More and more bright, clever talent is entering defense tech, not least because of the leadership from the European Defense Tech Hub (EDTH): From hackathons and niche events to a prominent presence at major defense conferences, the defense tech Zeitenwende is real.
As a Ukrainian, I like seeing Europe finally waking up. As a defense tech investor, I want to use my voice to direct this incredible potential to solving challenges overlooked by the current market. There’s plenty of value to be created, plenty of impact still to be made in Ukraine, and plenty of room for the ecosystem to grow. Ultimately, this helps make Europe more secure.
Defense tech has already reached the stage where some verticals are getting overcrowded. The 2022-2023 surge brought a wave of companies that captured low-hanging fruit, particularly drones. Ukrainian drone production alone is expected to exceed 4 million units in 2025. These obvious niches are now saturated, and teams need to look elsewhere: toward gaps that remain unsolved and can have a meaningful impact in Ukraine.
At D3 VC, we have looked at 700+ startups and scaleups from Ukraine and the EU. Our focus is on applied defense tech—solutions that could be prototyped and tested in combat within 12-18 months.
In the EU, 31% are building drones (in all domains), 22% are pursuing other hardware solutions, 13% are focusing on defense-related software, 9% are trying to make drones autonomous, and only 7% are paying attention to opportunities in the electromagnetic spectrum (EMS).
Two verticals stand out: autonomy and EMS startups developing radiofrequency (RF) technology. While they rank fourth and fifth by company count in the EU, their importance becomes clear when you look at Ukraine. Ukrainian startups—building under fire—prioritize them differently: 11% focus on autonomy (top-3), and 10% on EMS (top-4).
Things get even more curious when we compare investor preferences, i.e., the total VC capital invested in startups between 2023 and Q3 2025. According to the Ukrainian Council of Defence Industry (UCDI), autonomy startups accounted for 29% of all funding, making it the largest vertical spanning multiple operational domains. EMS startups raised 12%.
The signal is clear: autonomy and RF are extremely important for the future of defense.
So What Should Startups Actually Be Building?
Drones
Let’s address the elephant in the room: drones. Drones truly spearheaded the defense tech revolution in 2022. UAVs have a relatively low entry barrier, which has brought early talent to the industry and generated meaningful traction in Ukraine in a relatively short time.
This has made UAVs a commodity. Today, building UAVs is about manufacturing optimisation and supply chain efficiency. This is the territory for private equity and experienced industrialists.
But there is still room for innovation.
The proliferation of UAVs has everyone craving counter-drone solutions. And C-UAS is still not solved, even at the effector level. Think about the different use cases: kinetic anti-FPV, anti-fiber, anti-Mavic, anti-ISR, anti-Shahed. These are separate CONOPS, each requiring different solutions.
A critical consideration is remaining attritable against specific targets and not losing to the adversary on economics. The 2025 incursions in Poland give a sobering case study. The key challenge: perfect radar coverage is a utopia, so figure out sensing and targeting without that assumption.
Here’s the fundamental disconnect many drone startups miss: they like to build drones as devices, as platforms. Militaries want capabilities, which is their sophisticated way of calling missions. The military doesn’t need “a reconnaissance drone with 90-minute flight time.” They need “persistent ISR over this grid square with real-time targeting feeds to strike units.” That’s a capability. That’s a mission.
So if you have a strong urge to build innovative drones, build missions that integrate different drone types, hardware, and software. Build the orchestration layer that integrates three different UAVs, a ground station, and a targeting algorithm into a single unified capability. Think counter-battery missions coordinating ISR drones, loitering munitions, and EW assets. Think defensive perimeter operations, managing swarms of heterogeneous sensors.
It’s worth going beyond drones entirely. The advancement of commercial tools has started to blur the lines between drones and missiles. The next logical step for drone enthusiasts is to go for low-cost, mass-produced missiles.
Autonomy
Making hardware platforms autonomous and software-enabled is the core premise of modern defense tech. But despite numerous promises and claims, we’re not there yet.
Drones aren’t truly autonomous at scale. True drone autonomy means terminal guidance without human input, navigation beyond GNSS, edge-based decision-making, and coordinated swarming. Over a hundred teams across Ukraine and NATO are trying to build navigation and terminal guidance solutions based on “conventional” optical sensors. Very few are looking at alternative sensors, such as magnetic fields, celestial, or RF.
Equally, numerous independent teams and drone makers are building swarming capabilities. Swarming is a big buzzword. But if everyone tries to build a swarm, someone needs to start working on counter-swarm systems already now. Could this be swarms of interceptors?
RF
Even the pursuit of autonomous drones will not make the electromagnetic spectrum less relevant. Besides, drones will need humans in the loop (and humans would always prefer to maintain connectivity with their gear). Building unjammable, compact, and attritable radios remains a critical task.
As long as drones, humans, and other battlefield assets communicate, there will be a market for both jamming and RF intelligence. Jamming needs to be intelligent, power-efficient, low-footprint EW. We cannot continue the rat race of drones changing frequency bands and jammers with more amps and more power—this becomes unproductive for EW.
ELINT needs BVLOS targeting capability. (Yes, you need to learn the acronyms.)
One major gap in any counter-UAS operations is detecting incoming threats. Radars of prior generations are no longer suitable for the job: they suffer from resolution and range issues, are expensive, and very visible to the adversary. We need novel radar architectures. And there will be no single radar to rule them all; we need multiple radar types tailored to specific CONOPS.
Software
Palantir is no spring chicken anymore. It is becoming really corporate, and those are always ripe for disruption. Ukraine’s experience shows that its solutions are often suboptimal. Pick a use case and build an excellent product around it. And if someone tells you “Palantir is doing it”: make it 10x better. They’re beatable.
Data is a powerful weapon. A lot of power is hidden in unconventional, unexpected data points and sources. Get creative with using various data sources and points to gain insights relevant not just to “national security” but also to military applications.
These gaps are critical. They will have a meaningful impact in Ukraine today and be foundational to any European ReArm built around modern technology. Go after these blind spots now – and you can capture the same opportunity window the drone companies had in 2022-2023.
Pingback:Дорозвідка #38: Sub Sea Baby, Бабенко, “Сапсан” і Верховодов