Amazon's latest Atlas 5 launch added satellites to a constellation that is running years behind schedule and is now short on rockets. The company has 241 spacecraft in orbit. It needs 1,616 to meet the halfway milestone in its FCC license by July, and it has told the FCC it will not get there.
The company filed for a 24-month extension in January, asking until July 2028 to deploy half its 3,232-satellite constellation. The filing did not dress up the reason. "The shortage has been driven by manufacturing disruptions, the failure and grounding of new launch vehicles, and limitations in spaceport capacity," Amazon told the FCC. That is a rare piece of candor from a company that spent years insisting everything was on track.
The numbers do not cooperate. Amazon has reserved more than 100 launches across four providers — United Launch Alliance, Arianespace, Blue Origin, and SpaceX — for a combined contract value exceeding $10 billion. It has invested over $200 million in ULA facility upgrades at Cape Canaveral. It built a 172,000-square-foot manufacturing plant in Kirkland, Washington capable of producing 30 satellites per week — over 1,500 per year. Production is not the bottleneck. The rockets are.
ULA's Vulcan Centaur, which Amazon contracted for 38 launches, has been grounded since a solid rocket booster anomaly during a February Space Force mission. Months of investigation lie ahead before it flies again. Blue Origin's New Glenn has not yet launched a single Amazon satellite, despite 12 contracted flights plus 15 options. Ariane 6 is new and working through its own ramp. That leaves Atlas V — three more launches remain — and SpaceX, where Amazon purchased 10 additional Falcon 9 flights on top of three already completed.
The SpaceX bookings carry a history. Amazon originally excluded SpaceX from the bidding process, a decision driven by the Bezos-Musk rivalry that resulted in a 2023 shareholder lawsuit. By the time Amazon came around to buying Falcon 9 launches, SpaceX, according to its own disclosures, had deployed more than 9,000 Starlink satellites and acquired 9 million subscribers. Amazon has 241.
Amazon's January filing projected roughly 700 satellites in orbit by the July deadline — less than half the required 1,616. The FCC's mid-point milestone is not a soft target. It is the condition under which Amazon retains the right to operate the constellation at all. Missing it does not mean a fine. It means losing the license.
SpaceX filed opposition to the extension in early April, arguing Amazon violated orbital debris requirements. Amazon countered that SpaceX moved its own satellites into overlapping orbits specifically to create the collision risk. The FCC has not ruled.
What is structurally interesting is the position Amazon has put itself in. It spent $10 billion on launch reservations and $200 million on ground infrastructure, built a factory that can outpace any available rocket, and is still waiting on vehicles that have not arrived or are currently grounded. The company that spent years on the sidelines watching SpaceX build a working constellation is now paying a premium to catch up on a vehicle mix that does not yet exist at scale.
The FCC extension request is pending. The July deadline is not. At current deployment rates, with Atlas V nearly exhausted and Vulcan grounded, Amazon will enter the hearing with fewer than half the satellites it needs — and a SpaceX opposition brief on file.
