Picking Winners w/ Trae Stephens , Co- Founder Anduril

Lauren Bedula 00:38 Welcome back to Building the Base, Lauren Bedula, here with Hondo Geurts. And it's a big day as we're kicking off season three of our podcast where we highlight leaders who are looking to bring together the national security community with the tech industry to build a stronger industrial base. No one better than Trae Stephens to kick off our season. Trae is the co-founder and chairman of Anduril. He's also founder of hardware tech, and in the commercial side, too. So, juggling a couple of things as a founder, partner at Founders Fund. So, looking at these issues from an investment perspective, but has a really interesting background, having served in the intelligence community spent time on Capitol Hill, and was early on at Palantir. So, Trae, thanks so much for joining us today. Trae Stephens 01:21 Yeah, it's pleasure to be here. Hondo Geurts 01:22 So, Trae, it's, you've got one of the more interesting, you know, bio stories. Most folks when I think when they think VC that grew up in the Silicon Valley, and you know, start at Stanford or something like that, you actually started completely differently, what got you into kind of national security start? And then how did you transition kind of into these roles you've had now? Trae Stephens 01:46 Definitely a crazy path to it ending up here that I never would have expected. I grew up in rural Ohio, in the countryside, and was in high school, when 911 happened, I was a senior in high school. And I kind of thought until that time that I would probably go into journalism, which is really funny in retrospect, thinking about what's happened to journalism in the last 25 years. But ended up you know, committing to this idea that I wanted to work in national security, considered briefly, you know, applying to the Air Force Academy, because I really thought being a fighter pilot would be cool. But then I found out that not everyone that went to the Air Force Academy became fighter pilots. And it became a lot less interesting as a result. So, I ended up going to Georgetown, which was this kind of crazy adventure, trying to get into a top tier university, from my rural public school, but somehow managed to talk them into letting me come there. And then after college, went into the intelligence community, which was not the experience that I had expected, I think I came in believing that there was going to be some James Bond moment, like somebody was going to hand me the keys to an Aston Martin, and I would have like a watch that had laser beams on it. And I sat down at my desk, and I had a CRT monitor and Windows 98. And it was a shock to say the least, but ended up finding Palantir really early on, like before there was any revenue, and kind of just got lucky in falling into line with what ended up becoming the most successful venture backed software company doing working for government in American history, which was not intentional, I can't take credit for that. It was really just a really fortunate series of events that led to that. And then about six years into being at Palantir, I was kind of running our what we call the leverage team, which was kind of sales, the sales organization, and got to be good friends with Peter Thiel, who is known to many listeners of the podcast, I'm sure now, was the guy that started PayPal, and was the first investor in Facebook, he was one of the cofounders of Palantir. He asked me to come join him in venture capital. This is a huge surprise, given that I don't think I took a single finance class in college. And I knew nothing about venture capital. But to the founders’ credit, they hired me, knowing that I knew nothing about venture capital. And it's been a crazy ride since then, for the last 10 years. And we could go into the Android founding story and all that stuff if you guys are interested, as well. But that's kind of a very circuitous path that I've taken. Lauren Bedula 04:23 Well, we want to do that there are several avenues we want to go and one thing I wanted to point out is you're known for having a goal of an outcome from conversations being focused. Can you believe Trae said that so I'm waiting for that. Trae Stephens 04:37 We’ll undoubtedly come I just I just did a talk at the Ronald Reagan National Security Innovation Base Summit. And there were a couple of moments that I have friends in the audience that were looking at me like, please don't, just stop. This is not going anywhere positive. Well, Lauren Bedula 04:54 I mean you're someone who has been persistent at doing it. Your business is finding companies to fund and support the Department of Defense and national security community. And it's a frustrating market. You've been at it a while you've seen the evolution, what is it that draws you to this nexus between the two, especially from the private sector side. Trae Stephens 05:17 I just think it's important. You know, if we look back to the Cold War, Silicon Valley was running on the front edge of doing national security work, and the most talented people in the disciplines that mattered to national security, whether it's communications, or radar technology, or semiconductors, you know, like really anything that you can imagine that was critical to our national security efforts, were coming out of Silicon Valley, and everyone that was working in those sectors was working with the DOD, they were the first funder in many cases, they were the substantial source of revenue at the earliest stages of these companies as they're being stood up. And when the Cold War ended, you know, DOD went on its way. And there's, you know, a bunch of legacy history with how the defense industrial base and up today ended up being that Silicon Valley when its own separate way, you know, it started going down the path of building internet companies and learning ways to optimize sharing 140 characters with your friend and selling ads, to social media companies. And that was good for the US economy in many ways, but it hadn't almost no bearing or relevance to the national security mission. In a world in which you've reached a geopolitical end of history, that's probably a pretty good outcome. Like you get economic growth, you're not super worried about your ability to project power and hold superiority in in the domains that you're working in. But that's clearly not the case. And we found this out the hard way, with 911, where it turns out, the intelligence operations look very different in a world with digital communication than it would have worked when we had human spies, you know, taping letters to the bottom of benches and parks. And so that that was a big part of the learning from that experience. But more recently, with near peer conflict, and the kind of reality that there has been no geopolitical end of history, we're in a really weird situation, because we're relying on all these very expensive, low quantity exquisite systems that were in many cases developed like 50 or more years ago. And we're saying, we somehow managed to convince ourselves that it's fine, like a Patriot missile that was developed in 1968, that we pay two and a quarter million dollars for every time we fire, it is fine. It's not though - it's not fine. We have to figure out new ways of integrating modern technology to reduce the cost to create efficiencies. And doing that you have to bring along the people that have skills in the areas that that you're doing work in, which is primarily software, not hardware, and we lost those people to the internet over the last 30 years. And so, the reason I'm passionate is that I believe we have to get those people back working on the problems that matter the most, the DOD, and the DOD has to be honest about its lack of access to those people as well. And that's where I think some of that tension and friction has come up. Lauren Bedula 08:17 One follow up to that and specific to Anduril, I remember in 2016 or 2017, you were just launching Anduril, at the In-Q-Tel Summit, and got a lot of buzz at the time. So much has changed since then, I was reflecting on everything from the Google project, Maven debacle, and then the geopolitical environment. Can you talk a little bit about that journey, your vision for Android specifically, and how it's changed, just given the threat environment since as well, Trae Stephens 08:47 Totally, I mean, 2017, when we started, the company, was right at that transition point where we had just gone through 15 plus years of counterterrorism counterinsurgency operations. But it was just before the world as a whole figured out that geopolitical conflict with near peers was going to happen. And so, you had this weird window where, you know, engineers, or I hesitate to call them engineers, activists at Google had decided that the end of history had happened, and there were no threats. And the idea that the sheer idea that you should be concerned enough that Google should be participating in federal defense work was ridiculous. And like, that was a thing that you could say in 2017 that people would listen to believe and take action on. And so, we enter with Anduril at that moment in time, and, you know, we were out there talking about how Vladimir Putin was going to technical universities in Moscow and saying “he who controls artificial intelligence will control the world”. And again in 2017, it's possible to think that you listen to that and you're like, this is crazy. Like we're not close to AI having the sort of impact that Putin is suggesting. But forward six years from there, where Anduril is today and where the tech paradigm is today, I think it's very easy to look back on that and say like, well, Putin was more prepared for this than the United States government was. And that's pretty concerning. Simultaneously, I'll give credit where credit is due. Josh Marcuse and the Google federal team have done a really great job of saying, the activists of 2016 are not controlling this business anymore. And we have work to do to align the, you know, big tech companies, with the mission that's being set in front of us. And I think they've responded well, to riding the ship there. So Hondo Geurts 10:43 So, Trae, again, it's awesome to have you here, somebody who's seen things from so many different perspectives. You know, often we talk about the government of what the government needs to do more of to bring more people in realize, bring back the software folks from the internet and all that. I read recently, you talking about what the government should do less of which I think is a conversation that also needs to happen, but actually doesn't happen very much. What's your perspective on where the government's making some positive moves, and then where it can be a better partner to get after this issue? Trae Stephens 11:18 I'm not entirely sure which the government should do less of that I've said that you're referring to. But I'll give you an example of one thing. If you look back to the Obama administration, they did a bunch of things, right. With technology. You know, it was under Obama that Ash Carter was the SecDef. That started the third offset initiative that launched the defense digital service that, you know, started the Defense Innovation Board started the Defense Innovation Unit experimental at the time, like they really launched a bunch of important efforts. At the same time, the Obama administration also decided that the best way to reskill a blue-collar workforce was to train them to be software developers, that's the dumbest possible thing you could do, literally, the first job that's going to be eradicated is low level software engineers, by AI. Like it's already happening. If you need to build a website, and you need someone to code HTML, an AI is going to be able to do that. It's the really, really complicated stuff that you can't have an AI build, because it's not creative and generative in the same way, as like the lower skill level jobs. And the reality is, there's just not that many of those people. In fact, you could probably map that out on a normal distribution curve, just like everything else, a bell curve, where, you know, at the bottom of the curve, there's like, a bunch of people who are super, super skilled, that are like the NBA basketball players of computer science. And then you have people on the other side that are like, they're literally never going to be interested or good at that skill set. And then you have this bulk of people in the middle of that curve. And they're the ones that are like, you know, writing HTML, and, you know, doing, you know, pattern recognition with, you know, Python scripts and things like that. And the middle of that is just AI, the high end of that is, is rare, and like these are the people that are out there building, doing research at open AI, there are people at Palantir, that are figuring out how to do data management stuff, and really complicated and clever ways. These are like the people that are building companies that are shaping our generation of technology. But the DOD decided that they could build software factories, that they could just retrain Airmen, to build the critical applications that are going to move the needle for the Pentagon, which is just crazy. Like, there's just no reason that you should start a bunch of software factories. This is literally why the commercial software industry exists. You don't need to build a custom operating system, you get a license from Microsoft, you don't need to build an email platform, you can get that from Microsoft and many other companies as well. We should be leveraging the private sector skills that exist to get the software that we need. It's kind of bananas, that there's this idea that the government should be writing code. It just doesn't really track to me at all.   Hondo Geurts 14:10 In your experience. I think sometimes in a government we convince ourselves we're such a unique customer and to our own detriment that we think we're so unique. So, we have to have some things unique. Is that the way you see things now with modern commercial software, that is a government truly a unique customer. That's got to have unique applications and unique software and unique operating systems. Trae Stephens 14:36 Well, the government has this crazy tension that exists those misaligned incentives, I would say more than anything else where they if you were to go and talk to someone that's in procurement, procurement leadership, they would say it's really important that there's dual use technologies that we're using commercially available tech. And the reason for that is that we don't want to own all the risk behind this this product or company. We want it to be that risk has to be amortized across both the private and the public sector together, which makes sense, I get it. But then that at the same time, they say, but our needs are very bespoke and very unique. And therefore, the dual use technologies we wish exists aren't really usable. So, we're going to bring in a large contractor that we've done a bunch of business with before, to detail a list of requirements and then build bespoke technology. or the like, really weird version of this is you go to someone, and they'll say, yeah, I really care about dual use tech and like, what do you really need? What do you need industry to provide us and they're like, weapons, it's like, again, you say that you want to use technology. But really what you want is someone to build a custom for you, or you want something that explodes, where you're literally the only potential customer for that thing. And so, there's this crazy tension. And I think there are cases like the defense travel service, for example, like that was built by a defense prime defense industrial base, and it ran for decades, very poorly, but it ran for decades, Concur literally exists, you can just license software that does the exact thing that you want, for 1/10 or less the cost. But we don't do that, because we've convinced ourselves that we can't, because our needs are really specific. And I just think that's it's to your point, it's just a lie. It's just not true. Lauren Bedula 16:21 Okay, let's talk state of play. It's so interesting to hear your passion for the market, why you're involved, what you've seen in terms of evolution. I'm curious for your take something we ask about on our show quite often is this balance between the authorities and cultural issues? Maybe it's from Anduril’s perspective, but you're involved with many other companies that are trying to penetrate this market? What's the state of play? What are the greatest challenges? And do you think they are policy issues or issues with authorities are still cultural, like, what's your take on the environment, Trae Stephens 16:57 I mean, there are probably hundreds, if not 1000s of like policy knits that we could get into, like, change the way oversight is managed here, change the way accountability is managed there, change some like specific language to make it easier to reprogram funding, there's probably hundreds of those things that we can do. And someone is going to make all of those recommendations because this whole town runs on policy. And so, people are going to get paid to figure out the right sentence to get into the NDAA. Next year, I get it, that's fine. The reality is all of the policy and authority is that the DOD needs to make good decisions exist. They just don't make good decisions. And we get ourselves out of this quandary by convincing ourselves that, oh, if the policy were different, we would have made a different decision. But you've seen good decisions be made too. And so, it's like, it just doesn't really square, I think you need leadership with courage. And you need to be able to charge forward and just make the right call when given the opportunity to make the right call. And I don't think that the incentives are particularly well aligned, so that people feel empowered to do that. But I also don't think the knowledge exists to understand what the difference is between the options that are being presented. And I could totally see how a well-intentioned intelligent person could say, Yeah, I should hire Lockheed Martin to build a VA appointment scheduling software. Now, there's literally no world where that's true. But I can see where someone could come to that conclusion. And we just need to like, show people that there's a better way. Lauren Bedula 18:40 Well, on that front, I think there's a lot of energy around industry partnerships are teaming between traditional industrial base, the primes and new entrants, any take on partnerships and that kind of aspect of this. Trae Stephens 18:54 I mean, I would love to see more partnerships that were creative and valuable to all parties. I haven't really seen it work. And, you know, you look at the, you know, the best success stories in this space over the last 20 years companies like SpaceX or Palantir. And you look at the big wins that they've experienced as businesses. It's not through partnerships. You know, there there's a world in which ULA could have said, we should bring SpaceX in and have them do reusability. They're not going to because that's not the way they're incentivized. They're incentivized on cost plus contracts. They're incentivized to write really long white papers and convince the government that they can build something component by component unit by unit from ground up with a great level of detail where the customer gets to dictate every single technical decision that's made. That's what they're incentivized to do. And so why would they bring in a vendor that has a product that they sell at a higher margin than a cost-plus contract? It doesn't really work. They're only incentivized to squeeze that margin and make it difficult for those companies to survive because they'd rather just build it themselves. Hondo Geurts 20:01 So, so Trae, you've seen the market changing. The number of startup founders who are interested in supporting country through defense-oriented kind of work is like I've never seen it before. What do you tell those startups? Or where do you look for if you're put your VC hat on? You know, what advice do you give to founders, if they, you know, for all the right reasons, want to contribute to the country, but want to do it through startup, kind of oriented and defense tech, Trae Stephens 20:35 One word cynicism. If a founder comes in and says, I'm going to do $100 million of revenue in the next two years, my customers are banging down my door, you know, I've got these procurement officers that are already trying to figure out how to contract this. I've got Lockheed Martin reaching out trying to buy something, the answer is no, it's not true, like nothing that you just said is true. And the fact that you think that it's true demonstrates that you have no appropriate level of understanding of how this space works. And so, if a founder comes in and says, you know, I believe that I'll have a five-year path to my first credible shot and a program of record. And this is the amount of money that I'm going to need to raise to build the company to that point, and I'm going to burn very responsibly, knowing that real revenue is not going to pop until years three through five, then I'm willing to listen. But I think founders are trained with this idea that like, I'm going to build the most valuable company in the history of the world. And I'm going to beat all expectations, I'm going to blow out any prior record that's ever been set for doing this thing. It's just delusion. And delusion is not investable. I need reality. And I think that's the number one thing that is usually lacking that I'm pushing people on as give me a realistic portrayal of how this goes. And I'm going to be able to give you a pretty realistic read on whether or not that's true. Hondo Geurts 21:58 Yeah, that's, that's really useful. How about advice you give them to, you know, maybe have a little humility and learn the mission use cases, as opposed to just I have a solution to a problem that, you know, I'm very passionate about, but may not be a problem that the Defense Department prioritizes totally Trae Stephens 22:18 Yeah, I think this is one area where the DOD has significantly improved in the last five years, they're doing a much better job at communicating the problem spaces that they want people to work on. And, you know, one of the things that again, I don't think a lot of startups are aware of is that unlike many industries, the DOD tells you a five-year plan of what they're going to buy. And so, when a company comes in, and they're like, I built a perpetual motion machine, it's like is that in the POM, because if it's not, this is not gonna go the way that you hope it's gonna go. And so, I think, I think like, being responsive to those needs, and finding ways to connect with the end user, are really important. Now, they're not sufficient. It is really important, but you have to be able to push from the top down, as well as from the bottom up. And so, there's a kind of a complicated dance on the GR and business development side that has to happen. But it all starts with the problem, like you said, Lauren Bedula 23:14 I'm going to flip it a little bit and ask what excites you or interests you the most these days from an investment perspective. You talked about AI? I know you're an investor in open AI, is it? Are you very interested in the AI space or anything that stands out more? Trae Stephens 23:29 Yeah, one of my all-time quotes from Peter Thiel, who is my partner at Founders Fund, is whenever someone asks you what you're excited about from an investment perspective, it's their way of telling you that they're not excited about anything. So maybe part of this Lauren is that you're not excited about anything, and you're looking for someone to tell you. Yeah, very cynical. Yeah, I think, you know, we're at this really interesting paradigm shift moment in, in tech, where, you know, we had we went from web one 1.0, you know, web 2.0, which is kind of what we lived through for most of the 2000s with, you know, things like ride sharing, and social media and things like that. And the tech shift now that we're seeing is like, well, what's gonna happen as we go from, you know, like online on a computer, like a desktop to mobile to AR/VR, potentially, at some point in the future, although I think that's maybe a little further out than most people expect. And then, you know, this AI kind of phenomenon, where does that take us? How does that change the way that we interact with hardware? How does that change the way that we behave in our day-to-day life? And that, as with any new technology, there are risks and opportunities? The big question that we have as investors when we approach the AI space, is which layer you're attacking. There's, of course, the foundational layer, and that's where we make our bet and open AI, which is to say, like, there aren't going to be a million winners of foundational models, there's probably going to be a small set and unfortunately, due to the nature of the technology and the cost of training, the big tech companies are kind of leading the way in that space. And then you have like the next, the next layer up, which is like vertical apps. So, like, if you are working in a law firm, or you're working in a real estate company, or you're working in a software company, you’re the way you interact with AI might be slightly different, it might require specific products that fit into those verticals. And, and I think that as we look through that entire space, the biggest risk right now, in my view, is competition. Everyone is deciding kind of simultaneously, that they're an AI company. And when everybody is an AI company, you're competing each other to zero. And so at Founders Fund we’re trying to find really novel things, things that stand out, that could be either on the product side of things, where it's like, this is something that they have some sort of technical advantage, or some sort of mode on, or it could just be a founder edge, like we believe that a person or a set of people is so much more talented than the rest of the players in the market, that they have some ability to make a monopoly play on an industry. Hondo Geurts 26:09 So, Trae, you gave a really good perspective on what you look at when a founder comes to you. What would you advise if I'm a government program manager, and I have a startup coming to me, who claims they’re already a successful and commercial? What would be the two or three questions you'd recommend that government person start thinking through? If they are open minded, and they're trying to, you know, look for nontraditional players? I don't think a lot of them have a lot of experience. And so, they have the same risk of falling in love with a pitch deck, and not seeing what's under the hood, what what kinds of questions would you recommend they should be asking? Particularly if the company says they've got a commercial product already out there successful in the market. Trae Stephens 26:54 Yeah, I mean, they should be able to ask for details on financials, like, Okay, you have something deployed in real life. Tell me more about those deployments? Who are your customers? What sort of revenue are you generating off of this? You know, what, what amounts of capital have you raised? Do you have runway, like, how much, when's the next time you're going to need to fundraise. Just trying to understand the level of complexity of thought that the founder is put into that I think, can be super useful. The best companies are, like they stand out, they have a tendency to stand out. And it's hard to see that without a bunch of reps. And I think this is where sometimes, you know, on the government side people can get in trouble is that, you know, they see 10 companies a year or whatever in their sector. And they're trying to make absolute assessments rather than relative assessments with a very, very small number. It’s one of the benefits of being in venture capital is that, you know, I see 500 pitch meetings a year. And so, with a really large sample set, it's much easier for me to say, the top 1% of these 500 companies are these five companies. And so, I think there is an argument for having a closer alignment between the capital providers and the government decision makers as well. Because we do have a pretty good read on which companies are doing well, which ones have been well capitalized, which ones have really standout founders. Hondo Geurts 28:17 Yeah, I mean, to your point, when I was at SOCOM that was, you know, I came to that realization of venture capital firms - they're pretty good scouts. And they're pretty good at assessing. We see a lot of companies. Yeah. And so, for the government listeners out there, you should figure out a way to take advantage of that in an unbiased way. But there's a lot of data out, there's a lot of information out there that you don't have to go rediscover all by yourself, you know, on the ninth hour of your 15th hour of the day kind of thing. Trae Stephens 28:47 Totally. Yep. Couldn't agree more. Lauren Bedula 28:49 So we talked about this timeframe from when you founded Anduril to now, there has been a trend on the government side to have as many doors to knock on as possible, which is helping to have these conversations Hondo’s suggesting to with VCs with founders in the like, in a trend of small awards, but to as many companies as possible to give them a shot, SIBERS, pilots, prototypes. What's your take on this approach? What are you seeing from an Anduril perspective? Because Anduril is significant in terms of size now and competing with primes? What's your take on state of play on that front? What are positives and negatives? Trae Stephens 29:24 I think, you know, it's good to expose a wide assortment of players to the problems that the government is facing. But if you don't transition those things into production opportunities, you know, it's going to be bad. Right now, as Hondo mentioned earlier, there's like never in recent modern history, there hasn't been another time that's had as much momentum for drawing the tech community back into doing work with the government and some of that is been prompted by companies like Anduril, Palantir and SpaceX. Actually, I think a large part of that is people that are trying to catch the tailwind behind that. But there's also, you know, a geopolitical moment where people are more aware of this reality that they want to work on something that matters rather than, you know, selling monkey JPEGs, or, you know, optimizing ads at a big social media company. And like that tailwind is bringing us to the place where we have an opportunity, where capital is showing up putting money behind things, the founders that would otherwise be doing other things they're showing up. And the government is showing up with these SBIR is prototype OTAs, which is, as I said, it's good. But if none of them transition, and all these companies die, the venture ecosystem is not going to be back for a while until like another economic cycle is passed. That doesn't mean by the way that I think it's on the government to pass 100 companies through into production, that's also not a very effective way to do business. The really controversial thing that I keep saying that makes people really upset about this is that the government has to pick winners, we have to be in the job of saying we have looked at a bunch of things. And there are these companies that we want to put like back the truck up and start pushing really hard on. And that's the way that every centrally planned competitor that we have does things. There aren't 20 group one drone companies in China that are equally successful. There's one they said DJI is the one, we are backing this company, we are subsidizing it, we're making it impossible for anyone globally to compete with them. And we haven't done that, because we're afraid of it. The entire American governmental bureaucracy is built on the concept of fake competitiveness. It's not actually competitive. In fact, the most communist part of the United States market economy is the defense industrial base, we've metered out in perfect increments, just north of inflation level growth for the top five defense primes for 30 years now. So, it's totally centrally planned. But we pretend very carefully, that it's competitive. And so, we do this in the startup space as well, we say, well, the worst thing that we could do is pick a winner. We are not in the business of picking winners, we're in the business of like creating opportunities. But if you create opportunities, and don't pick winners, then there are no winners, and then the capital runs away, the founders that are interested in the market run away, and you end up right back where you started. And so, I think we need to have a national mindset around the importance of saying there are going to be companies that lose, and there are going to be companies that win. And as long as there are a few companies that win, there will always be 1000s of other companies lining up for the shot at being the next company that wins. But if nobody wins, there's no line. Everyone's gone. Hondo Geurts 32:55 Yeah, you know, we've talked in other episodes about this lack of procurement funding, lack of buying, we're great at prototyping. But we're really struggling with buying and producing things. There's talk earlier that the government still doesn't know how to buy software, in fact, and there's, you know, funny stories about Anduril buying hardware companies so they could sell software. Is that Is that still the perspective you're seeing? Or are there been improvements in the way that government is thinking about actually buying software products? Trae Stephens 33:28 No, I think I think it's true. You know, at Anduril, we understand how difficult it is to sell software, primarily because many of us on the executive team, were at Palantir for a long time. So, we saw how hard that was. And the government is kind of more programmed to understand the how to evaluate a bill of materials, how to evaluate margins based on that bill of materials, how these like platforms, these hardware platforms are sustained over some period of time. And so, it just makes more sense in the procurement process. To go in with hardware, even if the value that that hardware is providing is primarily driven by software. When you walk in with software, there's almost like this, you know, moral revulsion that happens where people are like, well, what am I even paying you for? Like, if I'm giving you $10 million a year for this capability? What is your cost to deliver that? Well, it's its software. Generally speaking, the cost to the company delivered is like de minimis, because it's just code that's being digitally shipped. That means it has really high margin. And then you get this like reaction from the government that's like, I don't like I don't like that. I don't like the idea that someone's making a bunch of money at really high margin on giving me something that they have no marginal cost to deploy. And so, this is why we get into these bespoke build things because the one thing they can count with software is butts and seats. And so, if they have, you know, 1000 hours of butts and seats, they can go back and you know, beat boop on the calculator and say their cost to deliver this was 9.8 million. So, paying $10 million. Okay, cool. That's a, that's a really thin margin, we feel like we've gotten a great deal. But that's not how the world works. That's not how commercial software works. Again, it is really important that there are winners, and until people in procurement roles, understand that it is good for people to make money, because that's how you bring other people into the ecosystem, we're never going to get the software. Hondo Geurts 35:23 So how should they shift their thinking about buying a software product? Trae Stephens 35:32 I think it's like evaluating and every opportunity is the thing that I need something that I cannot get even to like 90 to 95% effect like coverage on my requirement from the commercial sector. And if I can't find going hire, you know, a big defense company to build it for you. But 99% of the time, there's gonna be something out there that gets you nearly all the way there at a fraction of the cost. And I think it's just holding people accountable for like, you want to build a travel system, you know, you can't Concur exists. You want to build like a data management, like a data warehouse. No, you can't snowflake exists. Palantir foundry exists. But right now, there's very little accountability, about that final step in the decision-making process. Lauren Bedula 36:22 We could talk to you all day, but I know you have a flight to catch. So, I have one more and we've hit on talent. You've talked about the energy, both Hondo and you're seeing from founders and companies stepping up. What's your take on the overall talent environment, even Maven coming back in a strong way? Or sorry, Google, from the Maven debacle, if you will? Are you having trouble hiring? Are you seeing a lot of energy and interest in service on the private sector side? What's your take on talent? Trae Stephens 36:50 Yeah, again, going back to the bell curve, normal distribution, if you think about that as like, super patriotic, and super not patriotic, and a patriotic is a loaded word. Maybe it's not the right one. But I think everyone gets what I'm saying. On the far side of that curve with the Super not patriotic, they're not going to work on defense. And you know, that's okay, we live in a democracy. Not everybody has to work in defense, I'm okay with, you know, the activists from Google, like working on non-defense stuff, cool, great. Love that. On the other side, there are the super patriotic top talent engineers. These people already work at Palantir, SpaceX, or maybe in some other startups in this space, they've self-selected into working on this industry. And you know, if you're starting a new business, it's not like you're recruiting people who have never worked in the industry before that are like gung-ho, like, if they were gung-ho, they're already working in the industry. So now, and then you have the middle of that curve. And these are people that could be convinced, it's possible to convince them of the importance of the mission. That was harder, five or six years ago, it's easier today, because people have kind of more of an emotional connection to the geopolitical conflict that exists globally. But you still have to convince them, there's no free pass. It's not like Ukraine happened. And all of a sudden, these people are now super patriotic. No, it's more like, these people are currently like, you know, writing, like firmware at Nvidia. And you have to go to them and say, I know why you're doing what you're doing. This is important. We need GPUs that work for training ALS. But let's talk about national security. And I want to explain to you why getting the strategic national security side of this right, is critical to preserve your ability to, I don't know, say work at Nvidia. And those conversations are nuanced. They're idiosyncratic. You have to like to get to know the people to even engage in those conversations. And I've seen that I'm seeing that happen more and more as we scale in this space. And there's a lot of people really talented people in tech that are working on problem sets that are vacuous. And they just like they feel it every day, they go into the office, maybe they're making really good comp, you know, tuning ad optimization models and things like that, but they know if you can talk to them, they know that what they're working on is not actually moving the needle for humanity. And so, you just need to get them in a place where you can convince them that there's something more important, they could be doing with their time. Lauren Bedula 39:15 Well, thank you Trae, part of what we hope to do with this show is stimulate more energy and talent around working on these issues. So, you've certainly done that today. Thank you for being so frank and for your continued passion. Absolutely. Trae Stephens 39:26 These issues. Thanks for having me guys. Take care. Thanks, Trae.