ESA Lost a Spacecraft for 33 Days — Then a Miracle Happened

[00:00:00]February 14th, 2026.

[00:00:03]Somewhere over the Indian Ocean, 60,000 km above Earth.

[00:00:09]A spacecraft is running a routine procedure it has done hundreds of times before.

[00:00:14]Nothing unusual. Nothing alarming.

[00:00:18]Just a standard reset of its orientation system.

[00:00:21]And then something goes wrong. Not explosively wrong. Not the kind of wrong you see in movies with alarms and sparks and people diving under consoles.

[00:00:32]Just quietly, invisibly, catastrophically wrong.

[00:00:37]A line of code hits a combination of conditions [music] it was never designed to handle.

[00:00:42]The spacecraft loses its sense of direction.

[00:00:46]Its solar panels drift away from the sun.

[00:00:49]The battery starts draining.

[00:00:51]And the fail-safe that should kick in, the one that was supposed to put the spacecraft into a safe low-power sleep, doesn't fire.

[00:01:00]It just keeps draining.

[00:01:02]Within hours, the spacecraft goes completely silent. No signal.

[00:01:08]No telemetry.

[00:01:10]No response to any command from the ground. One half of a 200 million euro mission spinning through the dark, unreachable. And the team on the ground has absolutely no idea if it's even still alive.

[00:01:24]Before I tell you how this ends, I want to make sure you understand what was actually at stake. Because on the surface, ESA loses contact with a satellite. Sounds like a technical problem for engineers to solve.

[00:01:37]It's not. It's the potential death of something genuinely irreplaceable.

[00:01:44]P R O B A 3 is not a normal space mission. It's built around an idea that sounds almost absurd when you first hear it.

[00:01:54]What if you could make a solar eclipse happen whenever you want.

[00:01:57]>> [music] >> Not wait for the moon's orbit to line up just right, which happens over any given city roughly once every 400 years.

[00:02:06]Not chase totality across continents, hoping the weather cooperates. Just manufacture one in space on demand.

[00:02:16]Here's why that matters.

[00:02:18]The sun's outer atmosphere, the corona, is one of the most scientifically important and most difficult things in the solar system to observe.

[00:02:26]It's faint.

[00:02:27]Impossibly faint compared to the solar disk itself.

[00:02:31]Trying to see it is like trying to read a candle flame while someone points a stadium floodlight directly at your face.

[00:02:38]The corona is right there at the edge, but the glare obliterates it.

[00:02:42]Total solar eclipses solve this perfectly.

[00:02:46]When the moon blocks the disk, the corona becomes visible, and scientists scramble to observe it.

[00:02:52]The problem is those eclipses last 2 minutes. 7 minutes at the absolute maximum. And in terms of actual science time collected by all of humanity across all recorded history, we're talking about a few hours total.

[00:03:09]A few hours of clear observation of this thing that drives all space weather, all solar storms, every geomagnetic event that has ever knocked out a satellite or threatened a power grid.

[00:03:21]And those events are not theoretical.

[00:03:24]In March 1989, a solar storm collapsed the entire power grid of Quebec in 90 seconds.

[00:03:31]6 million people lost electricity. The blackout lasted 9 hours.

[00:03:38]Transformers were damaged so severely that some took months to replace.

[00:03:44]That storm was not even close to the worst the sun can produce.

[00:03:48]The Carrington Event of 1859, a solar eruption so powerful that telegraph wires caught fire across North America and Europe. If something like that happened today, the estimated damage to global infrastructure runs into the trillions of dollars.

[00:04:06]The corona is where all of that begins.

[00:04:09]Understanding how those eruptions form in the region closest to the sun's surface is not abstract science.

[00:04:17]It's the difference between a 9-hour blackout and something considerably worse.

[00:04:23]And we have almost no direct observations of that region because until very recently, we simply could not see it.

[00:04:30]Proba-3's solution is elegant and a little insane.

[00:04:34]Instead of one spacecraft with an internal blocking disc, use two spacecraft.

[00:04:39]One, the occulter, flies ahead carrying a 1.4-m disc that plays the role of the moon. The other, the coronagraph, follows 150 m behind, sitting in the shadow, and observes the corona unobstructed.

[00:04:57]A telescope 150 m long with its two halves flying as completely separate satellites.

[00:05:05]Now, here's the part that should make you pause.

[00:05:08]Those two spacecraft have to hold their 150-m separation with an accuracy of 1 mm, not 1 cm, 1 mm, across the length of a football field, autonomously in orbit, >> [music] >> while dealing with gravitational fluctuations, solar pressure, and the constant perturbations of space flight.

[00:05:31]I've been thinking about that engineering requirement a lot while working on this video, and I still find it almost incomprehensible.

[00:05:39]The margin for error is smaller than the width of a credit card, and they have to maintain it without anyone on the ground holding their hand, because at orbital speeds radio commands arrive too late for real-time control.

[00:05:54]Proba-3 launched in December 2024.

[00:05:56][music] By June 2025, the two spacecraft had achieved what no mission had ever done before.

[00:06:04]They demonstrated true precision formation flying.

[00:06:08]Then they created the first artificial solar eclipse in space.

[00:06:13]The first images [music] came back, and they were extraordinary. Detailed views of coronal structures [music] that no instrument had ever clearly seen.

[00:06:22]By the time February 2026 arrived, the mission had produced more than 250 hours of solar corona observations. The entire history of total solar eclipses observed from Earth doesn't come close to that number.

[00:06:36]The first scientific papers had just been published, and they were already overturning assumptions.

[00:06:43]Solar wind in the inner corona was moving three to four times faster than the models predicted. Proba-3 was, by every measure, working.

[00:06:52]It was doing something unprecedented.

[00:06:55]It was changing what we know about the Sun.

[00:06:57]And then, it went dark.

[00:07:00]The team found out on the morning of February 15th.

[00:07:04]I want you to think about what that moment feels like.

[00:07:07]You have dedicated years to this mission.

[00:07:10]You [music] watched it launch. You held your breath through commissioning.

[00:07:15]You celebrated the first formation flight, the first artificial eclipse, the first images.

[00:07:22]And now your spacecraft, one of only two, and you need both of them, has stopped talking to you.

[00:07:30]No telemetry. No health data.

[00:07:33]No way to know if the instruments are intact. If the structure is damaged. If the battery has any charge left at all.

[00:07:40]The other spacecraft, the Occulter, is healthy.

[00:07:44]That's the first thing. It becomes, in a strange way, a sentinel.

[00:07:49]Still orbiting. Still functional.

[00:07:52]Still connected to the ground, but watching its lost twin drift away in silence.

[00:07:58]At the moment of the anomaly, the two spacecraft [music] had been roughly 4 km apart.

[00:08:04]Not in their close formation configuration.

[00:08:07]4 km doesn't sound like much, but trying to locate a dark, unpowered, tumbling satellite from another satellite 4 km away is genuinely difficult. You can't just look out a window. So, ESA starts mobilizing every resource it can.

[00:08:24]The TIRA radar system at the Fraunhofer Institute in Germany starts tracking the object.

[00:08:30]Commercial telescope networks, Northspace, Sabila Technologies, provide optical observation data.

[00:08:38]ESA's own space debris office begins processing [music] everything.

[00:08:43]The Occulter starts performing maneuvers to try to get visual data on its partner.

[00:08:48]Day one.

[00:08:49]Day three.

[00:08:50]>> [music] >> Day seven.

[00:08:52]Every morning, the team comes in knowing exactly what they don't know.

[00:08:57]Is the spacecraft still in a recoverable orbit?

[00:09:00]Is it tumbling in a pattern they can calculate?

[00:09:03]Is there any residual charge in the battery?

[00:09:06]Or is it completely dead?

[00:09:09]You can't know any of this until you know the rotation.

[00:09:13]And you can't know the rotation until you have enough tracking data.

[00:09:17]And gathering tracking data takes time.

[00:09:21]Day 10.

[00:09:23]Day 15.

[00:09:24]The worst part, and I think this is genuinely hard to appreciate unless you've been through something like it, is that you can't speed this up.

[00:09:34]You can work harder.

[00:09:35]You can throw more resources at it, but you can't change the laws of physics.

[00:09:41]The spacecraft tumbles at whatever rate it tumbles.

[00:09:45]The orbit decays at whatever rate it decays.

[00:09:48]You watch and you calculate and you wait.

[00:09:53]Think about the specific uncertainty they were living with.

[00:09:57]They knew the spacecraft had gone into a spin.

[00:10:00]They knew the battery had likely died completely. What they did not know, what they could not know until the radar data accumulated over weeks, was whether the spin was fast or slow.

[00:10:13]Whether it was a clean rotation around one axis or a chaotic wobble across multiple axes.

[00:10:19]Whether the panels ever swung toward the sun at all, even momentarily, or whether they were locked in a position that would keep them in permanent shadow forever.

[00:10:29]A fast, stable spin is recoverable.

[00:10:33]A chaotic tumble in the wrong orientation is not.

[00:10:37]For the first 2 weeks, they genuinely did not know which category they were dealing with.

[00:10:42]And while they were working that problem, there was a second question underneath it that nobody could answer.

[00:10:48]What had 30 days in the cold done to the spacecraft?

[00:10:52]Temperatures in Earth's shadow at Proba-3's altitude can drop to minus 100 minus 120°C.

[00:11:02]Spacecraft are designed for thermal stress, but designed for cycling, not for sustained deep cold over weeks.

[00:11:09]Certain materials contract.

[00:11:11]Certain seals stiffen.

[00:11:14]Certain electronics develop faults that you only discover when you try to turn them back on.

[00:11:19]The team could track the spacecraft.

[00:11:22]They could model the rotation.

[00:11:24]What they could not do, not until contact was reestablished, was know whether any of it was still worth saving.

[00:11:31]By mid-March, they had the data they needed.

[00:11:34]The spacecraft was still in a survivable orbit.

[00:11:37]It was tumbling in a specific pattern.

[00:11:39]And there were brief moments, fractions of a second, maybe less, when the tumble brought the solar panels momentarily toward the sun, just long enough potentially for a tiny amount of charge to accumulate.

[00:11:53]Just long enough potentially for the receiver to flicker on if the battery had retained any charge at all.

[00:12:01]If the receiver was still functional after more than a month in temperatures well below -100°C.

[00:12:08]If the timing calculation was right.

[00:12:11]If.

[00:12:12]March 18th.

[00:12:14]Villafranca ground station, Spain.

[00:12:17]The team has done everything a team can do.

[00:12:20]They've tracked the spacecraft for weeks. They've modeled the tumble.

[00:12:24]They've calculated the window.

[00:12:26]The precise moment when the solar panels should face the sun.

[00:12:30]When there might be just enough power for the signal to get through.

[00:12:34]There's nothing left to calculate.

[00:12:36]Nothing left to optimize.

[00:12:38]You've done the work.

[00:12:40]Now you transmit.

[00:12:42]And you find out.

[00:12:44]They send the signal.

[00:12:45]And they wait.

[00:12:47]I don't know what the atmosphere in that room was like.

[00:12:51]I don't know if people were quiet or tense or making nervous jokes.

[00:12:56]But I know what happened next.

[00:12:58]Because the ESA Director General Josef Aschbacher described it at a press conference the following day.

[00:13:04]He called it a miracle.

[00:13:06]Not a successful recovery.

[00:13:09]Not a positive outcome following sustained technical effort.

[00:13:13]He said, "Some miracle happened because we reconnected with the spacecraft."

[00:13:19]>> [music] >> The coronagraph had been tumbling through space for 33 days.

[00:13:24]In that time, it had experienced temperature swings that should have stressed every component on board.

[00:13:30]It had drifted tens of kilometers from its partner. And then, in the exact window the engineers had calculated, a tiny amount of sunlight hit the solar panels. Just enough.

[00:13:41]The receiver powered on.

[00:13:43]The signal got through.

[00:13:45]The spacecraft sent back a packet of telemetry data.

[00:13:48]Temperature readings, voltage levels, system status. The most mundane possible information.

[00:13:55]>> [music] >> And it was the best thing that team had heard in over a month. The spacecraft was alive.

[00:14:01]Now, here's what I find most interesting about this moment. And I've thought about it a lot.

[00:14:07]Everyone called it a miracle.

[00:14:10]The director general said it.

[00:14:12]The mission manager said it.

[00:14:14]The press covered it that way.

[00:14:16]And in one sense, yes.

[00:14:19]The timing, the residual charge, the fact that the receiver survived the cold, there was genuine fortune involved. But a miracle implies randomness.

[00:14:30]It implies that luck alone saved the spacecraft.

[00:14:34]And I don't think that's quite right.

[00:14:36]That team spent 33 days doing the work that made them ready for the miracle.

[00:14:41]They tracked a dark tumbling satellite across hundreds of thousands of kilometers of space.

[00:14:47]They built a rotation model precise enough to calculate a sub-second window of opportunity.

[00:14:53]They were at the console, signal prepared, at exactly the right moment.

[00:14:59]The universe gave them one chance.

[00:15:02]They were ready for it.

[00:15:04]Mission manager Damien Galano was asked about the moment the call came in from Villafranca.

[00:15:10]He said the excitement was palpable.

[00:15:13]And then he said immediately, "The hard work is not over yet." He was right. Getting a signal back is not the same as having a functioning spacecraft.

[00:15:23]The first telemetry told them the spacecraft was in safe mode.

[00:15:27]Minimal systems active. Everything else frozen.

[00:15:31]The solar panels were now facing the sun and slowly charging.

[00:15:35]But after 33 days in extreme cold, the onboard systems needed time to warm up before anything could be commanded.

[00:15:43]Spacecraft electronics that have been thermally stressed don't just spring back to life when you flip a switch.

[00:15:50]They need to be brought back carefully, incrementally, with constant monitoring.

[00:15:55]So, the team waited again.

[00:15:57]Carefully, methodically.

[00:16:00]In late March, they reestablished the inner satellite link between the coronagraph and the occulter.

[00:16:06]For the first time since February 14th, the two halves of the mission could talk to each other again.

[00:16:12]The coronagraph was still spinning slowly around one axis, a rotation it had settled into during its dormant period.

[00:16:19]But it was stable, sun-pointed, and charging.

[00:16:23]On April 1st, I appreciate that the universe has a sense of humor.

[00:16:28]>> [music] >> The team commanded the transition from safe mode into manual mode.

[00:16:32]They fired the thrusters.

[00:16:34]They stopped the spin.

[00:16:36]For the first time since February, the spacecraft regained full three-axis stabilization.

[00:16:43]Then came the moment that the science team had been most anxious [music] about.

[00:16:48]April 21st. They power up the coronagraph instrument for the first time since the anomaly.

[00:16:55]They point it at a patch of background stars. Nothing dramatic, just a test image to verify the optics and detector survived the cold. The image comes back clean.

[00:17:06]Andrei Zhukov, the principal investigator for the instrument from the Royal Observatory of Belgium, said, "After performing very careful checks of the ASPICS coronagraph instrument. I'm very happy to confirm that it's healthy and fully functional.

[00:17:23]No damage. No degradation.

[00:17:26]The instrument that spent [music] 33 days in the dark, spinning and freezing, came back in perfect condition.

[00:17:33]There's one more detail I want to mention because I think it's remarkable.

[00:17:38]When both spacecraft received software updates after the recovery, patches that fixed the original bug, the engineers also used the opportunity to improve the formation flying algorithm.

[00:17:51]The very failure that almost ended the mission led directly to software that made the mission more capable than it was before.

[00:18:00]June 9th, 2026.

[00:18:03]ESA officially declares the coronagraph spacecraft returned to full service.

[00:18:09]The first formation flight since February has been completed. New images of the solar corona are coming in. The mission is back.

[00:18:19]Let me ask you something directly.

[00:18:21]When you hear a story like this, a spacecraft going dark, a team spending a month in uncertainty, a single second of opportunity, what's the part that stays with you?

[00:18:34]For me, it's not the moment of contact.

[00:18:37]It's the 33 days before it. Because during those 33 days, the team didn't know how this was going to end.

[00:18:45]They didn't have the ending we're watching now.

[00:18:48]They had uncertainty >> [music] >> and incomplete data and the real possibility that they were going to lose the mission entirely.

[00:18:56]And they kept working anyway.

[00:18:58]Every day.

[00:19:00]Not because they knew it would work, but because it was the only thing to do.

[00:19:05]I find something genuinely valuable in that.

[00:19:08]There's a version of this story, probably more common than we admit, where the team reaches a point where the odds look bad enough that someone in a meeting room decides the resources aren't worth the outcome.

[00:19:21]Where the calculation shifts from how do we save this to how do we document what we lost?

[00:19:28]That version of this story doesn't end with contact on March 18th.

[00:19:33]The version we got ends the way it does because people decided the spacecraft was worth fighting for.

[00:19:40]Proba-3 will complete its nominal two-year mission in December 2026, just a few months from now.

[00:19:47]The people who built it and the people who saved it are going to watch that date arrive, and I suspect it will feel different for them than any milestone date usually does.

[00:19:57]Because they know what it cost to get here.

[00:20:00]They know about the 33 days.

[00:20:02]They know about the moment in Villafranca when they sent the signal and had no idea what would come back.

[00:20:08]Right now, as of this recording, the mission is fully operational.

[00:20:13]The two spacecraft are flying in formation again, 150 m apart, 1 mm accuracy, manufacturing eclipses over the Indian Ocean.

[00:20:24]The coronagraph is imaging the sun.

[00:20:27]The science is flowing, and the early results are already surprising researchers.

[00:20:32]Solar wind in the inner corona moving three to four times faster than anyone expected.

[00:20:38]Structures behaving in ways that are going to force the models to be rewritten.

[00:20:43]None of that was guaranteed to happen.

[00:20:46]All of it came very close to not happening at all.

[00:20:49]When you look at those images, detailed photographs of the sun's inner corona, structures no human instrument had ever clearly seen before the anomaly, and no instrument would have seen after it if the team had stopped working.

[00:21:04]You're looking at something that exists because people decided it was worth fighting for.

[00:21:09]I think that's worth knowing about.

[00:21:12]If this story got to you in some way, leave it in the comments.

[00:21:16]I genuinely like to know which part.

[00:21:19]And if you know someone who'd find this interesting, share it with them.

[00:21:23]I'll see you in the next one.