More than 80% of critical infrastructure sites have no affordable radar watching the air above them. A drone that costs a few hundred euros can shut an airport, a port or a substation. The sensors built to spot it were priced for a different era, and Gallium Defence just raised €7.1M to attack that price.
Gallium claims 10 km detection from commercial 5G parts at a fraction of legacy radar cost.
The open questions are weather, false alarms and whether primes integrate a cheap sensor.
The round closed in June and was announced on 2 November during European Defence Week. Keen Venture Partners and 360 Capital co-led, with Omnes Capital and Polytechnique Ventures participating. Gallium says orders for its first radar open in the first quarter of 2027.
Detection is the constraint
A low-cost FPV drone can cost a few hundred euros. A high-end 3D radar can cost as much as the asset it protects. That mismatch pushed buyers toward a small number of high-value sites and left the long tail of infrastructure exposed.
Eric Fournier, Gallium's chief executive, frames the design brief as an operator question: how do you meet 80% of the need at 20% of the price? His answer is a radar assembled from parts that already ship in telecom volumes.
Critical infrastructure without affordable radar
Gallium and its lead investor use this figure to size the detection gap. · Keen Venture Partners, 2026
A radar built from telecom parts
The GDS-10 is an X-band active electronically scanned array radar. It carries roughly 512 antenna elements and steers its beam electronically, so the antenna has no moving parts. Gallium says it cut antenna unit costs by more than tenfold by adapting gallium-based amplifiers from 5G supply chains.
Small savings compound across the array. "Each time you're saving a bit of money on one antenna, it's multiplied by 510 times," Fournier said.
The processing chain runs on embedded NVIDIA compute. Gallium claims more than 100 times the processing capacity of some conventional radar designs, plus an optical link adapted from telecom switches. The software-defined architecture lets the company update detection and classification models without changing hardware.
A radar that improves after it ships behaves differently from one frozen at the factory. That is the commercial promise behind the seed round, and the hardest part to verify before customers deploy it.
| Parameter | GDS-10 | High-end legacy radar |
|---|---|---|
| Target | Class 1 and FPV drones | Wider target set |
| Detection range | 10 km, company claim | Longer, classified |
| Coverage | 360° azimuth, −10° to +90° elevation | Varies by system |
| Accuracy | <2 m depth, <0.5° azimuth and elevation | Classified |
| Reported unit cost | About €500,000 target | Millions per unit |
| Interfaces | SAPIENT, ASTERIX, REST | Often proprietary |
Company specifications and Tech.eu reporting, 2026. Legacy costs are qualitative; no like-for-like test data is public.
The 10 km claim comes with a weather asterisk
Gallium says the GDS-10 detected Class 1 drones at 10 km. The test ran under controlled conditions, with favourable weather and known target locations. That is a real milestone for a company founded in April 2026. It is not yet an operational detection guarantee.
We would have preferred rain.— Eric Fournier, chief executive, Gallium Defence
Public materials do not include detection probability, false-alarm rate or performance in clutter. Those numbers decide whether operators trust a radar or switch it off after a week of nuisance alerts. Gallium plans monthly field demonstrations and says it updates the system every two weeks.
Attrition is part of the pitch. A cheap sensor can be lost without wrecking the budget, and several units can cover a site where one expensive radar used to sit. "That doesn't mean that it's not sturdy, but we accept attrition," Fournier said. Operators can also move units in sequence to keep coverage while one repositions.
Sell to the primes, do not become one
Gallium is not building a full counter-drone stack. It wants to sell the sensor to the companies that already integrate command-and-control software and effectors. The company says it is discussing five potential MOUs, and its interfaces follow SAPIENT and ASTERIX so detection data can plug into existing systems.
That choice keeps Gallium inside procurement channels it could not build alone. "We are going at the speed of prime supply, not that of DoD supply," Fournier said. The model also leaves room for white-labelling and licensed local assembly, with three radar modules built separately and combined at final assembly.
The field is crowded. Robin Radar has sold 100 IRIS detection radars to a defence ministry. Cambridge Aerospace raised $300M at a $3.4B valuation and is working on its own radar. Stendr raised $5.4M in April for AI-based detection. Above them sit Raytheon, Thales, Hensoldt and Saab, whose high-end systems reach further and cost far more.
Gallium's wedge is the space between short-range sensors and million-euro systems. Whether that space is a durable market or a temporary pricing hole depends on how fast incumbents cut their own costs.
Unit economics depend on volume
The reported target price is about €500,000 per unit, close to a short-range alternative with a longer reach. Gallium aims for €6M in orders in 2027, according to a briefing summary of the round. It has no revenue yet and about 15 employees, with plans to reach roughly 50.
The market backdrop supports the ambition. European counter-drone spending was estimated at $1.24B in 2025 and projected to reach $4.16B by 2030, a 27.5% annual growth rate, according to Tech Funding News. That forecast describes demand. It does not say who captures the margin.
A supplier that sells hundreds of cheap radars can win on volume and still lose on support. Field service, calibration and software updates are where hardware margins leak. Gallium has not published those unit economics, and at seed stage it does not need to. Its first customers will price that risk for it.
Data centres sit on the customer list for a reason. A large campus needs perimeter awareness, and its operator already buys cameras, access control and monitoring. A radar that costs less than the roof it protects can become another line item in the security budget rather than a capital project.
Cheap radar changes the deployment model
Traditional air-defence procurement treats a radar as a capital asset with a decade-long life. That logic produces few, large, heavily protected sites. A €500,000 sensor changes the math: it can be replaced, duplicated and moved.
The shift matters for airports, ports and energy operators that never had a radar budget. They can start with one unit covering an approach corridor and add units as the threat picture sharpens. Networked coverage then becomes a planning problem, not a single procurement decision.
It also changes training and maintenance. A software-defined radar can push new classification models to the whole fleet. That turns a hardware sale into a subscription-like relationship, if Gallium can support it. The company has not published service terms, and support is where cheap hardware often becomes expensive.
The export question
Dual-use radar sits inside export-control regimes. Gallium says the technology is subject to European and national rules, and its commercial focus is NATO and allied markets. A supplier model helps here. Primes already hold the licences and the customer relationships, so Gallium can ride their paperwork instead of building its own compliance team.
That dependence cuts both ways. If a prime controls the channel, it also controls the price and the roadmap. Gallium's white-label option gives partners a reason to carry the product, but it can dilute the brand and cap margin. The company will have to choose how much of its identity it is willing to rent out.
Why three consultants built hardware
The founding team came from advisory and procurement, not from a radar lab. Fournier spent 17 years advising defence and aerospace clients at Boston Consulting Group. Florian Lemoine, the chief technology officer, worked on armament programmes and later advised NATO. Robin Mandel, the chief operating officer, ran a 400-person engineering organisation at Thales before joining.
Their advantage is process knowledge. They knew which parts could be bought, which interfaces a prime would demand and how slowly a bespoke radar programme moves. They also recruited advisers who had run procurement from the inside, including former NATO and national defence leadership.
Gallium began testing a small-scale antenna in December 2025 and started two prototypes after incorporating in April 2026. The first open-air trials followed four months after the seed closed. Radar development usually runs in years. This pace is the company's strongest argument and its biggest execution risk.
What would change the math
Three data points matter more than the funding headline.
Rain and clutter trials with a published false-alarm rate
A named prime or infrastructure operator converting an MOU into an order
Export approvals and a licensed-assembly partner outside Europe
A second product line that reuses the same antenna and compute stack
Until those arrive, €7.1M buys time, a prototype and a manufacturing plan. That is a normal seed-stage position. It is also why the first field reports will matter more than the announcement.
As we wrote in October, the sensor layer under counter-drone systems is where the cost curve bends. Gallium is now testing whether a telecom supply chain can bend it faster than the incumbents.