Blue Origin ended its launch webcast five minutes before New Glenn's upper stage was scheduled to complete its critical second burn on April 19. The company then went silent for an hour before confirming, via social media, that its BlueBird 7 satellite had separated and powered on, one hour later than advertised, into an orbit that was not the one it was supposed to reach.
That silence is part of the story. Blue Origin's hydrogen-fueled upper-stage engine, called the BE-3U, malfunctioned during the second burn, leaving BlueBird 7 in an off-nominal orbit, SpaceNews reported. The first stage booster, the bottom half of the rocket and the part Blue Origin has staked its business case on, landed successfully on a drone ship nine and a half minutes after liftoff. The upper half did not finish its job.
AST SpaceMobile, which owns BlueBird 7, is depending on New Glenn for its 2026 constellation expansion. The company's CEO told investors in March that AST expects 45 satellites in orbit by year's end, with 60 more ready to launch, Reuters reported. BlueBird 7, weighing 6,100 kilograms with a 222-square-meter antenna array when deployed, was supposed to occupy a 460-kilometer orbit at 49.4 degrees inclination, a path that would cover mid-latitude zones for the company's direct-to-device broadband service. Whether it still can is under assessment.
The malfunction is the first upper-stage failure on three New Glenn flights. The BE-3U performed as intended on the inaugural NG-1 mission in January 2025 and on NG-2 in November. Blue Origin has offered no public explanation for what went wrong on NG-3, NASASpaceflight noted.
"A reusable booster is useless if the upper stage can't reliably restart in orbit," said one industry engineer, who requested anonymity to discuss a client.
The business case for Blue Origin's rocket depends on a straightforward argument: if the first stage can be reflown quickly and cheaply, launch costs fall and margins improve. The company has landed boosters on three consecutive attempts and says it plans to reuse hardware from flight to flight. On NG-3, it swapped all seven first-stage engines, the BE-4s, and tested a thermal protection upgrade on one nozzle, demonstrating that reuse is physically working. The BE-4s are a different engine family from the upper stage's BE-3U. What the company has not demonstrated is that the whole vehicle works as a system.
The second burn of the upper stage is the moment a rocket adjusts its path to match the target orbit. A malfunction there, even one that leaves the satellite alive and powered on, can mean the difference between a useful spacecraft and a piece of hardware slowly losing altitude and eventually burning up in the atmosphere. For AST SpaceMobile, which has contracted for multiple New Glenn launches as it builds out its constellation, it is a scheduling problem. For the launch customer paying for a precise orbit, it is a financial exposure.
Blue Origin has said it plans to use engines recovered from NG-2 on future flights, a sign it is serious about reusability at the subcomponent level. That is sensible engineering. But it does not address the upper-stage problem on NG-3. The BE-3U has now misfired once in three flights. Blue Origin has not said what caused the malfunction or whether the engines recovered from NG-2 will change anything.
The company is constructing a narrative around hardware reliability and reusability. That narrative requires every stage to work. Right now, one of them didn't.
