Ben Horowitz on Investing in AI: AI Bubbles, Economic Impact, and VC Acceleration

0:00You know, if you want to change the world, you have to believe you can change the world. What you're really trying to find is, are they literally the best in the world at a thing? And that's always the thing that's worth investing in, as opposed to they're pretty good at a lot of things and I can't figure out what they're not good at. We just have a higher concentration of talent here than that's probably possible in a company in terms of just sheer IQ. There's a lot of VCs, very few who can actually help you succeed as a company. And so being one of those, I think, is still quite a special position.

0:39AI is changing how companies are built and how venture firms operate. And those changes are forcing new decisions about structure, judgment, and speed. Traditional venture models were designed for long timelines and slow feedback. In the current AI cycle, outcomes move faster, competition is denser, and firms have to evaluate people and ideas before results are visible. That shift raises practical questions about how to run a partnership, how to hold investors accountable, and how to stay close to the work without creating friction. In this episode, Ben Horowitz joins for an AMA to explain how A16Z approaches those problems.

1:11He discusses why managing a group of GPs is different from running a company, why the firm evaluates investors at the point of decision rather than waiting for portfolio outcomes, and how verticalization allows small investing teams to scale without internal politics. Ben also shares his current view of the AI cycle, including why AI should be treated as a new computing platform, why application design and model orchestration matter more than raw model size, how AI pressure is reopening M &A, and why the current market reflects unusually strong demand rather than pure valuation inflation. We hope you enjoy.

1:47So I'm going to start first with more of how you manage a firm and particularly lessons that you've learned over the years, obviously extrapolating lessons as a founder as well, and then how we think about running the firm on a day-to-day. So the first question I'll start off is from your book, but it's, I think, relevant to this conversation in part because when you wrote that in technology businesses, you rarely know everything up front, and we're sitting in this massive AI wave right now. It's still incredibly early, but the difference, of course, between a mediocre company and one that's magical is often the difference between letting people take creative risks and then holding them too tightly accountable.

2:23So there's many ways we could take that direction, this question, but maybe first start, how do you manage a group of GPs? And particularly, what's different about managing GPs versus a company? And what's the same? It's pretty different than a company in that with a company, like there are functions, there are things, there are very specific outputs that you're driving to. And then the people in the company, I would just say like, we just have a higher concentration of talent here than that's probably possible in a company in terms of just sheer IQ. So if you look at Chris Dixon and Martin Casados and Alex Rampell and so forth, I mean, these guys, one, they've all run companies.

3:04And then it would just be very hard to have that many people that high IQ on an executive staff. And so, you know, have somebody like Martine who is probably the best architect in networking software in the last 20 years plus like a really talented investor and so forth, I'm not giving him that much direction. I'm more kind of helping him understand like how the process of the conversation affects the process of investing and kind of how you work your way to the right answer, taking the right amount of risk. and the biggest mistake we make is we get too wrapped around the axle about some weakness that a company has as opposed to focusing on what they're great at and how great they are.

3:49So, you know, everybody is like kind of, you could talk yourself into as great at something, but what you're really trying to find is are they literally the best in the world at a thing? And that's always the thing that's worth investing in as opposed to they're pretty good at a lot of things and I can't figure out what they're not good at, that's generally a worse investment. So just kind of orienting around that and then kind of helping think through the platform, the personnel and how to deal with the conflicts and then how to close deals and that kind of thing. So it's very different, I would say.

4:25And then just, you know, like understanding when people run out of gas, like investing in technology, you really have to be in deep in the tech to be good at it. And I think that it's very possible as people get older, they get less into it sometimes. And so at that point, we got to make a change. But maybe just sticking on the topic of GPs, how do you also think about accountability? Because one question had come up over when to promote the right people, when to manage out the right people, and ultimately the vertical levels are making decisions. But you're also making decisions. You and Mark are making decisions at the firm level as well.

5:00What's your sort of thinking and framework? Has that evolved over the last 16 years? Yeah, I mean, I think it's evolved a little, but the main things are kind of the same. So I think it's dangerous to, in VC, kind of wait for the outputs because they're so far up to kind of wait and see, okay, does somebody have a great portfolio after 10 years or 15 years before kind of deciding what to do with them? it's just such a long time and you can make a lot of bad investments in that time frame where like you could miss out on a lot of good investments in that time frame if you don't put that person in position to do more stuff.

5:42And so I really try to look at, okay, at the point of attack, like how are they showing up? Like how good are they at finding opportunities? How good are they at winning those opportunities? and what do we think the quality is at the time of investment? Because some work out and some don't, but it's not all magic. Like, you kind of know, you know, how great an entrepreneur is Mira or how great an entrepreneur is Ilya. Well, like, those are pretty special people. So if you can win that deal, like, that means something, whether or not those companies work. Maybe just to focus in on verticalization, because verticalization, I think, was probably one of the seminal points in the history of the firm in changing the structure of it.

6:27And you and Mark have talked about this. And now in retrospect, it's clear that that was the right decision at that point in time. How do you avoid some of the pitfalls associated with verticalization, meaning the communications and strengthening communication across verticals and making sure you still have that connectivity as the firm continues to size and scale? Yeah, so look, I think that the most important observation, and this is actually a conversation Mark and I had with the late great Dave Swenson back in 2009. and what Dave said, which I thought was very interesting at the time, it's like an investing team shouldn't be too much bigger than a basketball team.

7:06Basketball team's five people who start. And the reason for that is the conversation around the investments really needs to be a conversation. And so I always had in mind that, okay, any investing team, we really don't want to be too much bigger than that size. And so how do we kind of maintain that? And the only way to do it is to verticalize. And then the other thing that was happening simultaneously in the industry was software was eating the world. So we had to get bigger in order to address the market. But I didn't want teams being bigger than a basketball team. And so that kind of led to the vertical structure.

7:49I think, in effect, the most important thing is that those teams should be good. And then the communication across the teams, we do in different ways. So one way is if the teams are very close, like AI, Infra, and AI apps, then we have people from each team going to the other team's meetings. And so there's like good kind of just hardcore connectivity that way. And then in addition, we've got like a kind of management meeting of that group in AD that we do. And then we have kind of the big thing is the GP offsite. So we just take everybody away for two or three days twice a year with not much agenda.

8:30I do think David Haber has this thesis of opportunity lies at the intersection. Like I think everyone not only culturally knows, but also economically is incentive to see everyone win. And so there isn't that level of politicking. I also think that sometimes can be prevalent at other organizations where it's very zero sum and like protecting fiefdom type of behavior as well. Yeah, that's kind of like a cultural idea that we have at the firm. And I would say, I mean, the feedback we get from people who come from other firms is we have less politics than firms with 10 or 11 people. And it's just, that's a cultural thing.

9:05Like either politicking gets rewarded and then you have everything from coups to infighting all the time and they don't like each other and so forth. Or politicking gets de-incentivized. And that's what we have here. Another point around culture, I always remark to you, like I'm always in awe of the fact that you always hear the gas around the firm, even detail around minor things, where I'm like, how do you know that, Ben? And Mark comments, and you often say, whenever you run something, it's gotta be in the details. That's the only way to do it. But maybe I'll ask the question slightly differently.

9:36How do you stay on top of the details, right? How do you hear all this? How do you find this subtle, but without also micromanaging? Because I think there's also the creative sort of process of letting things run its course as well? How do you balance those two things? Yeah, if you think about my job, a lot of it is kind of setting the direction and then making decisions when like things get into conflict or we're not sure what to do. And the kind of, if you think about, you know, what's decision making, it's, you know, what makes you good at it? It's a combination of intelligence and judgment, or it's judgment kind of, which is a combination of intelligence and knowledge.

10:17So like, what do you know and then how smart are you at kind of turning that into the correct judgment. And the knowledge in an organization tends to live with the people doing the work. So meaning, you know, not the managers, I would say. You know, like, so it's, you know, what are the deal partners doing? What are the, you know, kind of individual people on the IT team doing? You know, like what are the kind of accountants doing? You know, like what are people on, you know, when we go visit LPs, like, you know, that's where the knowledge is. Like in talking to people who are kind of at the point of attack, so to speak, or talking to entrepreneurs.

11:07and so I just spend, I try to spend enough time both, you know, kind of in the team meetings and, you know, we're doing that. I just end up knowing a lot about a lot of things and then plus I'm a founder so if something gets effed up in the firm, somebody calls me, you know, like that happens a lot. Like they're like, okay, I'm going to tell Ben, you know, like he's not going to like this. So this is, I think, a key thing for leaders is you never want people to think, oh, we shouldn't bother them with that because it didn't take me, but it took me like 14 seconds to resolve it. And, you know, generally people aren't looking for, you know, they're just looking for clarity.

11:51You know, like a lot of what an organization needs often is clarity, not like correctness. And, you know, if you have clarity, you can move. Yep, yep, yep, yep. So last question on verticals. So we have seven verticals today. One topic that has come up is how do we know these are the right verticals? And then how, maybe give some examples of vertical ideas that you've resisted to start, but maybe sound good on this premise, but doesn't either have the right technological legs or even entrepreneurial kind of capability around that we decided to not pursue. Yeah, so they're really kind of designed around the market.

12:33And, you know, kind of where are the entrepreneurs? So we try to match up to, you know, if there is a big cluster of important entrepreneurs that are going to create multi-billion dollar companies, you know, do we have a team that's going to win those deals? And, you know, different categories end up having fairly different needs. So, you know, the needs of a crypto entrepreneur, a bio entrepreneur, an American dynamism entrepreneur are very, very different. And so you kind of have to have a product that matches that market. You know, in picking markets, you want to not be too early and not be too late, right?

13:19So it's a little bit of an art. I think that what we've seen is I'm very confident those are the right markets because there is lots of very interesting activity in all of them. Now, we've got to then perform in each market. So, it's not like a given just because we show up and we're Andreessen Horowitz that we're going to win that market. But, you know, so we have to kind of evolve the team and like evolve our thinking and make sure that we win. But I think that the markets we're in are like pretty clearly very good markets. We have had, you know, a few that have popped up that, you know, people have proposed that we haven't done.

14:06I don't think we got that serious with them. But the other one was kind of ESG kinds of things, you know, clean tech, green energy, this and that and the other. And, you know, we thought that the right lens on that was much more going to be American dynamism one because it wasn't like weirdly constraining. And it was much more oriented around the kind of economic outcome as opposed to the do good, do good by doing well or whatever the fucking phrase is. you know like those things kind of can lead you into very weird decision making so like investing is hard enough without like introducing other criteria other than is this thing going to be a giant company and make a lot of money and you want to have a focus on that and I think the beauty of American Dynast like there's a lot of things that like would be you know, maybe good for America or whatever, but like there's so many opportunities in that space because the U.S.

15:15really does have to modernize the way they do defense. We really do have to get much better intelligence and public safety to keep everybody safer. That, you know, and then we've got, we've got to solve the energy problem. We've got to solve the rare earth mineral mining problem. Like there's very good problems to go dig into where, you know, if you look at like, okay, we're going to come up with an alternative energy source or an alternative fuel, you know, like that's like, will any of those work? Like maybe, you know, and that kind of thing. So we never did ESG. Yeah, I remember actually as you were going through that, I remember very early on in our discussion around American dynamism, I remember you pushing the team and asking them like, hey, is this a marketing message or is this like real technological transformative change?

16:08And like they went and did the work around it and like it was very clear now at retrospect that there was real tech change, especially happening on the supply chain side and the defense side and really how people actually even engaged with the government. But I think that oftentimes when people ask us this question, like it's both a combination of, is there a real technology change? Because that's when you make and generate venture returns. And then also is entrepreneurial talent actually there to build it? But yeah, AD is a good marketing idea. And I think when they presented it internally, they presented the marketing idea.

16:41And I was like, well, I want to know what the fund idea is. You know, like, how do I make money? Like, we have investors, we got to make money. Like, it's a great marketing story, but like, we're not doing all that. We're going to do the things that, you know, the fund is going to be less than the marketing in terms of its focus. It's going to be tighter focus. Yeah, and then we ultimately zoned in on three kind of core vertical areas that there was actually a tech change happening. Maybe so to get it. So Mark and I believe the best thing society can do for a person is give them a shot. Give them a shot at life, a chance to contribute, a chance to do something larger than themselves, and make the world a better place.

17:18That's the best we can do. So can you elaborate on this and how it's driving how we are evolving as a firm and particularly looking ahead as people think about, you know, what's going to be a set of funds that are deployed over the next two to three years, but also ultimately have an impact over the next 10 to 15? How do you think about that as you think about leading the firm? You know, I think it's important to put the work that people do into context. And, you know, we're in a like super special position. you know what I wanted to get at was like if you take a step all the way back and you say what's been good for humanity what's been good for humanity historically is when people have a chance to kind of do something larger than themselves and contribute and look and and I think there are there are many systems ideas like well what if like we could make utopia or everybody equal or this and that and the other.

18:12And, you know, that's kind of ended up doing the opposite. If you look at the history of communism or what have you, it's kind of everybody has an equal chance of getting no shot is much more what occurs. And so you really want to enable contribution and kind of the rise of America kind of coincides with that rise in kind of a free market, capitalistic, rule of law system. And that, you know, if you look at kind of the history of the country and the history of humanity, that, you know, the rise in wealth, lifespan, you know, population size of the earth all kind of grew spectacularly in the last 250 years.

18:59And so America has been very important in that. And then America today is still, I think, very clearly the country and the system where people are most likely to have a real shot at life. And, you know, like we've done some things to screw that up and so forth, but, you know, that's certainly still the case. And for America to maintain its importance in the world, it has to win economically. It has to win, which means it has to win technologically. It has to win militarily, which means it has to win technologically. And our job is to, you know, kind of help the country win technologically. And it's not only important for us, it's important for the country, it's important for humanity.

19:47And that's really kind of what I was driving at. And I think that, you know, for our people, it really kind of helps them go, okay, like this isn't, you know, these things matter. You know, creating these opportunities matter. And just to give you an idea of some of the things that leads to, so actually, you know, Jen and I were just in Mexico. And, you know, a lot of that was catalyzed by kind of a junior person on the team going, you know, what we're doing is so important and we need to kind of help with this alliance. We need to help secure the border. We need to help with our own kind of defense manufacturing.

20:34We've got to, you know, help with energy. Like, I'm going to get this meeting. And then, you know, we got the meeting. And so, you know, if you want to change the world, you have to believe you can change the world. And that's a lot of what it was about. It seems like little tech M &A is opening back up. What's your view on whether that is here to stay and whether that might actually extend to larger opportunities as well? AI is such a disruptive phenomenon that every company, every incumbent is under threat from like AI in general. And so a lot of the ways that you deal with a threat is you just acquire the DNA of the future.

21:19And so I think there's going to be a lot of M &A because I think that people need to reconstruct how they work if they're going to survive. So if you go back, say, three or four years, I think people believe that, you know, there would be the big foundation models would be these giant brains that could do anything better than anybody. It has not played out quite like that. The way it's played out is that the big models do provide a very important infrastructure that, you know, all of our companies, you know, end up building on to some extent. But often the kind of, for any particular use case, the long tail of not only kind of scenarios, but the long tail or the fat tail, I should say, of human behavior ends up itself being something that you have to model and understand very, very well.

22:31And so if you look at Cursor, Cursor consists, I think, of 13 different AI models, all which kind of model different aspects of how you program, how you speak to a programmer, et cetera, et cetera, et cetera. And those models end up being very important, and so important that they, in fact, released their own foundation model

23:00specifically for programming and for coding. So they have a coding model that, you know, you can swap in in place of Anthropic or OpenAI if you want, or you can use the OpenAI or Anthropic models with their kind of other set. And that thing has gotten great adoption. So, like, it's kind of going, And while maybe the application behavior is actually in some ways more important than having a gigantic, like the biggest model train with the most GPUs, you know, and it's not clear exactly how that plays out. But right now, currently, I would say that the complexity of the application itself is very high and is not subsumed in the foundation model.

23:43And so I think these things are not as straightforward as they appear. and the benchmarks can be misleading. And, you know, like I think that this also is showing up in every aspect of AI. So we've seen that there's a great post that Justine Moore from our team did on, you know, there's no God-level video model, which I would encourage you to read, which kind of gets into like, you know, look, different use cases end up needing different models, which is, again, not what we thought four years ago. For sure, yeah. That actually goes back to when I started this webinar with your quote from The Hard Thing About Hard Things.

24:27In technology businesses, you rarely know everything up front. So as folks are figuring out everything that's happening in AI, like the calibration also, benchmarking is changing. And also the expectations around the utility of those things are changing. And also the founders that are building that are changing. So it is a very fun time to say the least, but also one in which we're learning in real time as a part of it, which can be sometimes deeply unsettling, I think, for also folks as well who, you know, may be anxious about where valuations are and the market and environment as a part of that.

25:06Okay, I am going to switch gears actually to the future of VC investing, which is very related. So some believe leaner, more efficient businesses will allow founders to retain more on the cap table. Do you worry about that at all when it comes to ownership and what you've been, you know, working with GPs on in terms of expectations of ownership in this new environment? So I think that, like, what we've seen is we're getting pretty good ownership. So if you look at a lot of the recent investments where, you know, at 20 % or better, there are ones that we're not, you know, we don't get to that level of ownership.

25:43But those companies get so valuable so fast that it's been fine. And, you know, like there have always been special companies with very, very special founders, you know, at a moment in time where like, okay, you know, that is what it is. But, you know, for us, it's, you know, for a lot of the just like core infrastructure things, the core applications and so forth, like, you know, the ownership has been like pretty reasonable. Question around just the VC landscape today, because when you and Mark started the firm 16 years ago, you know, there weren't nearly as many VC firms. There's now 3 ,000 plus VC firms running around.

26:27There is still this rate limit of great entrepreneurial talent, but also increasingly there's more sources of capital flooding in, whether that be on the retail side, etc. How do you see the power dynamics between whether it be LP, GP, and founders also evolving as you think about the future of the firm and anything you're doing at the firm to prepare for that? I mean, I think it's still like very hard to build a company. And, you know, if you're just like an engineer, an AI researcher, and you're kind of, you invented something and you're jumping into the world, it's a very competitive world.

27:07You know, having a financial partner that can help you build a company. So, you know, does the initial valuation matter more of the partner? And I think that like, you know, most entrepreneurs who are smart realize it's a partner. And there aren't that many good, like there's a lot of VCs, very few who can actually help you succeed as a company. And so like being one of those, I think is still like quite a special position. So the one area that we're kind of, you know, you're upping our emphasis in is this kind of entrepreneur who's just like starting something and doesn't, you know, it doesn't quite yet qualify for VC money with our speed run accelerator.

27:55and you know one of the reasons why we're emphasizing that so much now is it is you know with the new tools it is possible to convert an idea into a product much more easily than it has been in the past so we just want to make sure that we keep a very very close eye on that and then and you know and then the brand is translating into that accelerator we're just getting a lot of talent in there and and so that that phenomenon we want to make sure that we're on top of. So building on earlier question about being multiple winners in AI, why did prior cycles of technology not play out the same way?