Amazon's satellite division asked federal regulators last week to authorize a brand-new constellation of 5,105 low Earth orbit satellites — a system built for one purpose that its broadband constellation was never designed to handle: reaching the phone in your pocket, anywhere on earth, without a cell tower in sight.

The application, filed by Kuiper Systems LLC with the Federal Communications Commission on July 24, 2026, requests authorization for the Amazon Leo Direct-to-Device system, which would deliver voice calls, text messages, mobile data, and emergency services to unmodified 4G and 5G smartphones starting no earlier than 2028. That three-year runway matters: SpaceX's Starlink Direct to Cell has been live on T-Mobile since July 2025, and Amazon is arriving late to a market that is already moving fast.

Spectrum Is the Strategy, Not the Satellites

The 5,105-satellite figure is the headline, but it is not the story. The architecture underneath it is.

When SpaceX launched Starlink Direct to Cell, it did so by borrowing T-Mobile's licensed terrestrial spectrum under a regulatory framework the FCC calls Supplemental Coverage from Space. That approach requires a separate spectrum-sharing deal with a licensed carrier in each country where Starlink wants to transmit to phones — the satellite operator uses the carrier's frequency allocation, in the carrier's license territory, for the carrier's subscribers. T-Mobile's approval is an American approval; replicating it in Mexico, Australia, or Indonesia means negotiating again.

Amazon's approach is structurally different. The FCC filing specifies that the Amazon Leo D2D system will use spectrum in Globalstar's existing L-band and S-band allocations — approximately 1.6 GHz and 2.4 GHz respectively — Mobile Satellite Service frequencies that were purpose-allocated by the International Telecommunication Union specifically for satellite mobile communications. These are not frequencies borrowed from a terrestrial carrier. They are dedicated satellite spectrum, covered by Globalstar's Mobile Satellite Service licenses in more than 120 countries, which Amazon is acquiring along with the rest of Globalstar's business.

That spectrum is why Amazon agreed to pay approximately $11.57 billion for Globalstar in a deal announced in April 2026 and expected to close in 2027. Without those licenses, Amazon cannot legally beam voice calls to phones in most of the world. With them, Amazon inherits a globally coordinated right to transmit to mobile devices without recreating Starlink's carrier negotiation process country by country.

How the Signal Gets From Orbit to Your Phone

The proposed satellites would orbit in five shells at altitudes between 510 and 580 kilometers (317 to 360 miles) above Earth, according to Amazon's D2D technical explainer. Three shells provide dense coverage over heavily populated mid-latitude regions; two higher-inclination shells extend service toward polar areas.

The satellites are not simple relay stations. Unlike "bent pipe" satellites that bounce a signal back to the nearest ground station, Amazon's D2D spacecraft will process signals onboard in orbit. Optical inter-satellite links — laser connections between satellites — let traffic route through the constellation before reaching a ground gateway, which reduces coverage gaps. On the phone-facing side, the satellites use digital beamforming to steer concentrated radio energy precisely where users are located, combined with beam-hopping (allocating transmission time slots across multiple beams), dual-polarization reception to capture more signal from a low-power handset, and adaptive modulation that adjusts data encoding in real time based on signal conditions.

This suite of techniques answers a physics challenge that is easy to understate: L-band and S-band frequencies offer good obstacle penetration and wide coverage, but they carry less bandwidth than the Ka-band spectrum used for home satellite broadband. That is why Amazon Leo D2D is positioned for voice, messaging, emergency services, and IoT — not gigabit streaming. The tradeoff is deliberate: voice and emergency connectivity in areas that have none is a different problem than high-speed internet, and it requires a different frequency tool.

The system is designed to work with any 4G or 5G smartphone that includes satellite-capable chipsets, a hardware category that has expanded rapidly across Apple, Google, and Qualcomm platforms as a result of the non-terrestrial network standards defined in 3GPP Release 17.

From the Start, Hundreds of Millions of Potential Users

Amazon's filing also confirmed an agreement with Apple to continue powering the satellite features on supported iPhone and Apple Watch models, including Emergency SOS, Messages, Find My, and Roadside Assistance via satellite. Globalstar currently provides these services under a multi-year Apple agreement — a relationship Amazon is inheriting rather than building from scratch.

That inheritance represents an immediate installed base that no other D2D operator can claim. Hundreds of millions of Apple devices already connect to Globalstar's network for emergency features; when Amazon takes over the constellation and spectrum, those users come with the deal.

On the carrier and partner side, Amazon has signed agreements with Vodafone, DirecTV, Herotel, and Australia's NBN — South Africa's Herotel and Australia's National Broadband Network among them — as early D2D partners.

Who Amazon Is Racing Against

Amazon is entering a market that has accelerated dramatically in the past 12 months.

SpaceX's Starlink Direct to Cell has been commercially available since July 2025, first as a text-and-emergency service for T-Mobile customers, with voice and data expanding from there. In January 2026, the FCC approved SpaceX's Gen2 expansion from 7,500 to 15,000 satellites, with FCC Chair Brendan Carr describing the approval as "transformative for facilitating next-generation services." SpaceX has also filed separately for authorization to operate up to one million satellites — a request Amazon has formally asked the FCC to reject.

The other significant competitor is AST SpaceMobile, a Texas-based startup that received FCC commercial authorization on April 21, 2026 for a 248-satellite D2D constellation providing broadband directly to smartphones. AST SpaceMobile uses 700 MHz and 800 MHz low-band spectrum licensed by its carrier partners — AT&T, Verizon, and FirstNet — an approach that operates under the same SCS framework as Starlink but stacks partnerships across multiple carriers rather than relying on a single operator.

In May 2026, AT&T, T-Mobile, and Verizon announced a joint venture to pool spectrum, develop common D2D integration standards, and jointly invest in satellite connectivity. All three carriers retain their existing individual satellite partnerships — T-Mobile with SpaceX, AT&T with AST SpaceMobile, Verizon with both AST SpaceMobile and Skylo — but the JV signals a defensive move to establish a common technical platform before any single satellite operator gains a structural lock on the carrier relationship layer.

A wild card emerged on June 29, 2026, when Rocket Lab agreed to acquire Iridium for $8 billion, the legacy LEO satellite phone operator. Amazon's FCC filing explicitly seeks access to frequencies in the 1,618.725–1,621.25 MHz range currently associated with Iridium, outside the United States. If Rocket Lab completes its Iridium acquisition — expected in mid-2027 — the two companies may find themselves in a spectrum conflict at the FCC before Amazon has launched a single D2D satellite.

Amazon's First Constellation Is Still a Work in Progress

The D2D application arrives while Amazon's broadband constellation remains in early deployment. As of July 2026, Amazon had approximately 396 broadband satellites in orbit — a fraction of the 3,232-satellite constellation it holds FCC authorization to build. The FCC granted Amazon a waiver of a July 30, 2026 deadline that would have required the company to have 1,618 satellites operational by that date; Amazon remains bound by a full-constellation deadline of July 30, 2029.

The reason for the shortfall is rocket availability, not manufacturing. A shared pool of launch vehicles — United Launch Alliance Atlas V, Arianespace Ariane 6, Blue Origin New Glenn, and SpaceX Falcon 9 — has not kept pace with production. Amazon says it plans to begin deploying D2D satellites in 2028, in parallel with continuing to build out the broadband constellation.

At full deployment, the combined Amazon satellite fleet — broadband constellation, D2D constellation, and the Globalstar satellites that Amazon would acquire through the merger — would total more than 8,300 satellites, making it the second-largest satellite operator in the world behind SpaceX, which currently operates more than 10,000 Starlink satellites.

The satellite-to-device market these operators are competing for is substantial. According to a March 2026 forecast from the research firm Omdia, global satellite D2D services are projected to reach 411 million monthly active users and approximately $11.99 billion in revenue by 2030 — representing annual growth rates of more than 80 percent for users and nearly 50 percent for revenue from 2026 levels.

Amazon's Partner Model vs. Starlink's Carrier Competition

One strategic difference that will shape the market is how each operator positions itself relative to mobile carriers.

SpaceX's Direct to Cell launched as a T-Mobile-exclusive arrangement in the United States, which gave it speed and regulatory momentum but created friction with AT&T and Verizon, who responded by accelerating their own satellite relationships with AST SpaceMobile. SpaceX is now pursuing international D2D expansion, including in Mexico, without a named MNO partner — a move that generates regulatory friction with local spectrum regulators accustomed to the SCS framework.

Amazon has explicitly framed the D2D service as complementary to terrestrial networks, not as a replacement. Its partner model — signing Vodafone, DirecTV, Herotel, and the Australian NBN — positions Amazon as a wholesale connectivity provider to carriers rather than a consumer brand competing with them. That framing is in direct tension with Amazon's long-term competitive interests: if Amazon owns the satellite layer, the device software stack, and the spectrum, mobile operators eventually become distribution partners rather than network gatekeepers. The current language is conciliatory; the infrastructure being built is not.

What This Means for the Satellite Phone Era

When Chris Weber, Amazon Leo's VP of Business and Product, wrote on X that "our plan is to start deploying in 2028, and we're excited to work with partners to make it happen," the brevity understated the complexity.

The FCC must approve the application. The Globalstar acquisition must close. A spectrum conflict with Rocket Lab must be resolved. And Amazon must accelerate its broadband deployment enough to show regulators it can handle two parallel constellation builds.

None of those are guaranteed, and every one of them has a multi-year timeline. But the strategic logic is clear: by anchoring D2D in purpose-allocated MSS spectrum rather than borrowed carrier frequencies, Amazon is building for a global market without a carrier partner required in every country — a structural advantage that, if the regulatory and acquisition timelines hold, could make the satellite phone era look very different in 2029 than it does today.

Frequently Asked Questions

How is Amazon Leo's satellite phone approach different from Starlink's?

Starlink Direct to Cell uses spectrum borrowed from T-Mobile under the FCC's Supplemental Coverage from Space framework, which means it needs a separate carrier partner in each country where it wants to transmit to phones. Amazon Leo plans to use Globalstar's Mobile Satellite Service spectrum — L-band and S-band frequencies that the ITU allocates specifically for satellite mobile communications. Those licenses cover more than 120 countries, which means Amazon could theoretically operate globally without replicating the carrier-by-carrier spectrum negotiation that constrains Starlink's international expansion.

When will Amazon Leo satellite phone service be available on my phone?

Amazon has stated that satellite deployment for the D2D constellation is set to begin in 2028, after the Globalstar acquisition closes in 2027 and the FCC reviews the current application. No consumer service launch date has been announced. For context, SpaceX's equivalent service took approximately two years from its first specialized satellite launches to commercial availability.

What does the Globalstar deal actually give Amazon beyond satellites?

The strategic value of Amazon's $11.57 billion Globalstar acquisition is not primarily the satellite hardware — it is the spectrum and the Apple relationship. Globalstar holds Mobile Satellite Service licenses covering more than 120 countries, which are what give Amazon the right to transmit directly to mobile devices internationally. Additionally, Globalstar already powers Apple's satellite features on iPhone and Apple Watch, including Emergency SOS, Messages via Satellite, and Roadside Assistance. Amazon inherits both the spectrum rights and that Apple partnership, giving it an immediate installed base of compatible devices and an existing enterprise customer in one of the world's largest consumer hardware companies.

Can Amazon Leo's D2D satellites deliver broadband internet to my phone?

Not at broadband speeds. The L-band and S-band frequencies Amazon will use for D2D are purpose-suited for voice calls, text messages, emergency communications, and low-bandwidth IoT applications — not high-speed data streaming. The physics of lower-frequency bands mean excellent signal coverage and obstacle penetration but less available bandwidth than the Ka-band frequencies Amazon uses for its fixed home broadband service. Think of D2D as connectivity where you currently have none, not as a replacement for your home internet connection.