If you google drone swarms, the military dominates the conversation. But just because it’s the primary use case doesn’t mean it’s the most effective one. It’s more a reflection of drones’ origins as military tools. So when new developments around drones pop up, military applications are the default association.
But the thing is, the way this technology is usually used doesn’t work well in warfare. I’ll explain the three key reasons in the piece. Meanwhile, in humanitarian demining, this tech might be the only viable path to real automation. That means slashing costs by up to 70%, speeding up clearance, and most importantly, saving the lives of deminers.
In a true drone swarm, each unit operates autonomously using shared rules and capabilities. There’s no central controller. Drones make decisions locally based on their surroundings and the positions of nearby drones. Coordination comes from collective behavior, not direct communication. This allows the group to act as one.
How the military typically uses drone swarms
Military drone swarms usually refer to groups of FPV drones with the cheapest possible setup, used to attack by either crashing and exploding or dropping explosives. Why go cheap? Because these drones are likely to be destroyed, and armies don’t want to spend thousands on tech designed to take out a single target.
Each drone is operated by a human stationed safely behind the front line. The operator sees through the drone’s camera, controls it in real time, searches for a target, and directs the drone to strike.
Another growing use case involves drones flying pre-programmed GPS routes. These have a simple AI setup that lets them confirm targets with their cameras once they arrive. They’re more expensive and usually deployed against high-value targets.
Why drone swarms (groups) fall short in offensive roles
Despite the hype, drone swarms, or more accurately, groups of drones, are a poor fit for real-world military attacks. Below are key reasons why.
1. Leveling the battlefield rather than shifting it
Attacking drone groups are designed to be destroyed once they strike their target. So they need to be cheap to keep warfare cost-effective. But what does cheap warfare really mean? It often means simple, standardized designs. When both sides can copy and deploy the same tools, it removes the technological barrier that might otherwise give one side a lasting edge.
When both sides use the same tech and the same strategy, it leads to a Verdun-style deadlock – the most costly and least efficient kind of war.
Quick look back. When machine guns and artillery first appeared, many believed these advanced weapons would give one side a clear edge. But it all led to Verdun in World War I, the war’s longest and bloodiest battle. France and Germany had produced so many guns that nearly a million people died, yet neither side made real progress.
Liddell Hart backs this in his book “Strategy: The Indirect Approach.”
“Throughout the ages, effective results in war have rarely been attained unless the approach has had such indirectness as to ensure the opponent’s unreadiness to meet it…
In strategy, the longest way round is often the shortest way home.
A direct approach to the object exhausts the attacker and hardens the resistance by compression, whereas an indirect approach loosens the defender’s hold by upsetting his balance”.
The essence of effective war is gaining a smart advantage.
When the Germans reached the English Channel just weeks into World War II, they still used machine guns, but their role had changed. No longer seen as the decisive weapon they were before World War I, machine guns became just one element in a broader shift: the rise of mobile warfare. Rather than placing them in static positions, German forces integrated them into fast-moving infantry squads that supported armored advances. The gamechanger wasn’t the weapon – it was the method.
The same principle applies to drone swarms. It’s not cost-effective to throw thousands or millions of drones at your enemy. They’ll respond in kind, and both sides will suffer – destroyed infrastructure, exhausted forces, and civilian casualties. With little or no progress.
Instead, you can use the same technology strategically, like Ukraine did in its recent Spiderweb operation.
On June 1, 2025, Ukraine launched around 117 FPV drones from covert sites deep inside Russia. Their targets: five key airbases. The result: up to 41 aircraft damaged or destroyed, including strategic bombers, all without using traditional airpower. The drones flew in waves along preplanned GPS routes and relied on onboard AI for autonomy in jammed environments. The mission showed how cheap, massed drone strikes can break through strong defenses, impose asymmetric costs, and shift the strategic balance.
2. Tech limitations in the military context
Drones matter because they can go where humans shouldn’t: enemy territory, mined fields, active fire zones, or contested airspace. But that value vanishes if they can’t fly far enough to reach those areas.
While today’s drones can travel long distances, communication becomes a serious challenge beyond 100 kilometers. The Earth’s curvature hinders live imagery beyond the horizon.
One workaround is using multiple retranslators to extend the range. But that creates two major problems. First, there’s a delay of several seconds. At 50 meters per second, even a short delay means the drone has already moved 150 meters. That kind of lag kills accuracy. Second, it drives up costs, raising the question of whether it’s worth it.
The only real solution is satellite communication. Right now, that’s only available in the US, where drones fall under an Iridium satellite spot beam.
3. Ethical and legal concerns around fully autonomous killing
Real drone swarms are meant to act autonomously and make their own decisions. But in attacks, that creates a problem. Military doctrine and international law require meaningful human control over weapons. As autonomy increases, so does the risk of “accountability gaps,” where no one is clearly responsible if a drone strikes the wrong target.
There’s also a deep ethical concern. When machines make life-or-death decisions, they can’t be trusted to reliably tell friend from foe. Many worry a drone could mistakenly target its own side or harm civilians if its AI gets it wrong.
Humanitarian demining could reach zero tolerance with drone swarms
While drone swarms fall short in offensive military roles, the same technology could become transformative in humanitarian demining. In this drone swarms use case:
- We don’t need drones to go far away from us, so communicating with a swarm isn’t a problem.
- There are no humans on the mined field, and the goal is mine detection – just gathering and processing data, not targeting people. So full autonomy works just fine.
- The swarm isn’t a one-off tool here, which means we can build it as advanced as needed to get the best possible detection results.
- Plus, we don’t have to waste budget on protection. In humanitarian demining, drones don’t need to hide, shoot, or find secure ways to share info about an enemy.
That’s a breakthrough humanitarian demining urgently needs. Despite gradual drone and AI implementation, humanitarian demining is long, risky, and expensive:
- At the current pace, clearing all explosive remnants of war will take more than a century.
- In 2023, demining accidents
caused 15 deaths across three countries. - Mines also remain a danger to civilians, with at least 5,757 casualties from mines and ERW in 2023.
- Removing a single landmine costs 50 to 100 times more than making one. The main driver of high humanitarian demining costs is the large number of false alarms treated as real threats.
So we need a solution that is:
- Fast
- Accurate
- Safe for humans
Drone swarms meet all three needs. Working in parallel and continuously optimizing their flight paths, drones in swarms can cover large areas at high speeds (50, 100, even 120 kilometers per hour). It means they significantly reduce the time it takes to scan a mined area.
Drone swarms elevate detection accuracy because they validate findings collectively. When one drone detects a possible mine, others nearby automatically rescan the area from different angles and with different sensors. This multi-angle, multi-band confirmation process helps filter out noise, reduce uncertainty, and improve target confidence – something human teams or single drones can’t do as efficiently or reliably.
Equipped with basic onboard logic, these drones can also adjust their flight paths and scanning speed in real time to stay efficient and thorough. Once launched, they divide the area, follow set routes, identify possible threats, and combine their findings into a clear, shared picture.
Drone swarms could move humanitarian demining toward a zero-tolerance approach and reach up to 98% detection accuracy and cutting research time significantly.
All this with minimal human involvement during the technical survey stage of demining. With this approach, by the time humans begin clearance, they know exactly where the landmines are, making the process much safer.
Here is how drone swarms could work in humanitarian demining:
- Each drone carries Ground Penetrating Radar (GPR) sensors operating in different wavebands like L-band and X-band. While longer wavelengths in the range of 0.5-1 meter can penetrate beneath soil and vegetation for broad subsurface scans, shorter wavelengths provide higher resolution imaging but with shallower penetration.
- Drones autonomously scan suspected hazardous areas. Onboard AI detects anomalies in real time, adjusting wavebands based on soil, ERW type, season, vegetation, and weather.
- When one drone spots an anomaly, others rescan the area from different angles and bands. All data feeds into a large AI model for deeper analysis and precise mapping.
I predict this solution could be available by 2030. Right now, we face limits in processing power, GPS accuracy, and visual odometry. Solving these challenges demands complex R&D.
Drone swarm demining is already under research
Here are at least 2 examples of companies researching solutions in this direction.
- The Department of Mechanical and Aerospace Engineering at Sapienza University is testing drone swarms equipped with ground-penetrating radar to scan terrain in a structured way, though for now, they use only one type of waveband (but they may have tested different wavelengths).
- Dropla is developing a Multi-Domain Swarming Platform that integrates drone swarming with AI-driven sensor fusion and scalable robotics to enhance speed, safety, and efficiency in demining operations.
Conclusion
If the dominant story about drone swarms has been written by the military, it’s time for a new chapter. I’m not saying drone swarms are completely ineffective in warfare – reconnaissance and logistics, for example, offer promising military applications.
However, civilian use cases, including humanitarian demining, may offer an equal or even better fit for what drone swarms do best.
