Ben Horowitz谈American Dynamism与AI的未来｜The a16z Show

Over 70% of people in China are optimistic about AI and less than 30% in America were optimistic about AI.

We have placed the largest bet in American history on the proposition that this country will win the next century of technology.

We won the Industrial Revolution and we did that cuz we had superior technology and here we are on the dawn of a new technological revolution, the AI revolution. Every interesting company that gets started is a technology company now. America does give everybody a chance and entrepreneurs can really count on that.

If America loses its technology edge and its dominance, then the entire world will lose as well. What makes the American system irreplaceable for the world at large?

Everybody's in the same game, but the rules of the game have changed. And what I mean by that is

We’ve had a lot of excellent discussions. We’ve talked a lot about policy, but we’re actually going to start this more on the venture capital and technology side. You recently shared a blog with us that the firm has raised over $15 billion in a new set of funds, the largest in a16z history. What do you think are the new obligations that come with that kind of scale? We are the largest venture capital firm in the world. How does that tie into the mission?

My old mentor, Andy Grove, said to me, “Ben, when you’re a leader in an industry, then the whole industry—the size of it, the ethics of it, the morality of it—kind of depends on you.” I remember thinking, when he said that to me—he was running Intel at the time—I thought, “Wow, I’m glad I’m never going to be the leader of an industry, because that’s a lot of responsibility.” But here I am. Here we are.

And I think for us, you have to take a step back and say, “Okay, for the technology industry, and particularly for new technology companies, what does it need to be?” You have to think about it on a large scale. The way we’ve been thinking about it is: if you go all the way back to humanity, what’s important for humanity? What advances humanity?

It really comes down to: do people have a chance to contribute? Because when people have a chance to contribute, they can have an impact, and then they can advance the state of things for everybody. There’s really been no country in the world that comes close to America at giving people a shot, giving them a chance to succeed.

It’s not a completely equal chance. Some people are born rich, some people are born poor, some people have better genetics, some people live in dysfunctional families, and so forth. But it’s a chance. In most countries, people don’t have a chance, or they have a very small chance, a very slim chance. That ends up being really important.

If you think about why America has the influence it has today—from a military standpoint, from an economic standpoint, from a cultural standpoint—it comes back to the fact that we won the Industrial Revolution. We did that because we had superior technology. Here we are on the dawn of a new technological revolution, the AI revolution. Are we going to win? Are we going to have that influence? Are we going to be able to give people a shot and continue to advance humanity? I think that’s an open question right now.

When I think about our role in the industry, it’s: what can we do to help America win technologically? That’s everything from what we invest in to how we integrate into the government and make it accessible to all the citizens, and integrate with our allies and so forth. It’s a big mission. It’s a lot of work. And it’s why we have American Dynamism and the conference and everything.

You have said that the American way—giving people a shot—means that if we lose this technology race, if America loses its technology edge and its dominance, the entire world will lose as well. What makes the American system irreplaceable? Why is our technological dominance not just good for America, but actually good for the world at large?

I think a lot of things go into it, but to me it always comes back to a line in the Declaration of Independence that is very well known, but also underrated: “We hold these truths to be self-evident.”

What does that mean? Well, it means that these truths didn’t come from us. They’re truths. They are self-evident. They came from a higher thing. They came from God. We didn’t make these rules. We didn’t give you these rights. We didn’t give you this freedom. It came from something else.

That’s really, really important, because it means that if I’m the president, if I’m Congress, no matter who I am, I can’t change it, no matter how much I want to change it. There are many who would like to change some things, who think they know better.

We’re going to talk about that.

Yeah, and that is the thing. But because of that, the freedoms that you have in America are really, really protected to a degree that they’re not in any other country, and they’re very hard to lose.

We’ve seen this with free speech, where countries are like, “Well, maybe that’s not such a good idea anymore.” But if it can’t be taken—if it didn’t come from people, it can’t be taken away from people. That’s been very, very powerful and very persistent, I think, over the entire history of the country.

It’s really unique. There aren’t other countries that have quite as strong a version of that.

Absolutely. It’s now been 4-plus years since we launched the American Dynamism practice. We’re on our second fund. You and Mark have been incredible supporters of this effort of investing in the national interest at the intersection of advanced software and hardware, rebuilding not just the defense industrial base but manufacturing and all the precursors that go along with it.

What’s been the most surprising thing you’ve learned about investing at the intersection of national security and venture capital? Along with it came a big presence in Washington. What do you think the biggest learning has been?

It’s been interesting because if you go back to pre-ChatGPT, the conventional wisdom was that China had this big AI lead on us. Then ChatGPT comes out and says, “Oh, well, maybe now we have a lead on them.” Then things will evolve from there.

But the thing that was true about the old, incorrect idea was that they were way ahead of us in integrating their AI technology with their government—on a military basis, on a bureaucracy basis, in all facets. When we started, we were coming from, I would say, very far behind in that area.

The thing that’s been surprising, though, is how fast we’ve been catching up. I think that’s been really heartening, both from the number of entrepreneurs who want to help and the willingness of the U.S. government to say, “Hey, we’re interested in these new technologies. We’re interested in these new companies. We’re not stuck in our ways. We’ll change the rules if we have to.”

That’s been amazingly fruitful, and I think we’re seeing it in the current conflicts that we have. It gives me great hope that if we keep at it, we’ll be very strong going forward.

I agree. I think this is a moment of optimism, to see just how quickly things can shift when there’s a sense of urgency and a sense of change-makers. I want to talk about AI and policy. We had Emil Michael earlier, the Under Secretary of War for research and development and CTO of the Department of War, here. He was obviously in the news talking about Anthropic over the last week.

But this isn’t a new issue. It may have bubbled up to the mainstream news in the last week, but this has been an ongoing issue. Let’s talk about Anthropic and AI directly, head-on, right now. Then we’ll talk more broadly about policy in Washington.

To me, the thing about the whole Anthropic kerfuffle that nobody’s really said out loud, but is the obvious thing that happened, is that if you understand deals, particularly software deals, you realize that that deal did not fall apart because of philosophical differences. It fell apart because Anthropic wanted out of the deal.

You know that because they had all the leverage. They were already deployed. We were about to go to war. Nobody’s ever had more leverage than that in a software deal. So anything reasonable that they would have asked for, they would have gotten—probably even beyond what they should have.

But the fact that they found something they could walk away on and then didn’t return Emil’s calls—for whatever, the 3 calls, and then the call to the administration, or all those things that happened—it’s just: why would you not return the call if you wanted in the deal?

They just wanted out of the deal. I don’t know why they wanted out of the deal, whether it was an employee thing or Dario’s thing or what have you, but that’s what happened. I think the fact that it’s getting reported on like there was some ethical conflict...

I don’t think there is. In fact, if you wanted somebody to use your AI safely, nobody has more rules than the US government, particularly the Department of War. And by the way, since I’m speaking to a Washington crowd, you know this: if you were to break those rules, there’s a 100% chance it’s going to get leaked to the press immediately. There are no secrets in Washington. So the whole thing was a little bit surreal for me to watch, at least at this point.

But what do you think? Is there anything that founders should be thinking about—people who want to sell to the government or work with the government? How should they approach it inside their own companies? Maybe Anthropic is a one-off—

Yeah.

But you and I have been in Silicon Valley long enough to know that these issues tend not to come up and go away. They tend to stick around. What would you say to the founders who want to build for the national interest but are still coming from a Silicon Valley mindset? What do they need to think about? How should they talk to their teams? How can they build the products that are needed for our government to effectively have the best and most capable technologies?

Yeah. It’s interesting. When I spoke to the firm about it the first time, when we started American Diamond Minerals and so forth, I said, “Look, you can have whatever geopolitical views you want to have, but the world is what it is. It’s not as if everybody has agreed on peace worldwide, and it’s not as if everybody has agreed never to commit a crime. We do have crime, we do have war, and we have these things.”

Then you have to think about the people we grew up with who have gone into the military and risked their lives to save our lives. Are you making the call because you’re ethically superior and know more about geopolitics than the people in the State Department, the people running the Department of War, and the people running the intelligence agencies? Are you going to decide that the people who are sacrificing their lives, putting their lives on the line to protect us, don’t get the best technology? I’m not doing that. I’m not doing that.

You don’t have to work here if you want to do that. That would be crazy—to think you know better than the State Department.

Yes.

Yeah, I mean, we have this issue. As we’re board members of these companies, the founder or CEO will call us and say, “I have some employees who don’t think we should do business with this country or that country.” And it’s like, “Are you the State Department? Do you know better?” That’s crazy. And by the way, what a slippery slope to go down. You don’t want to be in that position.

Yeah, these issues are complicated. They really are complicated. I think that this dime-store morality that companies come up with—it’s like vibe geopolitics.

Right, that’s right.

Okay, so we call it American Dynamism, but it is more than America. It’s about America and her allies. You have spent a lot of time—I know this because sometimes I join you, but I also get a little bit of visibility into your travel schedule—you’ve spent a lot of time on the road internationally.

A lot of foreign leaders come to you and say, “We want a startup culture. We want American Dynamism. We want to reinvigorate a startup community, do manufacturing, and build for defense.” How do you think about exporting our capabilities, our advantages, and our culture to these other countries, especially ones that we may have supply-chain dependencies on and that are our allies—particularly ones connected to us, like Mexico, Canada, and Latin America? How can we take American Dynamism abroad?

It’s an interesting question. I don’t have a full answer because we haven’t entirely figured it out, but I’ll say this: America gives everybody a chance, and entrepreneurs can really count on that. If you’re going to bet your whole life on an idea and build something, you really have to believe that the government isn’t just going to arbitrarily take it away from you at any point. There are very few countries that have that going for them. So we are reasonably unique.

There are a couple of places that have great technology entrepreneurship. Sweden is one, and Israel is one. But it’s pretty rare around the world. Having said that, there is great expertise in things that we need help with, particularly when it comes to industrializing.

In Mexico, they have really good, high-quality manufacturing expertise. In some ways, I think the American cars that come out of Mexico are better than Chinese manufacturing on many fronts. So they do have talent, they are partners, and they want to work with us. We need to help secure that country because it’s a border, but it’s not a thick border—it’s a thin border. We have a lot of people vacationing down there and whatnot, so that’s a real thing.

Similarly, Japan has always loved robots, and they’re really good at manufacturing. We have a big deficit in the robot supply chain, so we’re working with them. At the same time, they just went from 0% of GDP to 3% of GDP spent on defense. We have a lot of defense companies, and their interests and our interests are very aligned when it comes to China. I think there are some great opportunities, both for us as a firm and for our companies, and then for the country in working with allies around the world.

Yeah, the recent advancements in robotics—we haven’t spent a lot of time talking about them today. I’m sure we will next year, but it really feels like we’re on the precipice of a robotics revolution, and it’s going to be very exciting to see that play out. For sure, a lot of our companies and a lot of our allies will have a role to play.

In a recent deep dive, Packy McCormick did this long deep dive on the firm. I thought it came out pretty well. He described the firm as a power broker using capital and networks to shape markets and influence. How do you view that and balance that power with the need to remain authentic to your founders and to institutions? How do we leverage that power? Obviously, we’re here in DC. How do you think about that responsibility, and how do we use it to advance the state of the art?

What you lack as an entrepreneur is power. You may have a great invention, or you may have built a great technology, but you don’t have the power to call anybody. You don’t necessarily have the power to get the right meeting with the right people in Congress to help get the regulation correct. You don’t have the power to get an audience with the right CEO to buy your product, and that kind of thing.

Power is a feature of our offering; that’s the way I think about it. That’s very different from the culture that we want to have and project internally. As you know, our very first cultural idea is “first-class business and only in a first-class way.” What that means when it comes to entrepreneurs is the ultimate respect for what they’re doing. It doesn’t mean we’re always their best friend or anything like that, but we show up on time, we get back to them in a timely fashion, and we’re honest. We do all the little things to make sure that our overall behavior isn’t, “We’re the powerful and you’re the weak,” because that would obviously be a short ride for our firm at the top.

What do you think about people watching us raise so much capital? It happened during a period when the broader VC industry had slowed down fundraising, but we closed very quickly and raised a massive amount of capital. Do you think that signals anything about where venture is going as an ecosystem or an industry? What do you think venture capital looks like today?

This is something that we’ve seen coming for a long time and that we prepared for. If you go back to when we started the firm, there was a study that showed that in any given year, approximately 15 companies would be started that would ever achieve $100 million a year in revenue. The whole idea of venture capital was that you had to get into as many of those 15 as possible, and you really didn’t want to be in anything that wasn’t one of those.

The way venture capital firms were constructed was as these partnerships, in both economics and control, that were small. Why have 100 or 500 or 600 people like we do now if you’re only going after 15 companies? It doesn’t make sense.

But what we saw back then was that, even though that had been true for decades, it wasn’t going to be true anymore because, as Mark wrote—my partner, your partner—“software is eating the world.” What that meant was that there was going to be no industry, no company, that wasn’t going to be better if there was a great software team at its core.

And that’s even evolving now into AI. Basically, every interesting company that gets started is a technology company now. So there are many more interesting companies, and in order to reach them, you had to be able to scale your organization.

Here’s the problem: If you were one of those venture capital firms that started all those years ago, you were likely set up, as I said, as a partnership with shared economics and shared control. If you have shared control, it’s very hard to reorganize. Because, as a CEO, you know this, David: to reorganize, you need a single decision-maker.

There’s no way to reorganize if everybody gets a vote. It’s going to be bad. I could write a whole book just on reorgs, but reorgs are essentially redistributions of power. When you redistribute power, people are mad. If they get a vote, they’re going to foul that reorganization. You can’t scale without reorganizing.

As a result, we’ve been able to scale because we have centralized control. But most of our competitors have not been able to scale. Very few have. So I think what we have now is large firms that can cover all the technological areas, and then specialized firms that say, “I’m a specialist in AI infrastructure. I’m a specialist in bio, crypto, or games, but I’m not doing all of that.” I think the firms in the middle are getting squeezed out.

I think that’s right. People sometimes ask me, “How do you work at such a massive firm—600-plus people?” And I say, “Look, we’re actually a bunch of really small teams that leverage a bunch of shared platform services, much like portfolio companies. That allows us to move with urgency while having the might of a very, very large organization behind us.”

I think one of the other things that’s interesting, and people don’t always realize about the firm, is that there are probably more CEOs as general partners inside the firm than at any other firm. And I think all of us recognize having a hierarchy and a benevolent dictatorship is really appealing. You don’t

Yeah, it works. [laughter]

Yeah, it works. It makes our lives easy. My job is to focus on American Dynamism. It’s very simple. Martin’s job is to focus on infrastructure and be the best he can be—to figure out what we need. That dynamic is something that these death-by-committee venture firms just cannot adapt to.

No, definitely. You’ve talked on a recent podcast about how the media landscape has shifted and how reputation can shape markets. We had Alex Karp here earlier today, and he’s phenomenal at going directly to his customers.

How do you think about that shift? It seems like we’re seeing that shift in some ways in the government, with Emil Michael going directly on X. How would you coach founders—and maybe even people in the government—where narrative and trust are paramount, in terms of communicating and thinking about media and communications?

I think the media world has changed, and it can be confusing because you don’t necessarily need to think about it in terms of old media and new media. Everybody’s in the same game, but the rules of the game have changed.

In the old world, the key to a media strategy was defense. The reason defense was key is that there were a limited number of channels, and those channels had very strict formats. If you’re in The New York Times or The Wall Street Journal, you can get a few quotes in, and then you’ll be characterized based on the rest of what you say. If you’re on CNN, you get a short and hostile conversation with a bright light in your face, and so forth.

Because there was a limited number of channels, if you said something wrong, you couldn’t erase it. It was kind of like being on the permanent record. There are many cases of these old-school gaffes, like Howard Dean and so forth, where you can never erase it and never get over it.

In the new world, there are unlimited channels and unlimited formats. The key to winning isn’t not making a mistake; it’s being interesting. If you don’t say anything interesting, it’s not going to get any pickup because there are too many other interesting things out there.

People like Alex Karp and Donald Trump are just an entertaining show. Alex was super entertaining. I was captivated. I didn’t know what he was going to say next. I don’t know if he knew what he was going to say next, but he was really entertaining, and that wins.

If you make a mistake, you just do 10 podcasts tomorrow and flood the zone. It doesn’t matter whether the channel is The Wall Street Journal, a podcast, or whatever. It’s all the same thing now because that’s the way you’re overwhelmed with media. To win, you have to be interesting.

That doesn’t mean you can just be all over the place and not be on message. Alex is always very, very, very consistently pro-America. That’s what you can always count on him to be, and that’s his main message. That’s great, and he’s interesting about it. He has a great knowledge of history and all these kinds of things.

Awesome. We have placed the largest bet in American history on the proposition that this country will win the next century of technology. I do feel very optimistic. I think you and I are both optimistic. I think our firm is broadly very optimistic. In fact, I would say that’s ultimately a core value of the firm: being optimistic about the future.

Yeah.

But what keeps you up at night? What worries you? What would you ask of the audience, some of whom are founders, some of whom are policymakers, and some of whom are implementers? What worries you and keeps you up at night, and what can we do to help?

Actually, my biggest worry is the perception of technology in America. There was an interesting poll that came out showing that, I think, over 70% of people in China are optimistic about AI, while less than 30% in America, or something like that, were optimistic about AI.

It’s really because we always talk about all the bad things. We’re very focused on the dangers of AI here and not on the positive things. I just spoke to someone from Japan who said, “The whole Japanese startup ecosystem is restarting because everybody’s so fired up about AI. They’re all very positive on it. We love robots.” The Japanese love robots, by the way.

I think we really need to focus on that. Somebody asked me the other day, “What good can AI do?” And I was like, “We can—we’re going to—not just we can, we’re going to end traffic deaths. We’re going to cure cancer. We’re going to end poverty as we know it.” Those are pretty good things.

We need to think about that as much as we think about the AI overlords, mass surveillance, or any of these other things that we worry about. The positive uses of the technology are extremely positive. The negative ones are there, but we can manage them in the same way we’ve managed the negative aspects of every technology, starting with fire, which has very bad consequences. It can burn down your village. It’s horrible. But it also heats the house, cooks your food, and does some very wonderful things. I think we have to get back to that.