Debunking myths on defense tech
Myth #1: “Defense startups can’t scale quickly enough for VC returns.”
Defense startups are a mix of hardware, software and some projects observe dynamics of deeptech projects. However, there is now enough data points to debunk this myth entirely:
Anduril’s valuation just doubled to $28B valuation, with $800M in sales, 7 years after launch.
Saronic climbed from $50M to $350M in bookings in a single year, valued at $5B just 3 years post-launch.
Meanwhile, over in Europe:
Helsing is valued at $5B after just 3 years in business.
Tekever is pulling in c. $100M in revenue by selling drones and related services to the UK, France, Ukraine, and Portugal.
Destinus generates c. $70M in revenue, 3 years after launch, through a vertically integrated approach to autonomous flights.
No matter how you slice it, these aren’t slow-burn growth stories. The shift towards autonomy and AI — plus the fact that budgets are “sticky” once you’re in — allows for robust, predictable revenue streams that can outlast typical consumer or enterprise cycles. The main difference lies in the fact that, contrary to SaaS continuous growth curves, here we observe more stair-step like growth.
Myth #2: “Competing with established defense incumbents is impossible.”
Many of us look at Lockheed, Thales, or BAE Systems and think they’re unstoppable, bolstered by decades of cozy relationships with governments.
But technical shifts can make even the biggest incumbents vulnerable. In defense, the rise of autonomous drones, unmanned vessels, and AI is creating a whole new category of “modern primes.” This is why we’re seeing the US DoD now awarding multi-billion dollar contracts to next-gen players. The most salient example is Anduril securing a $9bn contract with the Air Force, beating out legacy names like Boeing and Lockheed Martin. Others, like Saronic are rumored to have secured hundreds of millions worth of contracts from the DoD’s Replicator Initiative.
This shift will also accelerate, and the French Head of Ministerial Agency of Artificial Intelligence in Defense Bertrand Rondepierre recently stated that the industrial base for modern defense won’t be the same as the historical one — pointing that in Europe will award more and more contracts to modern players vs incumbents.
Instead of going directly head-to-head with incumbents in tender offers, we’ve also seen partnerships pop up between legacy primes and next-gen players. Eg Helsing partnered with Saab, bridging that gap between an emerging AI software specialist and an established manufacturer. Other examples include Anduril partnering with Rheinmetall to conquer unknown markets.
The bottom line? If your technology is compelling enough, governments will carve out funding for you — even if you’re not a giant (yet).
Myth #3: “It’s impossible to truly scale internationally — each government will favor its local player.”
Defense is the most sovereign category, so it can be easily assumed that each player will be mostly restrained to its native market. We obviously see these dynamics in play — the French MoD was reluctant at first to work with Helsing because they’re publicly supported by the Bundeswehr. Even Anduril has been said to close its European office after struggling to secure major contracts.
But this is definitely not the whole story. Looking at defense incumbents in Europe, they make most of their revenue outside of their native markets: Thales, BAE Systems and Dassault each make 70–80% their multi-billion revenue outside of their home countries.
Different paths can be taken to achieve this:
Build a multi-country lobbying force: Helsing has been busy expanding its presence across Europe by opening local offices to leverage them as lobbying. For example, both hires Antoine de Braquilange and Marc Fontaine were given the autonomy to become public figures for all French-related topics.
Brand yourself as a neutral player: Safran ($40bn market cap) is the leading example of this strategy, some players like Harmattan AI (autonomous drones) are trying to replicate by selling to both Europe and UAE, a year into existence. Comand AI (AI command system) was also able to win its first contract in Germany simultaneously as in its native-market, France.
Unique expertise: Kraken Technology (autonomous vessels) signed contracts with 5 different NATO entities, in its first year of commercialization, as it faces almost no competition on a new, but urgent need for Navy operations.
Recent dynamics with the US increasingly being perceived as a less reliable ally will pressure Europe’s MoDs to favor non-US alternatives.
Myth #4: “Barriers to entry are too high, and government sales cycles are painfully long.”
I used to assume that if you wanted to sell into defense, you were doomed to a years-long ordeal of government procurement labyrinths. Indeed, we do see that founding teams need to hire insiders who have the network and the skills to navigate these organizations.
But it is wrong to assume that there is only one door for each DoD, and that governments are the only starting point. Just as SaaS B2B can often target tech companies to scale faster in their early days, defense startups can address smaller, more agile types of customers:
Sell to defense contractors as a component supplier before you pitch the big government contract yourself. These can be meaningful, like Helsing working with Saab for multi-million dollar contracts.
Target Special forces: these elite units 1) offer shorter sales cycles, 2) are early-adopters of new technology, and 3) are very respected in the militaries and serve as examples of excellence. Again, these contracts can be massive as seen with Anduril’s $1bn contract with the US’s SOCOM.
Pilot projects can be both short and financially meaningful. You’re not just installing a “demo”; you’re effectively delivering real value — and if it sticks, the multi-year contract is often the next step — something we’re seeing with Comand AI (combat management systems).
In many ways, it’s reminiscent of the early SaaS era, where the real trick was to reduce friction at the start and then scale up once you had proven ROI. Yes, the stakes are higher in defense, but the pattern is surprisingly similar.
Myth #5: “Defense is only mostly about hardware.”
While websites of defense startups show images of sleek drones, the reality is that most of the value lies in the software.
As defense moves towards the need of fleets of thousands of attritable and autonomous equipment, we’re seeing the rise of vertically integrated modern primes owning the entire stack and able to constantly upgrade the intelligence capabilities of the systems.
This offers much more comfortable unit economics: whereas traditional primes often operate at 15–25% gross margins, newly minted defense primes that blend custom software and autonomy can see 60–70% margins.
We’ve seen this approach become dominant in automotive, in space and with satellites, and now it is happening in defense. Even software-only players like Helsing are pivoting to this model with its newly released autonomous drones in Ukraine.
And now, what’s next?
As NATO budgets are set to double over the next 5 years, the need to invest in our future capabilities has never been more urgent. The good news is that new LPAs are starting to drop their restrictives clauses and opportunities abound and the next big things might come from:
Modern primes — Startups are evolving from specialized end-point providers into full-fledged primes, challenging the old guard. We see that MoDs prefer to buy a comprehensive system and work closely with trusted partners, so the ability to offer a catalog of products is an edge.
Trifecta of ground, air and sea autonomy — land vehicles like Overland AI (US) or Arx (EU), maritime vessels like Kraken (EU) or Saronic (US) are still early and a lot can be done.
Electronic warfare and RF in general — In an environment filled with unmanned sensors and platforms, reliable communication and situational awareness are the lifeblood of successful missions, and I admire what US startups like Epirus, CX2 and Chaos Industries are building in the space.
Production capabilities are still the main bottleneck — Demand far outstrips supply for advanced drones, sensors, and autonomous systems, and we can expect to see companies mastering in-house manufacturing capabilities at scale reap the rewards (like SpaceX did with the space boom).
On top of transversality between domains (air, land, navy, space, information), we’ll see more transversality between sectors (defense, security, intelligence), with non-defense native players winning contracts with MoDs.
To dive deeper, Musk’s recent interview at West Point is a must watch.