SpaceX conducted the twelfth integrated flight test of its Starship system from its Starbase facility in South Texas in the early morning of May 23, Beijing time. The mission featured the Ship 39 spacecraft and the Booster 19 Super Heavy booster. This flight marked the first orbital-level appearance of the Starship system's third major configuration, known as V3, and also saw the inaugural use of the newly constructed Pad 2 at Starbase. The entire suborbital mission lasted over an hour. Approximately seven minutes after liftoff, the Booster 19 Super Heavy booster failed to achieve a controlled splashdown as planned due to anomalies, including the premature termination of its boostback burn. The Ship 39 spacecraft successfully deployed 22 satellites and began atmospheric re-entry about 47 minutes and 47 seconds after launch, splashing down in the Indian Ocean near the coast of Western Australia roughly 65 minutes post-launch. Upon contact with the water, the spacecraft partially exploded. Based on disclosed information, despite the booster recovery anomaly, SpaceX completed core mission objectives including satellite deployment, suborbital flight, and spacecraft splashdown. This flight marked the resumption of Starship flights after a seven-month hiatus for major modifications since October 2025.
The initial launch attempt for this twelfth flight was scheduled for the morning of May 22, Beijing time. The SpaceX team loaded the 124-meter-tall rocket with over 5,000 tons of super-chilled methane and liquid oxygen on the new Pad 2. However, the countdown was halted with 40 seconds remaining. Following the scrub, an explanation was provided, citing a hydraulic pin securing the tower arm that failed to retract. After overnight repairs, SpaceX proceeded with the launch attempt on May 22. This was the third schedule change within about a week, with initial expectations pointing to a launch as early as May 16.
The flight path for this mission was adjusted compared to previous ones. SpaceX selected a more southerly maritime route, with the rocket traversing between the northeastern coast of the Yucatán Peninsula and western Cuba over the Gulf of Mexico, instead of the Florida Strait used in prior missions. This adjustment involved re-coordinating flight safety and splashdown zones, aligning with the new launch pad's location and the V3 configuration's performance characteristics. SpaceX confirmed it would not attempt recovery of either stage for this mission. The Booster 19 Super Heavy booster experienced an anomaly during its boostback burn about seven minutes after liftoff, failing its planned controlled splashdown. The Ship 39 spacecraft followed a suborbital trajectory for about 65 minutes, splashing down in the Indian Ocean after completing its deployment test. This conservative approach was directly related to the mission's nature, prioritizing data collection over recovery validation for the simultaneous first flights of a new configuration and a new launch pad.
SpaceX stated that a primary goal of the twelfth flight was the first in-flight demonstration of all new components in the V3 configuration. With all new elements flying for the first time, SpaceX opted not to attempt catching Ship 39 or Booster 19 to mitigate risk and prioritize data acquisition. Payload deployment was another key objective. Ship 39 carried 22 satellites, nearly double the number from previous flights. Twenty were standard-mass simulated Starlink satellites, testing an upgraded Starlink PEZ dispenser mechanism equipped with newly designed actuators and inverters aimed at increasing individual satellite deployment speed. Validating dispenser performance is crucial for the future pace of constellation deployment, a core task for Starship's commercial operations. The final two specially modified satellites performed a unique "self-inspection" task. They were designed to scan Starship's heat shield and transmit images to ground operators, testing a method to analyze the shield's readiness for future missions returning to the launch site. Engineers painted several heat shield tiles white beforehand to simulate missing tiles and serve as clear imaging targets. Additionally, one tile was intentionally omitted at launch to precisely measure aerodynamic load differences on adjacent tiles during re-entry if a tile is lost. This marked the first time SpaceX actively created a heat shield defect for a controlled in-flight test. Addressing the reusable heat shield is identified as one of the most significant remaining technical challenges for the Starship program.
Beyond heat shield testing, the flight originally planned to restart one Raptor engine in space to verify this "key capability." However, during the actual flight, only five of the six Raptor engines on the spacecraft functioned normally during the initial burn phase, leading SpaceX to skip this demonstration. In-space engine restart is a critical maneuver for future lunar, Mars, and deep-space missions.
The V3 configuration incorporates improvements across nearly every system. The Super Heavy Booster V3 features noticeable changes to its grid fins. The number was reduced from four to three, but each fin's area increased by approximately 50% with significantly improved structural strength. Each fin contains a new capture point and is mounted at an adjusted angle on the booster to support future lift-and-catch operations. Their mounting position was lowered overall to reduce exposure to the upper-stage engine exhaust during hot staging. Components like pivot points, actuators, and mounting structures previously exposed on the booster exterior have been moved inside the main propellant tanks for better protection. The hot-staging method was completely redesigned. An integrated hot-staging structure replaced the previous expendable protective interstage. During stage separation, the booster tank's forward dome is directly exposed to the Starship engines' hot exhaust, protected solely by internal tank pressure and a non-structural steel shell. Interstage actuators automatically retract after separation, further shielding against engine exhaust. This transforms hot staging from a "sacrificial" scheme into a reusable structural design. The most critical internal change in the booster involves the propellant delivery system. A feedline responsible for delivering cryogenic propellant from the main tanks to the 33 Raptor engines at the base was completely redesigned with a substantially increased diameter. The new design allows all 33 engines to start simultaneously and execute faster, more reliable flip maneuvers. The aft thermal protection system was also redesigned, eliminating the large individual shrouds for each engine in favor of added shielding on surfaces between engines and around the thrust vector control hardware for the center 13 engines. With the elimination of aft cavities and engine shrouds, the CO2 fire suppression system housed within was also removed, significantly simplifying the booster's aft structure and reducing maintenance points. The connection method to the launch pad changed as well. The Super Heavy Booster V3 replaced the previous single quick-disconnect interface with two physically separate connection points for fuel and oxidizer, providing additional redundancy for ground-to-vehicle fluid connections and allowing for smaller, simpler launch pad support structures.
The Starship V3 spacecraft's propulsion system underwent a "complete redesign," enabling a new Raptor engine start method, increased propellant tank volume, and improved attitude control systems for in-flight steering. Updates also reduced enclosed spaces in the vehicle's aft where propellant leaks could accumulate, enhancing safety. The routing of aft fluid and electrical systems was replanned, a prerequisite for eliminating individual engine shrouds. The actuation system for the aft flaps was changed from two actuators per flap to a single actuator with three motors, improving redundancy while the motors back each other up, reducing total system mass and cost, and directly aiding reliability for return-to-launch-site operations. To meet future deep-space mission needs, Starship V3 is explicitly endowed with "extended duration flight capability." Specific configurations include a more efficient attitude control system, high-pressure gas isolation valves, 100% vacuum-jacket coverage for the nosecone supply system, a high-pressure electrically-driven cryogenic recirculation system, and a system dedicated to managing cryogenic propellant and engine interaction during extended coasting in space. On the vehicle's leeward side, four new "docking funnels" and associated propellant line connections are direct hardware preparations for ship-to-ship propellant transfer, or space refueling. This is a prerequisite capability for any mission beyond Earth-Moon orbit. The Starlink PEZ dispenser mechanism was also upgraded with newly designed actuators and inverters, aiming to increase the deployment speed per satellite.
The Raptor 3 engine is the power core of the entire V3 architecture. Compared to the previous generation, thrust for the sea-level variant increased from 230 tons to 250 tons, and from 258 tons to 275 tons for the vacuum variant. This thrust increase was achieved not simply by raising chamber pressure but through deep structural simplification. Sensors and controllers were integrated into the engine itself, protected by the engine's own thermal protection system. This design decision allowed SpaceX to eliminate the external shrouds previously configured for each engine on the Starship and Super Heavy booster. All engine models feature a redesigned ignition system. Mass for the sea-level variant was reduced from 1,630 kg to 1,525 kg. Through simplification of the engine itself, vehicle-side support facilities, and support hardware, approximately 1 ton of mass is saved per engine at the vehicle level, a significant cumulative reduction for the 33-engine booster and 6-engine spacecraft.
Starship V3 debuts a completely new avionics architecture specifically designed for high flight rates, full reusability, and enhanced reliability. The combined booster and spacecraft systems contain about 60 custom avionics units, integrating batteries, inverters, and high-voltage power distribution into single components. The entire system can provide approximately 9 megawatts of peak power across the vehicles and features distributed fault isolation, where a failure in one unit does not cripple the entire system. The navigation system was upgraded to a multi-sensor solution with high redundancy across anticipated mission profiles and environmental conditions, designed to achieve precise autonomous flight. New precision radio-frequency sensors measure propellant levels in microgravity, crucial prior to space-based propellant transfer operations, as inaccurate level measurement directly impacts transfer precision and safety margins. The camera system was significantly upgraded, providing about 50 video feeds covering key areas of the vehicles. All video data is downlinked to the ground in real-time via a redundant, high-speed, low-latency 480 Mbps Starlink connection, marking the first time Starship utilizes its own Starlink network for broadband real-time vehicle video transmission.
The Starbase Pad 2 launch site was activated concurrently with the V3 rocket, located about 300 meters west of the previous launch point used for all prior Starship test flights. Propellant storage capacity was expanded, and pumping capability significantly increased to shorten vehicle loading times. The chopstick manipulator arms on the launch tower were redesigned to be shorter and capable of faster movement to better track returning vehicles during capture operations. Their primary actuators were changed from hydraulic to electromechanical drive, improving speed, redundancy, and reliability. The quick-disconnect arm for loading propellant into the Starship upper stage was structurally reinforced and repackaged, rotating to a position farther from the rocket during launch for protection. The launch mount structure and hold-down clamps were completely redesigned, focusing on improving load distribution, hold-down release reliability, and vehicle protection during ascent. Internally, the bidirectional flame trench and top flame deflector were newly designed with the goal of completely eliminating post-launch ablation, making these surfaces require no refurbishment after each launch. The quick-disconnect apparatus for Super Heavy booster propellant loading was moved to the opposite side of the pad and split into separate methane and oxygen mechanisms. Various vent, isolation valves, and filters were relocated to a hardened bunker on the side of the pad. This bunker significantly shortened the distance to the rocket while isolating oxygen and methane systems into separate rooms for safety.
In an IPO prospectus filed with the U.S. Securities and Exchange Commission (SEC) on May 20, SpaceX for the first time formally outlined subsequent operational plans for Starship. The company explicitly expects Starship to begin delivering payloads to orbit in the second half of 2026. The continued deployment of the massive Starlink constellation is highly dependent on Starship. The prospectus notes that SpaceX's "current operational rockets, including Falcon 9 and Falcon Heavy, cannot deploy V3 satellites and V2 Mobile satellites." According to plans, a single Starship launch could carry up to 60 V3 Starlink satellites or 50 V2 Mobile satellites. SpaceX plans for V3 satellites to provide up to 1 Terabit per second (1 Tbps) of throughput each and to offer more comprehensive direct-to-cell phone services via V2 Mobile satellites by 2027.
Propellant transfer is another critical technology that must be demonstrated in the near term. The prospectus acknowledges: "In-space refueling is complex, and we have not yet demonstrated or attempted it. We may be unable to develop, commercialize, scale, or successfully implement these or other strategic initiatives on our currently anticipated timeline, or at all." This risk disclosure underscores the indispensability of propellant transfer. NASA holds a contract valued at over $4 billion with SpaceX for a human lunar lander, requiring Starship V3 to demonstrate propellant transfer capability.
Regarding overall project progress, a statement made prior to launch remained positive but cautious, noting that Starship production lines are full, with approximately 10 or more ships and about half that number of boosters expected to be completed this year. Therefore, unless the launch pad is destroyed, issues would not cause major setbacks. The prospectus also outlines longer-term capability goals. SpaceX expects Starship V3 to be capable of delivering 100 tons to low Earth orbit, with future versions increasing to 200 tons, and ultimately achieving a total annual launch capacity of 1 million tons. Around this capability, SpaceX proposes远期应用场景 including orbital AI data centers, lunar rare resource mining with direct return to Earth, and global point-to-point high-speed passenger and cargo transport.
The timing of the twelfth flight closely overlapped with SpaceX's IPO process. On May 20, the day before the planned launch, SpaceX formally submitted its上市申请 to the SEC. The market expects this IPO could raise up to $75 billion, targeting a company valuation of approximately $1.75 trillion, which, if achieved, would be the largest IPO in history. The prospectus首次全面披露 financial data behind the Starship program. Cumulative investment in Starship development exceeds $15 billion. This included $3 billion in 2025 and nearly $900 million in the first quarter of 2026. These substantial expenditures directly impact company profitability. In 2025, SpaceX's space division reported an operating loss of $657 million. In Q1 2026, the division's operating loss widened to $662 million, explicitly attributed to continued increased investment in Starship and related infrastructure. SpaceX stated it expects to continue increasing R&D spending this year, with about 80% spent on in-house manufacturing.
In the risk factors section, the prospectus lists Starship as the primary risk. It明确指出 that failure or delay in the large-scale development of Starship or in achieving the required launch tempo, reusability, and后续能力 could delay or limit the company's ability to execute its growth strategy, including deploying next-generation satellites, global satellite-to-mobile connectivity, and orbital AI computing, and could have a material adverse effect on business, financial condition, operating results, and future prospects. The prospectus further列举具体挑战: "reliable high-cadence return-to-launch-site operations" for the Super Heavy booster and Starship upper stage, rapid and frequent reuse of the vehicles, and "managing public and regulatory tolerance for anomalies during the transition to frequent operational flights." It特别指出 that if Starship fails to achieve full reusability or rapid turnaround, it could face consequences like increased cost per launch, delays in large constellation deployment, delayed revenue growth, and increased capital requirements. Regarding the new business direction of orbital AI computing, the prospectus is more blunt: "Scaled AI compute satellites will require Starship's full reusability to be economically attractive."
Consequently, the outcome of the twelfth flight carries weight in capital markets beyond mere technical testing. An analyst commented that for an IPO so heavily reliant on narrative and symbolism, this flight is seen as the most important pre-IPO catalyst on SpaceX's calendar, noting that investor enthusiasm could "dampen sharply" with a poor outcome. The FAA administrator stated in a pre-launch industry forum that SpaceX leadership outlined a long-term vision of reaching 10,000 launches per year within five years. The response was cautious, indicating the FAA would need to see more reliability before approving such expansion and noting that while the FAA is not currently a limiting factor for space launches, underfunding could make it one in the future. A Stanford space expert and former NASA center director assessed the risk of this launch as "enormous," adding that with the government choosing a commercial contract model for the human lunar lander, it is now up to the companies to deliver. A partner at a strategy consulting firm analyzed from a commercial perspective that a successful launch would truly pave the way for more space infrastructure and lunar contracts. A senior investment portfolio manager whose fund holds SpaceX stock offered a more measured view: SpaceX's execution track record is unparalleled, but Starship itself is extremely difficult to get right. The team will solve it at the appropriate time, unlocking tremendous value. This judgment summarizes a common sentiment in investment circles: acknowledging the technical difficulty but betting that SpaceX's execution capability will ultimately deliver. The mission, where the spacecraft completed primary procedures but the booster experienced anomalies and a key ignition test was skipped,恰恰印证了这一判断: technical difficulties are real, but SpaceX continues to make gradual progress.