A few years ago, defense was still a taboo word in most European venture conversations. Today, it’s one of the most dynamic corners of the ecosystem: European defense and security tech funding reached record levels in 2024 and remained elevated in 2025, with defense emerging as one of the fastest-growing segments of the European VC market.
Taken under the NATO Innovation Fund’s wider definition of “defense, security and resilience” – which spans robotics, biotech, quantum and dual-use AI – these technologies now make up roughly 10% of all VC investment in Europe.
But if this boom is here to stay, the real question becomes: how should investors approach this space?
This shift is (not) extraordinary
For decades, defense was viewed as politically sensitive and largely the preserve of major contractors; venture investors simply had little room, or mandate, to participate. That makes today’s record-funding levels stand out not just as a temporary reaction to a single conflict, but as a deeper change in mindset.
Even if the war ended tomorrow and demand for certain battlefield systems, especially drones, cooled, the strategic need for Europe to rearm and modernise its deterrence capabilities would continue. What seems truly permanent is not the spike in orders, but the long-term focus on defense innovation itself.
Governments are actively pushing innovation into the venture space, acting as both buyer of technology and provider of capital. That dual role — where the customer and the capital provider are often the same entity — is reshaping the rules of the game. And for investors who have been close to defense-related innovation through broader ecosystem relationships, this shift is particularly visible: the dynamics, incentives and timelines differ sharply from traditional venture spaces.
Defense autonomy demands a different investor logic
We cannot simply pour venture money into the sector and pretend it behaves like SaaS. Autonomous systems, safety protocols, procurement logic and dual-use dynamics follow rules that diverge fundamentally from conventional software investing.
For anyone who has spent time close to defense-related innovation, these differences are immediately visible: timelines are longer, stakeholders more complex and the feedback loop between deployment and iteration is shaped by regulation as much as by technology.
Ignoring that reality would be the fastest way to disappoint founders, misprice risk and under-build the technologies that actually matter.
What follows is a simple argument: autonomy in defense is really about safety; the investor’s math must be more sober; dual-use mostly works when sequenced; and the future of the category will be shaped by autonomy, data and counter-drone — not by pitch-deck slogans. In fact, many of the most capable companies in this space barely market themselves at all. Some generate more than €100 million in revenue without even maintaining a public LinkedIn presence, a reminder that defense innovation often grows through operational credibility, offline networks and institutional trust rather than branding or narrative polish. It is a category where traction and product speaks louder than storytelling.
The investor’s math in defense is different — it starts with the customer
Investing in defense technology follows a fundamentally different dynamic compared to conventional VC. The customer base is extremely narrow: in practice, states, governments, and ministries of defense — which makes the market behave very differently from civil tech. Unlike software or AI, where products can scale globally across thousands of customers, defense startups ultimately sell to a handful of highly specific buyers.
That reality fundamentally shapes how early-stage investing works. I believe that at pre-seed and seed stage it’s a smarter risk decision to back a company only when there is a clear indication that a relevant programme or buyer actually exists. Without that, even strong technology often has nowhere to go.
The “let’s see what the market becomes” approach, which can work in commercial technology and has helped create some of the most significant outliers, including Airbnb and Uber, where anticipating customer needs before they are fully formed can be a winning strategy, has a far narrower window in defense. It is not impossible, but it tends to require much more capital, far deeper access to end users and a longer timeline to validate. With only a small number of credible buyers, the risk profile is simply different.
It also makes long-shot outcomes harder to assume. The kind of >1000× returns that are theoretically possible in early-stage software are structurally rare in defense.
That doesn’t make the category unattractive, it often produces highly resilient companies with deep strategic importance. On the other hand, it does require investors to accept a different return profile: growth that comes in step-changes rather than network-driven compounding, fewer customers, and a higher dependence on procurement decisions.
Among CEE investors who have spent time analysing the space, this shift in expectations is increasingly recognised: return profiles tend to look closer to PE than classic VC — less extreme upside, but also a more grounded downside when technology is tied to concrete operational needs.
Procurement pipelines are slow (unless there is a war)
Defense startups operate under constraints that simply do not exist in conventional tech. Their primary customer is the states, not private industry, which means procurement generally takes longer and requires far deeper technical and security validation. There is the notable exception of Ukraine, where wartime conditions have forced procurement pipelines to be rebuilt for rapid testing and deployment.
Technologies must pass stringent testing, meet classification and safety requirements, and navigate export-control rules that restrict where and to whom they can sell. These limits make go-to-market fundamentally slower and more complex.
There is no fast “land-and-expand,” no freemium motion, and no shortcut through early adopters in the conventional tech sense; even though some units or specialised teams may pilot new systems early, these are structured trials, not the kind of rapid user-led adoption seen in commercial software.
Ukraine represents a uniquely open model, where innovation happens through tight, iterative collaboration between frontline teams and founders, including founders from outside the country working directly with Ukrainian units. The path to revenue is real, but it is built around programmes, compliance and long-term reliability. The latter is defined less by a fixed product and more by the ability to operate, produce and adapt under continuously changing conditions.
Autonomy starts with constraints
Autonomous systems are often framed as a leap into the future — drones and robotic platforms that can navigate, detect and act without a human in the loop. But in defense, autonomy is more about safety, reliability and control.
Across ISR, targeting and counter-UAS, autonomy only works when surrounded by a dense layer of guardrails: geofencing, the ability to identify friendly forces, recognition of civilian infrastructure, no-strike lists, fail-safe behaviours under jamming or signal loss. These mechanisms are not “add-ons.” They define whether a system is deployable at all.
This is why autonomy in defense is still one of the most technically demanding fields for early-stage companies. The bar for accuracy, redundancy and predictability is dramatically higher than in commercial robotics. In many cases, the engineering challenge is not “how autonomous can we make it,” but rather “how can we make sure it behaves exactly within the limits we set.”
The scale at which autonomy is now being deployed shows how quickly the field is maturing. Auterion’s recently reported plan to deliver 33,000 AI-enabled drone kits to Ukraine illustrates this shift: autonomy has moved beyond prototypes and into real, contested environments, where systems must handle jamming, degraded GPS and rapidly changing conditions.
For investors, this creates both opportunity and complexity. Autonomy is strategically important and has broad potential, but many safety-critical layers require deep engineering, long testing cycles and close work with end-users. In many ways, this is precisely the kind of frontier work early venture capital was built to support — yet today’s conventional VC has shifted toward lower-variance software bets, avoiding hardware or long R&D cycles. As a result, the most promising ideas in defense autonomy can look “too early” by current venture norms, even when the underlying need is immediate.
Dual-use works only when sequenced, not simultaneous
Dual-use has become one of the most frequently cited concepts in European defense tech, often framed as a way to expand markets and make technologies more “venture-compatible.” For years, this label also emerged because many funds simply could not or would not invest in defense directly, pushing founders to position their technology as dual-use. In practice, though, early-stage teams rarely succeed when they try to serve both civil and defense customers at the same time. The two markets operate on completely different rhythms: different sales motions, different requirements, different procurement pathways, and often entirely different decision-makers and stakeholders.
Most founders in defense come from military or security backgrounds. Their networks, understanding of requirements and product instincts are shaped around defense use cases. That is an advantage — but it also means civil-market expansion usually requires building a parallel team with different skills, channels and a GTM strategy. The reverse is equally true: teams that start with civil robotics, sensors or AI often underestimate how complex defense procurement and safety validation can be.
That is why dual-use works best as a sequence, not a starting point. Early focus brings clarity: one core customer group, one core problem, one clear route to market. Only later, once the product and the organisation are stable, it does make sense to split into civil and defense lines with dedicated teams for sales, product and compliance.
A good illustration is Vermeer: the company originally built its GPS-free visual positioning system for film and industrial use — enabling highly precise drone flight for cinematography and inspection. Only once that core technology was proven did the team extend it into defense applications, where GPS-denied and jammed environments make visual navigation a critical capability. Rather than a rigid sequencing model, this reflects a pattern we see frequently: early teams often explore both civil and defense use cases, then double down where the initial pull is strongest. In practice, many companies that start on the civil side find their strongest demand in defense — while the reverse rarely occurs.
Where is the next decade heading? Towards autonomy, data and counter-drone
From the conversations and signals across the ecosystem, three vectors appear to be gathering momentum: greater autonomy inside decision-making loops, the centrality of data and navigation resilience, and the rise of counter-drone and energy-bound systems.
Greater autonomy in decision-making loops
The technology for more autonomous systems already exists; what limits deployment today are ethical, legal and safety considerations rather than capability. Elements are already in use today, most visibly within modern air-defense systems. The implication is clear: more autonomy is coming, because contested environments require faster sensing, interpretation and reaction than humans alone can provide. It is also driven by the sheer scale of what governments need to protect, from airspace and borders to critical infrastructure, where relying on continuous human touchpoints is simply not feasible.
Data as the centre of gravity
Precision navigation, perception, cyber resilience and on-edge computation all depend on data pipelines that are robust under jamming, GNSS denial and degraded connectivity. In this sense, Europe’s gap in space infrastructure and advanced chips is a practical limitation on the performance of autonomous and semi-autonomous systems. Investments like Helsing’s €600m round show how strongly capital is concentrating around sovereign AI-defense software. Moreover, the Vermeer’s GPS-independent visual navigation highlights how critical alternative positioning systems will be in environments where GNSS cannot be relied on.
Counter-drone systems and energy-bounded directed energy
Drone warfare has fundamentally changed the battlefield, and Europe’s defense priorities with it. Counter-UAS solutions are becoming a core strategic capability. Directed-energy systems such as high-power lasers are promising but constrained by extreme energy requirements and thermal management. That makes innovations in batteries, charging systems and energy efficiency as important as the interceptors themselves.
Taken together, these vectors point to a future where defense innovation is less about single “breakthrough technologies” and more about integrated systems — autonomy, data and energy — that must work reliably under pressure.
Clarity is what European defense needs now
Europe finally has the talent, capital and political will to build serious defense technology. The question now is whether we can approach this sector with the clarity it demands — recognising its constraints, respecting its safety requirements, and investing with a long-term view rather than borrowing assumptions from commercial software.
If we can do that, Europe has a real opportunity not just to catch up, but to shape how the next generation of defense capabilities is built. And that requires an open discussion — between investors, founders, engineers and policymakers — about what it will take to get there.
This piece is meant as an invitation to that conversation.
