On Saturday, July 26, 2026, Amazon submitted an application to the US Federal Communications Commission (FCC) to build a new constellation of 5,105 low Earth orbit (LEO) satellites specifically designed to connect ordinary smartphones directly from space — with no special hardware, external antennas, or device modifications required. The plan calls for the first satellites to launch in 2028, on a scale that surpasses the original Project Kuiper constellation by nearly 60%.
It is crucial to understand that this is not Project Kuiper, Amazon's broadband constellation of approximately 3,236 satellites that already has 375 units in orbit. This is a separate constellation, nearly twice as large, focused exclusively on direct-to-device (D2D) connectivity — the same type of service that Starlink (SpaceX) has been testing with T-Mobile since 2024 and that Apple already offers via Globalstar for emergencies on iPhone 14 and later models.
The connection to Globalstar is precisely the most strategic element of the entire plan. Earlier in 2026, Amazon announced the acquisition of Globalstar for $11.57 billion — the same company that provides Emergency SOS via satellite for iPhones. The FCC filing makes it explicit that the new constellation "will use Globalstar's spectrum" and that Amazon plans to "collaborate with Apple on future satellite services using the expanded network." This means Apple could benefit from a modern 5,000+ satellite constellation to replace Globalstar's aging fleet of 24 satellites, dramatically expanding the capability of iPhone satellite services.
Technically, Amazon's satellites differentiate themselves by processing signals in orbit before relaying them to Earth — a regenerative payload that reduces latency, increases capacity, and enables flexible beamforming. Instead of simply amplifying and forwarding (like traditional "bent-pipe" satellites), they demodulate, route, and remodulate signals onboard. For developers, this means lower latency for real-time applications (voice, gaming, remote control) and higher capacity per cell.
The filing also requests permission to use Iridium's spectrum outside the US, putting Amazon on a collision course with Rocket Lab, which is in the process of acquiring Iridium. L-band spectrum is one of the most valuable assets for direct-to-device communication due to its ability to penetrate buildings and work indoors. The carrier partnership model follows the Starlink/T-Mobile playbook: use the carriers' licensed spectrum, their customer relationships, and their billing systems. Amazon competes with Starlink (already in beta with T-Mobile), AST SpaceMobile (five BlueBird satellites testing with AT&T and Verizon), and Lynk Global. The 5,105-satellite scale, if realized, surpasses them all, but space is unforgiving — 2029-2030 is a more realistic bet for meaningful coverage.
Amazon's move represents a significant escalation in the commercial space race. Unlike Starlink, which started as a fixed broadband service and only later added mobile capability, Amazon is designing its D2D constellation from the ground up with the cell phone as the primary target. This means the network architecture, satellite payloads, and carrier partnerships are all born designed for the ordinary consumer. The Apple deal adds an extra layer of credibility: having the world's most valuable company as an anchor customer gives Amazon financial and regulatory leverage that few competitors have. The lingering question is whether Amazon can execute on its ambitious launch schedule — putting more than 5,000 satellites into orbit in three years would require an industrial cadence unprecedented in space history.
Sources: The Verge, Reuters, Ars Technica
