11:49
- In November 2024, Professor Todd Humphreys, a GPS expert at the University of Texas at Austin, got a mysterious tip-off. Look at two specific days, at exact times in a data set collected years earlier by a network of GPS monitoring stations. These stations continuously record signals from satellite navigation systems, measuring how strong they are relative to background noise.
It didn't seem promising. The data was from 2021, and it was publicly available online. So surely if something unusual was in there, someone would have already found it. But Professor Humphreys and his student Zach Clements looked into the data at the two specific times suggested, and they discovered something surprising. At one precise moment, receivers across the network all reported the same thing, a sudden drop in signal-to-noise ratio. In effect, the navigation signals had been overwhelmed. The measured signal-to-noise dropped by roughly a factor of 10. But when they started looking back through the data, they found 75 days since 2019 where these strong disruptions had taken place.
And again, and again, the pattern was the same. All these receivers recorded the same drop at the same time. - This effect was being felt all the way across Europe, all the way to the north, to Svalbard, to the south, to Spain, all the way to Canada in the west, as far east as eastern Poland. And in fact, it had a distinct pattern across Europe to where it looked like the blast center was in Poland or Kaliningrad in that area. - So, they wondered what could cause this. At first, they thought perhaps it's interference, something on the ground broadcasting a much stronger signal, drowning out the ones coming from the satellites, which made Kaliningrad an interesting candidate.
It's a heavily militarized Russian exclave on the Baltic Sea wedged between Poland and Lithuania. And in recent years, it's become a hotspot for this kind of electronic warfare. So maybe that was it, a side effect of the conflict in the region. But they realized that explanation didn't work. - Even the highest tower in Kaliningrad only affects aviation as far as, say, Denmark. So, we had ourselves a mystery on our hands. We were experienced with interference that you could call local, something on a tower, something on a plane that might cover hundreds of kilometers. But this was different.
This was continental scale. - The researchers mapped out all the stations that were affected and asked a simple question. What kind of source could have affected all of them at once? The source on the ground just wouldn't work. The curvature of the earth would block the signal over these distances. The only way to affect such a wide area simultaneously is from high above the earth. And even using the most conservative assumptions, that the source only has to just clear the horizon from the perspective of every station, the geometry shows it had to be
at least 1,200 kilometers up. That's way higher than the International Space Station. So, what was causing this? Well, there is one known source of radio interference from space, which is the sun. In November 2025, a major solar storm disrupted GPS positioning globally for several hours. So, could these interference events be caused by the sun? There were certainly question marks. Solar interference typically builds up and fades over tens to hundreds of seconds. But these events were abrupt, short bursts of only three to five seconds.
Another important clue was the frequencies that were disrupted. Solar interference typically spans a wide range of radio frequencies, whereas this interference was confined to a very narrow slice, just five megahertz wide, centered at 1,577.5 megahertz. That's right in the part of the spectrum used by GPS. And the final nail in the coffin is that solar radio bursts affect the entire sunlit side of the earth. But what the researchers were seeing here was much more concentrated, consistently centered over Europe. So, Todd and Zach realized the only thing that could be causing this was a satellite, which left them with just one question.
- Is it intentional or is it accidental? Both my student and I looked at this and said, unprecedented. We've never seen this before. The world has never seen this before. - [Host] This satellite seemed to be targeting GPS. - Or rather what most of us call GPS. That is just the shorthand. GPS is the American system, but there are others, including the Russian one, the European one, and the Chinese one. And your phone is actually of these satellites to try and tell you where you are.
These systems are known as Global Navigation Satellite Systems or GNSS. And when the U.S. military first developed the system in the 1970s, they wanted to prove how well it really worked. So, the GPS program office set itself an ambitious goal to use GPS to drop five bombs in a row into the same hole from a moving plane. - They set up a chair next to the site that the bombs are going to be going into. Wait, what? If anyone doubts that we can drop the bombs in the same hole, you're welcome to go and stand right there and watch out.
(bombs blasting) - So how does GPS work? Well, think about what your phone is actually trying to do to tell you where you are. It's listening to signals from satellites. And those signals can tell it two things. First, the position of the satellite in space. And second, the exact time the signal was sent. And then the receiver can compare that to when the signal arrives. And you can use that time difference, multiply it by the speed of light, and you get the distance. So, if it took the signal 0.07 seconds, that must mean you're around 21,000 kilometers away from the satellite. If you're only using a single satellite,
that puts you somewhere on a huge sphere around it. Now you can add a second one, and that will narrow your position down to this circle where the two spheres intersect. If you add a third, you will narrow it down to only two possible points. And only one of these is actually on Earth's surface. So, you might think three satellites is all you need. We have three equations, one for each satellite, and three unknowns that are my position, x, y, and z. So, on the left side, I have the geometric distance between me and each of the three satellites. And on the right side, I have the measured signal time, all multiplied by the speed of light.
Now, satellites have atomic clocks, so their timing drifts by only three seconds every million years. But my phone does not contain an atomic clock, so its clock can drift by that much in just a couple of days. And a timing error of just 100 nanoseconds throws off your position by 30 meters. So, in reality, each distance is slightly wrong, which means the position you calculate is slightly off too. So now, instead of just solving for position, I also have to solve for the clock error. So, let's call that b. But now I have three equations and four unknowns, which is unsolvable. So, I can fix that by adding a fourth equation or a fourth satellite.
Four equations, four unknowns, that makes the system solvable. Okay, but we've made one key assumption, that when I'm calculating my position, I know where the satellite is. And sure, it can tell me where it is, but its position around the Earth and its speed are constantly changing. So how does the satellite actually know where it is in the first place? - The GPS satellites need to be told where they are, and they're told where they are with ground stations that monitor them. And the ground stations have to know where they are to tell the satellites where they are. - [Gregor] But the Earth itself is constantly shifting as the continents move.
- So, how do we work out where the ground stations are? You work out using quasars where the ground stations are. - [Gregor] Because they're billions of light years away, quasars appear almost perfectly fixed in the night sky, which means we can use them as reference points, much like early explorers navigating by the stars. A small fraction of quasars are extremely bright radio sources. When those radio waves reach Earth, telescopes around the world detect them at slightly different times. From those tiny differences in arrival time, you can precisely determine the relative position of the telescopes. And by making lots of those measurements over time, you can build a precise map
of how the Earth itself moves, continents included. GPS ground stations are tied into that same map. - [Ramsey] The ground stations tell the satellites where they are, and then that's how you work out where you are. (Gregor chuckles) - You just blew my mind. - So, we're always navigating by the stars, just via GPS. - And in reality, your phone isn't connected to just four satellites. It's usually far more than that. Mine right now is using, it's like over 20 satellites. That's because the system is doing much more than just solving for four equations.
Say you're on this little street in Paris and you open maps and you try to figure out where you are. Turns out there are a lot of things that can throw this off. There's the relatively simple ones, like taking into account special and general relativity. If you didn't correct for that, your position would drift about 11 kilometers per day, and you can be anywhere within this entire area. So, the satellites are constantly being monitored and updated, and they send down timing corrections that your phone applies in real time.
Now, on top of that, in the fraction of a second that it takes for the signal to reach you, the Earth has actually rotated slightly. So, when your phone figures out where you are, it also has to account for the fact where Paris was when the signal was sent, not where it is now. If you don't correct for that, that adds about 20 meters of error, enough that you're not even on the same block. And of course, the signal has to pass through Earth's atmosphere. Now, the ionosphere will add between 5 and 15 meters of error to your position, and the lower levels of the atmosphere, with weather and pressure and humidity, will add another meter or two. So, if you don't account for that,
you'll still be on the wrong side of the street. And that's where the extra satellites come in handy. They give your phone extra measurements to separate out the errors and give you a better fix of your position. So ultimately, it can be accurate down to 3 to 5 meters, which is pretty incredible if you think about what it takes to solve this. And with the most advanced GNSS tech, you can push this even further to a few centimeters. But at this point, you actually have to start accounting for stuff like how gravity bends space-time and delays the signal. But all of this only works if you're actually able to hear those signals.
- [Ramsey] When the signal arrives from space, it's traveled 20,000 kilometers to get here, and it was only transmitted with about 50 watts of power. So, it's like an old school light bulb two Earth diameters away. - [Gregor] And because that power spreads out over such a vast distance, the received GPS signal at your phone is only about a tenth of a quadrillionth of a watt, or 10 to the negative 16 watts. - It's insanely quiet. - Which means it doesn't take much to drown it out. If you broadcast enough noise at the same frequency, you can overwhelm the signal until it becomes indistinguishable from the background.
This is called jamming. And a lot of these navigation systems use the radio spectrum, roughly between 1.1 and 1.6 gigahertz. Lower frequencies travel well through the atmosphere, but they require much larger antennas. Higher frequencies allow for smaller antennas, but the signals are much more easily absorbed and scattered, especially in bad weather. So, this band is a kind of sweet spot that navigation systems rely on. Which means if something interferes with it, it can affect multiple systems at once. That's why this part of a spectrum is tightly protected.
Nothing else is supposed to be transmitting here. And deliberately interfering with these signals is illegal in most countries, but something was. Something powerful enough to overwhelm satellite navigation across an entire continent. Which brings us back to the question facing Professor Todd Humphreys. Was this accidental? Or was someone flooding this band on purpose? - It's, you know, a mystery novel here where you're trying to figure out whether it was the butler in the kitchen with the crowbar, right? And we have to put all of the clues together.
If it had been random, occurring because of some failure in a satellite, maybe some amplifier gets too hot and misbehaves, then you'd expect it to happen equally likely on a Monday and a Friday, right? - [Host] But when they went back through the data, something strange emerged. The interference events were happening mostly on Tuesdays, Wednesdays and Thursdays during business hours in Europe. - So, tell me, tell me what amplifier tends to fail during business hours in Europe on Wednesdays. - So now we know this probably isn't random.
There's likely human input. But that doesn't necessarily mean it's intentional or malicious. So, the hunt was on to track down the actual source. But that turned out to be harder than you might think. Because there were still so many unknowns. In situations like this, uncertainty can quickly turn into competing narratives. We've seen it before. When the EU chief's plane was allegedly hit by GPS interference last year, over 150 news outlets reported on it. But depending on where you saw the story, your perception could be completely different. If you saw Le Monde first, you might have come away thinking it might be Russia's fault. If you saw Newsweek first, you would definitely think it was Russia. And if you saw this headline, well,
you might question the veracity of the story altogether. Three headlines, three entirely different conclusions. That's why we use today's sponsor Ground News. They're a website and app designed to make reading the news easier and more data-driven. Every day they pull in thousands of stories from around the world. And each story comes with visual breakdowns of political leaning, factuality rating and ownership. This is all backed by ratings from three independent media monitoring organizations. Let's go back to that story to see it in action.
Right away, you can see that over 150 news outlets reported on the story. Below that, you can see that coverage was split relatively evenly across the political spectrum. And below that, you can check the overall factuality and ownership. Even see who broke the news. They even have a bias comparison feature that highlights specific differences in the contents of the articles. One of my favorite things is their blind spot feed, which shows you stories that are underreported by one side of the political spectrum or the other.
Staying informed with accurate information has become a full-time job. But that's what makes Ground News so valuable. Their data-driven approach helps make information accessible and digestible for everyone in minutes rather than hours. So, if you, like us, care about getting to the truth, go to ground.news/VE or scan this QR code. Our link gets you 40% off their vantage plan. So, I wanna thank Ground News for sponsoring this part of the video. And now, back to the hunt for the source. - To do this, we have to set up a kind of elaborate estimation problem, but there are many unknowns. - A jammer works by broadcasting a stronger signal in the same frequency band.
It's kind of like yelling over someone whispering. So, if this were a jammer on the ground, you could track it down by looking at signal strength. The closer you are, the stronger the signal will get. But once you're dealing with something in space, that intuition breaks down. To work backwards from signal strength, you'd need to know how much power the satellite was transmitting, and the exact shape and direction of its beam. Even how each receiver's antenna responds to signals arriving from different angles. And at 1,200 kilometers above the Earth, very different satellite positions can produce almost identical patterns on the ground.
So instead of trying to pinpoint the transmitter directly, the researchers used a simpler filter. If a satellite caused interference at all these stations at the same time, then it had to be above the horizon for all of them at once. Right now, there are over 15,000 active satellites in orbit. But that one constraint lets you eliminate over 98% of them. That still leaves around 200 possibilities. What's left are satellites in very high orbits, many of them in or near the geostationary belt, where they move at the same rate as the Earth rotates, so they stay fixed over one point. From about 36,000 kilometers up,
a single one can see a huge portion of the Earth at once. However, if the same satellite caused multiple events on different days, then it has to satisfy that condition every single time, which narrows it down to just 14 possible suspects. - So, what kind of satellites and countries were in these 14? - Well, there were several interesting candidates. One was a satellite operated by Algeria, and it even had, on public documentation, a transmitter in the same frequency band we were seeing interference on. - So, they took a closer look, and on paper it fit. It was visible to all of the reference stations, and it could transmit in exactly the band they were seeing.
This looked like it could be the answer. But they had a nagging doubt that some of the most distant stations in Svalbard and western Greenland, this satellite was barely above the horizon, which meant if it was the source, the signal would have to be skimming across the Earth and just barely glancing into the antenna of these receivers. Now, that's not impossible. Low angle signals can be picked up, but the more researchers looked, the less certain they became that this satellite was actually the source. So, they examined data from GNSS receivers that were tracking this satellite during the interference events.
- It looked just like all of the other ordinary GPS signals that were getting hammered by this interference. So, the ratio of the signal power to the background noise also dropped and dropped in the same proportion that the other GPS signals were dropping at the same time. This was not the attacker. This was just another legitimate signal in the L band that had been interfered with. - The Algerian satellite wasn't the source. It was another victim. But what about the other 13? Any one of them could have been visible to all the stations at once. So geometrically, they could all be the culprit. But that's where the trail started to go cold.
I mean, none of them had clear public documentation showing they could transmit and jam the right frequencies. And some of them had barely any public information at all. - We couldn't narrow it down further because of all of these things we didn't know about the signals they were broadcasting, how intense they were, what their pattern was. So, there we were stuck for about four months. - And there was an even more fundamental problem. Everything they'd done so far relied on one idea, that all of these interference events were coming from a single satellite.
- It was under a really controversial assumption. Controversial in that my student and I just couldn't come to fully embrace this assumption. If we broke our assumption, then that would relieve the focus on geostationary satellites. Then we could open up the focus on many other satellites. But now instead of 13 remaining on our list, it would expand to, you know, maybe 100 or more satellites. - And suddenly, instead of closing in on an answer, it felt like they were back at square one. So, they needed a completely different approach.
All the data they had was from GNSS receivers that process everything internally. They lock onto satellites, track signal strength, and then output a clean simplified measurement, how strong that signal is compared to the background noise. And they do that just once per second. But these interference bursts only lasted three to five seconds, which means each event is captured in just a handful of samples. You can see that the signal dropped, but not precisely when it dropped. Not with enough timing resolution to compare one receiver to another, to see which one was hit first, or how the interference moved across the network.
What they needed was the raw radio signal itself. So, they started designing specialized receivers that could capture the data in much higher temporal resolution. The plan was to deploy them all across Europe. But building them, installing them, and waiting for new events to happen would take months, maybe years. So, in September 2025, they took their investigation public at the Institute of Navigation Conference in Baltimore. And the response was immediate. The room was packed. Everyone wanted to know the same thing.
What was this satellite and who was responsible? - Honestly, it was one of the most fun meetings I've been part of. It was almost like we had been brought together for a big brainstorming session to see if we could crack this problem as a community that my student and I had not been able to crack by ourselves. - [Host] The question of a mystery satellite propagated throughout the research community. The German Aerospace Center came up with an ambitious plan to take a large tracking dish, point it at candidate satellites, and try to catch the interference in real time.
- But remember, we're trying to catch a needle in a haystack that only shows up over a short interval and only happens once, maybe once or twice a month. - [Host] They tracked satellite after satellite and nothing. No signal, no anomaly, no breakthrough. Yet again, the trail went cold. Weeks passed, then months, until an email arrived. - I looked at the email and then I had to reread it because it was almost like it had come right out of my dreams.
You know, this can't possibly be true. He had beautiful raw data from two different stations. - At last, they had what they'd been missing, the raw radio signal itself. Samples of the raw voltage off the antenna capturing the actual digitized radio waves millions of times per second from two stations, one in Amsterdam in the Netherlands and one in Trondheim in Northern Norway. They captured these interference events on February 11th, 2026. And now, rather than asking how strong is the signal, they could ask, when did it arrive? With this raw data, they isolated a 2.3 second window where both stations recorded the same event.
And because they were sampling at tens of megahertz, they could align the two data sets to pinpoint the exact moment the jamming signal arrived at both stations. This meant they could measure the tiny difference in when the signal reached each station. Imagine the signal reaches Trondheim roughly 139 microseconds before it reaches Amsterdam. Well, that tells you that the source must be slightly closer to Trondheim than it is to Amsterdam. And that constraint gives you a shape, a curve in space of all the possible locations of the source where that exact timing difference would be observed.
In three dimensions, that becomes a kind of warped surface, a hyperboloid. Whatever transmitted the signal had to lie somewhere on that surface. And because the raw data was so incredibly high resolution, the margin of error, the thickness of that shape, stretching tens of thousands of kilometers into space, was only about five meters. - And because we were so interested in avoiding errors, we just determined that we were not going to talk to each other for a week. We would do our separate thing for a week. And then we would come back together, compare notes, and we would hopefully have not made the same mistakes. And I presented my findings to Zach and he presented his to me and they were not the same.
And I have to admit that because I'm the professor, I was pretty much insistent that my finding was correct. But then he pointed out a couple of other things, other checks that he'd run. And then I realized that I had an error in my code. And once I fixed that error, I ended up getting exactly the same time difference of arrival that he did. - And now they had a way to test every possible satellite, not just the 14 from their earlier predictions. For each satellite, they took its known orbit and asked if the signal came from here, what timing difference would we expect between Amsterdam and Trondheim? Then they compared that prediction to the real data. If it didn't match, the satellite was ruled out. And since they had a continuous two
and a half second recording, they could take this a step further. As the satellite moved, that hyperboloid would shift slightly through space. So, the real source had to stay aligned with it the entire time, which made this test incredibly strict. - And only one satellite was anywhere near this time difference of arrival measurement. And in fact, it wasn't just near, it was dead on. - The only possible culprit was a Russian satellite, Cosmos 2546. It remained perfectly on the hyperboloid, aligning with their data to within 200 meters, which is well within the uncertainty of the satellite's publicly available orbital data.
Now it's worth saying here, this work is still new and it hasn't undergone peer review, but independent teams in Europe have verified aspects of these findings. There was only one slight issue. Cosmos 2546 was launched on the 22nd of May, 2020. So, it couldn't explain events going back to 2019. But then they discovered that Cosmos 2546 is part of a constellation of six satellites. This constellation is part of Russia's early missile warning system. - It's basically part of their Golden Dome. If you've heard of the U.S. Golden Dome, this is part of the sensory apparatus for the Russian missile detection system.
- They sit in what's called a Molniya orbit. A highly elliptical path that carries them high over the Northern hemisphere, where they can slow down and linger. It's the way that high latitude countries like Russia can get the benefits of something like geostationary orbit across multiple satellites. But this also means this constellation covers large parts of the globe. So, they have the capability to jam GNSS across a far wider region than we've seen, including over the United States. The pattern already told us this wasn't random.
The frequency told us it wasn't natural. And now we know it's coming from a military system in orbit. But intention, well, that's harder to prove. - What is already being broadcast is enormously powerful. It's like hundreds of times more powerful than the GPS signals themselves. - But it's slightly offset from the GPS frequency, which seems odd. Because surely if you want to jam GPS properly, you would put your jamming signal directly on top of it.
Well, Todd actually has a theory about this. - If you're going to test this capability, then you test it in the neighborhood of the signal you intend to jam, but not right on that signal. And you test it only briefly just to make sure everything's still working. And then in the eventual future, where there is a hot conflict, they go ahead and tune their transmitter down to the GPS band. But it's much more damaging now that it lies right on that band. - [Derek] So, one possibility is this. What we've been seeing are tests, a system being exercised without fully revealing what it can do. And while just a theory at this point, there is another hint this might be the case.
The raw data also revealed a second interference burst aimed at a lower frequency, 1,558.5 megahertz, which overlaps with signals from the Chinese BeiDou navigation system. - I'll say it in the most conservative way possible. I can no longer say this is accidental with confidence. I'm leaning toward this being a periodic test of a capability that would be very damaging if it was deployed in anger. - [Derek] But when we spoke to a team that had independently traced the signals back to satellites in Russia's missile warning system, they had another idea.
- So that'd be very odd behavior to be continually testing something. So, our alternative theory is that those might actually be very short, very brief, very specific comms messages coming from those satellites. - [Derek] And if that were true, transmitting messages on these frequencies would provide a layer of protection because the enemy wouldn't want to jam them and risk disrupting their own navigation systems. - I'm not saying it's not the jamming. That's still a very strong possibility. I think the point is there are some odd things about that and there is clearly here an alternative explanation.
- But whether these signals were covert messages or tests of a space-based jamming system, the events themselves were incredibly brief. And that might explain something else. Why this went unnoticed for years. But if this system were ever fully switched on, the impact could be enormous. When we see maps of massive GPS interference today, you are almost entirely looking at reports from commercial aircraft affected by ground-based jammers. Typical airliners cruise at altitudes between 30,000 and 42,000 feet, which is high enough to remain in direct line of sight of jamming signals across vast distances.
But at ground level, buildings and terrain shield much of our infrastructure from that interference. - But everything is line of sight from space, yeah. It's an invisible utility that supports virtually every technology. And it's essential to the way we live every day. The United States made GPS its gift to the world. And clever engineers took this free thing that works really well and very precisely and incorporated it into all kinds of systems. It's one of those wicked problems. You really don't know what all the dependencies are and where all the interconnections are because it's just proliferated throughout the world.
- [Derek] Global satellite navigation systems underpin aviation and global shipping. It synchronizes financial systems, keeps cell towers on time. It is crucial for logistics networks, ride hailing apps, delivery systems, even dating apps. A disruption of these systems across an area as large as what these satellites are capable of could affect hundreds of millions of people. - It would send shockwaves of fear through the world, not just through Europe, because an entirely new weapon from space would have been revealed. And people would know that at any time of their choosing, Russia could deploy this over their country
or over their continent. I do think that this is a massive escalation in the electronic warfare background conflict that's going on right now. - [Derek] Now, this sounds scary, but the truth is we should have been worried about this fragile system already. - The threat has always been there. Severe solar events could ionize the atmosphere and keep all kinds of signals from space from coming in, or it could destroy satellites.
We could have a version of the Kessler effect in medium Earth orbit, and satellites be destroyed by debris. So, everybody should have already been concerned. They should continue to be concerned. And this is just another example of why. - [Derek] But the good news is, we already know exactly how to fix the problem. It involves building a future system that doesn't rely entirely on weak satellite signals. - We suggest that a resilient national PNT architecture would be signals from space, terrestrial broadcast, and fiber, because those are three completely different
phenomenologies in a vector that would impact one, would not impact the other two. - Some countries are already doing this. Places like South Korea, China, and the UK are building out backup networks that don't rely on signals from space. They're using fiber optic cables to securely share time from atomic clocks here on Earth. And for navigation, they're turning to systems like eLoran, huge high-powered radio towers that are much harder to overwhelm.
Other solutions are even being developed using magnetic and quantum systems that use subtle variations in the Earth's magnetic field to work out position. So, you don't need any external signals. But despite all of this, most countries, including the United States, still remain highly dependent on GPS and other satellite navigation systems. In this video, we've been talking about jamming, but there's another kind of GNSS interference, one that swaps your real signal for a fake one, quietly feeding you a location that looks completely real. So, if you trust it, it can guide you somewhere else entirely.
This is known as GPS spoofing. It's affecting over 1,500 flights per day. And at sea, we're seeing the same thing. Ships are jumping implausible distances, briefly disappearing or reappearing in impossible locations, even on land. It's a systematic, geographically clustered distortion of reality, both in the air and at sea. If you wanna know more about that, let us know in the comments. This story was based on research by Professor Todd Humphreys and his student, Zach Clements, at the University of Texas at Austin.
There's a link to their work down in the description. I wanna thank them for all their help and for sharing their findings with us. And of course, I wanna thank you for watching. (upbeat music fades)
