Amazon Leo targets faster deployment cadence as deadline pressure mounts

Amazon's Kuiper satellite factory in Kirkland, Washington, is running fine. The 172,000-square-foot facility can produce 30 satellites per week. The company has more than 200 flight-ready satellites stacked at its Florida integration facility right now. The problem is nowhere near manufacturing.

The problem is rockets.

Amazon, the e-commerce and cloud giant building out its Project Kuiper broadband satellite network, has 212 satellites in orbit. The Federal Communications Commission requires 1,616 deployed by July 30, 2026. On January 30, Amazon filed with the FCC for a 24-month extension or full waiver, citing a shortage of launch vehicles. By Amazon’s own projections, it expects to reach roughly 700 satellites by that deadline—about 900 short of the milestone. On March 23, the company published a blog post announcing it plans to more than double its launch cadence in year two: 20-plus missions, up from seven completed out of 20 planned in 2025.

The announcement reads like a confidence statement. The underlying math is not.

The year-two plan involves squeezing more satellites per rocket. A United Launch Alliance Atlas V will carry 29 satellites to orbit—the heaviest payload the workhorse rocket has ever flown—enabled by an upgraded RL10C engine. Blue Origin’s New Glenn, which flew its first Kuiper mission last year, will scale to 48 satellites per flight. Vulcan Centaur, ULA’s newer heavy-lift vehicle, will carry 40. Amazon said it has invested $200 million in launch infrastructure at Cape Canaveral to support the increased tempo. The company also flew its first Ariane 64 mission in February, delivering 32 satellites—the first of an 18-mission Arianespace contract.

What the March 23 announcement did not mention: Amazon quietly bought 10 additional Falcon 9 flights from SpaceX. That detail only surfaced in the January FCC extension filing, reported by SpaceNews reporter Jeff Foust. Amazon originally excluded SpaceX entirely from its 2022 launch procurement—a deliberate diversification of its launch supply chain away from SpaceX’s Starlink competitor. The company has since converted Blue Origin options to 24 firm New Glenn contracts. Now it is quietly buying rides on Falcon 9 anyway.

The timing creates a problem that goes beyond logistics. On March 11, FCC Chairman Brendan Carr publicly called out Amazon for opposing SpaceX’s proposal to operate a 1-million-satellite orbital data center, while simultaneously filing petitions for its own FCC relief. “Amazon will fall roughly 1,000 satellites short of its deployment milestone,” Carr said, adding that its opposition to the SpaceX proposal “won’t get much traction.” Carr has been a documented SpaceX ally who previously accused the Biden-era FCC of regulatory harassment of the company. Amazon declined to comment at the time.

That is the regulator whose approval Amazon now needs to extend its deployment window. The company is asking for a waiver from the same person who just told the press it is a thousand satellites behind. Amazon’s FCC filing also contains a redacted appendix with its deployment schedule—meaning the public version does not show exactly when Amazon plans to hit what numbers.

The FCC separately approved an Amazon petition in February 2026 to deploy 4,500 additional satellites beyond its original 3,236-satellite FCC authorization from July 2020, which suggests the commission is not entirely hostile to the program. But there is a difference between approving future expansion and excusing a missed present milestone.

The policy stakes are real. Comments filed with the FCC on March 20 by the Information Technology and Innovation Foundation, a Washington technology policy think tank, argued that Amazon’s launch shortfall is not unique to Kuiper—it reflects systemic spaceport infrastructure inadequacy. Cold War-era launch facilities at Cape Canaveral were not designed for the launch cadence modern commercial operators require. Roads, wastewater systems, and payload processing centers are all bottlenecks. ITIF analyst Ellis Scherer noted that launch range coordination among multiple providers is straining federal range capacity, and cited the FCC’s ongoing Space Modernization Proceeding as the appropriate venue for addressing it. If the FCC denies Amazon’s extension, enforcement options include license modification, bond forfeiture, or preclusion from replacement authorizations.

The factory makes the satellites. The range cannot launch them fast enough. That distinction matters for anyone assessing whether Amazon’s Kuiper problem is fixable or structural.

Kuiper’s satellite design gives it a technical baseline worth noting. The network uses optical inter-satellite links with 100 Gbps throughput at ranges up to 2,600 kilometers—on paper competitive with SpaceX’s Starlink Gen 2 laser mesh—per SpaceNews December 14, 2023 reporting on Kuiper prototype tests. Rajeev Badyal, who leads the Kuiper engineering effort, was a SpaceX vice president fired by Elon Musk in 2018. He helped found the Kuiper team alongside other SpaceX alumni. The satellites work. The supply chain to deploy them fast enough is the story.

Starlink, meanwhile, has roughly 9,000 satellites in orbit—CNBC reported around that figure in March 2026—and an active subscriber base. Amazon has not announced commercial service launch timing. The FCC decision on the extension request, which has no publicly stated deadline, will determine whether Kuiper gets another two years to close the gap—or faces spectrum consequences for missing it.

Watch for the FCC ruling and any Carr commentary in the weeks ahead. The January filing sits in a regulatory environment where Amazon is simultaneously asking for grace and fighting its main competitor at the same commission. That is a difficult position to hold.