Editor’s note: this piece covers the double-catch plan SpaceX described on 21 October 2026. Flight 15 has since flown, and the result was split — the upper stage was caught at Starbase, the booster catch was waved off. The plan is set out below; the outcome is in the final section.
SpaceX used its third-quarter earnings call to tell investors that Flight 15 would try to catch both the Super Heavy booster and the upper stage — which the company calls “Ship” — using the mechanical arms of its launch towers at Starbase in South Texas. Chief executive Elon Musk announced the plan on the 21 October call, and SpaceX played a simulation of the upper-stage recovery during the livestream on X, according to Engadget and Gizmodo.
The plan, as SpaceX set it out
- Flight 15 would fly Booster 22 and Ship 42, the fourth pairing built to the larger Starship V3 design standard, from Orbital Launch Pad 2; the full stack stands roughly 123 metres tall, the largest rocket ever built, per Techtimes.
- Boosters had been caught repeatedly since October 2024 — three times by Gizmodo’s count — but every upper stage had splashed down in the ocean.
- The simulation ran about 80 seconds under the label “Up Next: Ship Catch”, Gadget Review reported; Engadget pointed viewers to the 1:07:19 mark of the call and to an excerpt posted by investor Sawyer Merritt.
- Launch-window estimates ranged from Gizmodo’s no-earlier-than 8 November to the four-to-six weeks around Thanksgiving that Musk gave X Takeover attendees at Giga Texas, which bracketed mid-to-late November; the flight went ahead in mid-November, at the early end of that range and after a provisional 19 October date had lapsed.
- Musk put the odds of a first-try success at 50 to 60 percent in September, per Techtimes, and told the October call he thought it was more likely than not, per Gizmodo.
Why it matters
Flight 15 “will be the first time that an orbital launch vehicle has been fully recovered”, Musk told the call, in remarks carried by Gizmodo, which framed a double catch as the precondition for full reusability: fly the same rocket repeatedly and the cost of spaceflight drops far enough to make currently infeasible missions economically viable. Catching Ship at the pad also takes ocean retrieval out of the loop, which is what makes short turnarounds between flights plausible, Gadget Review added.
That arithmetic runs straight through SpaceX’s satellite internet business. Starlink V3 satellites are too large and heavy for any other rocket in the fleet, Techtimes reports, leaving Starship as the only ride to orbit for a generation designed to deliver significantly higher throughput than earlier Starlink satellites. Flight 14 carried 26 of them.
The consumer-facing consequence is capacity. Starlink’s sold-out cells, waitlists and congestion-charge pricing all trace back to how much throughput sits over a given patch of ground, and V3 adds that throughput only as fast as Starship can lift it. A recovered stage that flies again lowers the cost of each batch; a splashed-down one has to be rebuilt. None of the cited outlets put a figure on the capacity added per Starship flight, so the per-gigabit cost case remains an inference rather than a published number — but it is the mechanism that decides whether service opens up in constrained markets, and it is the reason the reuse data feeding the Starlink V3 buildout matters more to subscribers than the spectacle of a rocket landing in a pair of arms.
How the catch is meant to work
The simulation opens on one of two “Mechazilla” launch and catch towers at Starbase, structures fitted with chopstick arms designed to grab a returning stage, in Gizmodo’s description, then pans up to Ship falling back to Earth horizontally in a “belly flop” position after an apparently completed orbital flight, steered through the atmosphere by its forward and aft flaps.
Near the tower, Ship performs a short burn to shed speed and flip upright, then descends between the extended arms, which close around it and bear its weight on hard points along its sides, Gadget Review explained. The booster catch was not shown, Gizmodo noted, but would follow a similar procedure on the site’s other tower.
Ground testing
Ship 42 was lifted by the Pad 2 chopsticks into a simulated catch position in mid-September 2026 to verify hardware alignment and arm positioning, per Techtimes; Gadget Review, citing Teslarati, flagged that the test had not been independently confirmed by a primary SpaceX source.
The engine problem hanging over the attempt
Flight 14 launched on 28 September 2026, reached orbit for the first time and deployed its 26 Starlink V3 satellites, per Techtimes. A Raptor Vacuum engine then shut down prematurely, cutting short a mission expected to run roughly 10 hours; Ship splashed down in the Pacific about three hours after liftoff and Super Heavy came down in the Gulf of Mexico, Gadget Review reported, citing Reuters.
Starship has flown with upgraded Raptor 3 engines three times and they malfunctioned on every one, Gizmodo wrote — which is the whole risk to a catch attempt, because the sequence only begins if Ship completes its orbital flight and reenters over Starbase. Reentry, the belly flop, the flip and the powered approach then have to succeed in order, with no partial credit.
What actually happened, and what comes next
Gizmodo’s coverage since the flight describes a successful upper-stage catch alongside an aborted booster catch, with Super Heavy splashing down in the Gulf of Mexico rather than returning to the tower — the split outcome we set out in our report on the first Ship catch and the 26-satellite Starlink deploy that preceded it. On the plain terms Musk set on the October call, that is a partial result: the harder, never-before-attempted half worked, and the routine half did not.
Musk had indicated that recovered hardware could be turned around and reflown on Flight 16, per Techtimes, which also has NASA Administrator Jared Isaacman telling Fox News he expects near-weekly Starship launches going into 2027 and Musk projecting weekly or twice-weekly flights that year during a White House livestream alongside NVIDIA chief executive Jensen Huang. With the booster lost to the Gulf, the reflight test now rests on Ship 42 alone.
Gadget Review cautioned that a single catch establishes nothing about cadence, and repeated recoveries are what would move rapid reuse from stated target to operational fact — among the milestones Starship still has to clear before rapid reuse counts. Starship remains classified as a developmental launch system undergoing integrated flight testing, Techtimes noted, with 14 integrated test flights completed since the programme’s first launch attempt in April 2023.
Frequently Asked Questions
What engines power the Super Heavy booster, and what do they burn?
Each Super Heavy booster uses 33 Raptor engines burning liquid methane and liquid oxygen, according to Techtimes.
How many Starship launches is SpaceX allowed to fly from Starbase?
The FAA, the US aviation regulator, has approved up to 25 Starship launches per year from Starbase, according to Techtimes. The same report says SpaceX is also building out pad infrastructure at Kennedy Space Center’s LC-39A to support higher cadences across multiple sites.
What else must Starship demonstrate before it can fly crew?
Techtimes lists on-orbit propellant transfer and proving rapid reuse turnaround times at scale as additional milestones required before Starship is considered operational for crewed missions. It notes NASA’s Artemis programme has designated a lunar-adapted version of the upper stage as its Human Landing System, tying the test campaign to the timeline for returning humans to the Moon.














Comments 1