This budget request is either the most honest read of where war is headed, or a giant bet that makes it easier to start fights we can’t finish.
Because asking for $54.6 billion for something that got $225 million the year before isn’t a normal “we’re scaling up.” It’s a signal flare. The Pentagon is telling everyone—Congress, allies, rivals, and taxpayers—that autonomous warfare isn’t a side project anymore. It wants to be the main event.
Based on what’s been shared publicly, the target here is the Defense Autonomous Warfare Group, and the proposed jump is wild: a 243-fold surge by fiscal year 2027. The public reason is pretty straightforward too. Cheap, one-way attack drones—the kind people call “kamikaze drones”—have been painfully effective in real wars lately. The Russia-Ukraine war is the example everybody points to. The US-Iran conflict gets mentioned too, though the details in what you shared are thin and I’m not going to pretend I know exactly what incidents they mean. The point is clear either way: low-cost drones are changing what “power” looks like.
My read: this is less about futuristic killer robots and more about math. If a small, cheap drone can destroy something expensive, the old logic of defense collapses. It’s not that tanks, ships, and air bases suddenly don’t matter. It’s that they become harder to protect and easier to punish. And once that’s true, the people who plan wars stop dreaming about perfect control and start obsessing over volume—how many things you can field, how fast, and how cheaply.
That’s the part that should make you uneasy.
When you spend billions on a few high-end systems, you have built-in restraint. You can’t replace them easily. You think twice before risking them. But when you can build or buy huge numbers of cheaper autonomous weapons, you change the emotional and political cost of using force. Leaders might feel like they can “do something” without paying the price of sending pilots, losing troops, or risking a major platform. That might reduce casualties for one side in the short run. It might also make escalation more tempting, because it feels more reversible.
Imagine you’re a commander and you have a choice. Option A: put a manned aircraft near a dangerous target and risk losing people. Option B: send a swarm of unmanned systems and accept you’ll lose a bunch of machines. Which choice gets made more often? And once that’s the default, how often do we end up in situations we would have avoided when the costs were more personal?
The other uncomfortable truth: if cheap drones work, everyone can get them. This isn’t like building a stealth bomber. The barrier is lower. That means the Pentagon isn’t just buying an advantage. It’s trying to keep up with an advantage that may not stay exclusive. So the question becomes: are we pouring money into something that will be copied fast, turning into a permanent global arms race where the “winner” is whoever can produce and replenish the fastest?
There’s also a quiet institutional story here. A budget explosion like this creates new power centers inside the defense world. New contracts. New careers. New reasons to push for more. Even if the original idea was sensible, the machine around it can become self-feeding. Once you build a big autonomous warfare budget, you need threats big enough to justify it every year. That’s how you end up with a permanent posture of urgency.
Now, the case for doing this is real. If the lesson from recent conflicts is that cheap drones can smash expensive gear, then not adapting is basically choosing to lose later. If you’re responsible for defending people, “we didn’t like the implications” isn’t a great excuse after the fact. And if autonomy helps reduce the risk to service members, that’s not a small moral point. It matters.
But there’s a difference between building defenses and building a new default way of fighting.
Say you’re a small country watching all this. You look at a superpower investing tens of billions into autonomous attack systems. Do you assume it’s purely defensive? Or do you assume you’re now on a timer to build your own deterrent? That’s how you get more drones in more hands, more often, with more mistakes. And mistakes with autonomous systems can be fast and hard to walk back, especially if identification is wrong, communications are jammed, or the pressure to act is high.
I also don’t love the way “less expensive” gets sold as “more responsible.” Cheap can mean efficient. Cheap can also mean disposable, and disposable weapons can make disposable decisions.
And here’s something I genuinely don’t know: what exactly counts as “autonomous” in this budget pitch. Is it mostly about navigation and targeting help? Is it full independent strike capability? Is it logistics? Training? Counter-drone defenses? The label can hide a lot. The public deserves clarity, because oversight can’t work if the category is so broad it means anything.
So yes, I think the Pentagon is reacting to real battlefield lessons. I also think a 243x jump is the kind of move that can quietly reshape national habits: what we build, what we fear, what we choose first when tensions rise.
If you were in charge, would you rather slow this down to avoid making war feel “too easy,” or speed it up because the side that hesitates could be the one that gets blindsided?