Amazon is learning that the hardest part of competing with SpaceX is getting off the ground. The tech giant relies on outside launch providers to build its satellite internet network, while its rival flies its own rockets.
That distinction became painfully clear over the weekend when a failed launch destroyed 40 Amazon satellites just as SpaceX reached another milestone in reusable spaceflight.
Twenty years without failure
United Launch Alliance’s (ULA) Vulcan rocket failed roughly 10 minutes after lifting off from Cape Canaveral Space Force Station in Florida on Sunday. None of the satellites aboard, destined for Amazon Leo, formerly Project Kuiper, reached orbit.
The failure also dealt a blow to Boeing and Lockheed Martin, which jointly own ULA. The launch provider had operated without a total mission loss since its formation in 2006.
Amazon acknowledged the risks associated with first missions and said it remains confident in ULA’s long-term plans for Vulcan.
Vulcan Was Meant to Be Amazon's Workhorse
Amazon Leo satellites carried per launch, by rocket
Bar chart. Rocket. Atlas V: Satellites per launch 29. Vulcan: Satellites per launch 40. Ariane 6: Satellites per launch 36. Falcon 9: Satellites per launch 24. New Glenn: Satellites per launch 48.
Source: GeekWire, azmth.space Amazon Leo tracker
Trouble under the hood
KeyBanc analyst Michael Leshock said preliminary data points to Vulcan’s first stage, raising the possibility of problems with its two BE-4 engines manufactured by Jeff Bezos’ Blue Origin.Telemetry suggested the anomaly came after the first stage separated, pointing investigators toward the upper stage rather than the Blue Origin-built engines on the booster.
The same engines power Blue Origin’s New Glenn rocket, which exploded during testing in May following an oxygen-valve failure. ULA had retrofitted its BE-4 engines ahead of Sunday’s mission.
Vulcan was returning from a nine-month grounding for its fifth flight since debuting in 2024. For Amazon, it was supposed to mark the beginning of its largest launch agreement, covering 38 Vulcan missions.
Amazon’s shrinking launch options
Vulcan was expected to carry more than 1.5K of Amazon’s first-generation satellites, making any prolonged grounding particularly problematic given the company’s regulatory obligations.
The Federal Communications Commission (FCC) requires Amazon to have all 3.2K satellites in service by July 30, 2029. The regulator conditionally waived an earlier deadline on June 5 that required half the constellation to be deployed by July 30, 2026.
Amazon had completed 14 satellite missions before Sunday’s failure and has yet to begin commercial service, despite securing distribution partnerships for its broadband network.
Its next launch aboard Ariane 6 will carry 36 satellites and is scheduled for Nov. 23 at the earliest. Another 24 reserved launches aboard Blue Origin’s New Glenn remain on hold.
SpaceX widens its lead
While Amazon’s launch providers contend with technical setbacks, SpaceX is getting closer to making rocket reuse routine.
Starship lifted off on Friday, Nov. 13, and deployed 26 Starlink V3 satellites, following its first orbital flight in September.
At roughly 4:20 p.m. ET on Saturday, Starship returned from orbit and was caught by its launch tower, marking the first successful catch of an orbital rocket’s upper stage.
The mission wasn’t flawless. A low-fuel reading forced the Super Heavy booster to end its boostback burn early and divert to a splashdown in the Gulf of Mexico.
Analysts nevertheless viewed SpaceX’s progress and deployment of next-generation satellites as positive developments, particularly in light of its competitor’s setbacks.
The economics of owning the rocket
Replacing the lost satellites is relatively straightforward for Amazon. The company said in March that its Kirkland, Washington, factory could manufacture up to 30 satellites a week. Securing enough launches is proving considerably harder.
Starlink already operates more than 11K satellites and serves millions of customers worldwide. Owning its launch infrastructure allows SpaceX to expand that network without competing for space aboard another company’s rockets.
Successfully reusing Starship’s upper stage could further reduce launch costs and shorten the time between missions, giving SpaceX another advantage over competitors paying for rides into orbit.
Amazon can afford to build thousands of satellites. What it can’t buy back is the time SpaceX spends putting its own into orbit.

