Amazon didn't buy Globalstar for satellites: it bought a seat at the network table

Is a company that sells you everything from batteries to cloud servers really in the satellite business? After this week, the answer is yes, and it is paying real money to prove it. Amazon's purchase of Globalstar looks like a space deal. We think it is a spectrum deal, and spectrum is a far more durable kind of moat.

Amazon cleared the FCC, but the satellite count still favors Starlink

The Federal Communications Commission approved Amazon Leo's $11.5 billion acquisition of Globalstar on Tuesday. The deal was announced in April at roughly $11.57 billion. Amazon adds Globalstar's two dozen satellites and its mobile-satellite spectrum to a network of more than 200 satellites of its own. Telecompaper reports that the companies expect to close in 2027.

Amazon Leo has launched more than 300 satellites since April 2025 and wants about 3,200 in low Earth orbit by 2029. Starlink, by the same reporting, already has more than 11,000. On raw constellation size, then, Amazon is far behind.

The prize is the airwaves, not the hardware

Satellites can be built and launched by anyone with enough money and patience. Licensed spectrum in internationally harmonized bands is much harder to conjure up. The FCC approval covers about 25 megahertz of Globalstar's mobile-satellite service spectrum, and the agency says the combined assets will boost Amazon Leo's entry into the direct-to-device market, where smartphones and other devices connect directly to satellites.

That is why the price makes more sense than the satellite count suggests. In our reading, Amazon is not paying for 24 satellites. It is paying to skip years of regulatory queueing and to start the direct-to-device race with spectrum in hand. It is a shortcut, and we think a smart one. Amazon has also committed to invest more than $10 billion in Leo, so $11.57 billion is roughly the same size of bet again.

Apple is the quiet kingmaker

The most interesting detail is not about Amazon or Starlink. Globalstar powers Emergency SOS and Find My on iPhone and Apple Watch, and according to the FCC, Apple accounted for about 63% of Globalstar's revenue last year. Apple also holds a 20% stake after its roughly $1.5 billion investment in 2024. Amazon says it has signed an agreement with Apple to keep providing those services, and plans to acquire Apple's stake.

Read that carefully. Amazon isn't just buying a network. It is buying a very large anchor customer. In our view, that is a stronger moat than any satellite, because it comes with distribution. Whether Apple is happy to rely on Amazon rather than own a piece of the operator is an open question, and we would not assume the answer is yes.

Does this really pressure Starlink?

Somewhat, but let's not oversell it. Starlink has a huge head start, and the FCC has already authorized SpaceX to build a 15,000-satellite direct-to-device constellation. Amazon's own commercial rollout is only planned for the end of 2026, the FCC said. Competition is welcome, and FCC Chairman Brendan Carr framed the approval that way, but a rival with more than 200 satellites does not dethrone a leader with 11,000 overnight.

The bigger question is who controls the layer between your phone and the sky. If it ends up as two giants, SpaceX and Amazon, regulators will have to ask whether that is real competition or a duopoly with better marketing. The FCC's approval found no material public-interest harms. We would keep watching that claim as the market matures, especially with more D2D spectrum coming: the agency is set to vote this week on proposals to open additional bands.

Our verdict

The broader thesis holds: Big Tech no longer wants to rent the pipes. Amazon already runs a huge cloud business, and it now wants a say in how devices reach it from places no tower covers. For consumers, the near-term effect is modest and probably positive: more rivals, more coverage, a sturdier Apple safety feature. Our advice is to judge this deal in 2027, when it is expected to close, not by the headline number today. If Amazon turns spectrum into a service people actually use, this will look cheap. If not, it will be a very expensive license.