Russian occupying forces have long used highways like the M14 to drive to Crimea. Trucks flowed along the highway, beyond the reach of the drones that made movement near the front line so dangerous. That began to change in April 2026 when the Ukrainian military deployed a large number of low-cost, intermediate range and semiautonomous attack drones that could reach beyond the range of quadcopters—as far as two hundred kilometers—in its middle-strike campaign. By late May, the M14 was populated by overturned trailers and burned out vehicles.
The affordable, precise mass Ukraine used in its middle-strike campaign is invalidating the tactics militaries have traditionally used to survive enemy fires: dispersion and mimicry. Both are challenged by the growing reach of drones and their proliferation. Dispersion tactics work well against artillery but do not improve survival rates against drones. Mimicry’s value collapses as the number of munitions approaches the number of targets in an area. While hardening can still help fixed positions survive, the US military needs to quickly identify techniques that work in combat and that will transfer to US operations. And it must develop new techniques to protect and sustain its forces in depth even as the tactics and technology associated with middle strike proliferate.
The Diminishing Value of Dispersion and Mimicry
Artillery and other indirect fire systems typically separate sensors and munitions. They do not steer toward targets, with a few exceptions, and are also area effect weapons, making their attacks inherently imprecise. As a result, unguided artillery is most valuable when targets are clustered, maximizing the probability of hitting at least one target, or destroying many targets simultaneously.
Dispersing targets reduces the efficiency of artillery, and therefore the probability that a strike will take place. If an artillery attack does take place, spreading out reduces the number of people and amount of equipment destroyed, making it easier to treat casualties or continue the mission. These lessons are reinforced at many training events.
Mimicry can also play a valuable role in defending against traditional strike tactics. Mimicry, in this case, refers to disguising a high-value asset as something less valuable, such as making a command vehicle or breaching asset look like an infantry carrier or logistics vehicle. Mimicry works best in environments where there are fewer munitions than potential targets, and attackers therefore prioritize high-payoff targets. In these conditions, a high-payoff target may not be attacked, even when observed by an attacker. Mimicry differs from the use of decoys, which increases the number of potential targets in an area, and is still a valuable approach.
Abundant one-way attack drones do not have the characteristics that make dispersion and mimicry useful techniques against artillery, missiles, rockets, or attack aircraft. Combining sensors, aircraft, and payload in a single platform creates an immediate feedback loop that improves accuracy. One-way attack drones are also usually point weapons. Many carry shaped charges, which typically only destroy a single target, no matter how close together targets are. Increasingly, they can also attack in large numbers.
As a result, abundant drones like those used in the middle-strike campaign do not use the same target selection criteria as artillery and long-range fires. When these drones reach an area, they often strike whatever targets they find rather than having to prioritize clustered or high-value targets. As a result, dispersion does not work because a drone’s payoff does not increase as targets move closer together. And mimicry does not work because sufficiently abundant drones can attack every target, rather than just high-payoff targets.
Testing the Argument
This theory of the diminishing value of traditional tactics like dispersion and mimicry is backed up by empirical evidence from the battlefield. When Russian forces disguised fuel trucks as water haulers or milk trucks, Ukraine’s middle-strike campaign destroyed them anyway, knowing that the loss of any truck was worthwhile given the low cost of the intermediate-range drone and time required to search an area.
But to more rigorously test the theory, we ran a series of Monte Carlo simulations. The experiments had a staggered grid formation of twenty vehicles move through a corridor twenty kilometers long and two kilometers wide. In the first experiment, the distance between vehicles varied from ten meters to six hundred meters at ten-meter intervals (i.e., vehicles were ten meters apart in the initial round of simulations, then twenty meters apart, then thirty, and so forth). At each ten-meter interval, separate iterations were run in which the vehicles were attacked by six, twelve, eighteen, and then twenty-four munitions. And each was run with two separate munition types—first, 155-millimeter artillery shells fired in a six-shell volley with a 150-meter circular error probable, and second, drones that destroyed their targets in eight out of ten attacks, a ratio derived from our field research. Each combination of interval, munition type, and munition count was tested four hundred times.
The results were stark. Against artillery, dispersion worked exactly as US practice would predict. At ten-meter intervals, between one-third and two-thirds of vehicles survived. Survival rates increased with dispersion, before leveling off at 95–98 percent at three-hundred-meter intervals. When vehicles were within drone range and the drones had a high probability of finding the vehicles, dispersion had no effect at all. At every interval and every munition count, losses were set by the number of drones committed and how often each attack destroyed a vehicle. Spreading out only changed which vehicles the drones destroyed first. The model assumes that dispersion does not degrade a drone’s ability to find moving vehicles. Recent field research in combat zones supports this assumption, as moving vehicles are usually unable to conceal themselves from persistent intelligence, surveillance, and reconnaissance (ISR).
The second experiment tested mimicry. In the same corridor, twenty by two kilometers, four of twenty vehicles were designated high-payoff targets, and the attacker’s munition count varied from two to thirty. Mimicry was effective as long as there were significantly fewer attack drones than vehicles. As the ratio of drones to vehicles approached one to one, the drones began destroying every vehicle, and the value of mimicry collapsed.
What Works, and When it Doesn’t
Hardening still works well against abundant, low-cost drones. While dispersion and mimicry focus on shaping an attacker’s choices, hardening increases the probability that a target will survive being attacked. Ukraine’s antidrone net corridors are a striking example. Ukraine covered more than one thousand kilometers of roads with nets between the start of 2026 and mid-July, mostly to reduce the threat of one-way attack drones that have less than a thirty-kilometer range. The nets catch and sometimes detonate one-way attack drones before they strike their targets.
Nets have limitations against middle-strike campaigns. They require a lot of work to build and maintain, especially when they are regularly attacked. In May 2026, Ukraine built more than two hundred kilometers of new covering, but also had to repair thirty-eight kilometers of netting destroyed by Russian forces. As a result, netting and other types of cover work well for select routes near the front lines, where most nets are today, but would be challenging both to build and to maintain if extended out to hundreds of kilometers, the maximum effective range of many middle-strike drones.
Another technique is to move facilities underground. This approach also has significant limitations. It only protects stationary objects, leaving logistics convoys or formations moving toward the front lines vulnerable. Hardening also does not create indestructible positions. Weapons with heavy payloads can destroy many underground or fortified positions. However, weapons that can destroy underground sites are also typically more expensive than drones with a smaller payload. For example, the FP-5, a low-cost Ukrainian cruise missile, costs roughly $600,000. Middle-strike platforms, by comparison, can cost less than $10,000. Subterranean protection does therefore have the advantage of forcing attackers to make trade-offs and can affect the number of munitions available for an attack.
Low-cost interceptors also play an important role. Interceptors work well against one-way attack drones like Shaheds at a relatively low cost, accounting for roughly 70 percent of the Shaheds downed near Kyiv over the course of one month this year. Other interceptors have targeted Russian ISR drones that often cued Lancet one-way attack drones. This campaign was so effective Russia eventually had to develop new, less successful tactics to continue employing the Lancet.
It is unclear how well those cases will transfer to defending against middle-strike campaigns. Interceptors face a similar density problem as covered roads. It is expensive and logistically demanding to build and field enough launchers, ground control stations, and interceptors to cover the large areas threatened by middle strike. It is also true that interceptors don’t pose the same cost ratio on middle-strike drones as on Shaheds and ISR drones. The Ukrainian military routinely destroys $35,000 Shaheds with interceptors that cost less than $10,000, and ISR drones that cost between $87,000 and $120,000 with FPV interceptors that cost less than $1,000. Interceptors and low-cost middle-strike drones are much closer to cost parity. It should also be noted that no single system is able to defend a wide area against drones at a reasonable cost.
What the US Military Should Do
The US military needs to prepare to fight adversaries, including mid-level powers and nonstate actors, that have access to the technology and tactics used in Ukraine’s middle-strike campaign. Imitation costs less and is usually faster than invention, and middle-range drones have already appeared in Sudan. The US military must now choose between defending large areas from drones strikes, planning for a greater degree of loss and friction, or accepting stark operational limits. Partial solutions and a route to a more effective set of solutions are needed.
Layer Passive and Active Defenses
The most straightforward way to strengthen US defenses against middle-strike campaigns is to copy counterdrone techniques that work well, while keeping in mind their significant limitations. The US military should fortify command posts, fuel and munitions, maintenance sites, and other points that will occupy the same location for days or weeks at a time. Emplacing nets would be more challenging, and would be beneficial to US forces in a limited number of scenarios. Because any US military effort to build nets would likely take place in foreign countries, there is greater potential for hostile populations and sabotage. That would challenge the construction rate, make maintenance more dangerous, and, in the case of sabotage, make the nets less reliable. Also, unless the US military can create and maintain net corridors at a much faster rate than the Ukrainians, nets will leave significant gaps.
The US military should also continue to increase its investments in active defenses. Interceptors have worked well in a variety of environments. While it will be difficult to replicate the cost-ratio of Shahed interceptors or destroying ISR drones with quadcopters, interceptors can still help reduce the threat of middle-strike drones. It is also likely that gun-based systems and other active defenses that are emerging will continue to strengthen a layered defense concept.
The combination of passive and active defenses has the potential to force attackers into trade-offs. Hardening drives attackers to use larger and less common munitions, reducing the number of munitions available for an attack, and therefore restoring some of the value of mimicry and dispersion at certain ranges. While helpful for defenders, this leaves a gap between the size of the areas that can be covered by nets and that of the areas that can be threatened inside the maximum effective range of abundant, middle-strike campaigns—a gap where vehicles and formations are vulnerable.
Transition to Attritable Logistics
Another approach is to accept that, for now, a portion of middle-strike drones will make it through active defenses and attack targets that do not have effective passive defenses, particularly logistics vehicles. With that assumption, it makes sense to shift to accepting a predictable amount of attrition and building systems the US military can afford to lose.
Three factors can make the loss of a logistics vehicle unacceptable: the lives of the crew, the cost and difficulty of replacing the platform, and the cost and difficulty of replacing the cargo. Low-cost uncrewed ground vehicles (UGV) have the potential to address all three. The destruction of a UGV does not threaten human life, and many UGVs cost under $25,000, significantly less than a crewed military vehicle. The mass deployment of UGVs, facilitated by lower costs per vehicle, will allow vehicles to carry smaller payloads per trip, reducing the impact of the destruction of a single payload.
There are limits to UGVs. Today, Ukraine uses UGVs for the last several kilometers of logistics. They do not use them farther away from the front line because of their relatively slow speed compared to crewed vehicles, the challenge of operating UGVs, the maintenance burden they carry, and their smaller payloads. Relying on UGVs to move supplies for hundreds of kilometers will require better control systems, better autonomy, and a large investment in vehicles.
Deploy Field Research Teams to Combat Zones
The Department of Defense should invest in field research in active and recent combat zones to better understand events and their context, develop lessons learned, and determine which lessons should and should not transfer to the US military and its allies and partners, including to the different contingencies the United States might face. In the context of defending against middle-strike campaigns, field research would help identify the cutting edge of active and passive defense technology, new tactics or techniques, new concepts, and leading indicators of emerging threats.
Ideally, field research teams would be a combination of military personnel and defense civilians. They could embed directly with belligerent forces, as the US military did when it sent General Pershing to the Russo-Japanese War, or stand up dedicated field survey teams after a conflict, similar to the US Strategic Bombing Survey during and after World War II. If policy constraints prohibit military personnel and defense civilians from conducting this research, the department should hire outside entities to openly conduct field research on its behalf.
Develop New Tactics, Techniques, and Procedures through Training and Modeling and Simulation
The department should develop new tactics, techniques, and procedures through a combination of live training and modeling and simulation. Questions and assumptions based on field research can be tested against live drone and interceptor fleets at combat training centers. This approach is well suited to testing several open questions this piece has raised. It could also help resolve the tension between dispersing and concentrating active defenses. Modeling and simulation could help determine whether limited interceptor or directed-energy inventories are better spread across a logistics network or concentrated at a smaller number of critical nodes, and if concentration proves the better answer, where along that network those nodes should sit. And it could test whether UGVs can sustain long-distance bulk logistics under drone threat, and how their survivability and throughput compare to crewed convoys performing the same mission.
An Unsolved Problem
Proliferated middle strike is a challenge the US military needs to prepare to address. There is no definitive solution other militaries have developed that it can copy. The techniques that still work well have limits. The US military has the opportunity to invest in existing capabilities and tactics that will reduce the problem, and develop and experiment with new technologies and tactics to close the gaps they leave. That work can be done now, or it can be done on the receiving end of a middle-strike campaign.
Justin Lynch is the senior director for defense at the Special Competitive Studies Project. He teaches artificial intelligence policy as an adjunct professor at Georgetown and George Washington Universities and is a nonresident senior fellow at the Atlantic Council. He served in the US Army and in civilian roles in the national security enterprise across government, academia, and industry.
Kian Molani is an associate director for defense at the Special Competitive Studies Project. He holds a master of science in engineering in aerospace engineering from the University of Michigan, where his research focused on the development of advanced autonomy algorithms for dynamics modeling and control for aerospace robotics. He also holds a bachelor of engineering in engineering physics from Carleton University in Ottawa, Canada.
The views expressed are those of the authors and do not reflect the official position of the United States Military Academy, Department of the Army, or Department of Defense.
Image credit: Staff Sgt. Alex Manne, US Army
