On November 4, 2026, at 12:39 UTC, a NASA satellite entered the atmosphere over northern Sumatra. According to the U.S. Space Force, later confirmed by NASA, the entry point was at 3 degrees north latitude and 99 degrees east longitude — neatly in the middle of the 11:42–13:42 UTC forecast window the military had issued hours earlier. “NASA expected most of the spacecraft to burn up,” to quote the official update, “but some components may survive reentry.” The odds that a sharp fragment would make it to the ground and hurt someone were estimated at roughly one chance in 2100. Sumatra — the planet’s fifth-most-populous island, with an eight-figure population — had a ticket in that lottery and, apparently, did not receive any space junk.
This satellite was supposed to die at the end of 2006. Its primary mission was designed to last two years. It worked eleven times longer, rewrote an entire field of astronomy, and left only because it never had an engine — while the Sun, in recent years, had become seriously unruly: the upper atmosphere, heated by solar activity, swelled up, drag increased, and the orbit melted away. By early 2025, the spacecraft had lost roughly half its original altitude. It would not have made it into the 2030s engineers had hoped for, no matter how much it wanted to.
Its name was Swift. And the name was not decoration. It was the assignment.
To understand why Swift was built at all, let us go back to 1963. The Moscow Treaty banning nuclear tests in the atmosphere, outer space, and underwater had been signed. The signatories faced an awkward question: how do you check that nobody is cheating? The answer was the Vela satellite series — from the Spanish velar, “to keep watch.” Beginning in October 1963, they were sent up in pairs into an orbit 250,000 kilometers in diameter, so that no patch of Earth could escape their gaze; aboard were X-ray, gamma-ray, and neutron detectors designed to catch the flash of a nuclear explosion in space.
The detectors found no nuclear explosions. But they did find flashes that were neither nuclear nor solar, nor tied to Earth. At first, engineers at Los Alamos simply filed such triggers away in folders marked “not of our origin, investigate later.” In 1972, Ian Strong was asked to look through those folders; together with Ray Klebesadel and Roy Olson, he used the difference in signal arrival times between the paired satellites to calculate the directions of the flashes to within about ten degrees — enough to rule out both the Sun and Earth. In 1973, a short paper appeared in ApJ Letters on sixteen “cosmic gamma-ray bursts” recorded between July 1969 and July 1972.
So humanity, trying to monitor compliance with a nuclear treaty, accidentally discovered the most powerful explosions since the Big Bang. A single such burst releases, in a matter of seconds, energy comparable to the radiation from all the stars in its galaxy.
And here science hit a wall for a generation. A burst lasts seconds — while large telescopes take hours to point. By the time any ground-based or orbital instrument turned in the right direction, all that remained in the sky was a fading afterglow with no identifying marks. “Bursts came and went so quickly that no one could point big telescopes at them to learn more,” as Brad Cenko, the mission’s principal investigator, put it. Where did the burst come from, how far away was it, what powered it — the mysteries materialized and dissolved faster than anyone could measure them.
The idea for a solution was born, as mission operations director John Nousek recalled, in 1997 “in the basement of a hotel in Estes Park, Colorado, in the middle of a conference”: a handful of astrophysicists began brainstorming between sessions about what kind of spacecraft could solve the gamma-ray burst problem — “a very big mystery at the time.” The answer conceived in that basement sounded simple to the point of audacity: a telescope was needed that watched the whole sky itself, grabbed a burst by the tail itself, and turned toward it in a fraction of a minute, so that the eyes looking in other wavelengths could sketch everything before it faded.
That was how Swift was conceived. On November 20, 2004, a Delta 2 rocket carried it away from Pad 17A at Cape Canaveral. Structurally, it was a hunting pack of three instruments. The BAT telescope hung over half the visible sky like a wide-angle monitor and waited for a gamma-ray burst. The moment a trigger went off, the entire spacecraft — mirrors, panels, wiring — pivoted toward the burst, and within about a minute two co-aligned telescopes, the X-ray XRT and the optical-ultraviolet UVOT, delivered a precise position and the first images. The coordinates were immediately sent to observatories around the world: investigators are given a suspect’s description while it is still fresh.
Pointing, however, was not its only talent. Swift’s planning system could accept targets-of-opportunity from the ground within minutes and monitor specific sources from minutes to months — up to 75 independent targets per day. In its later years, it carried out five to ten unscheduled observations daily, hundreds to thousands a year. An instrument built for one class of mysteries turned out to be a key to the entire variable sky: active galactic nuclei, erupting stars, comets, asteroids, stars being torn apart by black holes. It even managed to produce an X-ray atlas of the entire sky.
But back to the main business. In May 2005, six months after launch, Swift caught GRB 050509B, a burst lasting 0.03 seconds. Three hundredths of a second. The spacecraft turned quickly enough to record eleven X-ray photons from the afterglow — and that was enough for the first precise localization of a short gamma-ray burst in history. Twenty-year-old theoretical models were confirmed: short bursts are born in the merger of two neutron stars — objects with the mass of the Sun compressed to the size of a city. Eleven photons — the price of a revolution in our understanding of how binary stars die.
Then came the cascade. Its very first month of operation brought a record magnetar flare — and Swift effectively reclassified these “stellar corpses” with monstrous magnetic fields from scratch. Then came confirmation of the two-class scheme (long bursts — the collapse of massive stars; short bursts — neutron-star mergers), followed by its complication: bursts that fit into neither box. The brightest explosion in observational history. Record-breaking highest-energy photons. Candidates for the most distant explosions in the Universe. In 2017 — two historic events in succession: a kilonova from a neutron-star merger, caught by gravitational-wave detectors too, and the first-ever tracing of a neutrino back to its source. More than 2,000 gamma-ray bursts, more than 1,400 supernovae, around 7,000 scientific papers using its data.
And, no less importantly, Swift made all of its data openly available from day one — even while the team itself still did not understand exactly what it had seen. “That was very novel,” said Northwestern astrophysicist Wen-fai Fong, crediting Swift with creating the open-data culture later inherited by other observatories.
On October 9, 2022, detectors aboard several spacecraft at once — Fermi, Swift, Wind — were drowned in a flood of radiation arriving from the constellation Sagitta. Gamma-ray burst GRB 221009A would later acquire the unofficial name BOAT — Brightest Of All Time. The signal had traveled 1.9 billion years: it was born in the collapse of a massive star, at whose heart a newborn black hole was visibly contracting, accelerating two near-light-speed jets of matter. The downpour passed through the Solar System and triggered detectors pointed at entirely different regions of the sky.
Swift was at the center of it too. Its X-ray telescope imaged the afterglow an hour after discovery — the image came out surreal: bright concentric rings, an X-ray “echo” scattered by layers of invisible interstellar dust on the way to Earth. In essence, one burst scanned the Galaxy’s dust structure. Fermi recorded the explosion for more than ten hours; NASA’s conclusion was brutally honest: there may not be another burst like it for decades.
There was also a quiet engineering drama in that observation: the afterglow was so bright that Swift’s X-ray detector physically choked on the flux. The image was ultimately reconstructed from “dead time” — the fraction of photons the detector had not had time to count. Space can be so bright that you measure it by what gets missed.
Rapid response is a race of its own, and it has its own marshal on the ground. In the Sierra de San Pedro Mártir, at an altitude of 2800 meters in Mexico’s Baja California, stands the robotic half-meter COATLI telescope — a joint project of UNAM’s Institute of Astronomy and Arizona State University. Its name comes from Nahuatl: coatl, “snake” — and also “twin.” Its ASTELCO NTM-500 mount accelerates to 30 degrees per second: an alert signal from an orbital detector arrives, and within seconds the snake is already looking at the right point, catching the fading optical afterglow. That is what it was built for (full operations since 2018, project lead Alan Watson): “rapid response to transients” is written into its specifications in plain text.
The protocol across this whole operation is the same, and it was the one Swift followed too: every gamma-ray burst observation is always a relay race. A satellite sees the flash for the whole world, scout spacecraft turn within seconds, ground-based robots within tens of seconds, and only then do giants like the VLT and James Webb arrive, with their hours of retargeting time. In this relay, Swift was both the runner on the first leg and the one handing over the baton — twice in a single race.
By 2025, it had become clear: Swift’s orbit was melting away irreversibly. Solar maximum had heated and expanded the upper atmosphere, drag was rising, altitude was falling faster than any forecast. Swift had no engine of its own — twenty years ago, that had been a rational decision for a hunter mission; now it had become a death sentence. For the entire 2026 observing season, it was even put to sleep, its instruments mothballed in preparation for a rescue attempt: scientists could already feel the loss then, before any actual death.
And then something happened that until recently had looked like science fiction. Arizona startup Katalyst Space built LINK — a robotic tug meant to fly up to Swift, capture it, and lift it into a higher orbit. NASA paid 30 million dollars under a contract run through an astonishingly fast procedure: only a few months passed from the decision to launch. On July 3, 2026, LINK set off for its target. Never before had NASA tried to save a scientific spacecraft in low orbit with a commercial outsider’s hand — it was the first attempt of its kind in history, and the very fact that it was prepared already changes the rules for the entire fleet of aging satellites.
It did not work out. Roughly three weeks into the flight, two of LINK’s three reaction wheels failed — the small wheels that control a spacecraft’s orientation in space. The tug broke into uncontrolled rotation, communications began to drop out, and there could be no question of a precise rendezvous and capture. On August 19, 2026, NASA officially canceled the attempt. Curiously, rendezvous and close-proximity operations were nevertheless continued and completed: not to save Swift, but to gain experience for future satellite-servicing missions. NASA Administrator Jared Isaacman put it without grandstanding: “This isn’t the outcome we were working toward, but it doesn’t change why the mission was worth attempting.”
The Swift team had one thing left to do: squeeze every last bit of science from the dying spacecraft.
On September 24, 2026, Swift recorded a gamma-ray burst named GRB 260924A — an event Judith Racusin of Goddard would describe this way: “It is a little bit like Swift’s swan song.” The explosion, roughly 12 billion light-years away from us (redshift 3.5), emitted radiation with an isotropic energy of around 10^55 erg — among the most powerful bursts ever measured. Its X-ray afterglow entered the ranks of the brightest in the comparison sample; optically, when adjusted to the same redshift, it surpassed even the afterglow of BOAT itself during the first day. The most energetic photon — 43.3 gigaelectronvolts — arrived 82 seconds after the burst began. A month later, on October 23, a 30-page multiwavelength paper appeared on arXiv, where the afterglow had already been broken down into forward and reverse shock waves, while the delayed material was explained by energy injection from catching-up ejecta and the star’s magnetized matter.
Swift conducted its final observation on October 20 from an altitude of around 275 kilometers — any lower was simply impossible: atmospheric resistance no longer allowed it to hold its pointing steadily. In its final weeks, it measured Earth’s atmosphere from altitudes few spacecraft can endure, and even took a look at a comet. As it left, it remained true to itself: its archive — roughly 12 terabytes of observations, one and a half million UVOT images, 22 years of continuous brightness measurements of everything that blinks — remains open to everyone and will feed discoveries for decades to come.
Nature abhors a vacuum. Several systems are taking up the baton at once, and nearly every one is built on Swift’s principles: the whole sky on a monitor, automatic pointing in seconds, coordinates distributed instantly.
On the ISS, since April 2022, the OHMAN pairing has been at work: the Japanese MAXI monitor catches a burst and automatically sends a trigger to the NICER X-ray telescope, which turns toward it within minutes without a single human in the loop — a stripped-down version of the BAT-XRT pairing, assembled from instruments belonging to two space agencies. OHMAN’s first major job was BOAT itself in October 2022: NICER reached the source three hours later, as soon as it rose above the station’s horizon. Then there is the French-Chinese SVOM, already operating as Swift’s direct heir in philosophy; China’s Einstein Probe with its X-ray “lobster eyes,” forever watching wide fields; the CTAO observatory under construction for the highest energies; NASA, according to Astrophysics Division Director Shawn Domagal-Goldman, is now sorting through ways to respond rapidly to cosmic events using the fleet already in hand — from fast commercial platforms to new small missions. And ground robots like COATLI are feeling cramped: a network of dozens of such snakes scattered across every time zone is a matter of the next few years.
There is a bitter irony in how this mission ended. A satellite built to catch the fastest catastrophes in the Universe died from the slowest process of all — viscous friction against a heated atmosphere, microscopic braking accumulated over twenty-two years. It was not killed by equipment failure or operator error: the Sun merely added air where there was not supposed to be so much of it.
Sixteen bursts from Los Alamos folders in 1973. Eleven photons in 2005, overturning ideas about the death of binary stars. One record-breaking burst in 2022, scanning straight through the Galaxy’s dust rings. And a farewell burst two weeks before the end. If the designer of this mission had been told in 1997, in that Estes Park basement, that his brainstorming session would turn into twenty-two years of work and a machine that turned faster than explosions fade — he probably would have asked for a piece of the action.
The snake in the Sierra de San Pedro Mártir still slumbers in its tower, waiting for the next trigger. The sky will not stop exploding.