Behind what looks like a homogenous cohort, deep strategic differences are starting to show. The sector is maturing, and distinct playbooks are emerging, as some companies triple down on their anchor country while others export outside Europe as fast as they can, and as M&A ramps up, though not for the same capabilities: some are buying vertical integration, others suppliers, others still entire new domains to expand into.

The German bet

Germany is building the largest army in Europe, which makes its procurement the closest thing on the continent to a captive market for a German supplier, and, more to the point, the fastest route there is from a prototype to a programme of record.

Stark is the clearest case. It runs teams in Germany and the UK and opened a plant in Swindon last November, but Germany has been by far its biggest win, with the much-publicised €500M strike drone contracts it shares with Helsing. A company founded in 2024 reached a programme of that size in under two years, on a timeline that is not available anywhere else in Europe at the moment.

Helsing seems to be refocusing on Germany, after having built a pan-European footprint from day one with stellar teams recruited in the UK and France. The bulk of its revenue is already German, and that share should keep growing as it wins large contracts on collaborative combat aircraft (CCA) programmes. The latest is a €500M award on the Combat Fighter System Nucleus, Germany’s national combat cloud, which ties crewed fighters, autonomous drones, satellites and frontline sensors into a single AI-driven network.

On top of its own domestic market, Germany is extensively financing exports of German equipment to Ukraine. That gives German companies rapid revenue growth, battlefield validation, experience operating in GNSS-denied environments, and so on — but it also creates a dependency on a single procurement channel. The weakness of the bet is that it rests on one customer’s appetite, and appetite moves: this year Berlin approved the first tranche of the strike drone programme while cutting the multi-year framework behind it by more than half.

The pan-European play

The alternative is to treat several national budgets as a single market from the beginning, which is what you do when your home budget cannot fund a massive ramp on its own, as is the case for most budget-constrained European economies.

Cambridge Aerospace is the cleanest example. It has been delivering its Skyhammer interceptor to the British Ministry of Defence since May, and it has recently been building commercial teams in Poland, Germany and Norway, a footprint that partly mirrors the first framework contract under LEAP, the five-nation effort on low-cost effectors that spans the UK, Poland, France, Italy and Germany. The company is betting on a rapid pan-European expansion, and has set up a team in Australia as well.

This is a slower route than riding Germany’s urgent rearmament, but it builds sturdier foundations, with several countries and several customers to spread the risk across.

The ones who exported early

The third playbook treats Europe as one market among several, and only one European company is running it convincingly. Quantum Systems sells to the US Army, to Australia, to India and Romania, and it manufactures inside Ukraine. I’m guessing this strategy is an inheritance, since the company comes out of the 2015 hardware wave and spent its first decade selling drones to whoever would buy them, long before European budgets opened, and what began as a necessity now leaves it the least exposed of the German companies to any single ministry and the only one with a US programme of record in progress.

Harmattan has learned the same lesson from the other side. It signed programs of record in France and then in the UK at record speed. But winning fast at home does not make home the biggest market: both France and the UK have to fund an expensive nuclear arsenal, which leaves less net new cash for modern capabilities than in Germany. So Harmattan moved into the Middle East and the United States early, rather than discovering the ceiling later.

Whether European companies should move into the US market quickly is still an open question. In theory, the combination of a US market that favours local players and a Europe that is spending heavily on defence should push European champions to triple down at home. Yet the US remains the best defence market when it comes to the speed and size of contracts awarded. Hence Helsing’s announcement in July of a strike drone factory in West Virginia, its first manufacturing site outside Europe; Quantum Systems already selling there; and Harmattan poaching a US CEO from Vast to lead its expansion.

Buying up the stack

The other main difference between these players is their M&A strategy. Nine acquisitions have been made between them in eighteen months, and they fall into two groups: vertical integration and domain expansion.

Quantum Systems is the clearest case of the first, having done five deals in eighteen months, more than anyone else in European defence: AirRobot, the AI firm Spleenlab, the autonomous trucking company Fernride, the electric motor maker Hacker Motor and the Estonian intelligence software company SensusQ. Read together they describe a hardware company buying its way up the autonomy stack, one layer at a time. Spleenlab brought perception and GPS-denied navigation and tripled the software organisation on its own; Fernride brought an autonomy stack for the land domain; SensusQ brought the layer that turns sensor output into something a staff officer can act on. Only Hacker Motor, electric motor manufacturer is a supply chain deal, Stark has bought the same way on a smaller scale, taking Pleno for autonomous flight and swarming and a European subcontracting network to feed Swindon.

Buying a portfolio

Helsing buys differently, and is expanding across CCA and the maritime domain through acquisitions. Grob Aircraft gave it an airframe manufacturer and the certification history to go with it, which is what its autonomous combat aircraft programme needed, and the acquisitions around that programme follow the same pattern. Blue Ocean, the British-Australian maker of autonomous underwater vehicles, gave it an entire domain: hulls, a production line and a customer base in the water, where building the equivalent from scratch costs three or four years before a credible demonstration.

This is modelled on the Anduril playbook, where several of the flagship products came in through acquisition: the autonomous fighter Fury from Blue Force Technologies, the large autonomous submarine Ghost Shark from Dive Technologies, and the loitering munition ALTIUS from Area-I. The logic is that a company with strong customer relationships can buy a product and sell it far better than its previous owner could, which makes acquisitions a way of filling gaps in the portfolio rather than back-filling the factory.

What Helsing does not buy is the supply chain, where it uses joint ventures and volume agreements instead, and it does not need to buy revenue, because distribution is the thing it already has. It buys capabilities it can scale through relationships it already owns.

The next deal

For all that activity, it seems that not one of these companies has bought a business in order to enter a national market. Every deal of the past eighteen months was about capability, components or domain.

That is the gap I expect to close this year, because the incentives now all point the same way. Both the French and German markets are notoriously closed to each other, so you can expect a German company buying a French one to access the market, or vice-versa.

Funding and deals

IPOs:

• Lyntris, the Virginia defence sensing company assembled from around a dozen acquisitions in antennas, radar, sensor control and C2 software since 2018, went public on the NYSE at a c. $2B valuation.

• Ursa Major, an aerospace and defense company building hypersonics, solid rocket motors and space mobility systems announced an agreement to go public via a $2.3 billion SPAC merger.

Acquisitions:

• Austal USA, the American subsidiary of the Australian shipbuilder, is the target of a $1.2B offer from Hanwha Defense USA, part of the Korean group’s stated exploration of ways to expand its American footprint.

• BlackSea, the Baltimore maker of unmanned surface vessels, was acquired by Aevex for $650M, taking the drone maker from aerial into maritime autonomy; BlackSea expects around $150M of revenue this fiscal year.

• Resonant Sciences, which builds RF and sensor systems for classified programmes, was acquired by Joby for $500M and becomes the eVTOL company’s dedicated defence business.

• Wisk, SkyGrid and Insitu, Boeing’s three autonomous flight subsidiaries, were acquired by Archer Aviation for an undisclosed sum, giving Boeing a 20 percent stake in Archer and giving Archer what it calls an end-to-end physical AI platform across commercial aerospace, defence and air traffic management.

• L3Harris’ space propulsion, power and electronics units were acquired by AE Industrial for an undisclosed sum and relaunched as Rocketdyne.

• Aran Defense, the defence division of the Israeli engineering group Aran, was acquired by Ondas for an undisclosed sum, adding local manufacturing and integration capacity in Israel for its autonomous platforms.

Fundraises:

• Hadrian, building AI-enhanced factories for aerospace and defence parts, raised a $1.37B Series D at a $7.87B valuation, led by WCM Investment Management, Washington Harbour Partners, Valor Equity Partners, 137 Ventures and Baillie Gifford.

• Castelion, whose Blackbeard hypersonic missile is meant to cost hundreds of thousands of dollars against the $41M of Lockheed’s Dark Eagle, raised a $1B Series C co-led by JPMorganChase’s Strategic Investment Group, Andreessen Horowitz and funds managed by Carlyle.

• Astranis, which builds low-cost satellites for high orbits, raised $450M led by Andreessen Horowitz, taking total funding past $1.2B.

• Cambridge Aerospace, building interceptors cheap and fast enough to match a Shahed or a Geran, raised $300M at a $3.4B valuation.

• Smack Technologies, building AI for national security that runs both in operations centres and on edge devices, raised a $61M Series B led by Costanoa Ventures and First In.

• Worldscape, an AI-native decision platform for mission engineering, raised a $10M seed extension led by Scout Ventures.

Deep strike used to be the business of a very small club - at roughly $2 million apiece, bought by the dozen and treated like platforms: precious, slowly replaced, rarely expended. That weapon is now becoming ammunition, bought by the thousand, judged on cost per destroyed target, and fully expected to run out. These deep strike capabilities can be both drone or missile-based - the line between the two being blurred by the fact that these drones are now powered by jet engines and mimic cruise missiles.

The $55,000 deep strike

Ukraine flies this campaign without air superiority, so everything of range leaves from the ground — the same condition any European war against Russia would impose, and one that makes the truck rather than the bomber the launch point that matters. The ground brings its own economics: a launcher hidden in an ordinary container costs a rounding error of an aircraft and survives by dispersing rather than evading.

Fire Point, the Kyiv company behind much of the campaign, sells the range as a product line. Its FP-1, a one-way strike drone estimated at $55,000, flies out to 1,600 kilometers and was coming off the line at more than a hundred a day by the middle of last year; it is the volume behind the daily refinery footage. Its Flamingo cruise missile carries a warhead of over a tonne toward a claimed 3,000 kilometers — the class of weapon needed against hardened plants rather than distillation columns — and strikes attributed to it have reached missile factories 1,300 kilometers inside Russia. The two layers work together, drones saturating and mapping the defenses that missiles then penetrate. (Ukrainska Pravda’s tally of known Flamingo strikes suggests a success rate closer to a third than to the brochure — tolerable arithmetic at a reported unit price under a million dollars). It is believed that their first ballistic missile strikes could reach Moscow in the next few months.

Destinus, the European entrant, climbed the same ladder from the other end. Its Ruta missile entered combat in Ukraine at 300 kilometers; a 700-kilometer version followed within a year, and the Netherlands ordered seven hundred of them in June; a 2,000-kilometer variant is now in development with Rheinmetall, powered by turbojets Destinus builds itself by the thousand. An order for seven hundred missiles, placed in months rather than over a decade, is not procurement as the last generation practiced it. It is an ammunition purchase.

850 Tomahawks in two weeks

In the opening two weeks of the Iran war, the US Navy fired more than 850 Tomahawks — over 400 of them in the first 71 hours — enough to alarm Pentagon officials about what remained in the magazine. Each of those missiles takes on the order of three years to replace; at minimum, Japan’s pending Tomahawk order slipped as a consequence.

The arithmetic generalizes. At the tempo of an actual war, deep-strike munitions are consumed like artillery shells, and an arsenal replaced over years cannot sustain a campaign measured in weeks. What decides how long a country can keep striking is not the quality of its best missile but the output of its production lines — a manufacturing problem before it is a military one, a topic I covered here.

The US also gave a final push to turn European interest into urgency. In May, the Trump administration cancelled the planned deployment of the US Army’s Typhon launchers to Germany — the system that was to put American Tomahawks on German soil from 2026 — as part of a withdrawal of some 5,000 troops. Berlin has since moved to buy around 400 Tomahawks of its own — an order still unconfirmed, and one that would in any case join the same years-long production queue that is still replacing the missiles fired at Iran. Overall, this forced Europe to realize that the capability should be owned, not borrowed. Ukraine and Iran created the appetite; Washington removed the alternative.

From niche to boom

That’s why a market that was a niche is now suddenly expanding. Britain and Germany have been developing a joint deep-precision-strike weapon since the Trinity House agreement of October 2024, with a range beyond 2,000 kilometers and service entry planned for the 2030s, and in July the effort widened at the NATO summit, where twelve countries led by Britain committed more than $50 billion over ten years to the category.

The United States is running its own version, though mostly launched from the air rather than the ground. Its ERAM program went from concept to contract in months, with Zone 5 Technologies and CoAspire building thousands of cruise missiles for Ukraine in the $300,000 class. The Navy, for its part, this spring picked Castelion, a startup founded in 2022, to build its next hypersonic long-range strike missile. It’s interesting to note that CoAspire is a bootstrapped business - apparently you can make a dent in the world without raising hundreds of millions of dollars in private capital, crazy.

Where this can go

The obvious question is why, with demand like this, the field is not already crowded — and the answer is that the barriers sit far higher than in drones. To design and test your first drones, you need a basic workshop, 3D printer and a field, which is why there are hundreds of them. A missile maker has to cross a wall that Europe is poorly equipped to host: test ranges are few and booked out, which is why so much of the continent’s missile development is in practice being flight-tested in Ukraine. Guidance has to survive a thousand kilometers of jamming. Warheads and propellants draw on an energetics base in chronic shortage, and the rocket motors and small jet engines beneath them are the oligopolies I covered in Who is building the picks and shovels in defense. Clearing all of that takes years and serious capital, and the short list of companies that have done it — while the flagship programs still promise the 2030s — is precisely what makes them valuable.

For a new entrant, the opening sits at the edges of the envelope rather than in the middle of it. A subsonic cruise missile at a prime’s price is the incumbents’ home ground, and MBDA will defend it. On a battlefield where everything is seen, speed is the new stealth, and so, at the other end of the ledger, is mass. The room is at radical price, which is the trade Fire Point and Destinus have made, or at radical speed — Mach 3 and above, where nothing affordable exists and where Castelion’s $300,000 quasi-hypersonic just won a US Navy program. Anything between those two edges is a fight with the primes on their own terrain.

My own bet is that this gold rush produces its first billion-euro deal within two years, and that the buyer will be a prime purchasing the one thing it cannot build in time: cadence. Rheinmetall’s co-development of the long-range Ruta with Destinus looks like the opening move of that courtship, though Destinus lining up banks for a possible 2027 listing suggests the startups have done the same math. Whether this generation grows into primes or is absorbed by them will decide who owns European deep strike at the end of the decade. Either way, the missile has stopped being a jewel. It is ammunition now, and the scarce thing is the factory.

If you’re building in this category - feel free to reach out.

Funding and deals

Going Public:

• Space-Eyes, a defense tech R&D company backed by Eric Trump, is going public in a $638M SPAC deal with McKinley Acquisition Corp.

Acquisitions:

• Raft, a Virginia-based maker of AI, data-fusion and mission software for defense, was acquired by Leonardo DRS in an all-cash deal valued at $450M, extending Leonardo’s push into American defense software. The company also won a $99M IDIQ from the US Army for its data integration software platform.

• Atropos Group, a maker of autonomous airborne platforms, was acquired by Chaos Industries for an undisclosed sum — a deal six months in the making that extends Chaos’ line from radars and interceptors into autonomous aircraft.

Fundraises:

• K2 Space, building large, high-power satellites for commercial and national-security payloads, raised a $500M Series D co-led by Kleiner Perkins and ICONIQ at a $6.8B valuation.

• Cathedral, a New York startup founded by former DOGE staffers to expand US military cyber operations with AI, raised a $160M round co-led by Andreessen Horowitz and Sequoia Capital at a $1.4B post-money valuation.

• Twenty, whose AI automates offensive cyber operations for the Pentagon, raised an additional $30M from Khosla Ventures at a $1.2B valuation, a month after closing its $100M Series B.

• Agon, a London startup building secure training-data infrastructure for defense AI, raised a $23M seed led by XYZ Venture Capital, Lux Capital and Sisyphus.

• Array Labs, building radar-satellite fleets that map the Earth and track aircraft and ships, raised a $21M round anchored by a strategic investment from Mitsubishi Electric.

In this episode, we get into what he saw coming: why European resilience is a generational shift and not another hype cycle. Where Europe has a right to win, What Helsing’s mega-round says about king-making in defense tech. Why he bet on Auterion and why he thinks The Exploration Company could become one of the most valuable companies in Europe.

Plus, the corners of the market that are still quiet: space tech, offensive cyber, and yes, bionic cockroaches.

00:00 — From the Internet Economy to European Defense

Louise Boucher: Alex, so nice to have you here. By any measure you’re one of the software GOATs — you famously led Shopify’s Series A, you backed Pipedrive, and for twenty years you were betting on the winners of the internet economy. Then, a bit more than a year ago, you published a famous piece on European defense saying you were going all in. What did you see that made the man betting on the internet economy decide the most important thing was to turn around and invest in national security businesses?

Alex Ferrara: Well, thank you for the kind words. One of the benefits of doing this for a couple of decades is that I’ve had a chance to have some reasonable successes — and many that haven’t gone as well along the way. A couple of years ago we started to focus more heavily on what we call European resilience. Truth be told, we’d been pretty active investors in resilience globally for about a decade out of the US. We didn’t really call it that, but we were fairly early investors in companies like Rocket Lab and Skybox Imaging in the space industry — businesses we consider resilience-oriented.

A couple of years ago in Europe, we started to see a growing urgency and imperative on the part of founders and politicians to focus on this category — in particular the increase in spending across the NATO member states, which have committed to getting to five percent of GDP on defense, including critical infrastructure. There was a combination of factors driving greater spend, more urgency, and more innovation. And a lot of it has been driven by AI: many of the companies we’ve seen are underpinned by AI innovation that has created opportunities that previously didn’t exist.

So we put out a piece about our focus on three — now four — areas within European resilience: defense tech, space tech, advanced industrial manufacturing and rare earths, and a fourth area we haven’t talked about historically but are spending time on now, energy and energy infrastructure.

02:47 — A Secular Shift, Not a Hype Cycle

Louise Boucher: So you think it’s going to be a generational shift, not a passing trend. Many people think it’s just another hype cycle. Why do you feel differently — how big do you think this is going to be?

Alex Ferrara: We certainly think it’s going to be a long-term secular shift. The world has changed. Regardless of which administration is in power in the US, what’s happened in the last year or two has been a real wake-up call for Europe. And it’s not just political — there’s a recognition that Europe has begun to fall behind in some really critical areas. The AI boom is another example, where so much of the innovation has been centered in the Bay Area. What’s happened geopolitically has created an urgency that I hope is going to persist for some time.

I also think this is a category where Europe really has a right to win. So many of the European resilience businesses that have gotten started here have unique capabilities that give them that right to win at home — and I don’t see comparable US competitors necessarily coming here to unseat them.

Another factor is the recognition of just how much dependence there is on the US. We see it everywhere — from our payment systems, where we depend on the Visa, American Express, and Mastercard networks, to certain areas of defense tech, which will always be the case. But there are subsectors where Europe really should build out its own capabilities. Also in space: space sovereignty is increasingly critical for national security, and there’s been a recognition that it’s an integral part of defense tech as well.

So the hope is that it will be long-term secular. And I’m hopeful that a lot of the innovations we’re seeing today around defense — and in some cases offensive capabilities — will be grounded in technologies with dual-use potential, and that much like we saw in the US in the 50s, 60s, and 70s, they’ll spawn additional technologies used in commercial industry rather than just defense.

05:40 — Can Europe Defend Itself Without the US?

Louise Boucher: On the reliance on the US — Europe is now spending billions to build its own deep-strike capabilities, and more recently its own ballistic interceptors, so it doesn’t have to rely on the American Patriot or Tomahawk. Do you really think that in the next five or ten years Europe will be able to defend itself without relying on the US?

Alex Ferrara: I think it would be foolish to try. Europe needs the US and the US needs Europe. It’s mutual — perhaps not evenly balanced today, but there’s a mutual dependence and relationship. The types of threats European countries face are partly similar to what the US faces, but there are a lot of differences too. Your proximity to the war in Ukraine, the proximity to Moscow, dictates the kind of capabilities you need to think about strategically. The US is perhaps more focused on things like hypersonics — long-range deep-strike capabilities — while in Ukraine they’re much more focused on short-range capabilities, and countries like the UK, Germany, and France are looking for things in between. It even depends on latitude — whether you’re operating in cold weather or not.

My hope, in this theme of European startups having a right to win: in drones, Europe has the potential to be far more innovative than what we’re seeing in the US, because the feedback loops from testing these capabilities — whether in Ukraine or other geographies — are far greater in number and far more frequent. So in short-to-medium-range systems, both offensive and counter-UAS, there really should be a right to win for European companies. Whereas in an area like hypersonics — there are some exciting companies in Europe, including a wonderful team out of Germany going after it, but there are also great companies in the US, and the US probably has more funding allocated. My guess is they end up with a bit of an advantage there. But again, there’s a mutual dependence — the two regions really need to work together.

08:42 — One Market or Two? Lasers, Drones, and Auterion

Louise Boucher: We both invest globally — US, Europe, maybe the rest of the world. When you see a European company and a US comparable, do you think of them as competitors? Or do you take the hypothesis that the two regions will buy different products, from different customers — so you could invest in two companies doing roughly the same thing in different regions?

Alex Ferrara: It’s a great question. We try to identify specific areas we want to prioritize for investment — usually because we’ve met founders doing really exciting things, and then we research those markets and realize they’re compelling spaces. But we are global, as you mentioned, so we really try to find the best company wherever it resides — Europe, the US, or elsewhere. I’m based in London and most of my time is focused on meeting European companies, but we’ve invested in both.

Just recently — we haven’t announced it yet — we were looking at the counter-drone area and came across a number of companies using high-energy lasers and high-power microwave. The economic asymmetry there is so profound that, as we’ve seen in Israel, if we can get those capabilities to work at greater scale and power, they could be really effective — especially against things like drone swarms. In our assessment, the best team positioned for that opportunity happened to be in the US, and we invested there.

At the same time, we have companies like Auterion, based out of Zurich and Munich with an office in the US. Auterion provides an operating system for drones, and when we looked at that area and the ecosystem they’re building, we felt the best company in the Group 1–Group 2 drone category was Auterion, out of Europe. It’s rare that we find two companies with exactly the same capabilities in each geography, at the same stage.

Louise Boucher: Congrats on Auterion, by the way — they’re really making a killing. Right time, right place.

Alex Ferrara: Lorenz Meier, who runs the company, is just an exceptional founder. It’s been an honor and a pleasure to work with them — hopefully they continue to scale.

Louise Boucher: So impressive: starting with software, starting from Europe — now they’re in the US, in Ukraine, doing their own drones too. The depth of this company is quite amazing.

11:59 — Investing Through the Defense Winter

Louise Boucher: Even if it’s a long cycle, there are still going to be ups and downs — and today we’re clearly on the up. Will you still be here during the defense — or resilience — winter?

Alex Ferrara: Gosh, well, I hope we don’t hit a winter too soon. Maybe it would be good for the world and not so good for investors — so maybe I shouldn’t speak so quickly. Look, I’ve been investing for more than two decades, and a lot of these categories have ups and downs. We invested in some Bitcoin-related companies in 2012 and 2013, and that’s a category that’s had many cycles — several winters. Right now we’re at a fever pitch of interest in the defense tech sector, and I don’t know how long it’s going to last. On the public markets, some of the leading primes have pulled back a little. At the same time, the cadence of news coming out every week suggests it’s likely not ending anytime soon — which is maybe bad for us as a planet, but from an investment standpoint it carries on. The answer is: we only focus on these themes if we think they’re going to be decade-or-more secular trends. Assuming we’re right, we’ll continue to invest in these categories.

Louise Boucher: I think the first winter will actually be quite exciting. I’m a much earlier stage VC, and during a winter you can still have great founders coming to build a category. If they come when the cycle is down, it means they really want to do it — it shows a different kind of resilience and ambition. And they’ll be standing on the shoulders of the giants of this first generation of neo-primes, with lessons on what go-to-market worked and what didn’t — whereas today we’re still figuring things out.

Alex Ferrara: It’s a really good point. In these so-called winter periods, I often get the question: surely you want to invest in a winter because valuations are lower. I don’t find that’s necessarily the case — or at a minimum, it’s not really why we want to invest. What I tend to find is that in winter periods you’re able to take a little more time to really understand the company, get to know the CEO, the founder, the team — form a relationship. In periods like now, at fever pitch, things move so quickly that you often have to make decisions on very limited information — and it’s a little less fun. I really do enjoy getting to know the businesses and the CEOs, and having the time to do that, rather than being forced to invest on very limited information. But I don’t see any signs right now that we’re heading into a winter in defense tech. What about you — have you seen anything out there that suggests we’re slowing down?

Louise Boucher: Actually, no.

15:03 — The Question Nobody Wants to Ask

Louise Boucher: A transition on a very sad note — do you think we’re going to World War Three in the next couple of years? It’s a tough question. But if we’re investing in this cycle for the long term — less reliance on the US, warfare that has changed so it doesn’t depend on whether we’re in conflict, stockpiles that need to be renewed — it also means we’re betting there are going to be more and more conflicts. And that can point towards World War Three.

Alex Ferrara: Look, it’s not a question I’m really qualified to answer. I do, from time to time, ask some of the people we speak to who are far more qualified for their view. A few months ago I had the opportunity to speak with General CQ Brown, the former Chairman of the Joint Chiefs of Staff in the US, and we talked a bit about this topic. The prevailing feedback I get when this question comes up is that it seems very unlikely. But it is a scary question, for sure.

When I think about the silver lining of the environment we’re in — I think we can all agree that warfare is probably the one job we’d all like to see automated. The more we see robotics and drones used on the front lines — and I recognize it’s still extremely dangerous, extremely messy — the hope is certainly that as these unmanned systems become far more widely used, their very existence provides the deterrence needed to avoid unnecessary loss of life. As I hear terms like controlled escalation and de-escalation, I think we’re heading to a world of far more automated systems, perhaps constantly testing one another — but hopefully more carnage among machines, and less in human life.

It certainly feels like the world has gotten less peaceful in the last five years. At the same time, over the last hundred years, the longer-term trend has been towards more peace.

Louise Boucher: It was a tough question, to be honest — and it’s always dark to talk about it so openly.

17:45 — Neo-Primes vs. the Old Primes

Louise Boucher: On a more positive note: the pie of defense investing is big enough for both the traditional incumbents and the neo-primes — just look at Rheinmetall’s share price, up twentyfold. What about you — do you own shares of the public primes, or are you all in on the challengers?

Alex Ferrara: We’re actually very restricted in our ability to buy shares in public companies. And so much of my focus is backing companies at the early stage that I focus a lot more on private companies. So no, I don’t have any investments in the leading primes — at least not directly. I probably have some through index funds.

Louise Boucher: I do have a bunch of ETFs from both new defence and, let’s say, old defence — and for now they’re going head to head. Though the new defence one is really just Palantir basically, so it’s a bit biased by one over-performing stock.

Alex Ferrara: Right — in new defense there are certainly some exciting companies like Palantir. The primes have had a bit of a run-up, and I would worry a bit about the long term for some of them: a number don’t really have dual-use capabilities. They’re purely leveraged to the defense tech trend.

19:23 — The Exploration Company and Europe’s Space Gap

Louise Boucher: Closer to us, and maybe to your portfolio — is there a defence company you’re really bullish on that you’d like to talk about? Maybe it’s Auterion.

Alex Ferrara: One that I’m very bullish about is The Exploration Company, which is a space tech company. As I mentioned, within European resilience we look at defense, space tech, and some other areas — and we’re big believers that space sovereignty is increasingly critical for national security.

In defense tech there’s this concept of shortening the kill chain. It’s a bit of a vulgar phrase — and it doesn’t necessarily mean loss of life; it could be taking out some system. Increasingly, space-based capabilities — constellations of satellites, whether electro-optical, SAR-based, or communications satellites — are important for that ability to shorten the chain: you detect some behavior of interest via satellite, troop buildups at a border for example; you run identification on the satellite, maybe using AI to make decisions on board; you transmit that down to a command-and-control system; and you deploy a capability in response — maybe a warfighter, maybe something else. That whole loop of being able to respond more quickly is becoming increasingly important, and space keeps growing in importance from an ISR standpoint — intelligence, surveillance, and reconnaissance.

The Exploration Company is based out of Germany and France, with a presence in the US as well. They built a capsule to initially do downmass transport from the International Space Station, and they’ve had a couple of successful missions. And now they’re moving into launch. That’s one thing that’s really missing across Europe: the ability to do low-cost launch. We’re still so reliant on SpaceX, and on some of the older programs that have existed for a long time and are pretty expensive. The Exploration Company is really focused on solving the problem of bringing down the cost of launch in Europe, so that we can operate more independently of SpaceX.

22:18 — The One That Got Away

Louise Boucher: I remember I met Hélène a couple of years ago — and in full transparency, mea culpa, I passed on the seed.

Alex Ferrara: Did you?

Louise Boucher: Yes. Really my mistake, because Hélène is incredible — she has the network and the ability to pull off what a space prime is, whatever that means. I was too focused on the capsules. I was like: I don’t see the business, how big can this be? The typical nitty-gritty questions, where you fall into the market sizes and the product specs and you miss the big vision. She’s doing incredible — good for her, bad for me.

Alex Ferrara: It’s been impressive to work with her. And obviously we don’t get a hundred percent of the ones we meet either. It’s even more difficult for you at the very early stage, at inception, because there’s so little information a lot of the time other than the founder. At the later stage, I get the benefit of seeing the arc of the company and some data on how they’re performing.

But this concept of a space-based neo-prime, filling what I think is a big white space across Europe — it’s an amazingly high-potential project. If she succeeds in executing against these big ambitions, I think it could be one of the most valuable companies in Europe — maybe, if she does exceptionally well, one of the top two or three. And that’s what we’re all about: supporting founders as they go after really big ambitions.

Louise Boucher: To be fair, it was back in 2022 and I should have been more visionary. Of course, now Germany is putting in thirty-five billion, now space is a domain in itself, now we’re trying to break the monopoly of SpaceX. So many things which indeed mean The Exploration Company could be worth a hundred billion dollars in the next ten years — it’s very possible.

24:58 — Germany: From No Army to Europe’s Largest

Louise Boucher: Speaking of Germany — when I think about this rearmament cycle, if it’s real, it’s going to change the world order, how local economies are powered, and the global balance of power. Germany went from seventy years of having basically no army to now building the largest European army. Do you think they’ll become a real military power and a global exporter? How much are they going to change the military landscape?

Alex Ferrara: I think it’s a pretty dramatic change. After World War Two, for obvious reasons, there was really limited investment on the part of Germany in defense. Even talking with founders — a lot of Germans grew up anti-defense-investment; it wasn’t a priority for such a long time. Now that’s clearly changed: the US is pushing Europe to cover its fair share of the responsibility, and Germany is the largest economy in Europe — and importantly, it has more fiscal headroom than most other European economies. The moves they’ve made, like getting rid of the debt brake — they’ve probably been the fastest of the larger economies to aggressively increase defense spending. It seems pretty clear they’re best positioned to build out the next generation of defense tech capabilities.

And I’m sure you’re seeing this as well: so many of the most successful emerging neo-primes and defense tech companies are coming out of Germany. I’ve spent more time in Munich over the last two years than I could ever have imagined. You see companies like Helsing, Stark, nuclear fusion companies like Proxima Fusion — and The Exploration Company has a big office in Munich too. These ecosystems get started and begin to build a flywheel, and the exciting thing is we’re really seeing that flywheel kick in, especially in Munich.

Louise Boucher: It’s true that Germany rearming itself brings you to interesting places. I was at IPC in Hanover last week — I didn’t know this place existed two years ago, and now it’s almost close to my heart.

27:13 — King-Making and the Series B Valley of Death

Louise Boucher: One last thing I wanted to touch on: the king-making approach. You come from a world of software where it can also happen, but I think it’s quite specific to defence investing — you can put a billion dollars behind a single company, and that funding itself is what puts them ahead, before the product is proven, before the battlefield decides, because it unlocks a network and a flag-planting that wins them big contracts from militaries. Does that actually work? Do you invest like that?

Alex Ferrara: I do think it works. It’s not the highest priority in how we think about investments, but it is a factor we consider, especially for a late-stage investment. We’re certainly seeing it with Helsing raising this billion-eight round at eighteen billion. Think about the message that sends to the market — the quantum of capital, the valuation, the set of investors behind them. By the way, we’re not investors in Helsing, but I really like the team they’ve assembled, and the ambitions they have are really compelling. You factor all of that in, and they’re attempting to cement themselves — I think fairly successfully — as the leading defense tech neo-prime.

And in a market where there’s a bit of this valley of death you hear about — defense tech companies emerge, they build a great product, but it often takes time before they get meaningful revenue, or a long time to get through the manufacturing hell people talk about. Especially in Europe, I find there’s a dearth of Series B capital. A lot of these defense tech companies have great teams and great products but are not going to be able to achieve the ambitions they set out with. So companies like Helsing, and others that raise large amounts at high prices, are going to be really well positioned to acquire a lot of those teams and capabilities.

So perhaps even more so in defense tech than other industries, I expect king-making will be a factor. And the same goes for space tech — it’s such a capital-intensive industry, with such long cycles. If you want to build reusable launch capabilities, you’ve got the design, the build, the testing, the certifications — very long feedback loops. It’s another area where king-making is particularly relevant.

30:40 — Escape Velocity and the Flight to Quality

Louise Boucher: Which also means at the early stage there’s a flight to obvious quality — people put so much value on Anduril and Palantir profiles, repeat entrepreneurs — almost leaving very little chance for the outsiders and underdogs. But when the underdogs do work, it’s amazing performance.

Alex Ferrara: It’s a bit of an S-curve, right? If you can get to that escape velocity, the performance can be amazing. And being able to attract capital from US funds — once you have the scale to do that — helps quite a bit. Maybe one more comment: in some of these subcategories, the larger funds will only make one, maybe two bets. I don’t think they’re going to make five bets on drone companies. So it’s beneficial to be the leader that’s able to attract capital from those investors — because unlike some other industries, you’re not going to see multiple meaningful bets in the same category.

32:01 — What’s Still Quiet: Space Tech and the AI Why-Now

Louise Boucher: Last closing question. If you had advice for people like me — I’m 30, and when I started out, defense was non-consensus; now it’s consensus. Everything is so noisy: news every day, companies raising all the time. What should we pay attention to? What is still silent — if anything?

Alex Ferrara: Defense certainly seems to have become more consensus. I still think space tech is a little less consensus and less obvious right now. A lot of investors have been scared away by some very high-profile failures — there was this company in the UK, Orbex, that went bankrupt, and those things create a lot of caution. But space tech is a really exciting area. The more we learn about satellite and ground communications, new forms of multi-spectral sensing, even things like space manufacturing and data centers in space — the more excited we get. That stuff sounds a bit out there, but as you get into it, I think it’s an area that’s probably going to become more consensus in the next five years.

For us, a lot of it is the AI underpinning. We always ask ourselves: why now? What’s changed? Because some of these businesses are fundamentally hardware businesses, and hardware is hard to scale. The advances in AI around perception are a big change, and also the advances around manipulation. And the advances in simulation: the ability to simulate multiple different versions of a product and test them digitally before you manufacture — especially for products that don’t have a closed-form solution. That has created a real why-now moment. I’m also pretty optimistic that quantum computing capabilities could further accelerate things, but we haven’t seen a real impact yet — it’s more over the horizon at this point.

But you focus on the earlier stage — you see things at the inception stage. What’s exciting to you these days? What do you think is the next over-the-horizon?

35:11 — Cyber Offense and Biodefense

Louise Boucher: For me, the quietest areas would be cyber offense and biodefense. When you look at it, it’s quite obvious — it’s just that the market is not very big yet. But the needs, the capabilities, and the threats are there, and it’s being talked about in government more and more. It’s the writing on the wall, basically. But there are very few teams going after this, because again, the market is small — there are going to be very, very few teams in the next twelve months going after it. That’s what I’m spending time on now, which is less obvious than the deep-strike capabilities.

Alex Ferrara: What’s a biodefense example? What kind of biological attacks would a company be looking to defend against?

Louise Boucher: You have the chemical weapons and the bioweapons, but it’s more the idea that you’ll have more and more biothreats — with AI models, you can engineer biothreats more quickly. So the idea is a full-stack solution, starting from detection: having a kind of bio-radar that detects very quickly if there’s an engineered signature; then characterizing it, seeing how it will spread; and then designing the right response and drug as fast as possible. As a country: how do you detect, identify, characterize, and protect against new types of biological threats?

Alex Ferrara: So — accelerate the kill chain around biological threats. How you detect and then rapidly engineer a solution. That makes a lot of sense.

Louise Boucher: Yeah. But is there a market? I don’t know.

Alex Ferrara: Right — so much of it is timing. That’s the risk you have to take at these very early stages. And you said offensive cyber too. Are you looking at companies doing offensive cyber as a deterrent, or what kind of companies are there?

Louise Boucher: It’s both. It comes back to the real why-now: cyber has been a threat domain for quite some time — every country has a cyber unit, budgets have been augmented, so that part is more obvious. And similarly with AI, there are a number of new threats, but also new ways to actually equip these cyber units with real tooling. It’s things like fake news — how do you react to it, how do you drown the narrative — but also the deterrent side: how do you put malware in place, how do you counter narratives, basically using the same techniques our adversaries are using, but against them. Like the Chinese Typhoon malware that was installed in the US energy infrastructure — there could be a world where we decide to do the same, so that they wouldn’t strike, because they’d know we control their energy infrastructure from a distance.

Alex Ferrara: It reminds me of the cognitive warfare capabilities you see — companies focusing on misinformation: how you spread it, and how you pair it with a kinetic strike. When a piece of critical infrastructure gets attacked, how do you insert a narrative that it was self-inflicted — that someone domestic who didn’t want the war gave up the location of the recruiting center? That area is pretty interesting, especially as you see the power of social media and how effective it can be at spreading misinformation very rapidly.

Louise Boucher: But then you manipulate the truth.

Alex Ferrara: That’s right — it’s basically telling lies. Using that as an offensive technique to disrupt the enemy, but pairing it with kinetic effect.

Louise Boucher: Yeah. And there are a bunch of interesting teams and projects going after this.

39:39 — Bionic Cockroaches

Alex Ferrara: There’s another one we saw that I thought was pretty interesting — SWARM Biotactics. Embedding cockroaches with the ability to be controlled — what a cool concept: almost bionic cockroaches. How do you create a swarm of cyborg cockroaches that can be controlled remotely to pursue some sort of task? This idea of the melding of biological with robotics over time, and what that’s capable of — it’s a bit out there today, but I think there’s maybe something to it.

Louise Boucher: And this company has the best word of mouth, because I was with the French Special Forces and we heard about it there — everybody wants to talk about the cockroaches that can do stuff for the army. It was a fun one. Okay — it was really great to have you, Alex. Thanks a lot for coming by.

Alex Ferrara: Thank you so much for having me — it was good to chat.

But the more telling story is what’s still missing - and across the categories below, from missiles to biodefense, the need is already loud, the money is only starting to arrive, and the winners are still up for grabs.

Missiles

Missiles used to be a bespoke category: expensive, produced in the hundreds, stored for wars expected to be short. Russia now fires thousands of them into Ukraine. In twelve days last June, Iran launched around 550 ballistic missiles at Israel; stopping them consumed a quarter of America’s THAAD interceptors, and the US fired 30 Tomahawks in a single night in return. A new kind of warfare is taking shape, in which missiles have gone from rare engineering jewels to good-enough expendables made in bulk.

Washington moved first. The Pentagon announced $12.6 billion for the Family of Affordable Mass Missiles (FAMM), to procure 28,000 low-cost cruise missiles. The doctrine behind the price is saturation: fire cheap weapons in volume, force the adversary to spend its exquisite interceptors on them, and enough get through. Production starts this fall, at a pace meant to replenish stockpiles by the thousands each year. Castelion, Anduril, the bootstrapped Co-Aspire and Kongsberg’s Zone5 are all winning large contracts here.

Europe is just getting started as well. At the Ankara NATO summit this month, twelve members signed on to Deep Precision Strike, a UK-led initiative pledging $50 billion over ten years for long-range strike, and the continent is funding its first interceptors against ballistic and hypersonic missiles - one program led by MBDA, another by Destinus. From short-range to deep-strikes, Europe is scrambling to build its own capacity. The first challengers are emerging: Fire Point, born in wartime Ukraine with its 3,000-kilometer Flamingo, and Hypersonica, which just flew Europe’s first privately funded hypersonic test after a €23.3M raise. For a category this large, the field is wide open.

Propulsion

Every one of those missiles, drones and interceptors needs a motor - electric, solid rocket, liquid or turbojet - and I covered how short that supply already runs in the picks-and-shovels piece. What is new is that speed is joining volume as the constraint. Russia launched around 11,000 Shahed-type drones at Ukraine in 2024 and over 38,000 in 2025, and the newest jet-powered variants fly at 500-600 km/h - too fast for the propeller-based interceptors that handled the earlier ones. A turbojet can catch them, but it takes up to two minutes to spool up and comes from one of the most constrained supply chains in the industry. Propulsion now decides whether an interceptor closes the gap on its target or not.

This is why the category is drawing its own generation of entrants - Ursa Major in rocket motors, Volund Manufacturing in low-cost gas turbines, Venus in rocket engines for hypersonic weapons, Greenjets in electric jet propulsion - while neoprimes bring production in-house and established suppliers expand.

Subsea

The first wave of naval innovation happened on the surface: Saronic had its big moments of the Iran conflict, striking and rescuing autonomously, and UFORCE, whose Magura drones destroyed a third of Russia’s Black Sea Fleet, became a unicorn in March.

The next front is underwater. A Pacific war would largely be maritime, and Russia is already conducting a hybrid war in the North Atlantic. Its fleet is surveying and damaging the cables and pipelines Europe runs on - tracking just three of its submarines at work this spring cost the UK’s Royal Navy 500 people, a frigate, 450 hours aircraft flying hours. Nobody can watch an ocean that way.

So governments are buying autonomous, expandable machines to do it. Britain’s Atlantic Bastion, funded at £1.5 billion over four years, will scatter contractor-operated listening drones across the North Atlantic from 2027, and eighteen EU states signed onto a joint seabed-defense project this month, with an indicative €43-72 billion of national investment behind it by 2036. All of this buys the surveillance layer - the microphones and the drones that carry them. Underwater communications, long-endurance patrol vehicles, torpedoes and defenses against torpedoes remain mostly unbuilt.

The M&A deals have followed: Lockheed paid $3.45 billion for sonar house Ultra Maritime, Kraken Robotics $615 million for Covelya, Thales took a stake in Exail, and Anduril’s Ghost Shark grew from a 2022 acquisition into an A$1.7 billion Australian program. Several startups are building in the space as well as Ulysses, Vatn, Albacore, AICE.

Space

Orbit is what makes armies function on the ground: communications, battlefield surveillance, navigation for every guided weapon, missile warning, targeting, GPS.

That is exactly why it is already under attack. An hour before invading Ukraine, Russia hacked Viasat, the American operator whose satellites served the Ukrainian army - the malware broke tens of thousands of modems across Europe and cut the remote control of 5,800 German wind turbines. In May, Russia parked five satellites within 500 meters of ICEYE’s spacecraft, the Finnish radar satellites Ukraine uses to watch the battlefield. Commercial satellites have officially joined power grids and factories in the category of civilian assets that double as military targets, and that now needs defending.

Notice the weapons used so far: a cyberattack, a tailing maneuver - to which lasers that dazzle imaging satellites will soon be added. Space weapons are deliberately quiet and reversible, because the brutal option, the antisatellite missile, fills orbit with debris that kills everyone’s satellites, the attacker’s included. Which is why mass in space can be the winning strategy: if losing ten satellites changes nothing because a thousand remain and replacements launch within weeks, the attack was pointless. Winning in space is a race of production - more mass launched than the adversary, satellites built in series, constellations too numerous to kill.

Golden Dome is the best preview of what a complete space-warfare architecture looks like. The headline is interception - officially $185 billion for an initial capability by 2028, with 7,800 orbiting interceptors. But around it the Pentagon is funding every other layer: constellations tracking maneuvering hypersonic missiles, surveillance of orbit itself, maneuverable satellites trained for orbital dogfights, hardened communications and navigation, and the launch capacity to lift it all. True Anomaly, Anduril, Turion, Impulse and Sierra Space are the ones already collecting on it.

In Europe the serious money is German: €35 billion for military space by 2030, committed before the doctrine for spending it was written. That vagueness is a boulevard for companies able to define the requirement themselves - Helsing, fresh from a $1.8 billion raise and a space-surveillance venture with OHB; ICEYE, whose joint venture with Rheinmetall took a €1.7 billion contract for some forty radar satellites; and The Exploration Company on space transport. Some companies, as Arqus Aerospace, is now looking to expand the warfare capabilities.

Emerging Categories

Nuclear energy

Energy has always been a piece of warfare, and the demand is now spiking. Radars, command posts, satellite ground stations, and the swarms of drones and robots that need recharging between missions all draw power a base which gets from two bad sources: a civilian grid that makes an obvious target, and diesel, whose fuel convoys were among the deadliest jobs in Iraq and Afghanistan.

A microreactor runs a base for years on one core, through any weather, in a short footprint. The US Army ran its own reactors until 1977; it is buying them again. Its Janus Program wants small commercial reactors on up to nine bases, the first mandated by executive order to run by September 2028. A generation of startups has formed to build them - Radiant, Antares Nuclear, Valar Atomics, and Oklo.

Cyber Offense

Cyber moved from quiet espionage to an openly offensive instrument of statecraft - election interference, bot networks steering public narratives, and malware pre-positioned inside critical infrastructure, as China’s “Typhoon” intrusions into American grids, water and telecoms, revealed in 2023, made plain.

States have funded the shift, fast. France almost tripled its military cyber budget since 2018. Britain created its National Cyber Force in 2020, and has since put a £2.5 billion force behind it. Germany has just made cyber the fourth branch of its armed forces, standing beside army, navy and air force. And the United States, the scale-setter, spends around $15 billion a year on military cyber.

All that money bought commands and headcount, not scale: offensive cyber is still run like a craft, one target at a time with bespoke tools. Where the old tools stopped at flagging a weakness and left the break-in to a human operator, the new ones send in autonomous agents that reason like an attacker - pairing precise threat intelligence with execution, and turning what used to be handcrafted offensive work into a product that scales like software. Companies as Twenty, Method Security or Open Minds are to watch here, and this article from Maggie Gray at Shield Capital makes a great deep-dive.

Biodefense

AI is reshaping biology - drug discovery, diagnostics, medicine - faster every year as data and adoption grow. The capability cuts both ways: as AI lowers the barrier to engineering a dangerous pathogen, biodefense faces threats that evolve faster than its legacy systems can adapt. The response has to move at the same speed - spot a novel threat, design its countermeasure, a new drug, almost instantly, then manufacture and deliver it to those most exposed before an outbreak becomes a pandemic.

Still small, this corner is bound to grow: the United States spends around $27 billion a year, Britain is adding £1 billion for a national biosecurity centre, OpenAI has opened a biology-tuned frontier model to allied governments.

The first companies are almost all American - Valthos, Perimeter, Red Queen Bio. Europe has the AI and pharma talent but no equivalent, and this is the layer it can least afford to import: the real asset is not the sensors but the biological intelligence they build up, a living map fed by a country’s own clinical and genomic data, and the contracts that fund it require that data stay sovereign. No government will hand that to a US company, which is exactly the opening for a European builder.

If you’re building in one of these seven categories, reach out — I want to hear from you.

---

Fundraises

Helsing, the Munich builder of strike drones, underwater surveillance and battlefield AI, raised $1.8 billion at an $18 billion post-money valuation, with JPMorgan, Lightspeed and Iconiq.

Kraken, the UK company which designs designs & builds autonomous surface & subsurface vessels , raised $175 million at a $1 billion valuation led by DTCP.

Venus Aerospace, the Houston developer of rotating detonation engines for hypersonics and space, raised a $90 million Series B led by Mercury Fund, with Lockheed Martin Ventures.

Singularity, the LA startup building low-cost air-defense interceptors at automotive scale, exited stealth with an $80 million Series A at a $400 million valuation, co-led by Khosla Ventures and Felicis.

Senra Systems, modernizing wire-harness manufacturing for aerospace and defense, raised a $65 million Series B co-led by Lowercarbon and Interlagos, with Sequoia, a16z and Founders Fund.

Digantara, the Bengaluru builder of satellites and sensors that track space debris and missile launches, raised a $50 million Series B led by Reliance’s venture arm.

Valarian, the London builder of sovereign, compartmentalized data infrastructure for governments and defense, raised a $50 million Series A led by NEA.

Greenjets, the UK developer of electric jet propulsion for drones, raised a $40 million Series A led by Blossom Capital with the NATO Innovation Fund.

Skapion, the Washington-DC-and-Israel counter-swarm startup founded by Iron Dome veterans, raised a $36 million seed co-led by UP.Partners and Khosla Ventures.

Project Q, the Berlin-and-Munich maker of HYDRIS — an open-source platform tying legacy and commercial systems together for multi-domain operations — raised a €15 million Series A led by Expeditions, with HENSOLDT.

Funds

Lakestar, the pan-European firm founded by Klaus Hommels and an early backer of Spotify and Revolut, closed Resilience I at €262.2 million ($300 million) — Europe’s largest institutional fund built solely for defense and dual-use.

Expeditions, the BAE- and NATO-Innovation-Fund-backed manager founded in 2021 by Mikołaj Firlej and Stanisław Kastory, closed Fund II at €197 million; its portfolio already spans UFORCE, Orqa, Frankenburg and Comand AI.

In early 2024 he founded Tencore with the contrarian idea to send ground robots across the most dangerous kilometers of the front line, carrying water, ammunition, and wounded soldiers. Everyone told him it was a toy.

Fast forward two years later, Ukraine wants 100% of frontline logistics done by UGVs, robots are conducting strike missions with no infantry involved, and Tencore has just signed the biggest UGV order in history.

We talk about iterating with soldiers in the trenches, designing a factory for scale under missile threat, moving off Chinese supply chains, and the joint ventures with Quantum Systems, Shark Robotics, and Insta Group taking this technology to the rest of the world.

00:00 — From Breweries to the Battlefield

Louise Boucher: Roman, thanks for taking the time.

Roman Tkachenko: Thanks a lot — happy to see you again.

Louise Boucher: Happy to see you again. We’ve met a couple of times over the past year, and when I first met you, Tencore was really just starting. Now you’re the leader in UGVs — ground robots — in Ukraine. First of all, congrats. But let’s come back to the origin story. You weren’t in defense before. What was the moment you knew there needed to be a company like Tencore?

Roman Tkachenko: It’s true we’ve met several times, and I hope it’s only the beginning of a big story. In our past life, my co-founders and I worked in the food industry. We built breweries and dairy plants worldwide — as a Ukrainian company working on every continent, alongside huge primes like Tetra Pak, GEA, Krones, for clients like Nestlé and Mondelez.

Then the war started in our country. In 2023, when UAVs began to have a big impact on the front line, we heard from a lot of friends that frontline logistics had become extremely complicated. So, as dreamers, we started testing whether some kind of robot could cover the last one or two kilometers — to bring water, food, or medicine.

That’s how it started. We realized it was a possible game changer, and at the beginning of 2024 we opened the company and named it Tencore. The first year we fought to prove that the UGV is the future — that it has a place on the battlefield, and not only the battlefield. It’s a solution that can help save lives. Our main mission is to save as many lives as we can.

Through 2024 we worked hard to modernize the robot and to understand with the front line how it could be used. And frankly, it was only at the beginning of 2025 that the huge ramp-up of UGVs on the front line happened. So that’s the long story short: from breweries, a huge jump into defense — and we knew nothing about defense. Only Max, another co-founder, had a small experience in it. He had promised himself never to work in defense again. Now we have no other choice.

02:57 — Building a Team of Dreamers

Louise Boucher: Since the ramp-up it went very fast — you’re at nearly 200 employees. Who did you go and find first? Were you looking for defense-specific talent, hardware talent? What was the thinking?

Roman Tkachenko: Even today we have almost no one from the military — only some retired soldiers who joined us because of health issues. At the very beginning, we didn’t have anyone from defense at all, I believe for the entire first year.

The first employees were engineers who dreamed with us — it was a team of dreamers. I did the first electronics and batteries myself, another co-founder did the mechanical part. Then the first employees came: someone from hardware, another straight from a car-repair shop. No one was connected to the military in any way. Just ordinary engineers — nothing special.

04:15 — Iterating With the Front Line

Louise Boucher: Still, it’s impressive. We often say the magic of Ukrainian companies is the feedback loop from the battlefield — you iterate fast. But sometimes it feels a bit theoretical. How does it actually work with the units? How do you get product feedback? Did you ever have to change the product based on what they said?

Roman Tkachenko: Yes. Right now we have an almost constant online discussion with the front line — about bugs, proposals, ideas. At the very beginning we worked with one, two, three teams just to validate hypotheses on how to use the robot. Now it’s a common flow of information from many different units.

We work very closely with the front line, and without them we could do nothing. We need to understand the pain points, what’s missing in our solutions, what needs to be added fast — and to prioritize with them what’s crucial to build first. They’re part of our team, really. Tencore is not only the company and the employees sitting here; a lot of our “workers” are in command centers and in the trenches. We work as one team. Thanks to them we have a lot of information — sometimes it’s hard to sort out what’s critical and what can be done a bit later.

Louise Boucher: When you say you’re talking online — is it literally a WhatsApp or Signal group where they send videos of what they liked and didn’t like, and you change the design based on that?

Roman Tkachenko: I can’t tell you the exact channel we use, but yes — it’s essentially an online chat, like WhatsApp or Signal. We have a standard way of communicating. And if a unit doesn’t have a connection with us yet, they can find us — we’re everywhere, including Instagram. A lot of brigades have our contact. We can help them, or they can help us.

06:42 — The Supply Line Stops Being Human

Louise Boucher: It’s been one hell of a year for Tencore, but the next one is going to be even more intense. For context: Ukraine has said it wants 100% of frontline logistics done by ground robots, and it’s contracting tens of thousands of UGVs this year. Are we watching the supply line stop being human in real time?

Roman Tkachenko: War is mainly logistics. The biggest number of missions is logistics — supplying the positions — and that’s the biggest part of what UGVs do today. Right now there’s a huge transformation in the army, from classical logistics to robots. President Zelensky has said we have to purchase a large number of UGVs to fulfill this task.

But at the same time, we have to replace the infantry with offensive robots. You may have seen in the news that attacks have been carried out using only robots — with almost no infantry involved. And alongside logistics, we’re doing a lot of evacuation missions today.

Louise Boucher: Coming back to the origin of the problem — what’s the main advantage of robotizing logistics, whether it’s evacuation or bringing supplies and munitions?

Roman Tkachenko: Cost per mission. First, you don’t have any human near the vehicle — if you lose it, you lose only the UGV, and you can replace it fast. Second, the cost per mission is the lowest available today. You can’t match it with classical logistics. Today we have to count every euro and every dollar per mission, and the UGV is the most cost-efficient solution for this task. And it’s replaceable — if a UGV is damaged, it’s far easier to replace than an armored vehicle or any classical platform.

09:02 — From Logistics to Strike Missions

Louise Boucher: Now the first strike missions have been done with ground robots. Zelensky announced it and wants to double down, and Tencore will play a huge role. I’m guessing strike is about the level of danger — you send robots where it’s too dangerous to send humans. Can you talk about the strike missions UGVs will do in the future?

Roman Tkachenko: It’s better for me not to talk about this topic — it will be announced by our president or ministers, and you’ll see it in the news. I’m not in a position to comment right now, sorry.

Louise Boucher: That’s fine. But even just moving from logistics to strike — did you have to change the product?

Roman Tkachenko: No — it’s modularity. We make a Lego-like design: the product consists of different blocks, and we have the robot plus different payloads, which can include a remote weapon station, for example. Modularity helps us build fast, replace fast, and keep flexibility.

10:23 — Why UGVs Are Hard to Build

Louise Boucher: There’s a misconception that it’s easier to build UGVs than drones, missiles, or USVs. What would you say to those people?

Roman Tkachenko: It’s not easy — it’s a different concept entirely, the same way it would be different for us to start building rockets tomorrow. We work mainly to automotive standards — the way you’d build a car — and that’s completely different from building an aircraft.

We have a lot of processes UAV producers don’t have: welding, cutting, turning, milling. And in my view the quality requirements are higher. Many UAVs are one-way — an FPV just needs to fly a certain number of hours and that’s it. Our machines have to drive hundreds of kilometers. And when we run an evacuation mission, there’s a life on board — the price of any mistake is much higher. That’s why I can’t understand people who say UGVs are easy to make.

12:09 — Designed for Scale From Day One

Louise Boucher: What I found impressive when I visited your factory is that you produce a lot with very few humans — and actually very few robots too. How did you design a product that was made for scale from day one? Did you have to make tough choices?

Roman Tkachenko: From the very beginning, our main task was to build a scalable product. We designed the concept from day one to allow automation and different robots in production — to keep the price low and preserve the ability to scale.

We come from serial production, so from the starting point the request was to make the product as simple as possible, and to keep the possibility of using robots and any kind of automation in production — to keep quality at the highest possible level and reduce the risk of problems with welding or turning.

13:30 — Bottlenecks: Supply Chain and Workforce

Louise Boucher: I think people in Western Europe or the US don’t always understand the conditions you operate under — your factories are under threat of Russian strikes, so you have to relocate them regularly. There’s something very mobile and modular even in how you design the factory itself. Tencore has been used a lot this year and everything seems to be flying — so what’s the bottleneck for a company like yours, or for UGVs as a category?

Roman Tkachenko: The bottleneck today is the supply chain. We’re in transition from a Chinese supply chain to a Ukrainian one — we need a sustainable supply chain with high quality. That’s one point. Another is the lack of workforce. And of course, to be very fair, the risk of getting hit by a rocket or a Shahed.

Louise Boucher: On the Chinese components — why do you care? Is it about compliance, needing Chinese-free components?

Roman Tkachenko: It’s not about compliance, it’s about sustainability. There’s a risk that tomorrow China blocks the delivery of motors, for example, or cameras. For short-term planning it’s okay, but we need to plan long-term — and long-term, we don’t feel safe relying on the same supply chain we have today. We cannot rely on China’s supply chain given today’s global politics. It’s already happened many times — with sensors, and with optical fiber today.

Louise Boucher: Is there a deep enough local supply chain in Ukraine to replace those Chinese components?

Roman Tkachenko: That’s exactly our strategy for this year — to replace as many of the crucial components as we can. In partnership, we’ll open several companies concentrated on the most crucial parts, and we’ll produce them in Ukraine.

16:02 — Joint Ventures: Germany, France, Finland

Louise Boucher: Do you want to talk about the partnerships?

Roman Tkachenko: I can tell you about the partnerships we’ve already announced. One is in Germany with Quantum Systems — we signed with their subsidiary and we’re now opening a company called Quantum Tencore Industries, located in Munich. We’ll co-produce the Termit for Ukraine and run joint R&D projects. Another one is in France with Shark Robotics, who deliver firefighting robots — we’ll work with them in defense and dual-use. And in Finland we signed an MOU with Insta Group.

We’re building different kinds of partnerships to scale and to exchange experience — what we have, and what Europe has. Combining two different companies from two different countries brings a lot of benefits to both — as companies and as countries.

17:09 — Is There a UGV Market Outside Ukraine?

Louise Boucher: How mature is the UGV market outside Ukraine? If you take France, Finland, or Germany — are the ministries of defense already buying UGVs? Have they understood that logistics and strikes are getting robotized?

Roman Tkachenko: Today we’re working with our partners, and alone, in different countries — with MODs and army staff — to show them the capability of the UGV. Frankly speaking, there is no market right now; some projects are just starting.

But we see huge potential, because UGVs here in Ukraine have already proved they’re the most efficient way of fighting today. That’s what we’re trying to do in Europe and in the UK — to convince the countries supporting us that UGVs are the future and they need to invest, at least in training and exercises. The sooner countries start experimenting and implementing UGVs into their doctrine, the better — because today, we don’t have much time. We need to move faster.

Louise Boucher: I sometimes wonder what the wake-up call will be. Iran was the wake-up call on interception for these armies. Maybe for UGVs it’s just a question of time. And I see you’re also selling to the US?

Roman Tkachenko: Not for defense right now. We’re delivering a proof of concept — mainly a logistics robot for firefighting crews. It’s a dual-use track, not a defense track.

19:07 — Eurosatory and the Ukrainian Wave

Louise Boucher: You were at Eurosatory recently, the big defense fair in Paris. What was impressive is that Ukrainian companies had a major presence — very big booths and stands. Do you feel a different pulse in how the world sees Ukrainian companies?

Roman Tkachenko: Indeed. Over 60 Ukrainian companies were presented at Eurosatory. I didn’t attend two years ago, but I was at MSPO, the big exhibition in Poland — and compared to that, even year on year, the number of Ukrainian companies is three, four, five times bigger. The booths are bigger, there are many more products. It shows two things: the foreign market is interested in Ukrainian products, and we now have a big number of strong players with something to show.

Louise Boucher: Why expand through joint ventures, and how do you pick the partners? I guess it’s part of the globalization of Tencore’s strategy.

Roman Tkachenko: To make a long story short — yes, we’ve announced several joint ventures worldwide. With Quantum Systems in Germany — and last week the implementation agreement was signed between the countries; it’s the biggest UGV order in history for the time being. With Shark Robotics in France, with Insta Group in Finland.

We’re trying not to arrive in a country and fight the local industry — not to come as a competitor. Every country and every company is strong in some particular products and technologies. If you combine the best from all of them, you strengthen both countries, and we can move faster together. We want to concentrate on the opportunities, not on the competition.

Louise Boucher: So with Shark Robotics, for instance, the idea is to sell whole multi-domain systems — Tencore UGVs alongside their own robots — directly to the French Army?

Roman Tkachenko: Yes, and it goes further — you can share not only products but also subsystems. Some of our subsystems we can share with Shark, and some of theirs they can share with us.

22:34 — A Global Company With a Ukrainian Heart

Louise Boucher: Tencore is hitting an inflection point — the speed of scaling in Ukraine, and now the global exports. Where do you see the future of Tencore?

Roman Tkachenko: It’s a global company with a Ukrainian heart.

Louise Boucher: So you keep the headquarters and the team in Ukraine?

Roman Tkachenko: Yes. We’re thinking about foreign sales and foreign presence, but at the same time we want to keep the Ukrainian heart — just not to lose it.

23:00 — Dreamers Proven Right

Louise Boucher: Is there a belief about this war that you held 18 months ago and have reversed on?

Roman Tkachenko: No. We have the strategy: we have to survive, and we have to build a defense-technology umbrella in Ukraine. Nothing changed. We still have to move forward.

Louise Boucher: Not even the fact that your UGVs are now being used more and more for strike missions? It shows how the conflict is evolving. But maybe you were always a big believer.

Roman Tkachenko: We were dreamers from the very beginning, when everyone told us it was a toy and they didn’t want to use them. From the very beginning we saw that this kind of technology could bring a huge impact on the front line — and that’s what we can see right now. We developed it from day one with the understanding that it could replace soldiers and take on the risk at the front line. And I can tell you, you’ll see more and more good news where UGVs are the main actor.

Today UGVs are the main player in logistics and some additional missions, but there are still a lot of opportunities to cover. That’s why we keep dreaming — combining different technologies, testing them, validating them on the front line. You’ll see many more videos in the news.

Louise Boucher: I’m looking forward to it. If you look at how few drones are used by Western armies — except maybe the US — relative to their size, UGVs similarly have barely touched the surface of what needs to be built, bought, and used.

Roman Tkachenko: The main difference is that we count money — cost per mission, cost per evacuation, cost per logistics run. In the current warfare we have in Ukraine, this is the most efficient way to fight; there’s no more efficient technology than what we’re using right now. That’s why all of our foreign partners say they’re drawing a lot of lessons learned from Ukraine, even in drone technology.

And you can see the same from the latest conflict in the Middle East. A year ago in Europe, everyone told me: “It’s your war — any other war will be different.” The conflict there showed the opposite: drones are having more and more impact. It’s not just the warfare we have in Ukraine — it’s a global challenge everyone has to accept.

Louise Boucher: Though Iran was much more about air defense — drones, interception, missiles. They didn’t really use UGVs.

Roman Tkachenko: Because there was no land operation — it was an air campaign. But even in that case, you saw the huge impact of drones and of intercepting drones. It shows the global shift from expensive equipment to mass products — mass drones that can overload any anti-drone or anti-rocket system. It will be the same on land.

Louise Boucher: And they did use autonomous surface vessels, since it was also a maritime conflict. Well — thanks a lot, Roman, for your time.

Roman Tkachenko: I’m very happy to see you.

These markets have attracted far less capital, for an understandable reason. A prime owns the program, the customer relationship and the pricing power, so it earns higher revenues and higher valuations than its suppliers, and in normal conditions the supplier holds the worse position. Current conditions are different. Munitions order books now stretch years beyond capacity, customers are investing in their own suppliers to secure allocation, and governments are paying for factories.

In a market so constrained, suppliers’ position improves considerably. It is driven by players scarcity (how many of them are capable to produce a part) and the size of the demand spike (how much of that part is suddenly needed).

Solid Rocket motors

Solid rocket motors are the most constrained of the four markets and the most concentrated, which explains why they have seen the most investment activity. A motor is an expendable. Every interceptor fired over the Red Sea and every rocket launched in Ukraine consumes one. Yet, its industrial base shrank from 6 American producers in 1995 to 2 by 2017, Northrop Grumman and Aerojet Rocketdyne, the latter acquired by L3Harris for $4.7B in 2023.

This is why the US Congress allocated $400 million to develop new entrants. X-Bow, founded in 2016, raised a $105 million Series B led by Lockheed Martin last year and holds a $191 million Air Force contract. Anduril entered by acquiring Adranos in 2023 and targets 6,000 motors a year, and Mach Industries just paid $50 million for motor maker Exquadrum, citing the shortage as its reason. Ursa Major just closed $100 million in November 2025 and is redesigning the motor of the Navy’s Standard Missile family. Meanwhile, Firehawk is leveraging additive manufacturing to radically solve the bottleneck.

Europe is no better supplied. Its solid-motor industry consists of essentially three houses: Roxel, wholly owned by MBDA and powering a portion of its missiles; Avio (trading at $1.8B) which just raised $100M with Advent to expand production; and Nammo - mostly known for its ammunition business. And they are now being stretched from across the Atlantic: when Raytheon needed a second source for the booster of its SM-3 and SM-6 interceptors, it picked Nammo - which is investing $200 million on a rocket motor plant in Florida.

Jet engines

Similarly, jet engines are supply-constrained even though they’re consumables. A small turbofan is probably the most difficult item on this list, requiring precision blades that spin at extreme temperatures and know-how accumulated over decades, and the result is that America has basically one supplier at scale: Williams International, a private Michigan firm with an estimated $0.6B-1B in revenue and who received $250 million from the Pentagon in 2024 to expand capacity. In Europe, the only large supplier is Czech PBS - 800 people, roughly $160M in revenue, 1,700 engines built this year - whose turbojets power Ukraine’s deep-strike drones, and which is following the demand west with a plant near Atlanta. Smaller players exist, as Alloyed, Sierra Turbines, Beehive and JetCat, or Acutronic Group but operate at a smaller scale.

And yet, demand is about to grow well past what these firms can serve. To beat interception, drones and Shaheds are getting faster: Russia has launched eight times more jet-powered Shaheds this year than in all of 2025 - and faster means a jet engine. This is why, the Pentagon just signed agreements to buy more than 10,000 low-cost cruise missiles over three years, each requiring a small jet engine.

The industrial difficulty is the barrier to entry: Anduril ran an internal engine effort for years before giving up, Mach Industries stood up an engines team two years ago but it’s not clear if this effort has succeeded, and among the new generation only Destinus has cracked series production, with a thousand of its T150 turbojets built this year - a very impressive feat.

Radios

Military communications are complex. A military link has to keep working while an adversary tries to jam it, which takes years of waveform engineering, and the category spans many distinct products, each an engineering problem of its own: video links, datalinks between autonomous systems, soldier radios, vehicle networks.

Ukraine made this a large market as electronic warfare changed the communications ecosystem. The global leader Silvus Technologies grew from roughly $100 million in revenue a few years ago to $475 million in 2025, and Motorola Solutions bought it in August for $4.4 billion upfront, guiding to $750 million of revenue for 2026. Two openings remain. First, Silvus sits at the expensive, hardware-first end of a market that is moving toward cheaper software-defined links, and several primes and drone makers are actively looking for that alternative. Second, the radio makers face a bottleneck of their own in the radio-frequency chips and amplifiers inside every link. There is room for a vertically-integrated new communications prime, and it has happened before: Harris built itself into one of America’s largest defense contractors on tactical radios before merging with L3 in 2019. Companies such as Sine Engineering, Himera or Zerophase are interesting to watch in that area.

Thermal cameras

Demand for thermal cameras has been growing steadily, because autonomous systems increasingly need to operate at night. So far the Western incumbents (2 US and 1 EU) have seemed to absorb that growth because cameras, unlike motors, are less of a market in shock. A camera is not consumed in use: a thermal sight or a drone payload serves for years, where every rocket fired destroys its motor. And the one segment where cameras are consumed, the FPV drones expended by the thousand in Ukraine, runs almost entirely on cheap Chinese modules. What would change the market abruptly is a ban on or from China.

China holds the strongest position here of any market in this piece, on both ends of the supply chain. Chinese firms produced around 60 percent of all thermal sensors, at prices up to 40 percent below Western rivals, and China refines most of the world’s germanium, the metal used for thermal lenses because ordinary glass is opaque to heat radiation. If Chinese thermal components are excluded from Western drones, or if Beijing’s germanium ban returns when its suspension ends in late 2026, roughly 60 percent of world supply has to be replaced at once, at Western prices, and every Western camera maker must redesign its supply chain at the same time.

The early effects are already visible. LightPath Technologies, the Nasdaq-listed maker of germanium-free lenses, saw its backlog rise from $20 million to over $100 million in months, and two American drone makers, Ondas and Unusual Machines, invested $8 million each directly into LightPath to secure their lens supply.

Outside China, real capacity sits with four companies: Teledyne FLIR and Seek Thermal in the US, and in Europe Lynred - the Thales-Safran joint venture with €225 million in revenue that is expanding capacity by half - and Exosens, which generates roughly €450 million in revenue. Four companies cannot absorb that redirected volume. Rocket motors looked the same way in 2022, two suppliers facing a surge in demand, and that gap is what financed the current generation of motor startups. Companies as Odd Systems are interesting.

Some miners bring their own shovels

The complication across all four markets is vertical integration. When a bottleneck threatens them, the neoprimes buy or build the component themselves. Anduril acquired its rocket motor maker in 2023 and a cooled-camera maker last October; Destinus built a thousand of its own turbojets this year rather than depend on suppliers. For an investor this cuts both ways. The company best placed to be a customer may become a competitor, but each of those deals also confirms that the parts are worth owning, and acquisition is how component investors have mostly been paid so far.

Where this can go

Harris built its communications business to roughly $2 billion in revenue before broadening into everything else. Since then, the realistic outcomes in components have clustered between $1 billion and $5 billion: Aerojet Rocketdyne acquired for $4.7 billion, Silvus for $4.4 billion, Karman and Exosens worth $3 to $4 billion on the public markets. The leading neoprimes are already valued at ten times those figures. TransDigm and Heico show that parts businesses can compound into companies worth tens of billions, though each took decades of acquisitions to get there. The two trades have different shapes, but they can meet. In categories this supply-constrained, a new entrant willing to take risk and acquire aggressively is the one that could turn a parts business into a venture-scale outcome.

If you’re building in any of these categories - feel free to reach out.

Funding and deals

Fundraises:

• Quantum Systems, the German reconnaissance drone maker, raised a $1.2B Series D at an $8B valuation, co-led by Blackstone, Noteus, Airbus and Advent.

• Stark, a two-year-old Berlin startup making attack drones for European militaries, raised $569.2M at a $4B post-money valuation from investors including Founders Fund, Sequoia Capital, the NATO Innovation Fund, Project A and Döpfner Capital.

• Peregrine, which builds a full-stack AI platform for governments, raised a $250M Series D at a $6.8B valuation, led by existing investors including Sequoia Capital, Fifth Down Capital, OG Venture Partners, Goldcrest Capital, XYZ Ventures and Godfrey Capital.

• Dominion Dynamics, a one-year-old Ottawa startup developing command-and-control software, Arctic surveillance sensors and autonomous drones for extreme environments, raised a $97.9M Series A led by Georgian, with Valor Equity Partners, Expeditions and Lakestar.

• Traysar Industries emerged from stealth with a $25M seed led by Silent Ventures to build the technology for military supremacy in the subterranean domain.

Acquisitions:

• KNDS, the Franco-German tank maker, is selling equal 40 percent stakes to the French and German governments at a valuation between €15B and €18B, ahead of a dual listing in Frankfurt and Paris.

• Thales has acquired a stake into Exail, the developer of autonomous maritime systems, which would value the company at €3.9B.

• Lockheed Martin snapped up anti-submarine warfare-focused defense company Ultra Maritime for $3.45B.

• Booz Allen, the government services giant, is acquiring Ultra I&C Mission Solutions for $720M. The company builds mission software and edge compute that is quietly everywhere.

• Kraken Robotics, the Canadian maker of subsea sensors, acquired underwater technology company Covelya Group for an undisclosed sum.

Sine Engineering is powering the comms backbone of Ukraine’s drone war — keeping hundreds of thousands of drones flying when GPS is jammed and the enemy is hunting every signal.

The company went from 0 to 200 employees across Ukraine and Finland in 3 years, raised with the best US investors and is the leading provider of communications for drones in Ukraine.

Andriy Chulyk sat with me to discuss why electronic warfare (not firepower) is the real edge on the modern battlefield, how they’re scaling globally, and why they were just chosen as the first-ever investment of the US–Ukraine Reconstruction Fund.

00:00 — From Standing Desks to Defense Tech

Louise Boucher: Hi, Andriy. I’m very happy to have you here. We met for the first time more than a year ago in Kyiv, and the progress you’ve made since then has been outstanding. Thanks a lot for taking the time.

Andriy Chulyk: It’s nice to be here, thank you. I’m glad to tell you more about our company and how we’re doing.

Louise Boucher: In February 2022 you were running a very different kind of business — a standing-desk company in Lviv. By April you’d founded a defense-tech company, Sine Engineering. You had no prior defense background, and you chose one of the hardest problems to solve: military communications in a contested environment. Fast-forward four years, and you’re a leading provider of communications solutions in Ukraine. How did that happen?

Andriy Chulyk: Everything we did connected to Ukraine and the war came out of necessity — the need to do at least something to make our military more effective on the battlefield. The full-scale invasion in 2022 was a huge shock for all of us here. So we simply tried to be helpful and to find some connection to the military. I had no relationship with them before, so I reached out through a Facebook page.

We started with fairly simple things, but over time we did a lot of work related to radio, to UAVs, and more. One of our very first projects wasn’t radio at all — it was electronic warfare. We helped create a really effective jammer against various Russian drones. But then we saw that a lot of companies in Ukraine had big problems with communications. Based on what we’d learned doing EW, we realized we could do something genuinely strong with comms. We started on that toward the end of 2022.

02:17 — Why Battlefield Comms Are So Hard

Louise Boucher: Let’s come back to the comms problem for the audience. What does it actually mean, and why is it hard? I think most people underestimate how important and how difficult it is.

Andriy Chulyk: There are three things that matter most for any UAV: reliable communications, positioning, and some level of autonomy. Comms and positioning have to work everywhere — that’s why they’re so critical for every team and every company trying to build their own flying vehicles. That’s why we focused on comms, and from comms we built our GPS-denied positioning.

Why is it so hard? In current warfare you’ll find a lot of electronic-warfare systems on both sides, so you have to provide good, reliable comms everywhere across the front line and in all conditions. That takes really high-level engineering. The same is true for positioning. You can get positioning on the battlefield today with very good CRPA antennas — arrays of 16 antennas or more — but they’re extremely expensive. So you have to find another way to get a position across the front line. Time-of-flight positioning, or positioning based on beacons, is one of the viable approaches.

04:06 — Is Electronic Warfare Actually New?

Louise Boucher: With Ukraine, the world has heard a lot about electronic warfare, GPS jamming, GPS-denied environments. But it makes me wonder — in previous conflicts, electronic warfare seemed like a smaller component. Is this something new that’s appeared with Ukraine and Russia?

Andriy Chulyk: No, it appeared long before — but the specifics were different. Before the full-scale invasion, everything was about big, expensive machines and platforms: Rafales, Eurofighters, F-22s, large drones. The comms on those were based on software-defined radio, and the systems were very expensive. But back then we weren’t talking about a million drones.

Now we’re talking about millions and millions of drones. It’s not cost-effective to put that same expensive equipment on a $300, $400, or $500 drone. So you have to use something more affordable that’s still fully operational on the front line. The whole approach to building these systems has to be completely different — and that’s exactly why we’ve had success with our product.

05:49 — Staying Ahead: Iteration in Hours

Louise Boucher: Russia isn’t just jamming — there’s spoofing of coordinates and positioning, frequencies shifting in real time, constant adaptation. How do you stay ahead of an adversary that’s also iterating every day?

Andriy Chulyk: First, on why we chose comms over EW: our team’s vision is that comms will win against EW. You can do a great deal with comms to defeat jammers, and that’s why we went this way.

As for staying ahead — it’s all about iteration. One of our main products has gone through around seven or eight hardware iterations, hundreds of firmware iterations, and I believe a few hundred software iterations on top of that. We iterate continuously. We have an integration team doing a lot of testing on the range, and we work directly with many military units who give us constant feedback. That helps us improve very fast — sometimes the iteration cycle is just a few hours. That speed helps us enormously.

07:24 — PASIKA and Mass Control

Louise Boucher: There’s also what comes next. You have PASIKA, a platform where one operator controls multiple drones, switching between them in real time. How close is that to being operationally real, and what changes when it’s mainstream on the battlefield?

Andriy Chulyk: We don’t call it swarming — we call it mass control. And it’s fully operational. You can use it with different models everywhere across the front line. It’s not a thing of the future; it already exists, and it can be installed on almost any platform — fixed-wing, multicopter, and others.

The idea came from a simple problem: we have a lot of UAV pilots, but each one usually controls a single drone, and that’s very hard to scale. So you need a different approach that lets you control many drones simultaneously. PASIKA is a great example of how that technology can deliver mass-control capability. But we don’t force customers to use only our PASIKA software — we provide the technology and an SDK, and they can build their own. For example, we have an integration with Mosaic from Quantum Systems. It’s an open platform, so it integrates easily with other systems.

09:22 — Finding the Talent

Louise Boucher: Coming back to talent density — building comms is hard, and very few people actually know how to do it. Where did you find your talent? Did you bring in people from the industry, or bring in outsiders and upskill them?

Andriy Chulyk: You’re right, it’s really hard to find the talent to build this. And finding talent is one thing — having a clear vision of exactly what to build and how it should be done is another. That’s why we look for talent in Ukraine, Finland, Sweden, and maybe Germany soon. We’re a global company now, trying to hire talent everywhere. But beyond talent, the crucial thing is understanding precisely what to implement in your product and how.

10:26 — Going Global

Louise Boucher: On the global part — you built all of this from Ukraine, with the war happening around you, and now you’re expanding. You’ve opened an office in Finland, you have a US holding entity, you’re selling to many countries. What did going global actually require in how you think and operate? Did you plan it from the start, or did it happen gradually?

Andriy Chulyk: If you’d asked me a few years ago whether we’d open an office in Finland, I’d have said never. But now we have one, with a lot of people doing great work there. So we’re open to anything. We’re thinking about opening a big R&D center in the US as well. We’re also expanding beyond comms into computer vision and AI. We just try to follow our customers and give them everything they need to build their best platforms and systems.

11:42 — Why Finland, and How They Choose

Louise Boucher: Was it more opportunistic — in Finland, for instance, there’s the Nokia heritage, so I imagine a deep pool of comms talent. How did it happen? Was it a customer pulling you in? How do you decide which country to expand into?

Andriy Chulyk: It started at a small conference in Tampere, maybe a year and a half ago. We visited the city, we liked it, we had conversations with a few people — and afterward we decided we should definitely do something there.

12:17 — Raising From US Investors

Louise Boucher: The global dimension is also in your cap table — you raised from major US investors. I think you were one of the first Ukrainian companies to attract that kind of quality foreign capital. How did that go? Were they open to Ukrainian defense? Did you have to educate them?

Andriy Chulyk: It’s genuinely hard for US investors to invest in Ukraine. They know exactly what to do with US teams, but not with Ukrainians. So it’s still quite unique — but it’s changing, and I believe it will soon be much easier for them to invest in Ukrainian companies, or at least companies with Ukrainian roots.

Louise Boucher: It feels like an ecosystem kickstarting — even if it’s new, each company can vouch for the next, and it builds from there.

Andriy Chulyk: There are more and more companies doing similar things.

Louise Boucher: Do you know what worked for you specifically, given you were one of the first? Was it the US entity? The talent you’d built outside Ukraine? How did you bring those investors on board compared with other Ukrainian companies?

Andriy Chulyk: It’s a combination of a lot of things. We’re more about the underlying technologies than the final products like UAVs — that’s one thing. We were also profitable and self-funded, with strong traction. And we were very intensive with our prospective investors, so we ended up oversubscribed. But it all depends — every company’s situation is unique. If someone asked me how to build something similar, I couldn’t give a single answer, because everything is unique.

14:19 — Keeping the Bootstrapped Mindset

Louise Boucher: As you said, you were bootstrapped early and became profitable quickly. Now, with this fundraising, you have plenty of cash. Do you still keep some principles from your bootstrapped days? Does that teach you something different?

Andriy Chulyk: I’d say we’re still essentially bootstrapped — we’ve just spent the investment on international growth. The point of the seed round was really to show that we’re an investable company. That was the idea.

14:56 — The First US–Ukraine Reconstruction Fund Investment

Louise Boucher: Recently you had big news: you were selected as the first investment from the US–Ukraine Reconstruction Fund — the fund born out of the minerals deal with the US — beating out more than 200 applicants. What did it feel like to become, in some ways, the symbol of a new political partnership?

Andriy Chulyk: I wouldn’t frame it as beating 200 companies. I think Sine was the best choice for both governments because we met all the requirements and we were very open throughout the due-diligence process. We also have great US partners who helped us a lot with the negotiations — with the DFC, and with the Ukrainian government. It was a unique process, but beneficial for all parties.

Louise Boucher: What does the partnership actually look like — both for Sine as a company and more broadly, in terms of this US–Ukraine agreement?

Andriy Chulyk: I’m sure it’s a big win for Sine to be the first, and to have both governments as partners. I’m also sure it will be very helpful in the future, because when an investor like that puts money into you, everyone knows you’ve passed serious due diligence — there’s nothing unclear about your company. That’s why we decided to go with URIF.

Louise Boucher: It sends a strong quality signal. More broadly, what is the US–Ukraine Reconstruction Fund about? What’s the goal for both countries?

Andriy Chulyk: It’s a private fund backed by two governments. Overall it’s a private fund, but with some features you don’t usually see in private funds — extra requirements and a few other things. But fundamentally it’s just another fund operating for profit.

17:43 — The Future of Ukrainian Defense Tech

Louise Boucher: Let’s zoom out, because this episode — and Sine’s success — is also about the future of Ukrainian defense tech. Ukraine now has hundreds of defense-tech companies, battlefield testing, 24-hour feedback loops, and a generation of engineers stress-tested like no one else. The million-dollar question: what happens to this ecosystem when the war ends?

Andriy Chulyk: I think it’ll be similar to Israel after its conflicts. All the big companies will have a presence here. It’ll be a huge win for the industry overall, because the best of what exists today for UAVs, you’ll find in Ukraine. I believe it’ll become a major part of Ukraine’s economy — a kind of Silicon Valley for drone companies.

18:39 — Where Sine Goes Next

Louise Boucher: And you’re right at the center of it — building communications for drones, UGVs, interceptors. Drones now account for around 80% of casualties. What does that tell you about Sine’s future? Where do you see the company in five or ten years?

Andriy Chulyk: We definitely want to stay at the frontier of the technologies tied to the commercial and dual-use markets, and to go deeper. For example, we’re now working on custom-silicon projects, so we can create alternatives to systems like DJI’s. There’s a lot of different work ahead.

Louise Boucher: So you see yourself keeping the global headquarters in Ukraine, with teams and customers everywhere in the world?

Andriy Chulyk: Yes, definitely.

It was not a shock to the people who run European defence.

The generals already knew

Defense leaders have been planning around this for two years, because they have already been on the receiving end for military systems: Washington slow-walked missile deliveries, held back Patriots, and made clear that permission to fire Western weapons into Russia could be pulled depending on how Putin retaliated. Britain’s Storm Shadow, which depends on American guidance systems and cartographic data, has been sitting unusable since late 2024. And the unreliability, even withdrawal, is only accelerating: a few weeks ago Trump announced he was pulling 5,000 of America’s 36,000 troops out of Germany and cutting what it keeps on call for Europe - fighter jets from roughly 150 to 100, all 8 aerial-refuelling tankers withdrawn, a carrier and a missile submarine redeployed. NATO’s own spokesperson called it “the end of an over-reliance on US forces.”

The growing disengagement from the US has sparked a massive shift inside European defence: to build our own alternatives, free of US components. Britain unveiled this week three prototype long-range missiles built with no American parts at all - MBDA’s “Crossbow” swaps in an in-house visual navigation system for the US guidance it would once have used. Ukraine’s own Fire Point, its largest missile and drone maker, is turning out ITAR-free strike drones and cruise missiles for the same reason — and has now flight-tested a ballistic missile meant to anchor a homegrown air-defence system, a potential answer to the Patriot.

FirePoint dominated the Eurosatory show last week - showcasing its successful deep strikes missions on Russian territory.

Not anymore just a French reflex

The same instinct has now reached the software layer. France is replacing Palantir at the DGSI, its domestic intelligence service, with a homegrown firm, ChapsVision - around €300m in revenue against Palantir’s $5bn. It is a striking reversal: France was the first country outside the US to buy Palantir’s Gotham, after the 2015 attacks. Yet, you could still shrug it off, because France has always leaned sovereign and backing a national champion is what Paris does (remember when OVH was going to replace AWS or Qwant Google?).

The new David vs Goliath.

Harder to shrug off is who else is moving the same way: Germany’s domestic intelligence agency, the BfV, has signalled it will pick the same French firm over Palantir - already embedded in police forces in several Länder. Even Britain is reportedly weighing whether to drop Palantir from its large, historic NHS contract.

When countries with no national champion to protect start making the same call as Paris, it stops looking like a French chauvinistic reflex and starts looking like a direction of travel.

Fear is the wrong lesson

All of this is being sold to Europeans as fear: an unstable world, an unreliable ally, exposure on every side. It is accurate. It is also paralysing, and it casts rearmament as a bill to be paid rather than the opening it really is. What Europe actually needs is the opposite story, told on purpose - a positive narrative around defence renewal - because fear funds a budget line for a few years before the backlash sets in.

Because the genuinely exciting thing, underneath the instability everyone keeps invoking, is that the moment forces Europe to build strong champions in the frontier models, the software that runs on them, and a defence industry pushed to innovate again. Defence is a tremendous innovation power and has a long habit of producing the technology everyone later takes for granted. The internet began as a Pentagon network. GPS was built to steer warships and missiles before it guided anyone’s taxi. The voice assistant in every phone grew out of a DARPA project, and Bell Labs gave the transistor to the world on military contracts. A technological revolution and a redrawn world order have arrived together - a one-in-a-century alignment Europe can turn into large new winners rather than just a heavier defence bill.

Germany is the one country betting its next industrial chapter - arguably its whole economic future - on defence, treating rearmament not as a cost to be endured but as the foundation of its next set of champions for the 21st century. It voted through the largest defence budget in Europe and used the moment to revive an industrial base of carmakers, chemical works and precision engineering. Those budgets became a clear demand signal, and in a few years Germany has already produced some of the most exciting defence companies on the continent - Helsing, now valued at €17bn, and Stark at €3.5bn. Carmakers are rushing into the new ecosystem: the interceptor startup TYTAN is teaming up with Mercedes-Benz, Continental has paired with drone company 24Industries, and so on. Munich is now the hotspot for physical-AI innovation in Europe - as every ambitious founder in Europe is now opening a local team or eyeing a full relocation.

Germany has already shown what that looks like. Europe spent decades importing its security; this is the moment to build the global champions that sell it back to everyone else.

Aaron Applbaum, co-founder of Kinetica, sat with me to talk about building a new Israeli defense fund at the exact moment that refresh begins — backing founders before they’ve even left the army, in partnership with 8VC.

We get into why October 2023 pulled Israel’s best engineers back into uniform and changed what they wanted to build, and where Israel has a real right to win — from the algorithms behind information warfare to its long lead in aerial defense.

Transcript (lightly edited for clarity and flow) is below.

00:00 — What Kinetica Is Trying to Build

Louise Boucher: It’s great to have you here. It’s been a while since we’ve seen each other in person. I was excited to have you on because you recently announced the launch of Kinetica — a brand-new fund that’s already deployed, I think, more than $20 million across six exciting companies. So set the scene for me: what is Kinetica trying to be that didn’t exist in Israel before you built it?

Aaron Applbaum: First, a quick update — we’ve actually just completed our seventh deal, so we’re moving at a great clip.

The Israeli tech ecosystem has long been a leader in developing next-generation solutions. It’s also a country with compulsory military service: every 18-year-old, male and female, serves. Sometimes that’s in a kinetic capacity, and often it’s in a technological one. Young people get to put their hands on real problems, innovate with effectively unlimited budgets, and solve incredibly difficult problems.

And yet, for some reason, the worlds of defense and deterrence on one side, and tech and innovation on the other, never really crossed paths in Israel. There were a few exceptions, but largely the innovation coming out of the Israeli military was redirected into purely civilian use cases. That’s how Israel became a leader in cybersecurity, fintech, advertising, and gaming.

Then, in October 2023 — a shocking moment in Israeli history — the country was thrust into a devastating, costly, and emotional multi-front war that it still finds itself in. Because the Israeli military is not only compulsory but largely reserve-based, it wasn’t just bright 18-to-25-year-olds who were called up — the entire tech industry descended back into duty all at once. They were suddenly working in the domains of life and death, of national resilience, independence, and security.

When many of these high-quality engineers, entrepreneurs, and founders finished their reserve duty, they found it hard to go back to building interesting but less meaningful things. For the first time, we saw a convergence of the Israeli capacity to innovate, iterate, and build quickly with a deep sense of duty and giving back at the national level. The best engineers and founders started to build in defense.

At the same time, global demand was ballooning — largely because of the Russia-Ukraine war, which in many ways has been even longer and more devastating in human terms than the Israeli conflict. You started to see NATO spend go from 2% toward 5%; the Americans moving from roughly $850 billion toward $1.5 trillion; Japan coming back online for the first time since World War II; and the Middle East thrust into multi-front wars, with Saudi, Emirati, Qatari, and Kuwaiti spend all rising. There was news today that Iran hit a Kuwaiti airport. Nobody is spared from this new reality where you’re responsible for protecting your own borders.

When global spend goes from two to three trillion dollars, you don’t see 30% increases in the size of militaries — you see a refresh in technology, platforms, and capability. So there’s a unique moment: Israel is beginning to produce the supply of next-generation solutions just as the world’s demand for them is surging. If you understand that intersection between supply and demand, and you can build a team that sits right at it, you stand to make a real difference. It’s a generational opportunity to start a new platform.

The United States woke up to this a few years earlier and is now reaping the rewards — the likes of Anduril and Saronic are winning hundred-million- and multi-billion-dollar US government contracts and future-proofing America’s capacity to fight a future war, whether with China or otherwise. We looked at this moment — and Europe got there slightly before we did — and said: now is the time when Israel can and will produce its own Saronics, Shield AIs, Helsings, and Epiruses.

To do that, we had to build the team that could take advantage of the moment. It was very much about local excellence — access to the absolute best and brightest entrepreneurs at the earliest possible moments, often before they’ve even been released from military service — married to global access and reach. So we built meaningful international partnerships: with John Jansen, former head of procurement for the Marine Corps, and Ken Braithwaite, former Secretary of the US Navy.

We also decided we didn’t want to build regional companies with regional capabilities — we wanted to build companies that can play on a global stage. For that reason, we partnered with a sister fund, 8VC, out of Austin and San Francisco, which has scaled many of the American success stories. The founder of 8VC is also a co-founder of Palantir, and he brought on Palantir’s first employee to run their defense business. They incubated Saronic, backed Anduril, and sit on the cap tables of many American success stories. They want access to the Israeli market and have chosen to get it through their relationship with us, so we represent them here in Israel. They have, I believe, in excess of $8 billion under management.

So through local access, global reach, and the ability to form the geometry of capital needed to scale these businesses, we realized we had the right place, the right time, and the right team. We set to work raising the fund about a year ago. We’re well on our way — and that’s really the origin story of Kinetica.

09:10 — How Kinetica Assesses Founders

Louise Boucher: There’s a lot to unpack. Let’s stay close to Kinetica today. I didn’t realize the 8VC relationship was by design — so the answer is very much yes. Coming back to the strength of Israel for defense founders: when I compare it to what I look for here in Europe, you want teams with military or deep tech experience, and a sense of obsession and aggressivity. Everyone building in defense seems to have a certain intensity. So how do you assess teams in Israel? What do you look for?

Aaron Applbaum: This podcast is for publication, so I have to be careful about how deep I go into specific classified units and capabilities — but I’ll net it out as granularly as I can.

One of our partners in Kinetica is the former head of Israeli Special Forces. What’s unique about that role in Israel is that it has three major legs. There’s a kinetic leg — think the American Delta Force or the British SAS. There’s a human-intelligence unit — somewhere between the CIA, FBI, and ground reconnaissance. And there’s a technology unit that services the other two and the rest of the military.

That third pillar is our focus. It’s one of the most selective units in the Israeli military — the goal of almost anyone who doesn’t want to go into combat is to get into this unit, or into 8200, which has become more declassified and is more focused on signals intelligence and cyber. But this other unit, which my partner commanded, does a lot of the hardware innovation: electro-optics, acoustics, ASIC and circuit-board design, reverse-engineering of sophisticated hardware. The best of the best are recruited into it, and many stay — some become lifers. So we’re seeing these engineers before they hit the market. That’s the proprietary nature of our deal flow.

On diligence: at some level it’s alphabet soup — this unit, that unit. But we have access to every commander who led these engineers, every cadet who served under them, every unit they built for. So we understand their capability, their capacity, and how they act under pressure, tight timelines, and relentless customers — skills that translate beautifully into the commercial world.

We also have top-down capability. Another partner, Fred, has worked with the Israeli military-industrial complex his entire career — with Elbit, Rafael, Israel Aerospace Industries, and their European, Italian, and Middle Eastern equivalents. He deals in contracts worth hundreds of millions or billions of dollars, so he understands what a founder or CEO looks like when they have to execute at scale. It’s not just “can they build something” — it’s “can they move product?” Through his relationships, we also know the integrators, buyers, and customers. So we know the B and the G — the business and government buyers.

The two of them combined also have access to the top brass of the Israeli military itself. Understanding the local customer, the global customer, and the builder lets us triangulate a company’s capacity to succeed.

I also have a co-founder, Yoav, who’s an F-16 pilot. One of his roles after graduating Flight Academy was to sit on the panel that assessed and accepted candidates, and later evaluated those who became fighter pilots, navigators, and helicopter pilots. His ability to assess talent — especially around complex missions, technology, and high-pressure scenarios — is second to none. So that’s the 360-degree view we have before we ever pull the trigger on an investment. We can really dig into the bits and bytes of our founding teams.

15:39 — Where Israel Can Win — and Where It Can’t

Louise Boucher: Compared to 8VC, or a fund like Lightspeed — would you say the defense teams you’re vetting in Israel are different from the ones US or European funds are looking for? Different personality traits or experience?

Aaron Applbaum: It’s a really good question. Israel has a right to win globally in certain areas and not others. In some geographies there will be competitors who beat us, and there are certain things for which there’s no reason my team would beat the American equivalent.

Saronic is a great example. The founder, Dino, said America had lost its edge in shipbuilding — it had been outsourced. But once the American government aligned with Silicon Valley, and you add America’s capacity to manufacture at scale — a 350-million-person country, the largest economy in the world, with infinitely more space than Israel — they will buy and equip shipyards and spin up manufacturing that can build entire fleets. Israel is nine million people. It has some decent manufacturing — a bit on the engine side, a bit on the drone side — but Israel will never be a shipbuilding powerhouse. So against an equivalent in Japan or the United States, we’d lose. Israel won’t create a Saronic — or it could, but it wouldn’t win in the long term.

Now take a company we have backed, in the domain of information warfare — a foreign government’s ability to influence an election or sway a population. The classic example: recent polls in Taiwan show that around 30% of young voters would vote to become part of Greater China. China is showing it can take over Taiwan without firing a bullet. How? They own TikTok, young people are on TikTok, and they have the knowledge to manipulate algorithms — whether OpenAI, Anthropic, Reels, X, or TikTok — to sway public opinion.

We backed a company out of 8200 — mathematicians who reverse-engineer these algorithms, understanding every weight, bias, and threshold and how they change over time. If you look at what Israel has built in offensive cyber — and there’s a reason Paragon, NSO, Candiru, and others lead the field — it’s because there’s a critical mass of world-class talent here. It’s not that people out of the NSA can’t build it; it’s that if there’s going to be a Palantir-like product in this domain, Israel can build it. For each of my companies, I can justify why the Israeli version can be a global leader — and it always comes back to what our military is particularly good at.

There are certain functions that will always be localized — you have to sell locally, sometimes manufacture locally, and work with local prime contractors. But even with Israeli IP and Israeli management, you’ll always have a global footprint in our domains.

20:16 — The Autonomous Bet: Subsea Drones

Louise Boucher: Coming back to cyber specifically — I’ve been looking for a company to invest in there, but I always conclude that if someone pulls it off, it’ll be an Israeli company, or at least an Israeli founder in the US. Israel has an edge. And when I look at the hyped categories, it feels like the world is trying to catch up with what Israel has innovated over the last ten years — Iron Dome, intelligence networks, laser and directed-energy. But coming back to Saronic: no, you won’t become a global shipbuilding power, but we’re shifting to autonomous, scalable systems. Maybe Israel actually has an edge there — because with such a small population you can’t rely on huge ships, but you could build thousands of small surface vessels. So maybe you could be a generalist defense exporter after all.

Aaron Applbaum: I couldn’t agree more — you’ve hit the nail on the head. In fact, we’ve voted with our feet and our wallets on exactly that thesis. It’s a great counterpoint to what I said, and an interesting way of thinking about what Israel can contribute even if it won’t build global shipyards and beat Saronic at its own game.

We’ve invested in a globally leading team in the domain of subsea autonomy. Take a one-to-five-billion-dollar submarine — more if it’s nuclear, less if it’s diesel-electric — and ask: why can’t it be smaller and autonomous, but with similar capabilities? Thirty tons of payload, thousands of nautical miles of reach — just take the humans out of it. It takes a decade to design and deliver a submarine. Why can’t you augment what already exists, so your maintenance cycles are longer and you don’t have to use the manned platform as much? And why can’t allied countries that could never own and operate a submarine flotilla have subsea capabilities anyway? Panama has a pretty important canal to protect — why shouldn’t they be able to defend against mines and threats?

So we backed the former head of the Israeli submarine flotilla, a Brigadier General, who partnered with the former CTO of Habana Labs — a two-billion-dollar exit to Intel — and one of Israel’s foremost shipbuilders. The three of them had actually been in a submarine together years ago. One became a general, one made money in private-sector tech, and one went into industry. We brought the band back together to build the future of subsea warfare. The company is still in stealth, so I’ll respect that and not go into too much detail — but it had follow-on financing, and then follow-on to the follow-on, because everyone realized there’s no clear winner in the domain yet, especially around sophisticated sensors, communication protocols, autonomy, and using AI and data. You can’t throw humans at the problem, because Israel doesn’t have enough of them. We think it can be a global leader — a tens-or-hundreds-of-billions-of-dollars outcome.

24:22 — Building the “Impossible Team”

Louise Boucher: You mentioned you put the team together. In defense we look for almost impossible teams — a visionary CEO, someone who’s built the hardware, someone who knows the specific defense domain. Do you sometimes operate at the incubation stage, or is it more relationship-building — you know a CEO, you’re thinking about a CTO, and you put them together? Or is it something more scalable, like an incubator?

Aaron Applbaum: We have both muscles. We can assess an existing team with an existing idea and judge whether they have what it takes to win long-term — that “impossible team” where someone is a deep technologist who’s seen battle, understands scale, can manage R&D, speaks multiple languages, and has the charm to lobby members of Congress. Who are these people? They don’t really exist. We never had those expectations of the big cyber guys — they were nerdy, they sold to nerdy people, and they had big, wonderful, nerdy outcomes. So when we assess a team, we look for that balance: can they build and sell?

When we incubate, here’s a peek into the process. We keep a running list of ideas for companies we think can be large outcomes — everything from certain space technologies to subterranean to spectrum dominance. We also keep a running list of the best of the best, and we label them: CEO, CTO, R&D, “best first employee but not leadership,” “will always need to be told what to do,” and so on. The list grows, and we play matchmaker.

When you and I spoke briefly the other day, I’d just come out of a meeting with someone who is a generational talent — globally, you can count on one or two hands the people who can do what he can. And he would run a company into the ground. He knows it, I know it, everyone who meets him knows it. It’s about finding someone who can be the yin to his yang. We’ve started working with him on “founder dating,” and we think we’ve found someone who’s exceptional in his own right.

It’s a balance, and so far we feel we’ve gotten it right — because one of the hardest things is to correct it later. You tell yourself, “they don’t know how to raise money, but I can do that for them,” or “they can’t scale and motivate, but I can.” We can help, but your DNA on day zero is really important, as you said. We pay particular attention to it. It will undoubtedly be proven that we got some of it wrong — but we’re very happy with where we stand today.

27:58 — Israel as a Global Defense Exporter

Louise Boucher: I agree — the incubation stage is just great for reaching out, meeting amazing people, and partnering them with other amazing people. It’s about aligning all the planets: timeline, level of ambition, risk appetite. It doesn’t really scale, but it’s not meant to — it’s about doing the two, three, or four incubations that are truly relevant and create huge outcomes.

Coming back to Israel as a global defense exporter: it’s a bit of a paradox. Historically there have been great defense products from Israel, but on the global scene it was almost nowhere to be seen. Just before this, by coincidence, I was reading an article saying Israeli defense exports exploded last year, especially in Q1. So is the dam finally breaking? And looking out five years — will the global defense-export landscape change? We’ve also seen South Korea and Turkey exporting more and more. Do you think today’s top five will be replaced?

Aaron Applbaum: Great question. Yes — but not “yes” to the top five being replaced by a whole new top five. I don’t think the United States is going anywhere. That said, the fact that Korea is working on a sixth-generation strike fighter is news — that’s important.

I keep coming back to cybersecurity because it’s an area of Israeli excellence and dominance — Israel is responsible for, I think, in excess of 20% of global cyber spend. For a nine-million-person country, that’s a wild statistic. Yet it’s markedly absent from the “defense deck” — not every defense package has Israeli sensors, command-and-control, software, or effectors. That’s starting to change, especially in areas of Israeli dominance. We’re as good or better at aerial defense as we are at cyber — between David’s Sling, Iron Dome, and Arrow 2 and 3, with 4 and 5 now coming online. We’re hammering out the tenth iteration of our kinks before others have even figured out the initial problems. So there are certain areas where it would be irresponsible not to take the Israeli version.

Look at the recent India-Pakistan conflict: it wasn’t the French jets that brought it to an abrupt end — it was Israeli electronic warfare that ultimately saved India’s bacon. For decades of relative peace, having something local and readily available was good enough. Now you need best-of-breed. And in certain things — coming out of Rafael, Israel Aerospace, and in some cases Elbit, and from the entrepreneurs spinning out to start their own companies — Israel produces best-of-breed.

So I think Israel comes to the fore. You’re already seeing it in the backlog — tens of billions of dollars of it. You’ll also start to see it from other manufacturing powerhouses. Germany was obviously a military power, then very much not for a long time, and now you’re seeing a redirection: a lot of automotive companies have the press shops, R&D facilities, and chemical plants, and once they turn their attention back to defense, I think they’ll bubble back up into global leaders — not just in manufacturing, which they already are, but in defense.

Turkey is interesting — a large exporter, partly because they’ll sell to countries the US or Germany won’t. A lot of their kit doesn’t work as well as advertised, but they iterate quickly. Still, those who can buy the best won’t go with second or third best. So I think you’ll keep seeing America and its F-platforms — the F-16, F-22, F-35. And especially NATO countries trying to minimize reliance, because the US directive has been: start paying your own bills, you can’t rely on us to the same degree. So I think you’ll see some centers of excellence emerge. Japan could very well be on the same trajectory as Germany — a little behind, but they have a China problem.

Louise Boucher: France is technically the world’s second-biggest arms exporter, so this is a big question for us. We have traditional primes like Dassault with the Rafale, and they remain leaders. Can a new startup catch up with them — and can France keep its crown jewels in defense exports? We’ll see.

Aaron Applbaum: By the way — those were the planes the Indians were flying.

34:15 — Closing & a Book Recommendation

Louise Boucher: I know, that’s exactly why I’m thinking about it. I believe it was the first time Rafales were lost in combat. We might look back at this in five years as the moment the world shifted in warfare technology — the signal.

To wrap up: if someone wants to understand the Israeli defense ecosystem a bit more deeply, is there a book you’d recommend?

Aaron Applbaum: Yossi Cohen — the former head of the Mossad — wrote a really good one. It’s called The Sword of Freedom: Israel, Mossad, and the Secret War. He’s a friend.

Louise Boucher: Is it about Israel’s intelligence networks?

Aaron Applbaum: Correct. It won’t give you coverage of the broader military-industrial complex, but it’ll give you insight into the Israeli mindset — the nature of asymmetric warfare, a small country in a tough neighborhood, reliance on unfair competitive advantages, whether that’s robotics, autonomy, big data, resourcefulness, creativity, or a better understanding of supply chains. It’ll help you understand that context and how it spills into everything.

Louise Boucher: Thanks a lot for that — and thanks for your time, Aaron.

Aaron Applbaum: Thank you, Louise. This has been delightful. I hope there’s some usable stuff in there — it was a lot of fun.

Yet, over the past five years, three have managed it: Israel, South Korea and Turkey have each built multi-billion-dollar export businesses from a relatively low base. They did so by reading the same opening: the established suppliers can no longer meet global demand. America’s defence firms are stretched, their order books filled with replacements for weapons used in the Red Sea and Israel or donated to Ukraine. Europe is rearming quickly but its manufacturers need years to rebuild the capacity they lost after the Cold War. And Russia has turned inward; its arms exports fell by 64 percent between 2015-19 and 2020-24 as it supplied its own war and sanctions cut off key components, prompting long-standing customers such as India, Vietnam and Egypt to look elsewhere. Three different sellers moved into that space.

Turkey competes on availability and price. It offers equipment built to NATO standards, sold without political conditions, to governments that cannot or will not buy from Washington. The Bayraktar drones proved the approach, with more than 500 sold to over 30 countries after seeing combat in Libya, Ukraine and the Caucasus, and the product range has since widened to armoured vehicles, missiles, radars and warships. Exports have risen from roughly $2B to $8.5B in 5 years. The notable change is geographic: a supplier once focused on Africa and the Middle East is now signing deals inside Europe, including a $1.6B agreement with Spain to co-produce trainer jets and an invitation to join Germany’s Sky Shield air-defence programme.

South Korea competes on scale and speed. Seven large firms work closely with the government, source most of their components at home, and can therefore produce equipment quickly and at competitive prices, which is exactly what a depleted Western manufacturing base struggles to do. Its main market is Europe rather than Asia. Setting the United States aside, South Korea is now tied with France as the largest arms supplier to European NATO members, and it has more global orders than the US for tanks and artillery. The anchor is a 2022 package with Poland, now worth about $22B. Romania has since ordered around $1B of howitzers (a sort of cannon). Exports reached $15.4B in 2025, and South Korea is now developing its own fighter jet to move into higher-value sales.

Israel competes on combat experience, and from the highest base of the three. Its exports reached a record $19.2B in 2025, up 30% and the fifth consecutive annual record, led by missile and air-defence systems and by surveillance and optical equipment. The selling point is that the systems have been tested in war. Israeli equipment has operated under combat conditions since October 2023, and air defence in particular, including the Iron Dome and its longer-range interceptors, has performed under sustained attack in a way no competitor has. Europe was the largest destination, taking 36% of deals. Israel is selling the categories that current conflicts call for: drones, missiles and defences against large-scale aerial attacks.

What this means for the future of defence exports

The wider lesson is that a country can now build a competitive, export-grade defence industry fairly quickly, and that it is a strong business to be in. These are high-value products that carry years of maintenance, training and political relationships with them. Israel sells more than $19B of weapons while under diplomatic pressure, because the systems work. South Korea moved from a minor supplier to one of Europe’s leading sources in under five years.

That points to an ever more competitive market rather than a settled new order. Almost every region is trying to depend less on foreign, and particularly American, suppliers, and wants to build its local champion. The likely outcome is a more crowded field.

None of this displaces the United States, which will stay dominant for the foreseeable future. The more interesting question is what happens to the incumbents just beneath it, and France is the clearest test. It is still the world’s 2nd largest exporter, and its established primes are thriving: MBDA booked a record €14B of orders in 2024 and now holds a €37B backlog, while Dassault Aviation’s revenue rose 29% on the strength of Rafale exports, lifting its order book above €43B. But France’s public finances are stretched, which raises a real question about whether it can fund the next generation of suppliers as generously as it funded the last. The risk is that contracts keep flowing to the giants while smaller firms struggle to reach scale, and that the stamp “DGA-approved”, France’s defence-procurement agency, counts for less in a market that increasingly rewards speed and combat results over pedigree. A new cohort of defense innovators is emerging: Exail naval-drone contracts have pushed its order book past a billion euros, Harmattan reached a billion euros valuation, and both Alta Ares and Egide are able to poach top talent. Whether this generation grows into primes or is absorbed by them will say a great deal about who stays at the top of the table, and who slowly slips down it.

My own bet for the next breakout is Germany. It has the largest defence budget in Europe, a huge industrial base that converts readily to defence production, and a captive market next door. Eastern Europe and Poland are rearming against Russia at speed and would generally prefer to buy within the European Union than rely on Washington or Seoul. They already procure tanks from Germany, meaning trust and relationship is already there. Germany is well placed to export on a large scale across Europe, and once its carmakers fully turn their capacity toward defence it could grow faster than any of the three exporters described here. Japan, now undertaking its largest military build-up since the second world war and loosening its export rules, is on a similar path a step behind.

---

Funding and deals

IPOs

• Applied A&D (NYSE: AADX), which supplies subsystems like missile bodies and satellite solar arrays, raised ~$650M in its IPO for a $3.6B valuation.

• Quantum Systems and Destinus, two European drone makers, are lining up banks for potential IPOs as early as 2027, with Quantum eyeing a €10B+ valuation.

Acquisitions:

• D-Fend Solutions, an Israeli counter-drone company whose systems detect and take over rogue drones, was acquired by Motorola Solutions for $1.5B; deployed in 30+ countries, with ~$185M expected 2026 revenue.

• Bluewhite, an Israeli AI startup that retrofits vehicles into self-driving platforms, was acquired by Elbit Systems for an undisclosed sum.

Fundraises:

• Iceye, space surveillance, raised €1B at a €10B valuation, led by General Atlantic.

• Impulse Space, building highly manoeuvrable in-space spacecraft, raised a $500M Series D co-led by 137 Ventures and Banner at a $4.26B valuation.

• Mach Industries, developing unmanned systems and distributed manufacturing, raised a $300M Series C at a $1.8B valuation, led by Infinite Capital and Ribbit Capital. Ribbit is a legendary fintech investor but I guess neoprimes are the new neobanks.

• Isar Aerospace has closed a €270M Series D with Island Green Capital.

• Allen Control Systems, maker of Bullfrog, an autonomous gun that shoots down drones, raised a $200M Series B at a $2.2B valuation, led by Smash Capital.

• Airis Labs, turning raw field video into machine-readable intelligence, emerged from stealth with $60M, including a $31M Series B led by PSG Equity.

• Picogrid, building an open data layer that fuses battlefield sensors and systems, raised a $45M Series A led by Bessemer.

• Layup Parts, a software-defined factory for carbon-fibre parts, raised a $42M Series A led by Marlinspike.

• Volund Manufacturing, building low-cost gas turbine engines, raised a $12M seed led by Root Ventures.

• Shifters, developing AI-native autonomous ground robots, raised a $10.2M seed led by Ace Capital Partners.

Philipp Kerth, co-founder and CEO of Hypersonica, sat with me to talk about what it takes to launch a hypersonic missile company at the end of a PhD, when defense was still an afterthought in most of Europe.

We cover why maneuverability is the hardest problem in hypersonics, what the new generation of hypersonic systems actually changes on the battlefield, and why building an entirely European supply chain - no US components - is as much a strategic choice as a technical one.

Transcript (lightly edited for clarity and flow) is below.

00:00 — The Founding Story

Louise Boucher: Before we dive into Hypersonica, I want to come back to your decision to launch this company. I met you more than a year ago, in January, and visited your factory near Munich — which is very impressive. But just to give some context: you grew up in Germany, went to Oxford for your PhD in hypersonics, and when you came back to Munich you built a missile company in 2023. Could you walk me through that decision? What was the moment you knew you were going to do this?

Philipp Kerth: There isn’t a single moment, so I’ll have to walk you through a longer story. For founding Hypersonica, three things came together. The first was technical fascination. The second was a very deep understanding of the values we have in Europe. And the third was really wanting to do something — wanting to have an impact and bring new things into the world.

The values piece is the most important one for me as the intellectual backbone of the company. I’ve always felt a deep appreciation for being able to grow up in a Europe that’s free, democratic, that has friendship between nations, that is safe and wealthy. The biggest contribution I could see myself make — coming out of a PhD in hypersonics — was to contribute critical technologies that will enable us to uphold those values in the future.

The technical fascination started very early. There’s a great science museum in Munich and my dad used to take me there as a child. I saw all those space capsules and spacecraft and from that point on I was lost — in a very positive sense. Hypersonics, making things fly really fast in a controlled way, combines physics, fluid mechanics, electronics, and materials. It’s a great topic.

The third thing is wanting to contribute something substantial and practical. When I studied at TU Munich, a couple of startups that came out of the university started to really take off. That inspired me to look deeper into entrepreneurship and see how big a difference you can make. All three things combined, and Mark and I decided to kick off the company.

04:47 — Hypersonics as Deterrence

Louise Boucher: You launched at a point where defense was less of an obvious topic for the general population. Did you experience pushback from friends or family?

Philipp Kerth: We started the company at a point where it wasn’t as obvious that hypersonics and deterrence would become such a big topic. But the concept of values that stands behind it was always clear in every conversation — that is what’s behind the company and essential to our DNA. Hypersonics is a possibility of having a shortcut to some very accurate and sovereign deterrence for Europe. That is exactly what we wanted to bring.

Louise Boucher: You mentioned your co-founder Marc. You met at Oxford with similar academic backgrounds, working on one of the hardest engineering problems in the world. Do you ever disagree?

Philipp Kerth: Marc was one year ahead of me in the PhD program. When he finished, he went to Bain for a year as a consultant, and then we founded the company together. We have a really good culture of disagreeing with each other in a very constructive and meaningful way. Speaking for both of us — we nailed that part. The results speak for themselves.

07:16 — What Iran and Russia Taught Us

Louise Boucher: Before we go into the technology, I want to get into the threat context. When we look at Iran and Russia, how they’ve used ballistics and hypersonics — are there lessons to draw for the future of warfare?

Philipp Kerth: There are lessons. I’ll stay careful on judging the specific capabilities of other systems, but I can give an overview of what hypersonic actually means. In aerospace, hypersonics simply means flying faster than five times the speed of sound — roughly 6,000 kilometers per hour. The reason there is a separate field is because temperature effects start to dominate the fluid mechanics at those speeds.

The new generation of systems combines hypersonic speeds with precise maneuverability. That’s where the real differentiator is. The Russian and Iranian systems are more on the less maneuverable side, but they do achieve hypersonic speeds. What we’ve seen overarchingly is that being able to strike very deep into enemy territory is an absolute essential capability. Multiple nations across Europe have published their demand for these systems.

Looking specifically at Iranian systems, they’ve been able to challenge air defense architectures with relatively high cost efficiency. Without making claims about their capabilities as singular weapon systems, we can infer that it is essential to drive down the cost of these developments. That is exactly what Hypersonica is after — bringing this core capability in a sovereign and cost-effective way.

09:56 — Why Maneuverability Is the Hardest Problem

Louise Boucher: You’ve talked publicly about maneuverability being harder than propulsion. Why is it so important, and why is it so hard?

Philipp Kerth: The importance of maneuverability comes from evading interception and unlocking the full capability of a hypersonic system. Let me differentiate from a conventional ballistic missile. Ballistic missiles have existed for decades — since the 1940s and 50s, we’ve been accelerating things to hypersonic speeds. But a ballistic missile is propelled by its booster, climbs on a very high trajectory out of the atmosphere, flies through space, and then falls down onto its target. During re-entry, it reaches hypersonic speeds again. But this trajectory is very predefined — you can’t really modulate it, you can’t really maneuver.

The second issue with a ballistic trajectory is that it goes very high. It’s well above the horizon, so you can detect it early on radar. Because it can’t change its path, a defender has quite a lot of time to put countermeasures in place.

With the new generation of hypersonic missiles, you eliminate both of those disadvantages. Whether it’s a hypersonic cruise missile or a hypersonic glide vehicle, both are accelerated by a solid rocket booster and then separate to start their hypersonic flight in the high atmosphere — a lot lower than a ballistic missile. They come over the horizon much later, and while they fly at hypersonic speeds they remain fully maneuverable. They can change their path, change their target, execute evasion maneuvers. That’s what really causes a step change in strike technology. It makes those systems very survivable — a very high chance of actually hitting the target, and therefore a very efficient way of executing what’s called a deep precision strike.

As for why it’s hard: the hypersonic speed regime is fascinating but very challenging for engineers. The most important issue is temperature. When flying at those speeds, everything gets very hot. You need to deal with that across your entire engineering stack, and you need very high precision, which at those speeds is a real challenge.

Our approach was to rapidly go from concept to test — to compress decision timelines on engineering questions, get data as early as possible, and move forward quickly. We demonstrated that with our first hypersonic test flight, and that is how we are getting from scratch to this capability, tackling all the engineering challenges for maneuverability step by step.

14:20 — Nine Months from Design to Launchpad

Louise Boucher: The test flight was only nine months from design to launchpad. How do you build such a fast-paced culture?

Philipp Kerth: Three key ingredients. First, the right people. Hypersonics has been a niche field for a while. In the 90s, there was still quite a lot of hypersonics research associated with the Space Shuttle program and with the European equivalent — the Hermes glider project. When those were cancelled, hypersonics died down quite a bit. There are very few people actually proficient in making things fly at hypersonic speeds. Marc and I, coming from this research background, became the gravitational center for hypersonics talent and expertise. That is key to developing this technology at pace.

Second, we mixed that hypersonics talent with people who are very proficient at rapidly iterating technology. We looked at where technology has been developed faster and cheaper in highly regulated environments. The new space sector is the clearest example over the last 20 years — timelines and costs got compressed, with SpaceX leading. We brought people from that sector to rapidly iterate from scratch to prototype.

Third, a purpose-built company. You build the company around the technology of hypersonics and around the mission of developing strike systems as rapidly as possible. You end up with really efficient structures focused entirely on this hardware-rich development. That is what allowed us to go this quickly.

16:33 — Building ITAR-Free

Louise Boucher: There’s a sovereignty dimension to how you build, not just what you build. You’ve built an entirely ITAR-free supply chain — no US-controlled components or IP. Was that a deliberate choice from day one?

Philipp Kerth: We’re not trying to replicate what was done in the past or legacy ways of developing these technologies. We’re fundamentally starting on a different operating model and starting with the given reality that we have right now. That reality mandates doing things in an ITAR-free and sovereign way. Very fundamentally, we don’t have 10 years or 10 billion to develop these systems, and we need to do it with a sovereign supply chain.

You need to be able to judge what components work at hypersonic speeds and what don’t — going from first principles, away from any legacy limitations. What enables us to do that is our backbone of hypersonics engineering expertise. For our first test vehicle, all components were sourced ITAR-free, and we’ve been successfully validating that further. In Europe, we do find the skills we need. A lot of research has been done in hypersonics here. There are components we can use from the ITAR-free market, but there are also components we have to pull in-house and redevelop from scratch because nothing on the market will work for this specific hypersonic application. It’s really about doing that back and forth between trade-offs — with ITAR-free as a hard constraint.

Louise Boucher: You’re often compared to Castelion, a US company working on hypersonic and supersonic missiles that’s been growing very fast. Do you even think about competition?

Philipp Kerth: The US startup ecosystem has produced some really impressive companies in the defense space, and Castelion is definitely one of them. The pace at which they’ve developed and tested is impressive and we congratulate them on their success. What is fundamental for us is that we’re a company built around hypersonics and around European sovereignty, European demand, and European requirements. I don’t see the necessity to draw a holistic comparison. We focus on what Hypersonica does.

20:08 — Anglo-German by Design

Louise Boucher: You’re headquartered in Munich, but you’re also a supplier to the UK’s national hypersonic program. How do you navigate the two procurement cultures?

Philipp Kerth: Hypersonica was really built around binational cooperation from the start. Marc and I met at Oxford. Marc spent even more time in the UK than I did — I did four years, he did eight, including his undergraduate studies in London. We were both trained in hypersonics in the UK and are really committed to contributing to that ecosystem.

Between the UK and Germany, there is the Trinity House Agreement, which is a defense cooperation framework that includes deep precision strike as one of its pillars — and hypersonics can be part of that. Very fundamentally, we’re looking to contribute to the ambitions of both countries within that framework.

In terms of what each brings: in the UK, there is hypersonics talent from our group and a national framework that committed several years ago to developing hypersonic strike capability. In Germany, there is strong institutional expertise and a hardware ecosystem in Munich that has given us a great platform to grow. Seeing those two countries develop close ties through Trinity House is particularly helpful for us.

Louise Boucher: Your website also mentions civilian applications — space reentry, faster travel. Is that a genuine long-term vision, or is it more of a dual-use framing?

Philipp Kerth: Hypersonics is absolutely a dual-use technology. Everything we learn about flying faster than Mach 5 in a controlled way contributes well beyond weapons development. It influences space reentry — when capsules come back from orbit and glow as they re-enter the atmosphere, that is exactly the hypersonic speed regime. Anything we develop for better control and heat management can make those reentries more efficient and reduce thermal protection requirements per payload. And eventually, yes, it can contribute to concepts of sustained hypersonic aviation. But those are a bit in the future. We would not pursue aviation concepts in the near term.

We simply see hypersonics as a core technology. Moving at beyond Mach 5 in a maneuverable and precisely controlled way is going to influence a lot of different things — starting with defense, and then certainly spreading into other areas.

24:53 — What Hypersonica Looks Like in Five Years

Louise Boucher: What does Hypersonica look like in four or five years?

Philipp Kerth: The ambition is to be one company concerned with everything that flies faster than five times the speed of sound. The starting point is this: Europe might be behind on a couple of core technologies, but let’s make hypersonics ours. Let’s make this the core technology that first solves our deterrence problem — or at least a large chunk of it — and then moves forward to lead in other areas: space, and maybe at some point aviation. That is the ambition for the company over the next five years.

Louise Boucher: If you’re listening and you’re a great engineer — Hypersonica is hiring. Thanks a lot, Philipp.

Philipp Kerth: Thank you. Awesome.

That is not what the factories look like.

A recent trip to Ukraine makes that clear. The shops producing land robots (“UGVs”) drones are small, often fewer than twenty people, assembling frames and soldering components by hand. Still, output per shop runs around 300 units a month without automation. This is not an isolate case and shows that scale can come from the design more than the assembly process.

Scale first, automate later

Anduril’s Arsenal-1 in Ohio is the most instructive example. When Fury (the unmanned combat aerial vehicle) production began in March 2026, the line had 22 workstations and no robots. The production floor, as one journalist who visited described it, resembled an Ikea more than a defense facility: bare bones stations, no large permanent tooling structures, nothing bolted to the floor that couldn’t be moved if the line needed reconfiguring.

At Arsenal-1, John Malone, Anduril’s head of production, came from Tesla, where he watched the Model 3 ramp nearly collapse under premature automation - cars ended up being assembled in a tent in a parking lot. His posture at Arsenal-1 is the inverse: run the line by hand first, and let it show you where automation is actually needed.

“When you go through early stage production, you’ll see what you should automate. It kind of presents itself to you,” John Malone, Anduril’s head of production.

Visiting Harmattan’s new factory last week, the same logic came up unprompted: production remains mostly manual today because designs are still changing too fast to justify locking in automated tooling. The moment you automate a specific assembly step, you’ve made it rigid. In a product still evolving, rigidity costs more than it saves.

Saronic embedded the same principle before building anything. Their head of manufacturing, recruited from SpaceX, was their third hire and joined before a single prototype existed. The reasoning: if production isn’t part of the design process from the start, whatever you build is only guaranteed not to scale. Their vessel platforms have 80% hardware consistency across the range and roughly seven core components per ship. The result is cheaper and more scalable products: their 24-foot Corsair costs under $1 million and can be produced at over 2,000 units per year.

Scale is a design decision

The reason human-heavy lines work at volume is that simplicity was engineered in upstream. Anduril chose abundant aluminum over rare titanium for the Fury airframe, standardized on commercial off-the-shelf components, and eliminated castings and forgings. Nearly 90% of Arsenal-1’s components are commercially sourced. The goal was an aircraft that automotive assembly workers, not aerospace specialists, could build consistently — a workforce decision made possible by a design decision.

Harmattan operates on the same principle. The ISR drone has four screws. The company minimized fasteners across the product line on the basis that every additional screw is assembly time and a potential failure point at scale. 90% of parts are shared across products, meaning a single production ramp covers the full portfolio. Like Anduril, Harmattan recruited deliberately from the automotive industry — people who think with a mass production mindset day one. The factory went from site identification to opening in 5 months, against an industry norm of 15, by doing almost everything in-house (as the design of the factory plan) and maintaining direct relationships with every supplier.

The Ukrainian shops I visited operate under a more extreme version of the same constraints. Output runs around 300 units a month per shop, assembled entirely by hand. But there is an additional variable that does not apply in the Western world: the factories move. Ukrainian drone producers operate across dispersed, constantly rotating sites because fixed facilities are targets for Shahed strikes. The production model has to be light enough to relocate on short notice - no heavy tooling, no fixed infrastructure, nothing that cannot be packed and moved. The reason it works at volume under those conditions is the same reason the model has to be mobile: the design is simple enough that humans can assemble it anywhere.

The automation question, in other words, is largely answered before anyone steps onto the factory floor. Assembly can stay human-heavy and still reach volume, provided the design decisions have been made correctly upstream.

Where software has the most to add

The assembly floor, in other words, is not where the industrial base bottleneck lives. The constraint is one level down, in the precision component shops that supply the assembly lines.

This $1.8 trillion industry is powered by small businesses whose owners are on average over 65, with 40% planning to retire within five years. That is where automation and software have the clearest value to add.

Hadrian, based in California, is building software-defined factories to address that gap. Its platform automates CNC tool path generation, quality inspection, scheduling, and adaptive programming, adjusting parameters in real time. A production cell goes from nothing to operational in under six months. New technicians can be trained in 30 days. As experienced machinists retire, the platform absorbs their process knowledge into the software layer. The company raised $260 million in 2025 at a $1.6 billion valuation with Founders Fund.

Isembard, started out from the UK, is working the same problem through a franchise model. It licenses MasonOS — an agentic AI platform integrating quoting, scheduling, supply chain management, manufacturing execution, and quality control — to existing machine shop operators and new entrants. The bet is that physical capacity already exists across thousands of small shops but what those shops lack is the software layer to run at defense-grade speed and consistency, which Isembard provides. Isembard raised $50 million recently with USV and is targeting 25 factories by end of year, spread across both the US and Europe. They’ve already signed an impressive list of clients, from Anduril, Tekever or Alloyed.

The two models differ — centralized automated factories versus distributed software franchises — but the diagnosis is the same. The factories that will determine whether saturation strategies are executable are not the ones producing the final product. They are the ones producing the parts.

Fundraisings:

• Hawkeye 360, a U.S. commercial satellite company that tracks radio-frequency signals from orbit, raised $416M in its IPO for a $2.4B valuation; shares rose 30% on the first day of trading.

• Anduril, a pioneer in autonomous military systems, raised $5B in a Series H at a $61B valuation, led by Thrive Capital and Andreessen Horowitz.

• Armada, building edge computing for military operations, raised $230M at $2B Valuation, led by Overmatch.

• Amca, a roll-up of critical component suppliers, raised a $300m Series B at a $1B valuation led by Caffeinated.

• HavocAI, the multi-domain robotics company known for its USVs, raised $100M in a Series A led by Cobalt Capital.

• SWEBAL, rebuilding TNT factories out of Sweden, raised €30M in a round backed by Sweden’s former defense minister Karl Engelbrektson, ex-EQT CEO Thomas von Koch, and several Swedish family offices.

• Darkhive, a tactical autonomy and edge software company built around small UAS, raised $30M in a Series B led by RTX Ventures.

• Heaviside, building autonomous precision munitions, raised $28M in a Series A led by Interlagos Capital.

• Anello Photonics, a silicon photonic optical gyroscope company, raised $25M in a Series B-2 led by MESH. .

I sat down with Doug Beck, former director of the Pentagon’s Defense Innovation Unit (DIU). He steered billions of dollars into turning Anduril and Palantir from startups into primes, and driving the Replicator initiative - the first real signal of the shift of procurement toward autonomous, expendable systems.

Now, his work continues, shaping defense in Japan and in Europe.

Transcript (lightly edited for clarity and flow) is below.

00:00 — Two parallel lives

Louise: I’m very happy to have you here.

Doug: Great, well I’m super excited to be here, Louise. Thanks so much for having me.

Louise: There’s so much to say and tell about your life. You were former director of DIU, and I want to come back to that. But before that — your path here is genuinely unusual. You served in Iraq and Afghanistan. You spent fourteen years at Apple reporting directly to Tim Cook. And somewhere in between, you became one of the most important people in American defense innovation. What’s the thread that connects all of that for you?

Doug: I actually started out as somebody focused on international relations and foreign policy. That’s how I ended up going into the private sector in the first place. I worked between college and grad school for the Office of the U.S. Trade Representative, and at the time I came away 100% convinced that the way global corporations, particularly in technology, were evolving was going to be a major determinant of international relations for the future. This was way back in the first couple of years after the end of the Cold War — that’s how old I am, sadly. And unfortunately, most policymakers didn’t really understand how global corporations made decisions, and as a result they made policy that backfired.

So I wrote my thesis at Oxford on — effectively, although we didn’t use these words at the time — global value chains and supply chains in technology, and US–Japan trade relations. The point being that we needed to get a lot more sophisticated.

After that I said: well, I’m going to be somebody who actually understands this. So I went to McKinsey, thinking I’d be there for just a couple of years. I quickly realized there was so much to learn and I was growing so fast in that environment that I thought I might be a more useful person in this debate eventually if I stayed on that track. But I didn’t want to wait that long to serve, because I’ve always been driven a hundred percent by purpose of service. So I took time out from McKinsey, went to officer training school, and ended up getting a commission. That’s when my sort of two-lives-at-the-same-time life began. That was nearly 30 years ago. I’ve lived simultaneous parallel lives ever since.

I spent a bunch of time in Asia. I lived in China for three years. I did a whole lot of things with my private sector life while that military career was evolving. And then I did end up going on active duty with a joint special operations task force — first in Iraq, then a couple of tours in Afghanistan. Because I lived in both these worlds, I ended up in a position to help build a bridge between them. And that became an increasing passion of mine.

04:00 — The merry cabal: how DIU was born

Louise: Now, of course, innovation in defense is more obvious. We’re in 2026, the Ukraine war has been raging for four years. Iran, Pakistan, India — all these conflicts, all the same things. Warfare has changed: technology, autonomy, expendable systems are everywhere. But when you helped create DIU back in 2014, how did you build that conviction that it was the right time, and that something needed to be done now?

Doug: Around 2014, I had the opportunity to be part of what I call the merry cabal of people advising then–Vice Chairman of the Joint Chiefs, Admiral Sandy Winnefeld, and then–Secretary Ash Carter on the idea of building a bridge to Silicon Valley from national security — and also helping to wake Silicon Valley, more broadly, to the need to support national security. That’s what ultimately led to the original creation of DIUx.

Secretary Ash Carter was prescient in seeing that there was no way we could meet the strategic imperative facing the United States or our allies and partners — particularly with respect to the People’s Republic of China — unless we could harness the incredible power of the commercial tech sector.

Even back in 2014–15, innovation in critical areas for national security — AI, autonomy, energy, biotech, space, cyber, telecom — was going to happen faster in meeting the demands of billions of consumers around the world than it ever could in the purely defense-specific pathways that characterized our more traditional defense sector, whether in the labs or in the defense primes. We needed to harness that alongside those traditional pathways.

Silicon Valley had actually been born out of collaboration between the public sector, private sector and academia in support of national security — first during World War II, then during the Cold War. But we had completely lost that connection. By the time we were all talking about this, your average person in the Pentagon thought that when they wanted some software, they should pick up the phone and call one of the traditional defense primes. Meanwhile, the people I was hanging out with in Silicon Valley mostly figured that you kind of unscrewed your cap and the free and open international system flowed out of it — you could go start your company, sell it, do whatever, and then go mountain biking and not have to worry about it. Both those things were wrong. So we really had to build a bridge.

The first iteration of DIUx back then was really just about building a bridge at all. It was almost like an embassy. I was first a civilian advisor, and then once we got the unit started from inception, I had command of the Joint Reserve Component, so I was part of the team there. Pretty quickly, DIU — once the X got dropped — evolved into what we call DIU 2.0, which really proved you could take real military problems and solve them with commercially derived tech, using authorities we already had on the books but didn’t know how to use. They’d actually been developed for the Apollo program decades before. And then get prototypes into the hands of the warfighter in weeks or months rather than five or ten or fifteen years.

That was heresy when it first started. DIU under my predecessors’ leadership — first Raj Shah and Mike Brown — did an admirable job, and did it lots of times. That was fantastic, but it wasn’t good enough yet. To meet the strategic imperative Ash Carter had been talking about, and that DIUx was established to do in the first place, you really needed to take that capability and apply it for strategic effect. That meant first focusing on the most critical operational capability gaps we must solve to deter major conflict or win if forced to fight a war — and then scaling those solutions so they could deliver that strategic effect.

That meant a completely different relationship with the rest of the department, different funding, all kinds of different things. That’s when Secretary Austin decided to elevate DIU to be a direct report. And that’s when I left my perfectly good, happy life at Apple to go full time and work on that.

Congress came in quickly behind the DIU 3.0 strategy we created — first, by making permanent DIU’s role reporting to the Secretary as what’s called a principal staff assistant. It’s like an under-secretary, but a non-partisan one. And then with a budget about 10x as big — about a billion dollars — but more importantly with really unique flexibility in the way the budget worked. It allowed DIU to move money quickly from idea to idea, from company to company if necessary, much more flexibly than traditional programs of record. It also let us partner very differently with the services — DIU could fund all the way up to initial operational fielding of technology to operating units. Which meant for the first time you could go from idea to prototype to initial fielding in one year, as long as the tech was there. And then do the scaling in the years that came after.

That’s basically what we’ve been up to for almost three years before I left the department seven or eight months ago — and what is now scaling under my successor’s leadership.

11:00 — Anduril, Saronic, and the startup-to-prime playbook

Louise: That’s very impressive. The book Unit X does a very good job of explaining what DIU did for defense innovation, with OTA and so on. But could you explain — companies like Anduril, Palantir, Shield AI went from promising startups to defining what American defense looks like, under your governance. Can you talk about what DIU did, what you did, to bring these companies from startups to the new primes they are today?

Doug: Yeah. Anduril is a great example. In the very early days they got a boost from In-Q-Tel, which focuses on early-stage. Once they had a capability — in this case counter-UAS — that was ready for deployment, DIU worked really closely with them and with Special Operations Command. That ultimately led to the deployment of their counter-UAS capability. Then SOCOM put a scaling contract in place — around a billion-dollar IDIQ — to help them really scale. That’s one of the things that helped get them off the ground. And there are multiple examples, hardware and software across the board, with Anduril then applying to all kinds of DIU programs and successfully having some of those go on to scale with the services.

Maybe I’ll give a more recent example. A company called Saronic makes small unmanned surface vessels. Saronic applied for one of DIU’s CSOs, was successful along with a couple of other companies for a prototype. That solicitation was put out in February 2024. Many, many companies applied. It was down-selected to about 20, and then three companies were awarded.

Saronic was one of them. The prototype worked. They got a success memo by December of that same year. DIU was able to use that flexible funding I talked about to buy the first operating units to provide to a new spotter created by the Navy for autonomous warfare. Then in mid-2025, the Navy followed up with a nearly $400 million scaling contract — which is now off to the races and live. Saronic, alongside that, has been able to do substantial raises that have allowed them to expand capacity, meet demand, and begin looking at other things.

That’s really how it’s supposed to work — a great example of what DIU 3.0 was looking to make happen. A capability that’s a clear strategic need from the warfighter, aligned with scaling intent from the services, great commercially derived technology, a venture-backed company that we’re able to help get to truly scaling — and most importantly, putting real capability into the hands of the warfighter. There are now lots of examples of that, and the examples are expanding every day.

16:00 — Ten times further. One-tenth of the way.

Louise: Now the rest is history. It’s crazy, because back in the day Replicator was a big bet — fielding thousands of attritable, expendable autonomous systems very quickly. Now the new administration seems to be doubling, even tripling down on the thesis you laid out. Are you happy about what’s being done today on defense innovation in the US?

Doug: Yeah, I’d say the answer is yes. My view is we are more than 10 times further than where I hoped we’d be when I came into office — and we are probably one-tenth of the way to where we need to go. So I’m very happy about the extent of change and the expansion of that.

I’m still concerned that too much of the change remains reliant on waivers, special processes, workarounds, or individual personalities driving it. Even with all the progress, I’m keen now to see more institutionalization of that capability.

To me that means a few things. One: we’ve got to keep reinforcing the capability of organizations like DIU to keep pushing the vanguard of change, redefining what it means to move fast, what it means to have the whole department work together — which is really what Replicator was all about. Replicator was partly about putting capability in place quickly, but even more than that, about building the muscle as a department to work together — taking steps that ordinarily happen in serial and doing them in parallel, so you can get much more done much more quickly, and actually achieve anything close to the kinds of objectives Deputy Secretary Kath Hicks had laid out for the timeframe.

Two: I’d love to see the flexibility in budget DIU was granted provided to the services — which are the real scaling entity of the department — alongside the same kind of transparency Congress required from DIU. Congress gave us incredible flexibility, which was really an expression of trust. But at the same time, they required that in order for us to execute on it, we had to provide a completely unprecedented level of transparency. I literally shared the spreadsheet we used to run DIU. Every single project, exactly where it was, exactly what we were spending on it, what was going well, what wasn’t going well — good, bad and ugly. There were a few projects where contracts hadn’t been awarded yet, so we couldn’t provide specific company names. But short of that, I shared the same spreadsheet with the appropriators that we used to run the business. That was necessary, I think, to earn that trust. I’d love to see that same flexibility and transparency in much more of the department.

Three: we need a dramatic increase in how much of this new capability and new way of thinking is being used for large-scale platforms. I love, for example, the move the Navy recently made to have Hadrian involved in advanced manufacturing for submarine production, which we need to dramatically accelerate. But that’s just the beginning of really changing how we think about large-scale programs.

Four: we need to really expand the way we generate and incorporate talent that is dual-fluency. Everything at DIU rests on this group of people who speak fluent warfighterese and fluent tech-sectorese. We just don’t have enough of them. We need to make it much easier to generate those people inside the department, and then keep them and promote them. We also have to make it easier for people from the outside who want to serve to come in. We have all kinds of people willing to take a 99% pay cut, in some cases, to come work for the department. But they’re not willing to sit around for a year and a half waiting to see if they’ll get the job, going through the incredibly slow hiring processes typical of the department. We’ve got to change that.

And last — we’ve really got to change the incentives for people in the procurement and scaling zone. There have been moves toward this under the current administration, but it’s got to go to a whole other level. Today it’s still culturally very hard, even with all the progress, for people in a lot of those jobs to take the risks we need them to take. The risk to their career is much greater if they try to go fast, use OTA and make a mistake, than if they go slow, run a very complex process, check every box — and then five years later some Marine doesn’t have a capability and we end up losing lives. Nobody’s going to go back five years and say: hey, who slowed that program down? We’ve got to change those incentives too.

So as you can hear, I’m incredibly excited about where we are. And I also feel like there’s still so much more to do.

23:00 — Japan, DISTY, and the international DIU network

Louise: What’s impressive is that your work is never done — you’re now taking your ambition and mindset to Asia. I wanted to come back to the importance of international partnerships, and to your work as an advisor to the Ministry of Defence in Japan. It’s an interesting country: they went from near-zero defense ambition to building their own version of DIU and a large military. Are they building a new military the same way we are in the US, or in Europe today — autonomous, expendable, scalable systems first?

Doug: First of all, you’re right — international partnership was a really critical part of what we did under DIU 3.0. One of the five main lines of effort was international partnerships. It was all about partnering deeply with allies and partners that already had an organization like DIU, helping those like the Japanese who didn’t have one to create one — which is how DISTY, the organization they created, was built. We actually worked with them from the beginning on that, then partnered with them deeply, and where possible helped knit these different organizations together to facilitate the network overall. We worked closely with the British, the Australians, the Japanese, the Singaporeans, partners across NATO and obviously the Ukrainians.

Speaking of Japan — they are in the middle right now of a really significant change in the way they think about national defense. It began under Prime Minister Abe and continues to accelerate, including under the current prime minister. They’re looking to play a much greater role in their own defense and in the defense of the system. They’ve got a long way to go in modernizing their system to be able to do that. They have incredible technology in large companies, increasing numbers of startups — although the VC sector in Japan is still not super well developed compared to the US or even Europe.

But they also have incredibly powerful companies at the medium-size, small-and-medium growth stage. A lot of the companies that make the technologies inside your iPhone — just thinking of a place I used to work — were Japanese companies, because they were the very, very best in areas like sensors. So the Japanese have huge strengths, but they’ve also got significant changes to make in the way they bring capability aboard and think about their own capacity. They’re obviously looking at areas like autonomy.

33:00 — Europe, sovereignty, and the free world

Louise: I firmly agree with you that in the end, it’s still the free world against autocracies, and that’s something worth fighting for. Even though it does feel we’re entering into a new world order, a multipolar one. Sometimes people forget that we are a free world and we should defend it.

Doug: The answer is there has to be — there must be. We have enjoyed 80 years of relative peace and prosperity, even with all the challenges we’ve faced. That has allowed the dramatic expansion of standard of living, the total explosion of things like the technology we’re talking about today. And that has been driven by the creation and maintenance of a free and open international system on which all of that peace and prosperity has relied.

So I’m unashamedly — both as a global citizen and as an American — a deep believer in the importance of that, and the importance of the United States playing a central role in both creating and securing that world.

Now, it is absolutely true that our allies and partners around the world both can and must do much more to support that alongside the United States. I think it’s great that Europe is now moving to do that, that Japan is now moving to do that, that countries around the world are moving to do that. And the massive pressure President Trump has placed on all those countries by shaking the tree as hard as he has has an enormous and important benefit.

On the other hand — if in the resulting evolution we end up shaking trust so deeply that we undermine the bonds that are the very things we are trying to secure and protect, then the negative impact on peace and prosperity will be so great it will actually be much worse than if we hadn’t started. So it’s absolutely imperative we all find a way to preserve the core of that trust, even as we change the dynamic through which we secure it. I know that’s hard to hear from a lot of countries, given some of the things that have happened over the last year. But I really think it’s in all our fundamental interest.

To answer your more specific question — all our countries together need to think about what we can truly share, and what we need to make sure we have independent sovereign access to or control of, even as we’re sharing. If we end up in a world where everybody develops their own version of the exact same thing, we will lose so much in terms of capacity, capability, and interoperability that we’ll put ourselves in a lousy place.

On the other hand, from a European standpoint, I 100% understand that Europe needs to be able to act when it’s in Europe’s interest, even if the United States isn’t fully on board. So managing that balance is going to take a lot of nuance.

The one thing I’d stress — well, two things. One: the extent to which countries can do that across Europe, rather than each individual country chasing its own parochial version — which too often winds up being captured by domestic political and economic interests. If Europe goes that route, Europe is going to have a really hard time generating the capability it needs. Two: even as Europe ensures European countries have autonomy and sovereignty in certain areas, we need to do it in a way that maintains full interoperability. If we don’t have interoperability, we’re not going to be able to do what we need to do together to defend against autocracies — which is the real challenge we should all be focused on.

What’s next

Louise: On a more personal note — looking at the next five years, where are you going to spend most of your time? Mostly Asia and Japan? Rearming Europe? Doubling down on your efforts in the US?

Doug: Right now, I’m spending a lot of time — I’m not sure I can predict the next five years, because I’m not sure any of the previous chunks of five years of my life and career have been great predictors of the next one. But right now, I’m very focused on doing what I can to help at the intersection of security and our allies and partners — going right back to what we were just talking about.

I’m spending a fair amount of time doing whatever I can to help the heroic efforts in Ukraine. I want to be able to help in that zone — and to do it in a way that is 100% in line not only with the legal requirements just put in place by the US government, but also with what I believe is fundamentally in the interests and objectives of the United States as well as our allies and partners around the world.

Doug: Thank you, Louise — and thanks for the work you’re doing in this space. It’s super important.

Three things stand out.

The first is the emergence of genuine champions. Swarmer (SWMR) went public on the Nasdaq in April — the first Ukrainian defense IPO. UFORCE (unmanned systems, best known for maritime drones), General Cherry (FPV and interceptor drones), and Airlogix (reconnaissance and autonomous strike UAVs) have all crossed unicorn territory. These are not wartime anomalies propped up by emergency procurement. These are businesses with real revenue, real customers, and capital markets now paying attention.

The second is a meaningful recovery of supply chain sovereignty. A year ago, DJI was omnipresent and unchallengeable — Ukrainian teams told me openly that nothing else came close. That dominance is eroding. Local manufacturers are filling gaps that only Chinese hardware could cover before, and the verticalisation I flagged as a direction in 2024 is now underway across multiple product lines.

The third is operational reach. Deep strikes have moved from concept to campaign — Ukrainian long-range systems have repeatedly hit critical Russian infrastructure, oil ports and logistics nodes, without depending on Western munitions for every mission. The volume of Ukrainian long-range drone launches grew from approximately 110 UAVs in January 2024 to 7,500 in March 2026, the first time in a while Ukraine surpassed Russia in drone launches. UGVs (the land robots I wrote about couple weeks back) have gone from sideshow to serious procurement category.

What these companies are not yet, despite all of the above, is being global leaders. And that gap is not about product maturity or revenue scale — on both counts many of them are already world-class. What’s missing is export, and the barriers are specific.

Dual-use companies are doing fine. Himera and Sine Engineering on communications, are selling globally with strong traction as Western armies scramble for the latest electronic warfare innovations. In general, drones and UGVs can be exported when they have been removed from their military capabilities.

For pure weapons-related companies, the picture is more constrained. Ukraine restricts exports to keep local production focused on the war effort. Some companies operate in a grey zone, opening foreign entities in Estonia, Germany, or the US to relocate IP and export from there — but that comes with real risk. The Ukrainian government has shown it can pull local contracts from companies that route business through foreign entities, reassigning them to domestic competitors. Deep strike companies learned this the hard way, to the direct benefit of FirePoint.

The Gulf illustrates the full complexity. Gulf states tried to buy Ukrainian interceptors: low-cost, battlefield-tested, exactly what they needed. The political barrier hit first — Ukraine wanted to keep its interceptors for its own conflict and use export approvals as leverage in negotiations, offering Ukraine’s drone technology in return for diplomatic support and energy deals. Some companies did sell into the Gulf: Merops was the most visible winner, largely because it is US-incorporated and wholly owned by Eric Schmidt. Then came the technical barrier: Ukrainian interceptors still require pilots, and the Gulf has almost no trained pilot pool. More fundamentally, always-on air defense cannot depend on humans being on the clock. The same logic that makes Patriot batteries work — autonomous, persistent, no shift rotations — is what Gulf customers need. Ukraine hasn’t built that yet (but is working on it). Yet, as part of the Middle East exports of interceptors, Ukraine has promised training, software updates and co-production lines for its technologies, laying the foundation for durable defence partnerships.

Teleoperation indeed is still widely used. This can be shown in army’s recruitment programs - advertising that you can defend your country from the comfort & security of cities:

Several opportunities remain open:

• Maritime and underwater systems are underdeveloped relative to demand — U-Force’s valuation is a signal, not a ceiling. For instance, Ukraine has started to conduct the first counter-drone interception launched from an unmanned surface vehicle. We should expect to see more integration of air defense and sea capabilities.

• Swarm behaviour at scale is still blocked, mostly by a communications problem.

• Full autonomy for military systems. The push for autonomy is getting stronger - and we’re getting closer. Zelensky has announced he wants to conduct 25,000 missions with UGVs - that’s too much compared to the pool of available pilots at any given time. Besides, latency or camera freezing altogether can still plague teleoperated missions. An inflection point may be close, driven by advances in computer vision and the scale of real-world battlefield data Ukraine has accumulated and is now starting to integrate.

• Ballistic missile interceptors are the biggest gap. Ukraine has become expert at intercepting drones and Shaheds but remains dependent on US Patriots for ballistic threats. A single Patriot interceptor costs $4M and is produced slowly. The Iran war has drained global stockpiles — the US has withdrawn batteries from South Korea, delayed deliveries to Switzerland, and asked Poland to transfer interceptors from its own inventory, a request Poland declined. A Ukrainian solution here would not just solve a local problem - it would be a global product.

• Attritable cruise missiles for offensive operations remain a related gap.

My conviction coming out of this trip is simple. The question for Ukraine’s defense industry is no longer whether it can build. It has answered that. The question is which of these companies will make the leap from battlefield champion to global defense player. Some of them will. The ones that solve the export equation will be the next Ukrainian neoprimes and will reshape how the rest of the world thinks about where defense technology comes from.

April Global Fundraisings

True Anomaly, developing space warfare solutions, raised a $650M Series D led by Eclipse, Riot Ventures, and Paradigm.

Skydio, the largest US drone manufacturer, raised $110M at a $4.4B valuation.

Firestorm Labs, which builds containerized drone manufacturing systems, raised an $82M Series B.

Kelluu, building autonomous airships, raised a €15M Series A led by Nato Innovation Fund.

Not Terminators

The systems involved were not science fiction. They are four-wheeled platforms that look closer to small tanks rather than Hollywood-terminators. They cost between $10,000 and $50,000 apiece. They are operated remotely by humans via Starlink terminals, from positions well behind the kill zone.

This changes reach: Such systems are expendables, and they can venture in places too dangerous for human soldiers. In Ukraine today, such capability is what determines whether an operation is possible at all.

Zelenskyy named several systems in his address:

• The first three (Tencore’s TerMIT,Burevii’s Ardal, and, UArmor’s Protector) are the heaviest of the group and can engage fire on enemy armored vehicles

• Rovertec’s Zmiy is a demining platform armored against mine blasts

• The Ratel S is a kamikaze vehicle designed to crack open reinforced bunkers

• The Rys Pro mounts a remotely operated machine-gun turret with AI-assisted thermal tracking

Together they constitute a combined arms stack, not a single weapon — a reminder that “ground robots,” like “drones” before them, is shorthand for an entire toolkit of distinct platforms, each handling a specific phase of an operation.

The Year of the UGV

When I visited Kyiv last June, Tencore’s team told me this would be the year of the UGVs. At the time it read as founder confidence. It turned out to be an understatement.

In 2024, roughly 2,000 UGVs were delivered to Ukrainian frontline units. In 2025, that number reached 15,000 — surpassing every procurement target. Already 25,000 units have been procured in the first half of 2026 alone.

More than 280 Ukrainian companies are now developing variants. Over 200 distinct models were in production by mid-2025, with 40 new ones appearing that year alone, each iteration incorporating front-line feedback in weeks rather than the years that define Western procurement cycles.

The geographical layout driving this is straightforward. Aerial drone saturation has pushed the effective kill zone out to 20-25 kilometers from the front. Any infantry movement in that corridor is hopeless. FPV drones that cost less than a decent laptop will intercept armored vehicles on the way to evacuate a position. Ukraine faces a manpower crisis compounded by a physics problem: the most important ground on the battlefield is also the most lethally inaccessible to human beings. UGVs dissolve that constraint. Lose one, absorb the cost, send another.

From Logistics to Combat

UGVs entered the war as logistics platforms — crawling along cratered supply routes where pickup trucks had become death traps. By late 2025, the BBC reported that up to 90% of all supplies to frontline positions around Pokrovsk were moving by robot, not truck. One UGV retrieved a wounded Ukrainian soldier 64 kilometers inside Russian-held territory, surviving repeated drone strikes and a mine blast on the return journey.

Then the missions got closer to combat — though not by central design. As most shifts in Ukrainian warfare, the change from logistics robot to weapons platform happened bottom-up, driven by brigade-level improvisation. Units already running UGVs into kill zones to deliver ammunition noticed that if a robot could survive the route, it could carry a weapon too. The 3rd Assault Brigade built a dedicated strike UGV company, explicitly to prove the concept and force the state to follow. The Ukrainian Ground Forces Command only began formally developing a doctrine for robotic defensive lines in November 2024 — after the battlefield had already moved.

“For the Russians, it’s cheaper to send people. We don’t have people — so we build robots.” said Yurii Poritskyi, CEO of DevDroid.

In July 2025, Ukraine’s 3rd Assault Brigade reported that FPV drones and a kamikaze ground robot loaded with antitank mines struck a Russian bunker entrance in Kharkiv Oblast. When a second robot approached the damaged position, two surviving Russian soldiers held up a handwritten sign reading “We want to surrender.” Fedorov subsequently claimed that over 100 Russian troops laid down their arms to Ukrainian unmanned systems during the winter alone. In January 2026, a ground robot captured three Russian soldiers — the first documented prisoner capture by a ground robot in the history of warfare. By March, a ground robot equipped with a machine gun had held a frontline position for 45 consecutive days, reloaded every 48 hours, with no human presence at the position itself.

What does it mean for the future of infantry?

For some - robotization means the end of infantry. Brigadier General Andriy Biletskyi of the 3rd Army Corps has argued that units fully integrating UGVs could reduce frontline infantry requirements by 30% by end of year, and up to 80% further down the line.

But reality is usually more complex. UGVs still fail to reach their destination four times out of five in certain conditions. Navigation under electronic warfare remains a serious constraint. Urban clearing, holding terrain after seizure, exploiting breakthroughs — these still require human judgment and physical presence.

What is changing is the threshold for when and where human beings need to be present. The first line is becoming robotic. Humans are moving back.

The manpower dimension matters beyond Ukraine. Ukraine is fighting an adversary with three to four times its available manpower pool, and UGVs are directly softening that asymmetry — each robot that holds a position or runs a logistics route is a soldier who doesn’t need to be recruited, trained, or replaced. The implications for Europe are significant. Rebuilding credible standing armies across the continent has stalled politically for a decade, in part because reinstating conscription is both deeply unpopular and arguably misaligned with how modern warfare actually works — what militaries need are professionals who can operate complex networked systems, not mass infantry. A robotized frontline changes that calculation. Smaller professional forces, augmented by large fleets of expendable machines, may be able to generate the combat mass that European governments cannot generate through recruitment alone.

The rest of the world follows

As Ukraine is building the world’s first robotized ground army, the rest of the world’s militaries are catching up. Western companies also start to scale:

• ARX Robotics closed a €42 million Series A in 2025, with its Gereon UGV platform now deployed across six European armed forces including Ukraine. These are now used for logistics, but should move into combat use cases soon.

• Quantum Systems and Tencore announced a joint venture — Quantum Tencore Industries — in Berlin last month, targeting thousands of systems for Ukraine and NATO armed forces.

• In the US, Forterra closed a $238 million Series C in November, pushing its valuation past $1 billion.

Massive contracts towards robotization of the battlefield are now really kicking off. Just this week, Estonia suspended its €500 million infantry fighting vehicle program and redirected the funds toward drones and unmanned systems, citing Ukraine directly.

Meanwhile, France and Germany are pressing ahead with the MGCS — their next-generation crewed main battle tank program — while Germany and Italy are developing a separate successor platform, and Poland is executing one of the largest tank procurement programs in NATO history. These are multi-billion dollar bets on crewed heavy armor at the precise moment the battlefield is moving in the opposite direction. They may not age well.

Anduril and Helsing deliberately chose a different one. Both companies entered the UK market directly, targeting urgent operational gaps that the legacy system was too slow to fill and without a prime as an intermediary.

Anduril’s first UK contract was a £4M deal with UK Strategic Command’s jHub to trial its Lattice operating system at the airbase RAF Akrotiri - small in value, but an efficient way to start owning the relationship with the MOD directly. Helsing incorporated as a UK entity from day one, with a single proposition of building autonomous military systems.

Both approaches were a deliberate inversion of the traditional model:

From pilots to kill chains

The move from small innovation contracts to strategic relevance happened through Project ASGARD - the British Army’s effort to build a digital targeting web, connecting sensors to shooters in near-real time via AI.

Anduril provided the Lattice command-and-control backbone and the Ghost drone. Helsing provided AI-enabled targeting and the HX-2 strike drone. Contracts were awarded in January 2025, with first deployments a few months later in May 2025 - a tempo that traditional primes rarely match.

On Project NYX, the British Army’s collaborative combat drone programme, Anduril is both the prime and lead systems integrator, with GKN Aerospace building the airframes to its specifications. Helsing is following the same logic with its Plymouth factory, manufacturing the underwater drone itself.

Why foreign entities hit a ceiling

To lead programmes like UK’s ASGARD, you need List X status: the ability to hold and process UK Secret information on your own infrastructure. List X is only granted to companies incorporated in the UK with at least one physical UK site. Bid as a French or German entity, and the classified integration work must flow to a UK partner - which immediately downgrades you from prime to subcontractor, regardless of your technology.

The second constraint is softer but real. The UK MOD’s Defence and Security Industrial Strategy weights social value and UK industrial benefit at 10-20% of bid scores. A foreign entity with no UK jobs, no UK tax contribution, and no local manufacturing starts every bid behind. To win in the UK, you need to be perceived as a local player.

Sovereignty doesn’t require scale

Helsing didn’t solve this by building a large UK operation. It built a lean, strategically structured one - and today the UK MOD perceives Helsing as a British company. That perception came from people, not headcount: their first UK CEO was Nick Elliott, a former Director General of the UK Vaccine Taskforce. Sir Chris Deverell, a four-star general, came on as an advisor.

Anduril took the visibility even further and launched an aggressive marketing campaign that dominated the London Tube for weeks.

In parallel, operating an extensive lobbying effort is also necessary: Anduril ran 19 meetings with government officials since 2023, former MOD staff hired into the team, industry groups joined, events sponsored. Winning contracts in a new market is about being structurally local — but heavy lobbying definitely helps.

Incorporate early, hire for credibility, start with quick wins that build MOD trust, and scale from there toward programmes of record. Anduril started with a £3.8M jHub contract. It now sits at the centre of the UK’s defence plans.

Fundraisings

• Shield AI, which develops AI pilot technology for military aircraft, raised a $2B Series G at a $12.7B valuation led by Advent.

• Saronic, an autonomous maritime systems company, raised a $1.75B Series D at a $9.25B valuation, led by Kleiner Perkins.

• Hermeus, a hypersonic aircraft company, raised a $350M Series C at a $1B valuation, led by Khosla Ventures.

• Zipline, which operates autonomous drone delivery systems for medical supplies, food, and retail, raised a $200M Series H at a $7.6B valuation, led by Fidelity.

• PAVE Space, which develops orbital transfer vehicles to move satellites between orbits, raised a $40M seed round co-led by Visionaries Club and Creandum.

• Egide, a French startup developing interceptor drones to destroy hostile UAVs, raised a $9.3M seed round co-led by Expeditions.

• Airbase, which automates spectrum allocation, raised a $5M round led by A16Z.

• Eclipse, a hard tech VC firm focused on manufacturing and defence, raised a $1.3B fund.

I sat down with General Philippe Moralès - former Major General of the French Air and Space Force.

A few things stayed with me:

• On battlefield transparency: stealth still matters, but the window of surprise is shrinking fast as low-orbit satellites provide real-time surveillance of the battlefield. Stealth is being replaced by the ability to move faster and deceive smarter.

• On hypersonics: the attacker has the advantage today, but a prepared defender with the right technology (including directed energy) changes the equation.

• On Collaborative Combat Aircraft and autonomous flight: when you have two trained pilots, you’re not twice as strong — you’re four times stronger. Now imagine that dynamic at scale, with robotized wingmen that enable you to have larger patrols.

• On AI and C2: General Morales watched his own notes become obsolete in 18 months. The OODA loop - Observe, Orient, Decide, Act - will accelerate. The question is whether your doctrine accelerates with it.

🎙️ Original episode in French:

🎙️ I played with ElevenLabs to translate the episode in English :

Transcript (lightly edited for clarity and flow) is below.

Introduction & career

Louise Boucher: General Morales, we welcome you. You’ve had a truly exceptional career — 38 years serving in the Air and Space Force, from commanding squadrons to ultimately becoming Major General. You’re known for having had a very modern vision from the outset, and you’ve witnessed the evolution of conflicts from the Balkans to Iraq and Chad. We’ll discuss drones, hypersonics, space, AI in command. But before we dive in, I’d like you to recount your journey in your own words.

General Philippe Morales: I’ll leave it to your appreciation of my background. I am a classic fighter pilot. I’ve flown older generation aircraft — third generation, not even fourth. I know the Jaguar and the Mirage F1. I had the chance to lead the Patrouille de France in 1998. After that, it was a classic fighter pilot career: alternating between operational and staff assignments. I commanded a defense base of 5,000 personnel in Nancy during an experimental phase, then held more significant command positions — deputy commander for special operations, and commander of air defense and air operations. I concluded my career as Major General of the Air and Space Force, meaning number two of the force. A bit like, proportions kept, what a managing director is to the chairman of a large corporation, or what a prime minister is to the president of France. He’s the one who implements the policy of the Chief of Staff — responsible for present operations, training, operational readiness, armament programs, and preparing for the future.

General Philippe Morales: And preparing the future is complicated now. Technology evolves so quickly that projecting 30 years ahead is extremely difficult. At the start of my career, there was a linear progression — a bit faster, a bit higher, a bit further. Simple. Now I can’t tell you what things will look like in 30 years. Three years ago, I was asking myself certain questions about artificial intelligence at the very start of my role. When I left, I looked back at my notes from 18 months prior — and they made no sense anymore. The questions had evolved so much. Being rigorous at 30 years out, I think, is illusory. But the intellectual exercise of imagining the environment in 30 years, and then asking what an army would need in that environment — that’s interesting. That’s foresight.

Lessons from Ukraine and Iran

Louise Boucher: I’d like to start by diving into the present. Ukraine taught us about saturation and layered air defense. Iran confirmed it wasn’t just a symptom but a recurring theme — waves of drones and attacking missiles. These topics have captured enormous media attention. But for you, what are the recurring themes we’re seeing across both theatres that we still underestimate today?

General Philippe Morales: We can observe two things, and it’s interesting to compare. In Iran, stealth is being used by the Israelis and Americans — and it gives them total air supremacy. The Americans and the Israelis do whatever they want in Iranian airspace. That’s not the case in Ukraine, where stealth has not been employed. You can draw the same lessons by example or counterexample from these two present-day theatres.

General Philippe Morales: And specifically, we see something emerging: total transparency on the battlefield, an increasing level of openness in the environment. Before, we could conduct operations with the principle of surprise — take your time, prepare carefully. Now we’ll have much less time, because the enemy will discover preparations more quickly. The important thing now is to be very fast. If you want to surprise, you must move fast.

Hypersonics: the advantage goes to the attacker

Louise Boucher: If speed is the key, we should talk about hypersonics. Iran reportedly used its first hypersonic missiles — though I hear conflicting reports. What is the real revolution? Is it the speed, the maneuverability, or the difficulty of interception?

General Philippe Morales: There are two things. The speed of the effectors — and the speed of command. We mustn’t forget the latter. With AI and computing power, we’ll be able to decide more quickly. That’s the hypersonic of command, if I may express it that way.

General Philippe Morales: On effectors: if you go very, very fast, it’s extremely hard to intercept. Between the moment you detect a hypersonic missile and the moment it reaches you, you have very little time. You need something extremely sophisticated to hit a tiny target arriving at full speed. There’s clearly an advantage for whoever employs their hypersonic effector. At the same time, the defender who is prepared — if the missile is purely ballistic and therefore predictable — can respond with a laser, for example. You point it, you fire it, with relatively little energy. The envelope melts, the missile disintegrates in the atmosphere. So there’s an advantage for the attacker, but a prepared defender with the right technology can fight more effectively than against a conventional threat.

Lasers: science fiction or reality?

Louise Boucher: Today we see a cost advantage going to the attacker — cheap drones countered with expensive interceptors. On lasers: for me it’s still a bit science fiction. What is the current state of the art? Is it actually used in combat?

General Philippe Morales: To my knowledge, there are experiments. You’d probably want to interview industry specialists in optics and laser technology — they could tell you whether prototypes exist and how far away scaling really is. But we do see, especially with mini drones, lasers that don’t require phenomenal power but are sufficient to disrupt a drone’s electronics and destroy it. There are many solutions. You shouldn’t be disheartened when you face mass-produced, low-cost threats. A mass, low-cost response can be very effective.

Battlefield transparency and space dominance

Louise Boucher: When you say the battlefield is becoming more transparent, I imagine that refers to the explosion of micro-satellites — commercial and military. Can we still hide? Do European armies have enough of their own constellations to maintain a spatial advantage?

General Philippe Morales: With the explosion of low-orbit satellite constellations, we’re starting to see many things. When we can see everything, there’s almost nowhere to hide — unless you’re under dense foliage, and even then you emit heat, or a radar satellite picks up a metallic mass. What’s happening, however, is a monstrous flood of information. Finding the needle in the haystack is difficult — except that now there’s something called automatic image interpretation with AI, which will scan and analyze at speed.

General Philippe Morales: We’re also emitting more and more electromagnetic waves. So we’re visible not just in the optical spectrum — infrared to ultraviolet — but across various electromagnetic frequency bands. By correlating all of that, you can create a rather comprehensive picture. That’s why there is much greater transparency on the battlefield now. Either you say: I cannot hide, so I must go very fast — speed is my strength. Or you try to camouflage, deceive, send false images, generate false targets, false radar signals. At some point, the enemy doesn’t know what to do. The answers are always the same: surprise, mislead, deceive. It just carries far greater importance than before.

Electronic warfare and the Kessler effect

Louise Boucher: There’s also a third technique: blinding the enemy. We’ve seen examples in Ukraine of jamming and spoofing — electronic warfare to prevent satellites from being used effectively.

General Philippe Morales: Absolutely. Before, we could jam a radar more easily. Now it’s satellite constellations. You can try to blind a satellite with a laser beam — as it passes over a target area, a beam is emitted, the satellite sees a white spot and can’t identify what’s underneath. You can also attempt to drown signals in random noise. And beyond jamming, there’s the question of physically destroying satellites. The problem is the Kessler effect: destroy one low-orbit satellite, and you generate thousands of debris that destroy others, exponentially. Low earth orbits become completely polluted — nothing can exist there anymore. This becomes a deterrence weapon, because whoever relies heavily on low-orbit technology has the most to lose.

CCAs and the future of air combat

Louise Boucher: Today when we talk about drones, people think of FPVs or Shaheds. But there’s a larger challenge around Collaborative Combat Aircraft — CCAs. In five to ten years, what does the next patrol look like?

General Philippe Morales: Let’s start with autonomous drones. You have platforms like the Reaper that are still remotely piloted. But imagine an autonomous version capable of basic evasion in response to a threat — turning away from an approaching enemy aircraft, triggering electronic countermeasures. That gives it meaningful survival capability. And then there are many missions we don’t currently dare to perform because we won’t send a pilot into that risk. With autonomous aircraft — the first aerial robots — we could.

General Philippe Morales: Here’s the key insight: when you have one fighter jet, it’s very effective. When you have two highly trained pilots flying together, you’re not twice as effective — you’re four times more effective. The multiplier compounds. With four planes, you’re sixteen times stronger. That’s a caricature, but the principle holds — smart patrol coordination between aircraft that support and decoy for each other is an enormous force multiplier. The grail is the robot patrol. If you have automated combat robots that know how to work in patrol, learning is just a software update. Even with basic software, you compensate with numbers — instead of two or four, you deploy twelve or twenty-four.

General Philippe Morales: If there’s an algorithmic breakthrough in this area, whoever achieves it will be able to project enormous air power at very low cost. And if there are losses, you rebuild and reinject the program. The question isn’t whether this will happen. It will. The question is when.

AI and C2: the OODA loop

Louise Boucher: When we talk about C2, it’s used for almost everything touching military software. What does command and control actually look like?

General Philippe Morales: C2 is essentially: establish a central hub, absorb intelligence, images and objectives, form a maneuver concept, identify targets, plan missions, execute, monitor, correct. It’s very manual. Very process-driven. Armies have genuine expertise here that I’d encourage industry to study — it’s perfectly adaptable to crisis management, which most companies frankly lack the methodology for.

General Philippe Morales: Now: if AI helps you catalog, detect, plan, and target — if it says, given your resources, do this first, then that, and here’s the simulation of the outcome — you have a revolution. This is the OODA loop: Observe, Orient, Decide, Act. The faster you run this loop, the stronger you are. If you run it faster than the enemy, you’ll overwhelm him. What AI will bring is the God’s Eye view. In a fighter jet, I used to read a radar oscilloscope and interpret a blip. Now you have a screen that shows your plane and the enemy maneuvering in real time. AI in command centers will do that for the entire battlefield.

Palantir, Anthropic and the air campaign in Iran

Louise Boucher: Concretely — we read that in Iran, the Americans used Palantir with Anthropic for better planning and targeting. What do you make of that?

General Philippe Morales: I have no more information than you do from the press. But when you use LLMs, when you use ChatGPT to research something, you quickly see that it helps you work much more efficiently. With Palantir structuring the data and Anthropic interpreting it, you can imagine an OODA loop that runs much faster. Where many people were needed, fewer will be. Where things were slow, they’ll go faster. And since you need fewer people, you can supervise more, reallocate staff, fight several battles in parallel. When I look at what appears to have happened in Iran — a very effective air campaign, executed at an extremely rapid pace — I can absolutely imagine that this kind of algorithmic assistance, targeting what is most sensitive or most strategic for the enemy, is what made it possible.

Underestimated threats: biology, multi-domain, robotization

Louise Boucher: What topics aren’t taken seriously enough yet, or will transform military doctrine?

General Philippe Morales: I think of CRISPR-Cas9 — the DNA scissors. We can imagine a future where confrontation is no longer about planes, tanks or ships, but about viruses and vaccines. My opponent makes viruses; my best response is to make vaccines faster. The shield and the sword will change their nature.

General Philippe Morales: And then there’s the multi-domain approach — Joint All Domain Operations. Ground, air, maritime, cyber, electromagnetic, information — all synchronized, overwhelming the enemy across every field simultaneously, including economic pressure. If done well, we can win the war before the war, to borrow the phrase of former Chief of Staff Thierry Burkhard. Collapse the economy, spread false information, take advantage of internal tensions — at some point the regime collapses without kinetic action. Utopian, perhaps. But intellectually serious.

General Philippe Morales: On robotization: people say it’s immoral to send robots to kill. But consider — if it’s robot against robot, and my algorithm is better than yours, you don’t fight. Like the Greek champions of antiquity, fighting so their armies don’t have to. The ethical calculus changes. The problem is: if conflict has no human cost, the threshold to wage it lowers. Man will always be a wolf to man. That’s the constant.

Techno-optimism and closing thoughts

Louise Boucher: You are genuinely a well of knowledge. And I didn’t expect, when we first spoke, for you to be so techno-optimistic.

General Philippe Morales: You’re right to push back on that label. Being techno-optimistic can be a flaw — it implies technology solves everything automatically. That’s not true. It’s about the use of technology. Things are evolving incredibly fast. When I first joined the Air Force, combat piloting looked nothing like what it is now. Technology has an enormous influence on behavior. The key is to keep thinking, keep improving, incrementally — POCs, MVPs, building up. At some point, several ideas converge and you see a real discontinuity. Someone gains the edge. Because it’s accelerating, the imperative to keep thinking actually increases, not decreases.

Louise Boucher: I use techno-optimism as a compliment — it shows you’re techno-curious. Techno-pessimism is a form of rejection. There’s a power in technology that changes everything, and we have to engage with it. It was a genuine pleasure, General Morales. Thank you very much for your time.

Walking around the room, you become quickly aware that you are in the presence of a new configuration of power. A small number of technology founders and investors have built direct relationships with the executive branch and are now shaping procurement, industrial policy, and regulatory frameworks. The tech elite is setting its agenda on the domains that matter most. Hill & Valley is where that new order convenes

The core thesis: the manufacturing mistake is the original sin

If there is one shared belief that runs through every panel at Hill & Valley, it is this: the decision to offshore American manufacturing to China between roughly 2001 and 2015 was a civilisational error, and the next decade is about correcting it.

This is not nostalgia for a lost industrial America, but more a strategic reappraisal. The argument is that you cannot maintain military and technological sovereignty without a domestic production base. You cannot defend Taiwan if you cannot build submarines. You cannot run AI models if you cannot manufacture the chips that power the data centres. You cannot win a drone war if your supply chains run through the country you are fighting.

But the more uncomfortable observation is that China has moved from out-manufacturing the US to out-innovating it — across AI, biotech, and energy simultaneously. These are no longer domains where China is the low-cost replicator and the US is the original inventor. The gap has closed, and in places it has reversed. The US is 232 times slower at manufacturing ships than China. China today produces more advanced robotics, and is moving aggressively down the cost curve on AI inference. The US tech-defence establishment has internalised this shift. The response is not complacency but a kind of urgent clarity: you need to rebuild the production base and stay ahead on the frontier.

The clearest illustration of the production problem came from an unexpected place: a science communicator turned scrubber manufacturer who tried to source a stainless steel nut made in America and was advised by Chat-GPT to import it instead. His company eventually had to route chain mail production through Canada because the one US manufacturer capable of making it was already fully committed to the Navy.

Which brings us to the kingmaking problem: Industrial policy is back but governments are notoriously poor capital allocators, and the line between “directing strategic investment” and “picking winners” is thin.

The new model being championed at Hill & Valley tries to sidestep this by structuring shared risk: private capital builds the infrastructure, government provides the long-term offtake commitment. The clearest example is the recently announced $2.4 billion joint Navy-Hadrian investment in Alabama — a 2-million-square-foot automated factory dedicated to Virginia and Columbia class submarine manufacturing. No free money from the Navy, as the Secretary put it on stage: if the company underperforms, it does not get paid. Private markets capitalise the factory on a long amortisation schedule; the Navy buys the output.

Space is already a battlefield — and a new economic layer

NASA is back and it announced the most ambitious overhaul of the US space programme in decades, under the banner “Ignition.”

The headline: NASA will build a permanent lunar base, committing $20 billion over 7 years. Alongside this, NASA announced the first nuclear-powered interplanetary spacecraft, slated to launch to Mars by 2028 to demonstrate nuclear electric propulsion. Isaacman is also reconstituting Mission Control as a vertically integrated capability, reversing decades of outsourcing.

The subtext is competitive. The United States is in the midst of a 21st century space race, with the goal of returning to the moon before China. “If you fail going back to the moon, people think you're failing short everywhere”: failing to return to the moon is not just a space policy failure — it is a legibility signal to allies and adversaries about American capability more broadly.

What is new in 2026 is the economic layer. Historically, the three traditional pillars of the space economy were launch, observation, and communication. They are now being joined by a fourth: the prospect of millions of AI-enabled satellites as distributed intelligence infrastructure.

For Europe, the implication is sharp. If you want space sovereignty, you need a launcher first — before you think about replacing Starlink or building observation capacity. Without sovereign access to orbit, every other space ambition is contingent on American goodwill. All eyes on Isar Aerospace to build much cheaper space access.

What the US actually wants from Europe

The American tech-defence establishment is not anti-European - or not really. The framing at Hill & Valley was consistently that the US-EU alliance is the most important economic relationship the US has — Europe remains its largest trade partner — and that the 60-year US military alliance structure is what provided the most stable interstate order in recorded history, however imperfect.

What the US wants, stated plainly, is for Europe to become stronger: GDP per capita in Europe has declined from roughly 90% of US levels to around 70% over the past two decades. The ask is: rebuild your industrial base, become more pro-business, grow again. A stronger Europe is a better ally. A Europe that is dependent on US security guarantees while running persistent trade surpluses and underinvesting in defence is not a sustainable arrangement — and the current US administration is making that expectation explicit rather than diplomatic.

The secondary ask — implicit but present — is that Europe not become a vector for Chinese technology or capital into sensitive supply chains. Every supply chain decision is now a geopolitical decision. European companies that remain deeply integrated with Chinese suppliers in semiconductors, energy infrastructure, or communications are a strategic liability from the American perspective.

What this means if you are building or investing in European defence

The bottleneck is physical, not digital. The US is rediscovering this. Europe, which never fully dematerialised its industrial base in the same way, has a potential structural advantage — if it can mobilise it. The question for European defence founders is not whether to build software, but whether the software they build is connected to real production capacity.

Industrial policy is back, but the kingmaking risk is real. The Hadrian/Navy model — private capital builds, government commits to offtake — is the template being promoted. Europe should be designing equivalent structures. Germany has been more vocal recently on choosing winners: Proxima for fusion, Hypersonica for missiles, Helsing and Stark for autonomous weapons.

The clean tech lesson applies. General Catalyst’s Hemant Taneja was direct about this: in the clean tech era, a lot of capital went into complex hardware companies, and the moment markets turned and the capital spigot stopped, it was very hard for those companies to show progress — because hardware takes a long time to reach the right unit economics. The implication for defence tech: back companies where manufacturing competency is core, not an afterthought.

The Hill and Valley Forum is, ultimately, a document of American ambition in a moment of genuine strategic urgency.

The people in that room believe the decisions made in the next 24 months will shape the structure of the global economy for decades. They are probably right. The question for European readers is whether we are part of that story as architects, or as observers.

Fundraisings:

Shield AI, a developer of AI-powered autonomous systems for military aviation, raised a $2 billion Series G at a $12.7 billion valuation. It also acquired Aechelon, a 27-year-old defense simulation company specialising in high-fidelity image generation and training data for autonomy software

Zipline, a South San Francisco company operating autonomous drone delivery systems for medical supplies, food, and retail, raised a $200 million Series H at a $7.6 billion post-money valuation, led by Fidelity, with Baillie Gifford, Valor Equity Partners, Tiger Global, and Paradigm participating.

PAVE Space, a Lausanne-based startup developing orbital transfer vehicles to move satellites between orbits, raised a $40 million seed round co-led by Visionaries Club and Creandum.

Egide, a French startup developing interceptor drones designed to destroy hostile UAVs, raised a $9.3 million seed round co-led by Expeditions and Eurazeo.

Airbase, a startup developing AI-driven spectrum allocation software for military and commercial communications, raised a $5 million seed round led by A16Z.

Simultaneously, Merlin — which designs autonomous flight systems — also had a successful IPO and raised $200M, valuing the company at $1B.

This past week tells you everything about where investor conviction is coming from right now: capital markets are hungry for defense exposure, no matter the actual performance of the underlying asset.

This appetite is also very visible in Europe

In January, CSG — a Czech constructor of armored vehicles — recorded the largest defense IPO ever recorded globally. It raised €3.8 billion on Euronext Amsterdam and shares closed 31% above the offer price, valuing CSG at nearly €33 billion.

Whereas software IPOs in Europe are nearly non-existent, defense is the exception. Over the past 18 months: TKMS went public at €4.5B, Renk IPO’d and peaked with a share price up 235% , Exosens listed at €3.5B with shares up 200% and Rheinmetall was one of the most performant stock globally (x4).

More is coming too:

• KNDS — the Franco-German maker of Leopard 2 tanks — will go public this year at a valuation of around €20 billion. It comes to market with a €23.5 billion order backlog, built on €11.2 billion in new orders in 2024 alone.

• Doncasters, a UK-based precision engineering firm, is preparing for an IPO later this year, aiming for a valuation of $4 billion.

• Quantum Systems, the €3B German autonomous drone company, has been rumored as the next European defense tech IPO to watch.

The appetite isn’t limited to public markets

Anduril secondaries are now trading at $90B. Just this week, Europe and Israel quietly minted three new defense unicorns (not yet announced).

Swarmer also tells a broader story: Ukrainian companies, which used to be discounted by foreign investors, are now stealing the spotlight. Last month, Ukraine minted its first defense unicorn with U-Force, the naval drone company, Airlogix passed the $500m in revenue milestone and several large fundraisings are underway.

Where future listings land is now a live strategic question.

Usually, European startups all incorporate in Delaware (US) as the setup is easy and globally recognized. In defense, founders hesitate for sovereignty and some believe they will have to IPO in Europe.

But Swarmer shows a different path: a company with no US revenue, founded and operating entirely outside the United States, that listed on Nasdaq — successfully. In a similar manner, Doncasters — a UK defense company — is expected to also go public in New York rather than in London.

The gravitational pull of US capital markets remains as strong as ever.

Fundraisings & M&A:

• AV (formerly AeroVironment) – drone manufacturer, acquired California-based ESAero for $200M, less than a year after its $4.1B acquisition of BlueHalo.

• Advanced Navigation – AI-powered navigation and positioning systems, raised a $110M Series C.

• Cape – privacy-first mobile carrier, raised $100M led by Bain Capital Ventures and IVP.

• Nominal – defense testing startup, raised $80M led by Founder’s Fund, and announced a teaming agreement with ground autonomy company Forterra.

• Arc – electric powertrains for maritime and defense vessels, raised $50M Series C led by Eclipse.

• Smack Technologies – frontier AI lab for national security, raised a $32M Series A, with the Seed led by Point72 Ventures and the Series A led by Geodesic Capital and Costanoa Ventures.

• Worldscape.ai – AI-powered geospatial intelligence, raised a $6M Seed led by Scout Ventures, with participation from Radius and Washington Harbour Partners.

Martin and I have known each other from the early days of Harmattan, France’s first defense unicorn.

Harmattan went from zero to 100+ people across Europe, the Middle East and the US.

We discuss how they built a talent culture that compounds on itself, how Harmattan is contributing to the next generation of combat aircraft, and why vertical integration is the real moat.

🎙️ New episode live:

Transcript (lightly edited for clarity and flow) is below.

Introduction

00:34

Louise Boucher Hi Martin, thanks for coming by. You built the first French defense unicorn — and you started out at Parrot, one of the leaders in consumer drones, before moving into defense. I thought the shift from traditional drones to defense-first drones would be an interesting place to start. Can you take us back to your time at Parrot — what did they get right in the first wave of European drones?

Martin de Gourcuff If we’re being honest, DJI is the only company that truly got it right back then. Many others died — 3D Robotics, GoPro — or became small players. That said, Parrot did get a few things right and ultimately survived the consumer drone bloodbath that happened between 2010 and 2020. They made a robust product, they were very vertically integrated — which brings cost-effectiveness and decouples you from the supply chain — and they acquired Pix4D, the photogrammetry company, which turned out to be a strong business move.

Louise Boucher DJI became ubiquitous on the battlefield. How come more traditional players like Parrot or Skydio didn’t manage to catch that wave as effectively?

Martin de Gourcuff It depends who you’re talking about. You could argue that Quantum Systems — initially focused on agriculture drones — is actually one of the winners of this defense cycle. As for Parrot, I believe they were initially reluctant to go full defense and got somewhat lost in their product roadmap. Skydio split their focus across security and inspection markets. Both companies’ ISR drones suffer from a relatively high price point and have faced drawbacks in contested environments, though I think that’s improving.

What Harmattan Took from Parrot

02:59

Louise Boucher When you founded Harmattan, what did you want to bring from the Parrot experience — and what did you leave behind?

Martin de Gourcuff There were a lot of useful takeaways. Vertical integration remains critical — even more so today. That was a strong lesson. Parrot was also a company that didn’t cut corners when building — as R&D engineers, we had the means and tools to build good products. And we were deeply product-oriented: get to something functional as fast as possible, then make it robust by testing it in every condition imaginable.

Vertical Integration as a Moat

03:58

Louise Boucher On vertical integration — beyond cost effectiveness, is part of the push also about supply chain sovereignty? Reducing dependence on Chinese components?

Martin de Gourcuff We’ve been vertically integrated from day one, for reasons that go beyond the sovereignty trend. Controlling your supply chain is always a key differentiator — it lets you rely on multiple sources, enables more integrated and customised designs, and reduces both price and weight. Of course, sometimes it doesn’t make sense to build your own components from a time-to-market or cost perspective. But in our industry, when you ask “should I make it or should I buy it?” — the rational answer is very often: make it. Even more so as geopolitical tensions have increased.

Louise Boucher The US is leading on this — drone manufacturers now need up to 65% American components to qualify for certain programmes. Has that affected Harmattan’s roadmap, especially as you expand into the US?

Martin de Gourcuff Yes. Even though we’d already anticipated the importance of supply chain control, this has a real impact — we have to update our sourcing strategy, adapt products, make trade-offs, and sometimes fundamentally rethink designs. One concrete example: replacing Nvidia components for your main processor and hardware acceleration. You have to rely on other players and get creative.

Louise Boucher Is removing Nvidia chips already a requirement from European MoDs, or is that an anticipation on your part?

Martin de Gourcuff It’s an anticipation. We also try to stay compliant with different export control rules, which don’t always allow certain components to be integrated into our products.

Louise Boucher What are the alternatives to Nvidia for high-performance embedded AI?

Martin de Gourcuff You have to be creative. You can approach Nvidia-level performance by combining several components — a strong CPU from one maker, dedicated hardware acceleration from another. That way you can assemble something that functionally resembles what Nvidia offers for embedded systems.

Scaling Harmattan: Talent, Speed, and Structure

07:37

Louise Boucher Harmattan went from zero to 100+ people across France, Switzerland, the UAE and the US. How did you maintain a high talent bar, fast iteration cycles and organisational coherence while scaling that quickly?

Martin de Gourcuff These are three distinct challenges. On talent: keeping a high bar has to be treated as one of the top priorities — if not the top priority — of the company. That means engineering the sourcing and recruitment process well and improving it continuously.

On fast iteration: start with a smaller feature envelope, make it work, then iterate, then make it better. If you try to reach the best product on the first attempt, you’ll almost certainly fail.

On scaling: if not handled properly, it leads to three problems. Mess — if you’re not organised enough, you scale into chaos. Bottlenecks — if you rely too heavily on a few key people, you slow down the whole company. And silos — parts of the company that drift apart and stop following the same standards. The first two are solved by continuously rethinking organisation, processes, tools and recruitment as you grow. Silos are addressed through top-down transmission of company culture that keeps everything bound together.

Louise Boucher Was there a company model that inspired how you structured Harmattan — DJI’s internal competition model, or Palantir’s small embedded teams?

Martin de Gourcuff The Palantir forward-deployed engineer model is interesting in industries where the number of clients is low and the economic stakes are high. It doesn’t apply to hardware, but for government software development, it’s something we’re looking at.

Where to Find European Defense Talent

10:19

Louise Boucher In the US there’s a clear premium on Anduril, Palantir, SpaceX alumni. Are there equivalent talent pools in Europe that are less obvious?

Martin de Gourcuff There are strong talent pools in Europe that haven’t experienced the same salary inflation as software and machine learning in the US. It depends what you’re looking for. For people who can build cost-effective hardware products fast, look at successful consumer electronics companies — Parrot, Devialet and similar. You’ll find very strong engineers who haven’t necessarily been discovered by Big Tech yet.

For top hardware talent — sensors, high-grade navigation, radio frequencies — look at traditional primes: Safran, Thales, MBDA, Leonardo, BAE Systems. For software and machine learning, look at European AI labs like DeepMind or Mistral, or the European offices of Google, Meta, Palantir and others.

Overall, the best ecosystems for defense tech and robotics are in Paris, London, Lausanne, Zurich, Munich and Delft.

Louise Boucher Do you hire young engineers straight out of school, or do you prefer people who already know the craft?

Martin de Gourcuff Age is a poor predictor of the quality of engineers. I look for curiosity, ownership, passion, motivation — qualities you can find at any age. That said, there are roles where experience makes a huge difference, simply because you need to have seen things fail before you can make them work. RF hardware design is a clear example — your first RF design is a long way from something that actually works in the field. In those cases, you need to understand that you’re standing on the shoulders of giants, and hire experienced engineers so juniors can learn from the best.

Autonomy, Ethics, and the Limits of AI

13:51

Louise Boucher European MoDs, AI ethics, and autonomy — how hungry are operators for more AI in their systems, and where do you see the real limits?

Martin de Gourcuff Contrary to what many people think, I believe AI — in the sense of autonomy — can actually make systems more ethical. If you build a system to perform a specific mission, you can lock its usage. Our drone interceptors at Harmattan are trained to detect and track specific types of drones. You couldn’t repurpose them for something else. Autonomy can be a constraint on misuse, not just an enabler of capability.

That said, full autonomy raises genuine ethical issues because of real technical limitations. The parallel with autonomous cars is useful. Two reasons we don’t have fully autonomous vehicles yet apply equally to defense systems.

First: machine learning algorithms struggle with rare events that weren’t well-represented in training data. As a human, if a monkey crosses the road, I can make sense of it — I can infer it escaped from a nearby zoo. A computer vision system may be completely confused by that same event and take a bad decision. Second: sensors have fundamental flaws. Cameras degrade when facing the sun. These limitations bound what any autonomous system can safely understand about its environment.

Until we solve those problems, we won’t be able to rely on fully autonomous systems — unless you don’t care about collateral damage.

Louise Boucher But we do have autonomous cars operating now, in parts of the US and increasingly in Europe. Is AGI-level autonomy coming to defense systems?

Martin de Gourcuff The level of autonomy achievable today is already a game changer for most applications. Whether AGI comes to defense depends on what AGI means. If it means the ability to learn new skills by interacting with an environment, I’m not sure it’s well suited to defense — you need systems you can control, bound, and explain. Something that has complete freedom of action is precisely where collateral damage and tragic events become likely. Explainability isn’t optional when lives are at stake.

That said, AI is clearly non-negotiable for perception tasks — detecting and tracking entities, extracting geospatial information, semantic understanding of a scene in 3D. And it’s a genuine game changer for complex mission planning, where you can mathematically model the problem and use powerful planning algorithms to find the optimal solution.

Dassault, Combat Aircraft, and the New Primes

19:56

Louise Boucher You recently raised a Series B with Dassault Aviation, with announcements around collaborative combat aircraft. How autonomous are we really trying to get there?

Martin de Gourcuff What I can say is that we are indeed collaborating on topics related to technical components for collaborative combat aircraft and autonomy. I can’t go further than that.

Louise Boucher More broadly — until recently, the assumption was that new primes would stay in expendable, smaller autonomous systems. Now we’re seeing companies like Helsing announce their own fighter jet programmes. New primes appear to be moving into territory that was previously the crown jewels of the traditional primes.

Martin de Gourcuff For us, Helsing moving into its own fighter jet programme is a signal of ambition, and proof that in this industry you very often need to develop your own hardware. In my opinion, that’s the direction of history. There is no way the fighter jet landscape won’t experience major shifts in the coming years — both in terms of who the players are and in terms of what a fighter jet actually looks like. I suspect we’ll see new combat aircraft emerge that don’t necessarily resemble what we know today.

Louise Boucher The US collaborative aircraft programme is a $9 billion effort bringing together Primes, Neo-Primes, software companies — all partnering. Does that model hold up?

Martin de Gourcuff A fighter jet is a collection of complex subsystems — motors, electronic warfare, avionics — each made by different companies and integrated by the airframe manufacturer. It’s very hard to believe a single company could master that entire stack, unless you’re building something fundamentally simpler. Which can also make sense for certain use cases.

Hiring and the European Talent Shift

23:22

Louise Boucher Who are you looking to hire, and where are you expanding?

Martin de Gourcuff We’re looking for engineers with a passion for building responsible defense systems. Today’s geopolitical landscape has made clear that to preserve the world we want to live in, we need to be able to defend ourselves and remain as independent as possible. We’re hiring in France, across Europe, in the US and in Morocco.

Louise Boucher Are you seeing European talent returning from US companies to work on European sovereignty?

Martin de Gourcuff Yes, definitely. Recent geopolitical developments have pushed people who were at US companies to come back to Europe or pivot toward European sovereignty projects. We’re seeing that movement. The challenge is that US compensation packages are very compelling — the activation energy required to make people move is sometimes hard to match.

Louise Boucher Thanks Martin — great conversation.

Martin de Gourcuff Thank you for having me.