ドローン優位性プログラム｜DIU代理ディレクター トラビス・メッツ氏

We actually have a very wide and obviously sophisticated drone industry here. I say wide because it's not deep. I'm sort of impressed by the quality of the entrepreneurs, the quality of the ideas that they have. If we're at war with a pure adversary, it's going to be terrifying how many of these things we put up in here. It's not that we don't have people using this stuff and using it and understanding it. It's just not happening fast enough. Travis Metz joins War Talk today. He was recently promoted to DIU's acting director after spending the past year running the drone dominance program. Previously, he spent a career in PE and VC, spending not too much time touching defense, but presumably had an interest in weapons of war ever since he named his high school crew boat the Viking Death Barge. Travis, welcome to War Talk. Thank you. The Viking Death Barge thing is a deep pool, so I'm impressed by your Google skills. That's awesome. Co-hosting today, Brian Clark, Justin, say who you guys are.

We're going to introduce hosts. I'm Brian Clark. I'm a senior fellow at Hudson Institute, and I'm a retired submariner, and I'm just glad to be here with you guys. Hey, everybody. Justin McIntosh, 20-year retired Green Beret, served in first and fifth group, and now I currently work in defense tech and write a sub stack blog that is read by tens of people called Mind of Things. So just so you're coming from from drone dominance. I think that's probably become one of the biggest, you know, points of reference that DIU can make for kind of how they're thinking about acquisitions in the future and how they're thinking about experimentation in the future. Can you kind of talk us through just, you know, big picture, the genesis of it, and then how it got from, you know, concept to first trial, second trial, now we're in phase three, like what that actually looks like. Yeah, I'd love to. So just by way of background on me, I don't have any military experience. I was an Army brat. My dad was in the Army during Vietnam.

Actually, two of my boys, one was in the Army, one is in the Army. So I've got sort of Army blood, but no background in the military myself and not in the Department of War before last summer. I came down from Boston to work with an old colleague named Owen West, who had been in the prior Department of War under Trump 45. And Owen had been asked last summer by the secretary to help him think about what it was inside the Department of War that had gotten in the way of us fundamentally procuring, scaling, and fielding thrones. And Owen went on effectively a research project, and I sat second chair with him on this, where we realized that there are, it is not that there aren't myriads of efforts within the Department of War to buy robots, to buy autonomous systems,

to figure out the new ways of using them, to do the hard work around the doctrine and the TTPs and the data map PF. But rather it was simply not happening fast enough at enough scale. And so I don't think that's a complicated observation. I also don't think it's a controversial one. So lots of smart people trying to do the thing, but just not going fast enough. And you can argue about the root cause of that. It's probably as simple to say the problem of war is a big, complicated place with lots of inertia, and it's difficult to react in relevant timeframes to new things that we're seeing on the battlefield. But, of course, we do have this horrible battlefield in Ukraine, and the Ukrainians in particular, but the Russians as well, have innovated at an absolutely blinding pace around the use of robots to kill each other. And even though we, the United States, will fight differently than the Ukrainians and the Russians do,

almost no matter where we fight because of our capabilities at much higher altitudes, nonetheless, it seems obvious in staring at all this stuff that you need robots, You need unmanned systems. You need the doctrine and the force structure to use them at a much broader scale than we were getting to. So the Secretary wrote a memo, effectively a deregulatory memo called the drone memo last July. The reconciliation bill, the OBAA, the one big beautiful bill, provided money for lots of things. But one of the things it was particularly focused on was scaling U.S. supply chains around small UAS, so small unmanned aerial systems. And Owen and I were given the mandate by the secretary to use a portion of that money to do that. So to scale the U.S. industrial base, the way that we decided to work on that problem set, the way we decided to address that question, that goal, scale U.S., small U.S. industrial base, was to actually go buy drones.

So there are lots of tools that the federal government has to shape industry. but our view and my view, and I'm a sort of free market capitalist to my core, is that the best way to get the effects you want around the U.S. industry is to actually just go place orders for drones. Create a problem statement that you want to solve, create a set of guardrails around where the drones must be made to sort of achieve our goals around the U.S. industrial base, and then just go place orders. And so that's what we started to do starting last fall. So we had $1.1 billion in reconciliation money. We published a framework, I'm sorry, intellectual framework for a four-step process. So every six months we would do this thing called a gauntlet. We would bring drone companies in. We would compete them. At the end of the gauntlet, we would place orders. They would have to deliver on those orders in the following six months. And then we would do it again. So we issued a request for submissions in November.

We had, sorry, we announced in November, we published a request for submissions in December. We got 209 submissions in January. We invited people to come to Fort, we invited 25 companies to come to Fort Benning in February. We placed orders in March. And those orders are delivered, mostly delivered. There's a few outstanding. And then we did it again. So that was 30,000 drones that were mostly unweaponized. So just think about the drone platform itself. And I should have said this particular program right now, at least, is focused very narrowly on what we call small one-way attack drones. So these are about the width. You know, you can hold your hands apart. They're about that big. They're quadcopters. They've got four propellers. They VTOL. And they go out 15 kilometers. They try to kill something, but don't come back. In Ukraine, $500 to $2,000. Our first batch of orders, unweaponized, was like $5,000. Our second batch of orders was $4,000.

And this is, you know, used at the squad and the platoon level. If the Ukraine battlefield is any indication, they're used at ridiculous scale. So 8 million, we think there's 8 million quads just on the Ukrainian side this year. So 8 million of these small one-way attack drones. That number is, it's easy to say 8 million. it's hard to actually internalize what that means on the battlefield. It's an extraordinary thing. Between that and the bombers that we can talk about, these little, these octocopter bombers, depending on who you ask and which data source you look at from Ukraine, 70, 80% of the casualties are, you know, in the last few months have come from these things. We don't, we need lots of drones. We need lots of different types of drones. I might even suggest that quadcopters, these short-range ones, are probably not the fulcrum issue for us when you look at our problem sets. But it's not a bad place to start. And if you keep it, the goal in mind, which is supply chain, I'm actually not, the goal isn't to field and deploy these 30,000 drones

because we need way more than 30,000 drones. But the goal is to put us in a condition where we can actually start to have supply chain resilient drone companies. I think that's super interesting. And I just want to talk one thing specifically that I've noticed, like looking through the gauntlet outcomes. And then I know Brian's got a couple of things that are more specific to like industrial based stuff. But like I was surprised. So when I look at the initial list of companies that were going to participate, I saw a lot of Ukrainian companies, you know, General Cherry and, you know, some of the Ukrainian drone manufacturers. I was surprised by the fact that those didn't do super well. And now does that indicate to you that actually our drone technology is not in the, the, the doldrums that people had thought that it was, that it really was just a production issue or a scale issue, or does that indicate that we're setting our test conditions differently than what they're doing on the battlefield? So it's not a good replication or a perfect replication, I should say.

Yeah. That's a really good question. I think you exactly hit the nuance there. I think it's, Both plus. So, number one, we explicitly went to get the Ukrainian companies to come compete. Why? Because they've scaled and they're battle-tested, right? And we are neither. We were neither. But the policy administrations were building an American drone industry. So in order for them to come do this, we published this thing called the Supply Chain Framework. It's a 10-page document. It's a policy document, not a statutory document. It says if you want to sell us drones, you have to be on this journey to being a U.S. manufacturer. Define U.S. manufacturer in lots of different ways. We're actually quite copacetic about different business models. But fundamentally, policy administration is that you need to be here manufacturing drones.

It is hard for the Ukrainians to come here and sell to us. We are messy and complicated and difficult. We've been successful getting them to come over here. I think it is hard to objectively test FPV drones. I think we have done as good, I mean, I think we've tried really hard to do it, and I think we've done a pretty good job. But it's certainly the case that we are, the ground truth could be, you know, modestly different from the way that we've sampled it using these sort of testings. So that's sort of part of it, which is, you know, there's absolutely some beta in the evaluation of these sort of test results. Having said that, you know, so Ukrainian drone companies are getting orders, just not all of them. So a company called Skycutter, a company called F-Drones. I'm actually just having a brain freeze on the third one from before. Oh, from a company that's called Hyperscale, which is fundamentally based on a Ukrainian company called Green Tech. Eric Schmidt's company, Prenny Autonomy, which is an American company,

but fundamentally built and battle-tested. Ukraine is getting orders. There's a Croatian company that's now in the United States. It's fundamentally Ukraine-tested. Yeah, so you're seeing a lot of this sort of stuff. Sorry, one more point. But the American companies are good. There's a lot of capital that's gone into these American companies. They have tried, many of them, the good ones, have tried to go test it in Ukraine. So they're getting the benefit of that battlefield, which is sort of hard to replace actually testing it. And they're learning how to scale. And so I'm actually quite happy with the mix that we're getting. We're getting some of the best Ukrainian companies. We've got these wonderful American companies with deep backing. They're trying to figure out how to scale. We're hoping the Ukrainian companies who know how to scale will maybe joint venture with some of these folks. Messy mix, but I'm actually good with that mix. So given the experience that you have working with Ukrainians or Eastern European companies, How are you using that as a way to address kind of the two big challenges over there, which is lack of GPS due to electronic warfare and lack of comms due to electronic warfare?

So how are they approaching the challenge of alternative sources for position, navigation, and timing and command and control in an environment where you're not really going to get reliable comms, especially the last tactical mile? Is this providing a way for the drone dominance companies to kind of think through how they do that? because obviously it's a new challenge the U.S. hasn't had to address very much. Yeah, it's a good question. I mean, one of the ferocious realities on the Ukrainian battlefield is it is deeply contested across the electromagnetic spectrum, both in terms of trying to stop the command and control links and also deny position through the GNSS denial. We are doing that in our gauntlets. We are replicating those conditions. It makes zero sense to buy, to compete drones in a non-contested environment because virtually any battlefield that we would be on would have some version of this. And so we have, so we're partnered with JATA 401, which is the Pentagon Department of Wars. So it's a joint interagency task force 401. It's led by General Matt Ross, but right here, General Matt Ross.

They are the coordinating function within the department or counter UAS across lots of dimensions, but in particular around FED quads. and they have brought their technology and their vendors and their soldiers to our gauntlets to actually throw up a bunch of jamming, both on the GPS side and on the RF side. Ukrainians have figured out, the Ukrainians are involved in this incredible cat and mouse game with the Russians, with these sort of innovation cycles around this. So, you know, I could be wrong about this. I don't think anyone in 2022 imagined you would control a drone with a fiber optic cable that spun off the back of it for, there's evidence now they're going out to like 80 kilometers, so 50 miles, right? So 50 miles of fiber optics spooling off the back of a drone. We routinely, because we see this, we tested this at Carson three weeks ago or four weeks ago. So we're going out 15 kilometers at Fort Carson with fiber optics spooling out behind it. And sometimes it breaks, but mostly it doesn't. Mostly they figured it out, and mostly they figured out the payload issues with it.

And so if you have a fiber optic thing, you have no risk of RF denial, right? You have a high bit rate connection to that drone. GNSS is a different problem, the GPS denial. I think, you know, the U.S. has actually had really good, very high technology on this question of how do you do inertial, how do you do terrain following, and so forth. and I would posit that the Ukrainians and the Russians have advanced the state of the art of that at the speed of war over the past four years. I mean, you see real successful alt P&T strategies over there. The terrain-following stuff using the tiles from the satellites and all this stuff is, you know, by the way, sometimes it's hard to figure out which ones are actually working. It's very hard to discern this, but the Ukrainians, it's existential for them, so their purchases discern it. They buy the stuff that works and don't buy the stuff that doesn't work. That's actually really interesting, Charles,

and I want to touch on that because I was in Estonia this summer and I spoke with some Estonian SAF commanders over there and one of the things that they told me is that they've really enjoyed, of course, they're deeply partnered with the Ukrainians. They work with them very closely and he said that one of the commanders told me that what he sees when they're in a country when they're in contact with the Ukrainians is that they'll have vendors come in and the Ukrainians will literally pick up a device, look at it, hand it back and be like, this won't work and walk off and like very quick to cut vendors down and say like, no, no, no. We already have tried this. It doesn't work. Do you start to see that as a as a way that you can inculcate that knowledge base that they have into the way you're doing preselection to make sure we're getting qualified companies into these experiments? Yeah. I mean, one of the reasons why I thought it was important to bring some of these Ukrainian companies over here, as long as they're willing to truly come over here, was because they have the benefit of this accumulated wisdom over, you know, now, what, five years, almost five years of, well, I guess they would say like 17 years of battle, right?

Did my math wrong, 12 years of battle. So, yeah, we want, I mean, I would say we want to take advantage of all the iterations they've had. But then maybe as importantly, how do you take advantage, how do you actually recreate a system that innovates at the speed they're actually doing it in a setting where you are either, you're not at war or you're in a different type of war that doesn't, you know, it's not relevant to that particular product. That is the big question to me. And I think we have some ideas. We've set some experiments running on that. but the ability for them to use and then discard or use and say to the vendor who's with them in these military units, like the collaboration is preposterously tight. You got to fix this thing. If you looked at that first 12 months of war from February 2022, there was actually a lot of American drones being used in that. but then quickly for a period of time, there wasn't because they, they, they were not able to iterate at the speed of relevancy for that war.

And the Ukrainians built their own indigenous capabilities or indigenous companies, you know, an indigenous supply chains, which is amazing. Sort of away from the Chinese supply chains. But, and then what you then saw is a lot of American companies who are clever were like, we need to be in Ukraine. We need to benefit of this terrible lab. We need to go over there and do this iteration. And so I get, you know, I get pitched every day by a drone company, and I say to them two things. One is you don't need to biz dev me. You just need to come and compete. And the second is, are you in Ukraine? Because if you're not in, and it's hard to go to Ukraine. It's expensive and painful. And by the way, like, it's dangerous. Like, these companies send me pictures of their facilities getting hit. But it's invaluable for this development cycle. And so I admit I'm a little skeptical of drone companies that haven't figured out how to test their stuff in some type of operationally relevant environment because we have them.

I don't think they're going to be as good. Along those lines, Travis, so one of the reasons that the Ukrainians and the Russians have both fallen to using small drones for a lot of their combat is in part because of the adaptability that they provide. And you can get into this rapid move, counter move competition. You know, and Dan, Pat, and I have written about this before, but how are you grading or assessing these drone companies in terms of their offerings ability to adapt? So how, you know, is it modular or is it got readily adaptable software? So it's one thing to go and test it out in Ukraine, but it's another thing to be able to say, I can grade you on some level of adaptability that you're able to provide because we know whatever it is now is wrong for whatever the future environment will be. Yeah, I don't think we've done a good job of this so far. the gauntlets are six months apart so we are imposing change in that six month cycle imposing basically increasing the requirements or increasing the difficulty of the problem set that we're imposing on people to come to the gauntlets

or succeed at the gauntlets but six months is a long time and on the one hand six months is a short time inside the Department of War we're not used to these cycles and we need to be better at it I'm actually an optimist about that I think we are getting better at it. But six months is too long for the iteration cycles. One of the things we're starting to experiment with or thinking about, and this is true both inside of the drone dominance program within DIU, but also thinking more broadly with the DERPM and Project Agincourt, is how do you create an operationally relevant, operationally realistic setting where you can pair the military units at enough scale and the vendors and the technology to do the feedback loops with the right contracting vehicles that incent all this and then select for the companies that succeed Because I would posit, and I think this is, I know you and Dan agree on this,

or I assume you do. I'm not, the point of the drone dominance program isn't per se to buy, isn't to find the drone because that drone will be irrelevant shortly after I buy it. It is to find the business. It is to scale the U.S. supply chain so it gets cheaper. because mass matters here, right? Like that's one of the convictions we all have. The cheaper it is, the more you can buy of them. And I need to find the companies that can change them, that can iterate them. But it's the right question, and I wake up anxious about it because I think we need to do more on it. And it seems like with Jaieta 401, it seems like we've got a red force or an opposition force, kind of tailor-made for this move-counter-move competition. It's just a matter of kind of maturing both sides and then getting them into a regular battle rhythm for making the gauntlet a two-way competition rather than just a one-way competition. Yeah, and you'll see, so Gauntlet 3, the one that will run in March of next year, will have an OP4 and more red-on-blue stuff. But it's hard to do statistically robust testing in that setting.

But there is, I don't want to imply that this isn't happening now. So you'll see, like, so the CTO of the Army is a guy named Dr. Alex Miller. He's running, so the Army is trying out lots of things here, running these much broader exercises, red on blue, enabling the system, enabling the units with various robotics units and trying to sort of, you know, turbocharge this type of adversarial setting. Yeah, yeah, transmission and contact. Yeah. Yeah, exactly. I just, I think, and I think, I don't think this is controversial, by the way. I think it just needs to be more and bigger. And bigger meaning some of the, you know, some of the systems that didn't exist pre-Ukraine war, but now exists are 150 kilometers out and 2,000 kilometers out with cheap masks doing things that we don't, we have the much higher exquisite things with much bigger payloads, but we don't have these things. Yeah, I think they change the way you fight

Right, it changes your ops Yeah, definitely, you think about structuring your fight differently If you have these available Well, that also kind of leads to There's kind of this belief that we're going to figure this out And then, like, Jyada 401 can go away And some of these things can dissolve But it seems like the natural evolution will be That those will drive towards An actual early Exploration phase Where they're actually starting to try to predict What the next evolutions of this technology Are going to be so that we can get ahead of it instead of playing the mouse in the cat-and-mouse game and trying to catch up to what the adversaries are doing. Do you see that being like a future evolution of Jyata 401 and the way they interact with DIU? Yes, yeah. Yeah, in fact, I would say more broadly the reason why the Secretary of War and the Deputy Secretary of War set up the DERPM-UXS. These are all terrible names, by the way, right? We're so bad with acronyms. Don't let people name things, yeah, exactly. Yeah. When Owen took the leadership role at DIU, one of his first observations was we need more nouns.

So projects need names. So we just did a CSO for, I won't tell you the old name because I don't understand what it means, but the new name is drones on a boat. Hey, that is at least descriptive if nothing else. You know what you're trying to do. Get your elevator pitch done. Yeah, exactly. You know what I mean when I say it, right? Right? We have drones on a boat, and then we have snakes on a boat, right? That's right. Yeah. So when we think about this idea of creating a capability for rapid evolution of the force, so drone dominance identifies the companies, creates the kind of organizational infrastructure, so we do that. One of the challenges that we might run into is the force would need to evolve in different ways, depending on who the adversary is and what the environment is in which they're going to be fighting. I mean, a lot of the same underlying technologies may be relevant, but to your point earlier, the 2,000-mile or 2,000-kilometer drone may not be that relevant in Europe, but it may be really relevant in the Western Pacific.

So would we think about drone dominance having to have different gauntlets that are maybe oriented around different operational environments or different adversarial situations? Yes, yeah, exactly right. So, you know, interestingly, I would suggest I think we're doing a good, I mean, completely biased. I think we're doing a good job of getting the wheel, getting the rock rolling down the hill on the FBV quads, the short range ones. But I would suggest it's not the most relevant platform for one of the operational problems that we have. And so, you know, in PACOM, in CENTCOM, you know, the distance reigns and we need to be running. First of all, we need to be thinking about the operational problems there and how we can help them with a treatable mass at scale. And then we need to run the gauntlets there, or at least in operationally relevant analogies. And so Owen and the team at the DERPM are spending a lot of time thinking about how do we pair up with the right military units who have these problem sets in their AOR and then figure out how to run the gauntlets there. So when I talked to my Navy friends, like yesterday I was talking to a bunch of them,

but the guys coming back from CENTCOM are saying, well, pretty much the CENTCOM problem is now like the PACOM problem because the range at which we're operating is in a lot of cases going to be so far that we need something that's comparable to what you'd use in PACOM. So the stuff that's coming out of Ukraine, while it used to be maybe relevant to CENTCOM, is maybe less and less relevant to CENTCOM because of the challenges operating inside the environment. Yeah, you're going to see – I would suggest that some of the long-range one-way attack that is blowing up the Russian oil fields and a lot of their logistic stuff is relevant for PACOM and CENTCOM. DIU is running a prize challenge called Drone Domino's Long-Range One-Way Strike Nouns. So we'll be testing 32 systems at Dugway in November, I think. on that topic. And the, and the, and the, and the consumer for the, for that is going to be pick on the CENCOM. Well, it'll be all the comms, but that is so that the, so the bottom there is 600 nautical miles.

So that's what a thousand kilometers. But of course you're seeing 2,500 kilometers in Ukraine for things that are not that survivable. They're low and slow, but they're, you know, less than a hundred thousand dollars putting meaningful payloads, you know, on a dinner plate. And that is the capability that I would suggest does have real relevance. Now, we have U.S. companies doing that, too, and we've got good U.S. companies doing that. But the Ukrainians, of course, have had this, what we talked about before, this reiteration cycle. Most pressing, open, analytical questions that, like, the broader think tank defense analysis community, you'd want to have think about and chew on. You know, I think we think hard about the problem sets in PACOM and what are the right unmanned systems. I don't think they exist, by the way. I think the problem sets in PACOM are specific enough and they have enough of their own flavor that we don't, you know, it would be trite to say you just transport some Ukrainian system over there.

That's the question. Just to state the obvious, all of us hope we don't have to fight there. But if you did, what do these low-cost, amassable systems bring to the fight? What do they look like? Because we, I shouldn't say I, we want to get after this, right? We want to actually create problems that sets for the adversary that make it less likely that we'll fight. because we brought a capability that we don't have there today. So I'm sure there's lots of thoughts out in your community about what those, what that might look like. By the way, that is such a non-trivial question problem that I just asked. We don't want your easy ones, Travis. Yeah, like I don't even know what clearance levels require, just even ask the question, right? Yeah. Yeah. Brian's got a paper for you. So on the sort of broad theory of advanced market commitments plus gauntlets to get, you know, get your supply, get your suppliers correct.

Like presumably you don't think we're close to diminishing returns from expanding that model. But I guess I'm curious, like how far do you want to put it? And are you already feeling like what some of the limits to this approach to acquisition? could potentially be once you start to spread it out to more and more things to buy. Well, I think we're nowhere near the guardrails on this, my own humble opinion. So the way that I think about it is the – so these are not F-35s, right? So on the one hand, and we have companies and a private capital community that is staring at the Department of War and have this, I think, this thematic thought that the United States is likely to purchase meaningful scale of autonomous systems across all the domains, and therefore they can stare at forecasts for the total addressable market and rationally invest capital to build out these companies on their own dime

because of the future orders that might be available to them. And I think it is in our interest to leverage all that private capital doing this development and then bring them and then our framework. But what they need from us is clear demand signal. They need clarity about the fact of what are the problem sets we're trying to solve, how are we going to evaluate them that are being solved, and then if you win, what do you get? And so we talk about this as wide aperture. So, open a wide aperture for, so describe a problem, wide aperture test, evaluate them, rank them, order them. And it is absolutely true that FBV quads, these short-range run-run attack, are pretty homogenous. You sort of know what they are. They're easy to test. I can objectively score them. I personally think we approach the ground truth with our scoring. The other ones are harder. They don't all, they're not as homogenous and so forth. But nonetheless, I think you could extend that to lots and lots of platforms in this sort of area. Could you do this with an F-35?

Obviously not. But in the context of these types of products, for sure. And I think the market's thirsty for it. And I think competition is good. Like, I think it works. I think, you know, because, of course, the price matters here, right? We have, like, the United States is so wealthy, but the Department of War has very finite resources nonetheless, and we've got to drive down this price curve. This cost exchange thing is a real thing. We get on the wrong side of that against pure adversaries, and it is a desperate problem for us. Yeah, and I guess just from a firm perspective, like, it's more fun playing this game than it is playing a lobbying game or playing, okay, let's make sure we're friends with a prime that can give us some whatever subcontract. If you want to build drones, you want to build drones and get paid on building drones, and you want your VC investors to judge the teams on how well they can build the drones, not how many friends they have who are connected. Honestly, I think the companies prefer this.

I mean, maybe there's some few exceptions of folks that have matured their business models in a certain way, But, no, this is what you want if you're a company. Tell me what to do to make profits. Yeah, I mean, like, anything to get you away from the normal business development model in defense tech is a good thing. Because, like you said, like, the RFP chase of over a year for some of these things has been, one, it's prohibitive to small businesses, I think, in a meaningful way. But then also it's one of those things that it just you're already baking in a year to 18 months of lead time for development of a new technology. That is, like you said, it's going to be obsolete very quickly thereafter. So what do we do, especially on these lower tiers? I have a question, I guess. Do you see the bottleneck? Because when I look at it, where I see the bottleneck, it isn't in the individual tech companies. Like I think the defense tech companies, there's a lot of them, and they're very innovative. I think it's in the tier two, tier three suppliers where you're going to start seeing the bottleneck. And do you see that as pushing those towards vertical integration and doing it themselves

and then having multiple companies that are able to be, like, completely vertically integrated and make all their parts all the way up? Or do you want to see a more expanded, like, tier two, tier three supplier base in the United States? And I guess the question would be, if that doesn't get solved, is that indicative that there is a bubble because we're not able to grow the base enough to actually feed the need that we have? Yeah. One of my theories is to try to not be too prescriptive about the market structure on the basis that if you just consistently apply demand, the market structure will end up in a sort of optimized sort of place. My own view is that this is something where vertical integration is not the goal from our point of view. So, you know, from the United States Department of Works point of view, you'd like to have a diverse set of vendor bases across all the way, not single-thread through anyone. It would be nice if some of these factories and component suppliers were dual-use so that we're getting the benefit of coming down the cost curve from commercial demand.

But my key issue is to broadly not be prescriptive. I do think, and you're quite right, that the components, you know, the component supply chains are the ones that are immature. we actually have a very wide and obviously sophisticated drone industry here. I say wide because it's not deep. I'm sort of impressed by the quality of the entrepreneurs, the quality of the ideas that they have, but it's not, none of them is scaled. Like literally, and everyone gets mad at me when I say that, but like, and when I say scaled, I mean like, tell me where you are on the log chart versus 8 million a year. Cause that's what scaled looks like, right? It's not having made a thousand last week. we'll get there on this. You already see it, and it's messy, but markets are messy, right? So Gauntlet 2, the one we just ran, so we just placed 60,000 orders last week or two weeks ago coming out of Gauntlet 2. So that was Fort Carson. We had 19 companies, 23 platforms. We placed orders with nine companies, 10 platforms.

That is the first set of orders the Department of War has placed. I think this is true. So 99% sure this is true. where the drones are not allowed to have in them Chinese brushless motors or Chinese battery packs. And by the way, the best brushless motors in the world, the cheapest, highest quality ones are Chinese. The best battery packs in the world, cheapest, highest quality ones are Chinese. So this is no small thing. But this is part of the point, right? In the long term, we, you know, I say this publicly before, so forgive me being repetitive, But like in the long term, we should not assume the Chinese will sell us one thing that we need to build a drone when we're at war with them. So if you are of the view, like I am, you should plan for being at war with your peer adversaries and hope you're not. You should not have you need to build a drone doesn't have one atom that goes through China. And the reality is the Chinese are fantastic at making drones. So all these supply chains have sort of centralized there.

And we have to figure out a way away from that. So how do you build a brushless motor outside of China? So first, outside of China, and then secondly, in the U.S.? And the answer is, I don't have the first clue how to do that, but I know that if I place orders for it, someone smart will figure it out and capital will float to it, and that is what's happening. Like, we are going to buy these 60,000 drones without Chinese brushless motors in it, and it's going to be messy, but it's working. So to build those, you need mineral supplies that are outside China and processing outside China because you need permanent magnets, You need lithium and other components that go into batteries. So how big do you think the market needs to be to incentivize the companies to then pay for suppliers to stand up the facilities necessary outside China to do it? And obviously, you're not the only thing. You're not the only customer that's looking for this kind of capability. So, and have you looked at how big is the total addressable market for permanent magnet motors that have no Chinese components to see how big does drone dominance need to get in order to make sure that that actually happens?

Yeah, that's a good question. You asked the exact question I never like answering. Because, of course, I was very careful with what I said. I said Chinese battery packs, not cells, right? Because the chemistry we can't do outside, we broadly can't do. The railroad magnets in those brushless motors fundamentally being refined in China. If you look at the supply chain framework for just the drone dominance program, so just this one little slice right now, it has us getting away from both of those things over time. Is drone dominance enough demand to fundamentally move battery chemistry away from covered countries? No chance. Zero chance. That's okay. There are lots of efforts in the United States government. And we are just a tiny little part of that that are focused on the resiliency of our battery supply chains. I'm an optimist about that. I actually think I'm amazed at the amount of press announcements. It was just a facility opened in Fargo. I should do the company credit by remembering their name. I think it's called Badlands.

I don't mean to single them out, the exclusion of the others, but people are after this problem. and the United States has lots of different funding sources, both in the DOW and across the commerce and other places that are getting after that. Same thing on rare earth magnets. So the United States Department of War has invested in multiple companies at meaningful dollar scale to try to find alternative raw material sources and, more importantly, alternative refining sources for these things. I am not going to solve that problem by buying 30,000 drones that have, what, 120,000 brushless motors and have, I don't know how many magnets are per motor. I'm not solving it with that demand signal. So there's certainly those elements. We're also not solving it with transistors, right? Drone dominance is just this tiny little part of transistors. In the long term, I would suggest we need to have non-Chinese supplies of all these cheap microelectronic things. My demand signal is not going to sort of get there. Right. Do you see U.S. government? Go ahead. That's what I was going to ask. So do you see efforts across the government to try to maybe coordinate the demand signals efficiently to drive that?

Because obviously there's all these areas that have Chinese dependencies like packaging and, you know, circuit board manufacturing or production. All those things are happening in China to a large degree. Yeah, no, I'm impressed by the unity of purpose on this. I have to say, yeah, I'm at these people on PCC's policy coordination committee meetings. that it's absolutely a whole-of-government activity on the drone space. You know, one of the key things that would help drive demand signal here and accurate demand is a consumer drone industry in the United States. Those of us who, you know, like to criticize post-talk would say, you know, if we'd banned DJI 10 years ago, we wouldn't be in this position. And it's easy to say. It was hard to do at the time. But we should have done it. But whatever. We're clear right now about this. So we've had, there's all these regulatory efforts, the FCC and the FAA, that are slowly but surely creating conditions where you are going to build American consumer drones.

And that'll help. It's funny. It'll hurt in the short term because we'll be competing up those supply chains, as you guys described. Like, it'll be harder to build enough brushless motors because consumers, but in the long run, it's what we need, right? We need these sort of consumer-based demand signals. Speaking of consumer demand signals maybe we can bring this back a bit your prior prior life i don know like lessons from uh you know doing doing buyouts not in this world i mean what what holds what doesn hold like if you could bring 50 more people with your sort of backgrounds like where would you sprinkle them across the department of war i mean what's amazing about coming in the department of war is the, it is an incredible place. We all complain about it, we criticize it and so forth, but when you stare at the level of the people, there's an amazing unity of mission, people who have foregone lots of worldly comforts for being there and are smart.

I think, you know, I mean, maybe I should have grokked this, but like our uniformed forces, they are generally talented people. And so, and, you know, when you stare at a problem inside the Department of War, when you stare at one of these things that we seem to have a hard time fixing, I don't, I think it is almost never malevolence or corruption or whatever. It is a systems issue. It is an incentives issue. It's an alignment issue. It's a how do you change from things that worked in the past issue. So I think it's great to have to sprinkle in some external folks into the building. I consider myself lucky to have been asked to sort of try to help on these things. I do think in particular when you think about buying a charitable mass, so buying large scales of things, I think that particularly lends itself to people who have been out into the marketplace more, who have been in venture capital land, who have been in private capital land,

where we can rely on other people's money to do the research on this, to do the development on this, to do the scaling on this. All we need to do is give them an objectively transparent system and a demand signal, and they will iterate like mad. It's no longer the case, right, that the United States, it used to be, right, that all of the leading-edge science and technology came out of the Department of War, basically, right, or NASA wouldn't. That's not the case anymore. And that's, I mean, we shouldn't cry about that. It's just the reality. It's a big world with lots of places that hard science is being done. So we need to take advantage of it. And if we can do it with other people's money, I say do that. But I do think you need some new ideas and new thinking to best leverage that. So that brings up a couple of points. So one is, do you see inside the Department of the War the willingness to change operational concepts to use the kind of technology that will come out of the commercial sector because you have to accept the fact that you're not able to build the specialized

whatever missile or weapon that you want to build. You have to sort of accept what is going to be available when you cobble together commercial technologies if you want to tap into that funding stream that comes from the private sector. So have you seen that willingness to change? And then do we envision that this defense market will continue to grow at a pace that's going to make it attractive for private capital to continue to come into it, or are we going to sort of top out here pretty soon when the budget starts to flatten out? Do you think there's a willingness to change? Absolutely. I see it every day. I think the Secretary's leadership on this topic has been blunt and good. I do think there's nothing like the evidence on the battlefield, and so whether or not it's staring at some of the tactics of our adversaries in CENTCOM or watching what's happening in Russia, Ukraine, or, you know, thinking about the problem sets we have in PACOM. So I'm an optimist on this. I think we're absolutely rolling downhill.

The long term, I don't, you know, this is where I get myself in trouble, right? It's one of these podcasts that's eventually going to get me sent home. The idea would be that the budgets, you can see already, on the horizon is going to be probably flat budgets or less money. It's not going to be growing at the same pace as it has been. This is a zero-sum game at some level. We have to buy less of the legacy stuff and buy more of the future stuff. Right, so that's what I was going to get, is that we're going to have to adjust our force. So have you seen the willingness also to think about adjusting our force designs to keep less of the legacy stuff that just is really expensive and buy more of this stuff that's going to be more relevant given the commercial sectors, technologies that we can tap into. I think it's fair to say that there's been modest progress on that, but not nearly enough, meaning we have not won the fight on that, on the one hand. On the other hand, if you were in the rooms I'm in, it is a vigorous argument every single day. And the fact that it is discussable and arguable now and we're having these very difficult conversations, I think is good.

Yeah. I mean, I think the more you use, the less you have to maintain, Brian. That's right, exactly. Part of the game, too. Well, yeah, like the Navy just made this announcement about their F-40 or F-A-X-X, the new strike fighter they're going to build. And they had to scrape that money together out of hide. You know, so there was no big, you know, bolus of funding coming in from Congress or from, you know, the Office of the Secretary of War to pay for that, which I think is sort of a case study in the change that we're seeing, where if you want to continue to buy the legacy stuff, you're going to have to figure out how to pay for it, which means you're going to have to make some hard choices. So hopefully more of those hard choices are coming the service's way. Yeah, I could not agree more. And I think any of us who claim to have the right crystal ball is an exercising hubris, but I also think that just to say it simply, right, Like the, I am, I will take the under on how many humans we're going to put in manned airplanes in 2040.

Yeah, yeah. And that's not to say that we don't need to continue to have this incredibly good, amazing stuff. We do. It's not this sort of hard cliff. But I don't think, you know, I simply don't think, I just don't think the balance of power is going to shift. It's just a question of the balance of, you know, how we use these weapons systems is going to shift. I don't think it's a question of if. I just think it's a question of when and how do we sort of adjust our sort of budgetary constraints accordingly. I think that brings up an interesting topic, and you talked about it earlier, dot mil PFT. We'll just use the easy names. So when you look at doctrine and procurement, do you see DIU's role as helping both inform the services doctrine as they're looking at these new techs? And how do you balance that with, like, the procurement side and trying to speed up and get away from these five-year POM cycles and these, you know, long lead time procurements? Like, where is the balance of DIU's focus and how do you kind of balance those two as you're trying to maneuver in that space?

primarily on the procurement side. So DIU has in-year responsive funding that allows the services to move left away from their palm cycles. We rely on the services to fundamentally help define the problem statements, the mission gaps that they have where we can go find commercial technology or non-prime vendors to go solve it. Our primary, and we have funding to do this. And we have a well-oiled procurement machine that knows how to do this quickly with competitive tension in it and driving to conclusions and prototypes. But we, the services, have to do the .mrpf on this. We can provoke it. We can help accelerate it by moving procurement left and create prototypes more quickly and giving them material to work with. My own view is, like, there's nothing like the forcing function of being sent. you know, we sent 10,000 drones to the Army out of that first gauntlet. Turns out, I think sending 10,000 drones to them helps the .amup.

You have to build. Those are prototypes, right? 10,000 drones is a drop in the bucket on the small ones. But it works. It helps. But no, the services are responsible for that. It would be presumptuous for us to imagine we know how to do that for them. Yeah. So part of our job is actually, like, partnering properly with them, right? So, D.I.U. actually partners with the PAEs and the program offices and the province set owners so that this is a joint effort with them. Yeah, because I think what we, you know, the services need is like an interlocutor to who's going to connect me with industry and act as like an honest broker, if you will. But, you know, so it's just not operator with industry. There's somebody there who can stand by and say, yeah, it's okay to do this from a contracting perspective, from a legal perspective or whatever. And they need that. And I think, you know, D.I.U. seems to perform that function. I mean, that was part of the goal of TIU 3.0, as I understood it, was to make that interconnection to allow that development operation cycle. Exactly right. Yeah, yeah, yeah. You said it better than I do.

Yeah, I think that, you know, when you look at the service, when you look at the services kind of writ large, and, you know, you alluded to this with, like, running kind of the Dugway event in November where you're going to have Paycom and Syncom feedback. So obviously you're setting up an experiment that's been informed by a GCC. I always, at least my time in, and this is an in of one, so, you know, it could be completely wrong. Doctrine was very slow to update, even as new technology came on. And one of the ways we saw that was in Syria. We were already using, this is 2016, we were using drones, both quadcopters all the way up to, you know, Reapers and Predators to be able to direct artillery fire and to help, you know, move, advance the Kurds in Manbij or advance the Kurds in Raqqa or, you know, Hajin down south. I see that as like what that should have been a template for is, hey, this is going to be part of like we're seeing this as a more effective use. You don't have to have FOs and forward observers out for cab scouts calling in

this artillery. We can use these attributable systems that don't put as many people in harm's way. There was a lot of resistance in the artillery community to moving in that direction during the early 20-teens. I guess the question then becomes, if we're moving this way with the technology, what is the come along that's going to pull those services, doctrine, capabilities, or is it just pounds per square inch of pressure on advancing the technology and just they're going to come because they're going to have to? I suppose it's a little bit of both. I think in many ways the Secretary's focus on creating the Drone Domino's program was downstream of what your experience was. It's not that we don't have people using this stuff and using it and understanding it. It's just not happening fast enough. And so it's not that the people in the G357, the Army, don't get this. They do. Like, they're smart, but they're just not fast enough, right? Just not enough institutional pressure at every level to say this is actually more important than whatever X, Y, and Z are.

And so, again, I view part of the, and this sounds so blunt, but I view part of the benefit of actually buying these things at scale to be a forcing function to that. I have a simplistic view in my head that if we're able to get these things, because part of the deal with the Army and the Marines in particular and SOCOM around the drone dominance program for the quads is that we bought $30,000, now we're buying $60,000. They're taking them all. We funded it. It's a bridge to their palms. They take it all, but they have to use it. That's the deal. And I view that as a forcing function. I think we have these wonderful junior officers and junior enlisted folks. You give them these things and you set them into battle missions of any sort. I mean, they'll figure it out. I'm an optimist about that. I know it's messy, but I think it'll help accelerate the actual implementation of all those doctrinal changes. And I do think, you know, I have this image in my head of like the Army being this sort of big diesel engine. So slow to crank, but a lot of torque. They're doing the thing

now. Like this is happening. Yeah, you could argue it's whatever, X years, slow, whatever. It's happening. No, I mean, like I said, I absolutely agree. When I talk to people who are still in the service, you know, at various ranges, they're definitely excited by it. They definitely push. And, of course, they're in the military, so it's never fast enough and it's never good enough. So it's like, oh, we should already be doing this or we should already be doing that. And I do think that's a really positive thing. Do you see, I guess kind of one of my final questions, you know, from like small company side, if you're trying to build up the industrial base, obviously you need small companies and you need large companies. when we talk about getting into Ukraine, testing in these realistic environments, is there a role or a mechanism you see with DIU to help, hey, this is an interesting technology or, hey, this is an interesting approach. We want to get it better vetted. We're going to help this company move into, like, the actual operational testing or get them out to where they get some experience. Do you see DIU having a role there? So sort of. And this is maybe a little bit of a difference in terms of how Owen and I have conceptualized

So the leadership at DAU prior to Owen leaving last week was the principal deputy director is a guy named Matt Shortle out on the West Coast, a Marine pilot, civilian, but former Marine pilot. And then we have a military deputy, a guy named Major General Solo Kunkel, who's a two-star Air Force general, who's an F-22, F-15 pilot. So I would say one of the things that one of the sort of framework changes we've made is that we don't view our mandate particularly as picking a company and developing it. And I'll speak for myself on this because I think folks have different views. I think it is presumptuous, presumptuous for us to pick companies. I don't have confidence in my judgment on that front. Now, part of that's because the defense tech space, the startup space, has exploded in size and capital raise.

It is extraordinary the diversity and the number of companies that are trying to do the thing here. So how do I pick winners? So I'm much more focused on creating processes that bring in multiple companies, and we do whatever objective evaluation is available for that particular thing, and then funding them, then funding them through the prototypes and hopefully funding them into, you know, volumes in some way, shape or form, getting them the money and the resources they need to do the operational testing or operational like sort of testing. But I we get approached every day by single companies and I don't feel confident to say yes, no to them. I think you have a lot of people who have novel tech and they try to push it and it's both untested and like having it. For lack of a better term, having a bureaucrat pick and say, like, oh, this is what we're going to go with, because once you start down that road of investment, it becomes one of those things where now this is my project.

Now I want to see it succeed. And you get into that, you know, some cost fallacy to some degree. And so, Travis, I guess that, you know, just like you said with the army, where you guys are basically acting as a bridge to their palm. In a lot of ways, this forces the services to think differently about how they structure their programs, right? You're not structuring it around the particular offering from a particular company anymore. You're structuring it around some capability area that they've got. There's a bucket of money that could go to a bunch of different companies depending on who actually performs when the time comes. So are you finding that that bridge works effectively? Because you guys have that open-ended approach. Are you able to carry that effort to the services now? And they build their budget a little bit with that kind of broader PE line, I guess? Yeah. I mean, they're still trying to work out how to put the right amount of money against these PE lines. But literally just before this, before I'm talking to you gentlemen, I was on phone with a colonel in the Army who's fantastic and who's thinking about this problem and, you know, getting ready for the transition, right? So we're going to run four – our plans are run four gauntlets, and then we're transitioning this.

And so if I've done my job properly, so we bought from, gosh, I've lost the numbers in my head. I think we bought from 11 companies in Gauntlet 1. We brought from nine companies in Gauntlet 2. But this aperture is now going to narrow because, of course, the Army doesn't need nine or 11 FPV one-way attack quads. That imposes a logistical tail on them, which is ridiculous. what they need is, as Travis's view, I think they need two or three. I think it would be a mistake to have one. Wait, wait, wait. But, I mean, how many does, like Ukraine, right, revealed preference. They got, like, nine at scale or something. I mean, it's crazy how distributed it is still. But each unit generally has one. And their units are, like, are very independent. And so just to be obvious, right? We're trying to get the market to standardize on these particular things because I don't want to impose standards.

It's one of my theories here. But battery packs, the way you charge the batteries, the ground stations, there's a lot of kit here, which if it is drone specific and then you say to a company and like everyone can build on it and it doesn't need like, you know, connection to the battlefield and like constant adaptation every three months to be functional is what you're saying. Yeah. But I, what I'm worried about saying you've got six drones in your, in your, at the company level. And by the way, the six drones have six different battery connectors and battery types. And no, that's not going to work. Right. We know that. We know that going from the theoretical use in my little lanes at Fort Carson to the battlefield is going to be a very difficult challenge. So we've got to get down to a discrete number of drones. I will have done my job if I give the Army the ability to choose, in my words, two to three quadcopters at a much lower cost than when we started,

that are much more proven competent because we put them through these four sprints and we have compared them against the battle-tested drones. And we've got these companies started on scaling. I'm a religious believer in the value of the cost curve here. Like if you get people starting to do scale stuff, the cost will come down. And so the Army will be able to buy twice as many or three times as many as if we hadn't gotten down this cost curve. That will be job accomplished for me. Then it's there. They have to implement it. They have to do the .OMPF. They have to field and deploy it. They're not going to do that with five or six drones with ten different weapons platforms on them. They're going to do it with two or three in this assignment. And I think that's actually a really good point because I think there's a lot of misinformation about what a requirement is when it comes out. And people talk negatively about requirements. And I think that that's really good to hear that, hey, there are going to be things that are going to be standardized and there are going to be requirements. Now, what we're not going to do is specify to the nth degree everything that this must do. We're not going to have a, you know, a TRD that goes, you know, a thousand lines long for this.

But we are going to say has to be able to charge with a certain type of a certain way, has to be able to meet certain battery requirements or things like that. I think those are intelligent. And I think that's actually a natural prevention from like PAEs and program managers kind of becoming justice steward where they're like, I don't know what I want, but I know it when I see it. and it just being like scattergories in the force, do you see any resistance to even that level of requirements that you're imposing or that you're talking about imposing? Good question. I mean, I have such a bias on this about trying to get the market to sort these things out. I think, and it has to be realistic about this, but given enough time and cycles, I think the market would standardize all these things for us. So I would never even have to write the requirements document. But realistically we going to end up so there this thing called the so you have a drone then you got an ESAD right the electronic safe and arming device then you got the actual payload

Realistically, we're going to have to standardize around some of these connections, right? We can't have a million different ways that you attach the thing that goes boom to the drone. I want the market to figure it out because I think the market is going to actually, like, do the optimization thing as long as we're testing them properly. As long as I'm putting them through mission-appropriate testing, I will get the right combination of quality and price and interconnectivity and so forth. But I think we don't have infinite time, and so I think we, and this wouldn't be me because I don't know these things, but the Army and the Marines will stare at where we're at with these different things and say that's going to be the one or that's going to be the two. So, yeah. No, I don't think there's – I think the – if you go and read our first two RFSs, for Gotland 1 and 2, the request for submissions, we're really careful to be, like, problem statement. Like, the quad doesn't have to look like this. I shouldn't even say quad, by the way, because a bunch of people showed up without quads. One of the companies that got orders in Gotland 1 was this dystopian-looking sort of central core

with counter-rotating things on it, and it was terrifying. But it did a good job inside the buildings in terms of killing the bad guys, so we bought some. Yeah, I mean, why was there no line for, like, aesthetics? Like, don't you want your drone to look scary? Well, you should come see it. So we're going to run a bomber competition in about a month at Camp Grayling. So a bomber is more of a group 2-ish type of drone. It's like a quad or octocopter or a sextocopter. It has six motors. It's not a tradable per se, so it comes back. So it drops munitions. It drops mortars or something. thing, it goes boom, drops like 10 to 15 kilograms. The dominant class, and this was not a thing four years ago, and now it's killing Russians and Ukrainians every night. It generally hunts at night, so it's generally got thermal sensors all over it. And in Ukraine, they're $10,000 to $20,000. Over here, they'll be more expensive, but they come back, so they don't need to buy 8 million of them.

Sorry, I mention that because they are terrifying looking. They paint them matte black. They hunt at night. When you see them and you see them take off, you want no part of it. And it makes me think about all the sci-fi movies that we've all watched. But the Army and the Marines need that bomber. Like if we are in ground combat with a flaw of any sort, even one that's more maneuver, not attrition, it's dynamic, we need a bomber. These things, they go out 30 kilometers and they kill things. It's quite extraordinary what they do. Travis how would you describe your um you know FPV drone driving skills dreadful I uh the marines were kind enough to put me on the sticks on one of their neuros archers the neuros archer is an incredibly good platform and it took about 40 seconds before the marines took back the stick because they were putting at risk their five thousand dollar investment in that Neuros Archer. But this is the whole thing, right? Like we need to be able to train

19 year olds in like three days. So, I mean, you're like some weird equivalent of that, right? Well, no, I'm much worse. Truth be told, it's a really good point because I think where that the Archer that they gave me was one that I've actually been there's technology in that thing, which is supposed to not require a drone racer. And even so, I was incompetent at it. But the kids, I'm not a gamer, so that's one of the reasons why. The kids, I call them kids, but obviously soldiers and marines. They're kids. They're kids. No, they're kids. They're my kids, so I think about it that way. They're much better at this right out of the box than people like I was. But to your point, yeah, in the medium and long term, these are not things where you were going to be exquisitely trained to fly them. This is going to be, you know, at most a few weeks school. We're going to put enough technology on these drones that they take over most of the cognitive load.

You know, one of the mistakes we made in Gauntlet One, with all respect to Justin's heritage here, is too many special forces folks were there. It's a common fail mode, I believe, with the military is to call on SOCOM to get guys because you're going to get people who, like, build quadcopters in their basement and know how to do it, and, like, that's what they live to do, and they're going to be the guy that shows up, and they're going to be like, this thing is great, and it's going to get into a private's hands, and they're going to be like, I don't know how to make it take off. Yeah. I won't use the name because of OffSec, but there's a SEAL, a well-known SEAL that came to help us evaluate these things, and he's smart. He's thought about drones a ton. and he's really helpful for people like me. And then you see him fly it, and you're like, oh, this is not the guy. He's Maverick, right? He's a top gun. He's exquisitely good at this. He's in the top. He's four standard deviations on this thing. I can't rely on his judgment about which drones to buy for that reason.

He wants to fly the drone, do backflips, and land it on a dinner plate. I don't need that. Yeah, we want the drone that takes care of itself. I have an AI ML background. And I would, I think if you look forward, I don't know how many years, these drones are going to mostly do the thing on their own. The Ukrainians are already proving this. They're fighting off parts of the terminal autonomy phase. The all-GNS, the all-PNT stuff we talked about before, the drones are taking that over. The contested environment stuff is going to be so brutal that if the drone can't do much of its mission without being directly C2, it's not going to be viable on the battlefield. Can we talk about this for a second? The idea of, like, inference on the edge. I mean, that's basically been, like, cell phones, right? Like, are we going to need to get Qualcomm on the case here? Like, who's going to – is there something, like, specific that needs to be designed for those chips that are going to live on these drones? Or, you know, is Moore's Law just going to take us there? No, no, Qualcomm's already done it. Thank you, Qualcomm. Yeah, exactly.

No, 100%. Yeah, yeah, yeah. I mean, I think Moore's Law is going to get us there nonetheless. But, yeah, we need inference on the edge on these things. the drones, when we went to Fort Penning we went out 10 kilometers to try to kill things some people use most folks don't use repeaters, it's a different question, problem set 10 kilometers off of a ground control station into the Georgia Pines, actually a little hard the people that did the best were the ones that had a terminal autonomy capability, so as soon as it got below visual line of sight if you put it up high enough 10 kilometers away you can see it right? But once you got into the pines and you had to hit the thing, there was very little ability to RF it. And the ones that won were the ones that had built ATR models or bounding box models that allowed you to for the pilot to identify it from what I think of as outside the bubble. So think about it as outside the EW bubble or the distance bubble. So you can see it, identify it, put your finger on it, tell the machine to go kill it, and then the machine goes, kills it. You don't even know it's killed it because you've lost

contact with it. That capability requires inference on the edge. That is non-trivial. It's increasingly trivial, not in terms of compute. You've got to have compute out there on that. And so we expect that to be increasingly the case. We, in fact, I don't know how much detail you want to get into this, but we actually changed the price. So from Gauntlet 1, Gauntlet 2, the price went down. For Gauntlet 3, the price for part of what we're buying actually went up to allow for people to invest a little more in the technology for the distance-track things so that I wasn't pushing us down a dead end where you couldn't get enough inference on the edge and can't do the thing because you've got to have the thing. And I don't know whether or not it's going to be a Raspberry Pi. I think the Qualcomm product is called a Voxel, I think. Someone might know this better than I do. People use Jetsons on these things, but you can't put a Jetson in a $2,000 drone, right? It just doesn't work. Maybe you can put a Pi. So I don't know what it's going to look like. I bet it's going to get cheaper. yeah um we have to have that stuff on the edge no because you think about the drone units right

like you you probably are still going to need humans to do like the manual repairs on your you know ten thousand dollar baba yaga um but the idea that you need to have a human being like like when when my chat gpt can like beat video games for me and i just like watch it on my computer and am amazed. The fact that we need human beings to like be in the line of sight or, you know, be connected on the other end of the fiber optic cable. I mean, it just, it, this is going to disappear sooner or rather later. And yes, people complain like where, you know, we were promised drone swarms and 2026 are not there yet. But I mean, this is like, yeah, you know, two, three, five years, not like 15, 20. I completely agree. I mean, those are fighting words, by the way, for a lot of people. But yeah, I mean, the future is dystopian on this, right? If we're, the whole goal is to get the price down so you can buy masks. And buying masks implies that you can't one-to-one

these things. Like, I don't, you know, depending on whatever modality, think about the mid-range strike, long-range strike, or the quads, it doesn't matter which one. If we're at war with a pure adversary, given the nature of our ability to build and procure things over long periods of time, it's going to be terrifying how many of these things we put up in the air. And if the distances are long, that's a long time that someone's going to have to be managing these things in flight. So that's what Ukraine's finding is they just have it totally autonomous until it gets to endgame and then somebody jumps on and manages it for the final tactical mile. So I imagine that's what we're going to have to look at here is it's autonomous for almost its entire flight. I think so. The Ukrainians are funny on this topic. partly because their existence is threatened tomorrow. I would say the one thing you could say about them, because I'm an admirer of their business system on this, is they don't actually have the luxury of thinking about the longer-term cycles on that.

So I do think they would be mad. Some of these Ukrainian companies are doing this, so they'd be mad. But I do think they're probably under-investing in some of those two- and three-year capabilities on this stuff. I get why. But they're mostly flying these. You're right. They do take over, but they're mostly still one-to-one on these things. And if I was them, I would be underinvesting on my team showing up in America and doing the gauntlet thing, right? Yeah. So that's a good question, and we're very sensitive to this question. Oh, and my boss and I went over to – sorry, my former boss – went over to Ukraine, and we had a bunch of government-to-government interactions. and our view is that just first principles, the Ukrainian companies need to help Ukraine figure out a way to end this terrible thing. But we have, we can get alignment on this through economics, right? These are national, I view these as national assets of Ukraine, strategic national assets of Ukraine. So we want to pay for this stuff.

We want to be profitable as Ukrainian companies. We want it to flow back into the Ukrainian system so that they're getting the benefits of the fruits of their blood. on this sort of thing. No, I think, I mean, I went to the Brave One event, expo, whatever you want to call it. It was in D.C. probably eight months ago or something. And it was interesting just talking to some of the founders and some of the, you know, for lack of a better term, CEOs. And they're literally like, yeah, we started in an apartment during the early days of the war, putting together fuselages out of, like, drain pipes because we knew we needed to do something. and, like, we had enough people that had enough electronics background in the apartment complex to, like, start putting these together and doing, like, low-grade manufacturing to have, you know, something to do in the war. I did think it was interesting, and probably this was, like, my last question at least. You said something that I fundamentally agree with. You don't need an MOS. You don't need a job in the military to do this. We hear a lot of the services talking about needing to set up jobs and needing to have, like, people who are specialized in this.

And where do you think the tension is there and how do you think we get across that? The like, no, actually, guys, like you're not going to need to be super specialized because this is going to do a lot of it for you. You need to be specialized in these other aspects of war that are going to give you something that actually brings to and complements the autonomous capability. Yeah, I don't think we're I don't think in the long term you need MOS to fly a drone. I think you absolutely need career paths and MOSs around what it means to set up robotics units that fight in a different way. And I think one of the, you know this better than I do, one of the problems is if you don't have those career paths and those natural progressions, you're disincenting people from actually joining those things. I don't have a prescription in my head for what the four structure differences need to look like. but I would just have, you know, having dabbled in Ukraine on this topic, I would suggest that they're going to look different. I don't know how, I don't know what the differences are, but I think the nature of the technical capabilities that are going to be

available to us with the, especially in the unmanned aerial systems is such that you will think about designing your maneuver brigades differently. I don't know how, but we have a lot of smart people thinking about that problem. But what they would say is, but I got to have it. I got to have the right career paths in this area to sort of give people the education, the schoolhouse, the doctrine, and then reward them for having done it and not have it be a sideshow, you know, apart from the normal progression in the military. Brian, do you have anything else? So I just want to circle back to one thing, Travis. I think I heard an invitation for War Talk to show up at a gauntlet. Can we get that on the record? We have had, I mean, we get way more people that want to come than we can take because these are hard to manage. But we have had media come. So, yeah, it's a possibility. It's going to be cold. We're going to Michigan in the cold. We're a hearty bunch here at work.

We'll send Travis an email. I was on a phone call this morning with the whole team, and I promised the logo earmuffs. So I'm on the hook for drone dominance earmuffs. We got to do one thing before we start. Stop. Oh, I do outro music for every episode. Like, is there a musical genre that you think, like, best captures the, you know, drone dominance spirit? Oh, my gosh. Yeah, obviously I enter Sandman. And there's a back story here. I don't think I'm allowed to tell this story. But we did a video with the secretary back in July on the lawn of the Pentagon where he signed the drone memo with drones rolling around him with the drone dropping it. And the backing music was Enter Sandman for the video. But things I don't understand, we hadn't cleared the rights or what have you. So you have this risk.

But Owen will claim it was his idea. It was not. It was my idea. But if you are creative enough to find the original, we had to take it down and replace it with no music, but you can still find the original version with Enter Sandman on there, and I would say it hits pretty hard. The famous, the titious, the talent. You guys got to get different social media folks. I think there are different corners of the Trump administration that are perfectly fine playing with musical copyright. But anyways, Justin, you got to pitch. Oh, so, by the way, we're starting a War Talk book club, Team Yankee. It's a Tony pick. Justin, do you want to make a case for it? Yeah, so I talk about Team Yankee. It's actually a Cold War novel written about 1985. I would consider it a classic. It's along with, like, Red Dawn-ish kind of, you know, classic 1980s trope. the U.S. and Russia get into a war, into a hot war, and we see –

No spoilers. Well, and then we get to see the frustrations that come from that. I will say that for a book that was written in 1985 and subsequently turned into both a comic book and a video – or not a video game, a strategy game, it does a really good job of talking about the problems with production, but as experienced from a frontline fighter viewpoint and not necessarily from the machinations of a political or bureaucrat. And I think that was something that was well done and something that clearly people were thinking about in 1985 on our ability to produce and man a force in World War III. So I think it's pretty relevant for a lot of the topics that we've talked about in the last couple of weeks. and and just from a writing perspective like i have a really low tolerance for like the shitty version of this book and this is not the shitty version of this book like there there are characters and they have like enough depth for me not to completely throw up in my mouth when like you know he talks to his wife and kids and whatever um so uh yeah yeah check it out check it out uh

we'll be talking about it uh probably next week uh on next week's show unless you know some other catastrophe happens in the world, which, you know, not zero chance. Travis, thank you so much for being a part of Wartime. Thanks. That was fun. I appreciate it. Thank you very much. Count the quads instead of sheep. Buy enough and they get cheap. Us little army, don't take a word. Time is gonna buy you a one-way bird. And if that bird don't come back, that's the point, son. One-way attack. Million-dollar missile, two-quick and fly to the map. Son, oh my God. Six drones, six drones in one jump draw.

None of them fit. Welcome to the war. Navy SEAL does that, but it's landing on a plate Handed into a pipe, it won't take off great Little start-up, skip the business, son I'm only buying the ones that want Hot little army, don't stay away Papa's gonna find you a one-way bird And if that bird's gonna come back That's the point, son One-way attack Sir, sir, we're gonna need the sticks back Now I lay me down to sleep I pray the Lord my cap table keeps If I die before I scale Let a prime buy me in a fire sale And if that pack or motor's Chinese made Papa's gonna spend it back And you don't get paid In the cells? Next question, next question And if the only can't afford the bomb

Papa's gonna bridge them to their palm And if that acronym makes you proud Papa's gonna buy you a brand new noun Papa's gonna buy you drones on a boat And if the boat won't be blown Papa's gonna buy you They start a boat Come still on me Don't stay away Papa's gonna buy you a one-way bird And if that bird don't come back That's the point, son One way of fact 30,000 $63,000. $8 million. I was working on it. Legal did clear this, right? Okay. This one was Owen's idea. Breaking! The! The!