On Saturday, 14 November 2026, SpaceX caught a Starship upper stage with the mechanical arms of its launch tower, the first time any spacecraft returning from orbit has been caught that way. The catch came about 22 and a half hours after liftoff, at the end of a flight that circled Earth 14 times and delivered 26 Starlink satellites. It was a genuine first — and it was only half of what SpaceX had planned.
That split result is the whole story of Flight 15. One stage came home and was gently set into the tower's arms. The other ran low on fuel, aborted itself, and ended in a fireball over the water. Both halves matter, because SpaceX's entire plan for cheap, rapid reuse depends on catching both stages, every time. This article walks through what happened minute by minute, why the booster turned away, how the Ship survived reentry and a tower catch, and what the flight does and does not prove about the future of Starship.
What actually happened on Flight 15
Flight 15 lifted off from Pad 2 at SpaceX's Starbase site in South Texas at 5:47 p.m. EST on Friday, 13 November 2026. The vehicle was Super Heavy Booster 22 under Ship 42 — the fourth launch of the redesigned Starship V3 configuration. Thirty-three Raptor engines pushed the stack off the pad, and about two minutes and 21 seconds into the flight the stages separated in a hot-staging manoeuvre, with the Ship's six engines igniting while still attached to the booster.
The plan, published by SpaceX before launch, was the most ambitious the company had ever attempted: return both stages to the same launch tower and catch them both. The booster was due back about six and a half minutes after liftoff. The Ship would stay in orbit for more than 22 hours, deploy 26 Starlink V3 satellites, circle Earth about 14 times at an altitude of roughly 275 kilometres, and then fly itself back to the same tower on Saturday afternoon.
The first half of that plan failed within minutes. The booster relit 32 of its planned 33 engines for the boostback burn, but the burn ended early. SpaceX's guidance system detected that the booster was not in the right position for a tower catch, aborted the attempt automatically, and steered the vehicle to a contingency landing zone offshore in the Gulf of Mexico. The booster performed a landing burn over the water, shut its engines down while hovering, and then its flight termination system triggered, destroying it a few feet above the waves.
The second half worked. Ship 42 reached orbit, deployed all 26 Starlink satellites, and spent nearly a full day in space. On Saturday afternoon it fired a single engine for a roughly 10-second deorbit burn, fell back through the atmosphere over Mexico, swung out over the Gulf, flipped around, ignited three Raptor engines for its landing burn, and slid between the tower's arms — known as chopsticks — at about 4:20 p.m. EST. The arms closed around it. SpaceX had caught a Starship returning from orbit for the first time.
The catch: 4:20 p.m. ET on Saturday
The moment itself lasted seconds, but it was the payoff of a flight that had lasted almost a full day. Ship 42 had left the same tower on Friday evening. It came back to it on Saturday afternoon, after travelling around the planet 14 times, and was caught on the first attempt.
Scientific American, which followed the flight live, described the final seconds like this: the Ship plunged through the atmosphere, performed a "belly flop" manoeuvre — turning end-over-end and swivelling its engine plumes downward — and slid precisely between the awaiting arms, which "swung closed around the midair vehicle, like chopsticks snatching the world's largest egg roll". The Associated Press called it the first-ever capture of a Starship flying back from space.
The catch time, about 4:20 p.m. EST (21:20 UTC) on Saturday, 14 November, matches SpaceX's own published timeline almost exactly. The timeline placed the Ship catch at 22 hours and 32 minutes after launch; the actual catch came roughly 22 hours and 33 minutes after liftoff. After months of delays and a launch licence that was only finalised on the day of the flight, the Ship's return ran to plan.
What the catch looked like from the ground was almost anticlimactic, and that was the point. A 50-metre spacecraft fell out of the sky, lit three engines, slowed to a hover beside a 140-metre tower, and was held in place by two steel arms. The webcast showed Ship 42 hanging between the arms at Starbase, intact, about 22 and a half hours after it had left the same pad.
The catch also topped Hacker News on 14 November, where the thread "SpaceX catches the Starship to end Flight 15" drew more than a thousand points within hours — a sign of how closely the spaceflight community had been watching a flight billed, before launch, as SpaceX's boldest test yet.
A 22-and-a-half-hour flight, 14 orbits
Most Starship test flights before Flight 15 were over in about an hour. Flight 15 lasted more than 22 hours, and that duration was deliberate. SpaceX wanted to prove that a Ship could survive a full day in orbit — the kind of long stay a real operational mission needs — and then still fly itself home.
The Ship's orbit was roughly 275 kilometres above Earth, low enough that atmospheric drag is real but high enough for a stable day-long flight. Over about 22 and a half hours it completed approximately 14 orbits, which works out to roughly one orbit every 97 minutes — the normal period for low Earth orbit at that altitude.
Scientific American noted that Flight 15 was Starship's longest stay in orbit yet. That matters because the longer a spacecraft stays up, the more its systems are tested: thermal control as it passes in and out of sunlight, power management, communications, and the health of the engines and heat shield that must all still work at the end of the flight. A Ship that can only survive an hour in orbit cannot do the jobs SpaceX has planned for it, from deploying satellites to carrying crews to the Moon.
The return began on Saturday afternoon. At about 3:23 p.m. EST, Ship 42 fired a single engine for a roughly 10-second deorbit burn, according to Space.com's live coverage, slowing itself just enough to drop out of orbit. Scientific American put the burn at just before 3:30 p.m. EST. Either way, the Ship then spent about an hour falling back through the atmosphere, steering with its flaps, before the final landing burn and the catch at about 4:20 p.m.
The booster that did not come home
The booster was supposed to be the easy half of the double catch. SpaceX has caught Super Heavy boosters before — three V1 boosters were caught on earlier test flights, and two of them flew again. But Flight 15 was the first catch attempt for a V3 Super Heavy, the redesigned first stage, and it did not go to plan.
The trouble started during the boostback burn, the engine firing that turns the booster around and sends it back toward the launch site. The booster relit 32 of its planned 33 engines, but the burn ended early. SpaceX's Kate Tice, the company's director of quality management and training, said on the Saturday webcast that early indications pointed to a low fuel mass alarm cutting the burn short, and that the root cause was still under investigation.
Because the burn ended early, the booster was not where it needed to be. Its guidance, navigation and control system did what it was designed to do: it recognised that the vehicle could not safely continue to the tower, called off the catch automatically, and steered the booster to a pre-planned contingency landing zone offshore in the Gulf of Mexico. SpaceX spokesperson Dan Huot initially described the problem as a guidance, navigation and control issue, then clarified on Saturday that the guidance system had worked correctly — it was the booster that was out of position, and the software correctly aborted the catch.
Elon Musk gave a different piece of the explanation on X a few hours before the Ship's return, writing that "slight underperformance of engines on this flight is the reason" for the booster's diversion. He also said separately that the booster was slightly low on fuel. Space.com also reported that one booster engine dropped out after stage separation. None of these accounts contradicts the others: an engine underperforming can leave a booster short of propellant, which can end a burn early, which can leave the vehicle out of position. But SpaceX has not yet published a final root cause, and the investigation was still open as of the weekend.
What happened next was visible on the webcast. The booster came down over the water, performed its landing burn, hovered briefly, and shut its engines down. Then the flight termination system triggered and the booster exploded a few feet above the waves. That was not an accident in the usual sense: once the booster had aborted to the offshore zone, the termination system was the planned end of the flight. The vehicle was never going to be recovered after the abort. The loss was decided the moment the boostback burn ended early.
Why the booster aborted: what SpaceX has said so far
Three explanations have been offered for the booster abort, and it helps to keep them separate, because they describe different links in the same chain.
First, the low fuel mass alarm. According to Kate Tice, the booster's boostback burn ended early because the vehicle detected that its propellant mass was low. A burn that ends early leaves the booster slower and further from the launch site than planned, and there is no way to make up the difference later.
Second, engine underperformance. Musk said on X that "slight underperformance of engines on this flight is the reason" for the diversion, and Space.com reported that one booster engine dropped out after stage separation. An engine producing less thrust than expected burns propellant less efficiently and changes the booster's trajectory — which is consistent with a low fuel alarm further into the burn.
Third, the guidance system did its job. Huot's clarification matters because early reports framed the abort as a guidance failure. It was the opposite: the guidance system detected that the booster was not in the right position for a tower catch and correctly refused to attempt one. A tower catch of a mispositioned booster would have risked the launch tower itself — the most expensive single piece of infrastructure at Starbase — so the software chose the safe option, an offshore splashdown zone, over the spectacular one.
What SpaceX has not said is why the engines underperformed or why the fuel ran low. The root cause was still under investigation as of Saturday, and SpaceX has a history of taking weeks to publish detailed findings after a flight anomaly. Until then, any claim that a specific valve, sensor or software bug caused the abort is speculation, and this article will not repeat it as fact.
One thing the abort does establish is that the V3 booster is still unproven at recovery. The three successful booster catches were all V1 hardware. Flight 15 was the first V3 catch attempt, and it failed — not proof the design is flawed, since a single flight with a fuel alarm is not enough evidence for that, but the next attempt will carry more weight than this one did.
Flight 15 timeline
SpaceX publishes a flight timeline for every Starship test, listing each planned event by mission elapsed time. The table below combines that published timeline with what actually happened, based on SpaceX's mission page and the live coverage from Scientific American, Space.com and the Associated Press.
| Event | When | What happened |
|---|---|---|
| Liftoff | Fri 13 Nov, 5:47 p.m. EST | 33 Raptor engines lift Booster 22 and Ship 42 off Pad 2 at Starbase. |
| Maximum aerodynamic pressure | T+58 seconds | Vehicle passes through max-Q as planned. |
| Booster engine cutoff, hot-staging | T+2 min 19 s / T+2 min 21 s | Stages separate; Ship's six engines ignite. One booster engine reportedly dropped out after separation. |
| Booster boostback burn | Minutes after liftoff | 32 of 33 engines relight; burn ends early on a low fuel mass alarm. |
| Booster catch aborted | ~T+6 min 35 s (planned catch time) | Guidance aborts the tower catch; booster diverts offshore, performs a landing burn over the Gulf, and is destroyed when its flight termination system triggers. |
| Starlink V3 deployment | In orbit, Friday evening | Ship 42 deploys 26 Starlink V3 satellites. |
| Ship in orbit | ~14 orbits, ~275 km altitude | Longest Starship stay in orbit yet: more than 22 hours. |
| Deorbit burn | Sat 14 Nov, ~3:23 p.m. EST | Single engine fires for roughly 10 seconds to begin reentry. |
| Ship catch | Sat 14 Nov, ~4:20 p.m. EST | Three Raptors ignite for the landing flip; the tower's arms close around Ship 42. First-ever catch of a Starship upper stage. |
Two details deserve attention. The Ship catch landed within about a minute of the time SpaceX had published before launch — 22 hours and 32 minutes after liftoff. And the booster abort happened at almost exactly the moment the catch had been scheduled, which is why early social media posts were confused: viewers watching for a catch at six and a half minutes saw the booster turn away instead, and many assumed the whole flight had failed. It had not.
Ship 42 and Booster 22: the V3 hardware
Flight 15 was the fourth launch of Starship V3, a redesigned version of the vehicle that differs substantially from the first two versions. SpaceX has not published a full public specification of every V3 change, but the mission page and pre-flight coverage confirm the broad picture: a new booster and a new ship, flying together for a full orbital test with recovery of both stages as the goal.
The numbers that are public tell part of the story. The Super Heavy booster stands about 71 metres tall and is powered by 33 Raptor engines. The Ship is about 50 metres tall with six Raptor engines of its own. Stacked together, the vehicle is about 121 metres tall, the largest rocket ever built, and the largest ever flown.
One detail worth noting is that Flight 15 was the first catch attempt for a V3 Super Heavy. The three booster catches SpaceX achieved before were all V1 hardware, and two of those boosters went on to fly again. The V3 booster is a different vehicle in important ways — new engines, new structure, new flight software — so its recovery record starts from zero. After Flight 15, that record is one attempt, one abort.
The Ship, by contrast, had never been caught in any version. Every previous Ship had either been deliberately destroyed, splashed down in the ocean, or — on Flight 14 — brought down early in the Pacific after an engine issue. Ship 42 is therefore the first Starship upper stage of any version to come back to land intact, and the first spacecraft of any kind to be caught by a launch tower after an orbital flight.
The 26 Starlink V3 satellites
Flight 15 was not only a recovery test. It was also a working mission: Ship 42 carried 26 Starlink V3 satellites into orbit and deployed them there. That is the second batch of the V3 satellite design to reach orbit, after Flight 14 on 28 September 2026, which also deployed Starlink satellites and reached orbit before its early ending.
The V3 satellites matter because they are the reason Starship exists in the near term. SpaceX's Falcon 9, the workhorse rocket that built the Starlink constellation, can only carry so many satellites at a time. Starship, once fully operational, will loft Starlink satellites in far larger and more frequent batches, according to Scientific American — and SpaceX has disclosed plans to supplement Starlink with huge numbers of orbital data centres for artificial intelligence, which would need Starship's payload capacity even more.
That commercial logic is also why the booster abort, while disappointing, did not spoil the flight. The payload reached orbit. The satellites were deployed. The Ship came home. A test flight that delivers its payload and recovers its most valuable stage is a productive flight even when the other stage is lost — and SpaceX builds a new upper stage roughly every three weeks, according to Ars Technica reporting cited in post-flight coverage, so the hardware pipeline is not the constraint. The constraint is proving that both stages can be caught and reflown quickly.
How a tower catches a spacecraft
The launch tower at Starbase is about 140 metres tall and is fitted with two enormous mechanical arms that SpaceX calls chopsticks. The tower is nicknamed Mechazilla. The arms are the same hardware that stacks the Ship on top of the booster before launch — they lift the 50-metre upper stage and place it on the booster — and the same arms catch the stages when they return.
A catch works like this. The returning vehicle steers itself to a point beside the tower, then performs a landing burn that slows it to a hover. The arms, which are mounted on rails and can move vertically and horizontally, close around the vehicle and grip it by small hardpoints — reinforced catch pins — on its sides. There is no landing leg, no runway, no net. The vehicle is simply held in mid-air by the arms, and can then be lowered onto the launch mount.
The reason SpaceX catches instead of landing is mass. Landing legs strong enough to hold a 200-tonne booster or a 100-tonne ship would add tonnes of weight that could otherwise be propellant or payload. Catching removes the legs entirely and lets the tower do the work. The trade-off is precision: the vehicle must arrive at exactly the right place, at exactly the right speed, with its guidance system healthy. If anything is wrong, the tower does not attempt the catch — which is exactly what happened with Booster 22.
The Ship catch is harder than the booster catch in one important respect. The booster returns within minutes of launch, from a suborbital arc, with its trajectory and fuel state closely known. The Ship returns after a day in orbit, having survived reentry heating, flown a long hypersonic glide steered by its flaps, and then flipped from a horizontal belly-first fall to a vertical tail-down hover in the final seconds. Every one of those steps has to work, and the tower only gets one attempt. On 14 November, they all did.
Flight 15 vs the flights before it
Flight 15 sits in a short, fast-moving history. Starship has flown 15 times, and the recovery story has changed with almost every flight. The table below puts Flight 15 in that context, using only outcomes confirmed in SpaceX's own mission pages and the live coverage of each flight.
| Flight | Date | Booster | Ship |
|---|---|---|---|
| Flight 5 | October 2024 | Caught by the tower — the first booster catch ever | Splashed down in the Indian Ocean as planned |
| Flights 6–13 | 2024–2026 | Two more V1 boosters caught; two reflown | Ocean splashdowns; no Ship catch attempted |
| Flight 14 | 28 September 2026 | Splashed down in the Gulf as planned | Reached orbit, deployed Starlink satellites, then brought down early in the Pacific after an engine issue |
| Flight 15 | 13–14 November 2026 | Catch aborted — low fuel alarm; destroyed over the Gulf | Caught by the tower — first Ship catch ever, after 14 orbits |
Two patterns stand out. First, booster recovery is still not routine. SpaceX has caught three V1 boosters, but it has also lost boosters on most flights, and the V3 booster has now failed its first catch attempt. The company's own history shows that catching a booster is possible but not yet reliable — and reliability, not possibility, is what a reusable rocket needs.
Second, the Ship's progress has been steadier. Flight 14 reached orbit and deployed satellites before an engine issue cut it short. Flight 15 did everything Flight 14 did, stayed in orbit a full day, and then came home to the tower.
Why catching the Ship matters more than the booster
It is tempting to score Flight 15 as a draw — one stage caught, one lost — but that understates what the Ship catch means. The booster is the simpler vehicle. It is a big fuel tank with engines: it goes up, turns around, and comes back within minutes. SpaceX proved it could catch one in October 2024, and has done it three times since. The Ship is the spacecraft. It goes to orbit, survives reentry at more than 27,000 kilometres per hour, flies a controlled glide through the atmosphere, and then has to hit a target the size of a launch tower after a day in space. Nobody had ever caught one before, because nobody had ever tried — and nobody had tried because it is the harder half of the problem.
The Ship is also the half that carries the value. The booster is expensive, but the Ship is where the payload goes, where the crew will go, and where the heat shield, the flaps and the engines that must survive reentry all live. A reusable booster with an expendable Ship is a half-reusable rocket.
That promise is not abstract. SpaceX is developing a modified Starship as the Human Landing System for NASA's Artemis programme, which aims to return American astronauts to the Moon before the end of 2028. Under the current plan, a Starship lander will rendezvous and dock in low Earth orbit with a NASA Orion spacecraft carrying four astronauts during Artemis III, targeted for next summer. Before that can happen, Starship has to demonstrate in-space refuelling — a heavy lunar round trip needs more fuel than one Starship can carry into orbit — and it has to demonstrate that its vehicles can be recovered and reflown. Flight 15 proved the recovery of the Ship. Refuelling is still untested.
There is also the Starlink arithmetic. A fully reusable Starship that flies roughly once a week — Musk's stated projection for the first half of 2027 — would change the economics of the satellite constellation completely, letting SpaceX launch far larger batches of satellites far more often than Falcon 9 allows. That is the business case that pays for the test programme, and the Ship catch is the clearest evidence yet that the business case is reachable.
What the catch does not prove yet
A single catch does not prove reusability. It proves that a Ship can be caught once, under good conditions, after a flight that went almost exactly to plan. SpaceX still has to show several things before Flight 15's catch becomes a routine operation.
First, the Ship has to fly again. Catching a spacecraft is only useful if the spacecraft can be inspected, refuelled and launched again without a rebuild. SpaceX has not yet reflown a caught Ship — there has never been one to refly — and the condition of Ship 42 after its catch has not been publicly detailed. The heat shield, the engines and the structure all went through reentry heating and a landing burn; how much refurbishment they need will determine whether the catch actually saves money or time.
Second, the turnaround has to be fast. Musk's projection is roughly one Starship launch per week in the first half of 2027. That requires not just catching both stages but turning them around in days, not months. Even if Flight 15 had caught both stages, the real prize — rapid reuse — would still be untested, as Scientific American pointed out in its coverage.
Third, the booster has to be caught too, and reliably. The V3 booster has now failed its first catch attempt, and the root cause is still under investigation. Until a V3 booster is caught and reflown, the double catch — the headline goal of Flight 15 — remains undemonstrated.
Fourth, the flight was lucky in one respect that should not be overlooked: the Federal Aviation Administration only issued the revised licence authorising the flight on the day of the launch, 13 November. A programme that needs a licence on the day of every launch is not yet a routine operation.
What happens next
SpaceX has not announced a date for Flight 16, and it would be premature to predict one before the company publishes its findings on the booster abort. The next flight will almost certainly attempt the double catch again — it is the goal SpaceX set for Flight 15, and it remains unachieved. Whether the next attempt uses another V3 booster, and what changes SpaceX makes to the boostback burn or the propellant margins, will be the first sign of what the investigation found.
Beyond the next test flight, the calendar is crowded. Artemis III is targeted for next summer, with a Starship lander due to meet the Orion crew capsule in orbit. In-space refuelling tests, which the lunar mission depends on, have yet to be demonstrated publicly. And Musk's projection of roughly one launch a week in the first half of 2027 sets a pace the programme has never approached — 15 flights in more than three years, against a target of about 26 in six months.
The politics around Starbase are also moving. Texas governor-elect Gina Hinojosa campaigned against energy-intensive data centres and has pledged to pause new approvals, according to Scientific American — awkward timing for SpaceX, which has disclosed plans for a proposed $24-billion gas-fired power plant and supercomputing complex at Starbase and is seeking local tax incentives.
For now, the scoreboard is simple. Fifteen flights. One Ship caught. Three V1 boosters caught, none yet in the V3 configuration. One double catch attempted, none completed. Flight 15 moved the programme forward by exactly one historic first — and left the other one still waiting.
Frequently Asked Questions
Was the whole Starship rocket recovered on Flight 15?
No. Ship 42, the upper stage, was caught by the launch tower on Saturday, 14 November 2026. Super Heavy Booster 22 was not caught: its catch was aborted minutes after liftoff on Friday, and the booster was destroyed over the Gulf of Mexico. Only one of the two planned recoveries worked.
When exactly was the Ship caught?
At about 4:20 p.m. EST (21:20 UTC) on Saturday, 14 November 2026, roughly 22 and a half hours after liftoff from the same launch tower at Starbase, Texas. SpaceX's published timeline had placed the catch at 22 hours and 32 minutes after launch.
Why did the booster catch fail?
The boostback burn ended early, which SpaceX's Kate Tice attributed to a low fuel mass alarm. The guidance system then detected the booster was not in the right position and automatically aborted the catch. Musk also said on X that slight engine underperformance was the reason, and Space.com reported one booster engine dropped out after stage separation. The root cause was still under investigation as of the weekend.
Has SpaceX caught a Starship before?
It had caught Super Heavy boosters before — three V1 boosters, starting in October 2024, with two reflown. It had never caught a Ship upper stage before Flight 15, and no spacecraft returning from orbit had ever been caught by a launch tower before.
What did Flight 15 carry into orbit?
Ship 42 deployed 26 Starlink V3 satellites, the second batch of that satellite design to reach orbit after Flight 14. The satellites are the commercial payload that the Starship programme is being built to launch at scale.
What is the next milestone for Starship?
The double catch: recovering both the booster and the Ship on the same flight, which Flight 15 attempted and did not achieve. After that, the bigger tests are reflowing a caught Ship, turning both stages around quickly, and demonstrating in-space refuelling for the Artemis lunar missions.
Sources
- SpaceX — Starship Flight 15 mission page (primary source: flight plan, timeline, vehicles, objectives)
- Scientific American — "Starship Flight 15 puts SpaceX's moonshots to the test" by Lee Billings, updated 14 Nov 2026 (primary live coverage: catch time, booster abort, Musk quote, Flight 14, Artemis, Starbase politics)
- Hacker News — "SpaceX catches the Starship to end Flight 15" (discussion thread; topped Hacker News on 14 Nov 2026)
- Drohnen.de — "Starship Flight 15: Ship caught by the tower" (live timeline: 21:20 UTC catch, booster abort detail)
- Weatherboy — "SpaceX Tower Arms Catch Starship Upper Stage for First Time" (booster abort detail, FAA licence, Musk fuel comment)
- USA Today — "SpaceX catches Starship for first time in Texas, making space history" (paywall) (catch confirmation, booster splashdown, Musk comment)
Related reading on Asad Webs: NASA Reworks Artemis III: Crewed Orbital Test Flight in Summer 2027, Moon Landing Moves to 2028, SpaceX Starlink 15-16 Set for Record 72nd California Launch of 2026 From Vandenberg, Japan's MMX Spacecraft Launches to Bring Home a Piece of Mars's Moon Phobos. No fitting tool exists for this topic on Asad Webs Tools — the tools index is linked here for readers who want to browse it.
Want help planning content around a space or technology story? Contact Asad Webs and tell us what you are working on.




Be the first to share a useful question or thought.