Where Edge Computing meets Precision Strike in European Defense: Stark Cofounders

Cold open

Uwe Horstmann: I think what came together is obviously an understanding in Europe that Europe needs to take more responsibility for its own defense. That is, of course, something the US has told us repeatedly. And they're right. Europe has to step up and needs to be able to defend itself, because that is our way of continuing to live in peace. It's deterrence, right? We probably understand now that the world is unfortunately not as peaceful by default as it ideally would be. It needs to be defended, and the only way we can be safe is by deterring a potential attacker.

Main conversation:

Pat Grady: My partner Luciana and I are here with Uwe and Johannes, the co-founders of Stark, a neo prime serving the European defense community. Gentlemen, thank you for joining us.

Uwe Horstmann: Thank you so much for having us.

Pat Grady: All right. To get things started: you guys landed billions of dollars of contracts within 18 months of starting the company, which I believe is unheard of in the defense business. What did you do right? How did that happen?

Uwe Horstmann: Obviously this is something we like to claim as a big success, but it's also a moment that is just optimistic for all of us, because something has changed beyond just what we built in our company, and we're moving in the right direction.

In Europe, I think what we did specifically is four things. First, we did innovation really close to the battlefield, really being integrated with the units and soldiers who actually use the systems.

Second, we're venture-first, thanks to Sequoia and others, which basically means we fund the innovation and the product development ourselves. We don't rely on government R&D spending.

Third, we focus on designing the products so they become really usable and deployable en masse. So beyond pure innovation use cases, all the way to certification, validation and safety of the system: really usable for NATO forces.

And number four, we designed the products with industrial scaling in mind from the get-go. How can we make sure these are not only a few exquisite systems that stand around somewhere, but that the factory is part of the weapon? Production is deterrence.

Pat Grady: Let me ask you about point number one, close to the battlefield. I'm guessing most military units would not let a couple of random guys traipse into their unit and understand their needs. How did you gain access to the battlefield so early in the life of the company?

Uwe Horstmann: We do that with a large presence in Ukraine. We have a large team in Ukraine, and we integrated into those units quite extensively. We also work with the European forces. There are more and more formats where that is enabled, so you can connect with the soldiers. You [can] show up with them. You can go to the battleground, the testing grounds, the training grounds, and that's where you get the best feedback.

That is the most important input factor for us. The capital we can get, again, thanks to Sequoia and other people. But what we really need is the input of the customer, which is not necessarily the buyer. The buyer is probably some high-level official, but the customer oftentimes is a relatively simple soldier, on an operational, tactical level. We need their input. That's more important to us than anything else.

Pat Grady: Just to level-set on what you guys do: what did you hear from those customers, and what's the product you've now given them in response?

Uwe Horstmann: At any point in time we probably run 10 to 15 different programs, almost like a venture-style innovation portfolio. One at the moment is what you would probably call a strike drone. It's actually a munition, but it's a weapon system that is very much software- and AI-driven, and it's relatively inexpensive compared to other weapon systems out there.

We got a lot of great feedback on usability. We're only beginning to understand, for example, what it does to the mental load of a soldier who is supposed to operate those unmanned swarms. How exactly does that work? How should a box be designed? Should it be [wood]? Should it be plastic? How do we visualize a swarm on a small screen so the soldier can operate it?

Pat Grady: Earlier you had this interesting anecdote about plastic versus wood for the boxes that go into the field. Can you share it? I think it really illustrates the difference between the lab and the field.

Uwe Horstmann: A thousand percent. You often have these great ideas about how a product should be. You came up with the idea on the whiteboard, and then you put it in front of the users, the soldiers using it in the field, either in real battle or in training and simulations. And sometimes it's very, very different.

For example, we showed up with this beautiful plastic box that is light to carry. And the soldiers looked at us like, "What do we do with this? Why don't you give us wood? Because wood, at least, we can put in the fire, and it keeps us warm in the winter." I didn't think about that.

There are many anecdotes like that. The feedback is so valuable. Traditionally you would say the most important soldiers in product development and procurement are the generals, who tell you, "This is where the big budgets are, and you should do this and that." We always like to say that the staff sergeants, the enlisted personnel who actually use those systems, are maybe even more important to us. In a true startup spirit, you want to be close to the customer and develop for that customer. For us, the customer is the buyer, yes, but it's also the user, and in a military situation those are not the same people.

And we're just beginning to understand, for example, how exactly you control a drone swarm. Do you have a mouse? A keyboard? A touchpad? How does it work when you wear gloves? What clicks are okay? Where do you need to be really precise? What is the mental load? How much information can I process at the same time? We're only beginning to understand.

And then there's how you treat soldiers. They are used to industry just showing up, dumping stuff on them and saying, "Here, this is the best thing. If you have issues with it, it's because you didn't understand it." You really need to turn that around. You need to treat them as your customer. The most valuable thing they can give you is their time and their feedback.

Sometimes we show up and they say, "You know what? This is absolute shit." But you can really win trust by saying, "Tell us exactly how it's shit, and we'll show up again tomorrow or the day after tomorrow with a new version, and it'll be a little bit less shit." That goes a long way.

One of the things that makes me proudest of this company is that we won a lot of minds and hearts in the community, with people saying, "You know what? Those guys approach it from the right perspective. They have a soldier-centric view, and they really want to build a product that's good for us." That's quite obvious for startup people. Probably everybody who sits here says, "Yeah, we want to build for the customer. The customer needs to be happy." Net promoter score and retention and NRR and ARR growth. For us, the translation of that is: listen to the soldiers, build with them, build for them, and hopefully good things will happen on the procurement side as well.

Luciana Lixandru: Can we take a step back? You mentioned there's a favorable moment in Europe right now for building defense companies. And Uwe, you and I have known each other for a long time. We started as venture capitalists in Europe, and you saw the opportunity in defense before other people. You were investing in defense way before it was obvious, I would say. And you also have a background in the military yourself. Can you tell us, first, what came together in Europe, and why there's an opportunity to build a neo prime today?

Uwe Horstmann: I think what came together is obviously an understanding in Europe that Europe needs to take more responsibility for its own defense. That is, of course, something the US has told us repeatedly. And they're right. Europe has to step up and needs to be able to defend itself, because that is our way of continuing to live in peace. It's deterrence, right? We probably understand now that the world is unfortunately not as peaceful by default as it ideally would be. It needs to be defended, and the only way we can be safe is by deterring a potential attacker.

So that understanding has really materialized in people's minds. And it came together with a point where people also understand that something is happening on the battlefield. New technologies are emerging. They're software-driven, AI-driven, relatively inexpensive to produce, small-scale.

Those two points came together. And we give a lot of credit to the Germans and other governments who also understood: if that's the case, we should, A, procure those solutions where our capabilities are lacking or there is a gap, and B, change the process of how we procure, because you don't procure a drone like you procure a battleship. It's just different.

So that came together, and that was probably also the starting point for us, coinciding with the founding of companies like ours: "Okay, now we need to lean in." Is this going to be one of those moments where, in a few years, we'll ask ourselves, "What did we do during that time?"

And you mentioned my biography. For me, two big streams really came together: an interest in military affairs, security, international relations, things like that, and then the technology part. So it was quite a clear moment for me to say, "Okay, I need to step in."

Luciana Lixandru: Do you think the procurement process changes are sustainable, or is this a moment in time for Germany and other European countries?

Uwe Horstmann: I think they are sustainable. Changes like that in Europe don't happen because a charismatic leader says, "We need to change. This is the way. Follow me. Here's a great narrative." That's not how Europe works. Europe works based on, "Well, I guess we kind of have to do it now." It's reality kicking in. The positive side is that, hopefully, the changes are more lasting, because they're based on actual structural, fundamental changes. We are in a new world now.

That's what gives me hope about the changes in procurement processes, which went from years to months. That includes certification, which is a big part of what we do. It's a very boring topic in a way, but it's very, very important.

I think those changes are here to last, because the real learning from the Russia-Ukraine war is probably not drones, or not only drones. It is that an innovation ecosystem is the actual wonder weapon. We always like to read on the internet about some new weapon system and how awesome it is. That doesn't really matter. The whole point is: how do we, as Western societies, set ourselves up so that we can continuously innovate in defense, with all the different ingredients? We already talked about the input from fighting soldiers. There's a talent aspect: our smartest minds should be willing to work in defense. There's a capital aspect, and all that stuff. I think that is here to last. I'm quite hopeful about it.

Pat Grady: What advantages do you guys have as a two-year-old company selling into the defense apparatus that a 70-year-old company selling to the same customer might not have?

Uwe Horstmann: In that segment of the market, the inexpensive, AI-driven, mass-scalable systems (our strike drones come to mind, but also, for example, unmanned boats), we have startup DNA. We built the prototypes pretty much like you would prototype in a startup, and we built them from the get-go to be mass-manufactured items, which is not something defense typically does.

Just a few days ago we had some operational experts in our factory, and they looked at us and said, "Oh, this is about as complex as a washing machine." I don't know if they were trying to insult us, but it is exactly where I want to be. If I tell you, "Today we're building a thousand washing machines a month, and we need to scale this to 10,000 a month, or to a million a year," you would say, "Okay, it's not easy, but it's doable somehow." That's where you want to be.

The established primes are not necessarily great at that. They're really good at building those exquisite, very important, very expensive systems, and I think they will still play a big role. But there's a part of the industry that's about rapid innovation, ongoing development close to the customer, understanding the customer and building for mass scalability, where I think we have a pretty good edge.

Pat Grady: Awesome. Johannes, can you say a word about the product engineering organization? As Uwe was saying, how do you execute on rapid innovation close to the customer? It's not the thing we typically associate with hardware companies selling into defense. Say a bit about the organization and how you move fast.

Johannes Schaback: At its very core, Stark is an engineering company. It's everything we do. The way we organized the company is by building programs that can act and build very autonomously. As Uwe said earlier, we're running the company almost like a venture fund, where we're building small startups within the startup.

So engineers who join Stark very quickly find themselves in a family, a program that works very autonomously and very cross-functionally toward the mission of the program. That can be fully autonomous maritime systems, fully autonomous naval or aerial systems, strike drones, electronic warfare, or software.

Pat Grady: So is there a platform layer underneath these different programs? How does that work exactly?

Johannes Schaback: The programs serve external customers. And then there's a platform layer that we call capability teams. They build capabilities that augment the ecosystem we're creating: for example, GNSS-denied navigation, object classification, target re-identification, communication links, how we go about flight control. There are certain pieces we then bring back as compounding value into the family of systems, into the programs. So as an engineer, you can either work on the platform and help the entire family of systems, or you can be more customer-centric and work on a frontline system.

Pat Grady: Which capabilities are you most proud of today?

Johannes Schaback: We're extremely proud of being very, very accurate at hitting targets. That's what we're extremely good at.

Pat Grady: What does that take?

Johannes Schaback: It took a very integrated approach, bringing hardware and software tightly together. Just building a hardware system is good, but just building a separate, isolated software stack that is able to cope in a simulation, for example, is not enough. Actual submeter precision for striking a tank, or a moving target traveling at 60 km/h, requires fine-tuning the settings and parameterization in terms of compute budget and memory allocation, all done perfectly for the system to fulfill its purpose. That is what we're very, very proud of, and what we accomplished. Also, just recently in certification, I think it's fair to say that out of several dozen shots, every single one hit.

Pat Grady: You guys had a moment about a year ago where there was an experiment you were running, and the experiment didn't quite hit the mark. It's an experiment, and experiments don't always hit the mark. But it was reported, and there was a big hubbub about how it failed. And you had, what, four weeks or something to turn it around and pass an even harder test. Can you say a word about how it felt to go through that, and how the team rallied to get the job done? Maybe the details on what you were trying to do?

Johannes Schaback: That was a forging moment for Stark. Stark, in many regards, was born out of fire, but that was a particular moment.

To get the context right: we had a major upgrade in one of our systems, in hardware as well as software. As every engineer knows, when you have an update, you have a regression. The center of gravity is off, or since the opening angle of the camera changed, you have to retrain some of your edge perception models. That's just how it is.

However, the Bundeswehr, the German army, asked us to participate in a very important limited operational experiment, as it's called, and we couldn't say no to that. So we went there with a system that we knew was not at our ambition level. At the same time, we couldn't say no, because we also thought the step wasn't so big, and we weren't able to fully test. That's number one. Number two, we tested in an environment that was not ideal for the setup we had at the time.

Now, the media was all over it. I think the Bundeswehr was actually fairly relaxed. They knew they wanted to bring us to the edge and stress us out, so in their minds it actually wasn't that bad. However, the media picked this up and went all in on it.

So what we did is essentially an all-hands-on-deck meeting. We said, "Okay, everybody, this is a make-or-break moment for the company. This is where true engineering excellence will show. The best minds of the engineering community at Stark have to come together."

We booked a three-star spa hotel, just to be close to the shooting range we had access to at the time. It was in November, actually. It was extremely cold. And we worked three shifts around the clock.

While the flight test team went out and collected data, they were sending it back. We had several containers set up at that time. We even had mobile generators and Starlink, basically bought from DIY stores very, very quickly to set up a base with the necessary infrastructure to collect data. As the flight test team was flying in the extreme cold, the software engineers took that data and kept optimizing the full stack: flight control as well as perception. At the same time, some of the hardware engineers helped improve the actuation and tuned the system.

In the course of only two weeks, we got to a hit accuracy that was perfect. But there was this moment, because we said we would only leave this place once we had three consecutive spot-on strikes.

By then the tanks we were shooting at were totally perforated. We wasted a lot of systems on that, and they're not exactly cheap. But that was clear: no questions asked, money doesn't matter, we have to get to the precision. So we worked 20-hour shifts. When the flight test team and the hardware engineers came back, they repaired some of the systems. Since we were flying and shooting so much, we basically had to salvage debris from our own systems and put it back together in order to get a few more shots.

So that was one important workstream happening on site. At the same time, the simulation team helped us further improve the flight dynamics modeling. We were already employing AI systems at that time to train flight dynamics models, to improve the accuracy of the simulation so our AI models could be trained better within the simulation. These two workstreams then came together in the end, after two weeks. As a matter of fact, it was less than four weeks. There was this one moment where we realized, "Okay, that's it." And from there on, we've always had [sentence trails off; check audio at ~21:10].

Luciana Lixandru: Johannes, if you're building a software company today, it's very clear where the talent pools are. If you're building a defense company, it's not as clear. Tell us a little bit about how you find the best people and where you look.

Johannes Schaback: We look for talent in classical industry, but also at universities. We look at startups. We also look at government institutions. We look everywhere, because the talent that has to come together has to come from everywhere. We need such a breadth of functions, whether very senior or very junior, that we have to bring all the greatest minds together. Someone from an established prime usually comes with knowledge that is very, very valuable. At the same time, the startup mindset is equally important.

Pat Grady: So they're coming from different domains and different cultures. How do you stitch all that together and get it to work cohesively, particularly as you scale so quickly?

Johannes Schaback: The programs build products for the outside world, and they're cross-functional. Within each function, there are function heads who bring cohesiveness and consistency across electrical engineering, mechanical engineering and the various other functions. They build communities of practice across the cross-functional teams. That's how we share knowledge and best practices, but also tooling decisions, AI approaches, AI workflows, skills (very important) and [MCPs] across the company.

Uwe Horstmann: There's definitely a cultural shift in the tech and startup community. A few years ago, it would have been unimaginable for people to go into defense. It was not something you would do. That is really changing. We observe it all over Europe, in all the countries. People understand this is a generational challenge, a generational task. It's time to step up.

We see people who have had no touchpoints with defense before saying, "Maybe this is something for me." And for them, people like us, or Stark specifically, can be a good first touchpoint, because we are as much a defense company as we are a high-growth tech company. We understand both worlds. For people who want to apply their technical skills or their general skills to defense for the first time, because they think, "Yeah, this is what I want to contribute to," we're a very good first home. You don't need to have worked in defense to work at Stark.

Pat Grady: As a way to illustrate the technology itself, could you walk us through a single strike, from launch to strike? What's happening? What are all the different systems? What are the inputs and outputs?

Johannes Schaback: For sure. I'm just wondering where to start, because I'd start at the logistics level. That's what matters. You don't win wars just because you strike; everything leads up to the strike. So working backwards from that, it always starts with logistics. We just heard about the wooden boxes, which are important. How do you disassemble and reassemble the system? That's extremely important. How do you transport it? It has to withstand the shocks and vibrations, and the pressure if it's transported by air, or by sea. There are various effects and environmental conditions the system has to withstand. In some cases it actually has a shelf life of, say, ten years before it gets deployed to the front line, where the system is then assembled or put on a catapult.

Then the concept of operations for all our systems is that an operator prepares the intent. The system always implements and follows the intent. Think of it as a mission. The intent is modeled as, say, no-fly zones or a kill zone, or certain other elements that make up the mission. When the system then takes off, it automatically and autonomously finds its way, evading no-fly zones and anti-air installations. If it's operating in combination with a swarm, it communicates what it sees across the swarm, so the members share situational awareness.

Pat Grady: Would that work only if the swarm is all Stark products? Or is there some communication that works across different systems?

Johannes Schaback: We're very open to that. Extremely open. We'll probably talk about modularity as well. But just to finish: the system arrives at the kill zone and starts looking for the [target]. Then, depending on how the customer implements human-in-the-loop, either the system fully autonomously seeks and destroys the target, or, if there is a human-in-the-loop element, it waits for confirmation by the operator and ultimately fulfills its mission.

Pat Grady: So it sounds like there's a lot of high-speed, very precise computer vision processing on the device.

Johannes Schaback: Correct. The units need to be fully autonomous. That requires perception, and this perception is fed into a reasoning engine that acts on the environment and the conditions in order to fulfill the mission. So there's a higher-order corridor, you might say.

Pat Grady: Do you need to use your own ASICs to make that work?

Johannes Schaback: We buy various components from suppliers to make that work. There's an element of recency: as we keep learning, there are new target types. The war in Ukraine is changing daily. So there's a lot of modularity and flexibility built into the systems. We buy a lot of components that are procurable in large quantities, such as infrared cameras or electro-optical cameras, bring them together and sensor-fuse them to give the system the situational awareness it needs to act autonomously.

Pat Grady: How do you guys think about where you need to be vertically integrated and where you want to leverage the ecosystem or the supply chain that's already in place?

Uwe Horstmann: The governing principle for us is: how do we solve the customer's issue? And scalability of output is a big customer issue. So that would be my first question on vertical integration. A strong argument for vertical integration is if we can solve the scalability issue through it. That's one force pulling on it.

There's another force that goes in exactly the opposite direction. It's so helpful for us, when we sell these systems to other countries, if they can plug in their own suppliers, because we want to root the production of these systems in the countries where they're being procured. So that's the ongoing balance. Principle number one: can we solve the scaling issue? Number two: how can we plug in different suppliers to make everything more resilient?

Luciana Lixandru: You've talked a lot about your strike drone product. What's your philosophy around being a single-product company that's the best in the world at that one product, versus being multi-product from day one?

Uwe Horstmann: We look at where the capability gaps are. Those are basically self-described by the military, saying, "Hey, we have a capability gap here." We observe them quite closely while they're being analyzed and formulated. Then we look at them and ask: Can we be a good supplier? Can we [provide] it? Do we have a right to offer a solution for this? And is there one on the market already?

That's the first exercise we do when thinking about a product portfolio, and it pretty quickly gets you to multi-product, because we don't see an abundance of offerings in the things we do. So it's almost a mandate: we have to try. We of course always have to counterbalance that against the complexity and loss of focus of multiple programs. But our first reaction would be: there is a capability gap, it's evolving, we try to contribute to the definition of that gap, and then we will always ask ourselves, can we help? Can we contribute?

Multiple things, on the technology side and on the concept-of-operations side, can be taken from the aerial domain to the maritime domain, for example. What Johannes mentioned, intent planning and mission planning, doesn't depend on whether the asset is flying or swimming. There's navigation in denied environments, whether it's jamming; that happens in both the maritime and the aerial domain. There's object detection and recognition, again the same. There's even some targeting, where I'm homing in on a target. So there's a lot we can use across domains, as we call them. Or dimensions, if you will.

Luciana Lixandru: Can we talk about a topic that may sound boring on the surface but is really make-or-break for a company like yours, which is manufacturing? How do you go from zero to manufacturing at real scale?

Johannes Schaback: When we start a program, there are two goals. We optimize for time to market, and second, we optimize for production readiness. They're often partially competing. If you're optimizing only for time to market, you may quickly end up with a product you can show and demonstrate, but that is very difficult to manufacture. Take a Formula 1 car: you only build one. You're very quick to be ready, but you hardly ever produce it in large quantities.

So from an engineering perspective, we shift left on manufacturability. That's extremely important, because otherwise you accrue a debt you have to fix later on. And that's really just the beginning, because the moment you decide which components you buy, supply chain resilience plays a massive role in the decision-making process early on as well. So, to the analogy of the washing machine, we may not always buy the greatest, most leading-edge camera, for example. We buy the ones that are very, very good, top quartile, but actually procurable in large numbers.

Luciana Lixandru: People in defense talk a lot about local supply chains. Is that realistic in Europe, or do you have to look at, say, China or other countries as well?

Uwe Horstmann: I think it is realistic, specifically for certain components where it's fairly simple, as a first step: "Here we can diversify and source from different suppliers and even from different countries." Conventional wisdom would ask, "Is that good from a bottom-line profitability perspective?" But it's probably good from a resilience perspective, and even more so [if you think about the resilience of the business as a whole].

There are still certain components where we, as the Western world, have dependencies, for example, on Chinese supply. We try, of course, to work on that, and we've reduced it dramatically. I think the current systems have hardly any reliance anymore. But there are also things where I think we as an industry all have to come together, with politics and the militaries, and figure out the strategy, where raw chemicals and ingredients are needed that maybe go beyond the scope of just one company. That's probably more of an industrial coordination effort.

Pat Grady: Speaking of supply chain, I want to ask you about a very different part of your supply chain, which is AI. Part of the future of warfare is not just the move from exquisite systems to attritable systems. It's also the move from humans doing everything to more autonomous systems. And we're sitting here in the wake of the Hugging Face hack, where these civilizations of agents, so to speak, broke out of their sandbox and wreaked havoc. So I'm curious: given the power of the system and the inherent dangers in the product you guys have, how do you think about the possibility of AI going rogue, or somehow not working exactly the way it's supposed to?

Johannes Schaback: Since the beginning, Stark has implemented NATO's Principles of Responsible Use of AI. That entails auditing our own systems, not only by humans but also with safeguards, layers that keep the AI within a corridor restricted to implementing the operator's intent. That is what governs the entire autonomy stack at Stark, in every system and at every stage of the mission.

There are hard-coded constraints that depend on where the system is. For example, if it's flying over a particular inhabited area, it may not arm. That's one example. So there are certain safeguards, IEC 61508, and several things we borrowed and learned from autonomous driving that also found their way into our autonomous systems.

Pat Grady: More generally, I know autonomy is interwoven throughout the product and the operations of the product. When we talk about autonomous systems, what does it mean? Where does the AI actually show up in the product or in its operations?

Johannes Schaback: It shows up in every bit of the product, and in every part of the mission, whenever there's a decision to be made, be it about what the system perceives or how the system interacts with its environment. Also in how the mission is planned on the ground before the system actually takes off. That's a very important aspect, because mission planning used to be very cumbersome and long.

Pat Grady: Can we talk about that? How does AI play into mission planning? What's a good example?

Johannes Schaback: Imagine, as a very hands-on example, that there's a target coming in, and the operator has three systems at their disposal. The first decision to be made is: how many systems do I shoot at this target? What's the most cost-effective way of addressing this threat? Once that trade-off has been made, in our case the ground control system, called Minerva, optimizes the mission planning to address the threat accurately and most cost-effectively.

Luciana Lixandru: You saw a gap in the market and you went for strike drones, maritime drones and other products. Why haven't the existing, traditional primes figured this out as well?

Uwe Horstmann: I think they do to a certain degree, sometimes more, sometimes less. Some of them have leaned into this. And some of them, including some who speak to us, say, "That's just not our DNA. We're not an organization that builds cheap systems. We're a company that focuses on a different end of the market." Some are trying. We welcome that. One of the core ideas of what we need to do now is to start a competitive process. We want to compete against those companies and their systems, and jointly we'll get better. At the end of the day, we're all on the same team. So some are trying, some not so much, but we welcome all kinds of competition.

Luciana Lixandru: Arguably, they have more capital than you. How do you think about that?

Uwe Horstmann: Yes and no. Maybe this is an interesting fact: if you just look at how much capital we intend to invest over the next 12, 24, 36 months, it's a sum of probably something like €600 million. Significantly more than half a billion dollars, potentially much more. That's all going to go into research and development and building up production capability. When you compare that to the orders of magnitude of R&D spending at the big European primes, and you subtract all the R&D spending that is government-funded and look at the actual core R&D spending those companies do, we're not way off.

That's not automatically a good thing. We want to stay nimble. We want to focus on our segment of the market, but we want to make a meaningful contribution. What I said earlier about understanding the customer, building exactly what the customer wants, getting that feedback from the staff sergeant: it needs to be different. Training, iterating with them, that's what we intend to do. And I think we're in a good position, with the startup spirit and now with the capital in our hands, to hopefully build a real pillar of the Western defense architecture.

Luciana Lixandru: You mentioned the capital you raised. You've done a lot in two years. You also went from zero to 700 people in two years. Is it chaos? How do you handle all of it?

Johannes Schaback: First and foremost, the blueprints of the company need to be crystal clear. We spoke about that earlier. Number two: yes, growth always comes with the trade-off that there are sometimes hard decisions to be made. However, that process is very manageable. [We keep improving how we grow our talent.] Generally, we feel the blueprint works very well. But we also need to be clear that the problem we're helping to solve is now. So we have to be fast. We have to make a dent now. Being slow is not an option. So allocating these 700 people (not all engineers, but the majority are engineers) to extremely important and valuable programs is key. That's what we set out to do. Are we at the end of the journey? Of course not. But we're very determined to succeed in these programs.

Luciana Lixandru: Is there a very hard problem you're trying to solve right now? What's on the roadmap that you haven't figured out yet?

Johannes Schaback: What we always see in Ukraine is that systems are flying faster. That's something we're leapfrogging into as well, a lot. We see compute becoming more accessible, so we're thinking a lot about how we can further drive AI on the systems in a safe manner and give it more capabilities: for example, to be adaptable to targets on the naval side, and maybe in the air, not only on the ground. There are various areas we're going into. In the electromagnetic space, too, there's a massive disruption happening, thanks to software-defined radios and affordable compute, which brings down the cost of sensors for the electromagnetic space. That is extremely interesting.

Uwe Horstmann: I'm really interested in solving the next big thing, which for me is contract design. I think there's a big deterrence possibility in the way we structure these contracts. It's one thing that we sell hundreds or thousands, or potentially tens of thousands or hundreds of thousands, of these systems. But what's really interesting is whether we can come to a contract design between the government and us, and potentially others, that already has everything in place for the massive surge in scaling we would need if there's ever a hot conflict. That would be very interesting. All of a sudden, we maybe think about this more in terms of cloud contracts, with surges and so on. If we get that done, that is probably where the real deterrence comes from. It's not about the individual capabilities of one product here or there.

Pat Grady: What would that look like? There's some stockpile, presumably, but then they prepay for manufacturing capacity that you can just spin up in the event something...

Uwe Horstmann: Exactly. One way or another. There's a big question about the supply chain. There's a big question about storage of systems. There's a big question about who updates those systems. If you look at a drone at the moment, the motors are probably good for three years, but the software probably should be updated every week with new training data and so on. So how do we structure all of that? That's something very interesting to me, because that's one way you can create real deterrence, I think.

Luciana Lixandru: Has anyone in the Western world figured this out?

Uwe Horstmann: I don't think so, necessarily. There's no blueprint. The same goes for everything we do in procurement and product development. It's not like hundreds of people have done this and can just teach us. I think we all learn together, and that, by the way, cuts across both sides: the government side, the military side, the industrial side. I do think there's an optimistic spirit of, "Okay, we need to figure out a few things, because we don't have the right toolset just yet." In that spirit, I think we need to solve it together. I don't think anybody has the perfect playbook: "Here's how it's done, just copy this." We don't have that yet, but it makes the whole thing so exciting. It's at the forefront of what's technologically possible, and it's very clearly what's required to be able to survive in those battles. Bringing those two things together allows us to make a real contribution to continuing to live in peace.

Pat Grady: What is the mission of the company?

Uwe Horstmann: The mission is quite clear. We want to be one of the pillars of our European and Western defense architecture. That means, yes, we want to be a sizable company that delivers the capabilities we talked about, but we also want to help orchestrate a real innovation ecosystem. We want to be a good collaborator with other people. We want to work with the large ones, the small ones, our peers, with everybody. Because when we think about the one mission for the company, it's very closely linked to all of our missions. That's very important for me to stress. We want to be a pillar (a dominant pillar, of course), but it's not a mission that can be achieved by one single company alone.

Pat Grady: And if you do achieve the mission, and we're sitting here 5 or 10 years from now, what does Stark look like? What does Europe look like? What does the world look like?

Uwe Horstmann: I think Europe will become a better partner for the US again. There will hopefully be a new chapter of transatlantic cooperation. Europe will be a helpful, strong, principled ally, among ourselves and toward the US and other partners in the world. And hopefully we'll live in more peaceful times, where we go back to how the world was a few years ago, where we've de-escalated some of the wars and situations around the world, and others have never flared up.

Pat Grady: Amazing. Uwe, Johannes, thank you for joining us.

Uwe Horstmann: Thank you so much.

Luciana Lixandru: Thank you.