General Catalyst CEO Hemant Taneja：将 GC 管理资产规模扩至400亿美元的经验

Our aspiration in venture capital is to be the best seed firm in the world. I actually have a strong belief that venture capital can't scale and performance at the same time. I deeply believe that just because we have more money doesn't mean there are more Patrick Collisons or Sam Altmans who are going to go build iconic companies.

I lost the Series A of Stripe, of Samsara, and of Snap. The first one I won was a Series A of Gusto. Triple, triple, double, double is definitely dead. I tell our investors, “Don't bring that to me.” Going from 1 to 3 to 9 to 27 is not interesting. You've got to go like 1 to 15 to 20 to 100.

Are you more bullish on the future of the United States with a Trump administration or not?

Ready to go. It is so good to have you here. Last time was 7 years ago. It wasn't in person. I've been so looking forward to this, dude.

It has been 7 years. Last time, I was a little younger and you were a little skinnier. You've gotten fit, and you have no glasses. I have my glasses right here.

1. Is Hemant a CEO or an Investor?

Dude, I was much skinnier. I think this was before I fell into a project. You've done well. That's good. My question to you is: You have built, in the last 10 years, one of the most defining firms that we have in venture. Do you consider yourself a venture capitalist? Would you consider yourself a CEO?

Harry, that's a great question. I carry the title of CEO and managing director for a very intentional reason, which is that General Catalyst is a business. But it wouldn't be a business if it wasn't venture capital at its core.

So I am a managing director and a partner, just like everybody else in our partnership, but I'm also the CEO. That's the duality that it's going to take to build an iconic institution in our industry.

Do you think GC is still a VC firm?

GC is very much, at the core, a VC firm. Not only that, our aspiration is that we want to be one of the best seed firms, like you. That truly is our aspiration, because the earliest relationship with founders and that trust is the key to actually doing the best work in building the companies that matter.

When you look at total AUM, can you realistically put the hours in and justify that commitment to seed when it's a $200 million vehicle in a $25 billion pool?

Culturally, this gets hard for VC firms as they scale. At GC, the thing we talk about is focusing on the ownership and the relationship with the company versus the sizes of the checks that you put in. When you reorient yourself to think that way, we get that only at seed.

If you think about the last 2 years—bringing on Jeannette and La Famiglia, Yuri and Wayfinder, Norrsken, and Venture Highway—we've tried to really make sure that, at our core, we remain very committed to doing the seed work with the same intensity and rigor that you do at 20VC.

Oh, zero rigor here. No, we're just blindfolded, throwing darts.

I didn't say a lot of the same rigor.

Oh, okay. Fantastic. I was feeling bad.

It's a genuine comment. Internally, we talk about early-stage venture capital as a core. Obviously, we want to leverage that core to have greater impact in the world, but if we don't do early-stage investing well, we will lose the right to exist, and we're paranoid about that.

Do you worry about the transition of venture? Doug Leone said that we've moved from a high-margin boutique community to a low-margin, commoditized industry. Do you agree with that?

Yes. I think—let me unpack that in a couple of ways. One is, if you think about the innovation in venture as the role of technology has scaled, all the innovation, for the most part, ends up being on 3 axes: stage, sector, and geography. Make the funds bigger, put them in different geographies, and put them in different sectors.

The reality is that the role of the companies that we're building is becoming far more sophisticated in society. The innovation in the industry was much more focused on how we deploy more dollars and try to keep as much of the return as possible, where the reality should be: How do we retool our proposition for founders so they can build the biggest companies possible?

When you think with that second lens, you have to innovate, and you have to think broader than just that fund-formation mindset. I think that is what allows you to break from, “Hey, going from high-margin boutique”—smaller funds, better returns—“to low-margin scale”—bigger funds, lower returns.

That's only happening because we're thinking about innovation in a constrained way in this industry versus being first-principles about how we transform our proposition for founders.

So many things to say there. You said bigger funds, lower returns. Do you disagree with that as a premise, then?

No. I actually have a strong belief that venture capital can't scale and performance at the same time. I deeply believe that. The reason is because just because we have more money doesn't mean there are more Patrick Collisons or, pick your favorite founder, Sam Altmans who are going to go build iconic companies.

So we're actually, in some ways, finding ourselves in the zero-sum game of founders who are naturally going and doing great things. That's not necessarily going to scale because we have more money.

But if we can create more tools to help more founders scale, then we can actually manufacture more outliers than the ones that naturally exist on the power law. That's our mindset: How do we expand the proposition to founders so that there can be more companies on the power law?

That is a very different way to think about it than asking whether we have enough capital to get everything that's on the power law.

2. Explaining Returns to Different LP Classes

If you accept lower performance with bigger funds, respectfully, Hemant, what do you tell LPs when you go out and fundraise for early-stage venture funds and much larger funds? Is it just a different LP class? Because I'm sure you hear the podcasts and the shows, and it's like, “Oh, they're just pitching sovereigns who are happy with 10%, and so it's graduating.” Is that true?

Going back to saying that we want to remain early-stage venture at our core, I actually reject being in a business that has lower performance. What we have done is, if you look at our overall assets under management, we've basically said we're not going to make our venture funds bigger.

What we're going to do is keep the size of the venture fund where we think it can be to create elite performance. To us, you've got to at least deliver 4 to 5x funds on the capital that you raise, and build bottom-up. Can you do that in venture?

Then we have the Creation Fund and Customer Value Fund, which are focused on other value propositions and other capital solutions for founders, so they can do M&A more effectively and invest in sales and marketing more effectively. But don't scale the venture fund itself, because that will degrade performance.

That is the way we have architected the capital that we provide to founders.

When you think about where you intersect with them in the journey, as much as I love the idea that we absolutely want to be focused on seed and build that relationship as early as possible, when you have the capital supply that you have today, arguably it's a much better proposition to just do a Kleiner and put $100 million into Anthropic at a $60 billion valuation, play the large check at late stage, and actually ride that wave.

I don't think so.

You don't think so?

Yeah, I mean, look, our best returns have come from seeding companies like Stripe and a company likely called Anduril, or creating companies like Kayak, Livongo, Commure, and others.

Livongo is insane, huh?

It was a great outcome for us, and that's something we built in our offices, really.

How much did GC make from Livongo?

A few billion.

How big was the fund?

Livongo sat in 2 funds. It returned one of the funds approximately 3 or 4x, and it returned one of the other funds maybe close to 1x. I have to go back and look at the numbers, but I think that's generally right.

Is that the best-performing investment GC has made?

I think you would have to give that to Stripe, still. We've been invested in Stripe since 2010. That's a big position for us.

My point is more that we obsess over either the companies getting incubated at GC or investing in the seed round. If we don't, we want to be in the iconic companies, and we will invest in them at growth stage as well.

To me, that's about believing in the companies that you think will compound for a long time. Take Stripe as an example. We seeded it in 2010, and I've invested in Stripe 14 times in the last 15 years, just to give you a sense.

That's one of our core philosophies: When you think something's going to be compounding for a long time, be strong supporters of the company along the way.

We invested in Helsing. Jeannette seeded that, if you remember, before, and we've invested in it in all the rounds that they've raised since then. If you look at Anduril, same thing. We seeded that, and we invested in every round that they've raised since then.

I think being part of these iconic companies and supporting them along the way is the reason to have the capital base. If we miss them at seed, we want to catch them as early as possible and then continue to help them with everything we've got—our entire proposition as a platform—to support these founders and help them all the way to endurance.

3. Mapping Markets at Scale

Can I ask, when you're at your scale, do you map out the industries that matter, the companies in them, and go, “We have to have a check in these regardless of entry”? Is that how you map markets and capital injection?

The business for me is about getting serendipity and intentionality right. You definitely won't know the industries that will become important in the future. I'll never forget this—one of my big messes was when Paul Graham asked me to look at the seed round of Coinbase, and I said to myself, “A Bitcoin ATM? What is that?” I had no idea what this industry was about to become. It still haunts me.

At seed, it's about being very much focused on just backing the great founders and not over-extrapolating what the returns are.

Did you meet with Brian?

Yeah.

Did you think he was amazing?

He is amazing. My little brain got ahead of itself in thinking about the world.

My point was, at the seed stage, you definitely don't want to be industry-focused. You want to say, “We're going to back founders regardless of our view of the world.” Then I think the reality is, if you take a step back and think about the tectonic shifts that are happening in the world, the theme that we call global resilience—where every region is focusing on defense, energy, industrials, health, and financial services, and how to be resilient from a sovereign perspective—does have an impact on industries. It has an impact on how businesses are going to get built.

As an example, I'll tell you: I think we're the only firm that's invested in a defense prime in the US, Europe, and India. We did, and we invested in Helsing, and we invested in a company called Raphe in India. Each of these regions needs to create its own AI deterrence solutions, and they want to see indigenous industries emerge from a resilience perspective. So we should make sure we're backing that theme.

I think it ends up being: embrace serendipity, be humble that these founders are going to take us into the world in a way that we just don't understand, and then be intentional where we think there are large macro shifts happening, so that we can play certain sectors with a bit more of a thematic lens.

What do you think is the most significant macro shift today that not enough people are talking about?

The most significant macro shift today that not enough people are talking about is thinking about jobs. I've gone around the world. We have a real focus on understanding how to help governments think about transformations, and the transformation of any country, we think, is in 4 parts.

One is, how do you apply AI to deterrence? Without peace, you don't have capitalism, and if there's no capitalism, then business can't really be a change vector. So you need peace. You have to think about transforming healthcare because we just came out of a pandemic, and we saw what it can do, and we're still reeling from it. You need to accelerate the diffusion of AI into business because that's ultimately what's going to lead to your industries being competitive. Then, if you get all that right, you have to think about jobs because there is immense reskilling that needs to happen.

People are starting to give lip service to it, but it hasn't hit people yet. We were talking about this earlier. We have a lot of these AI transformations we're doing of these service businesses, these AI roll-ups as they're called. I'm seeing what's going to happen to jobs as AI gets adopted to bring efficiency and productivity to these white-collar jobs all around the world.

4. No One is Talking About the True Impact of AI on Jobs

What are you seeing there? Because the MIT study was discouraging. It's like 95% actually doesn't have much impact. I read it and I'm like, “God, this is a bit of a downer.” So what are you seeing?

I do think there is merit to the MIT study for the following reason. When you think about transforming an enterprise with AI, you actually have to do 4 things correctly. First is, you have to get your data infrastructure ready so that your company can adopt AI. Your data readiness is huge, and infrastructure readiness is huge.

Second is, you need models that understand your business. You have to train these models in the context of your secret sauce, your business. Third, you have to think about a workforce transformation because now you're going to have humans and AI working next to each other. Some humans are going to manage AI agents. Some AI agents are going to manage humans. Imagine how the org charts have to change.

The fourth is that, for all this to work, you actually need courage at the top. The CEOs need to really get behind it to drive it. So the idea that all 4 of these things are happening in a company to make the adoption of AI go from beyond just prototyping an OpenAI or an Anthropic model to really creating change in your business is very, very difficult.

That's why these things are hitting a wall. That's why this MIT study is giving you the signal that it's giving you.

Now, one place where businesses already outsourced and let go of core operations was wherever they wanted to get labor arbitrage over the last 40 years. So our whole thesis around AI roll-ups was: everywhere you offshored for labor benefit, you're going to onshore for AI productivity. That's where we're seeing a lot of this.

We've bought call centers. I'll give you an example. We bought a call center in the Philippines, 3,000 employees, in one of our companies called Crescendo. I think it's going to have a huge change in the headcount. It's going to go down by quite a bit as this gets fully AI-enabled. My first question to the team was, “Well, what are those people going to do in the Philippines, and how many are there?”

Every country that built its middle class off of offshore labor—how do we really help them think about reskilling those people to be more successful in the world of AI? This is what's not being talked about enough.

Do you think this is a 12- or 18-month problem, or actually a 5- to 10-year problem? I always go back to Bill Gates: we overestimate a year and underestimate 10.

I get this question a lot. Is this adoption of AI into businesses going to be fast over 12 to 18 months, or 5 to 10 years? I would say this is a 5-year problem. I say that because, if you go back and think about the physics with which these companies are getting built, the companies we're building have to put these teams together. They have to go get some customers they can demonstrate progress on. Then they have to start accelerating growth.

Only after a few years of that do you start to make a dent in the industry enough that it becomes a problem. The diffusion of technology has its own physics. So it's not immediate, but 5 years is also not a long time.

What I'm seeing is that enough of these companies are going to start being successful in these different areas, and they're going to start impacting jobs in a material way.

I'll give you a really interesting anecdote that the CEO of one of the large consulting companies told me. He said one of their big clients came to them and said, “We have 50,000 employees today. Draw up a plan for us that, in 5 years, we are 100,000 employees, but only 10,000 of them are humans. The rest are AI agents.”

This is to be provocative, but they're saying, “If that was going to be our plan, how do we get there?” This is the kind of stuff people are thinking about now. It's not going to happen in the next 5 years, but are organizations going to potentially change that much over the next 10 to 15 years? It's a nontrivial probability that can happen.

That's a very forward-thinking CEO and business leader. To what extent are governments prepared, thinking, and equipped for this labor change that could be there within a 5-year period?

I think not enough. I don't think people are grappling enough with what AI really means or how fast it's going to diffuse. They're not even thinking enough about reskilling.

I'll leave you with one interesting thought on this particular topic. Imagine we're in London today. Imagine if every nurse, every lawyer, and every accountant that works in London becomes an AI agent of some company in the United States in the next 10 years. You're going to hollow out a lot of your labor productivity and give it to a US company or a Chinese company.

My point is that it could actually hollow out the service sector, just like we hollowed out manufacturing jobs through globalization before. So every region needs to think about this. This is actually a point that Jeannette makes with her European Champions Initiative a lot, which is: how do you retain productivity onshore in these countries so that, while you do the AI transformation, you're maintaining vibrancy—not only because your businesses get more profitable, but also because you're capturing the productivity gains onshore as well?

The governments need to think about this as they are architecting this next phase of their transformation with AI.

Which government do you think is most impressive, and which is most screwed?

I find folks in Singapore to be very impressive. I recently went there, and I spoke at their National Day, and I was just blown away by the depth of thought the politicians there have put into it. I have to tell you, the prime minister of Greece is very impressive.

He's thinking about how we can be pragmatic in deploying this. I've met with Prime Minister Starmer here. I think I know there are some announcements being made this week as well around AI. So I know they're making some moves, but I don't see enough of the, “Hey, let's think comprehensively about this.”

5. Investing in Capped Upside Companies

The answer I usually get when I talk to heads of state about diffusion of AI and this jobs issue that we just talked about is that they believe if it's going to be that disruptive, society will just slow it down. Because you just can't have a world where—I know in Silicon Valley we've created a billion-dollar company with a single employee—but you just can't have a world where that's what business looks like and people have no work.

At some point, the interplay of business and society will sort of force it to be a more stable scenario. That's what the governments are sort of thinking.

I think that's true. I think Adam Smith's invisible hand would tell you otherwise.

I think market forces are way stronger.

I agree with you, but I'm saying that's a bit of what they take comfort in: that we have time to figure this out. I don't think we have time unless we're a lot more intentional about it.

I vehemently dislike government intervention. I'm as free-market maximalist as possible. Is that the same for you?

I think you can't make progress if capitalism is not working. But I do think capitalism is a privilege. If you think about what happened in the last 15 to 20 years, a lot of the nationalism all around the world is because social media essentially struck a chord with the core issue that the productivity gains from all the technology didn't really get passed on properly to everybody in society.

It kind of created multiple multitrillion-dollar companies, but our own innovation ecosystem—how much did that capture, and how much does society really capture? Actually, a small percentage. We look back and say, “Wow, venture capital, there's been a real boon in the last 15 years.” But when you look at the overall value created in venture compared to the Magnificent 7, it's noise.

Are you really creating a world where there's opportunity and capitalism can kind of do its thing? You have to make sure you protect that. That part of it, I do think government has to play a role in. Beyond that, you've got to be very free-market-oriented. Let bottom-up innovation—stuff that you and I do—go create the future. I mean, that's what you want to see.

6. Is Hemant Concerned by the Concentration of Value in MAG 7?

Do you worry about the ever-increasing inequality of wealth? It plays in our favor in a lot of ways, but I look at so much of the next 10 years and I just see the concentration of wealth in very small networks, and I get very worried.

I am worried about that. The whole idea of whether we can build these companies to focus on being the most profitable and the biggest, but also in a way that they're inclusive, is something that I think a lot about.

I do think we have this opportunity. There's this moment—if you think about the last 5 years, we've had wars, we had a pandemic, and we had a situation where, because of wars, the US kicked Russia out of SWIFT, so financial infrastructure got questioned. Every part of our energy crisis happened. Every part of our core pillars of society, where capitalism may be starting to break, all sort of manifested over the last few years.

Then AI comes along as an answer to all this. So now the choice we have to make is: are we going to build these businesses in a way that the value accrues to very, very few, or can we actually do it with a mindset of abundance where everybody benefits? That's a choice that we have in the way we set up the companies of the future.

I do worry that if it's not a mindset of abundance, then that's not sustainable in the very long term. We won't feel that in the next 10 years, and you and I will make a lot of money. Our funds will do great, and our teams and our partners will generate great returns. But what do we create on the other side? I think that's the thing that we have to think hard about.

Is it actually a choice, though? If we think about, say, OpenAI and Anthropic—you know you're an investor in Anthropic—when we think about the value that they generate, those returns will go to a very small handful of people. Is it a choice that we're concentrating the returns and wealth?

Look, I think about innovation in AI in 2 parts. One is that every region is trying to figure out how to become a leader in core infrastructure, and we're racing to it. There aren't going to be many. There weren't that many cloud companies, and there aren't going to be that many AI model companies that actually get to scale and potentially become even bigger than what these cloud companies became.

I think that's the current course and speed. That's probably what's going to happen. But what happens on top? What's the ecosystem we're building, and how does it interplay with consumers across the board? What happens to healthcare? What happens to education? What happens to opportunity itself? How are we thinking about those things?

Is there a level playing field so there can be a vibrant, diverse ecosystem that gets built on top? That's what I think a lot about. Or is it going to be—for example, is the Amazon of healthcare going to be an ecosystem with lots of companies and a more resilient system, or is it going to be some company that comes along and just controls healthcare?

The latter is not good for us. So how do we create a level playing field for startups and founders, so that opportunity can manifest into new, successful businesses everywhere, versus there being a few concentrated ones? To me, that's the place where policy has to create conditions that allow for opportunity for many, as opposed to opportunity for a few. That's the role governments can play when it comes to technology.

7. Has Trump Done More to Hurt or Help the US?

I've been very public on my concerns around the Labour government in the UK and what it's done for the UK so far. It's the fastest exodus of millionaires out of any country. It's terrifying. Are you more bullish on the future of the States with a Trump administration, or not?

Look, my belief is that the US is very well positioned. We have energy, we have AI, we have the largest market, and we have the largest entrepreneurial ecosystem. In a lot of ways, we're very well positioned, and I think in the short term, in some ways, we're actually increasing our moats if we really focus on everyone investing in the US and creating more capital and whatnot.

To me, I look at the mandate of the companies that we fund. When we fund companies, whether it's in Europe or in the US, I always think about it as, “Hey, you need to go win your market, and then you need to become a global leader.”

The thing I worry about in the US today is mostly the sentiment and appetite of the world to embrace companies coming out of the US and let them be global leaders. I think that's where there's going to be work to do, because we're doing this one-time reshifting with tariffs and everything: “Hey, we need to realign commerce and trade.”

But we were also the keeper of the world order in a lot of ways. As we are disrupting that, what is the relationship going to be with European countries, and how well-positioned will American companies be to be global leaders? That's the place where I think rubber meets the road in our ecosystem.

I'm talking about companies funded in our world. Founders: how will they become global leaders, given that there'll be more friction? That, to me, is the place where there'll be some challenges.

When you walk around London now, as I do with my mother every weekend, all we see is Anduril posters on the sides of buses and on the sides of bus stops. I always send it to Matt Grimm like, “Thank you for decorating London with American posters,” which actually look incredibly British and wonderful.

By the way, I love your post with your mom. I think that's a great thing that you do every weekend, and I enjoyed your last post about breaking the idea that 90% of the time you spend with your parents is before 18. That was an amazing, insightful comment.

I find it incredible that people just accept that. It's like, you could change that. I really appreciate that.

But when we go to that, when people become global leaders, the element that's inserted is competition. When you back multiple geographic leaders, suddenly Anduril comes into Helsing's path. Is the age of not having competitive investments over? When you're at your scale, you just have to embrace that you're going to be in multiple players in the same space.

Look, I think when we invest in these companies, we always want to see that they will be the leaders in their own market. We have a lot of confidence that Helsing, with its mission, its execution, its access to capital, and its access to talent, should be the company that gets a disproportionate share of defense in Europe, and is doing that in the US.

We think this company called Rafi that we invested in is going to do that in India, and then they should go and be competitive in the global ecosystem.

I actually think maybe there's a new innovation in partnerships. We have not innovated partnerships ever. They're the same boring old metaphors. The question is: could these companies all think about engaging? It's not happening yet, but are there resilient ecosystems that get created where they have special advantages because of where they sit in which ecosystems, and they can leverage each other to gain more global share as well?

I'm very keen to see—while on one side we've created some structural inflation because of the need for global resilience, on the other side, I am curious: is the playbook going to change in how you become market leaders in this next phase?

You mentioned on the geopolitical side that it's this race for infrastructure, and I find there's often this rhetoric that's like, “Hey, it's China versus the US, and it's the war for AI.” Do you agree with that? Like, the race for AI—is there a destination? Is there a winner? What does that mean?

Look, the way we see it, we are in a bipolar world for sure. Despite all the recent turbulence around tariffs and the relationship between the different regions, the US, Europe, and India are very aligned in terms of core values. I do think it's going to settle down to be a place where AI gets developed with a common set of values.

I think the Chinese system is different, and the race, to me, is only in the context of which AI is better, because capitalism will force the adoption of that. Think about DeepSeek and the open-source models that have come about out of China. People are using them in the US because they’re better, and so now the question is which AI ends up winning.

There’s not much difference between China and the US. I think they’re very comparable in what they are in AI today. There’s a few months’ lead here and there. I’d even argue there’s always a second-mover advantage to people building on top, so I think they’re pretty comparable.

It is important that you see AI infrastructure in the West gain market share in businesses and be more pervasive. It’s equally important to make sure it’s done in a way that the compute productivity is captured onshore in each of the geographies, so it doesn’t leak from a lot of the places to a single company or a single country, because that creates imbalance. You want to be inclusive and abundant in your mindset with AI creation.

But is the competitive dynamic something we need to worry about in terms of companies coming out of the US versus China?

Yeah, I mean, that’s driving a lot of where the value is going to be and our ability to compete globally.

In terms of second-mover advantage, really interesting element. Do you primarily believe that just because of distillation and the benefits that we saw DeepSeek have as a result of being second?

I think you’re seeing this in different use cases. Take customer support, for example. When new models come out, if you start building your company in the GPT-3 era versus GPT-4 versus GPT-5, you just have more tools at your disposal.

The go-to-market advantage you may have created in a year, having started on GPT-4 versus GPT-5, might be anemic compared to the technology advantage that you have if you start in the GPT-5 era. The choices you make and how fast you can move, because the models are stronger, give you this potent-force unfair advantage.

Are you really going to rearchitect everything you did and take a step back, or not, because you actually have good momentum? What’s happening is that the good and greater companies are getting created with each new model, and the companies that start later end up having some advantages because of technology.

8. GC’s Anthropic Investment: Upside from a $60BN Price

The question is, can you not be bogged down by technical debt? Technical debt used to be on the order of a decade of coding, not a year of coding. Can you actually overcome that and make sure you remain well positioned on the new technology stacks or not? That’s the advantage that I think second-mover companies could theoretically have in these different verticals.

You mentioned the evolution of models there, from phase to phase. Anthropic is a big position for you. Can you talk to me about your first entry point into the company and the thesis that you had on first entry?

Yeah.

You’ve got to thank Ishan for this one.

Yeah, okay. Look, we invested in Anthropic less than a year ago, at the $60 billion round, where we saw that the use case around coding was becoming an interesting application that was going to distinguish them.

These models started to become distinguished. Obviously, everybody wants to do everything, but OpenAI, to me, is more of a consumer company with ChatGPT. I know they have enterprise ambitions, Codex, et cetera. I get it.

Anthropic kind of became an apps company in the cloud world, with coding as a use case, and they’re showing really good traction. That was the first time we felt, are we really betting on these companies with tremendous valuations, tremendous burns, and tremendous dilution toward some abstract AGI goal, or are they actually going to be businesses?

I actually think, risk-adjusted, that was the round that was interesting to do. As somebody who studies investment decisions in companies, you ask: where did people get lucky, and where did they actually make a great call?

Why do you think that was the round that was the best?

Because that’s when the use case that was going to draw them in and build a relationship with enterprises became very clear.

And it could scale, and it has.

When you invested, what was the revenue?

It was at the end of last year, so it was under—I think it was under $1 billion, if I recall. They’ve publicly said they’re going to grow nine times that. That’s not the forecast we modeled. They’ve done way better than we thought, which is amazing.

It’s an incredible team: very values-oriented and very execution-focused.

How much did you put in at $60 billion?

We put in a few hundred million.

A few hundred million. And then you did another few hundred million at $180 billion?

Yes.

Yeah.

I think you would. By the way, I would actually have 5x oversubscribed Anthropic.

This round was probably the cheapest round that got done this year on a multiple basis. Which company was raising capital at 20 times ARR? They’re all raising capital at 50 to 100 times ARR, and at a scale that’s 10 times bigger than any other companies that are raising capital.

Risk-adjusted—and I should say risk-adjusted carefully, because the durability of everything in the models is highly unclear—in that cohort, I would say that was the best-priced round you could have done. So I’m not surprised it was 5x oversubscribed.

Dude, risk-adjusted? That went out of the window years ago.

Welcome to “I eat risk for breakfast.” So I’m with you.

That is the caption for it. When you look out, do you put another few hundred million into the next round?

It depends on how the business does and depends on how they’re expanding. But when you get into growth, you have to be very fundamentally focused on actual economics, revenues, margins, profitability, scale, and market size.

The market size for these things is endless. There’s $500 billion of payroll for developers in the world, probably. I think it’s about $10 trillion of white-collar jobs, if I have that generally correct. It’s an insanely large market, and you’re naturally well positioned to be one of 2 or 3 players that’s going to go capture it.

So if you believe that this company could be 10 to 20 times bigger from here and is well priced even at this valuation, if it keeps on that trajectory, of course you would want to invest. Who would not want to invest?

Yes, but you also then look at the information. Did you see this incredible chart? I don’t know if you saw this pie chart yesterday with OpenAI’s distributions, or the value generated, and 1%—I think it’s like $5 billion—is to early investors.

You look at that and think, well done, founders and team. This is the greatest transfer of wealth from venture capitalists to founders and team members. When you look at the employee stock-based compensation that’s going out now, this is the greatest transfer of wealth ever.

Is it actually a great investment when you look at the dilutive nature and cash burn of these businesses?

Well, I think OpenAI—I saw that early round, and Sam’s a force of nature, and I’ve said this publicly. The guy can bend reality, and he has. It’s changed the world.

I just couldn’t get my arms around the structure. If those numbers are correct—I don’t know if they’re entirely correct—you would say $5 billion, the $200 million at the $1 billion round, only generated 25x.

Our best companies, like Livongo and Circle and others—Stripe and others—our first rounds were not 25x. They were hundreds of times in terms of return. So I agree with you: dilution took a huge toll here.

But there are 2 reasons. One is because that structure led to, “The nonprofit needs to be given a share.” The second is the compute that was needed to make this happen. The first one to provide that compute was Microsoft, and they had a lot of leverage. That was a good deal for Microsoft. They made a huge amount of return, because without that, this was never going to be a company.

So I think it’s just the sequencing of who really took the risk, risk-adjusted, with capital. Microsoft maybe took more risk in a lot of ways, and they did benefit. The dilution comes from a lot of that dynamic as well.

How do you analyze that relationship with Microsoft? Speaking of Anthropic, Microsoft has now openly said that, for the majority of its software engineering, they’re actually using Anthropic.

Yeah.

How do you analyze that relationship?

Well, I think that relationship was—if you remember when that was done, everybody said Satya was brilliant. That was an incredible way to essentially buy innovation.

It’s like what biotech and pharma companies do, right? They bought innovation in AI because the internal efforts maybe weren’t as productive. It gave them the halo effect of being the leading AI company and gave them an entry with Azure. Azure has had a huge lift because of OpenAI as well, and they’ve really gained market share in the cloud industry as a result.

So that was an amazing investment for Microsoft. Is that an enduring investment? No. Obviously, now they’ve gone at odds with each other, because there are ambitions that collide between the 2 companies, and they want to have more choice.

They want to have Anthropic at the table as well. I think that’s just normal, good business thinking. I get it, but taking that bet was hugely, hugely valuable for Microsoft, I think. If you look at the return, by the way, if they put in $20 billion, they’re the ones who have the highest multiple return as well. So it was a great investment on a financial basis, and way more on a strategic basis.

As an Anthropic holder, do you worry when you see Sam Altman talk about, “We’re going to need to spend hundreds of billions,” and then you see his deal with Oracle, where both of them are going to need to be levered up to the hilt to be able to finance it on both of their sides?

I think, look, never bet against Sam, but I feel like he’s doing a lot of things. He’s doing it in every single dimension: the phone, the data center, the infrastructure, and all that. I think Anthropic is a much more focused, product-oriented company. They have not taken as much money to get here, to this size, as well. I’d argue their enterprise business will be bigger than OpenAI’s, if it already isn’t, without having spent that much.

So I do think capital is a lever, but it’s not the only lever. Execution matters, and I would say one of the things I’ve learned—and Arthur Mensch at Mistral really taught us about this—is that you can waste a lot of compute, too. A more focused team with a focused agenda can probably get there much more effectively. Anthropic seems to be doing that, being very targeted in the bets they’re picking and doing them really, really well.

I’m a big believer that leadership is really maturing to be excellent company builders. I mean, Dario Amodei—if you think about that person a few years ago, running research, and the kinds of choices he’s made, both in business and for society, and how they’ve scaled, the bets they’ve made and how well they’ve paid off every step of the way, and the products they’ve launched—it’s really impressive. That’s why the investors there will probably end up doing better on a multiple basis than, again, in a world of what’s durable and what’s not, the early investors in OpenAI would, if that math is correct—the one that you were referring to.

9. Does Revenue Growth Matter in a World of AI

The hard thing is that growth is great until it’s not great, and at some point it does reduce and reduce and reduce. $1 billion goes to $9 billion, which is insane. I mean, nuts—nuts numbers we’ve never seen before. What is it next year? $27 billion, which would be 3x, which would still be great, but at some point growth does reduce, and then there’s the core business that sits beneath it.

I can just interrupt for a second. Let’s say it’s $27 billion. I don’t know what numbers they have shared, and they’ve always done better than they’ve said, by the way, too. But let’s say it’s $27 billion.

You’re not also Anthropic’s CFO—not a spoiler alert.

Yeah, yeah, yeah. I’ve got to caution you, but if you think about it, that business would still grow 200% to go from, let’s say, $9 billion this year to $27 billion. That’s tremendous growth, and you put any multiple on that, that is a very valuable company. Think about technology multiples on that kind of growth. That’s a really valuable company.

I think there’s significant headroom because the market size is so large, even from there, that they can maintain good growth. It doesn’t have to be this crazy growth. Even if you only price it at 20 times this year’s ARR, so 20 times $27 billion—

If 20 times $27 billion—

Take the same multiple, you know, which you could get in public markets with—

$550 billion.

Something like that. My point is that it’s a half-trillion-dollar company by the end of next year. I’m not saying that’s what it’s going to be, but if they hit their numbers, I don’t see why that won’t happen. I’m just saying, public-market comps—that’s what those things are going to be valued at.

I completely agree. Does margin not matter anymore today?

Margin matters, absolutely, and I actually think that’s another place where they’ve done a good job. The reality is that when you think about the ROI in the coding space, you’re doing the work. This coding agent is essentially a replacement for engineering, right? You’ll start with low-end, junior engineers and move to more senior engineers. Even a junior engineer makes $80,000 to $100,000 a year, so your pricing power there is actually pretty significant if you’re truly doing that kind of work.

Margins are not going to be an issue, and margins already are not an issue for Anthropic. They have a good command of it; they’ve been very disciplined about how they’ve built their business.

Do you worry about the competitive nature of the landscape when you look across at Codex, but then you’ve also got Cognition, and then you’ve got Cursor, and you’ve got your kind of lower, more consumer-end products—your Lovables, your Replits?

Great. I’ll take you back to the clouds. You could use the same logic on the clouds and say, “Hey, they’re going to get commoditized.” You’ve got 3; you’ll have 3, probably. There were 3 big telcos in every geography, probably 3 big clouds, probably 3 big AI models. I’m just sort of saying that empirically.

Think about the margins that the cloud companies have. It’s like 70s. I do think these companies will all figure out the margin structure really well and at scale, because there are so many different ways they can add value to hold on to that margin, and they’re all kind of getting specialized into different areas where they are going to be doing that.

There are, but there are only 3 cloud providers, really.

There are many today, but I think that’s going to shrink. I don’t think everybody’s going to make it.

So you think there’s going to be a real shrinkage in those products?

I think you’ll have a couple of global ones and a couple of sovereign ones in every geography. That probably is what will end up happening in AI, in my view. Not everybody’s going to make it. You have a lot of other models that have been funded, different approaches as well. We’ll see what happens to those. Is there a new architecture that emerges? But it’s not going to be that many.

Think about the size of the market. The labor market is the AI market, you know.

So you buy that massive sum: “Hey, there’s a $10 trillion labor GDP up for grabs, and if we capture $2 trillion of that, game on”?

10. The Risks of an AI-Driven Economy

I don’t see why not. I don’t see why technology is not going to do most of the work we do in companies better than humans do. Going back to your earlier point, market forces will take us there because it’ll be cheaper, it’ll be faster, it’ll be better, and it’ll give the businesses more leverage. I think that’s a real trend we’re going to head down the path of.

The economy is so on the AI hype wave—or not the hype wave, but the AI momentum wave—because a lot of it’s not hype. When you look at the concentration of shareholder-value accumulation, it’s solely predicated on the Magnificent 7 pretty much at this point.

Do you worry that we hit a speed bump and, when we hit a speed bump in the AI train, the world literally kind of crashes, given how much is predicated on that? Or do you think that just won’t happen because we are on such an exponential upswing?

I don’t think it’s going to be a speed bump that spirals us down, for the simple reason that with every new model and its capabilities, there’s a certain amount of the content of white-collar jobs—the $10 trillion you mentioned—that is now doable, and a speed bump is not going to reverse that.

I have this sort of visual in my mind where we’re cumulatively taking more and more of the labor content, and AI is better than us in more and more of it. Over time, it’s going to be most of it, and when robotics comes, it’ll be all of it. In that context, a lot can already be done, and we have the energy and the compute to be able to support that.

The question is the really frontier stuff: what are we going to do with that? I think you could see speed bumps there. The speed bump could be that the architecture doesn’t scale and that you need world models to take over, or some of the new architectures people are developing that are non-language-oriented could be needed. So I think the frontier could slow, but I don’t think this transition of labor, or the melting of labor into productivity, is necessarily going to slow because the economics are way too compelling for that not to happen, in my view.

You mentioned geo-winners in some respects. You mentioned a company likely called Mistral. I love Arthur. I think the world of him. I’m obviously a proud European. A lot of people are going, “Huh, it feels like they’ve been left behind.” Calling a spade a spade, do you think sovereignty is enough of a reason for Mistral to be a winner?

It’s a great question. Jeannette is on the board there, and we talk a lot about it. We’ve had many conversations and meals with Arthur about this as well. I look at that company and I see Arthur’s growth from a scientist to a CEO.

And remember, it's a 2-year-old company. Do you want to hear a funny story?

Yeah.

I got introduced to him by Jean-Charles at Alan. I was the first VC he ever met.

Yeah, sure.

He took a video call with me on a park bench in Paris, and I said, “Dude, I’m going to give it to you straight. I’ve never had such a bad pitch, and you are competing against Sam Altman, the mother of all fundraisers. This is not going to end well.” Now I see him pitching, and I’m like, “Well, he changed.”

He changed. But I think this is the point. It wasn’t only his ability to aggregate capital. He stayed focused on doing really disciplined work in the way they build models, and I think they were compute- and capital-constrained. So they fell behind, but I think they’ve caught up. Everything I hear now is that their models are now again sort of there as they’re investing.

He’s figured out how to aggregate capital. You’ve seen that, and he’s also figured out that his relationship with customers needs to be a lot more commercial than “if you build it, they will come.” What I see makes me bullish on what they will do, even though I was anxious about it, because he and his team are growing up. I think they’re learning how to be in this competitive world.

By the way, if OpenAI and Anthropic didn’t exist in this world, you would say this is the hottest startup in the world in terms of how fast they’re scaling, what they’ve accomplished, and their valuation and progress. You just have the overhang of these 2 monsters that got the flywheel going with capital and products and so on. That’s why we ask if this is interesting. I actually think they’ll build a pretty compelling business.

I see a lot of interest from companies in Europe, but all over the world, that want an open-source player. Who else is truly dedicated to open source and doing it in a way that enterprises care about? It’s not Meta; they’re not an enterprise company. In the West, it really is Mistral.

Can you name for me success stories where sovereignty was the number-one driver of their success? I don’t mean that to be tricky.

Many companies in defense. All the U.S. defense primes were built off sovereignty. We were the biggest allocator of spend in defense, and that’s what likely Lockheed Martin, likely Raytheon, and Boeing were largely built on. Sovereignty also dictated who and which countries they sold to. The State Department gets involved, but it was all dictated by that. I think AI is that strategic a technology.

Does revenue growth matter anymore? We just had the founders of a company likely called Mercor on the show.

We led the seed there.

Yeah, yeah. $1 million to $500 million in 17 months.

Unbelievable.

Well done. Thrilled for you. Does it matter anymore, though? Every week, there’s a new company going from $1 million to $100 million or $1 million to $500 million. How do you think about that?

It’s a great question, Harry. The internal conversation I had about this was, of course it matters, but the normal has changed. We used to think about it like, when we did Samsara and Gusto and some of these companies, “Yeah, yeah, triple, triple, triple, double, double.”

You look at these companies and say, “Wait a minute. Going from 1 to 3 to 9 to 27 is not interesting,” or 1 to 5 to 9 to 27 is not—you know, whatever the math is—not interesting. You’ve got to go from 1 to 15 to 20 to 100, and they’re all, on a revenue basis, more interesting than the stuff we thought was the most interesting 5 years ago.

That’s because it speaks to the way value concentrates in the hands of a few companies. It speaks to the fact that these technologies underneath are so high-leverage that they’re potent in making these companies grow fast when you actually get a product right. That’s what’s going on.

Durability is a question. The thing that’s unknown is that we never had so much scale without just taking durability for granted. What’s the question that we all grapple with today? Lovable is an amazing company. Anton’s done a great job. Is that going to be around?

Mercor is an amazing company. That’s what the naysayers say. The people that believe in it—we have a huge conviction in Mercor. We have our own thesis. Everybody’s grappling with this, and we all have our theories. Some people will get lucky, and I do think some of these companies that grow really fast in this space will also not be around.

It’s so funny. I’m very good friends with Rory O’Driscoll from Scale, who I think is one of the most brilliantly strategic but unassuming people, and I’ve learned a lot from him. He said the go-to-market has fundamentally changed in a world of AI where, bluntly, it’s a case of: you just go into a market, scream the loudest in the room, gain mindshare, and deliver from there. Your Harveys and your Abridges are great examples of that. Do you agree with that?

I think, if you look at Harvey—or whoever invested in companies likely called Legora and Judia—take the legal space, right? It’s not just that they screamed the loudest and won. The interesting thing that has happened with AI—I want to go back to 1 important observation—is that, for the first time, every CEO in every industry in every country is thinking about, “What do I do with this technology?” That never happened before. Cloud wasn’t like that. PCs certainly weren’t like that. The internet wasn’t like that.

Everybody was like, “What do I do with this?” All of a sudden, in every department, people popped up early, and they got to go around. There was resonance with the customers. There wasn’t as much evangelism for the earliest companies; everybody just wanted to engage. That’s why these companies got initial momentum so fast.

But then I want to go back to my second-mover advantage. Some of the ones that started after had a chance to take a step back and be like, “Oh, wait, there’s better technology now.” I think we’ve learned that the proposition needs to be better.

The initial diffusion in the zeitgeist was really fast, but the actual deployment, to my earlier points, was like, “How do we really use this?” That’s where this next generation of companies that are coming are just more sophisticated at that. So is the early-mover advantage in some of those companies really going to take hold or not? I think it remains TBD.

So when you look at today, is triple, triple, triple, double, double dead?

Triple, triple, double, double is definitely dead. I tell our investors, “Don’t bring that to me.”

What do you do with the generation of SaaS companies you have? I hope you have it, because I have it.

They’re good companies. By the way, they’re durable companies. They’re going to be around, and this is actually an observation I give Pranav on our team a lot of credit for. He said, “These companies—venture capital doesn’t like them anymore because they grow 20% and they’re not hyper-growth. Nobody wants to fund them. But there are some founders for whom this is their life’s work, and if you give them alternate ways to endure and scale, they will, and they will create value. It’ll just take longer.”

That’s where we’ve also made sure our Customer Value Fund supports those types of founders. We obviously want to support the fastest-growing companies in venture, but also the ones that are fundamentally good businesses and would be profitable if they weren’t investing in sales and marketing. We give them capital to scale their sales and marketing.

That’s what the Customer Value strategy does, and it’s entirely focused on founders who deserve to endure and compound because their companies are good, their customers like them, and the value proposition is growing. They’re just not in the zeitgeist.

11. Lessons from Covid-Era Market Shifts

I’ve never felt so much uncertainty in what I do. I’m trying to understand it. It almost feels like COVID, where we had these highly transient categories that were created, and it was like: do we actually just do exercises at home all day on Pelotons, or do we go back to gyms? You didn’t know what would be enduring, strong markets and what wouldn’t, and I feel that is the case here. Have you ever felt such uncertainty in investing, and what would you advise me?

Yeah, it’s a great question, and I’ll give you the same advice that we follow here at GC. This is peak ambiguity.

Everything we do to support founders, everything you do—you’ve shown me a lot of your cool ways of supporting founders and getting them excited—all the products and solutions we’ve created to support founders, the question is: to what end?

Having a true sense of a long-term set of principles that you believe in is all you can lean on in a world of ambiguity. The way you navigate ambiguity is by having a true north.

In the U.S., we have this enormous movement toward transforming healthcare post-pandemic, and so everything we do in healthcare, we go back to saying, “Is this decision going to make it proactive, affordable, and accessible or not?”

In Europe, the work that Jeannette is doing is very much about Europe’s resilience with AI. Everything we do, we look at and say, “Is this going to make the economy here more resilient?”—whether it’s an investment decision, a relationship decision, or a partnership decision.

Having that kind of sense of where you’re going, so you’re directionally aligned with your values, is the only thing you can lean on.

There’s so much uncertainty. It’s so difficult, and I feel bad for investors who are learning in this era because, in some ways, you have no signals to determine whether your decisions were right or wrong. You have this great revenue growth to lean on, but no durability. Then you have great margin structures, but revenue growth is just not there.

You have to be values-oriented. You have to have a sense of what you’re really trying to do. At GC, we say we build deep relationships with people and enduring companies, and we’re doing that to transform industries across the world.

What does that do? It gives our founders access to talent, policy sophistication, distribution, and differentiator capital. If we give founders all of that and have a set of values with which we want to march down these industries, I think we’ll be okay. We’re trying to take faith in that, and I would have everybody think about it because I do think we are building the future.

It’s an amazing time. We will shape what society is going to look like for probably 100 years. This shift is as big as what electricity was 150 years ago. We get to shape it, but we have to make those decisions: What do you want this to look like?

I think that intentionality should always be in the back of our minds as we make short-term decisions, as we deal with FOMO, and as we deal with how we’re scaling our business and supporting our founders. What are the things we choose to do and not to do? There’s far more opportunity than any one of us can pursue.

You’re the true north. You continuously mention the exponential market size, or the insane market size, that we have. It makes me think of—I wish we could edit video more efficiently—but it’s either a Buffett or a Munger quote: It’s better to buy a great business at a good price than a good business at a great price. Is there any point in being price-sensitive if markets are a trillion dollars?

One of my partners, Joel Cutler, used to say, “Price only hurts once.” It’s like buying a Gucci bag: Price only hurts once, but then you’ll never regret it.

It’s not true. Whenever I see my mother with a Chanel bag, I’m reminded of the dent it caused.

I think there’s actually wisdom in that comment, even though it’s a cheeky comment. First of all, when did we ever get price right? I’ve been doing this for 25 years. We’ve seen all these models, and I am yet to see an investor, at least in our firm, ever nail the price in the way they thought it was going to be. It’s usually worse than that, and we make all the money when it’s better than what we thought.

If all the money is made in what’s better than we thought, using price to pass—I’ll tell you, investors use price as a reason to pass because they couldn’t gain conviction elsewhere. It just makes them sound pragmatic. I generally get very ticked off when somebody says, “I love this company, but I don’t like the price.” I’m just like, “Well, then you don’t know if you love this company.”

You’re taking solace and trying to be a price-disciplined investor because you didn’t really understand the potential of the company or see what it’s actually going to be. If it’s destined for greatness, then jump in.

What about if it’s a capped-upside company? Let me walk you through this. There’s a data-providing company—I’m just making this up—and you say, “This is a good business, and I can easily see a $2–4 billion outcome here, either to a strategic buyer or as a public company.” I’m getting in at an $80 million pre-money valuation. That’s a great way to make a lot of money: a potential 25x on an early-stage check, and a $10 million check that can return my fund almost entirely. But if it’s $140 million, it’s very different from $80 million in terms of the multiples.

100% right. We say these things: “It’s a capped-upside company.” What does that mean? When I invested in Stripe, all the people I called who knew payments were like, “This is a niche thing. Why are you doing it?” I kid you not.

I just don’t want to listen to the experts about what they think their industries are or are going to be. Markets expand, too. The humility in this business is understanding that we don’t know what’s going to happen in the future.

If it’s truly capped-upside, then you shouldn’t be doing it anyway. It’s not a price question. We’re in the business of trying to build and back companies that can become enduring, very large businesses. That’s a precondition.

People get stuck either because their companies are completely mediocre and they’re not even going to be worth $2–4 billion, or because they’re actually great and you’re not willing to stretch because you’re not willing to believe what the world is going to look like.

I always think of Peter Thiel’s biggest investing mistake, which was not doing the next round in Facebook.

When did you not do the next round in a company where, with the benefit of hindsight, you thought, “I should have”? What did you learn?

I don’t think we have time for all my mistakes, but I’ll give you a very recent example. We have a company that’s a decacorn now. The investor who led it at GC called me and said, “Congratulations, you’re going to make over a billion dollars on this investment—and you’re an idiot, because you gave up making the second billion.”

You gave up making the second billion because you didn’t double down. I think that’s where you can get a lot of it wrong. As I mentioned earlier, I’ve invested in Stripe 14 times. If you’re in the best companies, you really should be buying into them constantly.

In fact, that’s the reason to scale capital. The reason to scale capital isn’t to be a low-margin business. It’s because you want to have capital to build the very best ones and really lean into them. That’s ultimately where you drive your best returns, and that requires courage, conviction, and belief in how markets are going to change.

Do I worry about capital concentration limits? I think I have invested in probably over 200 companies, and 60–70% of the returns are from around 10 companies over 25 years. Capital concentration is the way you drive returns. You just have to be right.

Which single company are you most capital-concentrated in?

Stripe. It’s about $1 billion.

Are there concerns around cross-fund investing? LPs often worry about it and don’t like it.

Cross-fund investing is an important consideration. We think a lot about it before we cross funds, but we do cross funds. You want to make sure you’ve put enough capital to work in the fund it’s in that you feel comfortable with the risk you’re taking before you cross it.

If you really believe in something, you want to make sure it becomes one of your largest positions in that fund before you go to the next fund. But I also try not to have more than 10–15% in a single company in a fund. At some point, if it’s truly a great company, you will be forced to cross because you should have multiple funds benefiting from it.

I totally get you, and I agree. I always remember Brian Singerman saying that capital concentration limits are the enemy of great venture returns, which is why they’ll often have 30% in a single company. I thought, “Wow, I need to get more courage.”

Yeah. Concentration is key to being great at investing. I genuinely believe that.

When companies go public, you have the choice to distribute the shares or not. How do you think about whether you’re better placed than your LPs to manage those positions once a company goes public?

I always look at it as a variety of things. One is: Will our time matter? If our time continues to be spent on this company, will it matter in terms of compounding from here or not? Some of the companies we started, for example, would make sense to stay involved with, so you want to stay on and do that.

The other thing I look at is how long we should hold it to make sure we drive the best returns for that fund. Is this the company that should be compounding the most to keep generating returns and driving performance, given our commitments to the LPs?

A lot of the LPs will have their public sleeve and their private sleeve. You give them stock, and they’ll sell it. It’s programmatic for them. But we want to make sure we give it to them at a point where we’ve really captured enough value. That’s another factor we think about in that context.

Which suggests you do think you’re better at managing it than they are.

They may not be managing it. That’s what I’m saying. In their private sleeve, they’re just not going to hold it. The private team is basically told that once you get stock, you sell.

12. Navigating Secondary Markets

The other thing is that, in our lead companies—where we led and where we go public—we have a lot of stock. You also have to be measured in how you distribute stock because, by doing too much at one time, you could hurt the price of the stock, which hurts the rest of the position. There’s also a pacing question: How do you liquidate?

How do you think about navigating secondary markets?

You know, when we look at you, there's a very strong chance that we have a trillion-dollar private company in an OpenAI of the world. How do you think about navigating secondary markets when the public market is sometimes not there?

Well, look, I think for the very best companies, private markets behave like public markets. There's a secondary market where you can liquidate, so your holders can take liquidity. Employees can take liquidity. You have access to credit. You can do M&A.

Your stock's value and your valuations are believed. I'm talking about the very best: the Stripes, the SpaceXs, and I think OpenAI is going to get there, Anthropic, and so on. Databricks is getting there slowly. So, for the very best, that's what happens.

Then there are the very good companies—not the, let's say, top 10 or 15 private ones, not the magnificent private 10, if you will. For them, going public and getting validated could actually be more helpful. It may be that the secondary market isn't behaving as well, or they can't do M&A as effectively, or they need to access a lot more capital than they can just by being in the private markets. I think those decisions are what ultimately push you to go public.

Then there is, as I said, this bloated set of companies that are good companies that'll compound at 20–25%, maybe forever, that have no access to public markets because they're too small. They're not a billion-dollar company growing 30% a year that the public markets would be excited about, and they're too slow for venture to fund. That's the purgatory where we need innovation, and that's where the Customer Value Fund resides: helping these companies get to that scale so they can go public someday.

Is the extension of private markets not an increasingly harmful thing to the distribution of wealth in society? Before, it would be your Fidelitys, your T. Rowes, and the pensioners of the world. My grandparents would pay them 20 bps, 30 bps, 40 bps. Now, with the extension of private markets, you get money managers getting 2 and 20.

It's an enormous issue. I'll tell you those stories. When I published my first book, Unscaled, in 2018, I had the leadership of Vanguard come by, and they talked about, “Look, Main Street doesn't have access to this asset class.”

So I spent a lot of time in 2018 saying, “How do we do that? How do we actually give retail access to our funds?” Now, with 401(k) changes and some of the ’40 Act evolution, you actually can do that. So I fully expect that you will start seeing products that give retail access to the best companies in technology, and we will definitely engage in that because it's the right thing to do.

Do you agree with that? When I saw this discussed, I was like, “Oh, wow, guys, well done.” But part of me is like, well, isn't that what we want—the democratization of access?

Yeah. Look, I think everybody wins in this. You open up large pools of capital for investing in technology, but you also open up large pools of opportunity for people that don't have it otherwise.

We don't need to look at it as, “Oh, we're doing it because it's sneaky.” I think it's good for the world, and we should do it. If you're oversubscribed, make room for it. That's the main thing, which is where this matters: when you have more capital than you want to take on in a fund, would you let it in or not? What I'm saying is that we should be making room for this.

Do you think fee structures need to change?

Look, I think I am very much focused on performance as the number one thing. That's why I said in the beginning that I measure us as: how good a seed firm are we? In the context of everything we do, is our core right? Are we doing the highest-risk, highest-reward work? Are we helping founders in the early stages?

You want to stay high-performance. If you want to stay high-performance, then your incentive should be much more focused on generating carry and making it a prosperous place for your team than generating fees, which to me can be a distraction. You said about carry. By the way, in our business, we don't distribute any fees. We invest everything back in the business, and that's a deep belief that we don't want to be in the game where the partners of the funds at General Catalyst want bigger and bigger funds because they can take bigger and bigger distributions. Whatever fees we get, we invest back in the business.

I'm so sorry to be so blunt, but don't partners make 3 or 5 million bucks?

Less than that.

Do you worry that you're not going to get the best partners because they're getting that at alternative funds?

It depends on if they're focused on performance or salary. I think that, to me, is a filter, and my commitment is that you go deliver your dream and you'll make more money than anywhere else, but it's got to be aligned. We have to be focused on performance and value creation versus being rich and fat and happy on salaries. That is just not the culture we want.

Do you think that is the same or reflective of the rest of the venture ecosystem?

I have no idea. I pay no attention to it.

You don't?

No.

Do you not pay attention to your competitors?

I don't know what people get paid there. I don't.

Which competitors do you most respect?

All of them. They all make us better.

If you were to choose 1, so I could say, like, a16z?

I think a likely a16z in Europe have done an incredible job really carving out their industry and knowing what is their type of deal.

Look, Andy Golden, who ran the Princeton endowment and is doing some stuff with us now, had a huge impact on me as I was helping build GC. He always said, “Run your own race.” It's actually a chapter in my upcoming book about that: play your own game. And that really—

Do you not think you can learn from others?

No, I want to learn from others, but I don't want to be in the game of competing in the zero-sum game of venture capital.

When you look at that product that you give to founders, you have a lot of products now—from your geographies in Europe, India, and the U.S. to seed, growth, the Customer Value Fund, and rollups. What product do you not have that you would like to have?

Do you have, like, a strong heritage in the wealth-management business?

We do.

Of course you do.

We do. Yeah, and that's actually growing rapidly. It's a fairly large business at this point.

I think we have a roadmap, to be honest, and I always look at it as: what do the founders need? We have a roadmap of things that we think about, that we will over time experiment with and see if we should bring in.

I mean, we have really 3 products, right? We have venture capital, starting with seed; we have the Customer Value Fund; and we have Creation, which is where we do the rollups and hatches, building companies from scratch, really being builders. We have 3 products today.

But which one would you most like to have?

I think we need to figure out infrastructure, if you ask me. The race in AI—the thing that I'm very focused on learning about, and we're early in our thinking—is this: in order to get AI right, you have to get energy right. Everybody knows this.

If you think about energy, it's a really interesting opportunity, with all the new demand, to actually move toward sustainability profitably as well. But in the short term, you don't have sustainable solutions; you really have natural gas in the U.S., for example. So what is that arc with which we're going to think about energy to really get AI right?

That's an infrastructure problem. That's an example of something where, if we care about using AI to change the world—all of our transformation work is about transforming industries and businesses with AI all over the world—then we need to figure out what to do with energy.

13. Changing the Capital Supply Base

Totally understand that. How do you think about how you need to change the capital supply base with the different products? I very much operate in the endowment, fund, and foundation world, which is lovely and nice, but does it change drastically when you move across products?

It's a great question. I think as we went through the succession at GC—from David, Joel, and me running the business for most of the last decade together, with the 2 of them before that, to me taking on as CEO—we also had a succession from a leadership standpoint. Ken Chenault came and became our chairman and mentored me, but we also had an interesting evolution of our LP base.

The LPs—the endowment foundations, many of which are huge backers of ours, and which I consider really part of our team—had the mindset: “We want managers to be dedicated to single strategies, and we will create the portfolio.”

The break in strategy we made was to say, “Well, no, we need to have all the strategies that make the founders successful, and you back us to make the founders successful because then we'll create alpha.” That's what we were on a campaign to convince enough of them to stay on with us and do that, and they did.

Then we went and got a lot of the state pensions in the U.S. because, going back to your point, I want to make sure we create wealth for everybody in the U.S.

There was a motivator there as well. Now we're actually very deeply partnering with sovereigns because, as I said earlier, AI is transforming countries. But there, the relationship needs to be more of a partnership. We're helping them think about what they can be doing in their regions, and we're saying, “Let us be a strategic partner to you, and you be a capital provider to us.” We're doing interesting partnerships in that regard.

You have to go to sovereigns because they're the only ones who can write a billion-dollar check.

I think states—some states—can as well, but I think there's scaled capital in each of these areas. The key is how your work fits in the context of their strategy, and GC is a flexible platform where people can engage in that context.

As we were joking about before, retail is another one that's about to open up. There's about $16 trillion of retail capital.

What will cause retail to open?

I think retail is opening up in terms of the ’40 Act regulations and the changes in what 401(k)s can invest in. It is going to open up, and it should open up. We need to be responsible about exposing retail to the right part of the risk curve in private markets.

I think it's very important that we do that. We need to be very thoughtful about whether and when we make GC available—we don't do that today—and what the right way to do it would be. I think the whole industry is going to think about that.

In my view, all these pools are there. If we are to service founders the right way and have all the capital and flexible capital solutions available for them to build their companies, we need to engage with all these capital sources that are willing to support different parts of that stack.

When retail opens, is it a trickle or is it a flood?

I think, hopefully, it's a trickle. It starts slowly and then scales. I do think it'll scale in a big way, but I think we should be careful.

I worry, when we go back to your very early statement, that there aren't many Patricks and Johns, or Sams or Darios. The problem is that there aren't enough truly generational, defining entrepreneurs for the supply side of cash. That will only get worse.

I'm not looking at retail opening up going, “Woo-hoo.” I'm going, “Wow, this is about to get harder.”

That's right. I think retail can open up to the very best companies at scale, in my view. I think Robinhood is working on some work there, for example.

The tokenization—

They just announced that they're going to create a way to give retail access to some of the top companies as well. I think it's a recent announcement.

You could be giving people access to SpaceX and Stripe. You're not going to regret it. It'll do right by them, and you'll feel proud of it. What you don't want to do is take retail and put it into the bottom quartile of venture capital funds that lose money.

I think that's where it needs to be trickling in, to make sure it goes where the returns are. We cannot—and should not—put retail into very high-risk situations where they lose money. I feel very strongly about that. We have to be careful.

14. Looking Back on Product & Strategy Mistakes

I think it needs to be trickling down the risk curve in terms of how retail accesses our asset class.

What did you do that you wish you hadn't done in the last 10 years? It doesn't need to be a deal. I'm more thinking about product, strategy, or firm-building.

The good thing about our culture—and I give the founders of GC a lot of credit—is that anytime I had a crazy idea, they supported it. I've usually been able to do most of what I wanted to.

There's one place where I look back and say, “Did I make a mistake?” When the financial services market took off, I was like, “I want to be in the best company.” I was like, “Let's invest in Stripe, but let's not do Square,” and whatever else.

When the AI stuff happened, I was like, “I want to be where I think I can actually make the most money.” In hindsight, we should have just gone and indexed those. Some other investors, including one likely called Yuri Milner, did a great job with it and did very well.

I was focused on, “I want to be in the best,” and I was like, “Aha, see how good I am? I did the best one.” If I could go back, I would understand that, in certain parts of the stack, indexing, if you can afford to have the capital base, is a better strategy than trying to pick in a world of peak ambiguity.

That's something where I'm a slow learner. I feel like I've been at it for 25 years, and I'm starting to understand that better.

You've moved to that now?

I have not moved to that, but I'm going to wait until the next time I miss it in AI. Our friends at a firm likely called Lightspeed did a great job in AI, for example. I think it's going to work out really well for them.

Why do you think that?

When you know the trend is going to win but you don't know which one is going to win, you're better off backing all of them than trying to pick, meaningfully play, and get it wrong.

Do you regret not being in OpenAI when you had the chance to, but didn't, because of the structure?

This is a daily conversation I have with myself and with my partners. I do regret it, because of the amount of learning we would have had if I had been at a front-row seat, really understanding what's going on. I wish I had that.

Would it have prevented you from doing Anthropic?

I don't think so. I think there are plenty of investors that are in both companies. I think many people overthought that structure, and I overthought it as well.

There's a lot going on in that platform that's changing the world, and I don't have a front-row seat. So, yes, I do regret that a little bit.

We mentioned focus on performance. Circle's IPO was nuts, and it did wonders for the fund in terms of returns. How did that fund return look? We were chatting about it outside.

That fund is one of our 2 or 3 best funds. Just to tell you what was in it: Livongo was in that, Snap was in that, Circle was in that, and Gusto. I think I'm missing a few others. It's probably going to end up being a 13–15x fund.

The story's not over yet. So how big is the fund?

It was $500 million.

Wow. Well done.

We need to keep doing it. That was a long time ago.

What is GC in 10 years?

I think GC is going to look like the most diversified solution for founders to build enduring companies. That's the lens with which we justify everything that's in it.

If you looked at it, GC as a business is going to feel like a strategic conglomerate, where every part of GC is in service of founders—whether it's giving them access to distribution, access to policy, access to capital, or access to wealth management. It's all about founders. It's the platform for founders.

How many team members do you have?

We are over 300 people.

300 people.

Yeah.

15. Boutique vs Mega-Funds in VC

Positively small compared to Andreessen Horowitz. Absolutely. We're fine. We're tiny.

On the future of venture, everyone does this binary. You're either the massive, massive AUM gatherer, or you're the boutique provider, and that's it. Do you agree with that binary view of venture, or do you think it's an alternate view?

I don't like that view. I don't like that framing.

My framing is actually Walmart and Chanel.

Totally. Totally. But I would say we want to have the biggest AUM in venture because that means we're doing the best job—not because we have a lot of companies or we've raised a lot of money, but because we're in 20 Stripes.

The biggest AUM doesn't mean you're doing the best job.

No, no. I'm saying the kind of AUM I want to be the biggest is the kind where we have the fewest number of companies, meaning our companies have created a lot of value. It's not the amount of money we raised.

AUM can be one of two things: How much money did you raise, or what is the value of the capital you raised? I want the value of the capital you raised to be the biggest, but the amount of money you raised to be the smallest. That's when you've created the most alpha.

That's why I think this “biggest AUM” is not a very informative way to look at it. If you're boutique, you could still have really, really big AUM. If your portfolio was only the top 10 companies that got funded, you actually would have the biggest AUM, even if it was a $500 million fund or a $300 million fund—whatever you call boutique.

The focus needs to be on being the support for the best founders to build the biggest companies, which will give you the biggest AUM and the biggest performance, and not on whether you can go raise the most amount of money for venture.

That's why I said in the beginning that our aspiration in venture capital is to be the best seed firm in the world—or second-best after you.

I have 2 more. What was the most memorable first founder meeting? Then I'll tell you why I laughed.

The most memorable founder meeting was with Patrick Collison. One of the questions I asked him was one of those moments, like in the movie The Sixth Sense when the ring falls and the guy says, “Oh, I’m the one who’s dead.” You just feel like they knew something about the world and how to think about it. That’s how I felt with him, because I asked him, “Who are your ideal customers?” and he said, “They haven’t been born yet.”

He was talking about the developer movement. Remember, this is 2010, right? He was talking about the developer movement and what was about to come, and I was just like, “Oh, crap. I don’t even have a complete view of the world and what’s happening around me.” I was like, “I have to back this person.” I didn’t even know payments.

Everybody in payments was telling me what was wrong with payments, and I kind of didn’t care. I didn’t know what was in this thing, but he saw something, and we had to be part of it. That moment taught me a lot about humility in terms of what this business is all about. It really is about backing the best people.

Honestly, it had a huge impact, because I saw him and John Collison build Stripe. I was like, “Why does our business run so badly when we’re in the service of helping build big businesses that can be run really well?” I want to run well. GC should be running with the same rigor these companies run with. I think we still run pretty messily—we’re still a very messy company—but the aspiration is that we want to run with the same rigor as companies like Stripe.

Which loss hurts the most? The reason I laughed when you just said, “I’ve only lost 1 deal in 3 years,” is that it’s not me being arrogant. It’s just the truth.

I lost to you.

Which is great. You know what I hate, by the way, when you lose? People do too. They phone up and they’re like, “Oh, it’s Sasha and Kian [?]. Do you know them?” “Yeah, out of Shopify. Amazing guys. Love them. Fantastic. Well done.”

That was Jeanette.

Yeah, awesome.

Great founders.

You know what I freaking hate, though? When other investors call up and they’re like, “Hey, can we share it and have some of yours?” You’re like, “Are you kidding me?”

No.

Yeah.

I never do that.

Yeah.

That one sticks in my mind as one that I’m annoyed about.

16. “What I Learned Losing the Series A of Snap, Stripe, Samsara”

First of all, Harry, you should be losing more. I’ll tell you a little story. When we moved to the Bay Area, I lost the Series A of a company called ClassDojo, where he did the Series B.

My partners from Boston came, and they just felt bad for me. They were like, “You can’t be losing, because that’s just going to emotionally devastate you, and then you’re not going to be able to compete here.” I was like, “Are you kidding me? If I’m not losing, I’m not winning, because the very best founders go meet all the 5 to 7 great firms, and they pick 1.”

Theoretically, your win rate, as long as it’s over 30%, means you’re actually maybe in the right fight. It’s very important to be in the right fight, so you actually want to feel like you’re losing more. Everyone on the team comes and tells me, “I haven’t lost in my career.” I’m like, “Well, then you’re just in the wrong pond, buddy.” That’s an important thing: there are so many smart people.

I just lost 1 the other day.

There you go. That’s my point.

That hurts too, but I was worried about saying that one.

I think we should enjoy the pain. I actually think if we’re losing, we’re winning. That’s something I genuinely believe about venture: you need to be in the right fights. Then you have to win your fair share and not dwell on it beyond that.

The things I lost—I have so many of these moments. Drew Houston, when he was starting Dropbox: I asked him to come work at GC with me because I had him intern for Chris Dixon back at Chris’s company. I said, “Come work with me.” He was like, “No, I’m going to go start this file-storage company.” I was like, “File storage?” That was the first $1 million. It would have been a $2 billion return. I was like, “What is this thing?” Even though I was willing to work with this guy, I didn’t give money to that.

But you didn’t lose it. That’s a miss.

Well, to my own brain—

That’s a miss. The one where you lost it.

I lost the Series A of Stripe. I lost the Series A of Samsara. I lost the Series A of Snap. The first one I won was the Series A of Gusto. If I look at my most competitive fights when I first got to the Valley, I’m like, “Oh my God, this just sucks. No one’s ever going to pick me. I’m getting out of Boston.” And I did. I lost a lot.

But the key was that I stayed on, and then I was able to do the next round. Those moments make you better. You just have to realize there are a lot of smart people in the industry. They’re very smart, and you’re going to lose.

I called up my mentor, who is one of the best investors in the world—a billionaire genius—and I called him up after losing this deal to what was likely Andreessen Horowitz and said, “Fucking hell. What the fuck?” He said, “Harry, Harry, I’ve never made any of my biggest returns when I won the deal. I always lost the deal.” Then I just had to scramble and buy secondaries from angels, operators, and founders. All of those were my best returns.

Yeah. Look, I think the thing is, in the very best companies, you don’t get the amount of ownership you want anyway, because founders command a premium. You’re constantly building ownership after that. As I mentioned, my biggest overall investment is in Stripe.

How much do you have in Stripe across everything?

It’s still sub-10%, right?

But like $2 billion, $1 billion?

No, no, more.

$5 billion?

More. But I think the key point is that it’s a lot. I think Stripe’s going to be a $1 trillion company. You just have to give it 10 years, you know.

It’s a compounding business. Patrick and John always say infrastructure is hard, but it also compounds. They’re steadily just doing that and making some really smart choices in this AI world. You have to have a long-term view. We’ll have a 25-year hold, probably, on Stripe in some form or the other.

When these companies are good, you want to keep buying in. At some point, you say, “Okay, I’ve done enough, and now the next 5x just seems extraordinary to think about. Way more beyond that—maybe I’m going to stop.” At some point, you have to say, “Okay, I’ve got to move on to the next thing.”

17. Quick-Fire Round

There are some firms that are trying to make all their money on SpaceX and keep buying SpaceX. That’s great, and I actually think Elon’s got a lot of runway with SpaceX. That’s a good investment, but I want to go after the next generation of entrepreneurs as well and figure out if we can generate alpha there too.

Listen, dude, I could talk to you all day. I need to do a quick-fire, because I’m sure you actually have someplace else to be in life. Tell me, what have you changed your mind on most in the last 12 months?

This idea of indexing—being in every company when you have macro trends. Should you index or not? I’m being open-minded to thinking that way about major technological trends or market shifts.

What has been the biggest challenge in changing your leadership?

For me, I’m very emotional, and as I lead, I need to dampen my emotions. I think it becomes really hard going from becoming a master at something to being a teacher at something. The reality is that when you can teach something, you’ve truly mastered it.

I don’t think I’m very good at that. Some of my partners will say gibberish comes out of your mouth when you try to teach. It’s much better to just watch what you’re doing and make sense out of it. I’m still trying to figure out how to crack that.

What would be your single biggest piece of advice to an LP navigating venture today?

The proposition for founders has to change, and you want to embrace entrepreneurial VCs that are innovating around that.

What worries you most in the world today?

The short-term misalignment of value creation in business with long-term prosperity, with everybody being inclusive and abundant in that.

What would you do if you weren’t scared? For me, if I wasn’t scared, I’d move to the Valley. I’d compete in the coliseum of greatness in technology, and I wouldn’t just be content being a big fish in a small pond in Europe. What would you do if you weren’t scared?

I’m not scared. I just don’t operate with fear. I’m doing what I would do. I think we take a lot of risk. We’re innovating in every dimension that we possibly can. We’re pushing ourselves as much as we can. I’d like to think I would do what I’m doing.

Does money make you happy?

No. Money is a byproduct of the impact I want to create.

What’s your biggest advice on parenting?

Teach them to be unique, and in the world of AI, teach them to ask questions, not solve problems.

Is college less valuable than ever?

I have a 16-year-old, and he’s definitely going to college. I have an 11-year-old, and I talked to him the other day and said, “Hey, AJ, you may not need to go to college. The world may change in how we think about developing skills.” He was happy about that.

Final one for you. What are you most excited about? I like to leave on a tone of positivity. What are you most excited for when you look forward?

Look, technology is neutral, and what I’m most excited about is that, over the next 20 years, if I look at GC, we’ll probably invest, what, $300 billion, $500 billion into the world, helping shape what AI does for society.

I, my partners, and my broader team have the opportunity to leave a mark, and I want to get it right. I want to get it right so that when I’m someday sitting in a senior living facility, I’m like, “Hey, I did right by the world. It’s actually turned out to be okay through the shift.”

Listen, Hemant, shows like this remind me why I love what I do so much. You’ve got to remember, I love investing. This is my true passion. Being able to speak with you and discuss the craft of what I love so much is such a joy. So, thank you for being so brilliant.

Thanks for having me. This was fun. I really enjoyed it.