Cry ‘drone,’ and let slip the fruit flies of war

Scientists study fruit flies for a simple reason: they reproduce fast. As a result, researchers can change the environment and observe 25 generations of fruit flies each year. 

Drones are the fruit flies of warfare, and their rapid evolution offers the military the chance to do something extraordinary: deliberately accelerate adaptation so that cheap changes on our side force expensive ones on theirs. 

Traditional weapons take years, even decades, to develop and then remain in service for generations. Drones in every domain can evolve in weeks. Their radios, sensors, software, payloads, propulsion and even structures can change at different rates. In under a year, a drone can morph in both shape and function.  

The clearest example is currently unfolding in the skies over Ukraine. Early in the war, Russia purchased and quickly employed simple Iranian Shahed drones with piston engines and a satellite receiver for navigation. It called them Gerans and within months, licensed the design and began building their own. Russian forces quickly began using them in inexpensive long-range attacks against Ukraine. Ukraine responded with electronic warfare to jam the satellite receiver and disrupt navigation. Russia countered with receivers designed to resist jamming. As Ukrainian jamming improved, the Geran grew new antennas in different patterns. 

Ukraine also got better at tracking and shooting down Gerans. Russia added cameras and communications that allowed operators to see what the drones saw and redirect them in flight. Ukraine began using cheap interceptor drones to hunt them. Russia answered with speed — it first flew an Iranian jet-powered variant, the Shahed-238, then built its own Geran-3 by fitting the same airframe with a Chinese turbojet. The shortcut had a drawback: at twice the speed, the old Iranian airframe couldn’t take the G-loads. So, Russia did what it had previously avoided and redesigned the aircraft itself, producing the Geran-4 and Geran-5 and retooling its factories to build them. By August 2026 it had stopped making the Geran-3 altogether.

Ukraine had pushed reported interception rates against propeller-driven Gerans above 90 percent. Against the jet-powered models, the rate reportedly fell to around 60 percent. Since the jets fly higher, above cheap defenses like gun trucks and shoulder fired missiles, they drove Ukraine to generate fighter sorties and fire missiles it cannot spare, including scarce missiles.  

In only a few years, the Geran family incorporated new navigation, communications, payloads, propulsion and airframe designs. Each Ukrainian response changed the threat environment. The Geran evolved to survive in it. Ukraine is now testing faster interceptors to catch them.  

The pressure which drove the Geran’s evolution came from intense combat. But the American military doesn’t need to wait for an enemy to supply that pressure: it can provide its own, in training and test events where operators choose where to apply selection pressure, aiming it at a problem instead of merely surviving it.  

The fruit fly comparison is not perfect, of course. Unlike with biological organisms, drones don’t reproduce or mutate by themselves — operators, engineers, manufacturers and military organizations decide which designs survive and continue to compete. But the underlying pattern is similar: short generations, variation, selection pressure, inheritance and recombination. 

Scientists use fruit flies to understand biological processes that extend beyond flies themselves, since living organisms share a surprising amount of DNA. The military could use drones in a similar way by experimenting with technologies, interfaces, manufacturing and acquisition processes that also govern more complex weapon systems. While modeling and simulation can help, nothing beats physical experimentation when you are looking for the unexpected. 

Today, the military generally tests systems to determine whether they meet requirements and to discover problems before combat. In competition-based acquisitions such as the Drone Dominance Program Gauntlets, testing helps select the best system. The program iterates but rewards the best drone rather than the teams that can produce successive generations the fastest.  

However, the military could use testing for another purpose: to deliberately drive technological evolution. Pick several drone problems and give multiple teams twelve months to produce four generations. Every three months, change the threat environment. Introduce an interceptor, jam the GPS or change the mission. Some changes should be unexpected. Give the teams the results, let them adapt and repeat.

The initial experiment could involve tens or hundreds of drones per generation, rather than replacing an entire operational inventory every quarter. Each generation should produce a limited batch that operators can actually test and use, while demonstrating a credible path to larger-scale manufacturing. 

Four generations in a year is, admittedly, an ambitious starting point, intended to expose the limits of the entire adaptation process. Whether three months is achievable is one of the questions the experiment should answer. But because the US — thankfully — is not in a Ukraine scenario, pushing the limits on what seems plausible is the best way to force rapid evolution.  

In a recent article for Breaking Defense, I argued that the military should crash 100,000 drones a year in training. Even a small share of those flights could be chained into successive generations with the results of one becoming the starting point for the next. The military would learn which technologies — and which teams — keep producing useful, scalable solutions as the environment shifts. A team with the best drone in generation one could lose by generation four to a different team with a better process for adapting and manufacturing every three months. Russia pays for its adaptation in combat losses; we can pay for ours in training crashes. 

The experiments should measure the time required for every stage of a generation. While a drone may be designed and prototyped in a week, the testing, manufacturing, contracting, delivery and training could take much longer. If all that takes a year, it means the generation took a year, no matter how fast the prototype came together. These experiments would reveal which parts of the overall system slow adaptation, and those are the processes which should be redesigned.

This is where the idea becomes a weapon. Once a military learns how to shorten the time between generations, it can study which changes force expensive responses from an opponent. The goal would be to identify which adaptations would be inexpensive to add but expensive and time-consuming for an adversary to counter: a software change that forces the adversary to change hardware; a hardware change that forces a system redesign; a system redesign that forces an overhaul of the adversary’s factories.  

A military that can reliably field new generations faster than its adversary can respond gets to decide where that adversary spends its money, its engineers and its factory capacity. Ukraine spent years developing a layered defense against propeller Gerans. A change in propulsion to a commercial Chinese turbojet, then a new airframe built around it, rendered much of that layer irrelevant, forcing Ukraine back onto its most expensive interceptors. The military needs to figure out which of its own adaptations would do the same. 

Imagine being an American adversary. You have finally developed a countermeasure to an American drone. Your engineers have solved the problem. Your factories are beginning to build the solution at scale. Your military is developing tactics and learning to use it. But the Americans are already two generations ahead with upgrades that your countermeasure may not defeat. Do you continue to produce the countermeasure? Redesign it? Or abandon your investment and start again?

And now, imagine that happening over and over again. An adversary could impose the same problem on the United States. The American military needs to learn how to create this advantage before an opponent does. And it can start by treating drones like fruit flies.

Col. Dustin “Whiz” Thomas is the Commander of the Defense Contract Management Agency’s Special Programs Command. He previously commanded DCMA’s Unmanned Systems–Experimental Command (US-X), where he led the Department of War’s Blue List program. He founded Game of Drones and previously led Black Phoenix, an initiative demonstrating the rapid design, manufacture and flight of unmanned aircraft. His work focuses on defense innovation, adaptive manufacturing and accelerating the fielding of emerging technologies.

Maj. Aubry J. Eaton, USAF, is an acquisition and intelligence professional currently serving as a Department of the Air Force Strategic Policy Fellow.  His work has previously appeared in Joint Force Quarterly and War on the Rocks. 

The views expressed in this article are those of the authors and do not reflect the official policy or position of the Department of the Air Force, the Defense Contract Management Agency, the Department of War or the U.S. Government.