From the newsletter (November 16; sign-up or follow if you want it sooner): ULA’s first launch failure in 20 years has destroyed 40 Amazon Leo satellites, dealing another blow to its delayed broadband constellation and exposing Amazon’s dependence on struggling third-party rockets to challenge SpaceX’s launch advantage.
United Launch Alliance (ULA), the joint venture between Boeing and Lockheed Martin, suffered its first outright launch failure in its 20-year history yesterday (Sunday) after its Vulcan rocket somehow fell to pieces – “suffered an anomaly” and “separated” – 10 minutes into after liftoff from Cape Canaveral in Florida. Its entire payload – 40 satellites belonging to Amazon Leo – was lost. The timing could hardly be worse. Amazon Leo, formerly Project Kuiper, has had a bruising run-up to its commercial launch. The Vulcan failure is another setback for a business that is already struggling to get enough satellites into low-Earth orbit (LEO). It has had FCC approval for 3,236 of them since 2020; it has so far deployed around 398. Starlink has deployed 13,122 since 2019, and its parent, SpaceX, is targeting 60 next-gen V3 satellites and around 61Tbps of capacity per Starship launch – more than 20 times its payload capacity at present (a Falcon 9 carrying V2s), and about 1.5 times as many satellites and over 20 times the aggregate bandwidth per launch versus Amazon Leo’s current-gen satellites on a problematic Vulcan rocket.
Amazon has the production scale; it reckons it can manufacture 30 satellites a week (over 1,500 annually). But getting them into orbit is a headache. Its alternative launch providers have also been struggling: Blue Origin’s New Glenn rocked suffered a catastrophic on-pad failure in May after an oxygen-valve problem in one of its engines; ULA’s Vulcan has twice suffered solid rocket booster nozzle anomalies on earlier flights; Ariane 6 doesn’t offer the launch cadence to meet its deployment ambitions. Amazon has also bought launches from SpaceX, its principal competitor, using Falcon 9. This was the first of 38 Vulcan launches contracted by Amazon, expected to carry more than 1,500 Leo satellites into orbit. It had also invested in dedicated launch infrastructure at Cape Canaveral, including a new integration facility and mobile launch platform. T
he intention was to get Vulcan launches turning around in days or weeks, rather than months. Amazon said its second Vulcan mission was already being prepared and could follow within weeks of a successful first flight. Instead, it now faces another investigation and potentially months of disruption. It was hoping to begin initial commercial services in 2026. The loss of 40 satellites, on top of the problems with New Glenn and Vulcan, makes that timetable challenging. The firm has invested billions in building a rival to Starlink, has the manufacturing capability to produce satellites at scale, and has assembled a huge portfolio of launch contracts. But it doesn’t control the rockets needed to execute its plan. No such bother for SpaceX – which owns the launch engine, and has just caught its 171-foot upper stage on its return from orbit using a pair of giant mechanical ‘chopstick’ arms on its launch tower in Texas.
