A one-of-a-kind NASA space telescope is about to fall out of the sky.

For more than two decades, the Neil Gehrels Swift Observatory telescope has gazed at the fading afterglows of some of the universe’s most violent explosions.

After launching in 2004, Swift initially circled about 370 miles above the Earth’s surface. That far up, there were hardly any air molecules for Swift to run into, but slowly, over time, its orbit descended lower and lower.

It is now at an altitude of about 210 miles, where the air is denser, dragging on the telescope. Within a few months, the friction will pull Swift into the thicker part of the atmosphere and rip it apart.

That is, unless a robotic spacecraft can pull off a daring, never-before-tried rescue: catching Swift and then nudging it back to a higher orbit.

“I am cautiously optimistic,” said Brad Cenko, the principal investigator for the Swift mission.

The rescue attempt is scheduled to commence on Tuesday, with the launch of a refrigerator-sized spacecraft from Kwajalein Atoll, one of the Marshall Islands in the middle of the Pacific Ocean.

Launch is scheduled for no earlier than 6:23 a.m. Eastern time. (On Kwajalein, it will be 10:23 p.m.)

A small start-up company, Katalyst Space Technologies of Flagstaff, Ariz., developed and built the spacecraft, called Link, in about nine months, a breakneck pace compared to typical space missions, which take years to get off the ground.

But to save Swift, the luxury of careful, thorough engineering was not an option. Ghonhee Lee, the chief executive of Katalyst, said NASA only laid two basic requirements: boost the spacecraft and do not run into it or otherwise damage it.

“That gave us a ton of flexibility in how we designed the program and how we designed the spacecraft,” Mr. Lee said.

A replacement for Swift, if NASA decided to build one, would most likely take years and cost hundreds of millions of dollars. Giving a $30 million contract to Katalyst to boost Swift was a smart bet, NASA officials said.

“The risk of us losing Swift — if we hadn’t done this — the odds were 100 percent,” said Shawn Domagal-Goldman, director of NASA’s astrophysics division. “That’s why, from a financial and management standpoint, this made sense to do.”

He was more emphatic during a later news conference: “No one thought it was going to be possible. No one thought we would get as far as we’ve already gotten today.”

Successfully rescuing Swift could add years of precise gamma-ray burst observations.

Gamma rays are a form of light like radio waves, the visible spectrum of light and X-rays, but they possess the very highest energies. In the 1960s, satellites that were built to monitor nuclear weapon tests unexpectedly discovered flashes of gamma rays popping all around the cosmos.

When another telescope spots a gamma-ray burst, Swift lives up to its name by quickly swiveling toward the newly discovered burst and making detailed measurements of the high-energy light as it fades away.

Those brief torrents of gamma rays tell stories about moments of extreme cosmic violence, like explosions of stars and the collisions of small, ultradense stars known as neutron stars.

NASA originally counted on Swift lasting just two years, and it did not include any contingency plans for boosting the orbit a couple of decades later.

The pace of descent is difficult to predict, because it speeds up and slows down depending on the ebbs and flows of the sun’s 11-year sunspot cycle. When the sun is more active at the peak of the sunspot cycle, it shoots out more intense solar flares that heat the Earth’s atmosphere and puff it outward, increasing the drag on orbiting satellites like Swift.

The most recent peak, at the end of 2024, was stronger than predicted. That greatly sped up Swift’s fall, and NASA suddenly faced the possibility that the telescope might re-enter the Earth’s atmosphere sometime this year.

There was not enough time to design from scratch a mission to boost Swift. Instead, NASA scrambled to find three companies, including Katalyst, that had already developed technologies that could be deployed for this rescue attempt.

Mr. Lee said he started Katalyst six years ago with the vision of deploying a fleet of space robots to build, repair and refuel satellites and other infrastructure in space.

NASA awarded the rescue mission contract to the company last September.

“I think we were the only ones that were maybe crazy enough to say, ‘Yeah, I think there’s a possibility that we could do this for that budget and that timeline,’” Mr. Lee said.

Katalyst was also fortunate in coming across a discounted, leftover rocket. The price tag for many rockets, including SpaceX’s workhorse, Falcon 9, would use up the $30 million that NASA allocated for the entire rescue mission.

Katalyst called up Northrop Grumman, the aerospace and defense technology company, and Mr. Lee said he joked that it would be nice if the company had an old winged rocket known as Pegasus lying around.

Unlike most rockets, the Pegasus, developed in the late 1980s by Orbital Sciences Corporation, which Northrop Grumman later absorbed, does not lift off from a launchpad. Rather, it is tucked under the fuselage of an airplane that flies to an altitude of 40,000 feet before dropping the rocket. Then, the rocket motors of Pegasus ignite, accelerating it to space.

Forty-five Pegasuses have been launched, but none since 2021, and the rocket seemed to have been discontinued. But Mr. Lee said Northrop called back to say that there was indeed one more Pegasus that had been designated for a different governmental mission.

“It was just in long-term storage,” Mr. Lee said. “They had basically shrink-wrapped it, put it away, and it was just sitting in the corner.”

The ability of an airplane to serve as a movable launch site proved a boon in this instance. Swift’s orbit is not easily reached from Florida and other places farther to the north.

The rescue looks like it will be in time.

Earlier this year, Swift stopped observing gamma ray-bursts in order to perform maneuvers needed to slow its fall. If Swift had continued to operate as usual, the telescope might be about 25 miles lower today, Dr. Cenko, the principal investigator, said.

After the launch of the Link rescue spacecraft, mission controllers at Katalyst will spend a week or two checking that its systems are working properly. Then Link will take about a month and a half to approach Swift and grab it.

After that, the spacecraft will gradually propel Swift upward for two months before letting it go 100 miles higher, enough of a boost for it to remain in orbit for another decade.

Swift will also prove key to solving new mysteries about gamma-ray bursts.

There are two types of gamma-ray bursts: short ones lasting a couple of seconds or less, and long ones persisting for minutes or hours. For decades, astrophysicists thought they understood the fundamental difference between the two. Short ones were generated by the merging of neutron stars, while the long ones were generated by the collapse of large stars.

But now they have found bursts that broke the pattern, when mergers of neutron stars produced long bursts and not the usual short ones.

“What we’ve seen in the last couple of years is that that simple picture is wrong,” Dr. Cenko said. “It’s more complicated than that.”

Astrophysicists hope data from Swift will reveal a new understanding of something they had thought impossible.