In eastern Ukraine last year, an interceptor built by a company barely a year old started knocking Shahed drones out of the sky. The system, called X-Lock, ran onboard computer vision that found the drone, held the track, and steered itself through the final seconds of flight while a human kept control of the engagement. The company is Alta Ares, founded in Paris in 2024, and last week it raised €50 million to turn that battlefield habit into an industrial one.

The Series A was led by Air Street Capital, the AI-first fund run by Nathan Benaich, with Cherry Ventures, OTB Ventures, and Harpoon Ventures alongside and existing backers renewing. It is a venture syndicate, not a defence-prime consortium, which is itself part of the story: the money that usually chases enterprise software is now buying interceptors.

The new arithmetic of air defence

The pitch starts with an economic problem, not a technical one. Modern aerial attacks no longer arrive as a handful of expensive missiles; they arrive as massed salvos of cheap drones, glide bombs, and cruise missiles, sometimes hundreds of airframes in a single night. Legacy air defence was built to answer scarce, fast threats with scarce, expensive interceptors, and that math has inverted. When a $50,000 drone draws a million-dollar missile, the defender can win every intercept and still lose the war on cost. The shield that works now has to be cheap enough to fire at scale, robust enough to keep working under jamming, and able to change as fast as the threat does.

Alta Ares builds across the whole chain rather than selling one box. Pixel Lock provides the onboard vision for detection, tracking, and terminal guidance while leaving the human in charge of the trigger; Gamma supports autonomous guidance and surveillance workflows. The two effectors are already in the field: X-Lock, with a roughly 15km radius, against Shahed-type one-way attack drones, and Black Bird, with a roughly 30km radius, against faster targets such as KH101 cruise missiles and FAB500 glide bombs. The company develops the models, the avionics, the guidance, and the manufacturing around one another, on the argument that in the physical world the advantage belongs to whoever owns the loop from data to deployment, not to whoever ships the cleverest model in isolation.

The receipts come from the field

What separates this from the long shelf of “AI for defence” pitches is that the system has been shot in anger. Interceptors carrying Pixel Lock began downing Shahed-type drones in 2025; the company has since demonstrated systems with NATO, which gave it an innovation award in March 2025, and tested Black Bird in arctic conditions with the Estonian Defence Forces. It says it is deployed across multiple theatres and has signed multi-million-euro contracts with around half a dozen countries across Europe, the Middle East, and Asia. In the weeks around the raise it added a counter-drone memorandum with Airbus Defence and Space and a French DGA partnership with the missile maker MBDA, aimed at producing several hundred interceptors a month from early 2027 through an entirely European supply chain.

The team reads the same way. Co-founders Hadrien Canter and Stanislas Walch have pulled engineers and operators from Anduril, Helsing, Palantir, Safran, Thales, MBDA, Embraer, and the French Army, and seated an advisory board that includes Philippe Lavigne, former Chief of Staff of the French Air and Space Force and former NATO Supreme Allied Commander Transformation, and General Corentin Lancrenon, a three-star French Army general. Production sits in Toulouse and in Ukraine, which is unusual, and deliberate: the company wants its factory next to the war it is learning from.

The crowded shield

The bear case is not subtle. Counter-drone is one of the most crowded corners of European defence right now, with Helsing, the German prime contractors, and a wave of newer entrants all chasing the same budgets, and a venture lead does not guarantee a startup can out-manufacture an incumbent that already runs assembly lines. Combat-zone deployment is hard for an outsider to audit, which means investors and ministries are partly taking the field record on the company’s word. And defence procurement, the actual gate to scale, still moves at a tempo built for a slower era; a decade-long programme assumes the threat will stand still long enough for the programme to arrive, which in drone warfare it never does.

The rebuttal is the field loop itself. A static dataset on a server is not the asset; the asset is the relationship that lets operators use the system, watch it fail, recover the evidence, and change the product before the next salvo. That is precisely what the old procurement model cannot absorb, and it is the reason the round exists. Europe has spent the past two years agreeing in principle that it must build its own defence capability rather than import it, the same argument running underneath Stark’s €500M defence-drone round; Alta Ares is one test of whether the continent will fund the building or just enlarge the budgets for buying.

Which returns to the arithmetic it started with. The threat is cheap, fast, and adaptive, and the only shield that survives it has to be the same. Alta Ares now has €50 million to prove a small French company can make interceptors at the price and pace the new math demands. The contracts are signed and the systems are flying. What it has to industrialise is the one thing that does not show up in a demo: the speed of its own next revision.