Amazon files for a 5,105-satellite direct-to-device constellation
What happened
Kuiper Systems LLC ("Amazon Leo"), a wholly owned Amazon.com Services subsidiary, applied to the FCC for authority to launch and operate up to 5,105 LEO NGSO satellites delivering direct-to-device connectivity to consumer, enterprise and government users. Orbital architecture is five altitude/inclination shells — 510km/56.3°, 540km/84°, 560km/58.2°, 570km/73°, 580km/55.7°. Note a detail the secondary chatter missed: the filing defines 11,045 total orbital slots across those shells while explicitly capping the system at 5,105 satellites ("While the total number of defined orbital slots in Table 1 exceeds 5,105, the Amazon Leo D2D System will not exceed 5,105 satellites") — that is a deliberate ~2.2x placement-flexibility envelope, not a headline count. Service links are the 1.6/2.4 GHz MSS band (1610–1618.725 MHz uplink, 2483.5–2500 MHz downlink), which Amazon acquires via the pending Globalstar transaction and which the filing states will be used "following the closing" of that deal. Feeder links are Ka-band (17.3–18.6, 18.8–20.2, 27.5–30.0 GHz) and V-band (37.5–42.0, 42.0–42.5, 47.2–50.2, 50.4–51.4, 51.4–52.4 GHz), with Ka/V TT&C. Two scope points matter and were misreported in the X-sourced version of this story: (1) Amazon is not requesting Supplemental Coverage from Space inside the United States — "Amazon Leo is not requesting authority for Supplemental Coverage from Space ('SCS') service within the United States at this time" — it seeks SCS only outside the US, across 1427–2690 MHz and 3300–4200 MHz; (2) the extra-territorial MSS request covers 1518–1525 MHz, 1610–1621.35 MHz, 1668–1675 MHz and 2483.5–2500 MHz. Amazon separately asks to operate uplinks in 1618.725–1621.35 MHz outside the United States "in any jurisdiction where Iridium is not authorized to operate," as Globalstar's successor, while conceding Iridium has exclusive access to that band inside the US.
Read-across for SES
Three separate implications, in descending order of importance to SES. First, 2 GHz. Amazon leads SpaceConnect, the Brussels bloc opposing COM(2026)311. A 5,105-satellite D2D application, filed with acquired MSS spectrum and a credible feeder-link architecture, is the strongest available evidence for SpaceConnect's core argument: that commercial D2D supply is abundant and an EU reservation of 2 GHz for IRIS² is a subsidy rather than a fix for market failure. Expect this filing to appear in SpaceConnect's EU submissions. This is the channel by which a US licensing action damages SES's most valuable EXISTENTIAL position. Second, the D2D field is closing at five. Iridium/Rocket Lab, Amazon/Globalstar, SpaceX, AST — and now this as a distinct, funded fifth architecture. None is European. SES's non-Musk, European-anchored identity is a structural asset only where the buyer is choosing on sovereignty; on commercial D2D the field is now crowded enough that scarcity arguments stop working. Third, a live spectrum collision with Rocket Lab/Iridium. Amazon's request to use 1618.725–1621.35 MHz outside the US wherever Iridium is unauthorized is a direct encroachment on the asset Rocket Lab is paying $3.6bn-bridge-funded money to acquire, and it lands before that deal closes mid-2027. Per priors' acquisition-cascade check, this is the kind of secondary effect more material than the headline: it bears on the EMSS/GMDSS/DoD PTT contracts where SES/Intelsat General competes with Iridium. Analyst read worth tracking but not asserting: @Megaconstellati argues the Viasat L-band overlap (1518–1525 MHz) makes Viasat/Space42's ~2,800-satellite Equatys JV "likely dead" and raises whether Amazon acquires Viasat next — plausible, unverified, and Viasat is already on the M&A watch.
As the brief filed it
[AGE: 3d] Amazon's Kuiper Systems filed FCC application SAT-LOA-20260601-00224 (Jul 24) for a 5,105-satellite direct-to-device constellation across five shells at 510–580km, running Globalstar's 1.6/2.4 GHz post-close plus L-/S-band abroad — a fifth funded D2D system, none European, and it reaches into Iridium's and Inmarsat/Viasat's L-band outside the US. Hands SpaceConnect's lead member a concrete "D2D supply is abundant" argument against reserving 2 GHz for IRIS².