Drones Are Becoming Cell Phones
Three governments just told us how the next decade of autonomy is procured: Like cell phones.
In the space of four months, the United States, Germany, and Ukraine have each done the same thing that was unthinkable:
The US Drone Dominance Program ran its first Gauntlet, ordered tens of thousands of attack drones from the top performers, and is now climbing from short-range FPV toward longer-range strike systems in its next round. Germany approved parallel contracts with three loitering munition makers at once and is having soldiers test them side by side rather than in sequence. And Ukraine, through its Defense Procurement Agency, just launched its first open competitive tender for strike and deep strike drones, buying not by model name but by tactical and technical characteristics.
This is not a coincidence. It is a signal about how autonomous systems will be procured for the next decade, and it has a clear consequence for who wins.
The cell phone model: Attrition built into the product
For seventy years, defense procurement worked one way. A government picked a contractor, co-developed a custom platform with them, and committed to sustaining that platform for thirty or forty years. The F-16 first flew in the 1970s and still flies today.
Start from first principles and you can see why. Every one of those systems was built around a human. A fighter, a tank, a ship has to carry a crew and, above all, protect it. Protecting a person means armor, life support, redundancy, escape systems, and all the structure to hold it together. Platforms can’t be lost without losing human life. That drives up size, which drives up cost, which drives up complexity. When a single platform costs tens or hundreds of millions of dollars and takes years to build, you do nothing or throw it away. You sustain it for decades, and you co-develop it with one contractor who will keep it running.
The sustainment model was a rational response to the cost of keeping a human alive inside the machine.
Take the human out and the entire cost structure collapses. An autonomous system has nothing to protect. No crew, no armor, no life support, no escape systems. It can be small, cheap, and fast to build. And once it is cheap and fast to build, it stops behaving like a fighter jet and starts behaving like a cell phone: largely similar across vendors, quickly aging, and expected to be replaced within two or three years. No serious military plans to fly the same attack drone in 2032 that it bought in 2026. The threat environment moves too fast, and the hardware is cheap enough that replacement beats sustainment.
When the product ages out every few years, the old model stops making sense. You do not marry a single contractor for a generation. You run a competition, you buy what works, and you run the competition again. That is exactly what the US, Germany, and Ukraine are now doing, each in their own way, at the same time.
Chips, then computers, then platforms
Every hardware industry matures the same way, in three steps. First the components. Then the integrated products built on top of them. Then the platform layer that standardized the whole industry. In every case, the durable value ends up in the top and bottom of the model: the chips powering every device and the information platform tying everything together.
Computing is the clearest example. First came the chip. It was the brain of the machine, it mattered enormously, but it was a component. Then came the computer: a wave of companies that each built their own box, top to bottom, and competed on the box. For a while that was the business. Then the boxes became interchangeable. They got cheaper, they were replaced every few years, and one machine was much like the next. The hardware makers ground each other down on price.
Microsoft did not build the best computer. It built the layer every computer ran on, and it compounded across every hardware generation while the hardware itself turned over. The value migrated up the stack, away from the thing that aged out and toward the thing that stayed.
Owning the operating system layer gave the West durable technology leadership in computing. In robotics, that same layer is more than a commercial prize. It is a requirement for credible deterrence.
Robotics is following the same path, just later. First came the inertial sensing chips that became cheap because smartphones used them for gaming. Then came autopilots: the Pixhawk flight controller (that I designed in parallel to my CS degree) arrived around 2014 and gave the industry its common brain, the Stanford computer club equivalent of its era. Then came the full systems: vertically integrated companies that each built their own airframe, electronics, and software, top to bottom, and competed on the airframe. That was the right strategy for that phase, the same way the early computer makers built everything in-house.
We are now entering the third phase. The procurement signals from three governments indicate a tectonic shift: when buyers treat drones as interchangeable and replaceable, they are telling you the airframe is being standardized, exactly as the PC was. The value is moving up the stack. The only question left is who owns the layer it moves to - and if specialized vendors that are focused on one domain (small drones or long range drones or boats) can withstand the pressure of big, vertically integrated players.
A faster prime is still a prime
The obvious answer, to many people, is the neo-primes: the software-first, fast-moving companies built to displace the legacy giants. And they were right about almost everything. Right that software wins. Right that the old development cycles were too slow.
But look at the shape of what they are building. The neo-prime model is to become the new prime: own the integration layer, build much of the hardware, and consolidate it into large, multi-year enterprise programs. That is the prime contractor model, run by a software company at higher speed. The structure is unchanged. One integrator, one bundled stack, betting to own long-running programs that are quickly going away.
That structure is exactly what the procurement shift breaks. When a government runs an open, repeating competition and buys by capability rather than by name, it is refusing to lock itself into one vendor’s bundle. It wants interchangeable hardware from many suppliers, and the freedom to swap the airframe every few years without tearing out the autonomy underneath.
This is good news for new entrants and family-run businesses: they get a chance to re-enter a market the neo-primes have tried to monopolize, with multiple opportunities a year to win on scrappiness and ingenuity. Much like the young Michael Dell, who built a computer company out of his dorm room and outmaneuvered far larger rivals. But like Dell, they need a dependable partner who delivers a competitive tech stack every year, so they can focus on what they do best: building the aircraft, boats, and vehicles that win the competition.
The honest objection here is that these companies say they are open too. They publish an SDK. They sign interoperability partnerships. Anyone, in principle, can integrate. But openness is not the same as neutrality, and the difference is the whole game. When the company that owns the integration layer also builds competing airframes, every other vendor is being asked to plug into a platform controlled by a direct competitor: one who sets the roadmap, sees their integration details, and has every incentive to make its own hardware work best. Rational vendors do not commit their future to that. They integrate defensively, hold back their best work, or go build their own stack. The ecosystem never forms, because no one trusts an open door owned by a rival.
That is precisely where the government’s goal fails. The point of buying by capability across many suppliers is to get true multi-vendor integration, a real and competitive ecosystem beneath one common layer. A layer owned by one of the competitors cannot deliver it. It delivers the appearance of openness and the reality of dependence on a single company, which is the outcome the new procurement model was designed to avoid. So the openness piece does not rescue the position. It is the position’s ceiling.
Computing already ran this experiment, twice. The operating system that won was the one that ran on everyone’s hardware, not one welded to a single manufacturer. The neo-primes have chosen the opposite position, the prime’s shape kept intact and only made faster. The shift now underway does not reward a faster version of the old shape. It rewards a different shape entirely: the open, neutral platform that sits beneath every vendor’s hardware and is owned by none of them.
A new era in defense tech
Auterion is the first company of a new generation. We are not building a faster prime. We are building the platform for a world where customers are no longer loyal to one vendor, one airframe, one boat: the operating system that any airframe can run, owned by no single hardware maker, that the next procurement cycle extends instead of replaces.
We are not an adapter layer that merely “integrates,” like many of our competitors. AuterionOS is the operating system for warfare: the capabilities and integrations in the warfighting tech stack that governments invest in and keep, while the hardware around it turns over.
On top of it, we bring together an alliance of manufacturers and specialized software companies, each the best in their craft: multicopters, long-range drones, boats, rovers. None of them covers air, land, and sea alone, and none of them has to. Each joined by choice, because together they outcompete the vertical players.
We made the deliberate decision not to bet the company on a single airframe. When a government buys by capability rather than by model name and replaces its drones every few years, it is commoditizing the airframe and concentrating the durable value in the platform beneath it. The drones are the handsets, which differentiate on cameras, battery life, range and ruggedness vs. cost. We are the operating system that brings them together into the swarm.
The three announcements this spring are not separate procurement stories. They are one structural shift, the shift the entire history of hardware predicts. The companies that win the next decade of robotics will not be the ones with the best drone in 2026. They will be the ones who own the layer that every drone, in every cycle, has to run on.
That is the company we are building.