Europe is buying a new class of weapon with an old rulebook. Christine Thaagaard of Robotto, a Danish autonomy company, argues that one-way attack drones have more in common with ammunition than with the reconnaissance drones they are usually bought alongside. Treat them that way, and the Nordics stand to gain: software over airframe, and cost per effect.
Ask a European procurement officer to buy a reconnaissance drone and there is a well-worn path to follow. There is a category, a procurement code, a set of standards, a doctrine that says what the thing is for. Ask the same officer to buy a one-way attack drone, a loitering munition, and much of that falls away. The weapon has arrived faster than the rulebook meant to govern it.
"The drone space today is pre-doctrine," says Christine Thaagaard, who leads government relations outside Denmark for the autonomy company Robotto. "There are only draft doctrines floating around. And when you don't have doctrine, you don't have a direct way of procuring and adopting the technology."
Robotto has come through NATO's DIANA accelerator and is one of ten companies picked for the European Defence Agency's Sentinel Strike Challenge, a prize contest set up to give defence ministries hard performance data on loitering munitions rather than to buy any of them. The question of how to buy them is exactly the one Thaagaard thinks is unsettled.
In practice, the burden lands on procurement officers. They largely source what they are told to source, Thaagaard points out, rather than deciding for themselves which capabilities to field. When a capability is pre-doctrine, there is often no line item to buy it against. The procurement catalogues, she notes, carry no dedicated code for one-way effectors. Instead they are filed under the broader drone categories, which means they are bought, and evaluated, as drones.
That, in her view, is the crux. A reconnaissance drone and a strike drone deliver very different capabilities, and call for equally different evaluation. ISR drones are a mature, widely adopted technology, and Thaagaard credits that part of the market with doing "a wonderful job" of interoperability, range and finding its place. Strike drones are something else. They are coming in to take over part of the work once done by howitzers and other artillery, and she argues they should be procured much as artillery shells are: as a consumable, designed and produced at scale, where the value lies not in the airframe but in what it can actually do.
Robotto's answer is to separate the software from the airframe, and Thaagaard reaches for an everyday comparison.
"You don't get one driver's licence for a Mercedes and another for a Suzuki. You get one licence so you can drive them all," she says. "That's very much how we see it. You shouldn't procure just two suppliers. You should procure a system that has suppliers underneath it, so that you can distribute that risk."
The interview is continued below.
In practice that means an edge-level control layer that sits between command-and-control systems such as Delta and ATAK and the drone itself, giving it navigation when GPS is jammed or denied and guiding it onto the target, the same software whether it is flying a quadcopter, a fixed wing or something larger, and, Robotto says, combat-proven on all of them. The same approach, she argues, could extend to any one-time-use UAV, counter-drone interceptors included, which are likewise flown once to achieve an effect and, in her reading, best procured as a software-led system rather than one closed platform at a time.
Buy the new class with the old template, she warns, and the risks compound. A state that evaluates strike drones the way it evaluates ISR drones tends to land on one or two suppliers, each running its own closed system.
"Anytime you replace one of them, you have to retrain for that new drone," Thaagaard says. "So you're creating a constant need to support these companies long-term, because if you have to replace them, you have to retrain your entire force." The buyer also ends up carrying more of the manufacturing risk, and in a war that turned hot, she notes, a single supplier's supply chain could become the weak link.
The point of a common software layer, in this reading, is not only resilience but variety. Thaagaard credits Ukraine's drone programme with succeeding precisely because it is diverse, running thirty-odd types of airframe rather than two, which keeps an adversary guessing about what comes next. She sees the clearest sign of where procurement is heading in the United States, where the Air Force chose an autonomy software, Shield AI's Hivemind, and had it integrated onto a separate uncrewed airframe: the software and the hardware selected on their own merits and asked to work together.
If airframes are consumables, then the way to compare them has to change too. Thaagaard's preferred measure is cost per effect: the airframes consumed, plus the licence, the updates, the ground equipment and the training, divided by the number of targets actually struck, rather than the sticker price of the drone. It is, she argues, the artillery way of thinking, turned onto a new weapon.
The difference that makes is not small. FPV drones flown without autonomy, she says, land somewhere around a 30 to 40 per cent hit rate. At a recent operational experimentation in Czechia, Robotto's software, a legacy build the team had not touched in a year and a half, came in at 80 per cent, a figure she is quick to say the company holds itself above on its newer fixed-wing work. Even that gap, she notes, is a large improvement in cost per effect.
None of this, Thaagaard is careful to say, is a complaint about the people doing the buying. Given a pre-doctrine field with no settled standards to lean on, she thinks Europe's procurement bodies, Denmark's own among them, are "doing a fabulous job of trying to figure it out." Her argument is aimed one step upstream: at the question itself. Take the lessons coming out of Ukraine, and the software-first way the EDA's own challenge judged these systems, and translate them into something that fits NATO and the Nordics. Settle what a strike drone is, and how to measure it, and the procurement follows.

