The Observatory — August 4, 2026

I · Now observing

A telescope falls, and the robot sent to catch it is failing too

A NASA gamma-ray telescope that cannot save itself is sinking toward the atmosphere, and the commercial spacecraft sent to grab it and push it back up has just suffered a breakdown of its own.[1] The Neil Gehrels Swift Observatory, in orbit since 2004 and built without any engine, has fallen to about 373 kilometers as a busy Sun puffs up the thin air it flies through, and NASA now puts the odds of an uncontrolled reentry before year’s end near 90 percent.[5]

NASA rendering of the Neil Gehrels Swift Observatory in orbit above Earth — The Neil Gehrels Swift Observatory, launched in 2004 without onboard propulsion. NASA rendering, public domain.

01 What happened

In September 2025 NASA gave a startup, Katalyst Space Technologies, roughly $30 million and less than a year to build a robot that could grab an observatory never designed to be grabbed.[3] That servicer, called LINK, launched on July 3, 2026 on the final flight of Northrop Grumman’s air-launched Pegasus XL, dropped from a carrier aircraft over the Pacific near Kwajalein Atoll.[4] Within weeks the rescuer needed rescuing. NASA’s July 28 update reported that LINK had spun up, fallen into intermittent contact, and lost the use of two of its three reaction wheels, the flywheels a spacecraft spins to point itself without firing thrusters.[1] The echo is exact, since Swift itself lost a reaction wheel years ago, and now its would-be savior has lost two.[4]

373 km

Swift’s altitude now, from ~600

~90%

odds of reentry this year

$30M

servicer, vs ~$500M telescope

2 of 3

reaction wheels now dead

02 The machine, in numbers

Swift launched into a 585-by-604-kilometer orbit and has no way to fight drag.[4] To buy time, controllers fly it edge-on to the oncoming air to shrink the area the atmosphere pushes against, trying to hold it above roughly 300 kilometers, the floor below which a rendezvous loses its best chance.[3] LINK carries three Hall-effect thrusters, electric engines that ionize xenon gas and fling it out with an electric field, trading very high fuel efficiency for very little push.[2] By one spacecraft catalog’s figures it launched at about 425 kilograms with roughly 60 kilograms of xenon aboard.[6]

03 Why the catch is hard

Two spacecraft in nearly the same orbit do not simply drift together. Push straight ahead to catch up and the added energy raises your orbit, which slows your angular pace and drops you farther behind, the counterintuitive rule that governs every rendezvous.[4] LINK must instead close along a controlled path, measure its position relative to Swift to within centimeters, and match rates with a target that may be tumbling slowly.[5] Because Swift has no docking port and no grapple fixture, LINK is meant to spend two to three weeks mapping it with a laser range-finder before three independent arms attempt a rigid grip, all without striking the instruments or the solar panels.[5] Raising the joined stack afterward is its own slow problem, since the climb from 373 back toward 600 kilometers needs only about 130 meters per second of velocity change, trivial for the xenon aboard, but a Hall thruster delivers that as a continuous months-long spiral rather than a single burn.[4] The current fault bites hardest here, because detumbling and fine pointing are exactly what one working wheel and a degraded gas system cannot easily provide.[1]

04 What to watch

The next marker is the following NASA commissioning update, which should say whether Katalyst has arrested LINK’s spin and restored steady pointing on its remaining wheel.[1] The real gate is the start of the two-to-three-week laser survey at Swift, the go or no-go before any grab, running against a hard clock set by the telescope’s own decay.[3]

Sources

II · From the record

Magellan and the invention of braking on borrowed air

The Magellan spacecraft had mapped almost all of Venus by 1993, and it needed a lower orbit it could not afford to reach with fuel, so its team at the Jet Propulsion Laboratory used the planet’s own atmosphere to brake instead, the first spacecraft to do so at another planet.[1] The trick they proved on a shoestring became, four years later, the standard way NASA orbiters reach their working orbits, and on one mission it removed an entire class of rocket and about a quarter-billion dollars.[3]

NASA/JPL artist depiction of the Magellan spacecraft with its high-gain antenna and solar panel — The Magellan spacecraft, deployed from Space Shuttle Atlantis in 1989. NASA/JPL artist concept, public domain.

01 The maneuver

Magellan left the cargo bay of Space Shuttle Atlantis in 1989, and by 1993 it had imaged 98 percent of Venus with radar from a stretched, elliptical orbit whose high point sat about 8,500 kilometers up and whose period ran three hours and a quarter.[1] The next goal was a gravity map, which needed a low, nearly circular orbit so the craft flew at a steady height, and circularizing with thrusters would have cost far more propellant than remained.[2] Starting May 25, 1993, the flight team lowered the orbit’s low point until it grazed the top of the atmosphere near 140 kilometers, so that every pass shaved a little speed off the far side of the orbit.[1] Over more than 75 days and hundreds of passes the high point fell, the period collapsed to 94 minutes, and on August 10 JPL announced a nearly circular orbit, calling it a first for a planetary spacecraft.[2] An earlier Japanese probe, Hiten, had already dipped through Earth’s own atmosphere in 1991, so the distinction Magellan earned was the first aerobraking at another planet.[6]

02 The hard part

Aerobraking is corridor-flying with almost no margin. Skim too high and the air does nothing; dip too low and heating and pressure spike on a structure never built to fly through an atmosphere.[1] The binding difficulty is that the air itself is not fixed, since density at the top of an atmosphere swings with local time, dust, and the Sun’s activity, so the drag on the next pass is a forecast rather than a known quantity.[3] With the density models and ground-in-the-loop navigation of the era, each low point had to be nudged by small thruster burns using the drag measured on the pass before, hundreds of times over.[2] Mars Global Surveyor later showed how thin that margin was, when a solar panel that had never locked after launch bent well past its stop as the Martian air came in about twice as thick as expected, forcing engineers to pause and then resume about a month later at roughly a third of the original pressure.[4]

03 Why it mattered, measurably

The payoff shows up as mass, and mass converts to rockets and dollars. On Mars Global Surveyor, aerobraking removed the need to carry roughly 1,500 kilograms of braking propellant.[3] That saving let NASA fly the mission on a Delta II instead of a much larger Titan III, a difference the agency put at about $250 million on that mission alone.[3] The gain repeated, as the Mars Reconnaissance Orbiter aerobraked through about 445 passes in 2006 and, by JPL’s account, halved the fuel it needed to reach a science orbit, dropping its period from roughly 35 hours to under two.[5] Before Magellan, a low circular science orbit had to be bought with propellant carried from Earth, which sized the whole vehicle upward; after, it could be bought with the destination’s own air, at a cost measured in weeks of operations rather than kilograms of fuel.[3] The $250 million figure assumes the same orbit was a hard requirement, so it is best read as the price of not compromising the science rather than a universal saving.[3]

04 Echoes today

Aerobraking went from stunt to routine. Mars Odyssey, the Mars Reconnaissance Orbiter, and later Mars orbiters all braked into their working orbits this way, and the Reconnaissance Orbiter’s 2006 campaign remains the clearest published measure of what the maneuver buys.[5] The logic Magellan proved, that a mission should not carry propellant it can borrow from the planet it is visiting, is now assumed in mission design rather than argued for, and its more aggressive cousin, braking to capture in a single pass, is an active study for future orbiters where carried fuel would be even more punishing.[3]

Sources