Apr 17, 2025 · 1h 35m · 20vc
a16z's $20BN Fund & Founders Fund's $4.6BN & Why Josh Kushner Has Mastered the Game · 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 engages venture capitalists Rory O'Driscoll O'Driscoll and Jason Lemkin in a rigorous debate covering megafund dynamics and the decline of traditional SaaS in favor of AI. The conversation also explores the unique investment strategies of top-tier firms like Founders Fund and Thrive Capital, alongside shifting realities in corporate ethics and liquidity.
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 13.8% of the talking time here. How this is scored →
Harry as informed peer 4.7 Guest teaching 5.3 Guest disagreement 2.3 Harry pushing back 3.3
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 34:39 Jason declaring seed investing is for suckers
Jason Lemkin forcefully rejects the seed thesis, arguing that writing late-stage checks into proven mega-winners offers far better absolute returns and carry without decades of stress.
Hardest push from Harry ▶ 29:06 Harry's hard pushback on a16z deal flowHarry Stebbings directly refuses Jason's claim that Andreessen Horowitz sees 100% of S-tier venture deals.
Biggest teaching moment ▶ 1:06:33 Rory reframing long-term private holdings as a policy failureRory O'Driscoll educates the host by demonstrating that tech companies staying private forces ordinary retirement savers to pay 2&20 VC fees instead of 70 bps public fund fees for the exact same underlying growth asset.
Harry holds his own ▶ 1:19:42 Harry explaining the downside of capital stuffingHarry Stebbings demonstrates sharp market insight by arguing that growth funds over-capitalizing companies before they are ready degrades potential $10B outcomes into $4B outcomes.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Opening Titles and Branding Sequence | 3 | 2 | 2 | 1 | Host opens with light banter about Twitter billionaires before citing quotes from Victor Lazarte and Nabeel asking whether traditional SaaS evaluation metrics are dead. | |
| Market Saturation vs. Volatile AI Product-Market Fit | 3 | 5 | 1 | 2 | Rory and Jason explain how AI product-market fit changes rapidly compared to 20-year SaaS stability. Host presses with a targeted follow-up question on what structural changes drive this transience. | |
| Underwriting Risk and the Danger of Bimodal Fund Returns | 6 | 4 | 3 | 6 | Host directly challenges Rory's point on underwriting upside, arguing inflated entry valuations prevent investors from getting compensated for high risk. | |
| Portfolio Construction Debate and the Benchmark Strategy | 6 | 5 | 2 | 5 | Host cites Victor Lazarte's 8% fund check in HeyGen to question benchmark portfolio construction, then asks how traditional 3x/2x growing SaaS companies fit into current venture appetites. | |
| Managing the Two Trillion Dollar Private Asset Backlog | 5 | 6 | 1 | 2 | Host frames the macro liquidity challenge for $2T in private SaaS assets. Rory provides an in-depth breakdown of why private equity playbooks fail on horizontal sub-scale SaaS. | |
| Vertical vs. Horizontal PE Playbooks and LP Forgiveness | 5 | 4 | 1 | 3 | Host names specific overpriced 2021 buyouts like Coupa, Anaplan, and Zendesk to ask if PE buyers get off the hook for bad deals. | |
| Megafund Economics and the 20-Billion-Dollar Strategy | 6 | 5 | 2 | 5 | Host highlights the mathematical disconnect between a $20B fund size and traditional $1B exit returns, forcing a discussion on megafund economics. | |
| Universal Deal Flow and the Mechanics of Capital Deployment | 5 | 4 | 5 | 7 | Host delivers a sharp, direct pushback against Jason's claim that Andreessen Horowitz sees 100% of top-tier deals. | |
| Josh Kushner's Monopoly Strategy at Thrive Capital | 4 | 6 | 2 | 2 | Host asks about heavy dilution in foundation model rounds, leading Rory to break down Thrive Capital's real-estate-like strategy of buying the best property on every block. | |
| Why Seed Investing is for Suckers | 4 | 6 | 4 | 3 | Jason provocatively asserts that seed investing is for suckers in a $100B outcome world. Host questions the underlying correlated valuation risks of buying high-priced late-stage winners. | |
| Entry Price Math and the Micro-VC Counter-Argument | 4 | 5 | 2 | 3 | Host asks Rory for a concrete example where early entry prices protected returns while late-stage investors got burned. | |
| Conservative Fund Sizing and Lessons from Tech Downturns | 4 | 6 | 1 | 2 | Host asks Rory why he keeps fund sizes disciplined despite a strong track record. Rory shares reflective insights on surviving tech crashes between 2000 and 2010. | |
| Macro Inflation and the Strategy of Fund Sizing | 4 | 6 | 2 | 4 | Host questions the viability of $50M seed funds in modern markets, while Rory points to 3x GDP growth since 1999 to explain fund size scaling. | |
| Micro-Seed Fund Failures and the Rush to Founders Fund | 5 | 6 | 2 | 2 | Host cites massive institutional demand for Founders Fund, leading Rory to analyze Peter Thiel's strategy of long-term compounding in SpaceX. | |
| The Risks of Capital Concentration and Singerman's Strategy | 5 | 5 | 2 | 3 | Host quotes Brian Singerman on capital concentration limits and presents a detailed LP budget allocation scenario. | |
| Why Founders Fund Avoids B2B SaaS | 3 | 6 | 1 | 1 | Guests discuss Founders Fund's general aversion to B2B SaaS. Rory explains the difference between backing N-of-1 technological singularities versus broad SaaS markets. | |
| GP Wealth and Capital Commitments as Risk Drivers | 4 | 6 | 1 | 2 | Host asks for a 3-to-5-year outlook on LP capital commitments. Rory explains historical market sentiment overshoots and recalls when LPs turned hostile in 2009. | |
| The Reinvestment Cycle and the Necessity of Major Tech IPOs | 3 | 5 | 2 | 1 | Jason asks if total exit volume must match venture inflows over time. Rory notes market foolishness continues until capital dries up. | |
| Why Late-Stage Companies Stay Private and the Public Policy Failure | 5 | 7 | 3 | 3 | Host references Stripe and Klarna delaying IPOs. Rory articulates how companies remaining private represents a public policy failure that penalizes pension investors with 2&20 VC fees. | |
| Evaluating Safe Superintelligence and the Secret Recipe of OpenAI Founders | 6 | 4 | 3 | 4 | Host strongly defends investing in Safe Superintelligence at a $5B valuation, citing liquidation preferences and guaranteed M&A interest in Ilya Sutskever. | |
| The Illusion of Liquidation Preferences and Aggressive Corporate Dev Strategies | 6 | 6 | 5 | 7 | Jason and Rory explain how corp dev acqui-hires bypass VC liquidation preferences. Host pushes back, arguing that bypassing VC preferences burns long-term reputational bridges. | |
| Founder Secondaries, Capital Stuffing, and the Obsession with Winning Deals | 6 | 5 | 3 | 5 | Host explains how growth funds stuffing startups with $200M before they are ready can shrink $10B outcomes down to $4B through capital inefficiency. | |
| The Deel vs. Rippling Corporate Espionage Scandal and Financial Euphoria | 5 | 6 | 2 | 2 | Host introduces the Deel vs. Rippling espionage story. Rory cites John Kenneth Galbraith's economic concept of 'the bezel' to explain how market booms conceal bad behavior. | |
| ARR vs. GAAP: Why Recurring Revenue Metrics Are Being Questioned | 5 | 6 | 3 | 3 | Host asks why ARR is losing favor compared to GAAP revenue. Rory explains that GAAP represents verifiable legal facts while ARR contains high variance. |
Statements from this episode (68)
O'Driscoll: Tech and investing expertise is not transferable to politics
“I think what it shows is just because you really understand one domain, investing or technology, it doesn't automatically make you understand a totally different domain politics, and I think you're just saying, look, It turns out YQ is not transferable, and yo…”
Lemkin: Scaling $1M to $10M ARR in five quarters is S-tier growth
“One to 10 and five quarters or less is S tier. One to 10 and five quarters or less. And so that's, I think I copied that with attribution. I used that as my investing yardstick for years, right?”
Lemkin: Cloud VC offered nine-figure term sheets without speaking to founders
“Literally, I had two startups in my portfolio, two of them that a VC, all three of us know really about, one of the best cloud VCs, he offered them, I remember term sheets at high nine-figure valuations without talking to the founders, just immediately in late…”
Lemkin: Pure spreadsheet-driven SaaS investing is dead
“You didn't actually need to know what the company did in SaaS for a while, right? That, and so I think that is dead, right?”
O'Driscoll: Traditional SaaS markets are saturated and growth rates have flattened
“The existing market's saturated, so all the growth rates flattened out.”
O'Driscoll: AI startups gain and lose product-market fit repeatedly in two years
“I've had companies acquire and lose product market fit two or three times in a two year period.”
Lemkin: Startup product-market fit duration compressed from five years to five weeks
“It used to take you five years to follow the product market fit. Now it can be five weeks.”
Lemkin: Top portfolio startups work 12-hour days, seven days a week in-office
“All the best startups I've invested in are working seven days a week, 12 hours a day in the office, seven days a week, 12 hours a day.”
O'Driscoll: AI outcomes will likely surpass traditional SaaS outcomes
“In other words, do I think the outcomes of these companies can be huge? And arguably even bigger than some of the SaaS companies. Yes. So quote, the upside is there.”
O'Driscoll: Venture capital returns will become increasingly bimodal
“Yes, by definition. I think, and for two reasons, and we'll talk about it. One is, each individual deal has more risk in it. And then on top of that, the whole, it's a totally separate thing, the holding periods have elongated, right?”
O'Driscoll: Portfolio construction matters because consecutive bad funds hurt brand-name VCs
“I think it always matters. You know, your people actually lose money as distinct from thinking about losing money, and particularly when they lose money, two funds in a row, even a brand name firm can hit a bump, right?”
Lemkin: Will invest in triple-triple-double-double SaaS if CEO is exceptional
“For me, the triple triple double double, I'm totally into it if the CEO is amazing.”
Lemkin: 80% of legacy SaaS VCs now exclusively invest in AI
“I would say 70 to 80% of the folks I grew up with that were SaaS investors are not, won't do those normal triple triple double doubles. You would, and you're, you are a top tier performer, but they're just, they're momentum investors. And they want to put two …”
O'Driscoll: Almost everything that could be done in SaaS has been done
“Anything that could have been done 20 years ago in SAS probably has been done. So I'm not rooting around in SAS land looking for a good deal, right?”
O'Driscoll: Private venture assets total $3T, including $2T in stalled SaaS
“The reason it's a three trillion dollar question is because that's the rough fair market value of privately held venture assets. And, you know, maybe Half a trillion to a trillion of that is high growth, new stuff, and the other two trillion is mature, slower …”
O'Driscoll: VCs must do hard operational work to resolve $2T backlog
“So what's going to have to, there's a huge amount of really grim industrial work that's going to have to be done on everyone's portfolio to manage these companies through to a meaningful exit. Cause you know, as I said, you can't walk away from two trillion do…”
Lemkin: Private equity firms are not inquiring about mediocre-growth SaaS startups
“Folks that in this sort of mediocre growth level, they're getting no, no tire kicking. Are you seeing lots of tire kicking? It's not happening. Cause I ain't seeing it.”
O'Driscoll: Private equity prefers niche vertical monopolies over venture-backed SaaS
“PE guys ironically love the things that we don't love. Let me tell you what I mean by that. They love a boring ass software company. In a teeny tiny vertical, with 40% market share, where they can screw the customers for the next five years by raising prices b…”
O'Driscoll: Cutting R&D in horizontal SaaS degrades retention and kills products
“If you cut off the R&D and the sales and marketing, your gross dollar retention will be 80%, you won't be selling any new shit, you'll be declining, and your product will become irrelevant in two years.”
O'Driscoll: LPs will forgive PE firms for one bad fund, but not two
“I don't think someone's going to say, you know, Tomo Bravo, one of these guys, you're an idiot forevermore. Because you did that dumb thing, right? They're probably not. They're going to say, Ooh, that funds up. Now you put two funds back for our conversation.…”
Lemkin: Frozen PE-owned software products will lose 30-40% revenue in two years
“You fire half the engineering team, Blend set that sales team into someone selling something else, right? And leave the product frozen time in two years, you've lost 30, 40% of your revenue, right?”
O'Driscoll: PE firms will acquire AI startups to dress up legacy SaaS
“I'm willing to bet that all those PE shops have a, let's acquire some new AI Pixie Dust to put on top of Zendesk, to put on top of Anaplan so we can, you know, tart up the story with acquisitions.”
O'Driscoll: Longer private lifecycles drive sustained demand for megafunds
“I mean, I think the biggest thing they have in their favor Is the fact that so many companies are staying private for longer, right? Which by definition means more need for capital, right? Which by definition means if you have that capital, you should be able …”
O'Driscoll: Megafunds will deliver mid-to-high teen returns instead of 3x multiples
“The returns mightn't be three X venture returns, but you know, the competition is the small cap return of 11%. You know, if you're delivering high, mid, high teens, it may be that the LPs think that's great, and that's the bet they're taking.”
Lemkin: Andreessen Horowitz and Sequoia see every single deal in venture
“So if you're Andreessen, a beauty to Andreessen at this point, certainly been true of Sequoia since we started is they see every deal. Andreessen sees every deal.”
O'Driscoll: Higher deal flow makes deal picking more, not less, critical
“You see every deal, which means you see every bad deal, and there are 99 shit deals for every one good deal, so the more deal flow you see, the more important picking is.”
O'Driscoll: Megafund growth ends only when institutional CIOs halt VC allocations
“It won't stop because the venture guys will be mature. It won't stop because the founders will kind of be more careful. It won't even stop because the LPs will stop. It will stop because the LP's bosses, the overall CIOs will stop allocating capital to venture…”
Stebbings: Investors backing AI models at $4B saw only 3.5x at $60B
“I know, and we both know many investors in some of the model providers who came in at four billion, it's now sixty billion and they're three and a half acts up because employee stock Dilution was so heavy, and the funding rounds coming in were so heavy.”
O'Driscoll: Sole investment criterion is making money across the entire cycle
“The criteria is not, is it venture or not venture? There's only one criteria. Is it going to make you money? And is it going to make you money across the cycle?”
O'Driscoll: OpenAI is currently valued at $300 billion
“You can clearly compound from three hundred billion Which is where OpenAI is today to a trillion because five other companies did it, right?”
Lemkin: Late-stage fund matched my seed stake with nine-figure unicorn check
“I just have a company that's just become a unicorn and a late stage fund just put in almost nine figures. Okay. And they own as much as me.”
O'Driscoll: Megafunds with forgiving LPs should pursue late-stage investing over seed
“If you have access to capital that's large and forgiving, which is what these mega funds have, then you should play the big balls game. Cause you're exactly right. You don't do it. You have to do less work. If it works great, you make out the same as Jason who…”
O'Driscoll: 70% of tech investors in 1999 left industry by 2004
“When, you know, 70% of the people I knew in 99, 2000 were out of the business four years later.”
O'Driscoll: Tiger Global and SoftBank are out of the venture game
“I mean, the big, I mean, the big momentum, the two biggest momentum players Of the last decade, Tiger and SoftBank are already out of the game.”
O'Driscoll: Insight Partners survived mega-fund risk through savvy management
“You know, huge credit to someone like Insight, who were putting out a lot of money, but managed to survive because of Savvy, right?”
Stebbings: Insight generated $2.6B Wiz return in $8.5B fund
“It was a 2.6 billion dollar reported return to them in an eight and a half billion dollar fund, which is like one of the best investments of all time.”
O'Driscoll: Fund size dictates strategy in venture capital
“And you've got to size the fund for the strategy because fund size is the strategy.”
Stebbings: $50M seed funds cannot lead 20 deals at modern sizes
“Because actually when you take away fees, you've got 40 of investable, and when you think about the average seed round today being three to five million dollars, if you want to lead it and take real ownership like they say they do, there's no way that you're g…”
Stebbings: Founders Fund sees unprecedented institutional LP demand
“I've never had such institutional demand for any single fund asset than I have for founders fund. Every single LP wanted Founders Fund, and wanted Founders Fund gross, which is even more rare.”
O'Driscoll: Founders Fund is to a rounding error the best VC fund
“They may be, to a rounding error, the best fund, so no surprise to get the most amount. I mean, you know, it was great to see the leak now. They're astonishingly good.”
O'Driscoll: 15-year holding periods yield 8x to 10x venture fund returns
“If you compound at 30, 40% gross, not for eight years, but for 15 years, because you don't give a damn about giving the LPs money back early, you're just going to compound the thing to make money, then you end up with an eight or 10 X fund.”
O'Driscoll: Capital concentration limits prevent fund wipeouts but cap peak returns
“And it is the, and it is exactly right. But now again, to take both sides of that, it is the enemy of greatness and it is the protector of massive wipeouts, right? And it just boils down to the personal choice of where in that dimension you want to be.”
O'Driscoll: Founders Fund scored 5x on Stemcentrx before program was canceled
“Got a five X, get money off the table. We're pointing out three years later, the acquirer canceled the program. You know, huge amount of risk. Made, I mean, you know, made it work, but never forget the risk was there.”
O'Driscoll: LPs cannot replicate Founders Fund's returns with other managers
“An LP saying I'll do founders fund because they got these returns and 10 other funds, they'll be just like founders fund. That sentence doesn't make sense. They won't be just like founders' funds because They're not the same people with the same approach.”
O'Driscoll: Founders Fund leads large-scale VCs in generating 8-9x returns
“Based on observed data at scale, in other words, not including seed funds, but in terms of turning industrial quantity of money into eight and nine X's, let's put founders fund at the top.”
O'Driscoll: Venture LPs should target repeatable public market outperformance
“You build a portfolio of companies whose strategy says not that they can replicate the best number, but that they can comfortably outperform the public markets with a strategy that's repeatable and differentiable.”
Lemkin: Founders Fund intentionally avoids investing in B2B software
“Founders Fund doesn't do B to B intentionally. They make exceptions, but they don't believe in B to B. They don't believe the outcomes justify it. It's explicitly, they don't do it.”
O'Driscoll: B2B software has produced 200 to 300 venture winners
“Very few companies in B to have the same level of untrammeled, competitive free space that something like SpaceX does. That's the negative on B to B, but the positive on B to B is there's been two to 300 SaaS winners”
Lemkin: Founders Fund categorized Ramp as fintech to maintain non-B2B stance
“They told them they don't do B to B investing, then that viewed ramp as a fintech.”
O'Driscoll: Best venture investment window is when sentiment is unanimously hostile
“It will not be a great time to invest in venture until people spit at you when you mentioned the word.”
O'Driscoll: Market participants stop bad behavior only when capital runs out
“In general, my observation is people don't stop doing stupid shit because they intellectually figure out they stopped doing stupid shit. They generally stop doing stupid shit when there's no more money to do stupid shit.”
O'Driscoll: Capital flows to venture will fall if major IPOs stay delayed
“If on the other hand, if those keep pushing out, then at some point it becomes really hard to have all this money in private illiquid assets when you have pressures on your endowment and all that. And then at that point you would see the money go down.”
O'Driscoll: Late-stage startups staying private is a massive public policy failure
“And it's a massive public policy failure because what's happened here is this. It is more attractive for companies to stay private and access capital from GPs who are paid two and 20 plus to make those investments than it is for those same companies to go publ…”
O'Driscoll: Late-stage private investments will eventually underperform public markets
“At some point, what logically will happen is the late state private investments will underperform equivalent public investments by the amount of the fees. And then it'll switch.”
O'Driscoll: Non-OpenAI foundation model startups have underperformed Anthropic
“If you fast forward three years, all the foundation model companies that weren't populated by people who were at OpenAI haven't done great, and Anthropic that was populated by people that came from OpenAI has done pretty well.”
Stebbings: Safe Superintelligence is guaranteed to cover its liquidation preference
“With a LickPref, there is zero chance this does not get bought for at least LickPref.”
Lemkin: Acquirers in nine-figure deals aggressively circumvent VC preferences
“My limited visibility recently in M&A is that every acquirer is looking for ways to get around all the VC preference. It's aggressive. Like it was always true, but now it's like super aggressive. It's like, we just don't Even give a rat's ass in corp dev how t…”
O'Driscoll: Scale's portfolio companies bypass VC preferences in acqui-hires
“When I'm on the other side of the table and my late stage companies are trying to buy early stage companies, I do exactly the same thing. I don't give a shit about Jason and his bloody preference. I want to hire those five great engineers. Let's just give them…”
O'Driscoll: Lost a deal by refusing founder equity refresh concessions
“Yeah, I've lost a deal to not doing that, and you're right, I'm going to be sympathetic to the investor now. I've lost a deal to not doing that, because again, it sticks back to my comment, I tend to be perhaps stuck in the, we, it's not war, stuck in the mud …”
O'Driscoll: Enterprise customers must replace payroll providers facing criminal liability
“If you're relying on them to manage your payroll, to move money on your behalf, you possibly can provide them having some kind of civil liability. But if it trends over into criminal liability, you probably have to find a new payroll provider, right?”
Lemkin: A hundred corporate espionage and founder fraud cases will surface
“I think you're going to hear a hundred, just like fraud. Like every day now we pull up the media and there's another founder that stole thirty million from the investors and we shrug it off. Right. There's going to be a hundred of these in this environment.”
O'Driscoll: Scale Venture Partners shifts focus from ARR to GAAP revenue
“I mean, we've started really focusing on GAAP revenue now because ARR is a made up number and GAAP numbers are fact.”
Lemkin: Most reported ARR in modern startups is neither annual nor recurring
“Yeah, most of the A and the R and the R aren't real. Yeah, it's not really annual. The third one revenue. Yeah, there may not be revenue definitely doesn't recur and no way it's annual if everyone can get out after a month or two. So it's neither a nor.”
Lemkin: 93% of B2B sales professionals admit lying to win deals
“I asked 2000 folks in SASTR how many folks lie in deals to win deals? 97, 93% said they lied over 2000 to win deals. If you're lying, if 93% of 2000 B to B folks are lying to win deals, lying about features, lying about feature gaps, okay.”
Lemkin: Every departing sales rep violates laws by taking client contact lists
“How many folks take their Rolodex with them when they leave, which violates many laws, all of them.”
Lemkin: Corporate scandals impact new B2B sales but cause minimal churn
“It's going to hurt you for new customers. Because it's a weapon for the sales team to use against you. I say two percent, like even just churn, you know how much work it is to change payroll providers? It ain't worth it.”
Lemkin: All my venture investments above $100M valuation have performed poorly
“I recently took a look at my investments. I just, I can't make any decision well north of a hundred, so I'm out. All of my decisions are bad. North of a hundred, they're just all bad for a variety of reasons.”
O'Driscoll: Venture capital economics depend entirely on generational outlier companies
“The whole reason this business is awesome is there are singularly amazing companies in every generation. And maybe these are they, and when you do those companies, everything works and you're just so glad you bought it at any price, right? That's why this game…”