Science2 publishers3 min readPublished
SpaceX puts 26 working Starlink satellites on Starship's first orbital flight
Flight 14 opens its window on 22 September carrying real payload for the first time, though neither stage will attempt a landing and Starship is flying at roughly four a year against the 25 test flights SpaceX officials once hoped for.
The Scientist · Science desk
What happened
- SpaceX has opened a launch window at 7.15am CDT on 22 September at its Starbase site in Texas for Starship's 14th flight. The flight is to reach 275 kilometres and splash down in the Pacific west of Chile.
- All 13 earlier Starship flights were suborbital at best and carried only dummy satellites, while flight 14 is to release 26 Starlink V3 satellites, the first working payload the vehicle has flown.
- Three of the 26 satellites have been modified with cameras that can scan Starship's heat shield while the vehicle is in orbit and transmit the images to engineers on the ground.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- capability Alconcel's reading makes the deployment a sizing test as much as a delivery test: the V3s are bigger than earlier Starlinks, and the Starmind data centre spacecraft SpaceX has proposed are substantially bigger again.
- exposure Both stages are still third-generation hardware inside a fail-fast development programme, so functioning constellation hardware is now what a bad flight destroys.
- constraint Refuelling in space remains undemonstrated. The Moon and Mars use cases stay out of reach whatever happens over the Pacific next week.
- contradiction AFP calls July's flight largely a success where New Scientist records a booster that hit the water hard, and the gap decides how much the announced hardware and software changes have to cover.
SpaceX said on its website that "Flight tests until this point have intentionally flown passively safe suborbital trajectories to maximize public safety." The company also said: "By going to orbit, the next phase of developing Starship to be fully and rapidly reusable can begin." [21][22] Ten hours over six orbits is about 100 minutes a lap at 275 kilometres [29]. That is enough time to release satellites and then commit to a re-entry.
The 26 satellites are SpaceX's own, joining the Starlink constellation the company already operates [30]. Leah-Nani Alconcel at the University of Birmingham said the V3s fly in a lower orbit than existing Starlink satellites, have higher data capacity, and are larger than previous models [16]. "I don't think it's a huge jump in operational complexity for the launch vehicle, but demonstrating the Starlink V3 delivery to orbit is a significant step," Alconcel said [14].
Neither the booster nor the upper stage will attempt a controlled landing [8]. So the flight will not measure recovery or turnaround. That is where the cost per launch is decided. Davide Amato at Imperial College London said an orbital test indicates confidence inside the company that the rocket will at least survive the journey to orbit and the release of satellites. Re-entry, where recent tests have run into problems, may be another matter, he said [13]. "In short, they're making good progress towards introducing Starship as a successful commercial launcher," Amato said [11].
Cadence is the other open variable. SpaceX officials hoped for as many as 25 Starship test flights last year and five took place, one fifth of that number [27]. Flight 14 will be the third of 2026, which if it goes on 22 September is a pace of roughly four a year [9][28]. "Whether Starship will be 'revolutionary', as promised by their CEO, will depend on whether they will be able to achieve a high launch cadence and on whether their design is cost-effective," Amato said [12]. Alconcel framed the commercial gain as holding two vehicles at once. "It may not be as flashy as the 'chopstick catch' for the vehicle stages, but it would be a big boost to their commercial capabilities to have both the already-established high-cadence Falcon 9 and a proven Starship in their repertoire." [15]
On the previous flight the booster climbed as planned. Then only eight of the 13 engines meant to slow its descent reignited, and it hit the water hard instead of making the planned gentle simulated landing [7]. SpaceX did not respond to New Scientist's request for an interview; on its website the company says "there have been several modifications to hardware and software to address issues seen on the previous flight" [18][19].
What to watch
- Whether a 15th flight follows before the end of 2026. That would lift the annual rate above four.
- Whether SpaceX books a payload on Starship for a customer outside its own Starlink programme.
- Whether the deployed V3 satellites enter operational service, not just separate cleanly from the vehicle.