NASA’s Neil Gehrels Swift Observatory has spent more than two decades watching for gamma-ray bursts, the most powerful explosions in the universe. It is now sinking towards the atmosphere, and the plan to save it is about to leave the ground.

The agency has hired Katalyst Space Technologies, a startup in Flagstaff, Arizona, to fly a small robotic spacecraft up to Swift, take hold of it, and push it back to a safer altitude. That spacecraft, called LINK, was fitted to its rocket on 9 June at NASA’s Wallops Flight Facility in Virginia. According to NASA, launch is expected later this month.

If it works, it would be the first time a commercial robot has captured a government satellite that was never built to be serviced in orbit.

Why Swift is coming down

Swift launched in 2004 and works as a kind of dispatcher for the high-energy sky. When something violent happens out in space, a star collapsing or a sudden burst of gamma rays, Swift catches it quickly and points other telescopes at the spot. Like any satellite in low Earth orbit, it has been losing altitude to atmospheric drag for years.

What moved the timeline forward was the Sun. The current peak in the solar cycle has heated and swollen the upper atmosphere, increasing drag and pulling Swift down faster than expected. By accounts of the mission, the observatory has lost roughly a third of its original altitude, falling from around 600 kilometres towards 400. Reporting on the rescue has put the chance of an uncontrolled reentry at about even by the middle of this year, rising towards the end of it.

To buy time, the Swift team paused most of its observations earlier in 2026 and held the spacecraft in an orientation that minimises drag. That slows the fall. It does not stop it.

The seven-month rescue craft

NASA awarded Katalyst US$30 million under a Small Business Innovation Research Phase III contract in September 2025, with the boost then targeted for spring 2026. The striking part is the speed. The company built LINK in about seven months, a pace the journal Science described as almost unheard of for a NASA mission.

NASA has been open that this is a deliberate bet. Nicky Fox, who heads the agency’s Science Mission Directorate, framed the project as a chance to show it can go from concept to implementation in less than a year. The astrophysics division’s acting director, Shawn Domagal-Goldman, called it a forward-leaning, risk-tolerant approach, and pointed out that a boost is cheaper than building a replacement for what Swift does.

Those are the reasons for trying. They are not a guarantee that it will work.

How the boost is meant to happen

LINK will fly on a Northrop Grumman Pegasus XL, an air-launched rocket released in flight from a modified L-1011 airliner called Stargazer. NASA has used the same type of vehicle before, to launch its ICON and NuSTAR satellites. For this mission, Stargazer will carry the rocket from Wallops to Kwajalein Atoll in the Marshall Islands and drop it there.

Once in orbit, LINK is meant to find Swift, match its motion, take hold of a spacecraft that has no docking port or capture fixture, raise its orbit, and let go. The capture is the hard part. Swift was built to study the sky, not to be grabbed, so there is nothing on it designed to help another spacecraft hold on.

What is actually being tested

It helps to be clear about what this mission is and is not. It is a demonstration of satellite servicing on a craft that was never prepared for it, against a deadline set by physics rather than by a roadmap. Nothing quite like it has been completed before, which is part of why NASA has described the attempt in the language of risk rather than routine.

If the boost succeeds, Swift keeps working, and NASA gains real evidence that a commercial robot can extend the life of an existing satellite, including one already in trouble. If it fails, the most likely result is the outcome the mission was meant to prevent, an uncontrolled reentry of the observatory.

The wider interest for the industry is the model. A government science satellite in difficulty, a small company under contract, and a rescue assembled in months rather than years is a different way of operating from the one that built Swift in the first place.

What to watch next

The immediate milestone is the launch, anticipated before the end of June, followed within weeks by the rendezvous and the capture attempt. Only then will it be clear whether the boost can be carried out as planned.

At the time of writing, LINK was on its rocket and waiting. The hard parts are still ahead of it.