20VC：David Tisch谈高持股与高集中度为何是胡扯｜为何没有基金管理人真想一直做小｜为何与超级平台竞争注定是一场败仗｜为何《精益创业》是假的？

I’ve never understood trying to compete head-to-head against who I would deem to be the best VCs in the world. Every concentrated manager will say that we win 90% of the deals we want to win, and I think that’s total BS. I think it’s 20%. I think The Lean Startup was a total fake.

There are people right now who, to me—and this happens in hot market conditions—are playing a short game. I believe the right to invest in subsequent rounds is earned. I work for our LPs, and my customers are founders. We don’t allow grenades at early stage. You’re not allowed to throw a grenade on someone else’s deal.

This is 20VC with me, Harry Stebbings. Now, over the years, David and I have done many shows, and I’ve always questioned his model: low ownership, very diverse portfolios. And do you know what? I was wrong, in many different ways. David has absolutely crushed it. Yes, he was an investor in Cursor from one of its earliest rounds to a $60 billion sale to Mr. Elon Musk and X. What an incredible outcome that was for David and BoxGroup, returning the fund multiple times.

But he’s done it with so many companies, from Base10 to Clay to Rogo. David has crushed it. He’s one of the best early-stage investors of our time. He’s a true craftsman in a world where everyone has become a capital accumulation machine. David is one of the best. This was a phenomenal discussion, and I can’t thank David enough for being so open and, in all honesty, willing to go against the grain with some of the things he said today. I loved making this happen.

But before we dive into the show today, what do Uber, Cursor, and Harvey have in common? Well, they made the really wise decision to build on Fireworks. Fireworks is the specialized intelligence platform behind many of the world’s leading AI products. Companies use Fireworks to deploy the latest open models, specialize them with their own data, and run them in production with the speed and reliability modern AI applications demand. But getting models into production is only the beginning. As AI usage scales, the best model for one request isn’t always the best model for the next. Well, that’s where Fireworks Nexus comes in. Nexus connects to the AI coding tools and harnesses your engineers already use and intelligently routes each request to the best model for the job, helping you optimize for quality, latency, and cost without changing how your engineers work. To see intelligent model routing in practice, visit fireworks.ai/20vc to create a free account, and use the promo code 20VC50 to claim $50 in credits for 20VC50. While Fireworks powers AI, Laradin helps you measure its impact. Every company now runs two workforces, humans and AI agents. Laradin is the first unified dashboard across both workforces that measures their work and optimizes their impact. Laradin answers four questions. Who’s using AI? Which tools? What is it costing? And what’s it actually being produced for every team across every vendor? AI ROI stops being a guess and becomes a number that you can consistently and confidently measure. Gainsight, Vertiv, SurveyMonkey, and many other enterprise organizations run their AI programs on Laradin. Thousands of AI tools and millions of dollars of AI spend are measured. You bet on AI, Laradin proves that it’s working. Visit laradin.com to learn more. That’s L-A-R-A-D-I-N.com. While Laradin tracks AI impact, Zero Hash powers your crypto infrastructure. Every great software company eventually runs into money. Uber had to move it. Shopify had to hold it. Airbnb had to settle it. Money movement stopped being a fintech problem. It became a software problem. And increasingly, God, it’s an AI problem, too. Zero Hash is the infrastructure that makes global instant money movement seamless. One API integration for stablecoins, digital assets, and modern payment rails, so builders can stay builders. Zero Hash powers some of the world’s largest enterprises and financial institutions, including Calshi, Stripe, Morgan Stanley, Gusto, and Interactive Brokers. If you’re thinking about stablecoins, digital assets, and the future of money, it’s time to talk to the team at Zero Hash. Visit zerohash.com/20vc to learn more. You have now arrived at your destination. David, I’m so excited for this, dude. It’s been two and a half years since you joined me. You very kindly reminded me of a debate that we had a couple of years ago, and we’re going to get into that. But thank you so much for joining me first.

Thanks for having me back. Nothing’s happened in between the past two and a half years and now, so we can just sort of regurgitate all the old conversations.

Has our job completely changed, David?

1. The Market Has Changed

In some ways, totally, and in other ways, not. I’ve been doing early-stage investing for 17 years now, which makes me old. A lot of the time, people try to relate the current moment to a historical moment. So, is this like 2021? To me, no. It has nothing to do with ’21. It’s a totally separate moment, and this one’s funky and interesting and maybe frothy, maybe insane.

Whatever it is, it’s a standalone new market condition that you have to figure out in today’s world. In ’21, there were a bunch of other macro and micro factors creating that environment. And if you go back to ’18, ’15, and 2010, each one of these moments is just different. So I could argue that our job has totally changed and equally explain that our job is just to do whatever the current market conditions present.

The actual job is: see, meet, get to know great founders. Make a decision: should we invest or not? And beg them to take your money. That’s the funnel that we operate in.

I totally agree with you there. In 2021, I think we can acknowledge it was just incredibly bubble-like. It was a good time to sell; it was a bad time to buy. There was no fundamental technology innovation. There was short-term behavioral change with COVID in different respects: communications, buying. Everyone says, “Well, we’re now starting to feel like 2021.” Are you in the camp that the craziness is real, or are you in the camp that, no, we are so early and there is so much room to run?

I don’t think it’s as simple as that. In each one of these moments, there are good companies and bad companies. As a VC generally, and especially as an early-stage VC, I think the macro is less impactful than the micro. Is the company that I am fortunate enough to be an investor in the good one? Is it the great one? Is it the most important one? That transcends the macro. To me, the macro is an industry analysis, a portfolio analysis, but it doesn’t come down to a specific company analysis.

So, do I think there are not-perfect companies raising a lot of money at big valuations? Yeah. But I don’t know that that’s ever not the case, and I think you’ve seen it historically in our industry. There are incredibly public blow-ups of some of these companies that happen in catastrophically messy ways, sometimes creating TV shows, movies, or documentaries. And then there are companies that die quietly in the night.

But I think underlying all of these moments of frothiness are probably a handful to a lot of great companies building high-quality businesses in a long-term-oriented way. Are they raising money quicker? Are they raising money less dilutively? Are they doing it at big valuations? Absolutely. But I don’t think that is an industry-wide harmful set of conditions.

It’s just: is this company a great company? Can this founder—can these founders—navigate incredibly complex, changing environments almost in real time today to create an enormous amount of value? I think you are, to your point, in the early innings of a deployment of a horizontal new technology that will have tremendous impacts. But like all other technologies, what you see is this initial wave of impact, an initial set of reactions, some version of plateauing, more innovation, plateau, innovation, plateau.

The difference this time is those cycles are two to three months. It’s the new model release. It’s the new revenue growth of a vertical AI business. It’s a huge M&A deal. All these moments create these mini cycles that feel like they have waves across the industry.

Yeah.

2. Seed Rounds Get Bigger

We’ve seen seed rounds of $10 to $20 million very consistently.

Or also $50 million, $100 million, or $1 billion.

Oh, for sure.

But I’m saying, if we don’t take the real outliers, like $10 to $20 million, it’s really been quite normalized as a round size for good seed companies with pedigreed founders. I get that our job is just to find them and beg them to take our money, but does the business not really change in that respect? Suddenly, how do you think about sizing? Bluntly, there’s nothing there, and you’re having to write a large proportion of your fund as an entry check. How do you think about that with the massive explosion of round size?

You can either come on a podcast and complain about it, or you can do the thing. Make a decision: yes or no. If you’re going to play the game that’s in the market of deals, you don’t get to change the competitive dynamics of a free market. So if great founders or potentially great founders are going to raise $10 to $20 million because someone else is going to do the deal at that price, you get to say yes or no.

I think if you think that is an important company and one that you want to be part of, you have to say yes. And there are two conditions that allow for those starting points to not materially change the economics in venture if the outcome sizes are bigger. Think about 10 years ago: the idea of building a $10 billion company was far-fetched. Today, you have to sit and look at an early-stage company and try to squint and get to $10 billion as a sort of normal outcome, versus $1 billion.

So I think there’s a change factor there, and then you see anomalies well past that in terms of valuation, acquisitions, or future IPOs. The second one is the lack of dilution. As an early-stage investor, in the companies that tend to hockey stick very quickly, the Series B, C, and D rounds are dramatically less dilutive to the founders and the early cap table than in the past. They’re raising these later-stage rounds at 3% to 8% dilution instead of 12% to 20%. If you take that across multiple rounds, your early-stage investment just gets diluted less, and that is an economic factor.

Now, we at BoxGroup invest in a wider set of companies, which you’ve criticized many times before for not having an ownership focus. If our entire portfolio is filled with $20 million seeds at $200 million pre-money, do I think the business is fundamentally changed? Yes. In this market of varied entry points for early-stage investing, I think you have to have a portfolio balance, but not on an individual-deal basis.

Totally get that in terms of portfolio balance, blending entry prices and making it a little more palatable. You said there about my criticizing your portfolio.

I mean, speaking about it. I didn’t use the word “criticize” on purpose.

No, I do. Trust me, I think one of the most important things to look for in anyone, personally or professionally, is someone who’s willing to admit they were wrong. I have no problem admitting that I was wrong. What I was wrong about, very specifically, is not seeing outcome-size expansion the way that we have. That’s very rigid-minded of me, which is my fault. But it’s enabled incredible returns from companies with smaller checks and without ownership, so to speak. How would you characterize what I got wrong? Did I do a good job of it there? You can own me for it.

I want, like, a soundtrack as you’re doing that. So if we could add post-production music in—

Oh, yeah, we will. It’s like Eminem—

I think it would go a little wider—

Yeah, yeah.

—on the themes.

It’s kind of like “Lose Yourself.” You know “Lose Yourself”?

Yes.

The first kind of 30 seconds of the beat.

Yeah. Yeah.

Yeah. We’ll do a little—

Just a little background music. Harry getting deep. I think I’ve always believed that venture is a pure power-law-dynamic industry. You have to be in the outliers. The idea that you or I, to pick on both of us, are going to be able to find and buy 10% to 20% of a true outlier in the seed round and beat the rest of the market—I truly think many of them, not all, are great at their jobs—is a lot of steps that you have to believe in to get that business right. You’re going to fight them head-to-head and convince a founder to pick you, and that’s the company that matters.

I’ve never understood trying to compete head-to-head against who I would deem to be the best VCs in the world. I think these are brands and people that have spent decades building what their organization stands for and attracting the best set of founders to want them in their deal. I think those multi-stage firms that have been around for a long time are really, really good at seed investing.

In a way that they weren’t before, huh?

They’re more concentrated and focused on it, more accessible, and more efficient. The market’s more efficient, too. I think founders know how to reach them and how to run a process very differently than in the past. So I think post-COVID there’s a tremendous efficiency in the fundraising market. Competitive deals that used to have three to five term sheets in an early-stage round can these days have 10 to 15. Or it’s purposely only three to five, but those three to five are really competing against each other. That’s something you can complain about or just accept as the market conditions.

For us, the belief has always been that we have to be able to invest in the people we meet who we think are the best, and we can’t let our business get in the way of it. To me, the thing that you got wrong is two pieces. One is the math. Our business math, on our side as a venture manager that has LPs and needs to return our capital to LPs in an outlier way, is that I think you didn’t do the math right. I think a lot of people don’t appreciate that our business makes sense, and that’s okay. We like what we do. We think it’s really interesting.

And I specifically didn’t do the math right because I didn’t see the outcome size expanding. Is that it? You’ve got the numbers.

Well, the second part is, yeah, maybe that’s it, or maybe the lack-of-dilution part I mentioned before. What you get at the beginning actually stays pretty consistent, whereas I think in the past you would probably put a negative 80% on that number.

Well, Jason Lemkin, who I do a lot of shows with every week—we do the weekly show—is like, “100%, 60% you’ll be diluted.” Is that not true in your—? I’m not saying in your—

I don’t think that’s true in the best companies. Not every company goes up and to the right and never runs into bumps. The macro can cause some bumps when you get big. You’re seeing it right now in some of the biggest labs going out and trying to raise rounds and reevaluating the price they’re raising at. If you go back in history, Uber, Facebook, Airbnb, and Ramp, which I think is one of the best modern companies ever, have all had a bump, a down round. So I don’t think it’s always just up and to the right, but I think in this specific three-year window, there is a lot of continued momentum in terms of valuations.

The second piece, which I think is related to the first piece and that I think concentrated funds dismiss, is that there’s adverse selection in the concentrated model of what gets in versus what you want to get in. Every concentrated manager will say, “We win 90% of the deals we want to win.” I think that’s total BS. I think it’s 20%. In a given fund cycle, if you’re funding 30 companies that you have this immense conviction in, you need 10% of the company in order to make the deal work. I think you actually offer that deal to upwards of 120, maybe 150 companies to get your 30. I don’t think people admit that.

So I do believe that the best companies, the ones you’re most excited about, don’t always make it into the portfolio. Just because you have 30 that you own 10% of doesn’t mean those are the 30 you would’ve liked to be in that portfolio. I think in our model, what we’re able to do is get into more of the companies we’re excited about than other people.

I get that. And you know what worries me, though—and I’m asking here—is that I completely agree with you about that supposed win rate. What happens now, though, is they just say, “David, okay, fuck it, I lost to Sequoia and BoxGroup. Let’s do another round. I’ll do it at $800 million. I know David just did the first tranche at $400 million, but I’ll do it at $800 million today.” For the founder, it’s like, “Wow, good deal. Nothing’s changed in the company. I get more money at double the price.” I think it distorts companies and really encourages—

Who are you worried about?

I’m worried about founders losing focus and becoming distracted with way too much money way too quickly.

That is a babysitting problem. Sometimes that’s going to happen, for sure. But to attribute that fear to every founder—some founders are going to tremendously benefit from having more runway, a higher valuation, more momentum, and more buzz in order to recruit people, and to show that there are believers who think this can be huge, so they can justify it to customers and get early partners. If you’re a great founder—and it doesn’t mean all of them are going to do it properly—you have to take all the conditions you’ve chosen to take on. So if you took on two tranches, high valuations, or a lot of money, you’re going to have to use those variables and execute your company-building in a different way than if you’re super capital-constrained.

3. The Lean Startup Was Fake

I think The Lean Startup was a total fake. It was stupid brainwashing that happened for 10 years with professors and books and tweets and blog posts, and it was totally awful. I think it sent people the wrong message.

Why was it totally awful?

Because name one company that actually ran The Lean Startup playbook and worked. One. Every company, if you’re going to be actually successful, needs a lot of money, especially today.

But 37signals is one that’s not so much.

Yeah, some bootstrap companies, but if you want to get on the venture flywheel, lean is not the answer. Inevitably you’re going to be fat. You’re not going to be lean. You have to be. It doesn’t mean undisciplined. The difference between the word “lean” and the word “well-funded” is not a lack of discipline. It’s just not automatically attributable. Does that mean every company that raises a ton of money is going to be well-disciplined? No. There are going to be, again, back to where we started, bad companies.

Will you do tranched rounds if the founder said—

Yeah. I don’t like them, but again, there’s the emotional reaction and the business decision, and those are very decoupled.

But you will do the second or third tranche—

Don’t put this on a—I don’t want founders to know that we’ll just rip checks into tranched rounds, but yeah, let’s do it.

Well, no, but I won’t.

I don’t want to. I’d like to invest at the lower price.

Fuck off. This is a partnership.

I get it, but that’s your problem. Get a shrink.

This is a partnership, and you’re saying you’re inherently less valuable than them. Fine, just don’t take me, then.

Okay. That’s their choice. And guess what? They’re not going to take you, and you can go back and choose. That company is not going to be… That’s a yes-or-no decision. That’s it.

But you’re happy, then, to take that?

I didn’t say happy. That’s an emotional attribute.

But you will take it.

I will be sad and take it. That’s a different thing. You will—

I will be sad and take it.

You will succumb. You’ll just admit that this is sad and fall in line. But I think it’s our responsibility to earn the right to do what we want to do versus what a founder is offering us. You can beg an entrepreneur to do what you’re asking them to do. So if there are 2 tranches, I’d like to enter in the lower tranche. I need to convince the founder to allow me to do that. And if I don’t, it’s a me problem.

I would like the lower tranche, please.

Yeah, yeah.

Cake or death? Cake, please.

Yeah. And then they say no. Then you’re like, “Pretty please,” and they say no again, and you have to beg differently.

Come on, please. I’m just a little—

Yeah.

…seed fund. Come on.

Right. You have to use all your little sales tactics, and then you get rejected. But is that an enjoyable market condition that’s been emerging more and more in the past couple of years? No. If I’m sitting there and I view it as not the founder that tranched us, but the other VC, it’s good to know. I also have a long memory.

Oof. Scathing. Oof.

But it is, right? You’ve decided that I will pay another price, not the founder. You’ve created the structure of this deal such that I’m not sitting on the cap table with you. I’m pushed to a different product than the one you get to invest in. That means when we have early-stage investments and are going out to market to help them raise capital, if I get to be a voice in a founder’s ear, I have opinions, too.

Remember, I was on your side, David. I was—

No, you weren’t.

…I was always on your side. I was on your side. Come on. It was—

But I mean it, right? There are people to me right now—and this happens in hot market conditions—who are playing the short game, and there are people who play the long game.

And sorry, they’re playing a short game because they’re cutting people out, pushing them into the second tranche and stuff?

I just think there’s a set of relationships that you build in this business because we’re all dependent upon each other in very different ways that ebb and flow at different times, in different conditions, and in different companies. If you’re in the business of maintaining and building long-term relationships, you’re going to choose certain behaviors over others in different moments. The thing that we’ve done at BoxGroup that I think is different from a lot of other firms as they’ve built is that we’ve stayed very true to who we are. We do early-stage investing. We try to be there on day one. We don’t put our problems in front of founders. We try to be as helpful as we can and founder-friendly every step of the way. Does that work perfectly every time? No. But that’s the intention.

So we write checks that don’t need ownership. We don’t sit on boards. We don’t write competitive term sheets. We’re happy to lead a deal. We’re happy to help a founder get in business with some money. We’re happy to be totally creative in how we invest, but it’s oriented around early-stage investing and getting in at that seed level.

What’s the average ownership?

I don’t know. I’d have to… That’s a math question.

But generally, is it 5%, 2%?

There are deals where we own 1% at the initial check. There are deals where we own 10%, and everything in between.

But there’s no point that’s too low?

Again, that’s an emotional thing, not a business thing. I’m sad if we have to write a really small check, but it comes down to yes or no. When does that happen most frequently? When we get to a deal late. When you meet a founder and the deal’s in process, you’re begging for scraps, as you should. That’s our fault for getting to something late, and it’s something we can do better as a firm: meet founders earlier in their journey. If we didn’t, we need to figure out internally what we could have done better to meet a founder before there was a company.

Can I ask you, what’s your win rate?

It ebbs and flows with different market conditions, probably between 70% and 90%. I think the single factor that causes us to lose the most is getting to something too late. The speed of our decision-making and accessing the opportunity are the 2 factors we control that hurt us the most. There’s a subset of companies that just meet us and don’t want us, but that’s harder to course-correct because you almost don’t know why they didn’t like you. It’s an emotional outcome, or they didn’t think you were valuable enough, or maybe they referenced you and got something that said you weren’t as good as they needed you to be.

I think we care a lot about those situations, but that happens sometimes. The ones we can work on organizationally are: How do we get to everything before everybody else? I think that’s the obsession. We’re a team of 10 investors. There are 6 partners at the firm. We’ve all worked together for a long time. We all like each other, and we all do the work. Our job every day is to wake up and meet the next founder we’re excited to back.

I think both of us having the flexibility not to have to lead—I admitted my wrongdoing. I turned down great businesses over time because of ownership, and those were terrible decisions. Now we invest very differently. We’ll absolutely do 1% or 2% if we think it’s an unbelievable founder and an amazing opportunity. Great. That means I never lose because I’ll just do a smaller check. Or not never, but you know what I mean. I rarely lose because—

Yeah.

—but if I had to only lead, that would come down dramatically. So my question to you is: Do the pure-play, lead-only seed funds still have a business?

What are the pure-play, lead-only seed funds left?

I mean, you’ve got your Uncorks, your Floodgates, your—I mean, Susa… no. First Round.

I think they’ve moved into a different style of investing: more flexible, and playing a little bit later when necessary, for strategic purposes.

So what you’re saying is that business doesn’t exist anymore.

I don’t want to be in that business.

And why is that?

Because I don’t think you can compete against the greatest firms in the world who are running that strategy as part of what they do.

With much less price sensitivity and much bigger checkbooks.

Yeah. I don’t even think that’s a factor. They’re great at what they do. I like Pedestal very much.

I do think it is a factor. We’ve lost to a GC of the world, with respect, and a junior there. I don’t mean that disparagingly, but someone who’s more junior, because they offer treble the price and treble the check. I agree with the founder taking it. It’s wholly within your right to do it, but that’s why.

Yeah, but it’s an offer in the market, a free market, and I think they take the majority of the air out of the room if you’re trying to compete with, “I’m going to lead, sit on your board—or not sit on your board—but I need 10% to 15% of your seed company.” I just think there’s greater efficiency in accessing the multi-stage firms, and the multi-stage firms are happy to do those deals.

I also think what you’ve seen happen is that firms used to say, “Okay, I’m not going to do this. I want to wait to see this at the A,” or, “I’m not going to do this at the A. I want to wait to see this at the B.” What’s happened is they don’t get to see it. That might have been your only shot to do the deal, because there’s an efficiency in the market and there are other people who also want to do it at the B. You might lose there, and you might lose at the C.

So what you’re seeing is, if Sequoia wins, it might mean that Andreessen or Kleiner or someone else needs to immediately come into the next round because they know that’s the best probability they have of getting a real check into that company. They do it quickly and aggressively. If you lose head-to-head in a seed, in an A, and in a B, you can either put your tail between your legs and call it a day, or try to figure out a new entry point with minimal progress and minimal developments to get your check into that company, if you believe that was an important one.

I think it’s an amazing point, that there was a willingness to wait before that’s now gone. I think that’s absolutely right and very pertinent. Can I ask you, though? I speak to so many LPs. As we said, they love to talk. Every LP wants one category of venture. We want the pure-play seed investor in Silicon Valley.

Small fund, under $100 million. Pure-play seed. And I go, “That is the single worst fucking segment to be in. That’s the one segment you don’t want to be in.” They all look at me rather horrified. What advice would you give an LP when they look at the segments today and go, “Ah, I want pure-play seed in San Francisco as my bucket”?

I’m going to answer your question in a roundabout way.

Sure.

I think there are 3 different stakeholders in our industry, and their incentives and current conditions are very decoupled. We’re in a founder world. I think GPs are struggling to have a very static and focused strategy because we’re in a founder-friendly world. And I think LPs are frustrated because they want certain versions of exposure or portfolio construction that basically aren’t available.

I don’t think there’s an easy solution to the LP problems, and I don’t think there’s a clean way to operate as a GP in an ever-changing market that’s very complicated right now. Founders get the leverage. They get to raise capital from people they like at prices they like, and you as a GP can say yes or no.

On the LP side, they’re going to look at your portfolio construction, your entry point, and your ownership. A lot of the multi-stage funds have multiple products that LPs will tell you and me they don’t like to invest in, but the majority of their allocation into venture is going to go to the established multi-stage firms. They’re a way to get scaled capital into the industry and, frankly, exposure to some of the best companies in the industry.

Your question is about LP conditions. I believe my relationship with our partners is one of truth. We invest out of a $275 million seed fund right now. It will invest in around 150 companies, so we are pure-play seed, with a lack of focus on ownership and concentration. There’s a large cohort of LPs that are confused by our math, don’t think it works, and don’t believe in our model.

You do 50 companies a year?

A little more, probably.

No, but, David, do you remember them all?

Yeah. I don’t know what the other people who do 1 or 2 deals a year do all day.

Podcast.

That also confuses me.

Podcast, too.

Yeah, exactly. It sounds lovely. But there are LPs that don’t believe in our math. Our math works. I know our math works. I see our math, and I think our partners also believe our math works, and that’s awesome, because our model’s terrible if the math doesn’t work.

From a liquid returns basis, what has been the single best investment?

We were fortunate enough to have met Cursor before they were working on the business that scaled. My partner Claire got to know Michael when they were investing at Dorm Room Fund together, and we were in the earliest round of Cursor.

Did that return the fund?

Yes.

Multiple times.

Yeah.

I told you this model doesn’t work, David.

I know.

But you didn’t listen.

I know. I’ve got to do better math. That’s where our Excel sheet broke. I tried to update it one day, and—

Do you know what’s astonishing about that? And I’m really pleased you chose that as an example, actually, which is the speed to that enterprise value. So, in 4 years. I thought every LP was moaning that venture’s taking too long. Now we’re seeing OpenRoute. Was OpenRoute one of yours?

It was not.

Oh, bugger. It was on my mind—

I know. Next time.

Next time. But, you know, this was, whatever, 2 years to $8 billion.

Yeah. I think the speed of company creation, growth, and explosion is different. And I also think the underlying… I’m blessed to have been in the early part of the Cursor story. What they did and accomplished is truly sensational, and it’s a generational outcome. It’s the biggest venture M&A deal in history. We were just lucky to have been there early.

In the fund that Cursor sits in, the early-stage fund, what I love is that there are 10 other fascinating companies that are taking a slightly more reasonable timeline to grow and compound. Do I think any of them will accomplish what Cursor did, in terms of how big and great it was? I don’t know. And that’s not a binary “they did it or didn’t.”

What you have to appreciate in early-stage investing is that things take time. We were the earliest investor in a company called Clay. Clay has had a bumpy 4-year road to explosive growth, and now it’s growing and compounding what it’s doing in an amazing way. But it’s not a clean, perfect story. It takes learning, iteration, and developing their go-to-market to get them to where they are today.

We were in the seed round of Base10. You needed inference to become a space, and Baseten today is a full outlier company. They’re running at full speed and hitting that hockey stick fully. But it wasn’t just start-up and to the right.

Do you think the best companies are generally immediately obvious, or do you think the value is generally in the messy middle, in the ones you don’t expect?

I think there are signs early on that a founder is running the experiment at a world-class level. They’re either hiring total outlier people to join them quickly, finding customers that are inaccessible early, or you’re almost confused as to how they’re getting so far ahead so quickly. A lot of the time, it’s some combination of the product and the go-to-market conversations they’re having, or they picked the right idea.

But I think there are always challenges and bumps. Sometimes those are prolonged, and sometimes they’re immediate. What you’re really identifying is whether this founding team can navigate. Again, I think the market is so complicated today. We spend time with young people. We go to college campuses a lot, and we were up at Harvard and MIT recently, talking to undergrads.

I asked, “Raise your hand if you have a vision for what the world will look like in 10 years.” That’s our job. If you go back in time, young people and old people could say, “I can see what the world will be like in 10 years,” or at least make some guess. No hands went up. I said, “5 years?” One hand. “2 years?” That’s when you started to get some hands.

I think that’s what people are feeling today: uncertainty about the timelines of what the future looks like. In this moment of AI, you have people saying we’re all about to die from AI, and other people saying we’re entering an age of abundance and free-flowing markets that will change the way labor forces work. That’s a pretty wide gap between predictions of the future.

Then you have the models rolling out frequently, with major innovations that are important to a subset of companies. But in the end, not every company is going to need the cutting-edge, bleeding-edge model. You’re watching this age of AI deployment and AI adoption, and it’s going to impact people differently at different speeds.

It’s complicated. You’re seeing AI impact the physical world, science, and all these spaces that felt hard to access before, which are getting opened up because of this horizontal technology shift. There’s just a ton of uncertainty. I think you feel it in the public markets, and you see it with compute access, compute pricing, energy, data centers, and the politics around all of it.

This is complicated. When you’re asked to predict the future, I don’t know that anybody sitting here today actually has an answer. It’s cool that people come on podcasts or tweet or do whatever they do and say, “I know where the world’s going,” and you’re like, “No, you don’t.”

I think what you’re looking for in today’s founders are people capable of navigating, in real time, ever-changing conditions that are insanely fast and insanely complicated. To your question of whether you fund somebody and know it immediately: no. I think the world is changing too quickly today for that to be a condition.

Do you think the founders that succeed today are different from those in the past? If I look at the founders I meet, they’re inherently a lot more technical. There are a lot fewer sales-led founders today than there are—

You’re in a moment of that, right? The question to me is: 2 years out, does sales matter again in some way, and does go-to-market become a differentiator? If that’s the answer, you should fund those people today, because their exposure to the bigger markets isn’t going to be purely relevant for 1 to 4 years, when their company is a bit further along and their product’s ready.

If that is your bet, it actually makes sense to start rotating a little differently today. As an early-stage investor, you can’t solve an immediate opportunity or need. Nothing goes fast enough.

But I do think the founders today are more technical in a non-offensive way, smarter than ever before. I think there’s less opportunity for mediocrity to succeed. You can’t just identify a gap in a business market and go attack it with laissez-faire approaches. You have to be so good in order to win right now.

You saw—I don’t know if you saw—the Scott Wu video that went viral, where he’s doing counting a c- You’re like, “I can’t compete with that.”

I saw a great one, which is, “Cognition is just a wrapper around Scott Wu.”

But you have to look at every company that’s becoming an outlier right now. To me, the founder quality of each of them is really impressive. There was an era when everybody was told they should start a company. Everybody should start a company. I don’t think we’re there anymore.

Can everybody start a business? Sure, and I think that’s okay. I think the words “company” and “business” are different in some way. For probably 10 years, venture capital was funding a lot of categories that weren’t capable of becoming as big as people thought at the beginning, because they were aiming for something a bit smaller. Today, the appetite for scale from the venture product is uncompromising, because you need to get huge to make the math work. I think founders’ ambitions today are huge, and the opportunity set feels unlimited.

I don’t want to single out a company, but it’s one we both know, and everyone will know it, which is a testament to its success. Would you fund a Harry’s today?

I think Jeff and Andy from Harry’s are world-class founders. If you put them inside any company, they would build something important and meaningful. We at BoxGroup have stopped funding brands, not because I think they’re small. I don’t think you can fund brands at Seed. To me, brands are built off your initial data set: Does the brand resonate? Does the product resonate? Do people buy it? Do you have customer LTV or CAC-to-LTV ratios that make sense?

You don’t get any of that data at Seed. I think one of the more interesting investor products out there is brand investing at Series A. I don’t think it fits in traditional venture firms today, but I think there’s a subset of investors that fund brands. The beverage industry is a great example. Look at some of the QSRs, or personal care. There are scaling brands that are fascinating businesses. I just don’t think we’re equipped to invest pre-momentum, given where we sit today.

Can I ask about something I see? Earlier, you said it’s very much a founder’s market. They can choose the price they want and the people they want. To me, it’s a real haves-and-have-nots situation.

Yes.

There are three categories where, if you’re in that category, hell, 50 times you—

There are three categories of haves and have-nots? You only named two.

No, three categories of interesting companies that investors just go nuts for. Obviously, there are more than three, but there’s a small number. If you’re not in one of those, even if you’re in AI for X, investors are like, “Nah, not interesting. Not in compute, not in inference. Nah, not in data centers, not in chips,” whatever. And you’re instantly dismissed. Do you see that too?

Yes. But I come back to founder quality, founder ambition, and execution. We were in the earliest round of a company called Rogo. They were building AI for finance. Gabe finally told this story, but after the Seed round, I helped him try to raise a Series A. I put my name behind this company fully, saying, “Rogo is going after something huge, and Gabe is sensational.” I sent it out to—he said 40 Series A lead investors. He got fully passed on by 39 of them.

Did Keith do that round?

Keith did that round. So, 39 nos. This is now a consensus company. It’s now been funded by Keith, who led the A; Sequoia did it; Kleiner did it.

Do you doubt yourself?

Yeah.

You do? Or do you just think, “God, the mimicry of this market is ridiculous”?

Both. I just get sad. I’m like, “I’ve done this a long time. I’m telling you to spend time here.” Some of the passes were quick: “We don’t think this is interesting,” “We don’t know that this could get big,” or “They don’t have enough traction,” whatever they were.

But I think there are companies in this moment, with the king-making haves and have-nots you mentioned, that are misread at different stages and then become one of the other categories. To me, that means early stage is not a perfect market, as it should be. There isn’t perfect information. There are going to be companies that aren’t in the zeitgeist, that are off the radar, but that compound value and momentum and build something very important outside the immediacy of those funding conditions. I think Rogo was totally misinterpreted by a series of investors at a specific point in time. They kept building, kept getting better, and then, in the right way, were rewarded for their execution.

And you do follow-on checks into just the next round? The next 2 rounds? How—

We operate 2 funds today. They’re both $275 million: a core Seed fund, and then a follow-on fund. We write 20 to 30 lean-in checks into companies, almost all from our portfolio.

And the lean-in checks are $5 million?

$5 million to $30 million.

And that’s across the A and the B, the next couple of rounds?

Yeah.

What’s the most you’ve ever invested in a company across the two?

In aggregate?

Yeah.

Probably $40 million.

Which company was that?

We’ve written substantial checks into Ramp, Clay, Base10, Rogo, and Tandem.

And on reflection, what did you not see?

Sometimes you can’t get more than your pro rata in companies, and that’s a fight for allocation in future rounds. The founder might choose not to take a bigger check from you. That’s one condition. The other is that we misread the company and didn’t lean in hard enough when we could have. We don’t make perfect decisions. I wish we did. I look back, and there are so many opportunities with great companies that we were fortunate enough to get into early, but we were too slow to react to.

Is there one you think of first there, and what did you not see?

I think of a lot, but I think we’re incredibly fortunate to be in business with those founders at the earliest stage. Our job is to continue to help them in the ways they ask us to, and make our customers, who are the founders, happy.

David, is pro rata a right?

No. Do I think a bunch of the early-stage VCs out there saying, “We don’t need pro rata. We’re going to give it a…”? It’s all marketing garbage. It’s just nonsense. I can write a blog post, too. I just don’t believe it’s that black and white. I believe the right to invest in subsequent rounds is earned. Whether you have it in the documents or not, it’s earned. There are multiple factors that go into it.

Having it in the documents is more helpful for the founder than for you as a firm. If you’re asked by the founder to step back for whatever reason, it’s up to you to say yes or no, knowing that if you say no, you’re getting into a fight that you’re likely going to lose in some form. You’re either going to lose financially, in that you’re not going to get your pro rata because of whatever mechanism can be used to remove it, or, more importantly, you’re going to lose on reputation. You’re going to piss off the founder, and they’re not going to like you anymore. If you say no to a founder and want to take that reputational risk, that’s on you.

Why is it important to have it in the documents? I think funds that follow you into rounds in the A and the B are happy to squash you in any way they can. The more you have it in the documents, the more leverage a founder has to negotiate the terms of their next round. If a firm comes in and says, “We want 22%,” and you look at your cap table and say, “I have 6% of my cap table that needs pro rata,” you can create more leverage with your follow-on investor and say, “That math doesn’t work for me. I’d like to redo this.” Then you can use that to help with your negotiation.

Do you see large firms being more ownership-hungry, or more willing to accept less ownership?

Haves and have-nots. Sometimes there are firms that need 20% in a deal, and then you’re like, “But last week you did a deal at 5%.” I see it, because I’m in that deal too. I always think it’s funny when the word “need” comes up in venture. I know what “need” means, and it doesn’t mean “need.” It means “I’d like” or “I hope for,” but it doesn’t mean an actual binary need.

Within the big firms, there are very different kinds of actors. Some people are on the way up or are competitive inside their firm, and they behave one way. Others are very confident in their position at a firm, their standing, and their success, and they behave a different way. You can say “junior” or “senior,” whatever terminology you want to use, but I think relationships matter.

There are times when we send a deal to a multistage firm, they write the term sheet, and then they try to squash us. I’m like, “That’s weird.” Those situations tend to reflect an orientation toward short-term decision-making and a lack of appreciation for the consistency of a relationship. I’ve been doing this a long time. There are people you build great relationships with who tend not to have overly sharp elbows with you. Then sometimes you catch an elbow out of nowhere and you’re like, “Ooh, I didn’t see that one coming.”

Who do you pay most attention to when they send you a deal?

Great founders.

That's who I pay the most attention to. I don't think it's only the best founders; I think it's the founders we believe have high-quality taste bars and high-quality networks. When those people send you someone they know, someone they like, that's when you pay the most attention.

When you look at one part of our business that's investing, another part is also, bluntly, making money for our LPs and selling. We have very liquid secondary markets today. How do you think about that and taking cash off the table when there is so much demand for some of these names?

I work for our LPs, and my customers are founders. I think that's a totally dislocated set of relationships. My job is, as you said, to give the people who believe in us and invest money in us as much money as I possibly can. One way to do that is to try to get out of positions at different times. There's a financial risk to it and a reputational risk to it, and I think that's a really important thing to balance.

In general, we don't sell early. We believe we are investing at the seed stage to join the founder for their journey. That could be 15 years, 20 years, or 4 years. But selling early to me risks two things. One is that if a founder doesn't want you to sell early, you're risking your reputation and your relationship with that founder.

In this market, a lot of founders are taking secondary, so you could argue, “If you're doing it, I can do it.” I don't think that's necessarily the right framework. I really don't, and I can talk about why, but I don't use that framework to think about it. The second thing is that I think the best companies compound the most at the end. Going from a $2 billion to a $4 billion to an $8 billion company is easier than going from zero to a billion.

If you believe you're in a great company, I don't see why you would get out early. I think you're leaving the most money on the table by getting out of something early, and I don't want to do that. I want to see it through. If I still believe in a company, or one of my partners believes in a company, I don't think de-risking it is the right answer. We're in the business of risk, and we should stay investors in the companies that are going to keep compounding, because every turn you have at the end is a big win for your LPs, your firm, and the founders.

Why do you not think that if a founder's taking money off the table, investors should maybe do the same?

Because I think founders have a different set of economic conditions that are not related to what investors have. Let's say a founder has debt, or is living in a place that creates stress and cognitive overhead in their life. In reality, having some money to improve things outside of work allows them to be better at their company.

Knowing that I have $10 million, $20 million, or $30 million in my pocket might make me willing to see this through. I'm willing to have a bigger risk appetite and go for it because I feel I have a safety net. I don't think that's the same risk analysis a VC uses to say, “I need to take 20% of my position off here to make my fund math look good or tell my partners I took some money out of this thing.” I just think they're totally different.

Or giving early employees a tender offer and allowing them to have some early liquidity. I think it could help with retention. I think it could keep a lot of those employees happy, engaged, and excited to work at a company, knowing that they offer tenders every couple of years. We have a bunch of companies in our portfolio that are regularly doing tenders for their employee base, and I think that's an incredibly valuable culture to set and a retention tactic. I just don't believe they're always related.

What did you believe about venture that you now no longer believe, David?

Well, it just took a while to see. Every firm thinks they're in the top decile in an industry—

I've never heard any firm be like, “No, we're bottom, bottom quartile.”

Right. So I'm waiting to meet all the bottom three-quartile firms in the venture industry. When you hear all the LPs explain that, and you look at the data and it says all the returns in venture are aggregated to the top 5% or 10% of firms, I haven't met the others yet.

Do you think we'll have more or fewer VCs in five years' time?

There's less noise in the venture market than ever before, and there's more competition.

What do you mean, less noise and more—

There are fewer new firms.

Really? It feels like there are more spin-outs than ever before.

I think there are fewer. If you actually look at the data, it takes a long time for firms to go away, even if they're still alive. They become irrelevant. They're not investing a lot. They might not have raised a new fund. Nobody's taking them off the map, so the names remain, but the activity goes away.

You had asked before about sub-$100 million concentrated seed funds in San Francisco. I don't think very many people get into venture and want to stay consistent in doing what they do. The starter VC is a seed fund, right? If you're new, if you're unproven, you'll get a $20 million, $40 million, or $80 million fund to start. But your goal is to raise a way bigger fund one day.

When you raise a bigger fund, scale your business, and have partners and scale AUM, your investment style, stage, concentration, and check size are all going to fundamentally change. When I meet new investors, I always ask, “What's Fund 4 look like?” Fund 4 is actually your vision. It's the same as when you meet a founder and ask, “What's the big vision?” I want to know not what Fund 1 is doing today, but where you're going with this.

To me, the advice you give that person should be in line with how they get to their actual ambition, not how they execute their starter fund. I think what you see in new firms today is a big ambition. They want to start with something small and get bigger and do something different really quickly. What's lovely about what we do here is that this is who we are. This is who we've always been. We are an early-stage firm, we write checks at the very early stage, and we don't have any ambition to do anything else.

To be clear, you did just raise, what, $550 million across two funds? I mean, the—

We scale—

I get you. Yeah. The scale has changed.

Sure. But the investment style is exactly the same. I think we see founders today come to us and say, “You're doing the same thing. You're exactly where you were 10 years ago when I started my last company.” Maybe the check size is different, but as a firm, our DNA, our craft, is oriented towards being there at the very beginning, at the dream stage.

That's what I wake up and get excited about. When you look at new venture firms today, their ambition is to be a multi-stage firm for the most part. That's a very competitive part of the market.

It is. I don't actually think they know what their dream is. I know I sound really old, but one thing I would say to young people is, “Be careful what you wish for, because you might just get it.” The truth is, if you know the people at the multi-stage firms, a lot of the people running them are not happy and wish they were at the smaller firms.

Yeah, but I think there are different form factors of the multi-stage firm. You see Benchmark scaled today, USV scaled today, and the DNA and relationships inside those firms are very different from some of the other scaled firms. I think what Conviction has done is amazing at scale. I think what Ribbit has done is amazing at scale.

Then you go to Thrive. Thrive has created a generational venture firm in a different flavor in a rapid amount of time and challenges the very top of venture. The way they've built that firm is so different from your traditional multi-stage firms. I don't think it's as simple as categorizing every firm the same way. You have to look at each firm individually and at how they've executed.

I think what Daniel and Nat had done was incredible, and what Elad has done is incredible. All those firms I just mentioned are new. In the grand scheme of things, I guess Benchmark and USV are not, but the other ones are new. The speed at which they've gotten to where they've gotten is incredible, and the quality of what they've built is incredible. That was ambitious.

When you look at the new firms today, I think the ones that are great are going to execute well, have real ambition, and make a real space for themselves in this industry. The rest of them are going to go away.

When you think about finding, picking, and winning, if I were to ask you to review yourself one through three, with one being best, two being mediocre, and three being worst, how would you review yourself?

I just think we're mid constantly. No, that's what the kids say. Our obsession at our firm is to see things. We all wake up every day and try to see. I think we see somewhere between 50% and 75% of what we want to see. That's my guess.

It’s hard to know, but we can always be better there. There’s always something out there that we didn’t see that we want to see. Getting in front of people—and finding a way not just to get in front of them, but to have them take you seriously—is what we wake up and obsess over.

When you don’t win, it hurts. It makes you ask, “What could we have done better?” That’s the area in venture where I think you can tactically work to improve. You can try different things and figure out what you can do better, and I think we spend a lot of time on that.

What has that led to? You’ve spent a lot of time on that. What have you changed as a result?

Everything. You have to know what tactics to bring to the fight. That’s reading a room and guessing which relationship-building playbook you want to use. What does a founder want from you? What are they looking for, and how do you convince them that you’re the answer to what they’re looking for?

It really is so founder-dependent. There are certain founders who don’t want to talk a lot. There are certain founders who want transactional help. There are certain founders who want a friend, a set of experiences, or somebody who’s around all the time to bounce things off. It’s a services industry. The win condition, to me, is selling your service to the point where a founder wants it, and that’s complicated, especially in this market.

The picking is the art of this business, and it’s so hard. You can look at our firm from afar and assume we say yes more than other firms. We don’t. We’re hyper-focused on saying yes when we should, but that’s a really, really hard thing to do at the early stage in real time.

I want to tell you I know how to get better at that, or that we can get better at that. And yet there comes a point in your career when you’ve seen enough reps that your taste is your taste. Your butterflies show up, you’re going to say yes. They don’t, you’re going to say no. We think a lot about how to make the right decision and don’t know the tactics to get better at it.

David, how many companies do you meet a week as a firm?

A lot. Fifty to 100.

So people are meeting 20 companies a week.

Yeah, at least.

And then every Monday you come for an IC and you talk about the five best?

Tuesdays and Thursdays.

Twice a week. Okay, different.

We use Mondays for work. We’re agile with our decision-making. I think groupthink is terrible. We don’t vote. Somebody here has to want to do a deal if we’re going to do it. Somebody needs to say, “This is something that I’m excited about, and I think we should do it.”

Decisions start with a human on our side being excited about something. Whether that happens inside or outside our meeting isn’t important. If we get to our meeting and one of my partners is like, “I did a deal yesterday,” great. Equally, if I’m leaning into something and want to talk it through, great. We don’t allow negativity in making early-stage decisions.

What do you mean by that?

If you aren’t the person who has spent time with the founders, you aren’t allowed to be negative about it.

But what if you know something about the market that you feel—

I don’t think anybody knows anything about markets at the early stage. Markets evolve. If you know something about the founders, we’ve probably diligenced that before we’re taking meetings and leaning into something. If there are ethical issues or reputational issues with a founder, I think that’s important to share.

We don’t allow grenades at the early stage. You’re not allowed to throw a grenade on someone else’s deal. Our job is to get to yes. If you name a company at the early stage, I can tell you why it’s going to fail. Any company. Here are the five risk conditions, and here’s why it’s going to fail. It’s the easiest thing to do. If you’ve done this for a year, you can guess why a company’s going to fail.

The memos every firm writes make it very obvious why something isn’t going to work. That’s not the job. And, by the way, on a probability basis, it’s not going to work, because 70%, 90%—whatever the number is—of early-stage companies don’t work. So explaining why something isn’t going to work is incredibly obvious and incredibly irrelevant.

The magic of our business is squinting and seeing why something is going to work. It’s believing in the thing that you see, that the founder sees, that you’re aligned with the founder in seeing is going to work, and saying yes. Our job is to say yes to the right things, not to be fearful of saying yes to the ones that don’t work.

Totally get that and understand that. I always love that Google will just build it, or OpenAI are—

Whatever it is.

Yeah.

But you can’t hire. You don’t know how to hire. You’re not going to get—It’s not big enough. All these things, they’re just not factors for the win. Why is it going to work?

You said that 70% to 90% don’t work. Keith Rabois said before that’s dramatically overestimated, and in fact, if you’re investing in great people, your loss rate—in other words, it actually really not working, being a zero or close to zero—is much lower than people think.

I think at Series A that’s true. I think at seed it’s a real scale of numbers.

You do?

But I think you go one stage ahead and that number is dramatically lower. I also think Keith is a generationally talented investor, and if his loss rate is, I don’t know, a tenth of everybody else’s, that wouldn’t surprise me.

Yeah. Listen, David, I want to do a quick fire. I’ll say a short statement, and you give me your immediate thoughts—

How short?

Because that’s not big. Sixty seconds per one.

Let’s go.

I’m going to set a timer.

All right.

You can choose one seed fund, one Series A, and one growth fund to be an LP in. Which one?

I think Y Combinator has invented the only product in our industry. For all the people who say anything critical about them, it’s probably just rooted in some version of jealousy. I think what they’ve built is a world-class brand, a world-class attractor of great talent, and the best business model in our whole industry. So, seed fund: YC.

Fantastic. Completely agree with you. Series A.

I think there are five to seven equally impressive multistage firms that just continue to find themselves in the best companies. You can choose Benchmark, you can choose Sequoia. I think the machine Andreessen’s built, and what Thrive has done, are impressive. What Lee Fixel does at Addition is fascinating and off the radar in an interesting way, and sometimes it’s a great Series A investor.

I think what Ribbit has built over the years, if you look at what they’ve done, is incredible. And then you have Founders Fund, which operates in its own world and is incredible. Keith Ekosfa has done generational-level investing. I actually don’t think it’s easy to isolate one at Series A—

Oh.

—because you’re looking at this combination, if you put your LP hat on, of fund size, accessing early stage versus growth, and these blended products. I don’t think it’s a black-and-white answer. The two funds that were very narrowly focused on Series A over a historical period were USV and Benchmark.

Mm.

I think they’ve changed their fund size today, but not really the partnership dynamic.

Okay, fine. Yeah, absurd punting there.

Yeah.

Growth—

I weaved in and out of your forced question.

Growth fund. You’ve got one growth fund.

I think Spark Growth comes to mind as an amazing, slightly off-the-radar, unique outlier in decision-making. I think what Meritech did for a while, and what Menlo’s done. These are new names I’m weaving into my answer. I think what Coatue’s doing or Dragoneer’s doing right now is interesting.

You’re a fucking—you’re—

Yeah.

—you’re fucking Switzerland.

I am Switzerland.

Yeah, why? Show us your Toblerone, David. Come on.

But I mean it. There’s a tier. Greenoaks, right? I could easily just say Greenoaks as the answer to your question.

No, because I’m asking you. I want money. Dollar for dollar, where am I going to make the most money?

I think it changes in every market—

No, but I can give you mine. YC, I agree with you on ownership. No one gets the ownership they get from the founders.

But I think YC is the only differentiated product in venture.

And then I think Benchmark for Series A, and then I think Sequoia for Series B and beyond.

So you’re discounting Sequoia’s early-stage investing?

I’m suggesting I would rather be in YC’s than Sequoia’s seed funds, yes.

I think the real answer to your question is that you’re in a period of consolidation in venture, and the best firms are going to continue to compound in every way because they’re going to access the best founders and the best scaled LP capital, and it’s going to turn into the best set of returns. There are going to be haves and have-nots in venture. I think there are probably 20 to 30 names that are competitive, and probably 10 outliers among those 20 to 30.

What was the single most memorable founder meeting in the last 12 months? First founder meeting.

My partner Greg got to spend time with Noah from Instinct early, and I think, for him, that was the mind-blowing initial meeting and conversation. We were lucky enough to be part of that early round.

Woo, baby. Well done.

I think the things we’ve done in the past 12 months aren’t mature enough to talk about yet because they’re coming.

But which founder did you find most mind-blowing, where you left thinking, “Wow”?

I have to be as excited every time I say yes. I don’t operate like that. I operate like this: If something rises above this, you lean in, and if something doesn’t, you lean out. I view my job as a continuum of steadiness in order to make the right decisions.

What satisfies me the most, to this point, is the relationship that comes out of it. We funded this 22-year-old who’s building a single game that’s going to release in 2 months. Level of difficulty, one trillion, right? To win a game business. He’s been thinking about this game for a long, long time. We’re going to see if it hits, and if it does, it’s going to be awesome. But the probability of the market is low. All these factors are negative, but he’s awesome.

Then I met this woman who’s building a really cool piece of hardware that goes in your home. It’s a point solution in a huge market. Everybody has a need for this. But again, the level of difficulty is impossible. Then you meet veteran founders who are building a foundation model for a physical-world application, and you’re like, “Well, that’s crazy, and you guys seem like totally the right people to do this.” Then you have someone in science who’s doing something.

I really do believe you have to be excited about everybody in some form equally. Who do I like hanging out with? I have a consumer mindset. I find consumer founders the most interesting because they’re nuts. They’re going after the biggest, most confusing end user, and I learn from them.

We’re an early investor in a company called Triumph. Spending time with Jake is amazing. I love talking to him. I love hearing about what they’re doing. I think they’ll start making a lot of noise in the coming year and sort of be one of those haves.

David, what are you most excited about in the next 10 years?

I think seeing this play out. Where are we going? Where is the world going? There’s such curious uncertainty right now. That’s what makes this job a special opportunity: You get to meet the people who are going to answer that question and be part of their journey. I don’t know.

What is certain in the next 5 to 10 years?

What should be? Autonomous driving. It irritates me that it’s not here today. It would save lives. The data on autonomous driving says it would save hundreds of thousands of lives worldwide if we just mandated it. But freedom and all these other factors go into holding that back. The technology is a solved thing.

I’m not being mean. The technology isn’t quite a solved thing for ambiguous cities or challenging weather conditions.

It’s pretty solved.

From a safety-profile perspective, it’s pretty solved. It’s not perfect, but it’s pretty solved. If I told you “pretty solved” equals an 80% decrease in the deaths that automobiles cause, that’s pretty solved. I don’t like that, as a society, we don’t adopt future things because we hold on to the wrong establishments.

It’s because incentive design is skewed in a lot of frameworks.

I think modernizing society long-term is the right way to think.

I agree with you, and I think it’s kind of like the roll-up play versus an entirely new-company play. I think it’s much easier to create an incredible AI-first school from nothing than to try and incentivize innovation in stale architectures.

Or we should do both.

You can try both, but it’s much easier to do the first.

Yeah, but just move forward, right? I think the politics, the social dynamics, and the economic dynamics of all the challenges in the world would be best served by practical decision-making rather than charged decision-making. I hope that’s where we find ourselves over time.

If you end up in the age of abundance, I think it’s important to figure out how that benefits everybody and not just a few. I think technology generally is an equalizer and a positive force. Hopefully AI can move in ways that impact health, science, and every industry, such that it benefits a global population in great ways.

I think this was probably the most low-friction conversation we’ve had. There wasn’t a dramatic dip.

Should we fight at the end?

I don’t know. I feel like we should at least try.

We fought over you deciding Benchmark would be the only good VC firm.

It was the most nerdy venture fight. We fought about ownership versus diversification in portfolios. You were like, “All right, guys, calm down.” Dude, I loved chatting with you. Regardless of me being wrong, I have so much respect for the business that you’ve built.

Thanks for having me. I like coming on here. But before we leave you today, what do Uber, Cursor, and Harvey have in common? Well, they made the really wise decision to build on Fireworks. Fireworks is the specialized intelligence platform behind many of the world’s leading AI products. Companies use Fireworks to deploy the latest open models, specialize them with their own data, and run them in production with the speed and reliability modern AI applications demand. But getting models into production is only the beginning. As AI usage scales, the best model for one request isn’t always the best model for the next. Well, that’s where Fireworks Nexus comes in. Nexus connects to the AI coding tools and harnesses your engineers already use and intelligently routes each request to the best model for the job, helping you optimize for quality, latency, and cost without changing how your engineers work. To see intelligent model routing in practice, visit fireworks.ai/20vc to create a free account, and use the promo code 20VC50 to claim $50 in credits for 20VC50. While Fireworks powers AI, Laradin helps you measure its impact. Every company now runs two workforces, humans and AI agents. Laradin is the first unified dashboard across both workforces that measures their work and optimizes their impact. Laradin answers four questions: Who’s using AI? Which tools? What is it costing? And what’s it actually being produced? For every team across every vendor. AI ROI stops being a guess and becomes a number that you can consistently and confidently measure. Gainsight, Vertiv, SurveyMonkey, and many other enterprise organizations run their AI programs on Laradin. Thousands of AI tools and millions of dollars of AI spend are measured. You bet on AI, Laradin proves that it’s working. Visit laradin.com to learn more. That’s L-A-R-A-D-I-N .com. While Laradin tracks AI impact, Zero Hash powers your crypto infrastructure. Every great software company eventually runs into money. Uber had to move it. Shopify had to hold it. Airbnb had to settle it. Money movement stopped being a fintech problem. It became a software problem. And increasingly, God, it’s an AI problem, too. Zero Hash is the infrastructure that makes global instant money movement seamless. One API integration for stablecoins, digital assets, and modern payment rails, so builders can stay builders. Zero Hash powers some of the world’s largest enterprises and financial institutions, including Calshi, Stripe, Morgan Stanley, Gusto, and Interactive Brokers. If you’re thinking about stablecoins, digital assets, and the future of money, it’s time to talk to the team at Zero Hash. Visit zerohash.com/20vc to learn more.