Vulcan failure traced to weak first-stage booster, putting BE-4 engine in focus
ULA says the trouble began just after liftoff, not at the 10-minute mark. The engine family involved also powers Blue Origin's grounded New Glenn.

Key Takeaways
- ULA says the booster underperformed right after liftoff, so the stages separated too low and too slow to reach orbit.
- The solid rocket booster nozzle fix and the Centaur V upper stage both appear to have worked as intended.
- The core booster flies on Blue Origin's BE-4 engines, which also power New Glenn, a key Artemis vehicle.
- The flight was the first of 38 Vulcan launches Amazon booked for its Leo constellation.
United Launch Alliance has traced Sunday's loss of a Vulcan Centaur rocket and its 40 Amazon Leo internet satellites to the rocket's first-stage booster, which the company says underperformed almost from the moment it left the pad.
Controllers lost telemetry and declared the mission lost roughly 10 minutes into flight. In a Tuesday update, however, ULA said early data point to "low booster performance" just after liftoff. Because the booster was not producing enough thrust, the stages separated at too low an altitude and below the speed needed to reach orbit.
That change in the timeline matters. A failure first noticed late in flight could have pointed toward the upper stage. ULA's account places the problem on the core booster instead, and that stage flies on two BE-4 engines built by Blue Origin, the rocket company of Amazon founder Jeff Bezos. Amazon's satellites were lost on a rocket whose main engines come from Bezos' own space venture.
ULA has not identified the specific fault. The investigation has only started, and the company said finding a root cause will take more time. Nothing ULA has said so far blames the engines rather than some other part of the booster, and that distinction will decide how far the consequences reach.
The fixes that held
The parts of Vulcan that had drawn the most scrutiny came through this flight without trouble. On the rocket's second launch, in October 2024, and again on its fourth, in February 2026, a nozzle broke off one of the strap-on solid rocket boosters about 30 seconds after liftoff. Vulcan still delivered its payloads both times. After the second incident, which happened on a Space Force mission, the Space Force paused national-security launches on Vulcan until ULA resolved the nozzle problem.
In an Oct. 28 update, ULA said it had made "design and process improvements to the nozzle." Those changes appear to have worked. "All indications are that the solid rocket booster performance was nominal," the company said Tuesday.
The Centaur V upper stage, powered by two RL-10C engines, also performed as designed, according to ULA, but it could not make up for the shortfall left by the booster. That is a welcome data point for NASA, which plans to integrate the Centaur V into its Space Launch System moon rocket.
Still, Vulcan's record now looks shaky. Three of its five flights since its January 2024 debut have had problems, and this is the first that cost a payload. The rocket is meant to take over from ULA's long-serving Atlas V, which has six flights left before retirement. A successor needs a run of clean missions to earn trust, and Vulcan is instead grounded with a fresh failure under investigation.
Why the engine question reaches beyond ULA
If investigators find a serious flaw in the BE-4, the effects would extend well past one rocket. The same engine powers the first stage of Blue Origin's New Glenn, a central piece of hardware for NASA's Artemis lunar program. New Glenn is already out of service after a large explosion on its test stand on May 28.
That leaves two of the country's most powerful rockets sidelined at once, and the timing is poor. SpaceX plans to retire its Falcon 9 and Falcon Heavy and move its launch business to the fully reusable Starship. Space.com expects SpaceX to devote Starship's early flights largely to its own projects, including a next-generation Starlink constellation and its Starmind AI data center network. Taken together, that points to a tight market for launch capacity over the next few years.
In that setting, a long Vulcan investigation would hurt more than it would have when Falcon 9 had room to spare. The best outcome for ULA, Blue Origin and NASA would be a cause specific to this one booster, such as a manufacturing defect, rather than something in the engine design. An engine-level problem would put two launch vehicles and parts of the Artemis schedule on hold behind the same fix.
For Amazon, the loss hits the network meant to compete with SpaceX's Starlink. The company has booked about 100 launches to build a constellation that will eventually exceed 3,200 satellites, spread across Vulcan, Atlas V, New Glenn, Falcon 9 and Arianespace's Ariane 6. Vulcan holds 38 of those bookings, and Sunday's flight was the first.
Amazon has about 400 operational Leo satellites in orbit from 14 earlier launches, so the 40 lost on Sunday amount to about a tenth of the current fleet. The satellites can be rebuilt. The larger concern is the 37 Vulcan launches still on the manifest. With New Glenn also grounded, two of the rockets Amazon was counting on most are unavailable, and the rest of its build-out depends on Atlas V, which is near retirement, Ariane 6 and a SpaceX rocket owned by its direct competitor.
The next meaningful milestone is ULA's root-cause finding. Until then, every Vulcan customer is waiting on the same answer: whether Sunday's failure came from one bad booster or from the engines that both Vulcan and New Glenn depend on.




