Caught From Orbit: SpaceX Snatches a Starship Out of the Sky for the First Time

Jump to comments

At 4:20 p.m. Eastern on Saturday, Nov. 14, 2026, a stainless-steel spaceship that had spent 22 hours circling Earth fell out of the South Texas sky belly-first, lit three engines, swung upright and stopped beside its launch tower. Two steel arms snapped shut around it. The last engine went quiet.

That was Ship 42, the upper stage of SpaceX’s Starship Flight 15, and no rocket stage like it had ever come home this way. Boosters have been caught before. An upper stage, the half that goes all the way to orbit at about 17,500 mph and has to survive the heat of reentry, never had.

What Flight 15 did for the first time

First upper stage caught by a tower. Every Starship before this one ended its flight in the ocean, on purpose, so engineers could push the heat shield and flaps hard without risking anyone on the ground. Ship 42 came down into the arms of a launch tower instead.

First Starship to return to its launch site. It landed at Pad 2 at Starbase, the same pad it left about 22 and a half hours earlier.

First Starship reentry over land. Earlier ships came down over empty ocean by design. This one flew back across northern Mexico and South Texas, steering around the Rio Grande Valley’s cities on its way to the pad.

Longest Starship flight yet. About 14 orbits and more than 22 hours, more than double the roughly 10-hour Flight 14 in September, which was cut short after an engine problem.

First catch for Pad 2. The newer tower’s shorter chopsticks, switched from hydraulics to electric actuators so they can move faster, had never caught anything before.

And it did a real job on the way. Before heading home, the ship delivered 26 Starlink V3 satellites, SpaceX’s largest Starlink model, to orbit.

SpaceX’s reaction, posted on X about a minute later, was short: “Mechazilla catches Starship! Congratulations to the entire SpaceX team on this incredible achievement!” For a company that has spent more than a decade trying to make rockets fly like airliners, it is the first real proof that the top half of the rocket can come back too.

The flight, hour by hour

Friday, 4:47 p.m. CST: liftoff. Starship lifted off from Pad 2 at Starbase, Texas, on Nov. 13 under more than 18 million pounds of thrust from the 33 Raptor 3 engines on Super Heavy Booster 22. Stage separation came about two and a half minutes later, with the ship lighting its own six engines while still attached, the "hot staging" move that has become routine.

About T+6 minutes: the booster waves off. The plan was a double catch, booster first. Booster 22 was cleared to come back, but its onboard guidance called an automatic abort on the way in and diverted to a splashdown in the Gulf, where it exploded just above the water. SpaceX commentator Dan Huot later explained that the system "detected that the booster was not in the place that it was supposed to be to proceed with the catch, did an automated abort, and then headed out to that offshore splashdown location." In other words, the safety logic did exactly its job and protected the tower the ship would need the next day.

T+26 minutes: orbit. A short burn of a single Raptor raised the ship into a roughly 275-kilometer (170-mile) orbit. "We’ve got a nominal orbit insertion, so Starship now in orbit around Earth for the second time ever," Huot said on the webcast.

T+35 minutes to about T+1 hour 17 minutes: Starlink delivery. The ship deployed all 26 Starlink V3 satellites, its real payload. SpaceX says each V3 adds about 1 terabit per second of capacity, so this one load is worth about 10 Falcon 9 launches of the older V2 Mini satellites. Three of the satellites carried cameras pointed back at the ship to photograph its heat shield.

Overnight: the long coast. Flight controllers had three chances to end the mission early with a splashdown in the Pacific, near Hawaii or off Chile. About seven hours in they passed on the Chile option. "Starship’s going to stay in space. We are one step closer to an eventual return to launch site," Huot said. By Saturday morning SpaceX reported the ship was still healthy and the pad and tower were go for catch.

Saturday, late morning Pacific time: go for return. Teams polled go to bring Starship back to Starbase, and Elon Musk posted that all systems were green for the deorbit burn. A single Raptor 3 fired for about 10 seconds, and SpaceX confirmed a "good deorbit burn" with the catch about an hour away.

The last half hour: reentry and catch. Starship hit the atmosphere over the Pacific thousands of kilometers from home, then used its four flaps to fly a series of gentle turns over thinly populated northern Mexico and South Texas, skirting the Rio Grande Valley’s cities, passing north of South Padre Island and swinging out over the Gulf for a final turn south to Starbase. People in Cameron County were told to expect sonic booms.

How you catch a falling spaceship

The final minute is the part that looks impossible on video, and it follows a script SpaceX published a week before launch.

The belly flop

Starship falls the last few kilometers lying almost flat, belly to the ground, using its flaps like a skydiver uses arms and legs. SpaceX first proved this "belly flop" with high-altitude hops of early prototypes and has refined it across numerous suborbital flight tests.

The flip

Just short of the tower the ship lights its three sea-level Raptor engines, swings itself vertical in a few seconds and folds its flaps out of the way. Then it shuts down one engine and steers between the tower’s arms on two, and drops to a single engine as it slows to a near hover.

The grab

The arms close fast on two catch fittings just below the ship’s forward flaps. The moment the ship senses contact with the arms, it shuts down its last engine. The chopsticks on Pad 2 are shorter than the first tower’s and are driven by electromechanical actuators instead of hydraulics, a change SpaceX made so they can move faster and track a vehicle more precisely during exactly this kind of catch.

Every step had an off-ramp. If anything looked wrong during reentry, controllers could deliberately tumble the ship so it broke up high over empty country. If a problem showed up late, with the ship or with the tower, it could divert into the Gulf instead. None of those were needed.

Why this is a bigger deal than the booster catch

The upper stage is the hard half of a reusable rocket. Recovering boosters is now routine for SpaceX’s Falcon 9, but every Falcon second stage is still thrown away. Starship is built so that nothing is.

To come home, a ship has to survive temperatures of up to about 2,600°F during reentry. It is covered in roughly 18,000 heat-shield tiles, and for years those tiles were the weakest link: some fell off on ascent, and plasma found its way into gaps. For Flight 15, SpaceX added extra retention on the tiles most at risk, closed off paths that let hot plasma get behind them, and used a curved tile design over a large part of the shield to cut heating in the gaps. The Starlink cameras and ground-based cameras gave controllers a look at the shield in orbit before anyone committed to bringing the ship over land.

A rocket that can land on its own pad does not need a recovery ship, a long tow back to port or a trip on a barge. In principle, it can be inspected, refueled and stacked again right where it lands.

That is the real prize. A caught ship skips weeks of ocean recovery and salt-water damage. SpaceX’s stated goal is "airplane-like cadence," with the same vehicle flying again quickly, and Musk has said the company is aiming for 10,000 Starship launches a year by 2030 to build out Starlink V3 and orbital data centers. None of that works if every upper stage ends up at the bottom of an ocean.

The booster that got away

Flight 15 was supposed to be the first full recovery of an orbital rocket, both stages caught on the same mission. Losing Booster 22 means that milestone is still waiting, and the new V3 Super Heavy has yet to be caught. SpaceX has caught three earlier-generation boosters, on Flights 5, 7 and 8, and flown two of them again.

The wave-off may still have been the right call. Musk noted on Saturday that SpaceX has "the next booster almost ready for flight, but losing the tower or launch mount would potentially cost us 2 to 3 months." Because the booster’s own abort logic kept it away from the pad, the tower was intact and ready for the ship the next afternoon.

How SpaceX got here

October 2024, Flight 5: the tower catches a Super Heavy booster for the first time.

January and March 2025, Flights 7 and 8: two more booster catches; one of the caught boosters is reflown on Flight 9 in May 2025.

May 22, 2026, Flight 12: the first flight of Starship V3, Raptor 3 engines and Pad 2.

July 24, 2026, Flight 13: the ship makes a soft, intact splashdown in the Indian Ocean, giving engineers their first close look at a heat shield that survived reentry. Musk wrote that SpaceX would attempt to catch the ship with the tower "unless we discover problems after mission data review."

Sept. 28, 2026, Flight 14: Starship reaches orbit for the first time and delivers its first 26 Starlink V3 satellites. After an engine shut down early on ascent, controllers played it safe, cut the orbital stay short and splashed down in the Pacific.

Nov. 13–14, 2026, Flight 15: orbit again, 26 more satellites, and the first ship catch.

What comes next

A caught ship is a ship engineers can walk up to. The first job is inspection: how the tiles, flaps, engines and catch fittings came through a real return from orbit.

After that, the missions that matter most depend on Starship flying often. NASA picked Starship as a lunar lander for its Artemis program, and that plan needs several Starships to meet in orbit and move propellant from one to another, a demonstration SpaceX has said it wants to fly. Reusing the upper stage is what makes that many launches affordable. Catching the booster and the ship on the same flight is the obvious next headline.

Questions people are asking

Is this really the first time an upper stage has been caught?

Yes, for a rocket’s upper stage caught by a launch tower. SpaceX has caught boosters before, and the Space Shuttle orbiter glided back to a runway and flew again, but the Shuttle threw away its huge external fuel tank every flight. Starship’s ship is the whole second stage, tanks and engines included, and no orbital upper stage like it had come back to its own launch pad and been caught.

Why catch it instead of landing on legs?

Legs add weight that the ship has to carry to orbit and back on every flight. Leaving the landing hardware on the tower saves that mass for payload, and it puts the vehicle right back where it gets refueled and restacked.

Was the whole rocket recovered?

No. The Super Heavy booster aborted its catch attempt and was lost in the Gulf. The ship is the part that came home.

Where can I watch it again?

SpaceX streams every Starship flight on its website and on X, and the replay of Flight 15 covers the launch, the orbital views and the catch.

The bottom line

Two years after the tower first caught a booster, it caught the spaceship on top. The booster still has to come home on the same flight, and this was one test, not a routine. But the hardest piece of a fully reusable rocket, a stage that goes to orbit, survives reentry and lands back on its pad, has now happened for real.

For more on the plan going in, see our Flight 15 preview.

Further reading