The machine flew. It deployed its cargo. It survived re-entry.
From the Texas Gulf Coast on a Friday evening, SpaceX launched the most powerful rocket humanity has ever built — a 407-foot vessel carrying the ambitions of a species that has long looked moonward. The Version 3 Super Heavy-Starship lifted off with 18 million pounds of thrust, twice that of NASA's own lunar rocket, as part of a test campaign that must succeed before astronauts can return to the moon in 2028. The flight surfaced both the promise and the fragility of the endeavor: engines failed, a booster fell short, and yet the upper stage endured the fire of re-entry and delivered its cargo to space. Progress, as it so often does, arrived incomplete — but it arrived.
- Multiple Raptor engines shut down prematurely during ascent, sending the Super Heavy booster splashing down far short of its intended target in the Gulf of Mexico.
- The upper stage lost one vacuum engine but its flight computer compensated in real time, extending the burn of remaining engines to salvage an acceptable trajectory — a sign the system can adapt under pressure.
- Twenty-two Starlink satellite simulators were successfully deployed from an upgraded dispenser, and onboard cameras transmitted the first self-images of the vehicle's heat shield — a capability critical for future mission safety assessments.
- SpaceX now faces the defining technical challenge ahead: mastering autonomous orbital refueling, without which no Starship can carry astronauts to the moon, with a first refueling test targeted before year's end.
- Both SpaceX and Blue Origin must land unpiloted missions on the moon before NASA will risk a crewed landing in 2028, and neither company has yet placed a lunar lander in space.
On a Friday evening along the Texas Gulf Coast, SpaceX launched its Version 3 Super Heavy-Starship — a 407-foot rocket generating 18 million pounds of thrust, twice the power of NASA's own moon rocket. NASA Administrator Jared Isaacman watched from the ground as 33 methane-fueled Raptor engines ignited beneath the booster, sending the vehicle skyward after brief delays from a pad glitch and weather. The launch was designed to validate a generation of upgrades and build the confidence needed for NASA's Artemis lunar missions, which call for a Starship-derived lander to carry astronauts to the moon's surface beginning in 2028.
The flight exposed both advancement and vulnerability. About two and a half minutes after liftoff, the upper stage separated and the booster attempted to reverse course toward a Gulf splashdown — but engine failures during the return burn sent it well short of its target. The upper stage fared better: though one vacuum Raptor shut down early, the flight computer extended the burn of the remaining five engines to reach an acceptable sub-orbital path. In space, the vehicle deployed 22 Starlink satellite simulators and tested new cameras designed to inspect the heat shield from orbit — a tool that will matter greatly on future missions.
Re-entry proved the most encouraging chapter of the flight. The Starship endured the descent with far less thermal damage than earlier test vehicles had shown, executed a structural stress maneuver on its rear fins, and performed the banking approach future operational Starships will use for landing. It restarted engines, flipped to vertical, and splashed down before breaking apart in a planned fireball — the designed conclusion to a test flight.
The cause of the engine shutdowns remains under investigation, but SpaceX and NASA both framed the outcome as meaningful progress. The larger challenge now is autonomous orbital refueling — the ability to transfer propellant between vehicles in Earth orbit before a Starship lander can depart for deep space. Version 3 carries the hardware those operations will require, and SpaceX has targeted the first refueling test before year's end. With Blue Origin also developing a competing lunar lander under NASA contract, both companies face demanding schedules and an unforgiving deadline. The machine flew, the cargo was delivered, and the heat shield held. The moon remains years away, but the distance, at least, is measurable now.
SpaceX sent its most powerful rocket skyward on Friday evening from the Texas Gulf Coast, a 407-foot stack of steel and fire that would either prove the company's engineering ambitions or expose them in spectacular fashion. The Super Heavy-Starship V3 lifted off around 6:30 p.m. EDT from a newly reinforced launch pad at Starbase, generating 18 million pounds of thrust—twice what NASA's own moon rocket can muster. NASA Administrator Jared Isaacman watched from the ground as 33 methane-fueled Raptor engines at the base of the booster ignited in a brilliant torrent of blue-white flame, pushing the machine upward after delays from a launch pad glitch and weather.
The flight was meant to test a generation of upgrades: more powerful engines, refined control systems, and the infrastructure needed to make the rocket reliable enough for the missions that matter most. One of those missions, from NASA's perspective, is critical. The space agency has contracted SpaceX to build a lunar lander version of the Starship upper stage, one that will carry astronauts to the moon's surface starting in 2028 as part of the Artemis program. Before that can happen, SpaceX needs to prove the rocket works—repeatedly, reliably, and in ways that build confidence in the engineers and the machines.
The test flight revealed both progress and problems. Two minutes and 24 seconds after liftoff, as the rocket climbed out of the dense lower atmosphere, the Starship's six upper-stage Raptor engines ignited and the Super Heavy booster fell away, flipping around to reverse course toward a planned splashdown in the Gulf of Mexico. But the booster's engines did not perform as expected. One of the 33 Raptors shut down early during the climb, and additional engines failed to fire properly during the booster's attempt to return home. The machine dropped into the Gulf well short of its target, a setback that engineers will need to investigate and resolve.
The upper stage fared better. One of its three vacuum-optimized Raptor engines shut down prematurely as well, but the flight computer compensated by keeping the remaining five engines burning longer than planned, steering the Starship onto an acceptable sub-orbital trajectory. Once in space, the vehicle deployed 22 Starlink internet satellite simulators from an upgraded dispenser—a test of the deployment mechanisms that will eventually seed the next generation of the Starlink constellation. Two of those simulators carried cameras that transmitted images of the Starship itself, a capability that will be used on future flights to assess the condition of the vehicle's heat shield tiles, a critical component for surviving the violence of re-entry.
What triggered the engine shutdowns remains unclear. SpaceX engineers will spend the coming weeks analyzing telemetry and data to understand the failures. But the upper stage's performance in space and during re-entry suggested the core design is sound. The Starship endured the fiery heat of descent in apparently good condition, with far less thermal damage than previous test flights had shown. During the descent, the vehicle executed a maneuver designed to test the structural limits of its rear fins, then performed the dramatic banking maneuver that future Starships will use during normal landing operations. Just before reaching the Indian Ocean, the Starship restarted two engines, flipped to vertical, and descended toward a splashdown. It then tipped over as expected, broke apart, and exploded in a spectacular fireball—the planned end to a test flight.
Elon Musk, SpaceX's founder, posted on X that the company had scored "a goal for humanity." Isaacman, watching the test unfold, called it "a hell of a V3 Starship launch. One step closer to the Moon...one step closer to Mars." The rhetoric reflects the stakes. SpaceX is working to transition from sub-orbital test flights to orbital missions while perfecting the systems needed for lunar landings. A major hurdle lies ahead: the ability to autonomously refuel a Starship lander in Earth orbit before it can depart for deep space. The V3 now includes the attachment points and fuel-transfer systems those operations will require. SpaceX says the first in a series of orbital refueling tests is planned before the end of the year.
NASA and SpaceX are racing against a timeline. The agency plans to launch its next Artemis mission in 2027, sending four astronauts in an Orion capsule to rendezvous in Earth orbit with SpaceX's lander and an alternative being developed by Jeff Bezos' Blue Origin. Both companies must successfully land unpiloted missions on the moon before NASA will attempt to land astronauts in 2028. Neither has yet put a moon lander in space. Both face daunting test schedules. With Version 3 now available, SpaceX has the hardware to accelerate that work. The engine failures on Friday are setbacks, but not fatal ones. The machine flew. It deployed its cargo. It survived re-entry. The path forward is clearer now, even if the destination remains years away.
Notable Quotes
One step closer to the Moon...one step closer to Mars.— NASA Administrator Jared Isaacman, on the V3 launch
Scored a goal for humanity.— Elon Musk, SpaceX founder, on X platform
The Hearth Conversation Another angle on the story
Why does this particular test flight matter so much? Rockets have been flying for decades.
Because this one is built to be fully reusable and powerful enough to do things we haven't done before—land people on the moon, eventually Mars. NASA is betting its lunar program on it. If SpaceX can't make this work reliably, the whole timeline slips.
The engines shut down early. That sounds like a failure.
It is a failure, but a survivable one. The upper stage compensated and still reached orbit. The booster missed its target. Those are problems to solve, but the vehicle didn't explode on the pad or during re-entry. For a test flight, that's progress.
What's the hardest part still ahead?
Refueling in orbit. You can't send a lander to the moon on a single tank. You need to launch a tanker, dock it in Earth orbit, transfer fuel, then send the lander on its way. SpaceX has never done that at scale. It's the engineering problem that could delay everything.
How much time do they actually have?
NASA wants astronauts on the moon in 2028. That's less than two years away. Both SpaceX and Blue Origin need to land unpiloted missions first. SpaceX is ahead in some ways—they've flown more test flights—but neither company has actually put a moon lander in space yet.
What happens if they miss that date?
The program delays. NASA's already pushed Artemis back multiple times. Congress gets impatient. The political will to fund lunar exploration isn't infinite. But more immediately, it means the astronauts don't go to the moon when the agency promised they would.
So Friday's test—despite the engine problems—was actually good news?
Yes. The vehicle flew, deployed its cargo, survived re-entry in better shape than before. The engine shutdowns are data. Engineers will figure out what went wrong and fix it. That's what test flights are for.