One of the hottest investment themes of the last few years has suddenly gone cold.

The iShares U.S. Aerospace & Defense ETF (ITA) recently traded around $208, down over 17% from its all-time high reached less than two months ago. Some individual defense stocks have fallen even further and are already in bear market territory. And after years of strong gains across the sector, investors suddenly seem a lot less excited about defense.

That’s worth paying attention to.

Because while defense stocks have been falling, the technological arms race certainly hasn’t slowed down. In fact, some of the developments over the past couple of weeks suggest it may actually be accelerating.

And the biggest changes are happening in drones, artificial intelligence, and autonomous weapons.

Faster, Higher, Harder to Stop

Ukraine has become the world’s largest real-time laboratory for drone warfare technology, demonstrating that a relatively inexpensive unit can destroy equipment costing millions of dollars.

But now we’re seeing the next phase.

Russia has begun deploying more jet-powered Shahed drones that fly faster and higher than previous versions. Reuters reported last week that some operate at altitudes of roughly four to seven kilometers, putting them beyond the reach of many mobile air-defense teams.

The result is alarming. Ukraine had previously been able to intercept roughly 90% of the older drones. Its interception rate against the newer jet-powered versions has fallen to around 60%.

That creates an obvious problem.

You could shoot them down with expensive air-defense missiles. But firing a missile that can cost hundreds of thousands or even millions of dollars at a relatively inexpensive drone is not exactly a sustainable business model.

So, Ukraine is increasingly experimenting with another solution: Drones that hunt other drones.

Ukraine is testing new interceptor drones, including systems designed to attack incoming Shaheds. And this is quickly becoming a technological cat-and-mouse game. Every improvement on offense creates demand for a new defensive technology.

That cycle could last for years.

Battle-Tested Training Data

Here’s where the story gets even more interesting.

Earlier this month, the British Army successfully tested a drone swarm in a realistic military environment. Rather than controlling every aircraft individually, autonomous systems can increasingly coordinate with one another and operate as a group.

Then Britain took another major step last week.

Up to 12 British companies will be given access to Ukraine’s battlefield drone data to train new AI models for autonomous swarms. The database contains imagery and video collected by Ukrainian drones using both daylight and thermal sensors.

Think about what that means.

AI companies are constantly looking for better training data. Ukraine has accumulated something incredibly valuable: years of real-world battlefield data showing drones operating under some of the most difficult conditions imaginable.

Britain wants to use that data to train swarms capable of continuing to operate even when communications are disrupted or GPS isn’t available. A small number of soldiers could eventually control large groups of autonomous drones performing everything from reconnaissance and logistics to attacking enemy targets.

This is AI moving off the computer screen and onto the battlefield.

And Then We Have Lasers

Perhaps the most fascinating development involves AeroVironment (AVAV), one of the companies at the center of the next generation of defense technology.

AeroVironment is already known for its Switchblade loitering munitions and autonomous systems. But earlier this month, the company received a $464.8 million U.S. Army contract for its LOCUST high-energy laser counter-drone system.

According to AeroVironment, it’s the first production contract for high-energy laser weapon systems in U.S. history.

Then, just days later, the company announced its first international LOCUST order worth more than $50 million.

For over a century, laser weapons belonged to science fiction. H.G. Wells gave his invading Martians a Heat-Ray in The War of the Worlds back in 1898.

In 1983, Ronald Reagan unveiled a missile-defense plan built partly around lasers, which critics quickly nicknamed “Star Wars.” The space lasers never materialized. Now they’re moving into production.

And drones may be the reason.

Once a laser system is deployed, the incremental cost of firing it can be dramatically lower than launching an expensive missile. That’s potentially a much better economic equation for shooting down waves of relatively inexpensive drones.

In other words, drones aren’t just creating a new weapons market.

They’re creating an entirely new anti-drone market.

You gotta love the irony at play here, too. Locusts are history’s most famous swarm; they were one of the10 plagues of Egypt. Now LOCUST is the system designed to knock swarms out of the sky.

Tomorrow’s Arsenal, Today’s Pullback

That’s what makes the recent correction in defense stocks so interesting to me.

I’m not suggesting every defense stock is suddenly cheap or that investors should blindly buy the entire sector. Some defense companies became expensive after their big runs, and the market is beginning to separate the companies benefiting from tomorrow’s military spending from those tied primarily to yesterday’s technologies.

But that’s exactly what creates opportunities.

The next generation of defense spending won’t just go toward bigger aircraft carriers, more tanks, and increasingly expensive fighter jets. More money is going toward autonomous drones, AI, sensors, electronic warfare, satellites, directed-energy weapons, and the software connecting all of them.

Companies such as AeroVironment, Kratos Defense (KTOS), L3Harris (LHX), Red Cat (RCAT) and Palantir (PLTR) are attacking different pieces of this emerging defense-technology ecosystem.

And remember what’s happening in the background.

ITA is about 17% below its recent high. Yet the technological transformation of warfare hasn’t reversed 17%. Drone development hasn’t declined 17%. And AI hasn’t suddenly become less important to the military.

If anything, the events we’re watching around the world suggest the opposite.

That’s why I’m paying very close attention to this pullback.

The stocks have retreated. The arms race hasn’t.