How Startups Are Reshaping European Defence

The war in Ukraine has shown that military advantage increasingly depends on whether new systems can move from development into field use fast enough to adapt to changing conditions. A new generation of European startups is emerging around this faster model of defence innovation. This landscape looks at the companies reshaping the European defence technology stack and the capabilities they are building for a more contested security environment.

Category: Battlefield AI, C2 & Sensor Fusion

What is this category about?

• The command layer that turns battlefield information into coordinated action. Modern military operations generate a constant flow of data from soldiers, vehicles, drones, sensors, intelligence systems, and external sources. The role of this category is to bring that information together so commanders can understand the situation and decide what should happen next.

• The first major problem is fragmentation. Military information is often distributed across systems that were developed by different suppliers and were never designed to communicate with one another. Operators may therefore have access to valuable data without being able to combine it into a single operational picture.

• The second problem is the speed of decision making. Traditional command workflows depend heavily on manual analysis and information moving through several organizational layers. This can create delays between detecting a change on the battlefield and responding to it, especially when the volume of incoming information exceeds what a human team can process.

• The category also addresses the growing complexity of coordinating autonomous systems. A military unit may deploy many unmanned platforms at the same time, but each one can require its own interface and operator. Command software helps transform separate machines into a coordinated capability that can be supervised through a smaller number of human decision-makers.

• Reliability in contested environments remains the defining constraint. The software must continue functioning when communications are disrupted and data is incomplete. It must also preserve human control over consequential decisions, which makes traceability and predictable behaviour more important than simply producing the most sophisticated AI output.

What do products in this category do?

• The first product approach creates an AI native workspace for headquarters. Comand AI’s Prevail supports military planning and preserves the context created during earlier missions. Hadean’s dominAI adds simulation to the command workflow so users can compare possible courses of action before choosing how to respond.

• The second approach provides an interoperability layer for military systems. Intelic’s Nexus connects unmanned platforms from different manufacturers through one command interface. Project Q’s Hydris provides an open integration layer that can connect a sensor with an existing command application without forcing the customer to replace either system. Arondite’s Cobalt applies the same principle to more complex human-machine operations.

• The third approach helps intelligence teams turn fragmented information into a common picture. Adarga’s Vantage processes large volumes of internal and open source information (OSINT) so analysts can identify relevant relationships more quickly. SensusQ’s Winning Mind organizes operational intelligence around a shared picture that commanders and intelligence teams can continuously update.

• The fourth approach brings situational awareness closer to the tactical user. Vegvisir uses mixed reality interfaces to give vehicle crews and remote operators a clearer view of their surroundings. Its Virtual Command Station extends the same interface to commanders supervising several platforms. Unbound Autonomy begins with wearable blue force tracking and adds acoustic intelligence that can identify critical battlefield sounds.

• The fifth approach combines command software with the systems that execute the mission (platformization). Helsing has expanded from AI software into complete autonomous systems, while Delian connects its command layer to its own surveillance and electronic-warfare capabilities. These companies are trying to control more of the operational chain instead of remaining a software component inside another vendor’s product.

What types of customers are using these products?

• National armed forces are the most important direct customers. The Norwegian Army selected Six Robotics for a multi-unit pilot and later received the first Valkyrie swarm. The Royal Netherlands Army used Intelic during a live fire exercise, while the Bundeswehr tested Comand AI during military planning. Vegvisir has also supplied its technology for an Australian Army vehicle trial.

• Defence ministries increasingly buy the software as strategic infrastructure rather than as a limited experiment. The Dutch Ministry of Defence entered a partnership worth more than €30 million with Intelic to make Nexus the foundation of its unmanned systems ecosystem. Adarga secured an agreement worth up to £12 million with the UK Ministry of Defence. The German Ministry of Defence selected Helsing and Saab for the Eurofighter electronic warfare upgrade.

• Established defence companies provide another route into military programmes. Saab integrates Helsing’s Cirra software into its Eurofighter sensor suite. Airbus contributes the aircraft integration layer. Arondite works with Babcock on maritime autonomy, while Vegvisir is integrating its interface into DOK-ING robotic platforms.

• Exercises and multinational programmes often provide the first environment where the products prove interoperability. Intelic demonstrated Nexus in a NATO aligned experiment after earlier work with the Dutch armed forces. SensusQ contributed its intelligence-fusion capability to the British Army’s Project ASGARD. Hadean uses its partnership with Palantir to bring simulation and command products into the wider UK Armed Forces environment.

• Several companies enter through public safety operations before expanding further into defence. SensusQ has supported the Estonian Police during a large public event. Delian supplied autonomous monitoring systems to the Hellenic Ministry of Climate Crisis and Civil Protection. These deployments allow the companies to prove that the same command and sensor-fusion infrastructure can support civilian agencies operating under time pressure.

What are the major trends shaping this category?

• European defence spending is creating a much larger market for command software. European NATO members and Canada increased their combined defence expenditure by nearly 20% in real terms during 2025. NATO members have also committed to reach a broader 5% defence and security investment target by 2035. The European Union’s Readiness 2030 programme is pushing additional capital toward capability gaps that cannot be solved by purchasing more traditional equipment alone.

• Interoperability is becoming a strategic requirement rather than a technical feature. NATO’s Data Strategy calls for common standards that allow information to move securely across allied forces. Its Federated Mission Networking framework pursues the same objective at the operational level. Ukraine’s experience has shown the cost of deploying rapidly developed systems that use incompatible command protocols, which is increasing demand for open integration layers.

• Defence procurement is moving closer to continuous product development. Ukraine has created direct feedback loops between frontline units and engineers, allowing products to change as battlefield requirements evolve. BraveTech EU is bringing this approach into European defence innovation, while NATO’s Rapid Adoption Action Plan aims to shorten the path between testing and procurement. This benefits startups that deploy engineers alongside military users rather than delivering a fixed product after a long development programme.

• The proliferation of autonomous systems is making the coordination layer more valuable. A military may be able to purchase drones from many suppliers, but it cannot operate them effectively if each platform requires a separate control system. NATO now treats drones and autonomous systems as technologies that are reshaping the capabilities required for deterrence. The strategic bottleneck is therefore shifting from access to hardware toward the software that coordinates it.

• Responsible AI requirements will increasingly influence which products reach deployment. NATO’s revised AI strategy emphasizes testing and traceability before AI is integrated into allied capabilities. It also calls for greater interoperability between military AI systems. Companies will therefore need to prove that their software behaves reliably under operational conditions while preserving human control over consequential decisions.

How does the funding environment look?

• The funding market follows an extreme power law distribution. Helsing had already raised €1.36 billion through its Series D before announcing a $1.8 billion Series E in July 2026. The latest round valued the company at $18 billion and placed it on a scale that no other European startup in the category currently approaches.

• A second group of companies is beginning to raise larger software rounds. Comand AI’s €32 million Series A followed deployments with operational units across several allied countries. The round suggests that investors now see command software as a standalone platform opportunity rather than a specialist feature that must be sold through a traditional defence contractor.

• Most companies remain much earlier in their financing cycle. Six Robotics raised €12 million after the Norwegian Army had already selected and received its product. Project Q and Intelic initially raised smaller rounds while using partnerships and military exercises to establish credibility. The pattern suggests that operational validation often arrives before a large institutional financing.

• Capital is increasingly used to move beyond pure software. Helsing acquired Grob Aircraft and began developing complete autonomous platforms. Delian is following a similar direction by building its own systems around a shared autonomy and command architecture. Investors appear willing to support this expansion when software ownership gives the company a credible path toward controlling the complete capability.

• Strategic investors can provide market access as well as financing. Saab invested in Helsing before deepening its product partnership around the Eurofighter. The company later participated strategically in Comand AI’s Series A. In a market where integration with an existing military programme can determine commercial success, this relationship may be more valuable than the capital alone.

Category #2: Autonomous Platforms & Robotics

What is this category about?

• This category covers physical systems that can perform military tasks with less direct human involvement. These platforms operate in environments where sending personnel is dangerous, slow, or operationally inefficient. Their value comes from moving the point of risk away from the soldier rather than simply replacing a crewed vehicle with a remote controlled one.

• The first problem is exposure to danger. Reconnaissance missions can reveal a soldier’s position, while logistics and casualty evacuation require personnel to cross areas under fire. Robotic systems allow forces to maintain these activities while reducing the number of people placed in vulnerable positions.

• A second challenge is that many unmanned systems still require too much human attention. A drone may remove the pilot from the aircraft but continue to depend on a dedicated operator throughout the mission. Greater autonomy allows one person to supervise several systems and intervene only when the situation requires judgement.

• The systems must also remain useful when the surrounding infrastructure stops working. Satellite navigation can be disrupted and communication links may disappear without warning. A platform therefore needs enough local intelligence to continue the mission safely rather than becoming ineffective as soon as it loses contact with its operator.

• Another challenge is turning prototypes into equipment that armed forces can deploy at scale. Military users need systems that can be repaired in the field and adapted without redesigning the entire platform. This makes production capacity and modularity as important as the technical performance of the first vehicle.

What do products in this category do?

• One approach builds aerial platforms around a specific operational mission. For example, Skyeton and Ukrspecsystems focus on long range reconnaissance that can continue in contested environments. Buntar Aerospace places similar emphasis on electronic warfare resistance, while Elistair uses a tethered architecture to provide continuous observation from a fixed position. At the larger end of the market, Dronamics and Acodyne apply autonomous aircraft to logistics missions.

• A second approach removes soldiers from dangerous ground tasks. Ratel Robotics and Tencore initially focused on carrying supplies or evacuating casualties. Roboneers and Frontline Robotics extend the same platform concept toward reconnaissance and remotely operated fire support. These products often begin with a narrow mission before expanding into a modular ground robot family.

• Maritime companies apply autonomy where crewed operations are expensive or hazardous. Maritime Robotics develops surface vessels that can carry and coordinate underwater systems for mine clearance. SEABER takes the opposite approach with compact autonomous underwater vehicles. Kraken Technology Group builds faster surface platforms for naval surveillance and force protection, while Hard Cat Drones focuses on shallow waters where larger vessels cannot operate effectively.

• Some companies are building persistent systems rather than individual vehicles. RIFT combines long range aircraft with autonomous launch stations so missions can be initiated from a remote operations centre. Elistair’s Khronos provides a similar push button model for continuous surveillance. This reduces the need to deploy a specialist team each time an aircraft is launched.

• The most ambitious companies are adding a software layer that coordinates the complete robotic fleet (platformization). Quantum Systems uses MOSAIC UXS to control systems from different manufacturers through one interface. Milrem Robotics developed ARCOS for collaborative ground operations, while ARX Robotics uses Mithra OS to add connectivity and autonomy to both new robots and existing military vehicles. The product therefore becomes an operating system for unmanned operations rather than a single piece of hardware.

What types of customers are using these products?

• National armed forces remain the most important direct buyers. The U.S. Army selected Quantum Systems’ Vector AI, while Lithuania’s Ministry of National Defence ordered FPV systems from Granta Autonomy. The Royal Navy awarded Kraken Technology Group a contract for 20 surface vessels, and a Dutch-led initiative ordered more than 150 THeMIS vehicles from Milrem Robotics for Ukraine.

• Frontline military units also purchase or receive systems directly. Tencore says its ground robots are now used by more than 50 Ukrainian brigades. Buntar Aerospace has supplied its Skyhopper system to Ukrainian military intelligence, while Ukrspecsystems has published deployments with the 73rd Naval Special Operations Center. These customers provide unusually fast feedback because the operators using the product are often close to the manufacturer.

• Defence primes increasingly act as integration and distribution partners. Dassault Aviation is working with Harmattan AI on autonomy for future combat-air systems. Rheinmetall is helping ARX Robotics and Destinus move into larger industrial programmes. HENSOLDT provides the sensor layer for Dronamics’ defence platform, while Anduril gives Kraken a route into U.S. production.

• Security agencies use many of the same platforms outside conventional military operations. Evolve Dynamics has run trials with UK Border Force, and TEKEVER has sold systems to Spain’s Ministry of Interior. Skyeton delivered a radiation  monitoring configuration to Ukraine’s State Emergency Service. These deployments usually emphasize persistent observation rather than direct battlefield activity.

• Industrial operators provide an important dual-use market. Maritime Robotics has delivered vessels to DEME and the Flanders Marine Institute. RIFT has demonstrated remote monitoring for NaTran and SFDM, while SEABER works with Xylem on water-quality monitoring. These customers help companies validate autonomy in demanding environments without depending entirely on defence procurement.

What are the major trends shaping this category?

• Unmanned systems are moving from specialist equipment to mass consumed military assets. Ukraine planned to procure 4.5 million FPV drones during 2025 after purchasing more than 1.5 million the previous year. This scale changes the economics of the market because production cost and replenishment speed become central product requirements.

• Procurement is moving closer to the operational user. Ukraine’s DOT-Chain Defence marketplace allows units to choose the systems that match their immediate needs while the state manages contracting and delivery. By November 2025, the platform had delivered more than 100,000 drones with an average delivery time of seven days. This model rewards companies that respond quickly to battlefield feedback rather than waiting for a centralized multi-year programme.

• Autonomy in degraded environments is becoming the technical benchmark. NATO’s programmes increasingly test whether systems can navigate and cooperate when satellite positioning is unavailable. The central question is no longer whether a vehicle can follow a preplanned route under ideal conditions, but whether it can continue operating when communication and navigation are actively contested.

• Military value is shifting from individual vehicles toward coordinated groups of systems. The European Defence Agency’s autonomy plan places cooperative systems and manned / unmanned teaming at the centre of future capability development. A platform becomes more valuable when it can share tasks with other machines and fit into the workflow of an existing military unit.

How does the funding environment look?

• Capital is highly concentrated around a small number of emerging platform leaders. Quantum Systems’ $1.2 billion Series D created an $8 billion company, while Harmattan AI raised a $200 million Series B less than two years after being founded. Most companies in the landscape remain at seed stage or have never announced an institutional equity round.

• Large rounds tend to follow evidence that the product has moved beyond experimentation. ARX Robotics raised its €31 million Series A after deploying systems with several European armed forces and establishing a production roadmap. Harmattan AI’s Series B followed programmes of record with the French and British defence ministries. Investors appear more willing to finance manufacturing once procurement traction reduces the risk that a platform will remain a demonstration.

• Strategic investors increasingly connect financing with market access. Dassault Aviation’s investment in Harmattan AI came with a combat-aviation partnership. TEKEVER brought the NATO Innovation Fund and national security investors into its €70 million round as it expanded production and entered additional markets. For autonomous platform companies, an investor that helps the product reach an established military programme may be more valuable than a conventional financial backer.

• Public funding remains essential for projects that require several companies to cooperate. The European Defence Fund committed nearly €50 million to the Milrem-led iMUGS2 consortium because no single startup can independently establish a European standard for ground robotics. The European Innovation Council has also committed significant capital to Dronamics. These programmes finance shared technology and interoperability work that is difficult to support through standard venture returns.

• The capital stack is expanding beyond venture equity as the companies industrialize. Destinus secured a €50 million commercial bank facility after raising equity and convertible financing. This is an important signal because large factories and testing infrastructure eventually require forms of capital that are better suited to physical assets than venture funding.