After the Gold Rush: If/When Drone War Dominance Fails

There is a gold rush underway in military drones. Defense companies are offering increasingly lethal first-person-view drones, autonomous systems, loitering munitions, unmanned surface vessels, interceptor drones and artificial-intelligence-enabled weapons. Governments are creating specialized drone formations and accelerating procurement, while almost every new system seems to arrive accompanied by some variation of the same powerful phrase: combat proven in Ukraine.

There is good reason for the enthusiasm. Drones have become extraordinarily important in the Russia-Ukraine War. They find targets, direct artillery, attack vehicles and personnel, strike infrastructure hundreds of miles behind the front and increasingly hunt other drones. At sea, Ukrainian unmanned surface vessels have contributed to forcing much of Russia's Black Sea Fleet away from Sevastopol.

None of that should be dismissed, but neither should it end the discussion. Before spending heavily in this new technological gold rush, governments should be asking a considerably more uncomfortable question: What exactly has Ukraine proven about drones, and what have we merely assumed from Ukraine?

The distinction matters because the military history of the modern drone remains surprisingly specific. Much of what we now describe as "drone warfare" matured during one particular conflict, against particular opponents, over particular terrain, and increasingly within a positional battlefield that created unusually favorable conditions for persistent surveillance and repeated drone attack. We may indeed be witnessing a revolution in warfare, but we may also be extrapolating universal conclusions from an unusually specialized laboratory.

The Myth of the Drone

The "myth of the drone" does not mean that drones are ineffective. Rather, it describes the increasingly common image of the drone as an almost supernatural weapon: inexpensive, nearly invisible, omnipresent, precise and extraordinarily difficult to stop. The same broad category of weapon can seemingly kill an infantryman, destroy a tank, sink a ship, attack a refinery and hunt another drone, all while costing a fraction of the weapon or platform it destroys.

Each of those things has happened, but the fact that they have happened does not mean that they happen automatically. There is an enormous difference between demonstrating a capability and producing a reliable combat effect under varying conditions.

Battlefield videos naturally reinforce the more dramatic interpretation because they show successful attacks. They seldom show the drones that failed to launch, lost communications, crashed, were jammed, were shot down, failed to find targets, struck without destroying their targets or simply disappeared. That denominator matters.

The Royal United Services Institute has cautioned NATO specifically against replacing traditional firepower with drones based upon Ukraine's experience. RUSI notes that only a fraction of the enormous numbers of FPVs launched actually reach targets and fewer still inflict decisive damage. More importantly, it argues that Ukraine's increasing dependence upon drones resulted partly from shortages of artillery, precision weapons and other conventional capabilities, rather than necessarily reflecting a conclusion by Ukrainian commanders that drones were inherently superior replacements for those systems.

That is a very different lesson from saying that the drone has replaced the tank, artillery or conventional airpower.

Ukraine Is an Extraordinary Laboratory, but It Is Still One Laboratory

The development of drone warfare in Ukraine did not occur all at once. Drones existed before Russia's full-scale invasion in February 2022, Ukraine had employed UAVs during the Donbas fighting, and both sides used reconnaissance UAVs and larger systems during the opening stages of the invasion. The dense FPV environment now associated with the Ukrainian battlefield, however, developed later.

As the conflict became increasingly positional, drones became extraordinarily well suited to the environment. Established fronts allowed operators, communications networks, repeaters, repair facilities, launch sites and logistics systems to develop behind relatively stable lines. The battlefield also produced repeated target patterns. Roads remained roads, supply routes could be watched, fighting positions could be mapped, and vehicles repeatedly traveled through observable areas. Artillery positions, command posts, communications systems and eventually the drone operators themselves became part of an enormous reconnaissance-strike competition.

Ukraine consequently developed an extraordinary unmanned warfare ecosystem, and that achievement deserves careful study. Indeed, the Ukrainian procurement system itself may contain lessons that every Western military should examine. CSIS reports that Ukraine created a commercial-first acquisition structure, decentralized purchasing authority and incorporated battlefield feedback directly into procurement. By 2025, roughly one-third of Ukrainian weapons procurement spending was directed toward commercially developed capabilities, with unmanned systems comprising a major portion.

The success of Ukraine in adapting to its battlefield, however, does not automatically establish that every military should recreate Ukraine's force structure. RUSI has made precisely this distinction, warning that NATO forces preparing for maneuver warfare should not automatically copy Ukrainian formations optimized to conduct area defense and attrition under the particular conditions prevailing in Ukraine.

The appropriate lesson should therefore be straightforward: study Ukraine intensely, but do not assume that every future war will look like Ukraine.

What Happens When the Battlefield Starts Moving?

This question led us to a thought experiment. Rather than asking how drones perform on today's Ukrainian battlefield, we used AI-assisted modeling and conceptual wargaming to consider how the first generations of contemporary FPV-heavy drone warfare might function against two very different Eastern Front operational problems.

The first was Operation Citadel at Kursk in 1943, in which concentrated mechanized forces attempted to penetrate an extensively prepared defensive system. The second was Operation Bagration in 1944, which involved a successful Soviet breakthrough followed by rapid operational exploitation.

These exercises were not attempts to "re-fight" World War II with quadcopters, nor should they be mistaken for empirical evidence. They were deliberately extreme stress tests intended to expose assumptions. Kursk gave us density, while Bagration gave us velocity.

At Kursk, German forces attacked a defense in depth containing extensive minefields, fortifications, artillery zones and antitank defenses. Despite committing enormous armored strength, German formations achieved only limited penetration before the offensive stalled. Such an environment initially appears almost ideal for modern drones. ISR systems would probably have made concentration extraordinarily difficult, while FPVs could have attacked engineering vehicles, immobilized armor, logistics vehicles and command elements. Drone-corrected artillery would have compounded the problem.

Our exercise therefore did not suggest that drones suddenly become ineffective against massed armor. Quite the opposite: it suggested that Kursk might have been an exceptionally lethal drone battlefield.

The conclusion began to change when we turned to Bagration.

The Bagration Problem

Operation Bagration presented a fundamentally different operational problem because Soviet forces achieved deep penetrations and then exploited them rapidly. Instead of asking how drones might attack dense formations moving through prepared defensive zones, we now had to consider what happens when the entire battlefield begins moving.

The tanks move, but so do the reconnaissance elements, artillery, headquarters and supply columns. The difficulty is that the drone ecosystem must move as well. Operators must displace while antennas, repeaters, batteries, generators, replacement aircraft, warheads, repair facilities and communications equipment somehow remain close enough to support them.

A drone capable of striking a tank is therefore not necessarily the same thing as a drone organization capable of stopping an armored formation advancing twenty or thirty kilometers. The relevant question is no longer simply whether a particular drone can destroy a particular vehicle. It becomes whether the entire drone force can continue finding, identifying, engaging and repeatedly reengaging enough moving vehicles to disrupt the operational tempo of the formation while the drone organization itself is being forced to relocate.

That is a much more demanding test, and contemporary Ukraine is already producing evidence pointing toward this problem. RUSI reports that Russian forces increasingly attempt to identify and attack Ukrainian UAV operators, electronic-warfare positions, radar, command posts and artillery before offensive action. Russian forces then attempt rapid movement when Ukrainian ISR has been degraded, particularly during periods of poor weather.

This does not demonstrate the defeat of drone warfare. It demonstrates adaptation, and adaptation is precisely what the mythology surrounding drones sometimes leaves out.

The Drone Has an Ecosystem

A drone is not merely an aircraft carrying a camera or explosive charge. Its combat effectiveness can depend upon reconnaissance, communications, satellite navigation, operators, antennas, repeaters, electronic protection, maintenance, batteries, launch positions, target identification and sometimes external intelligence. An opponent therefore does not necessarily have to destroy every drone. If enough parts of the supporting chain can be disrupted, the drone can simply become less useful at the moment when it is needed most.

RUSI's examination of FPV employment in Ukraine identified several important limitations. Electronic warfare can deny areas to radio-controlled FPVs, weather can seriously degrade operations, payload limitations prevent many FPVs from destroying some target classes, and operators requiring low-latency communications may remain relatively static while flying, making them targets themselves. RUSI also found that most FPV kills in the Ukrainian data it examined occurred within roughly three kilometers on either side of the forward line.

None of these limitations makes FPVs unimportant. They demonstrate instead that their effectiveness is conditional, and that distinction should profoundly affect procurement decisions.

What Happens When the Enemy Learns?

Perhaps the greatest danger in extrapolating Ukraine's drone experience is forgetting that the opponent gets a vote. Every successful weapon encourages a countermeasure, and every successful countermeasure encourages another modification to the weapon. Radio-controlled drones encounter jamming, fiber-optic drones appear, defenses improve, autonomy increases, interceptor drones emerge, signatures change and operators move farther from the front. The cycle continues.

Russia's Geran program illustrates the process particularly well. CSIS describes a continuous development cycle in which modifications to navigation, communications, payload and propulsion are tested operationally, successful adaptations are standardized and failures discarded. The resulting advantage increasingly appears to lie in the speed of adaptation rather than in possession of one supposedly decisive drone.

This may ultimately prove to be the more important revolution. The revolutionary development may not be the drone itself, but rather the speed at which relatively inexpensive weapons can be modified after encountering enemy countermeasures. Those are very different propositions, particularly for governments deciding what they should buy in large quantities.

The Maritime Version of the Same Question

Ukraine's maritime drone campaign provides another warning against simplistic conclusions. Ukrainian unmanned surface vessels have achieved genuine operational effects in the Black Sea. They have damaged and sunk Russian vessels and contributed to pushing important elements of the Russian fleet away from Sevastopol. RUSI accordingly regards the Ukrainian maritime unmanned campaign as an important success.

Yet once again we should ask precisely what has been demonstrated. CSIS notes that Ukrainian USVs operated under favorable circumstances that included excellent intelligence concerning Russian dispositions, relatively constrained geography, known Russian ports and ships frequently spending time at anchor. It also observes that many Russian naval losses have resulted from long-range precision missile attacks against stationary ships rather than surface drones.

This does not invalidate the maritime drone; it changes the procurement question. A spectacular video of an unmanned boat striking a ship demonstrates that the attack worked, but it does not necessarily demonstrate that the same weapon can penetrate an alerted naval formation that knows it is coming and possesses airborne surveillance, helicopters, electronic warfare, escorts, defensive drones, guns and layered sensors specifically organized to stop it.

Surprise is legitimate warfare, but surprise and penetration are not the same test.

This suggests what we might call the Second-Attack Test. Rather than judging a supposedly revolutionary weapon primarily by its first spectacular success, military planners should examine what happens after the defender understands the weapon and begins reorganizing specifically to defeat it. Subsequent engagements reveal whether the weapon can survive an adaptation cycle and whether its users can modify their own tactics rapidly enough to preserve the advantage. By the time the same weapon and its countermeasures have encountered each other repeatedly, we are no longer merely testing surprise; we are beginning to test doctrine.

Beware the Drone Gap

This brings us to perhaps the most dangerous part of the gold rush: procurement psychology.

Stanley Kubrick provided an excellent analogy in Dr. Strangelove. Faced with the possibility of Soviet survivors retreating into underground complexes, General Turgidson becomes alarmed about the possibility of a "mineshaft gap."The absurdity works because the underlying logic is recognizable. If the other side might possess an important capability, then we must possess it as well, and because the consequences of falling behind could be catastrophic, questioning the requirement itself becomes increasingly difficult.

Military establishments today should guard against creating a drone gap in much the same fashion. Russia has enormous numbers of drones, Ukraine relies heavily upon drones, China is developing drones, and analysts continually describe drones as the future of warfare. It becomes easy to conclude that every military therefore needs more drones, followed by autonomous drones, drone swarms and whatever new system is being advertised as the latest combat-proven development.

Eventually, the procurement process risks moving away from the most important question: What military problem are we trying to solve?

That is particularly important because the manufacturer should not be allowed to define both the problem and the solution. A company selling the newest and "most lethal" drone naturally has an incentive to emphasize the battlefield conditions under which its system excels. The military customer has the opposite responsibility. It must deliberately create the conditions under which the weapon is most likely to fail and then determine whether the system still provides enough combat value to justify the investment.

Combat Proven Where?

The phrase "combat proven" consequently needs qualification. A weapon may be combat demonstrated because it successfully attacked a real enemy, but that is different from demonstrating sustained performance from an established drone infrastructure against targets operating within a constrained battlefield. Both are different again from proving that the system can remain effective while friendly and enemy forces rapidly relocate.

The most demanding standard would be something approaching combined-arms demonstration, in which the system continues producing useful effects against an enemy deliberately integrating electronic warfare, deception, counter-UAS, artillery, air defense and maneuver specifically to defeat it.

These are very different achievements. A drone that destroyed twenty vehicles around a static front has demonstrated something important, but it has not necessarily demonstrated that it can stop a mechanized formation conducting rapid exploitation. Similarly, a USV that strikes a vessel at anchor has proven a useful capability without necessarily demonstrating that it can penetrate an alerted surface action group.

The Pentagon, or any other defense ministry, should know precisely which proposition it is purchasing.

The Drone Dependence Paradox

There is an even greater danger if enthusiasm for drones begins influencing force structure itself. Suppose a military becomes convinced that drones have fundamentally replaced significant portions of conventional combat power. Artillery inventories might decline, armored forces could shrink, traditional reconnaissance might receive less investment, aircraft procurement could change and expensive missiles could increasingly be judged inefficient because inexpensive drones appear capable of attacking many of the same targets.

The force gradually becomes optimized around persistent surveillance and inexpensive precision attack. The danger appears when an opponent discovers how to degrade the drone ecosystem.

The opponent does not have to eliminate drones. It merely has to reduce their effectiveness below the level upon which the opposing doctrine depends.

This is the Drone Dependence Paradox. The more completely a military reorganizes itself around the assumption that drones will reliably provide persistent observation and inexpensive precision strike, the greater the operational consequences if an opponent can temporarily suppress that capability.

If one thousand drones previously generated one hundred useful engagements and countermeasures reduce that number to thirty, the drones have not suddenly become useless. The doctrine designed around receiving one hundred engagements, however, may no longer function as intended. If artillery, armor, aviation or other conventional capabilities that were supposed to provide the second layer have already been traded away to finance the drone revolution, there may be very little available to compensate for that decline.

RUSI's warning is therefore particularly important. Ukrainian reliance on drones should not lead NATO forces to substitute them wholesale for conventional fires. Ukraine's own commanders continue to value artillery and advanced antitank weapons because these systems can remain responsive and reliable against breakthrough attempts, while FPVs provide particularly useful tools for sustaining attrition because they can be produced and fielded in enormous numbers.

After the Gold Rush

None of this constitutes an argument against drones. Militaries that fail to understand unmanned warfare are inviting serious problems, and drones should be purchased, operators trained, counter-drone systems developed and autonomous technologies subjected to aggressive experimentation. The Ukrainian acquisition model also deserves serious examination, particularly its ability to connect soldiers with manufacturers and incorporate battlefield feedback rapidly.

Governments should nevertheless resist allowing an important military technology to become an article of faith.

The appropriate lesson from Ukraine may not be that drones inherently dominate modern warfare. It may instead be that drones have dominated important portions of a particular war because the battlefield, available technologies, force structures and resource constraints created conditions in which they became extraordinarily effective. A different war may create very different conditions.

The military that eventually gains the advantage may therefore not be the one possessing the largest number of drones, but rather the one that first understands why the opponent's drones work and then systematically attacks those reasons. That could mean suppressing reconnaissance, attacking operators, disrupting communications, employing deception and obscuration, exploiting weather, moving faster than the opposing kill chain, forcing drone organizations to relocate and integrating counter-UAS systems directly into maneuver formations. Just as importantly, it means retaining sufficient conventional fires to continue fighting during those periods when the drone network cannot deliver the effects expected of it.

This brings us back to Kursk and Bagration. Our modeling did not show that drones would suddenly become irrelevant on a mobile battlefield. At Kursk they might have been devastating. Bagration, however, exposed the other side of the equation by forcing the drone organization itself to keep pace with rapidly moving combat formations.

A weapon's lethality is only one component of combat power. Operational tempo can itself become a countermeasure.

That is why governments should be wary of purchasing tomorrow's force based entirely upon the most spectacular images from today's war. The drone is unquestionably here to stay. What remains unanswered is whether drone dominance is equally permanent.

Before governments spend their way into a drone gap because everyone else appears to be doing the same thing, they should find out.

Eventually, someone will.

The military that discovers the answer first may have purchased something considerably more valuable than another drone: an understanding of how to fight after the drone gold rush is over.