“I mean, the scale of how much you have to win, right, is different based on your price. It's pure math. So, uncapped notes suck at the seed stage.”
Frankel: Co-Founding CTOs Are Often Redeployed After Scaling Past 50 People
“The CTO at some point up to 50 people, you're golden. And then at some point you go, actually we could bring in better, better technical skills. And if you've got a good co-founding CTO, that person becomes like a Swiss army knife and is deployed in different …”
Frankel: Jeff Bezos Stated He Spent 50% of His Time Hiring
“I had lunch years ago, decades ago, with Jeff Bezos. I was invited to a lunch, and someone smarter than me said, what do you spend your time doing? And he said, 50% of my time is bums on seats. That's never left me.”
Stebbings: Non-CEO Co-Founders Usually Leave Within Three Years
“The other co-founders most often not there in three years.”
Frankel: CEO Salesmanship Outweighs CTO Due to Role Fungibility
“The CEO being a good salesperson and being a real entrepreneur is actually more important because the CTO role can be fungible depending on how complicated it is.”
Stebbings: Scaling From $1.5M to $15M ARR Is No Longer VC-Fundable
“I turned down a company the other day that went from one and a half, and they were gonna go to five, and then they were gonna go from five to 15, and it's just not enough anymore. It's just not interesting. I'm sorry for venture. We have an opportunity cost of…”
Frankel: Standard 10-Year Venture Capital Funds Now Take 18 Years to Mature
“You know, these ten-year funds are taking 18 years.”
Frankel: Founder Collective still holds every share in 2010 investment SeatGeek
“Harry, we still own every last share in SeatGeek. That was an investment I made in 2010.”
Frankel: OpenAI, Anthropic, and SpaceX Are Today's Metas and Googles
“And I think in, you know, OpenAI and Anthropic and SpaceX, we're seeing that already. Like these are the Metas and the Googles of our era, highly likely.”
Frankel: Seed Investing Is Crowded and Largely Commoditized
“I think seed is crowded and to some degree very commoditized.”
Frankel: Mega-Fund Seed Checks Are "Call Options" Bad for 95% of Founders
“Well, the problem is the mega platforms are taking call options. So is this good for the mega platforms? Is this good for the LPs or is this good for the entrepreneurs? Well, probably for 95% of the entrepreneurs, it's not good.”
Frankel: Junior champions leaving mega-funds leaves portfolio startups orphaned
“And the more junior, the principal at that big fund moves on. They start their own fund, they move to another fund, happens all the time, right? So the person who invested doesn't have mandate. They can't sit around with a partnership and say, look, let's just…”
Stebbings: Wix is trading at 1x revenue with a $2.1B valuation
“The market can stay irrational longer than you can stay solvent, and when I look at, like, a Wix today, trading at 2.1 billion on 2.1 billion of revenue, it's a great example where, like, there's obviously rationality at play, but it doesn't matter.”
Frankel: The jury is out on post-2020 venture fund DPI performance
“A little less since twenty-twenty. Like, if you look at the DPI analysis, like, the jury's out from twenty-twenty onwards.”
Stebbings: 20VC Lost Hundreds of Millions Passing on Deel Over Ownership Targets
“I look at ours, and our biggest mistake, and I can look at Deal, 11 Labs, I can look at Granola, and Starcloud, Fractile, could have done them all, but would have had one to two percent, and all of them we turned down purely for ownership, and that is hundreds…”
Frankel: Founder Collective Has Never Passed on a Startup Over Ownership Percentage
“I've never turned it down. Never.”
Frankel: Pro-rata investment rights are generally bad for entrepreneurs
“I think pro rata is generally not great for entrepreneurs. It's a call option against you.”
Frankel: Founder Collective has never led a follow-on round
“We've never, ever led another round. So we have this view of like, it would be negative correlation bias. It would be unfair to everybody if we didn't be somewhat kind of uniform.”
Frankel: Following On in Every Round Boosts Absolute Returns, Not Multiples
“If we follow, our view would be we'd had, we would have had to have followed on in everything, and I think that the absolute return would be better. I don't think that the multiple would necessarily be better on the fund.”
Stebbings: The $1 Billion Valuation Has Become the New Series A Standard
“I think a billion dollar valuation is the new series, eh? And you're like, whoa, Harry, whoa, whoa, kiddo, calm down, listen to the facts. We used to do a fifty million post and hope it would become a 1,000,000,020 x, without dilution, blunt. Now, you enter at…”
Frankel: Billion-Dollar Early Rounds Are Momentum Plays, Not Seed Value Investing
“I think that that's the momentum business, and I think knowing how and when to get out quickly with some of those really, really matters, and that's not really my business. So my business is value, is getting involved early, and trying to find value opportunit…”
Frankel: Founder Collective seeks exits after capturing 80% of value
“Our knee joke tends to be when this has gotten across a certain point, like we're out of here and credit to Eric at a point for going like we've captured 80% of the value. We could capture another 20%.”
Stebbings: Mega venture platforms will achieve venture returns via massive outcomes
“I think with the outcome expansion that we've seen, cursor at sixty billion trillion dollar companies in a matter of years with OpenAI and Anthropic, you will see
Venture returns with mega platform sizes.”
Frankel: Sovereign wealth funds prioritize IRR over fund return multiples
“Sovereign wealth funds and public investment corporations are looking for IRR. They're not measuring this in how many times do you X the fund.”
Frankel: Post-2020 mega funds boast high TVPI but lag on DPI
“Subsequent to twenty-twenty, like the TVPI is there, and some, in some cases, they're on steroids. The DPI is less there, if you look at the actual stats.”
Frankel: Larger fund sizes make achieving high DPI much tougher
“And we're working for DPI, and the bigger we make the fund, the tougher it is on the DPI.”
Frankel: Seed investors must enter themes 5 to 10 years before momentum builds
“The job is to be in there five years or 10 years ahead, and it's not where the momentum is. It never is.”
Frankel: Embedded, Mission-Critical SaaS Workflows Will Survive AI Disruption
“The more embedded you are, like the more difficult you are to dispense because real time, Thousands, billions of orders are being run in your system, or like mission critical biotech research is being run in your system. The more embedded you are, I think the …”
Frankel: Buying top beaten-down SaaS stocks will deliver good returns
“The contrarian in me, this is not what I do, would say buy a basket of like the top SaaS stocks that have all lost huge market cap. You're going to do okay.”
Stebbings: I invested in Palantir momentum over beaten-down SaaS
“Rory O'Driscoll, who we do the show with every week, has done that, and I put my money into Palantir and said, I'm a momentum surfer. I did better.”
Frankel: Founder Collective Requires 'I Love It Because' to Invest
“We have at our team meeting, I love it because dot dot dot. If you can't complete that sentence, you can't invest.”
Frankel: Founders raised inside specific verticals possess unmatched operational edge
“There are folks who have been in these verticals since they were kids. He watched his uncle in this vertical. It's like, there was nothing else he was going to do. And I go, like, they have more edge, and they know what to do with.”
Stebbings: Seed managers will be defined by secondary market navigation
“I think this generation of seed managers will be defined by their ability to access and navigate secondary markets effectively”
Frankel: IPO market will probably remain open through end of the year
“An IPO market that will probably be open for the remainder of this year. And then these IPO markets always close.”
Frankel: Seed VCs Should Sell 20% of Winners Early to Generate DPI
“Even in your top names, sometimes taking 20% off the table, if you can return 25% of the fund, particularly if it's a newish fund, so if it's a, you know, if it's a twenty-twenty-four fund and you can give back 25%, like, why wouldn't you do that? And you're s…”
Stebbings: VCs Should Take 50% Secondary Discounts Over Waiting Years for IPOs
“I just think we don't think about the velocity of cash enough, and what I mean by that is like, you know, yes, there might be another double, but if I have to wait five years, and then the IPO, and then an 18 month lockup, Jesus, give me 50% of that now, and I…”
Frankel: Founder Collective sold some Uber early near $10B valuation
“You know, in retrospect, we probably sold a little too early, so this was early on, you know, this is a business that's getting close to ten billion dollars in valuation, and there's an opportunity to take some off the table, and you're very new.”
Frankel: Suno experienced far lower dilution than Whoop due to rapid momentum
“Like we're so proud to be in both, but sooner, you know, sooner has been a very quick journey. So if you look at like how lower, how much lower the dilution is, part of it is just how quick the momentum of that has been versus a whoop, which is hardware took a…”
Stebbings: LPs Have Reverted to Prioritizing Paper TVPI Over Real Returns
“I speak to a lot of LPs now, and do you know what, honestly, we can say what we want. They've gone back to wanting TVPI. They've gone back to wanting big numbers. And yes, they want DPI, of course, they always want DPI. But they are still very impressed by TVP…”
Frankel: Founder Collective Is Too Small for LPs Requiring $50M Checks
“A lot of them, like, have minimum size checks now, so we're too small for quite a few of them. It's like, if I can't put fifty million dollars to it, and I think there's, it just reflects the inflation of the entire environment, and there are a bunch that real…”
Frankel: Fund of Funds Are Liquidating Entire Billion-Dollar Portfolios via Secondaries
“By the way, we've seen fund of funds do secondaries of their entire fund, So we go, oh my god, you're in fund two, or you're in fund four, like, you should never sell, right? Like, this is what's, and it's like, it's not about you, right? Like, you're a roundi…”
Frankel: Founder Abandonment and Lack of Focus Remain Rare Minority Cases
“I still see that the vast minority of the time. Like, I think it's easy to extrapolate and go, that's a trend, and I could be very Pollyannish about this, but for the most part, like, I see founders wanting to make it work.”
Frankel: Moderately Successful Second-Time Founders Outperform Prior Massive Winners
“And I think overall we've done a little better on second time founders who didn't do that great up front. They didn't, they did okay, right? It's life changing. Like the first million dollars is so life changing. But they're really hungry. They've learned some…”
Stebbings: OpenAI and Anthropic Will Reach $1 Trillion to $2 Trillion Valuations
“Both will be trillion dollar plus, potentially close to two trillion.”
Frankel: Microsoft Has Done a "Crappy Job" With AI
“So I think Microsoft have done a crappy job of AI generally.”
Frankel: Google Is in Pole Position in AI Due to Search Context
“Like I think Google's, Google is actually, if anything, in pole position because they come from that environment and the ability to search with context, the ability to apply AI with context is just like incredible, but they're having to fight like crazy for it…”
Frankel: Top-tier tech returns will spill into venture and broader investing
“So the returns at the top are going to be Incredible. They have to be now. And I think that that capital is going to spill over into venture and all sorts of investing.”
Frankel: The Tech Market Is "Definitely" Headed for Another Dot-Com Crash
“Are we headed for another dot com crash? Definitely. If, like, Is not a question. When? Nobody knows, right?”
Frankel: Angel networks can increasingly bypass traditional venture firms
“It may just be in, like, angels putting money all over the place, and some of those angels are going to know people that worked with them or for them, and they're going to, you know, I think you can bypass traditional venture to a great extent, and that's the …”
Frankel: AI Will Not Cause Mass Unemployment, Only Massive Productivity Gains
“Do I think that we're going to have mass unemployment because of AI? And you're seeing a lot of leadership now agree with the viewers. No, I think we're going to see tremendous productivity gains.”
Frankel: Governments and Defense Are Prime Markets for Predictive AI Simulations
“I think that governments and defense organizations, some kind of speculation with data of the future I think that's a very interesting play.”
Frankel: Tesla received massive non-dilutive government funding under Biden
“Is the non-dilutive government funding that Musk got for Tesla from the Biden administration was huge.”
Frankel: OpenAI and Anthropic Will Eventually Be Disrupted by Chinese Competitors
“That anthropic and open AI are going to be disrupted. It's like unequivocal. Like our whole careers are about disruption. Those platforms never, ever stay forever. Where is it going to come from? Excellent chance it comes from China. It's coming.”
Frankel: Optical chips will solve data center energy drain
“You're going to see optic chips with which are very, very energy compliant.
So when people talk about the data centers and the energy sucks, that's going to change.”
Frankel: Photonic Computing Will Disrupt or Be Acquired by Nvidia Within a Decade
“In my view, if you say in 10 years time, and I am not a thematic investor, but I am such a deep believer in
The status quo being changed always, and not, and like nothing stays the same.
I think photonic computing is coming down the line, and I think that's go…”
Frankel: The US underinvests in energy R&D compared to China
“If you look at the amount of money that's being spent in China on energy efficiency and energy research now, I don't think we're spending enough.
And by that, that's a negative of the Trump administration is we need much more money being spent on R&D.”
Frankel: The early seed-stage valuation arbitrage has completely disappeared
“There was an arbitrage. There was a real arbitrage because the risk premium at the seed stage was way overstated. That has changed completely.”
Frankel: VCs who claim they knew a company would become huge are dishonest
“Anyone in my seat who says, I knew, is just full of shit.”
Frankel: Kalshi and Polymarket are essentially gambling like DraftKings
“It has to be, right? Like it, I mean, if you look at like Kelsey and Polymarket, what's the difference there between DraftKings and Betway? And they seem very similar to me.”