Drones are Machine Guns
It was a sunny June morning in Brooklyn. Henry Gladwyn and I were walking to get a coffee and a bagel. Having met in the military, we got talking about the changing nature of warfare, the footage from Ukraine that we had seen, and the dronification of the battlefield.
We observed that almost all casualties in Ukraine are now caused by drones, and that they have forced changes of tactics on the front line. We wondered if the drone would be a technology that would change strategy forever. And so we looked to past conflicts to see which technologies emerged from the crucible of war, and the impact of each.
A soldier in 1916 might be forgiven for thinking that the machine gun was the decisive technology to emerge from the war. It had transformed the battlefield completely. Defensive lines held by a handful of men with a Vickers gun could stop entire battalions and the offensive had become nearly impossible.
If you had asked a general at the Somme what technology would define the next century of warfare, he would almost certainly have said the machine gun.
He would have been wrong. Not because the machine gun wasn't revolutionary, but because revolutionary and strategic are different things. The machine gun changed tactics, but it didn’t change strategy. It made the infantry more lethal, and then it was absorbed. Today every infantry section carries automatic weapons as standard. The machine gun became furniture.
The technologies that actually restructured the strategic landscape were the ones that looked marginal at the time. The tank made its debut at the Somme in 1916: slow and unreliable, it was more terrifying than effective. The aeroplane was still a canvas-and-wire reconnaissance tool, useful for spotting artillery but little else.
Neither looked like a world-changer, but both became one. The tank enabled manoeuvre warfare and was the foundation for the blitzkrieg that would tear through France in six weeks in 1940. The aeroplane became strategic the instrument that projected power across the world. It just took twenty years to get there.
Drones are tactically dominant right now but the trajectory of the machine gun suggests they will be integrated rather than decisive. In the near future, every infantry section will probably carry one the way they currently carry a rifle. Margins will collapse. The drone will become furniture.
So what is the strategic technology emerging from the war in Ukraine, if not the drone per se?
To answer that question, it is worth reflecting on the evolution of tactics and technology through the conflict.
The war began in a conventional way. In the first days of the conflict Russian forces conducted an air assault onto Hostomol airport near Kyiv. It was repelled by Ukrainian civilians wielding AK47s.
As the front line took shape, each side fired colossal amounts of artillery. Russia, at its peak, discharged an estimated 60,000 rounds per day. Ukraine fired a fraction of that but still enough to hollow out Western stockpiles built up over recent decades.
As the months passed, Ukraine began to run out of artillery ammunition, and men. Despite foreign aid it was not able to match Russian artillery inventories, or manpower. But it could replace artillery with drones.
At first these were commercial quadcopters dropping grenades down tank hatches. Then came the FPV drone: a racing drone with a warhead, costing $400, capable of defeating a $4 million armoured vehicle. Ukrainian volunteers were building them in garages. The economics were bleak for the defender of heavy armour.
Both sides learned to jam GPS signals and blind targeting systems. There followed a constant cycle of software update and counter-update, firmware pushed to the front like ammunition. That had never happened before. But now software iteration happens in almost real time.
Then Ukraine turned the Black Sea into a drone battlefield, forcing the Russian Black Sea Fleet to relocate its headquarters to Novorossiysk. A country with no functioning navy had achieved effective sea denial using uncrewed boats.
By 2025 the front had become something that military theorists struggled to categorise. It was too attritional to be called manoeuvre warfare, too technological to be called a meat-grinder in the WW1 sense.
Satellites meant that nothing larger than a small group of soldiers could move in daylight without being observed. AI-assisted targeting was compressing the kill chain from hours to minutes. Every movement was contested.
Fast-forward to today.
In 2026 the soldier's primary skill is increasingly electronic (operating drones, countering drones, managing communications in a jammed environment) rather than marksmanship or fieldcraft. The battlefield has, in short, become transparent and lethal in equal measure.
The question it leaves hanging is what kind of force wins on a battlefield like that. And what technology conveys advantage.
Some candidates might be autonomous decision-making at the edge, electronic warfare, space-based ISR, counter-drone technology. Or simply the introduction of tech deployed at the edge for the edge.
The soldier can now command drones, manage electronic warfare, gather and share intel, and target in faster loops. The strategic picture could build from the bottom up, rather than top down. And the side that can update their strategic decision cycle faster (and consistently faster) is likely to win.
Certainly Starlink is a strong candidate. It has provided Ukraine with reliable internet connectivity where conventional communications have been destroyed or jammed. It is fitted to many reconnaissance drones to stream live video back to operators, and enables commanders to securely communicate with units across the battlefield, even if mobile phone networks or fibre-optic infrastructure is unavailable.
Ukraine has also been using Starlink to coordinate resupply, evacuate casualties, and even connect civil infrastructure like hospitals, schools etc. Its inception as a company could not have come at a better time for Ukraine.
But Starlink doesn’t feel like quite the right conclusion to draw.
The autonomy stack (the collection of software that allows a machine to make decisions and act independently without constant human control) strikes me as the truly strategic technology in the decades to come. And it is a fitting analogy to the aeroplane or tank as it is still in its infancy, not quite ready for prime time.
Hardware follows the machine gun trajectory: costs fall, capability spreads, the technology becomes ubiquitous and undifferentiated. Software does not follow that trajectory. An autonomy stack that allows a drone to navigate, identify a target, and engage without a human in the loop is a different kind of weapon.
It removes the electromagnetic signature of the control link that currently makes drones vulnerable to jamming. It removes the human latency that limits reaction speed. And it changes the arithmetic of scale: a single operator can direct a swarm of drones, each sharing targeting data through a mesh network.
This is where the tank and aeroplane analogy holds. What made those technologies strategic was not the platform itself but the new doctrine that the platform made possible. Many veterans, myself included, grew up with the principle of combined arms warfare, of power projection across distance.
The autonomy stack, the targeting algorithms and the swarming protocols are the early blueprints of a doctrine that will restructure how wars are fought.
I don’t think even a clear-eyed observer in 1918 could have predicted combined arms warfare. It is similarly difficult, in 2026, to predict the doctrinal changes resulting from the autonomy stack. But the concentration of force that was necessary for combined arms warfare may not be possible in the future.
The aeroplane was innocuous in WWI, but iterations of it quickly led to an enormously powerful strategic capability. Since 1945 air superiority has been the primary goal of nation states at war.
The autonomy stack today seems similarly innocuous. But applying the same iterative process could see software superiority as critical to strategic success. In the future, the side that has software superiority will hold the strategic advantage.