Yes, another space post. I didn't even write this post this week. I'm on vacation, the news broke last Tuesday, and I've apparently become that guy.
Amazon is paying $11.57 billion for Globalstar. The headline framing is "shot at Starlink." The actual framing is simpler: Amazon paid twelve billion dollars for 24 satellites and some radio licenses, and the radio licenses are what cost twelve billion dollars.
Globalstar has roughly 25 MHz of licensed mobile satellite spectrum - 8.725 MHz in the L-band and 16.5 MHz in the S-band, authorized in 120+ countries. L-band in particular penetrates weather, works direct-to-device without specialized hardware, and took Globalstar three decades of regulatory grinding to assemble. You can launch satellites. You cannot launch spectrum. Amazon is also facing an FCC deadline to have roughly 1,600 Kuiper/Leo satellites in orbit by July 30, 2026 - and as of early April had 212. Acquiring someone else's licensed orbital footprint buys time against a regulator, which is a very expensive but very effective way to keep your authorization.
This connects directly to what I wrote about the FCC's "Weird Space Stuff" NPRM and satellite market-access reciprocity proceeding. The regulatory surface is where the next decade of satellite competition gets decided, not the launch manifest. Amazon just proved it by writing the check.
SpaceX has every reason to oppose the transfer. Iridium, whose L-band sits adjacent to Globalstar's, hit a 52-week high on the announcement. DoJ antitrust plus foreign direct investment review across 120+ jurisdictions is not a 2026 close. The deal is expected to close in 2027, and that assumes nobody successfully objects in the middle.
Read it as an infrastructure move, not a constellation race. The satellite race isn't about altitude. It's about which high-band licenses you can light up, in how many countries, before your competitors get there first.