Manufacturing 1 million drones a year | Olaf Hichwa, Neros

Today I'm sitting down with Olaf Hichwa, Co-Founder & CTO of Neros, and their new quarter-million-square-foot factory that's on track to produce a million drones a year by something like 2028. I was here about 4 months ago, interviewing Soren, when this was an empty box.

Big empty box.

Big empty box.

Not a thing in it.

Nothing in it. And now you guys are almost, I think, completely out of the old place, or soon to be.

We're completely out of the old place. That is now a small empty box, and we have a slightly less empty box here.

But this is going to be at, I think, 130,000 run rate drones per year.

This year.

By December.

This year.

So tell me about just starting up this entire production line from scratch, literally in a couple months.

Man, ground-up redesigns. They are awesome. We were so cramped in the old building, and so much of our kind of production process was dictated by initial decisions by the line of folding tables that we put together about 2 years ago.

So getting to design something fresh, design, you know, a production line with component feeder stations that go into each assembly station, it's been huge. And the numbers have been huge. We've been just consistently ripping, you know, 1,000 drone weeks, 1,200 drone weeks, and more soon.

Can you put it in context? What is 1,000 drones per week?

Yeah.

Because there's, like, no company in the United States producing that manyright now.

Right now, we build more drones in a week than any other drone company in America builds in a month. Every 2 minutes during business hours, we have an Archer coming off the production line. And that's, you know, America's state-of-the-art FPV strike drone.

It's a world-class piece of engineering.

There was another podcast that you did where you talked about the number of casualties in Ukraine and Russia from drones versus any other weapon. And I thinkright now it's something like 70% of all casualties are from FPV drones, I believe.

Yeah.

This dwarfs any other weapon in history.

Yeah.

As far as, like, number of deaths.

We're living in a crazy time. Never in human history that I'm aware of has a piece of technology come out that has had this big of a swing in how conflict is fought. You know, 70% was this big stat popularized by a Reuters article and some research.

Zelensky came out and now is quoting higher. You know, I hear 80, I hear 90. But at the end of the day, when has there been a time in human history when a piece of technology has come out that has been as impactful to the battlefield as drones?

You know, maybe nuclear weapons, but even then, it was, you know, not as big of a percentage as drones. The front line is just the drone zone. When you drive around, all you hear is drones. Drones are going to be, and already are, the defining weapon system of our generation.

And, you know, it's really critical that America and the free world has a play in that. If we let, you know, another group of folks define the weapon system of our generation, I don't think I need to exactly walk through what happens.

You know, Pax America falls.

I would like to talk about, like, the current drone market today and where it can kind of transition to maybe 5 or 10 years from now. You said earlier today something on the order of, like, 14 million drones are being sold from China to both Russia and Ukraine, like 7 million to each side, and they're basically just blowing each other up.

Talk about a business model. I mean, it's terrible. And I, you know, I feel bad for the Ukrainians that have to put up with, like, you know, their vendor. You know, generally, like, we don't like it when our vendors supply to our competitors.

We put up with it. Imagine, you know, you're giving feedback. You're a Ukrainian drone manufacturer. You're giving feedback on your design based on the blood of your comrades, and that information is going directly to your enemy. And often the Chinese give Russians the newer drone tech,right?

The first terminal guidance videos I saw on Twitter were from Russian drones. These are AI-automated drone strikes. Fiber optic was also, like, I saw it on Russian X accounts first.

How do you think that's going to shift over the next couple years as you guys start to ramp?

I think the kind of Western drone base is slowly, slowly coming online. People are starting to realize that, you know, drones are the most lethal weapon system of our generation, and they will define who runs the world. And we're starting to put two and two together.

We're like, okay, we should probably build them. And so you've got factories springing up. You know, we're not the only company with this million-drone dream. I'm seeing new announcements, and other people are saying, "We're going to build a million drones a year."

And I think that's awesome. I think the West needs a lot of million-drone factories because China has that as well. So I'm hopeful, you know, in a reasonable amount of time, Ukraine will be able to have a credible alternate to China in the West, in Europe, in America, in Millennium One, in this building, but all over.

Earlier today I was in a meeting with you, and you're talking about basically pulling in timelines and making sure that you can actually ship products.

For some reason, it's always in, never out.

Yes, yes.

Always in.

What, like, trade-offs did you make on the design of Archer so that you could actually ship it in a fast, rapid pace and scale your production quickly?

Reality is a very powerful thing,right? When your engineering project is in simulations or in low-volume demonstrators or in demo mode, you get to have a lot of control over the world, and things are pretty perfect. When you build things that serve a, you know, critical national security objective and are made in massive volume and are used by people whose lives depend on them in volume, you don't have very many luxuries.

And you're kind of forced, like, the kind of I've noticed the Elon algorithm is kind of forced upon you. The insane deletion isn't really like a luxury. It's like a, you know, you will fail. You will not deliver anything.

Your volumes will be tiny. Your timelines will slip. You have to be vicious in the deletion. Every requirement gets questioned to shit. We push back on all of them. You know, you want picture-perfect drone swarms that, like, you know, operate, like, in, you know, thousands of units and behave like a wild animal swarm, like a fish swarm.

You don't actually get that requirement. That's a demo requirement. And you have to be much more real. You have to think, like, what do the users want? What has worked when we deployed? What can we manufacture? And those kind of realities are just hammered into the engineers that work on Neros products.

A lot of people talk about, like, autonomous weapon systems now, but you guys, very specifically from the beginning, are basically trying to focus on creating the best possible thing for an operator.

Yep.

And now you're, like, moving into slightly more autonomous systems, but mainly still focused on the operator. Why did you do that? Why didn't you just go straight to autonomy?

We got to look about how the first Archer ever was designed. You know, Soren showed up at my office and was like, "I know some Ukrainians from Twitter. Just get on a call with me, and they will give us some good feedback on what we need to do."

So it immediately was, like, user obsession from day one. And then what did the users want? They didn't ask for an AI-powered drone swarm with 1,000 units. They asked for, "How do I set off the bomb? How do I make the radio not get jammed?

How do I go longer range?" And so it started there,right? And then we really believe in getting something that works to get feedback. Because if you spend too long just designing in a vacuum, when you have that contact with reality, everything falls apart.

Your assumptions are wrong. So we generally focus really heavily on getting something to the field quickly to iterate on. With that being said, you can't just continuously operate on those 6-week engineering timelines. We touched on it last time.

You need a little bit of time to have breathing room. And we do that by first usually deploying, like, an initial system, getting feedback, and understanding what the true pain points are. Some of the true pain points we've noticed is, like, operator burden is a big deal.

If a user has to look at the drone and carefully babysit it for the full, like, 30-minute strike, you know, you take off and you're commuting for 25, 30 minutes, and then going for that.

Is there some, like, a loss of focus?

There's a loss of focus. And, you know, you got to take care of your buddy. You got to watch your surroundings. You have to monitor maybe another drone. We start to learn what breaks with a full manual solution, and we fix the parts that break.

And I think it's, like, people assume autonomy is like a switch. They go straight from no autonomy to, you know, 1,000-unit AI swarm, and things fall apart so quickly. Those swarms are so finicky, and they just do not survive first contact with reality.

They're so, you know, perfectly designed for demo environments with GNSS, with other, like, unrealistic features. Like, they assume that your targets won't be moving. They assume your targets won't be occluded, and they assume they won't have, like, destructive patterns.

All these, like, edge cases of battle aren't considered. So generally, when we're working on products, we're focused on what gives America an edge in the fight. And that is usually things like extreme manual precision and things that decrease the operator burden.

This has been, like, one of the big design philosophies behind Archer AI. This is the product here, and this is actually a product we're launching today as part of this. So Archer AI is an Archer with an autonomy stack.

So in the front of Archer, we've got a computer vision system that can use AI to lock onto a target and fly itself into the target once you get jammed or go over the RF horizon. This is a very, like, high-user load part of the equation,right?

And it allows the user to just command a drone to go into the target. The user can take over. It uses our same industry-leading anti-jam radio solution. The user can readjust where the AI thinks the target is in real time.

If the AI messes up, if, you know, there's something, some countermeasure that the enemy has figured out, the user can readjust. But the drone will go into the target even if you lose comms, even if your GPS denied.

And I think this is, like, a really big step forward for what works. Like, really regularly, like, let's say you wanted to, you know, go use Archer against an enemy in the battlefield. One of the first things people run into is you go over the RF horizon,right?

If you try and go for the strike, there'll be trees or a mountain or something in the way, and then the radio won't go through that. Archer AI gets around that because once you, you know, once you find your target from up high, you could tell the drone to go strike even if it doesn't have communications back to the user.

So you can dip below that radio horizon. But these kinds of kind of decisions on what we build are always really informed by what the user really needs and really works. They're not often informed by, like, what, you know, certain folks in the government that maybe don't understand the users really well want in a demo or what Congress wants in something.

We're much more focused on just, like, build something that works for the users and then let the users be your salespeople.

Okay, this was actually something that you talked about earlier. You don't necessarily want to listen to government because the government doesn't know what to build.

Yeah.

How do you build something that you sell to government when the government doesn't know what it wants to buy?

Good question. It's a good question. We want to build something that the users really, really, really want. And of course, like, it is a massive generalization to say the government doesn't know anything about what they want. There are exceptions to this rule, but generally.

They're, like, directionally correct?

They're well-intentioned. They're well-intentioned. I'll say that. Yeah, very well-intentioned. And there are smart people in the government that are directionally and actually correct. We find it most useful to do the digging ourselves. We go to Ukraine. We meet with users.

We reach out to users. We respond to every customer support ticket very publicly. We have feedback sessions every month with that team. And then we tell the government what they want whenever they'll listen to us. And we'll say, like, "Hey, this worked in Ukraine.

This is going to be deployed in Ukraine. This is already working in Ukraine. Here is our library of videos of the system working in Ukraine." You know, we have hundreds and hundreds of successful missions flown on Archer today.

There's very few companies that can say that. I'm not aware of another defense tech startup that can say that, period, at least at this magnitude. And smart people in the government listen to that. They go, "Oh, well, they did the user study.

You know, don't trust, you know, any one company. Go talk to the Ukrainian users. We'll make an introduction." And the Ukrainian users will be like, "Yeah, check it out. I'll show you my video of this thing working." And the government, the smart people in the government will be like, "Oh, yeah, maybe that holds some weight.

Maybe we should set the requirements around what they're talking about and allow for some open competition from there." So I'd say folks are really well-intentioned. They just might not have the resources to run, like, the user studies that we run and the trials that we run.

When you first started the company, you and Soren built, like, 30 drones, then just literally put them in your backpack and went to Ukraine.

Suitcase. Yes, yes.

Suitcase.

American Tourister.

Okay, suitcase.

The cream of the crop.

With this sort of, like, with launching a new product after you've already built a brand and built a reputation, and you also have the credibility to say, like, "This thing will work," is there a different decision process for launching the product if you're going to do something completely new?

Totally. I think we've built a brand of products working, and that brand can come down just as fast as it comes up,right? Like, we only have to have, like, one bullshit demo product for users to no longer trust us.

People's lives are at stake with this thing and with products like this. So we have to be a little more careful maybe than just, like, "Hey, let's try it." We also have to be a little more intentional in what we're doing with a product.

You know, we'll have user groups that we're really close with and we really trust, and we'll be like, "You know, we don't know if this thing's going to work. We don't know if this is a good idea yet, but let's give it a shot.

Let me get your input." We take that input very seriously before we, you know, go public with something like Archer AI.

So there's this idea of, like, focus.

Yeah.

Right? And it's very easy to get distracted and work on, like, multiple things before you're ready. How do you decide now is the time to start building multiple products in parallel?

That's a good question. I think you need to kind of earn yourright to take the distraction. I feel like Archer is finally on a path to a million drones where it seems plausible. You know, we're, like I said, building 1,000-plus per weekright now.

By the end of the year, we'll be doing 130,000 a year run rate. And we've got, you know, great plans for next year and the year after. We have some, you know, preliminary designs on that million-drone line. It feels like we're ready for, you know, Archer is set up to really hit its stride where we can, you know, actually start spending a little bit of our company bandwidth, our brain power on some other innovative products like Archer AI.

I think that it's kind of a necessity to offer multiple products because not every battlefield solution can be solved by one product. You know, it has to go back to what are we doing so America can credibly maintain its position at the top of the world, you know, stage in deterring conflict, in saying, you know, "Don't fuck with us because we have the largest, most powerful military in the world, and we're going to maintain peace.

If you don't want to maintain that peace, if you don't want to maintain Pax America, we have something to say about it and something to do about it. We want to be able to maintain that position." And I don't think America will be able to maintain that position as well if Neros just builds one product, if we just build the Archer regular variant.

I think America needs small, low-cost attack drones. It needs AI-powered small, low-cost attack drones. It also needs interceptors. Base defense is a huge problem,right? It is, like, I'm sure you've seen the news in the Middle East. We need to be able to blow up incoming drones, and that's why we're building Bandit.

We're kind of expanding to products because America needs it, not because it's a cool, you know, demo or it's going to kind of make an investor happy. We build it because America needs it.

They're on shooting drones down. Why did you guys decide to go with just, like, an interceptor instead of, like, lasers or some other form of, you know, interceptor?

It's a good question. I think from the first trip in Ukraine, it was just so clear. Like, the entire country had transformed to a drone country. You know, FPV has always been this, like, niche kind of thing that, like, weird people on the internet do, and there's, like, random YouTube channels and that kind of thing.

Walking around in Kyiv on that first trip, like, there were, you know, shirts and signs of these weird niche YouTubers. Like, the entire country had kind of rallied around this, like, and seeing that, it was clear that these drones were more than just kind of a fad or a toy,right?

These drones were the country's survival. It was kind of the manifestation, the early manifestation of this, you know, once-in-a-lifetime, once-in-a-generation shift in how military activity works towards drones. I think from that point on, it was really clear that the only weapon that can stop a drone is another drone.

The only thing that can stop a, you know, a group of drones is another larger group of drones. And that's really the logic behind drone interceptors.

So are you going to have to have, like, way more autonomy in the system for intercepting drones in there?

Interceptors really, so autonomy does best in structured environments. Like, you know, you drove my Tesla Model 3 today. The FSD on that, you know, still at this point, like, if someone walks out from behind a bus or, you know, walks behind one side and out the other, it's not really good at predicting that, you know, they're going to come out the other side.

It might run into them. Those kinds of unstructured environments are where it really struggles. If you think about, like, you know, humanity has spent close to a trillion dollars developing, like, full self-driving, and a Tesla Model 3 has more compute than, you know, hundreds, thousands of times more compute than we could put on a $2,000 drone, that makes the, like, go, you know, hunt an individual soldier task very hard.

But the, like, hit-a-dot-in-the-sky task is much more reasonable. Like, it's much more easily automatable. It's a much more controlled environment. It's much more math-defined,right? You just need to make these intercept paths. So Bandit will be a much more autonomous solution.

We're still going to have that manual precision. I won't go into details, but we have ideas on where it's really helpful. But Bandit will be a much more autonomy-focused product.

What was the biggest, like, technical challenge on building that?

I mean, we're, look at this thing. It's a 3D print. We're super early. The biggest technical challenge will be manufacturing it, as always. And we've built 10 or 20 of these things. This is serial number three. So the biggest challenge has not been hit yet.

Butright now, one of the big challenges that we're working on is making sure that we can build a component base that's compatible across these two products. So, you know, we've invested a ton into this production facility, and we have gotten really good at making drone flight controllers, the guts behind drones, drone motor drivers, radios, these components that are made in volume.

No one else has a factory like this. If we want to be able to really mass-produce Bandit, which is what is needed, of course, we should leverage this factory to build Bandit as well. We should, you know, use 80% of this production line to make components, and then those components should be shared.

So one of the big challenges was, like, working on, you know, a spec for Archer components, spec for Bandit components that overlap. Bandit goes a lot faster than Archer. Bandit goes north of 250 kilometers per hour in this prototype state.

It'll probably get a lot faster. And kind of misusing those components to go that fast has been a big challenge. I think also figuring out the kind of flight dynamics of this product. As you go faster and faster, the aerodynamics change a lot.

That impacts the control systems, and that needs to be really carefully tuned. But this has a huge set of problems. It's a really fun project because it's still very much like early days Neros. It's a tiny team developing this.

You know, we're flight testing all the time. We're crashing. It's failing. You know, there's a bunch of hot glue in this thing. It's very much an early stage, you know, mess-around-and-find-out product that will eventually, just like Archer, hit mass production.

What was it like when you first decided to start working on this? Like, what was the first few weeks and iterations of this product like?

Yeah, I remember that. I mean, it was nights and weekends at the very beginning.

And when did you decide? Like, what made the flip in your head?

Really, it was taking that call with the Ukrainians and just getting on those calls, really, and seeing the need. You know, I grew up in D.C., very normal, nice bring-up,right? Went to, you know, a good local school in the area and never really thought about, like, the dangers of getting bombed or anything like that.

It was just so foreign. And I kind of had the thought, like, why work on weapon systems when I could work on, you know, Bay Area stuff, like, you know, getting a burrito to arrive 30 seconds faster at a house or something like that.

Talking to Ukrainians really quickly changed my mindset. I can't believe I, like, how quickly it became abundantly clear that if you don't build weapon systems, everything falls apart. And not building weapon systems is a massive luxury. Like, being able to, you know, sip your matcha and take your electric skateboard to work to your Bay Area cushy job is something that, you know, every Ukrainian dreams of, I'm sure, but does not have,right?

They are worried about getting blown up by Russians, as, you know, that's a real threat. I think if Americans kind of remembered what history was like and what the generations before us did to make sure that we can live this good life, people would be much more interested in building weapon systems.

And I think that realization was really what made me, like, want to stay up all night designing the motherboard, the main PCBA, printed circuit board assembly in Archer at the time. So, yeah, that was, like, the first big engineering project, just ripping the motherboard.

I was a full-time board designer, so I had been doing it for quite a while. And I remember that motherboard. I bought it on my credit card. It was the first PCBA of my career that I designed that worked for a spit.

And that was really important because there was not time for another spin. That was our shot. So really, really thankful that that worked out. Thankfully, it was a lot of, you know, knowledge pulled from the drone racing days.

It was the exact same flight computer that I designed in high school as an example. But, yeah, those early days were awesome. It was just Soren and I, you know, cooking. He was up all night doing mechanical CAD behind the initial birds, and then we, you know, asked our friends to come and help and camped out in his basement in Vermont.

But great, great times.

You mentioned that you went to Ukraine recently. What was that like? What's changed since the first time you came?

The conflict has become more intense. Like, on the previous trips, like, my reaction in Kyiv, which is, you know, pretty far from the front line, was just, like, you can probably almost forget the country is at war. Like, there's long periods of time where it's totally normal.

And life always continues. That's one of the big learnings. Like, even, you know, 20 kilometers from the front line, you can have, I've had a great steak dinner with friends and discuss, you know, random topics. Like, life always continues.

But during the first few trips, it was like, Kyiv seemed, it was like, it's like Paris, but, you know, in Eastern Europe. I'd say the war has intensified a lot. The strikes on civilian infrastructure by Russia, it's awful.

It's, you know, an act of terror. And they're really, they're not pulling the punches,right? They're purposely targeting apartment complexes of civilians. And, like, you hear Shaheds, you know, up in the sky just buzzing by, and that's just, like, a thing that people have had to live with.

What is a Shahed?

It's a Group 3, which is DOD speak for big drone. It's kind of like the size of, like, a minivan, and it's just packed full of explosives. Just packed. And they're super cheap. They're, like, tens of thousands of dollars.

They have little rocket motors, so they launch up. And they're little Deltas. They're not little, big Deltas. It was actually an Iranian or a German design initially a long time ago, and the Iranians perfected it. That's why the name.

And then there was this, like, crazy deal where Russia shipped bars of gold, literal physical bars of gold to Iran for the design, and then Russia started producing them. And now they send, you know, hundreds of these things every night to Ukrainian cities to blow up Ukrainian civilians.

And worse than just hitting the apartment buildings often, they're really, they're not holding any punches. They're blowing up power, which if you, like, if your power station goes down, people freeze. They're really just trying to blow up anything that makes the lives of ordinary Ukrainians trying to defend themselves hell.

And those impacts are more clear in Kyiv now. On one of the nights I was in Ukraine last time, I actually went out with a military unit that was hunting Shaheds, and that was a really impactful experience. Kind of hanging out with the guys.

It's a much less official operation than you might think. It's, like, four guys that know each other really well. You know, two guys, they're nerdy. They sit back. They build the drones. I remember at the time they had, like, a fleet of five interceptor drones that they were very proud of and, you know, like, talking about maybe soon they'll get a sixth, but these are, like, really prized possessions.

And they've got a van, and they take the, you know, pack the van full of, you know, a Starlink, a few of these interceptors, some really cool software, some radio tech they've been experimenting with, and they go park in the middle of the field.

And our shift was, like, 8:00 p.m. to 5:00 a.m., and we were stationed basically just looking for Shaheds. And, of course, like, you know, it's a lot of waiting, a lot of fucking with each other and maybe, you know, rotating, sleeping a little bit.

But I remember, like, there were, you know, a few false alarms. There was a Shahed headed towards us. Often these anti-aircraft or air defense positions, like, you want to take them out,right? You want to blow them up so that way you can break a hole in the air defense and go through.

So we got on high alert, but another unit went and killed the Shahed before we could go and take it out ourselves. So, like, a few false alarms. The Shaheds, they, like, we saw them on a map, and they would take the most non-trivial routes.

Like, the goal is to hit Kyiv, usually, most of them is to hit civilian infrastructure in Kyiv. Like, it depends, but a lot of them are trying to get there. And you'll see a Shahed, it'll go, like, two-thirds of the way to Kyiv, and then it'll turn around.

And the reason for that is it just learned that there's a gap in the air defense, and it's going to send the rest of the Shaheds to go exploit that gap. Or it's trying to find another gap from the inside.

That way they could send more. So sometimes it doesn't make sense to just go strike directly, even if you're close. It makes more sense to gather information on where the air defense units are and where they could be taken down.

I remember close to the end of the shift, one was getting closer and closer, and it actually got in our bubble. And so we launched. We launched the interceptor to try and blow up the Shahed, and we got close, but another unit actually blew it upright before, and we were able to hear the boom.

That was terrifying. But just the kind of realness of, like, it's kind of like a video game in that you just have a map with, like, all these birds or drones flying over Ukrainian airspace, like, you know, hunting.

And it feels very, you know, fake. Like, it's just like you're looking through a screen until you hear the bang, and you're like, "Oh, shit." It's very, very real. Like, there is a real world on the other side of this, you know, camera lens.

But I think it's the future of conflict. Like, I think people are going to be, you know, more and more just fighting through screens. That's really obvious, but it's more and more going to be, you know, drone on drone.

And drones are, like, you know, this kind of, like, apex predator of the weapon ecosystem. It's like, you know, rock, paper, scissors, drone, all the way at the top.

Because it costs almost nothing, and it's super, like, precise and able to deliver a lot of force.

Yep, yep. I think people will soon realize that, like, basically, you know, the only way to defeat a drone is another drone, and then there's going to be this drone war where, you know, one side is going to deplete all of their drones, and then the other side is probably just going to surrender.

There's no point in fighting a, like, man-on-drone war. You'll lose. And when people know they lose, like, they generally don't fight.

That's one thing that I saw earlier today. You collect some footage when the operators, like, send it back, and I saw something that I just, like, I don't know, never even occurred to me is, like, a soldier getting targeted by a drone, turning around, and then trying to shoot at it before it blew him up.

And you said that, like, 40% of drones get taken out this way or.

Something like that, yeah.

Something like that. How has kind of, like, warfare evolved and the way that, like, people? Because I imagine, like, being that guy, I'm like, "What the fuck am I doing out in this field, you know, getting targeted by a drone?"

I think this is a good lesson in, like, contact with reality. Like, I'm sure if, like, before Ukraine, you sat down, like, a weapons expert, they would tell you about, like, "Oh, we're going to do laser defense. The lasers are going to work.

You're going to use, you know, directed energy weapons to take out the drones. You're going to hack the drones or something. You're going to use AI to take out the drones somehow." I'm sure there would have been, like, a really, really complex, like, plan of what defeats the drone.

When you look at, you know, reality, like, I saw this shotgun that came out branded, like, Drone Buster AI or something like that. AI stands for Advanced Impact. It's just a regular shotgun with anti-drone rounds because that is the most effective weapon system in taking out FPVs close range.

And in war, you have to kind of make do with what you have, and kind of the non-optimal solution is often what you have to work with. And so now Ukrainians have mastered the art of, like, anti-drone schools.

Like, these, you know, infantry members, they'll go to these schools, and they'll learn how to shoot down a drone with, you know, an assault rifle that they have or their main rifle that they have. And, yeah, you got to make do with what you have.

I watch a lot of, like, YouTube about how these weapon systems are used, and it's really heavy stuff,right? Like, behind these videos, there's lives that are lost. It's really impactful, like, heavy, heavy content. But you learn a lot about the tactics.

And obviously, I have, you know, tremendous respect for the people that actually fight. They learn so much more. But you can learn a lot via just, like, watching YouTube. And these, you know, highly trained Ukrainian operators, they're insane.

They'll hear an FPV, dang, dang, dang, dang, five bullets later, the drone's taken down.

Is there anything that you've taken or learned because you watched that footage that you just would never have figured out if you hadn't?

100%. I mean, like, half the features of our product are based on those videos. I think, like, watching users, like, put their rucksack down and get in position quickly to deploy the drone, that was one of the reasons we came up with, like, it sounds simple, but color coding the antenna mounts.

Like, it's very important you get the antenna on in theright place. You don't want to be like, "Ah, which one goes to where?" Reading some code, like logos to, "Let's screw that. Orange to orange. White to white." I think that was made very clear via, like, one of these videos.

Things like how the ground station displacement kit works, like being able to just super quickly plug in these connectors and making sure that they're keyed to work very carefully comes from video inspirations. And then also watching how, like, the, you know, usually it's not just, like, one dude with the drone.

It's a team. So, like, two, three guys. Watching how the navigator talks over the shoulder of the pilot. We set up this, like, secondary screen. There's a port on our ground station so that you can just do USB video.

You can plug in your iPhone and use that or your ATAC device. Like, details like that have been absolutely uncovered via these YouTube videos.

What other things are you guys, like, making tweaks on, making iterations on to just improve the operator experience?

We recently redid the ground station software interface. There's, like, six different radios in the system. You got to configure justright. There's different modes for these radios. We try and automate it as much as possible, but there's a lot of systems that go into.

People are manually configuring six different radios every single time they launch?

There are auto modes, but, yeah, like, if you're an expert operator, you got to mess with these different radios, and we work very hard to.

Is it, like, mission-specific?

Yeah. They'll find out there's jamming in certain areas and then pick a certain mode. There are manual modes to things that you want to use sometimes. You might want to change your car to manually shift gears as an example.

Or really, you should have a stick by default, but anyway, just how those, like, menus work out, they're so important. Like, someone's life is in the balance on whether this menu is a good user experience or not. So we care so much about every page.

And that's just, like, one tiny example how the drone interfaces with various other systems, like what battlefield management softwares we interface with make all the difference in the user experience, and we care a lot about it.

Is there anything that you can see is a product or project that you're definitely going to want to tackle in the future but are not going to tackle today because you just don't have the bandwidth or the necessarily, like, technical know-how yet?

Yes. Consumer photography drones.

Okay.

Right now, most of American weapon systems, like, obviously, it's going to sound obvious, but they're funded by taxpayers. Thank you, Tai, for your support of Bandit and Archer. You know, it is your dollars.

I actually don't make money, so I don't pay taxes.

Okay, fair enough. Well, thank you, my other friends, for paying taxes. That sounds very obvious. Like, weapon systems are funded by the government. China has kind of, like, transcended this to a certain degree. I, as a DJI customer, am funding China's military industrial base.

You know, they make in China cargo ships. They're required to meet the PLAN, People's Liberation Army Navy standards for military vessels. So when I buy that thing from China, I'm actually funding their navy. So it's crazy that they have, obviously not their entire military budget, but a good chunk of their military budget is actually funded by, like, American and other Western taxpayers.

That is smart. We should start doing that. We've fallen so far behind that it's almost impossible. Competing with DJI, the world's largest drone producer, is nearly impossible. They own 70-plus percent of the world consumer drone market, and I don't know of another consumer company with a bigger advantage on all the competition than DJI.

Like, you look at an iPhone and, I don't know, a Samsung, and, like, you could see the difference in quality of product is, like, this much. To DJI, it's DJI is the next competitor. It's insane. They're just so far ahead.

And kind of building an American competitor to that takes time. We're not ready for that yet. I think it's going to take a million-drone factory first to get to the unit economics where that makes sense. But, yeah, my answer to that is consumer drone products.

That will be a challenge one day. It's a great way to have credible deterrence. China will start buying them. They'll start putting cameras pointing at their critical infrastructure, and that'll be a really interesting data source. Maybe I'm dreaming there, but having a consumer product is critical to maintaining credible deterrence.

One of the other things I heard you talk about today was the different suppliers that you guys are going to work with. And I was kind of thinking to myself, there's enough of them where clearly you have, like, a strategy versus just fully vertically integrating every single thing that you're going to do.

Why not vertically integrate everything?

You can never be 100% vertical,right? We're not designing our own screws. We should probably use the same screws that SpaceX uses and everyone else. So there's a big strategy. The make versus buy strategy is very important. The way we think about vertical integration, it's first about, like, well, if nothing exists, like, when we started, we wanted a drone radio for 50 bucks.

We reached out to the existing players, and we heard, like, $2,000, $3,000, $5,000. So, okay, go vertical. Easy answer. Next, it's about, like, does this have a massive impact on the quality of your product? For example, like, if we were using someone else's flight computer and their flight computer was a little worse, it had a little bug, like, all of a sudden our drone flies like crap, and our user's life is at risk, we should make that ourself.

We want to control every variation of that thing. We want to make sure every defect is, like, directly reported and not hidden because contract, you know.

And also it's, like, on you if, you know, if someone fucks up.

Yeah, if someone fucks up, it's on our team, and they get, you know, they get to see what happened to the production line, and, you know, they don't get to say, "Oh, well, we'll just work with other people."

Like, it's their job to be full-time. So when it's important to have control, that's important. It's also, like, there's a financial aspect behind it. Like, let's say the radio example. If there's one very low-cost piece of a radio and that has a really long lead time, that needs to be stockpiled, it's much more efficient to just stockpile that one, two, or three small parts of the radio instead of stockpiling a ton of fully assembled radios.

When there's some sort of, like, supply chain advantage you can do there, that's another great reason to vertically integrate. So, like, on our radio, there's a few chips that are long lead that we care a lot about making sure we have a lot of.

So we'll stockpile those, you know, couple-dollar chips in massive, massive volumes. And then the other, like, super common, you know, parts, we won't stockpile in such large volumes. If we were buying off-the-shelf radios, we'd have to stockpile entire radio assemblies.

It's really expensive. There's a bunch of reasons to vertically integrate. If you don't tick any of these boxes or others, that's when you get to, like, just don't worry about it and focus on other stuff.

When you think about, like, the first version of the Archer or the first version of your drone versus the one today, and you also think about, like, the amount of time, I think I saw this statistic that most of the, like, drone competitors that are selling a similar product take, like, 12 hours to make one unit, and then you're taking something like 45 minutes.

And I imagine that a huge part of that is literally just wringing out a bunch of, like, just deleting parts, you know, just making it a simpler thing.

Reality forces the Elon algorithm. It's crazy.

How much, like, more efficient have you been able to make the production of drones through just constant iteration over the course of the last three years?

Oh, man. We spent, like, the better part of a month or maybe more in Soren's parents' basement building the first 30. And those were, like, those were long days. Not much.

So a day per drone.

Yeah, minimum. And there were a bunch of people there. So, like, many man days per drone. And I thought that design, that first design was, like, I thought at the time that was a really design for manufacturing design as a, you know, engineer that thought they knew what they were doing.

But nothing teaches you how to make a design manufacturable except having a production line that you can test it on. So, you know, production techs here go up to engineers like, "Why is this painful?" Engineers come to the production line and they watch.

Oh, my gosh, a technician has to route this cable underneath because it's more comfortable? That's ridiculous. We should have designed it differently. There's just this, like, constant back and forth that needs to happen domestically. Like, we need to have the line hereright next to the engineers to really keep that loop tight.

And I think that is what gives us the kind of hope to get to a million drones per year. Like, we have learned so much by producing 37,000 Archers, and every lesson is captured and is going to go towards that million-drone design.

There's different levels of scale, and so certain techniques for building the drones at, you know, 5,000 drones a year won't scale to, like, 50,000 drones a year, won't scale to 500,000 drones a year. What things have you guys had to stop doing or, you know, massively improve efficiency on in order to actually scale production?

So when we were making, like, hundreds of drones a month, we got to just not use connectors and direct solder. You just take a wire and a board and melt the two together, essentially. That was a huge, like, it just makes everything simpler and tighter.

You now have to train users very carefully, and they have to, you know, get lucky to make sure that solder joint is up to quality, etc., etc. Obviously, that doesn't scale. So we were like, "Okay, must not have any manual solder joints."

That was a big step forward. Most FPVs, even today in Ukraine, have manual solder joints. So we avoided that. When you get up to, you know, I'd say high hundreds to thousands of units a month, that's where you have to get really worried about things like PokeYoke.

You have to make sure that if you can, like, there's no way to install a part wrong. You know, you'll notice the two camera slots don't have the same interface because you could flip them and then things would be messed up.

Little details like that are really important. Like, the two sides of a connector, they either have to be reversible or they have to use a different connector on each side. There's all these, like, little extra gotchas. When you get up to, like, the thousands of drones per month, like, you know, 5, 10,000, this is where you have to be really smart about your supply chain.

Like, we could buy every chip in America of a certain part number that goes in Archer and then not have enough to sustain a month or two of production because it's just, like, all the distributors combined only have, you know, 5,000, 10,000 chips.

So now you have to design around the manufacturer, like, you know, TSMC's lead time in Taiwan. Or better, you have to design two different boards. You have to design board A, do all the work to design board A, and you have to start again with board B.

And it has all the same requirements as board A. It must behave exactly the same, but you have the simple gotcha of you can't use any parts from board A. You have to figure it all out again with different engineering strategies, a different microprocessor, that kind of thing.

That's the kind of, like, hidden gotchas. And the product seems the same. Hey, my, you know, my iPhone seems the same as every other iPhone, but there's actually a bunch of different variants of iPhones that use different memory chips and that kind of thing.

That way, you can have stable supply chain.

They'll have almost exactly the same performance and stuff.

Yep. Yeah, and they look identical. Like, you don't even see the work that that, you know, team, entire team of engineers put through to redo all the work of design A, have it function exactly the same and not use any of the same parts.

And these are, like, just a few of the kind of, like, really hidden gotchas that it takes to get to a million a year of something. And I think that's why, like, you know, no other defense company in America makes a million of anything that isn't ammunition.

Like, we make a million a year of bullets. Name another product that is made in a million units per year.

Starlink terminals, I think.

Starlink terminals, totally.

I think it's like 5 million a year or something.

Yeah, I mean, Starlink is breaking records there. I wouldn't, well, they don't make a million Starshields per year. I wouldn't consider that a defense item. The Starlink terminal production is insane. It's a consumer product, but it's probably the only really full-fledged example of China beating consumer electronics production, like, consumer electronics production that would beat China here in America.

That team is incredible. We're very lucky to have engineers that have worked on that production line. The Starlink user terminal line is an incredible, incredible feat for America.

For the people that come from that line and come work at Neros, how do they think differently than everyone else from everywhere else?

I've been told, like, Elon always said, like, we're going to make this line run so fast that we're going to have to consider the aerodynamic drag on units as they whiz through the production line. That's how fast this thing is going to run.

Like, you know, there'll be in design reviews and, you know, we'll talk about, you know, manually aligning a connector to plug in. It's like, no. Like, it's completely unserious. This needs to have, you know, alignment tools or just not exist at all.

The kind of bar for, like, when a design is done and ready is just so high. And, like, also the understanding of, like, getting the thing to work once is one-tenth of the work. Like, it is so simple to, you know, build the drone in the basement and get it to fly.

The challenge is in the iterations to get it manufacturable. Those learnings are really, they come with time. And we're really lucky to have great people that have had that time.

One of the things I think I asked you last time was this idea of, you know kind of where the puck is going, and you want to be able to ship the best thing that you can ship at any given time, but obviously, you can't, like, immediately make an improvement and then scale that across, you know, your units.

Is it improving with scale? Like, are you getting better at shortening the time between making an improvement and having it in users' hands, or is that getting longer? Like, how does that change?

In some ways, it's getting shorter. In some ways, it's getting longer. I'd say we're engineering faster at this point. Like, one of the things I really love about engineering at Neros is that it is a bunch of, like, really small teams that are all behaving like an early startup.

You know, we're only three years old, so we have a lot of that DNA in us. But also, we're, like, perpetually terrified of big company bullshit. So we keep teams that are super, super tight that, you know, iterate quickly and are able to ship quickly.

With that being said, big production lines, they don't transition overnight. It is a big chip to steer when you're making these kinds of quantities. Starlink doesn't come out with 10,000 variants, and it builds one thing, and it builds one thing well.

So I'd say we are faster at iterating on new designs and kind of making, adding features when you're in engineering land.

Like prototyping.

Exactly. Faster prototyping, but slower mass production changes. The line is fixed.

In your mind, when you're kind of doing these two different modes of, like, mass production or prototyping, which one's more fun?

Well, I'm a massochist, so mass production is more fun for sure. It's just the weight of the problem is much bigger. Like, you see, you know, a station go down on the line, and then literally, like, a drone comes out of the building every two minutes.

So you just, like, drones just start stacking and stacking. You're like, we got to fix it. We got to fix it. And the thing with production is that we pay for materials to come into the factory, obviously, and we only get paid when materials go out of the factory.

But generally, the materials coming in don't stop. So you just start piling up, and you just see, like, the inventory piling, piling, piling and getting perpetually closer and closer to bankruptcy as the company's balance sheet just gets converted into parts.

It's a stressful, it's a crazy stressful activity, but it's so fun. Like, the magnitude of the challenge is so real. And you have to be so worried about, like, every little part of the design that can get you.

You know, there's so much, ah, yeah, it'll fly. Like, ah, we'll take the risk in prototyping, but in production, you got to be a lot more careful. It's like a game of 3D chess. You got to, you know, anything that can bite you will bite you.

Of course, they both have their fun. You know, some engineers much prefer this stage of development where you have more freedom to mess around. And we have both. But if I had to pick one, it is production for sure.

I know that Elon has this idea of, like, walk to the red when he was going through production hell in 2018 for the Model 3 production ramp. I think there'd effectively be screens all over the factory of whatever was, like, slow on the line.

Yeah.

And he would just go to one of the red screens and then just unfuck the bottleneck and then go to the next red screen and unfuck that bottleneck and just continue doing that until the line worked. When you, like, go on the line and you see some bottleneck, how often are you, like, figuring out there's some part or process that you don't even need to be doing?

I mean, all the time. Like, you'd be really surprised how simple the solutions are to some of these problems,right? Like, especially as the company grows, like, there's a lot of juice to squeeze in simple solutions. You know, you got to try the idiot checks.

Do we even need this part? Can we delete this? Like, you know, who qualified this part? Show me the test data. Did we make sure this part works? But walking to the red, I'm a big fan of it.

Like, Soren and I, we both try very, very, very hard to put incredible people in charge of basically every responsibility we have by default. And then we spend a lot of our time just unfucking. Like, in theory, if the company's running well, our calendar should be free, and we're able to just pick the most fucked area and dig there.

That's, of course, not always possible. There's some things you want, you know, high touch on, but that's generally our philosophy.

One of the biggest challenges with scaling this type of business is you have a, like, almost single customer, and that customer has to scale with you or else, again, just like you'd have inventory piling up or, like, parts and stuff piling up, you can't scale without people buying the drones.

So how does that relationship with the US government and other governments work? And are they kind of moving ahead of where you can scale, or are they not?

It's such an important and complex relationship. There's a lot of really, really impressive, impactful work being done in that domain. If we want to build this factory that credibly deters conflict, it needs to be profitable. It can't be like a charity item.

You know, our drones are $2,000 each. A million drones, it's $2 billion worth of drones. And that's just the drones. Like, what about the production, the engineers? There's a lot of stuff that goes into that. That's not like something that happens without a profitable enterprise around it.

And the government really wants this too. Like, our incentives are super aligned,right? Like, we exist to serve the American warfighter and to provide capital deterrence. And most of the people that are in the government, they have dedicated their lives to this,right?

Like, they could get way better pay outside the government. Often, they served in the armed forces. They care so much, deeply, deeply about making sure that America has credible deterrence. So the, like, incentives are often aligned. It's just the government is a big organization.

You know, it's a really, really big organization. Largest employer.

It's like the ultimate aircraft carrier. It takes fucking, you know, 200 miles to turn.

Forever to turn. And there's a lot of details that go into, like, writing a company a $2 billion contract to build a million drones,right? It's not like something that happens overnight. So that relationship, it's often it's a lot of strategizing together.

It's a lot of, like, you know, working with the customer. It's explaining the customer. Like I mentioned in the beginning, it's explaining to the customer what the warfighter needs. Explaining to the government what the warfighter needs. It's also explaining to the government how to contract, you know, a million drones and what they need to do.

And these, like, very, very well-meaning folks in the government, we sometimes have to explain to them, like, what we need. And they also explain to us, like, what their blockers are, what makes it hard to actually, like, buy these things or field these things.

Like, you know, they tell us, as an example, listen, Marines want to fly this drone, but they can't without crazy sign-offs, very, very, very senior, you know, exception to policies because of spectrum regulation. So we go, great, the answer is a JF-12.

That is like approval to use the drone on any military base. And then we work, we figure out how does the JF-12 process work? Who do we need to talk to to move through that quickly? What engineering resources can we give?

We start doing a bunch of testing ourselves to make it so at the end of the day, the Marine can fly the drone and they don't have to stress about it. But it's very much a symbiotic relationship, and it's a partnership that goes very deeply.

We really appreciate government partners that

understand the importance of explaining the what, but not the how. I think government is very good at, like, they told us, you know, the Marine can't fly the drone or we need to strike a target, you know, far away, XX kilometers far away.

Those kinds of, like, high-level capabilities are something well described by the government, and then we work on how that's done. Like, we own the implementation, and we work best with customers that understand that really well.

Get to fast forward a couple of years, and you guys do have this thing fully operational and it's making a million drones a year. What does that business look like structurally versus the one that you have today?

To build a million drones a year, you basically have to build a city. You have to build an entire ecosystem of components, and each little detail in that drone has a massive impact. You know, let's say you choose to use a chip in the drone, and that chip is just $10.

All of a sudden, you have a $10 million PO for just one year that you're waving around. You probably want to do a few years of this. So there's a $30 million contract to be won for one chip in this drone.

And there are hundreds of electronic components, probably thousands if you include all the SMT components in this system. You have to build a city of suppliers. I mean, a lot of it should be vertical, but eventually, you know, someone's got to sell you the raw materials probably, unless you're going to open a mine.

So you're building a massive ecosystem. And then internally, the kind of level of intention that has to be put into every part of the design at a million units per year, you assign, like, everything becomes a challenge. Like, just this, you know, this cable becomes a massive challenge.

Like, how do we design this cable to be, you know, really compatible with a million unit per year run rate? How do we make sure that the motors are compatible? Each one of these almost becomes like its own little company with its own, you know, responsible individuals and folks that control the destiny of that part.

So you build a massive ecosystem with, like, areas, more specific areas of extreme ownership. I think that's the best way to think about how to structure a million drone per year company.

Are there any lessons that you can take from a company like BYD in China where this is the only, this is the best example that I can think of of an entire city factory?

Yeah. I mean, these are ICE tenant. SkyFactory.

How big is their factory?

It's this massive skyscraper and a bunch of buildings around it. There's an insane tour that I've probably watched 20 times on YouTube of a drone flying through their campus. It is mind-boggling. But that's what it takes to, you know, build a million units per year, and it took them over a decade.

When you think of, like, Chinese production lines and factories versus the thing that you guys haveright here and in general, just United States factories, what are the biggest differences? Like, what enables them to be that much more efficient?

So, you know, one of the first things I did after getting back from Ukraine was going to China.

You went to China?

I went to China.

Was this before getting blacklisted?

This was before getting blacklisted, before getting sanctioned. But yeah, I mean, I wanted to build drones. I was a professional drone racing pilot before this, and I wanted to see where the drones were made. I had friends in China, I mean, sponsors that would send me free stuff.

That's kind of the extent of the professionalism, but that would send me free parts. And I knew them, and I knew that they got into building drones for war. They didn't know that I was going to get into this, but I knew that.

And I DM'd them. I was like, "Hey, I'm going to be in Shenzhen. Do you think I could see your factory?" And they were like, "Sure." They knew, you know, Soren had a consumer business selling Chinese drone parts.

I also sold Chinese drone parts. And they thought I was a customer. So they showed me around in detail. They told me where every step of the process had quality checks and what the tolerances were. They were so happy to brag about their specific tolerances on their quality checks and the names of the high-end machines they used to build these various parts and the different, you know, steps in the process for ensuring the best quality drone.

And, you know, they were showing me drones that were destined to Russia. They showed me these production lines making 2,000 Russian kamikaze drones a day. And honestly, insanely impactful trip,right? Like, that's the end state. Or not the end state, but that's where we're going.

One of the biggest takeaways there was that their manufacturing is pretty simple. It's a lot of people at desks doing pretty simple labor. And I think it's not exactly how we're going to run our labor. I think we're going to care a lot more about the workers.

Like, you know, I'm sure you've seen the images of the Foxconn suicide nets and the, like, you know, dorm rooms for only men or only women seven days a week. Dorm room factory, dorm room factory. The production, you know, obviously the number of people and the scale was impressive, but I think we can do much better here.

I think we have, you know, much smarter automation. And I'm sure that's changing in China. Like, it's moving away from manual labor and towards automation, but I think we can beat it by far. What I want to copy from China and what I was most impressed with is actually not the production side.

I don't think China is that advanced. They're advanced in the design side. You could literally go to this place, Huashang Bay. It's like a bazaar, like a street market for electronics. And you could buy chips. Like, a dude will sell you a watermelon and a chip and an FPV drone jammer, but we'll set that to the side.

And you could prototype with that, and then you could buy a new chip same day. And you can go to a different guy down the street, and he will, in his, like, sketchy, like, you know, 1,000 square foot building, build you a printed circuit board with that chip.

And you can go to another guy that will build you a CNC enclosure same day. And you have this, like, massive ecosystem that's all down the street from itself that allows you to iterate on products. And this is how they come out with an iPhone every year.

This is how they kind of make super rapid progress on designs so that they're manufacturable. And I think that's the secret sauce. It is this industrial base, this city, this ecosystem that allows engineers to iterate at ridiculous speed.

Like, you know, we spend six weeks waiting for boards generally for our radios. In China, that is one day. You get one day board spins. It is wild.

How are they that much faster?

Well, it's just volume. Zhuo Lisheng JLC, it's a large printed circuit board vendor. They rib through boards. They make so, so, so many boards. And it's a lights out. They're winning in automation. It's a lights out automated production line.

I tried very hard to go. They would not show me that part. But the engineering infrastructure, being an engineer in China, it must, you know, you have a lot of luxuries. You know, maybe you have to deal with Big Brother, but you do have a lot of luxuries in how fast your design works.

I think that is what is needed to build the million-drone ecosystem. You need this city of industrial services almost, where you can go down the street and get your board made same day, get your camera made same day, get your lens for your camera cut same day, get your injected molded part done.

That ecosystem is so powerful. And I think that's what the kind of million-drone company looks like to a certain degree. It's an ecosystem.

I don't think it's fair to say that the US has, like, credible deterrence against a war against China trying to take Taiwanright now. And so if you had to fast forward a few years and say we actually are in a state where the US can credibly deter China trying to do that, what does that look like?

Like, what does that require?

I really like to think we do currently. And here's why.

World leaders generally are pretty good at, you know, games of chess, so to speak,right? They don't get to be president or, you know, chair or whatever they want to call it by not being, you know, sly.

When an opportunity opens, world leaders take it, like, 99.9% of the time. If there is a strategic opportunity for a country to win, to get more territory, to get more influence, they're going to take it. I mean, maybe there are some exceptions of extreme incompetence, but it is very rare.

If America loses its credible deterrence, if America cannot defend Taiwan, 30 seconds, China is going to do it. That intelligence memo hits the desk of Xi Jinping, he's going to do it. So I think we still have credible deterrence.

It's just the rate at which it is going away is terrifying. And I think she knows that as well,right? Like, she has spies in Ukraine that are a lot smarter than me and that go to Ukrainian drone vendors or drone users and hear the same thing.

They hear, you know, quotes like, "Silly Americans, they think we speak English here in Ukraine. We speak Chinese. We like the Chinese drones, not the American drones, because they actually work." Like, they hear these same things that I hear.

And those quotes are going up to the highest levels, I'm sure. And basically, there's just this, like, ticking time bomb. So I think at the moment we do have critical deterrence. It is just, or credible deterrence. It is just going away very, very quickly.

And I think what reversing course looks like is a return to building. It's a return to, you know, El Segundo being the aerospace capital of the world, not of America. Becauseright now it is just the aerospace capital of America, in my opinion.

The real capital is probably Shenzhen, China, especially for electronics. But it is a return to being a country of builders, a country of people that make physical things. And it's also a return of a country of patriots. It is so easy to kind of

listen to the group think almost and just ignore weapons. Say, "I'm too good to build weapons. You know, it's just not my problem. Someone else can take care of it." I think that that view, like I said, comes from just a misunderstanding of the world.

And it's an easy mistake to make. But I think the return to a time where we, you know, really put a lot of effort into being the world's police and being the people that maintain the most peaceful era in human history, which has been the past 100 years, is going to look like a return to folks building systems like Archer.