The defense-tech boom has a visible part and an invisible part. The visible part is drones, satellites, sensors, the occasional autonomous boat, things you can photograph and put on a magazine cover. The invisible part is the software that sits above all of it and answers the only question that matters in a war, which is: given everything we now know, what should we do?
That layer is called command and control, or C2, and this week a Paris startup named Comand AI raised €32 million to own it. The round was led by Blossom Capital, with a strategic check from Sweden’s defense prime Saab and renewed backing from Expeditions. It was announced at Eurosatory, one of the largest defense trade shows on earth, which is the industry’s way of making sure the right generals were in the room when the number dropped.
€32 million is, by the standards of this sector, not very much money. Helsing has raised at an $18 billion valuation. Anduril recently closed a round that valued it at $61 billion. Against those figures a Series A is a rounding error. But Comand AI is not trying to out-raise the giants. It is trying to occupy the one piece of territory the giants have left strangely under-defended.
Why the decision layer is the prize
Here is the logic, and it is good logic. You can have the best drones and the best sensors in the world, and they will still drown you. A modern battlefield generates an absurd volume of data, more feeds, tracks, and alerts than any human staff can read, let alone reason about. The bottleneck is no longer collecting information. It’s deciding.
Comand AI’s product, Prevail, is built around what the company calls a “digital command staff”: specialized AI agents, each handling one function a human officer would, mission analysis, logistics, after-action review, with a person still making every final call. Founder Loïc Mougeolle’s pitch is that “we are moving from a battlefield governed by words to a battlefield governed by mathematics,” and that with Prevail a single laptop becomes a command node that can coordinate manned units and drone swarms at once.
Build the drone and you’ve built a tool. Build the thing that decides what the drone does, and you’ve built the place every other tool has to plug into. One of those is a feature. The other is a moat.
You can see why Saab cared enough to invest rather than build. Saab makes hardware, including GlobalEye, its airborne early-warning aircraft, and rather than write the brains itself, it is backing a startup to do it and integrating Prevail into a next-generation C2 ecosystem. French AI, Swedish airframes, a deliberately European stack. That pairing is the whole story, and it is aimed squarely at an awkward fact.
The awkward fact
Europe’s most prominent command-and-control deployment today is American. NATO has rolled out Palantir’s Maven, and European officials have said out loud, more than once, that there isn’t a homegrown alternative. Comand AI is positioning itself, with Saab’s hardware and a tricolor’s worth of sovereignty messaging, as exactly that alternative. The timing helps: Palantir’s standing in Europe has wobbled, and “do you really want your kill chain running on someone else’s software” is a question that sells itself in the current climate.
The thing Comand AI has that the giants can’t claim on this specific turf is reps. Prevail has been deployed with operational units in France, Germany, Ukraine and other allied nations over the past year, actual battlefield feedback at the decision layer, not a demo in a hangar. The company is vague about how extensive those deployments are, which is both standard operational discretion and a reminder to discount the marketing accordingly.
The catch, because there’s always a catch
The field is crowded with names that are far better capitalized. Anduril has Lattice. Palantir has Maven. Helsing has Altra. Shield AI has Commander. They are all circling the same prize, and €32 million does not buy you many engineers relative to companies valued in the tens of billions. Comand AI’s bet is that focus beats scale here, that being the company whose entire reason to exist is the command layer matters more than being a giant for whom it’s one product line among many.
That’s a real bet, not a sure thing. The decision layer is sticky once it’s installed, which is exactly why everyone wants it and why incumbents will fight hard to keep their hardware from depending on someone else’s brain. The funding will push Prevail across NATO markets and into new domains, air is underway, maritime is promised before year-end.
The honest summary is that nobody knows yet who controls the decision layer when the current defense-AI rush settles. What’s clear is that a Paris startup just convinced a Swedish prime and a serious venture fund that the answer might be a piece of software running on a laptop, made in Europe, that no American company has to be asked permission to use. In this market, that sentence alone is worth €32 million.