Amazon Files to Build a 5,000-Satellite Network for Direct-to-Phone Service

The FCC filing lays out a plan to beam voice, texts, and emergency alerts straight to ordinary phones by 2028, putting Amazon in more direct competition with Starlink and SpaceX.

Amazon wants to put more than 5,000 satellites into orbit for a service that would let ordinary cellphones connect straight to space, no cell tower required. The company filed an application with the FCC on Saturday for a new Leo (low Earth orbit) constellation built specifically for direct-to-device connectivity, covering voice calls, text messages, data, and emergency services. If the FCC signs off, Amazon says deployment would start in 2028.

The plan calls for partnerships with mobile carriers, the same basic model Starlink uses with T-Mobile to offer satellite texting in places with no cell coverage. Amazon hasn't named specific carrier partners yet, but the structure it's describing, satellites acting as orbital cell towers that hand off to a phone company's network, is now a familiar one in the industry.

What makes this filing notable is the spectrum behind it. Amazon says the new constellation will run on spectrum licensed to Globalstar, the satellite operator Amazon agreed to acquire earlier this year. That deal also came with Globalstar's existing contract to provide direct-to-device satellite service for Apple, and Amazon says it intends to keep working with Apple on satellite features using this expanded network. So this isn't a from-scratch venture. It's Amazon building out infrastructure on top of spectrum rights and a customer relationship it just bought its way into.

There's a wrinkle, though. According to a report from Advanced Television, Amazon's FCC application also asks to use spectrum tied to Iridium's satellite network for operations outside the US. That's notable timing, because Rocket Lab is currently in the process of acquiring Iridium, with its own plans to expand that network. Two well-funded companies angling to expand satellite constellations on overlapping spectrum outside the US is the kind of regulatory tangle that could take a while to sort out.

How this fits into Amazon's bigger space bet

This direct-to-device network wouldn't stand alone. Amazon frames it as a complement to the broadband satellite service it's already building under its Leo project, the constellation better known by its former name, Kuiper. That existing broadband constellation is planned to include roughly 3,200 satellites, and as of July 2nd, Amazon says more than 375 of them are already in orbit.

The new direct-to-device constellation would be nearly double that size, over 5,000 satellites, and serves a different purpose. Broadband Leo satellites are built to deliver internet service to fixed terminals, dishes on rooftops or in remote areas. Direct-to-device satellites are built to talk to the phone already in your pocket, with no extra hardware needed. Amazon says these new satellites will process signals in orbit before relaying them down to improve performance, rather than simply acting as a passive relay.

Beyond consumer phone service, Amazon is pitching the network for disaster response and emergency messaging, use cases that have become a selling point across the industry as satellite-linked phones prove useful when terrestrial networks go down during storms, wildfires, or other disasters.

Racing an already-moving market

Amazon isn't entering an empty field. SpaceX has been rolling out direct-to-cell Starlink service with T-Mobile, and Apple already offers satellite-based emergency messaging on iPhones through its Globalstar partnership, the same partnership Amazon is now inheriting. AST SpaceMobile has been building its own direct-to-device constellation aimed at similar carrier partnerships. The pitch across all of these efforts is the same: turn satellites into an invisible extension of the phone network, so coverage gaps stop being a hard boundary.

Amazon's advantage is scale and capital, plus a spectrum position it acquired rather than built from zero. Its challenge is time. A 2028 deployment target for a 5,000-plus satellite constellation is an aggressive build schedule, especially for a company that's still in the early stages of launching its first Leo broadband constellation. Getting from 375 satellites in orbit to a fully operational broadband network, while simultaneously standing up a second, larger constellation for phone service, is a heavy lift even by the standards of a company with Amazon's resources.

The FCC application is a first step, not a green light. Regulatory review, especially with the Iridium spectrum overlap in play, could shape the timeline as much as the engineering does. But the filing makes clear that Amazon isn't treating satellite-to-phone service as a side project tacked onto its broadband ambitions. It's building a second, dedicated network for it, and betting that owning the spectrum and the carrier relationships now will matter more than launching first.