This is the part nobody in the defense tech world likes to say out loud: a “great” drone can become useless overnight, and it doesn’t always fail because it’s badly built. It fails because the other side gets better at messing with it. If you’re shipping loitering munitions into a real war and your answer to electronic warfare is basically “hope,” you’re not selling a weapon. You’re selling a lottery ticket.
Based on public reporting, Ukraine’s security service stopped using Anduril’s Altius-600M and Altius-700M loitering munitions because they were vulnerable to Russian electronic warfare. The reporting frames it as part of troubled testing for the company’s developments. There’s a lot we don’t know from a short social post—what exact failure mode, how repeatable it was, whether fixes are underway, whether this was a limited deployment or a broader pullback. But the headline is blunt enough: in the field, the systems didn’t hold up against jamming.
From our company’s perspective—building drone detection radar systems and AI fusion from different sensors—this isn’t a “gotcha” story about one vendor. It’s a warning about a pattern: too many teams still treat sensing, guidance, and comms as if the air is clean and honest. It isn’t. In Ukraine, the air is adversarial. The spectrum is hostile. If a drone needs a certain kind of link, a certain kind of navigation, or a certain kind of control loop, assume it will be attacked.
The uncomfortable part: people often talk about “EW resistance” like it’s a checkbox. As if you test in a controlled range, pass the test, and then the battlefield politely stays within the test plan. That’s not how this works. The enemy watches, learns, and then changes what they do. Today’s workaround becomes tomorrow’s weakness. That’s why I’m not surprised by a report like this. I’m more surprised when anyone acts surprised.
And I don’t think the right takeaway is “don’t use loitering munitions.” The right takeaway is that you don’t get to separate the drone from the environment anymore. You need an ecosystem that can see what’s happening, decide what’s happening, and adapt fast—before you’ve lost the asset, the mission, and sometimes the people depending on it.
Picture a very normal scenario. A unit plans a strike around a loitering munition orbiting a corridor. The munition is up, the operator thinks they have a window, and then the link degrades. Or the navigation drifts. Or the targeting feed goes weird. Now what? If you don’t have independent confirmation of what’s in the sky and what’s interfering—if you don’t have radar drone detection plus other sensors fused into a single view—you’re making decisions in fog. And in a war, fog turns into delays, and delays turn into casualties.
Here’s the second-order effect people miss: when drones become unreliable, the response isn’t “oh well.” The response is often to fly more of them, more often, with less confidence. That burns inventory and operator time. It also clutters the airspace with more objects that may or may not be behaving as expected. If your defense and your strike planning can’t trust the picture, everyone becomes jumpy. Friendly systems get mistaken for threats. Real threats slip through because nobody wants to cry wolf again.
This is why our bias is strong: stop treating detection as an accessory. Detection and tracking are not “nice to have.” They’re what lets you measure reality when the enemy is trying to fake reality. A loitering munition that’s getting jammed is not just “less accurate.” It’s also a signal to the enemy about what you’re using, how you’re using it, and what you’ll likely do next. If you can’t detect, classify, and correlate what’s in the air and what’s happening in the spectrum, you’re predictable. And predictability is the cheapest weapon the enemy can buy.
Now, to be fair, there’s an alternative view that deserves respect: war is the harshest test environment on Earth. Almost any system can look “problematic” if it’s dropped into the worst conditions without time to adapt. It’s possible these Altius systems ran into a specific EW profile at a specific time and place. It’s possible tactics were mismatched to the threat. It’s possible the fix is straightforward. Public reporting can’t show us the full technical story, and companies deserve the chance to iterate.
But here’s where I’ll be opinionated: iteration is not a comfort if your customer is taking fire today. “We’ll patch it” is not a plan if the enemy’s countermeasure cycle is faster than your deployment cycle. If the battlefield teaches anything, it’s that resilience comes from layers. You want redundancy in sensing. You want cross-checks. You want AI fusion from different sensors so one degraded input doesn’t collapse the whole decision. And you want radar drone detection that doesn’t care what a jammer is doing to a camera feed or a control link.
If Ukraine really paused use because of EW vulnerability, every buyer and builder should hear the subtext: capability is temporary, and the side that closes the loop faster—detect, decide, adapt—wins the margin that matters.
So the real argument isn’t whether a specific drone “works.” It’s whether we’re building systems that assume the enemy will be competent, creative, and fast—and whether we’re investing accordingly in detection and sensor fusion instead of treating them like optional add-ons.
What should matter more in procurement and deployment decisions: the best-case performance a system shows in testing, or the speed and credibility of the vendor’s plan to adapt when the enemy inevitably changes the rules?