The FCC signed off on Amazon’s purchase of Globalstar on Tuesday. The Space Bureau found no public-interest harms, waved away the single objection (a contract fight that’s already in arbitration) and said the deal puts a well-financed operator behind phone-to-satellite service. Amazon agreed to pay $90 a share back in April, about $11.6 billion, and still expects to close in 2027.
Read the price the way you’d read any acquisition. What is Amazon getting that it couldn’t build itself? It isn’t the satellites. Globalstar flies 24 of them, an old and small fleet next to the thousands Amazon Leo plans to launch. It isn’t the revenue either; Globalstar is a niche business selling messaging, tracking and IoT links plus wholesale capacity. The thing Amazon can’t make in a satellite factory is a licence.
Globalstar holds about 25 MHz of mobile-satellite spectrum in the US: 1610 to 1618.725 MHz for the uplink and 2483.5 to 2500 MHz for the downlink. Both sit in bands that ordinary phone radios can reach. That’s the asset. The constellation, the ground stations and the customer contracts come along with it.
Why owned spectrum matters for direct-to-device
Satellite-to-phone service has two ways in. You can borrow a mobile carrier’s terrestrial spectrum and beam it down from orbit, which is how SpaceX’s Starlink Direct to Cell works with T-Mobile. Or you can own spectrum that’s licensed for satellite use in the first place. The first route is cheaper to start, though it runs on the carrier’s terms and on a non-interference basis, country by country. The second route gives you protection from interference and a service you can sell without a carrier partner if you want to.
SpaceX worked this out too. In September 2025 it agreed to buy EchoStar’s AWS-4 and H-block spectrum for roughly $17 billion, then got FCC permission for up to 15,000 direct-to-device satellites. Amazon’s Globalstar deal is the matching move. After it closes, the US will have two companies holding licensed satellite spectrum that phones can use, and both have a billionaire founder with a rocket company standing behind them. Everyone else in the market will be a tenant of one of them or of a carrier.
The FCC was fairly open about why it didn’t mind. It said the two companies don’t compete head to head today, since Globalstar does narrowband mobile service and Amazon Leo does fixed broadband, and it pointed to a “high likelihood” that Amazon would actually use the spectrum hard. Spectrum that sits idle is the regulator’s old worry. A buyer with Amazon’s capital budget answers it.
Apple is the quiet party in this deal
Globalstar’s biggest customer is Apple. Emergency SOS via satellite on every iPhone since the 14, and on the Apple Watch Ultra 3, runs over Globalstar, and Apple put about $1.5 billion into the company over the years, ending up with roughly a fifth of the equity. When Amazon announced the deal it also signed Apple up as a customer for satellite links on current and future iPhones and Watches.
So part of the $11.6 billion buys a contract with the one handset maker that sells hundreds of millions of devices a year with satellite messaging already switched on. That’s distribution. Starlink got its phone users through T-Mobile; Amazon gets them through Apple, and the Apple route doesn’t depend on any one carrier’s subscriber base. For Apple it’s a hedge too. Its satellite supplier goes from a small company that needed Apple’s cash to keep launching, to a company that can fund a 5,105-satellite direct-to-device constellation, which is what Amazon has asked the FCC to approve.
There’s a real security angle here as well. Satellite links to plain handsets matter in disasters, in war zones and in places where governments cut the ground network. Starlink in Ukraine since 2022 showed how fast a commercial constellation becomes part of somebody’s military plan, and how much leverage the owner gets. Two American companies controlling the licensed spectrum for this service, globally ambitious and both close to Washington, is a fact that regulators in Brussels, Delhi and Beijing will price into their own landing-rights decisions. Globalstar’s licences in other countries don’t move automatically with an FCC order, and international review was always expected to be slower.
What is still open
Approval of the transfer isn’t approval of the new constellation. Amazon’s application for the 5,105-satellite system is still pending, and the comment cycle closed only on October 5. Amazon is also behind on its own broadband buildout; it missed the FCC milestone of having 1,616 Leo satellites in orbit by July and is working under a waiver that runs to October 2027. Globalstar, meanwhile, got authorization earlier this month for its C-3 system of 48 satellites being built by MDA Space, which will carry the service until Amazon’s own direct-to-device fleet goes up. Amazon is aiming to start that service in 2028.
If you want the short version of what the FCC approved, it’s a spectrum sale with a satellite company attached. The same pattern showed up in last week’s LSEG and Octus roundup and in our piece on what an acquisition price tells you about a data business. Buyers pay most for the thing that’s slowest to rebuild. In satellite-to-phone, nothing is slower to rebuild than a licence the regulator won’t hand out again.
Amazon paid for the right to be one of two. That was the point.