The satellite race becomes a spectrum race

AMAZON'S $11.57 billion acquisition of Globalstar, announced on April 14th, is easy to cast as catch-up. Elon Musk's Starlink has roughly 10,000 satellites in orbit and over nine million users; Amazon's Leo business has 241 satellites and no paying customers. Counting satellites, the headline version of the deal is that an e-commerce behemoth has bought itself a running start against a well-entrenched rival. The real story is that counting satellites no longer tells you who is winning.

Globalstar is a 34-year-old mobile-satellite-service operator with 24 working spacecraft, a handful of ground gateways, and a portfolio of spectrum rights in at least eleven countries. Chief among them is Band 53 (and its 5G variant, n53), a swath of 2.4-GHz mid-band spectrum that Globalstar holds on an exclusive, coast-to-coast basis in the United States and on a globally harmonized basis elsewhere. According to the company, it is the only such mid-band not owned by an existing wireless operator. Apple paid $1.5 billion in 2024 for a 20% stake in the firm, largely to secure capacity for its Emergency SOS feature. Under the new terms, Apple will exit the cap table at $90 a share and become a customer of Amazon's, with a new deal covering current and future iPhone and Apple Watch products.

Band on the run

But look at what the hyperscalers have actually been buying. In September, SpaceX agreed to pay EchoStar roughly $17 billion for its AWS-4 and H-block licenses, a deal it followed in November with a further $2.6 billion for the remaining AWS-3 uplink spectrum. SpaceX also committed $2 billion to cover EchoStar's debt interest through 2027. Add Amazon's $11.57 billion for Globalstar, and the two U.S. direct-to-device entrants have written roughly $31 billion in checks for spectrum rights in seven months. They have repriced an entire asset class. Spectrum has become the most expensive thing in the LEO stack.