SpaceX’s fully integrated Starship and Super Heavy launch system lifted off from South Texas on Thursday and flew for roughly four minutes before breaking apart, completing the first flight of the largest rocket ever launched while falling well short of the near-orbital trajectory the company had planned. The vehicle rose from Starbase at 8:33 a.m. Central time, survived the period of maximum aerodynamic pressure and climbed to roughly 39 kilometers, but multiple booster engines shut down and the spacecraft never separated from the Super Heavy first stage.

The test was the first time the 33-engine Super Heavy booster and Starship upper stage flew together. The FAA launch license, issued April 14, authorized SpaceX to conduct Starship/Super Heavy operations from Boca Chica under federal safety, environmental and reporting requirements. Six days later, the rocket left the pad in a column of fire and debris, establishing that the integrated vehicle could at least survive liftoff from its new orbital launch mount.

A developmental test, not an operational mission

SpaceX had deliberately framed the flight as an experiment whose success would be measured in data rather than payload delivery. The nominal plan called for Super Heavy to accelerate Starship through the lower atmosphere, shut down and separate a little under three minutes after launch. The upper stage would then continue toward a near-orbital trajectory before reentering over the Pacific near Hawaii. Neither stage was intended to be recovered.

Space.com reported that several Raptor engines were visibly out during ascent and that the combined vehicle began tumbling when stage separation failed to occur. The stack then broke apart in what SpaceX calls a rapid unscheduled disassembly. A second Space.com analysis noted that the rocket nevertheless cleared the tower, passed through maximum dynamic pressure and reached an altitude of about 24 miles before the flight terminated.

Those milestones matter because Starship is not a conventional expendable launcher. SpaceX is trying to build both stages for rapid reuse, an approach that requires the company to solve propulsion, guidance, thermal protection, stage separation and recovery problems at a scale not previously attempted. Each test therefore validates some systems while exposing weaknesses in others.

Engine losses and failed separation define the next engineering problem

The 120-meter vehicle is designed around 33 methane-fueled Raptor engines on Super Heavy and six on Starship. During Thursday’s climb, the number of active booster engines dropped. Ars Technica reported that roughly 15 percent of the first-stage engines were no longer operating by the latter part of the burn. The vehicle continued climbing but eventually lost the planned flight profile before separation.

The inability to separate the stages is particularly consequential. Starship’s architecture depends on handing off from a powerful booster to an upper stage that must then accelerate itself almost to orbital velocity. The first integrated test never reached that transition. Engineers will now have telemetry from the real aerodynamic, vibration and propulsion environment to compare with ground tests and simulations.

collectSPACE reported that 30 of 33 Raptors appeared to ignite at liftoff and that additional engines were lost as the rocket climbed. That performance was sufficient to lift the nearly 400-foot stack, but the engine-out pattern, control authority and separation sequence will be central to the postflight review.

The launch pad is part of the experiment

The test also placed unusual stress on the ground system. Starship lifted off without the kind of large flame trench or water-deluge system used at many heavy-lift launch facilities. Video and photographs showed chunks of concrete and a broad plume of dust and debris thrown from beneath the orbital launch mount. The condition of the pad after liftoff will be part of SpaceX’s evaluation of how quickly the site can support another attempt.

The Planetary Society described the flight as a developmental success despite the explosive ending because the vehicle achieved several first-time objectives and generated data from an environment that cannot be fully reproduced on the ground. That interpretation reflects SpaceX’s iterative engineering model: fly hardware, identify failure modes, modify the design and fly again.

But the consequences of a launch failure extend beyond engineering. Because Starship is a licensed commercial launch system, SpaceX will have to work with federal regulators before returning the vehicle to flight. The FAA is responsible for ensuring that commercial launches protect the public and property, while environmental requirements at Boca Chica are tied to the agency’s authorization.

NASA has a direct stake in Starship’s progress

The program also matters well beyond SpaceX. NASA selected a lunar version of Starship as the Human Landing System for the Artemis program, making the vehicle part of the agency’s plan to return astronauts to the Moon. Starship must eventually demonstrate far more than launch: orbital operations, propellant transfer, long-duration flight and a crew-capable lunar landing architecture are all required for the NASA mission.

Wired’s launch-day account emphasized both the extraordinary scale of the vehicle and the fact that SpaceX entered the test with limited expectations of a complete mission. A first flight that ended in destruction is therefore neither a routine success nor a definitive failure. It is a data point in a development campaign with unusually high technical stakes.

The immediate result is straightforward. Starship proved it could leave the ground as a complete system and remain intact through the most aerodynamically stressful portion of ascent. It did not prove reliable booster propulsion, clean stage separation or near-orbital performance. Those missing demonstrations now define the next test. For SpaceX, the value of Thursday’s flight will be measured by how quickly engineers can convert the telemetry, launch-site damage and visible failure sequence into design changes that improve the probability of success on the next attempt.