Amazon Files for 5,105 Direct-to-Device Satellites on Globalstar's Big LEO Spectrum, in the Same Week It Misses the Milestone on the System It Already Has

What happened

Amazon filed application SAT-LOA-20260601-00224 with the FCC on 24 July 2026 for a direct-to-device system of 5,105 satellites, operating at 510, 540, 560, 570 and 580 kilometres across five shells, with the filing reserving orbital room for more than 11,000 spacecraft in total. Service links would use the 1.6 and 2.4 GHz bands assigned to Globalstar, with 1.5 and 1.6 GHz outside the United States, and Ka- and V-band for gateway and command links. Amazon told the Commission the system would operate in concert with the HIBLEO and C-3 systems acquired from Globalstar, and described the target service as mobile data, text messaging, emergency communications, in-vehicle connectivity and IoT.

The spectrum is not Amazon's yet. Amazon agreed in April 2026 to acquire Globalstar in a cash-and-stock deal valued at roughly $11bn, expected to close in 2027 subject to regulatory approval, and the application rides on that closing. Deployment is planned from 2028.

Six days after the filing, on 30 July, Amazon reached the FCC milestone requiring half of its separate 3,232-satellite Ku and Ka-band Leo system in orbit and arrived roughly 1,285 satellites short. Under the conditional waiver the Space Bureau granted in June, every Leo satellite it launches into that system now operates on a non-interference basis until it certifies half deployment or March 2028.

The band the FCC closed to everyone else

Big LEO is a small, crowded and deliberately exclusive piece of spectrum. In April 2026 the Commission reaffirmed that structure, dismissing requests from SpaceX, Kepler Communications, Satelio IoT Services and AST & Science for market access or sharing, together with SpaceX's rulemaking petition RM-11975 to revise the 1.6/2.4 GHz sharing plan and Iridium's petition to optimise the 1.6 GHz band. The stated reasoning was that sharing Big LEO is impractical, that the interference risk to incumbents is significant, and that the existing structure supports investment.

That order is what makes Amazon's application possible on these terms, and it is also what makes it contestable. An exclusivity granted on the premise of protecting incumbent investment now underwrites a 5,105-satellite mass-market system operated by the largest new entrant in the band, acquired rather than earned through the sharing process the same order foreclosed. Operators told to stay out have an obvious argument about what the premise was worth.

Why it matters for dispute formation

Three approvals have to line up and each is a separate proceeding with its own record. The Commission must consent to the transfer of control of Globalstar's licences, grant the new system authorisation, and be satisfied that a constellation of this size can be operated within the interference envelope the Big LEO plan assumes. A party unhappy with any one of them can press the others, and the sequencing gives competitors more than one place to file.

Amazon's own record supplies the second front. A milestone regime works on the premise that an authorisation is a promise to deploy, and Amazon has just demonstrated, in the same fortnight, what happens when the promise runs behind the schedule. Any condition the Commission attaches to a 5,105-satellite grant will be read against that, and the natural response is milestones tighter than the first system carried, which creates its own future dispute over whether they have been met.

The third exposure is contractual rather than regulatory. Apple's leverage over the network its emergency features depend on has been an equity stake and a capacity prepayment; after the sale it is a supply agreement with Amazon. Nothing about that is unusual commercially, but it concentrates a safety-of-life service, a consumer-device rivalry and a spectrum authorisation into one counterparty relationship, and the terms that govern it are not on the public record the way a licence condition is.

Who's exposed

What to watch

• Oppositions and petitions to deny against SAT-LOA-20260601-00224, particularly from operators dismissed from Big LEO in April 2026.
• Whether the FCC attaches deployment milestones to any grant that are tighter than those on Amazon's Ku / Ka-band system, and what the bond looks like.
• The transfer-of-control proceeding for Globalstar's licences, and whether the D2D application is held pending that closing in 2027.
• Any move to revisit the Big LEO exclusivity now that the band would carry a constellation of this scale, including renewed rulemaking petitions.
• Globalstar's ITU position for HIBLEO following the Radio Regulations Board's refusal to extend its milestone, since the international filings underpin the domestic case.