现代战争的现状：Palantir 与 Anduril 高管谈无人机、AI与传统战争的终结

Trae Stephens, Shyam Sankar, welcome to the All-in podcast at the Hill and Valley Forum. Thank you guys for being here. What's up? How are you guys doing?

Thanks for having us. Doing great.

Great to be here, too.

You guys are friends, right? You go back a long time?

A very long time.

Okay, tell us how you guys know each other. Palantir, Anduril—what's the connection and the history?

Well, I'll start. You can fill in all the gaps. I feel like I know the story you're going to tell, and it's going to be very uncomfortable.

Every few years, a new ad channel opens before the market catches on. That's axon.ai right now. The AI ad platform behind one of the biggest runs in tech with access to over a billion daily active users. Full-screen video ads in mobile games watched for a median of 35 seconds. Businesses are profitably spending hundreds of thousands of dollars a day on it, and most advertisers don't even know it exists yet. The window is open at axon.ai/allin.

So, Trae, I think in the early days of Palantir, I was roaming around giving demos to anyone who could possibly want to see them. Trae was working in an intel agency and happened to see one of these demos. He should tell you his version of it, his side of it, and his frustration with bureaucracy.

I think he realized, “Hey, this might be really cool. Maybe I should leave the hellhole I'm in, in the basement of this building, getting nothing done, talking about sports with other people, to go join this crusade.”

Trae reached out and applied. He made a big faux pas. He came all the way to Palo Alto, and he wore a full-on suit, tie, cuff links—CIA cuff links, mind you—to interview with, I don't know, probably 20 people who wore T-shirts and second-hand, hand-me-down clothes.

He was intercepted in the lobby by a receptionist who really cared about him and told him to ditch the tie, try to dress down, and not screw it up too badly. But we loved him immediately, and he helped us build the government business.

So, this was at Palantir. You were employee 13, 2006, right? This is pretty early on.

Yeah, I came in in early 2008, but there were still 25 or 30 people at that point. Pretty small.

And Peter kind of incubated it and was involved at the beginning, right? Peter Thiel. Maybe you could just recap—I know a lot of folks know the history of Palantir—but the early stand-up of Palantir and how things got going during that era with that small group, and how you guys figured out how to build the business.

Yeah, it was one of these things that was kind of a slow start. There was a real idea among the 5 co-founders, including Peter, that it's kind of insane to live in a world post-9/11 where people are arguing about what's more important, privacy or security. Aren't they both really important?

Who is actually spending time pushing out the efficient frontier? For any given amount of security, you should have more privacy than you had before. Or, for any given amount of privacy, you should have more security. This sort of changing the dialectic was really the entire impetus of what we started with.

There is a technical approach that follows from that. There's an approach to privacy and civil liberties around that. But really, we started as a business that was pretty myopically focused on solving a handful of problems in counterterrorism for a handful of institutions in the world.

Let me start by asking a question that I think is important to ask, particularly for a broad audience, and for the two of you to frame your personal philosophical views. Is war good? There's a lot of conversation about whether there is a military-industrial complex that has an incentive for war. What's your view, your philosophical motivation for why you do what you do, and what your view is of war and defense and the work that your businesses have pursued?

Well, anyone who's been to war would tell you that war is awful. War is bad—categorically bad. That doesn't mean it's always avoidable. There will be people who want to use might to make right, to define a set of rules, and you have to be in a position to protect your people and your interests accordingly.

At the end of the day, it's all about deterrence. You don't want to go to war, but you want to be prepared so that, if you do have to go to war, you will win decisively and quickly.

I've never met a general who has said, “You know what I really want to do today? I want to make phone calls letting parents know that their children have died in combat.” Nobody wants to do that. I think that's really the goal of everything we were working on at Palantir. That's what we're working on at Anduril: make it unthinkable to your adversaries that they should ever challenge you.

Why do you think it became taboo, and became so negative, particularly in Silicon Valley, that building a defense company or a company building technology to serve the defense sector was viewed with such disdain and was an untouchable sector for so long?

I think it's a beautiful consequence—and I mean it unironically—of the peaceful world that we lived in. The origin story of Silicon Valley is actually defense. Lockheed was the largest employer in Silicon Valley in the 1950s. The Corona spy satellites were built there.

So, in a world gripped by the threat of the Soviet Union, you had a very different posture in Silicon Valley. But in a world after the end of the Cold War—the end of history, and a view of globalism that it's all going to be great, that we exist beyond the confines of our country—these things were viewed much more cynically, and the threats didn't seem very real.

I think you could even back-test this to the moment when Silicon Valley really woke up and realized, “Hey, maybe these things are real”: when the Russian tanks rolled across the Ukrainian border. That's when they realized—and there are a lot of Ukrainians and Eastern Europeans, a lot of people who were affected by that in Silicon Valley—maybe there's a more nuanced issue here than just the simple Eisenhower quote about the military-industrial complex.

I think many of the people who were protesting Silicon Valley's involvement in working on national security priorities were the exact same people who had a Ukraine flag in their bio. So, there's clearly a policy mismatch or an understanding mismatch.

The other thing that I would mention is that these large tech companies, unlike many epochs prior, were global technology companies. They didn't necessarily view themselves as American. If you go and look at where these protests came from, you have to keep in mind that a lot of the signatures on these protests were not coming from U.S. citizens.

So, there was a global character to these companies that probably was not the case during the Cold War.

Given that that's changed recently, I mean, to Shyam's point, I think that there's just an increased awareness of the complexity of the geopolitical situation. I think it's certainly a lot less controversial to work on these topics than it was in 2017 and 2018.

It's a 20-year overnight success, right? Anduril is now reportedly raising money at a $60 billion valuation and just landed a $20 billion Army contract. Palantir today is worth $400 billion.

The flip side of what I just said is that some people are saying that Silicon Valley is taking over defense, and Silicon Valley is the next story of the American war machine. This has become a popular narrative. Can you respond a little bit to where we were coming out of World War II from a defense-industry perspective and where we find ourselves today, when Silicon Valley seems to be at the center of this?

The industrial base that won World War II and the early Cold War was not a defense industrial base. It was an American industrial base. Chrysler made the Minuteman ICBM. They were the prime contractor on it, and they made minivans. So, missiles and minivans.

General Mills, the cereal company, had a mechanics division. Everything they learned doing R&D to process grains, they used to build torpedoes and inertial guidance systems. Ford built satellites until 1990.

The entire economy was invested not only in economic prosperity, but also in underwriting the freedom that allowed us to have economic prosperity.

It's really a consequence of the end of the Cold War. When the Berlin Wall still stood in 1989, only 6% of spending on major weapon systems went to pure-play defense specialists. Ninety-four percent went to what I call dual-use companies. A missile is single-use; it's not a dual-use product. You're not going to buy it at Walmart. But these companies were invested in both parts of this.

That figure today is reversed: 94% goes to defense specialists, and 6% goes to dual-purpose companies. So, we have a very different structure of the U.S. economy as a result.

I think it leads to very perverse narratives. What would it be like to mobilize? If things got really bad, we'd just flip a switch, and our auto factories would magically enable us to provide for our defense and security. A cold-eyed look at history shows that it took 18 months to do that.

We actually started ramping up to provide this stuff for our allies—the Brits and the Soviets—in the context of World War II. We've missed some of those signals today. When Ukraine went through 10 years of production in 10 weeks of fighting, that probably should have been a five-alarm fire that we got the fundamental calculus on deterrence wrong. We thought the stockpile was going to deter our adversaries.

It was always the factory. It was the ability to generate and regenerate the stockpile.

Well, let's look at where we are today. In terms of defense readiness relative to our adversaries, there's a 10,000-to-1 drone production gap versus China and a 223× shipbuilding-capacity disadvantage. And I think you've set a 2027 Taiwan window of danger. At a high level, we spend more than any other nation on defense. How ready are we, and how long will it take us?

Our joint force is the best in the world. I think we have to couch the alarmism and commentary in that context here. But if you look at the rate of change, our adversaries are moving very quickly.

If you just look at the empirical loss of deterrence, you had the annexation of Crimea in 2014, the militarization of the Spratly Islands in 2015, and Iran with breakout capability to get the bomb in 2017. You've had a pogrom in Israel. You have the Houthis holding trade hostage in the Red Sea, not to mention the present conflict. So when you look at that, how can you not feel that deterrence is eroding?

To Trae's earlier point, we need capabilities that are so scary that picking a fight isn't worth it. People are investing significantly on the low end of the mix. Our high end is still unquestionably amazing, but you can't keep shooting $2 million interceptors at $20,000 drones and have that math work for very long.

I think the second part—why does it matter that we lost the American industrial base and have a defense industrial base—is that you lose volume. You lose the R&D stimulus to come up with creative ideas, like using the methodology of building a bathtub to build the next-generation low-cost cruise missile. And you get stuck in these platforms that are absolutely eye-wateringly amazing but are $2 million a pop.

You just can't produce them at the scale or with the speed that you really need to. And, Trae, you've talked about this. You wrote an article, “No Solvency, No Security.” Just share with us your view on the importance of this industrial base and what it takes for us to have a strong defense, which requires a strong manufacturing industrial base.

We really sent most of these capabilities away during the last 30 years of globalization, and that gutted entire communities in the United States. I mentioned in this article that, just in my immediate family—grandparents, aunts, uncles—I have family members who worked at GM, Ford, Frigidaire, National Cash Register, and Armco Steel. Every single one of these factories in Ohio closed. Not a single one still exists. They all have different places they've gone around the world.

One of the things we're doing with Anduril right now is building out a 5-million-square-foot factory campus in Columbus, Ohio. We have the benefit of going back and tapping into that knowledge base that exists and is underemployed at the moment because the factories that kept these people busy for many years are closed.

I think this is the story all over America. If you think about at-scale manufacturing for a new company started in this century, in the 2000s, there's really only one that comes to mind: Tesla. That's it. So we haven't really built any muscle around new manufacturing capacity.

As you spin things up, to Shyam's point, we don't have the ability to do that quickly because we've atrophied massively by turning all this away globally. I think readiness is a real problem, and we have to start investing ahead of the conflict being there, because you can't just start at the moment that it's absolutely needed.

Tell us about Arsenal-1 and the manufacturing behind it. Then I want to understand a little bit about whether this is product-led or production-capacity-led. Do we need to make selections and have capital committed to bringing a product to market before we start to build out the base, or is there an alternative way to get ready?

Arsenal-1 is the factory campus that we're building in Columbus. The operating system for that—the Arsenal platform—is really intended to reduce the cost and overhead required to automate and handle these processes in the most efficient way possible. There's a software layer that sits behind that. We're actually working with Palantir's Foundry platform on some of these programs as well.

The idea behind this is that you want it to be as modular as possible. If we build out the factory campus and say, “We're just going to build Furies here, or we're just going to build Roadrunners here, or we're just going to build Barracudas here,” that provides a tremendous limitation in a moment of conflict, when we need to be responsive to the demand for specific things that are relevant to that conflict.

We're thinking about this more like contract manufacturers think about building assembly capacity. They say, “We make this VR headset for Facebook. We make this VR headset for Samsung. We make this VR headset for Apple.” We build a skill set around building optical systems, but we're going to build a bunch of them, and we're going to be much more effective at doing that by hitting economies of scale. You get lower costs on all the individual parts, and you're going to build out a supply chain to work through that.

That's roughly what we're doing in this case. We're saying, like a contract manufacturer, that we want to be able to pivot on a dime into ramping up production of Roadrunners if we need Roadrunners, or ramping up production of Barracudas if we need Barracudas.

The counterexample to this, of course, is what we saw in the early days of Ukraine, where what they really wanted was more Stingers and Javelins. The problem is that once we burned through our inventory in the warehouse of Stingers and Javelins, the assembly line to build Stingers and Javelins didn't exist. All the people who worked on those assembly lines were retired. The primes were literally calling people out of retirement to come back and teach them how to build Stingers and Javelins again.

These are problems that we're trying to avoid through the design of the factory as an initial concept.

Who funds it? Is there a government contract that supports the production of these facilities? Are multiple companies going to be able to afford to stand these up and make these investments, or is this just going to end up being Anduril because you've got most of the capital and you guys are going to roll these facilities out? How do we make this a national-interest effort?

Our business model is definitely very different from the primes, which are basically responding to requirements directly from their customer. They're not really investing a whole lot ahead of the capability. Anduril, on the other hand, is doing all of this as private R&D investment and then selling the outcome—the output of that—as a product. So this is fundamentally a different business model.

I do think that there's a tremendous amount of capital required to pull this off as a new entrant to the space. I don't think that the market is going to support 100 new primes or something like that. I'm hopeful that it's not just Anduril. I think it would be healthy if there were more competition in the space.

Being able to raise that amount of capital and hire the people that you need to actually pull it off—this is not like building a normal tech company. It's not just the development of a product. There's a lot that goes into this.

This point might seem obvious or simple, particularly to an audience that's largely coming from tech or Silicon Valley, but that's not how defense works. People don't build products. People build to a spec that the government says, “This is what I want to buy,” and then you say, “Yes.”

You have to think about defense as a monopsony, where there's a single buyer for the thing. That concentrates an enormous amount of power in the buyer, whether they're right or wrong. If you look at the history of defense innovation, typically the monopsony is wrong.

It was Churchill, as the head of the Royal Navy, who built the tank because the British Army was not smart enough to realize that having horses in the next battle was not going to work. Every one of these innovations is an act of heresy. There's a founder-like figure who is so committed, almost pathologically, to a different heretical concept that they see it through. Only in combat, or only when it meets the moment, is it actually validated.

A grave threat like the Soviet Union is a sort of forcing function that allows you to innovate. When we were building ICBMs, we had 8 competing programs. Today, in peacetime, that would be viewed as wasteful. Why should you have 8 competing things? Why can't you just pick the right one and spend less, or put all your wood behind one arrow?

Of course, that kind of turns its back on our fundamental belief in America and the free market: that there's a competition of ideas, there's fundamental uncertainty, there's variance in the quality of the execution, and that you need this competition. It's not just competition among the primes or new entrants; it's also competition inside of government.

At one point, there were 51 major defense contractors. Just for the audience to understand, they were reduced to roughly 5 or 6 primes. That's the term “prime.” It's one of these prime contractors, and then they subcontract out to other companies to develop component parts.

My understanding of Anduril was that it was founded to do exactly this: build a product, show up with something that's cheaper, better, faster, and actually compete on the merits rather than be caught in this kind of prime landscape.

And that worked clearly. But over time, there seems to be a lot of capital concentrating into the new primes: you guys, SpaceX, OpenAI, maybe 1 or 2 others. Is the landscape going to emerge where we're going to just look like the same old way of operating, where we have a handful of primes that all have these trusted relationships with the government agencies? Or how do we create that competitive environment to continue to drive innovation and make things affordable for the US government and the taxpayer going forward?

Well, just like venture—and I interview Trae in my new book on this—there’s going to be a power law here. One of the grave mistakes we’ve made as we think about innovation, which has resulted in essentially innovation theater, is this idea that you’re going to just peanut-butter spread around the capital you have for innovation. Every company is going to get roughly the same amount. It’s not enough to hit scale, and it doesn’t reflect the relative performance differences, because there’s an authentic power-law curve, as any venture capitalist would realize: your biggest winners are going to return your whole fund. There’s something authentic here as well, which is, yes, you may have 10 bets, but at some point you’re going to have to get smart about concentrating down on the things that are actually working.

Yeah, I mean, this is a great point about any category. If you’re a space tech investor and you didn’t invest in SpaceX, you probably lost money. If you’re a crypto infrastructure investor and you didn’t invest in Coinbase, you probably lost money. If you’re a social media investor and you didn’t invest in Facebook, you probably lost money. And yet, for some reason, capital allocators have a very short memory. They don’t have the ability to go back to the prior boom cycle and say, “Oh, wait. I remember what happened that last time. We had to concentrate capital down to the winner.” I think this isn’t any different.

What was the motivator, then, for change in the construction of the landscape of defense contracting? Was it software that took us from the old primes to the new primes? Is that really what triggered this cascade evolution?

So, in 1993, having won the Cold War in 1991—or, I think more accurately, the Soviets lost the Cold War—by 1993, we expected as a nation a kind of peace dividend. We didn’t have an adversary anymore, so we should be able to spend less on defense. The department had this famous dinner where they brought 15 of the 51 primes together and said, “This is going to happen. The budget is going to get slashed. We’re not going to save you guys. You have our permission to consolidate. Some of you are going to go out of business. Some of you should try to make a commercial business,” which didn’t really work. That’s what led down to 5. Actually, there was going to be—we were going to go from 5 to 4 in 1999, but the Justice Department put its foot down and said, “We’re not going to let Lockheed and Northrop consolidate.”

And then today, the changeover: what gave you guys the window to build the business that you built? What did you see early on, 20 years ago, and continue forward on the hardware side that made you say, “This is the moment”? Was it software that enabled this?

Well, I’m going to throw you under the bus a little bit on this. It was not obvious in the beginning that this was going to work. I can’t tell you how many meetings we even had together where we’d walk out and be like, “This is not going to work. This is very bad.” We were not welcomed with open arms.

Right. Early days at Palantir.

Yeah, early days, mid-days, maybe even late days.

Yeah.

Famously, we had to sue our customer just for the right to compete. That’s the strength of the monopsony, right? And really, I’d say our entire business was validated from the field backwards. It was in DC that the doors were closed to us—the area where the monopsony is strongest. At the margin, in the field, people were saying, “Well, I’m on this deployment. I’m on this rotation, and there is a more free-market system. What I’m being given doesn’t work. I’d like to come home. How do I bend the rules and figure out how to get the software I need?” That allowed us to empirically show—create facts on the ground—that this stuff worked. It was a long row to hoe. I think this is part of why it took 20 years to get to this sort of point here.

I think that a lot of those lessons that were learned were not only educational for Anduril, but they also created a precedent that already existed by the time we showed up, where we didn’t have to go through all of those same growing pains. What took Palantir probably about 5 years to get to $10 million annually in revenue, we did that in 22 months. But I don’t think that was a credit to Anduril. I think it was mostly a credit to me, Matt, and Brian having lived the 5 years of Palantir.

5 years plus 22 months.

Yeah, exactly. So we kind of had a bit of a cheat code, but I don’t think it’s culturally shifted to the point where it’s actually just easy to do this now. It’s not like the government has fixed all of its problems. I think it’s just that there are enough people out there who have seen it and understand how to make it work. You can work within the system rather than, as Shyam said, trying to constantly fight from the outside in.

Was there a view at that time, in these years building Palantir, where you saw an opportunity in hardware? Is that the motivation for Anduril—that we’ve got software, but there’s more to do, and hardware could be reinvented and we could build systems? Help us understand the connection.

Yeah, in the very early days, actually, I came by to pitch Shyam before we officially started the company. Part of that pitch was, “We know how hard software is. We don’t want to do that again. We believe hardware would be less hard.” And I think we ended up being right from a getting-going perspective.

But the government doesn’t know how to think about the value that’s created by software. Maybe you feel like they’re a little bit better now, but it was really hard in the early days of Palantir to convince them to think about it as anything other than lines of code, for example, or some metrics that anyone who works in tech knows don’t matter. But for hardware, they know there’s a bill of materials. In fact, the government made stuff for a long time. During the Cold War, there were literally ammunition factories that were owned and operated by the United States government. So they can go through a spreadsheet and say, “Okay, I know how much it costs for you to build this. I know how much margin I’m comfortable paying you.” It’s a much easier crossing over the Rubicon. With software, it never felt like that.

Yeah, it’s still very hard. People want to pay for software in government on a cost-plus basis, which doesn’t make any sense if you think about the R&D that you put into software. The marginal price you’re paying is a small fraction of the R&D we’re amortizing across commercial and government customers.

And I think maybe, Trae, it’s worth telling the story of how, when you got to Founders Fund, you were looking for defense companies to invest in, and hitting so many dry holes is part of what led to the idea.

Yeah. I had—maybe still don’t have—any interest in venture capital. And so when Peter asked me to come over to Founders Fund, the only thing I knew to do was just to look at the category of things I felt like I understood pretty well, which was GovTech. Incidentally, you can look up who bids on federal contracts. I pulled that list, and at the time I was using Crunchbase. Then I just started outbounding to companies that were bidding on federal contracts.

I met with hundreds of companies in the first 3 years that I was at Founders Fund. We may ended up making one investment in KDM, which was renamed Expanse and acquired by Palantir Networks. But otherwise, that was it. There was nothing else. Even looking in retrospect, we didn’t miss anything. There was just nothing worth investing in. It turns out no one was building in this category.

So I went back to the Founders Fund investment team and said, “I feel like someone should be building a next-gen prime that builds hardware, but it’s software-defined and hardware-enabled, rather than being hardware-defined and software-enabled.” To my surprise, the team was like, “Yeah, sounds like you’re probably the guy who would know that that’s the gap. I don’t know, maybe you should start a company.” And I’m like, “No, you guys hired me to work on the investment team at Founders Fund.” They’re like, “Yeah, that’s fine. You can do that too. It’s all good.” So, yeah, it was kind of an accidental founding in many ways.

The capital requirements for a hardware business are notably higher. If you didn’t have the success you had building the manufacturing facility you’re building—and I’m assuming all the R&D cycles are longer and require more capital to get to a point of product—does it make sense for venture capital to be the funder of the next generation of hardware defense tech companies? How do the economics and the capital markets work?

So, I think if you back up a little bit, if you look at Fairchild, if you look at integrated circuits in 1968, 96% of all integrated circuits sold were sold to the Apollo program.

There was effectively a monopoly. There was 1 buyer for this thing. But Bob Noyce, who was at Fairchild at the time and co-inventor of the transistor, was so maniacally committed to a future where semiconductors—integrated circuits—would be in everything that he never let more than 4% of his R&D be paid for by the government. He was going to be like, “Why should I have some PM tell me what my R&D roadmap is? I invented the thing.”

So he kept executing that. He chased Moore’s law mercilessly. That meant in the ’80s, when the government needed precision-guided munitions for Assault Breaker to compete against the Soviets, we had the price-performance we needed. So, the idea is that when you get stuck inside the government loop of what they want, you never hit the price-performance.

By the way, there’s a good modern analogy, which is Jensen Huang and NVIDIA, right? In 1993, he sold graphics chips as an OE, almost like an OEM-type solution, until here we are. The one that speaks to my heart as a child who grew up in the shadow of the Space Coast is the shuttle. The shuttle is beautiful, but it’s $50,000 a kilogram to get to orbit. Starship’s heavy reuse will be under $20 a kilogram.

There’s no way that you will achieve that vision if you’re in a cost-plus world. Every day, you’re actually deleting your costs, which means you’d be deleting your profit. Instead, you want to delete your costs, turn that into margin, provide better price-performance, and keep riding that curve down.

Although, to be fair, SpaceX sued its customer, the Air Force, on the basis of the exact same federal procurement law that Palantir did. And so, you know, even with that massive costs costing down, they were in the same position as Palantir was in the early days.

Let’s go back to venture capital. At this conference, there are dozens of VCs now. This has clearly become a hot space, as defined by your successes. What’s going to happen? These VCs are plowing money into hundreds of defense tech companies, from drones to satellites to weapon systems to software tools. How is this going to play out over the next couple of years as a venture investor with your knowledge about Anduril?

I remember when I first got my offer letter from Palantir. We had this high-low mix of salary versus equity compensation, and it gave 3 examples of how much the equity in your offer would be worth at different valuations. I think it was 1, 5, and 10, maybe—or 1, 5, and 20, something like that. I remember talking to people around the company, and they were like, “$20 billion? That’s ridiculous. We’re never going to be worth $20 billion.”

I feel like now people are going out and saying, “Our seed round will be priced at $20 billion.” So, I think there’s been a big shift in the way that people think about these things. Some of that is good, and some of it’s bad. On the good side, they’re able to look at companies like Palantir, SpaceX, and Anduril and say there is actually a path to success. This isn’t an impossible market to innovate in.

At the same time, there’s a real tension around the amount of capital raised and the valuations that are being applied to these rounds, which add a tremendous amount of risk to the company that doesn’t have to exist. There’s this famous scene from HBO’s Silicon Valley where a CEO gets fired for underperformance on the plan that they had presented. Richard Hendricks, the main character, says, “You could have raised less at a lower price.” And he said, “What do you mean? No one ever told me I could raise less at a lower price.”

I think that’s ultimately what my advice to these companies is: “Look, your product might be awesome. It seems like you’re building a good team. I like the vision for what you’re doing and the mission, but you can also raise less at a lower price.” I think that avoids playing chicken, ultimately, with trying to hit numbers that frankly don’t even make sense.

One of the things that we’ve been focused on at Anduril that I think maps back to our experience at Palantir, actually, is climbing down the multiples tree with every round. We never want to go and raise the next round at a higher revenue multiple than the prior round. Even the Series A that you mentioned before is down pretty significantly from the Series G. We’re not doing that because investors wouldn’t be willing to pay higher multiples. We’re doing that because I believe the discipline is really, really important, especially heading into a medium-term IPO.

Emil Michael talked about the $200 billion of capital he wants to deploy. It’s been reported that he’s hiring bankers to help him deploy it. To counter your point, if there’s a lot of capital flowing from the monopsony, if there’s a big market that’s growing aggressively, you’ve got a $200 billion TAM to go after now. Shouldn’t that justify venture capital coming in, maybe even justify going at a higher price? Is this Department of War investment activity really going to change the landscape and increase the flow of venture capital?

I think it’s going to help out tremendously. First of all, I love the nickname this team has, which is Deal Team 6. I’d be honored to be part of that team. But I think if you look back at the titanium supply chain in the ’50s and ’60s, it was bootstrapped by the Air Force in a similar way, where they strategically injected capital down the supply chain to enable the aerospace industry to be created.

I think there’s a tremendous opportunity with the Office of Strategic Capital to think about where the structural bottlenecks are in production. You can have a lot of drone companies, and they’re all going to bottleneck on brushless motors. There are going to be certain key parts of the supply chain that we don’t have enough capacity to produce here in the US, and they’re going to need a bolus of investment. Just like integrated circuits, the first customer is going to be less economical than the end customer. So, I think that’s 1 part.

Then the other story that I love is where we really screwed this up. The drone is an American birthright. Abe Karem and the Predator—we built it at General Atomics. But, of course, then the government got in the way. The government said, “Hey, a drone is a flying missile. This thing needs to be ITAR-controlled.” Also, the FAA got in the way and said, “No beyond-line-of-sight operations.” So, you basically killed the domestic drone market.

There’s a counterfactual world where General Atomics had a consumer subsidiary called DJI, the consumer drone market was entirely owned by the US, and it provided economic prosperity for America and rode us down the price-production curve, so we could be using these things for innovation in national security as well.

What’s the next market where you’re worried that’s going to happen?

I have a lot of markets. We’re talking a lot about weapons of war here. I’m worried about things that go beyond weapons of war that affect our will to fight. Pharmaceuticals is 1 that’s close to my heart. My father was a pharmacist, and 1 of the things he always wanted me to work on was bringing pharmaceutical production back home.

80% of APIs for generic drugs are produced by China. What do you think the American people are going to think when they have to choose between defending the free world, defending American sovereignty, and their 5-year-old dying of an ear infection that we would have thought of as trivially curable?

I think there are these things where we’ve just imbibed the globalist vision that, hey, we’ll do the innovation and they’ll do the production, without realizing that innovation is a consequence of productivity. Probably to put it in terms that the tech community would really understand, what motivated Google to do the research in 2017 behind the paper Attention Is All You Need? A desire for a 3% incremental improvement in Google Translate. You cannot think of something more banal leading to something more revolutionary. And we’ve ceded all of those opportunities to realize and harness that innovation.

Now, if we stick with pharmaceuticals, there’s a reason 50% of all new clinical trials are happening in China and not in the US.

Do you have a view on a technology or product that’s developed in the US that we’re at risk of getting regulated out, or are we going in the right direction at this moment?

I would say semiconductors is actually a really interesting one, back to your point about Fairchild. We were the home of the semiconductor industry for many, many years. Once TSMC got a head of steam in Taiwan, they kind of ran away with it. Rather than investing in a domestic competitor, there are all sorts of reasons for this that we could spend hours talking about, but we basically just allowed that to happen somewhere else.

And now we’re in a position where almost no amount of money is going to fix this problem for us—certainly not on a timeline of relevance as it pertains to the risk that Taiwan has in 2027. So, I would say that’s another really big one.

Getting back to defense, how does the hardware and software fit together in this emerging war technology landscape? You guys are building software and hardware. Help us understand a little bit about what the systems of war look like in the decades ahead.

Well, I think the starting premise would be: where does it give us an advantage? We had the first offset, which was nuclear weapons; the second offset, precision-guided munitions and stealth.

Really, the third offset is decision advantage. How can we outthink and outexecute the adversary? That’s where these things fit together. That’s the thesis of even having them to begin with. I think the reality is that there’s a messy overlap of these things. All innovation is messy and chaotic, and maybe the department suffers at times because it tries to have a framework-based approach.

We’re going to have MOSA, modular open standards architecture,something like that. It’s like we’re going to avoid all the pain and messiness by having some sort of process. But in reality, the process always destroys all of the innovation. There’s obviously a very tight thesis where these things need to coexist and build off one another. They need to be interoperable, but you have to earn that. You have to earn that interoperability in the exercises, the test and evaluation, and in combat, would be my humble suggestion.

Do you think that outside of what Anduril’s building, there are actions that this administration or the private market should be taking to solve the hardware manufacturing and capacity gap that we have right now?

My view is that they’re actually taking a lot of actions. Not all of them have come to see the light of day here, but even something as simple as reimagining munitions and drones as consumables: they’re not things that you build and then stock on a shelf. They’re things that, at the time of ordering them, you already have an exercise and test plan where you plan to expend them. That means you’re going to have to replenish them, which means there’s a demand signal to industry to keep going and a buying cycle that means you can buy the next generation rather than having the old generation.

I think this department sees that in a very clear-eyed way and is doing the yeoman’s work of stitching that through all the services and all the portfolio acquisition executives. To the point of not believing in process, they’ve moved from a world where you have a very rigid, “Hey, I said I was going to buy X of Y, and I can’t change my mind,” to a world where there’s more autonomy and authority for the people who are buying things to say, “I told you I was going to buy something that accomplished this goal. I can change my mind about how I’m going to accomplish this goal.” Can you imagine trying to run any private-sector business without that degree of autonomy and flexibility in your decision-making?

To echo Shyam’s point about what the administration is doing, there are a bunch of swings that they’re taking. It wasn’t that long ago that, in the Obama administration, there was the Solyndra failure, where they invested in this solar-panel company and it ended up being a bad investment for the U.S. taxpayer. But now the Office of Strategic Capital is taking a really hard look at critical minerals. They’re looking at the refining of minerals, and they’re engaging directly with the private sector to come alongside them in doing deals to get offtake agreements on these things.

I think there’s some really clever thinking going on. In addition to that, I think the procurement process is up for revision constantly. Ash Carter started pulling a lot of these threads during the Obama administration. That was the Third Offset Strategy. We’re in a world today where there’s a renewed vigor around asking private industry, “What do you need? Tell us what you need to be changed about the way that we do business in order to streamline this and make it go faster.” I think there’s a lot of positive momentum.

Do you guys think these changes have been institutionalized, or are they political-party dependent? If the Republicans lose the midterms and there’s a Democratic president in 2028, do you think there’s a reversion to the old way of operating that’s going to have some political influence and change things? Or do you think we’ve really changed how things are running in the government here?

I don’t think it’s political, but I think it’s people. You can go back to, again, Ash Carter—a Democratic administration—highly focused on fixing these problems. The current Trump administration is highly focused on fixing these problems. There are a lot of people who were in the middle of that who didn’t prioritize this. It didn’t rise to the top of the stack. I think we always think of the bureaucracy as being a political infrastructure or an institution, or, you know, a bureaucracy. I actually just think that at the end of the day, it’s about leadership. Do we have the right people who understand the set of problems that we’re facing?

I couldn’t agree more. I call them heretics and heroes. The entropy of the bureaucracy is always toward some sort of sclerosis. It’s not a political statement. It’s just what bureaucracy does.

When David Packard, who’s probably the last major technology co-founder who served in the Department of Defense, came up with the 5000 series on acquisition—which today we view as just sclerotic BS that’s tied us down—his document was 7 pages long. In the years between when he wrote that and now, it’s 2,000 pages. Did he really screw us, or did the entropy take hold and there was no strong leadership in between to do the bushwhacking, mow the lawn, and make the right decisions?

This administration has blown up the JCS, which is one of these insane bureaucratic processes. Can you just take—forget a scalpel—take a machete and start clearing the jungle and re-earn some of these lessons as you go through it?

If you look at Kelly Johnson, who was the founder of Skunk Works, he built 41 airframes in his career, including the SR-71, still the fastest-flying manned aircraft, and the U-2, which we still fly. One of his rules was that he had to play defense to keep the government bureaucrats out of his program. I think you could—okay, it’s not a critique of government—you could think of big corporations. When the bureaucracy of big corporations gets into the innovation folks, the innovation stops.

The heretics and heroes were really founders. They were founders. From a tech perspective, that’s exactly what they were. To the Kelly Johnson point, the U-2 did not start as a U.S. military aircraft because they didn’t want it. He ended up going through the intelligence community to get his start there. It was about finding the right person who was willing to take the risk to do the thing, rather than relying on the system or the institution to do that.

This is the same with Bennie Schriever and ICBMs, and Admiral Rickover with the nuclear Navy. It used to be about people. But today, if you were to go to the Pentagon and say, “Who’s responsible for the F-35?” I’m not sure they even know. Who is responsible for the F-35? We’re building systems by committee rather than trusting the founders and people of the world.

Like a proclamation. It’s so great. It’s this X, Y, or Z project built in all 50 states. The objective is the money being spread around versus, how do you achieve the objective? The F-35 has components manufactured in 400 congressional districts. Right? I mean, that’s a political project.

This A/B test with SLS and Starship, right? SLS had to have subcontractors in all 50 states. But this A/B test has been played out.

I do think there’s something about the Calvinist spirit of America that sometimes gets us to misunderstand. We call it the Apollo program, but it was probably—it’s really the Gene Kranz program. It’s the F-16, but it’s really John Boyd’s plane. There’s a humility where we never want to call it that, although maybe with the nuclear Navy we kind of do call it Rickover’s Navy.

There’s an element of it being obviously both things. It’s bigger than the person, but the starting conditions actually require the founder figure.

Tell me about where we are in the state of readiness. We highlighted some of the statistics, but if you guys were to think about what we need to accomplish—the infrastructure we need to stand up to have production lines that can be turned on to support munitions capacity for conflict in, you know, pick your region—let’s say more than 1 region around the world, and the U.S. needs to sustain those conflicts to defend the United States. How far are we from being ready, based on what’s going on and the investments that are being made today across the sector?

I think the way I think about it is, if you thought about this as a spear, the tip of the spear is incredibly sharp. The shaft of the spear needs a lot of work here.

The Department of Defense is such a big organization. Structurally, supply and demand are not integrated. The demand side is the real-world events that happen at the combatant commands. The supply side is what the services and the industrial base build—the manning, training, and equipment, and how you’re actually producing that material. The ability to drive consensus, which is really the beating heart of any private-sector company—how supply and demand come together and what our view of it is—is managed with great difficulty in the Pentagon.

Your agility to respond to scenarios is weak. That’s where we can have a lot more precision, because if you can start changing your mind and saying, “Actually, I thought I needed X Barracudas, but now I need Y Barracudas and Z Furies,” how quickly does that take to percolate through the supply base? Instead, we get locked into these, like, “Hey, we’ve made a decision.”

The decision really can’t be revisited. Then, over time, the entropy is like, “Hey, I need—I have an unexpected bill. Let’s just take a little bit of money away from these programs,” and everything starts getting down to minimum-rate production, which is this idea of, “What can I produce just to keep the line open?” Which is not a deterrent. That’s how we frog-boiled our way here.

And this is where my idea is: You’ve got to tie this into consumption. Everything you’re building, particularly on the munition side, needs to be consumable, such that you know you’re going to replenish it, such that the primes and the neo-primes all have the demand signal they need to keep going and have a reason to produce them cheaper in order to make more money.

My most contrarian idea is the cynical view that you have the military-industrial complex and they’re just in it to make money. Well, it’s kind of a crappy business. These companies trade at less than 2 times revenue. The problem is it’s not profitable enough, actually.

But Trae, Shyam said in the past that we have 8 days of munitions on hand for a major conflict with China versus 800 days needed. How far away do you think we are from having the supply chain and production capacity built to meet that objective?

Well, one of the other things Shyam just said in this conversation is that there’s a high-low mix question as well. There are exquisite systems that are like the multimillion-dollar interceptor missiles, and then you have dumb munitions—things like bombs that are being dropped and things like that. But it’s not the same across that entire stack. There are some munitions that are woefully undersupplied and where we have very low readiness, and there are others where I actually feel pretty good.

From an Anduril perspective, we’re looking at ramping as fast as we can. We just opened our Arsenal-1 yesterday to start producing Furies. Looking out over the next 18 months, we’re not taking our foot off the gas, and yet there’s still a run-rate capacity that we’re running into as we ramp. It just takes time to get all these lines stood up.

If we were to start today with unlimited cash, I think over the next 18 months, as a country, we could get to the point where we were on track for having a sustainable industrial base. But we’re not going to do that. We’re going to trickle it out over time, and I think that unless there’s real political leadership that steps up to drive this forward, we will likely be in a similar situation for a long time.

So I think we really need to take advantage of this specific moment, where there’s clearly urgency that is understood by leadership across the administration.

How much of the spear is autonomous systems? Is everything going autonomous? Is it just drones, or is there on-the-ground autonomy?

Well, I would actually argue that the wars of today are fought with the weapons of yesterday. That’s just a fundamental truth. Some of our weapons of yesterday are awesome. B-2s are incredible. The bombs that we dropped last summer in Iran—incredible. These are very high-tech, exquisite systems. Patriot missiles are incredible. They’re very performant. They hit their targets almost every time. These are great systems.

The problem is that they’re incredibly expensive, and we can’t resupply. We’re way behind the 8-ball on that. Thinking about the wars of the future, it means that we need to start building these attritable mass systems today so that they’re in inventory to be used for those conflicts.

There are a number of Anduril systems at a level of readiness and deployment where they’re active and being used in combat today, but it’s still a very small percentage of the way that the wars are being waged. Over the next 5 to 10 years, that high-low mix is going to have to shift massively. This is just clearly evident in Ukraine. It’s clearly evident in Iran.

I think the department is taking steps to ensure that it has a better divide between low-cost, attritable systems and those high-cost, exquisite systems. But, as I said, it’s going to take political leadership to get there. We have the ability to do it.

The reason I ask is that I want to talk a little bit about the ethics of technology in war, the Anthropic run-in with the Department of War recently. I think you guys have both talked publicly about this. Anthropic refused to let its Claude model be used in Maven without human-oversight constraints. This is what’s been reported and what Emil Michael has said. The Pentagon labeled them a supply-chain risk.

Both AI and, generally, autonomous systems beg the question: What role should humans have, and who should have the right to play the role of hitting the kill switch? What’s your view on the role of the technology vendor to the Department of War, and where do you guys draw your line on what your responsibilities are with respect to ethics?

Well, my view on my responsibility to ethics is a slightly different question than my belief in democracy, which is a different thing. Maybe starting on the democracy point, I believe that the people of America have elected representatives to make really hard decisions about how we engage in combat. Full stop.

Fully autonomous weapons are not new. We’ve had autonomous systems in operation, like CIWS, which is deployed on naval vessels and shoots down aerial threats fully autonomously. You don’t have time to make decisions about incoming missiles or threats to your ship; you just have to shoot them down. So that’s what CIWS does.

CIWS has accountability in the system. There’s a person on that boat who is responsible for whatever actions that weapon system takes. I believe that this is the future of autonomous systems. Just like any other system—whether it’s a soldier carrying a gun, who is accountable for what happens with that gun, or the captain of the naval vessel, who is responsible for what happens with that CIWS—all of your autonomous systems are going to have accountability baked in.

Now, ethically, how do I think about this? I believe that what Shyam was saying about the first offset, the second offset, and the third offset is important. We as a society went from rocks and sticks to knives to guns to bombs, and then, in World War II, we kind of plateaued with nuclear weapons.

We all looked around at each other and said, “Wow, that’s pretty crazy. I don’t think we want to make more and more powerful nuclear weapons forever.” And so, actually, our engagement in combat has come back down the chain. We’re using precision-guided weapons. We’re shooting nonexplosive missiles into windows of apartment buildings, avoiding casualties and unintended casualties. I think that’s really the goal.

If you look at AI as the command center for making better decisions with better precision, with better discrimination, and with fewer civilian casualties, this is good. It’s actually ethically far improved from just dropping dumb bombs on areas of cities to eliminate military facilities.

I don’t think abstention from participating in the building of technology for national security is a morally neutral decision. You are making a moral decision when you decide to abstain. I am making a moral decision as a private citizen building a company in this space because I believe that this is ethically just. And I trust in our democracy to deploy those tools with the interest of the American people at heart.

There’s a lot that’s said about Palantir enabling a surveillance state. We had the All-In Summit in September. Alex Karp, your CEO, spoke, and there was a protest group outside protesting Palantir powering a surveillance state. I just want to give you a chance to respond to that, number 1.

And number 2, if you saw that Palantir’s tools were being used in an illegal way, where’s the responsibility for Palantir as a technology vendor in addressing those concerns?

It’s almost hard to respond to because it’s very unclear what surveillance people think we’re doing. There’s just a broad, almost maybe an outgrowth of Terminator fear around technology.

To get into that cultural question next, I think it’s very important to understand—

We don’t collect data. We don’t have any data. It’d be like accusing Excel of being a surveillance tool, right? It’s a way of bringing your own data that you have lawful authority to collect together to make decisions. Sounds a lot like Excel.

But because we are unabashedly patriotic and serve the US military, I think people have a colored view of these things. I would argue that it’s actually Excel, cell by cell, with much more security and auditability.

As Karp says, it’s the most insane platform in the world to try to do something illegal in because you’re going to be caught. That was part of the idea: How do you enhance privacy and security? How do you build more civil-liberties protections into it? How do you have a normative view that enables a democracy to say, “These are laws and rules; the system will enforce them”? We’re not just relying on people happening to do the correct thing.

That’s one piece of it. The other piece of it, to come back to this broader point, is the need for epistemic humility. To the point that we have elected officials, they are accountable for these policy decisions. I think, at the limit, it’s actually kind of indefensible to have a perspective other than lawful use, because if you are salami-slicing the policy, that’s actually tyranny by tech bro.

A small number of people are constraining the maneuver space of a democracy with no accountability to the populace. I think that's a pretty challenging perspective to be in.

We've been in this perspective where, if you go back to how the Soviets got the nuclear bomb, there are really 2 sources of treason. One was a committed communist like Klaus Fuchs, who was always a spy. But the other was people like Theodore Hall. He was one of the youngest scientists in the Manhattan Project, at 18 years old. His brother, Edward Hall, built the Minuteman missile. His brother was, in my terms, kind of a heritage hero.

Theodore said, “I'm one of the best physicists in the world. I'm probably also one of the best geopolitical strategists in the world. I think the only way to have world peace is if 2 countries have the bomb.” So, in 1944, Theodore Hall walked into the Soviet trade mission in New York and gave them critical secrets to the bomb. Theodore thought he was going to deliver world peace. Instead, every death from communism since 1949 is actually on his hands. And there's no accountability for that.

Yeah, I think one of the other things that comes up in these conversations about powering the surveillance state is this belief that we have policies, but we don't actually want our civil servants to have the best technology to enforce those policies. It's almost like a lack of belief in the institution of democracy. It's like traffic cameras, for example. I'm sort of libertarian in some ways. I don't love traffic cameras, but if traffic cameras say, “I'm going to send a ticket to every single person who blows through this red light,” maybe we have to change the policy around red lights if you don't like that.

I've gotten 3 speeding tickets in San Francisco from autonomous cameras in the last month.

Well, they did ramp it up significantly, which is incredible, because they're enforcing laws at all, if we're being honest. Here's my ticket. I'm literally watching the guy, while I'm doing 3 miles over the speed limit, take heroin needles out of his arm and put them on the floor next to an elementary school.

But it's the same thing with using tech for better enforcement of tax law or something like that. Is what we're saying by criticizing what Palantir is doing that we don't want our civil servants to have the best tools possible? You can have that position. I tend to think that position is pretty morally bankrupt, but I guess you could have that position. Ultimately, I think that's what it is.

So, scenario-play this for me. There's a public perception—I don't know how far-reaching it is—that there's some type of corruption between government officials who use technology that's super-advanced and hard for people to understand, and the technology vendors. If you saw government agencies using your technology, either of you, in an illegal way that you knew broke the law, would you report it?

100%. Yeah, to the IG. There's an entire mechanism there. There is an IG in every agency.

Right. So, can you just explain that for a second? The Inspector General, maybe?

Yeah. Every agency has an Inspector General, who's an independent organization that you can provide anonymous or non-anonymous complaints to. They then have the statutory ability to conduct an investigation in an unfettered way inside that organization, whether it's the Department of War or Housing and Urban Development—literally every single agency. This mechanism is used.

I actually saw it weaponized against one of my favorite heretics and heroes, Colonel Drew Cukor, who invented Maven—the founder of Maven, really. People would file complaints claiming that he was hiding illegal immigrants in his basement, a basement that he doesn't really have. The Naval Criminal Investigative Service went out to his house and actually looked into these things. People take this incredibly seriously.

Where does the culture come from—the anti-defense-tech alignment culture? Is it because of the peace era that we had and folks took national security for granted?

I think the first schism was really during Vietnam, where people felt like they were lied to about the war. It drove a fundamental schism between academia and defense, and we've never really healed from that schism. So, there's this distrust that's brewed there and kind of escalates in your society.

The second schism is that the number of people who are prior service or connected to this community, who actually see these people as humans and have a fully informed mental model of how diligent they are, how the work actually gets done, what these words actually mean, and what the process is like, has evaporated. So, their own fears fill in how they think it happens. It seems like maybe it's happening in more of a cowboy way, maybe it's happening in a way without oversight, that there is no such thing as doctrine. There's kind of a cartoon version of what's actually happening that I think you're unable to reconcile.

Yeah, I think if you even look back at what happened with Snowden—what was that, 15 years ago now?—there was almost no discussion about the investigations that went into whether the data collection was actually abused. The answer was basically not at all. There were less than 12 documented cases where someone got access to data they shouldn't have, and it was because of technology errors, not because of the policies that were implemented.

We can have disagreements about whether the intelligence community should have collected the data and stored it, but that policy was renewed multiple times by multiple administrations, with multiple political parties that had the majority in Congress, over decades. So, apparently our elected representatives thought it was important enough to keep in the system. I think it ends up being this kind of weird policy discussion.

But the second point that you make, I completely agree with: the distrust of the institution and creating the stories for themselves. I oftentimes go and do guest lectures at Stanford, and I always try to ask, “Raise your hand if you have an immediate family member who serves in the military.” No one ever raises their hand. It's crazy. At Stanford, there are veterans who come in and go to the GSB and things like that, but in the undergraduate student population, it's incredibly rare for anyone to have any connection whatsoever in their immediate family to the military.

I think some of this goes back to criticisms that people like J.D. Vance, the vice president, have made about elites in America and things like that. But there's just this incredible divide that has happened. We're kind of losing touch with that salt-of-the-earth, middle-of-the-country veteran community that I feel like was way more present during the Cold War and World War II before that.

Do you think there's any external influence driving this culture? Are there influences on social media or in mainstream media? Maybe just talk about destabilization and the attack vectors.

I've seen it in other areas of science. I don't want to spend time on this show talking about it, but there have been traces that I've found of external folks who want to destabilize American science and industry progress. They create fear, put out articles, and then go social; they become viral, and suddenly everyone believes it even though it's not true. Do you see that, and have you actually seen it in the sense of attacking the tech companies that are now supporting American defense?

100%. Even if you go back to—let's stick with Vietnam—the Soviets spent $7 million in 2026 dollars funding the peace movement, the anti-war protest. Now, obviously, there is an organic element to it, but this is just dumping gasoline on the fire to sow division and discord.

In the present day, certainly we see it against Palantir, where there's CCP money flowing to organizations that are protesting us for various domestic issues here. It's not isolated. It's a broadly successful strategy for adversaries to sow division. No one will believe it because no one wants to believe that they're being influenced.

I think Shyam's point is exactly right. It's actually a brilliant strategy. Good for them. They are not our friends; they are adversaries. What would you do if you were in their shoes? It makes a ton of sense. Do we do that?

I mean, look, we have all sorts of counterintelligence operations operating around the world. The modern art movement was really funded by the CIA to undermine the kind of Soviet control of art. It was broadly funded, and it wasn't directed by the CIA, but you can see how we have cultural values that we want to inculcate and spread.

Do you think we have a shot at recasting the defense industry—the tech industry that's addressing defense—and aligning it with a notion of patriotism? What is it going to take to make that happen?

I don't know. Maybe writing a book.

Oh, yeah, you have a book. That's right.

That's right. I'm trying to do exactly that. If we look at it with a clear-eyed sense of the world and how much deterrence we've lost, you could really say, “Maybe World War III has already started,” and 10 years from now we'll look back and be able to perceive that.

Right now, I think you don't need literally everyone to view this, but if you can create a more clear-eyed view of what's at stake here—not only for you but for your children and their future—I think you can get people to show up and participate. This has been America's story all the time. Usually, when we start these things, we are the underdog. All periods of American greatness have started when we realized that we're the underdog. We were the 17th-largest army in the world at the beginning of World War II.

What did a ragtag bunch of farmers and random tradesmen have to take on the world's largest army, the British, during the Revolutionary War? I think having some clarity about what's at stake and what the counterfactual is matters. It's so easy when you're successful to let the nihilism grow and say, "Look how imperfect we are." The self-loathing creeps in.

This is where I come back to your point on the legitimacy of our institutions. Should these institutions work? Do they deserve the best software? Whether they're public or private, you can't have doors falling off planes. Our government organizations need to provide the basic services they've signed up to provide without fraud, without corruption, and in an efficient way.

The reason is not just aesthetic. In the absence of that, it breeds nihilism. The younger generations look at that and say, "We should just tear all this down." Things will absolutely get worse in a world that looks like that. So it's incumbent on us to wake up every day and fight for the legitimacy of these institutions and make them more functional.

America's story is never written. Every chapter seems to be a whole new arc, and we're in one right now. There's a rising socialist movement in the United States. Can you guys comment on how much you think that socialist movement—whatever term they want to use for it—is going to affect our capacity for defense and resiliency going into the next decade, particularly as our adversaries are rising?

The argument I've always made is that our greatest threat as a nation is not homicide; it's suicide. It's in this vein: internal discord, division, self-loathing, and things like the socialist movement, which I think are symptomatic of this internal discord and of failing to ensure opportunity for our people to the point of having a functioning elite.

If you were to go backward and think about the root cause, maybe we don't have an elite that cares enough about the prosperity of the American people. We've made decisions through globalization. We were told NAFTA was going to be a great thing and that, if you lost your job in manufacturing, you should just learn how to code. There's a certain sort of callousness in that.

I'm not an unabashed free-market person, and I'm clearly not a communist, but there's a sense in which the decision on the margin really does matter. That comes down to leadership. I think one thing that makes me a little hopeful is that socialism literally doesn't work.

You look at what happened in San Francisco, where we elected Chesa Boudin as the district attorney, and then we're like, "Oh, wow, this is not going well. Recall him." We had the board of education.

It's rooted in a selling point of empathy.

Right. Right, fundamentally. What happened with the board of education getting recalled, and Daniel Lurie coming in as the mayor? We felt like we went into this valley and hit rock bottom. But I do think that people eventually realize that it fundamentally doesn't work.

Seattle and Washington State are likely going to lose a large number of their biggest employers. As that happens, they'll probably come through on the other side. It might take them 7 to 10 years to get there. It might be painful and take them a long time, but I think they'll eventually come to their senses.

What do you two disagree on?

That's a great question. We actually bickered with each other quite a bit at Palantir. Usually, it was Shyam being right about something and me taking a long time to come around to his point of view. Any of these cultures that are well-functioning are rooted in debate.

But you're avoiding my question—whether it's actions or outlooks. Whether you should show up to your interview at a tech company in a suit.

I know. I will say this because it actually is part of the story of Anduril. I thought it was a bad decision at Palantir to be as quiet as we were. I thought that we needed to get out there and tell the story so that there would be data that says things like what Shawn said about like we don't have any data. We are Excel.

We had a very quiet, reserved communications strategy, and we went back and forth on that a lot. When I started Anduril, I thought, "All the positive things we learned about doing business with the government from Palantir—the one lesson that I learned that we didn't implement at Palantir is that we're going to go out there and tell our story."

I think that's worked for us incredibly well. I'll be honest: I think Palantir has come around to my side of that debate.

Last question: If we don't do things right, what does 2040 look like? And if we do things right, what does 2040 look like, economically and in defense?

The economic part is the critical one, because national security is not an end unto itself. It's a means to an end, and that end is economic prosperity—the prosperity of the American people.

I think no country has done more to develop the world. How did integrated circuits and microelectronics get to Southeast Asia? We sent them there. We benefited in terms of trade, but which other winner in a war spent its own capital to rebuild the conquered in Japan and Germany? Now you have the stability as a result of it.

When I think about what could go wrong, you could actually have a Chinese century that we never recover from—that literally everyone in the world is a vassal state to China and might makes right in their world. We shouldn't forget that it's very clear, even in the present moment, that for the CCP, it's not enough for China to prosper. America must fall. That's an explicit part of the strategy.

Look, it is a business decision if you want to buy American or Brazilian soybeans. I don't begrudge you one iota which decision you make there. It's an entirely different decision when you're smuggling agricultural fungicides into the United States so that we can't grow soybeans. That's the sort of zero-sum frame that I think 2040 will look like if we get this wrong.

If we get it right, I think what we actually see is a massive reindustrialization of America, followed by the West. We see a thriving middle class, which I define qualitatively as a middle class that believes their children's future will be better than their future. That's a fundamental promise that I feel has broken down over the last 30 years or so. And we see a belief in our institutions once again.

I also think there's a big component around education that we haven't talked about, where we figure out a way to educate and successfully enter our young people into a marketplace that needs them and benefits from their services. I don't think we've quite nailed that.

But I do think that the reindustrialization point is going to be the central point in making sure that China's many-decade Belt and Road strategy doesn't put us in a position where we literally just can't do anything.

Do we need military and industrial primacy, or can we operate in a multipolar world where the United States can share influence and economic prosperity with China, perhaps Russia, perhaps 1 or 2 other nation-states?

The challenge with not being the leader is that you don't get to set the terms of engagement. The benefit we've had since the end of World War II is that we've had the primary seat at the table to say, "This is how we're going to do semiconductors. This is how we're going to do supply chains. This is how we're going to protect trade lanes."

The moment that you step back from that and someone else has all those incentives, you start playing by the rules of their game, and it doesn't stay multipolar for long.

I really appreciate the 2 of you being here. Trae, Shyam, thank you guys. This has been great.

Thank you.

Thank you, guys. I'm going all in. I'm going all in.