This week, Amazon (AMZN) announced it will buy satellite service provider Globalstar (GSAT) – the company behind Apple‘s (AAPL) iPhone Emergency SOS feature and a key player in satellite spectrum – for close to $11.6 billion.
Alongside the acquisition, Amazon announced a long-term agreement to power satellite features on future iPhone and Apple Watch models – including Emergency SOS, Messages, Find My, and roadside assistance.
The obvious takeaway is that Amazon aims to become a major player in the direct-to-device satellite internet market.
But the more interesting insight — the one that we think matters more for investors — is that Amazon just acquired the infrastructure prerequisite layer for running data centers in space. And it did so before SpaceX went public and before Musk got tens of billions of dollars to start launching his own orbital compute nodes at scale.
A new tech titan has entered the orbital compute arena.
Amazon just fired the starting gun. The race is on – and so is the clock for investors.
Amazon’s Globalstar Acquisition: The Obvious Play
Amazon gets three things from Globalstar that money alone can’t easily buy:
- The ability for the Amazon Leo network to connect directly to smartphones on the ground via satellite internet, without the need for terrestrial cell towers.
- A moat of already-licensed wireless spectrum – a scarce, tightly regulated asset that will, , give “Amazon the ability to launch satellites that connect directly to devices sooner, rather than having to go through its own country-by-country approval process.”
- Positioning within the Apple franchise: Amazon inherits Globalstar’s role as the backbone of iPhone and Apple Watch Emergency SOS, with a new long-term agreement to power Apple’s satellite services going forward.
Taken together, Amazon didn’t just buy a satellite company. It bought a credible seat at a table that Musk has been setting alone.
Satellite connectivity and satellite spectrum are a massive, multi-hundred-billion-dollar growth market over the next few years, and Amazon needed this deal to remain credible in that race. Starlink already has more than 10,000 satellites in LEO, serves paying subscribers across more than 150 countries and territories, and a growing government contract pipeline – including multibillion-dollar agreements with the U.S. Department of Defense and Space Force to provide secure satellite communications and battlefield connectivity.
Amazon was running behind. It needed to leapfrog, not catch up incrementally.
Globalstar was the leapfrog.
The Real Story: Orbital Compute Infrastructure
But that’s not the part of the story that matters most.
The world’s largest cloud providers have been quietly moving toward the same long-term thesis: the next wave of AI compute doesn’t live in a ground-based data center. It lives in space.
Satellites that don’t just route packets but process them. Edge nodes in space, running inference closer to the data source, operating outside the jurisdiction of any single government, consuming zero terrestrial land, and immune to most of the physical constraints that make building data centers on Earth increasingly painful.
That’s the orbital compute thesis. And to execute it at scale, you need three things:
- a dense, trusted LEO constellation
- licensed spectrum to route the traffic
- an anchor customer base large enough to justify the economics
Globalstar just handed Amazon all three alongside a single check.
Of course, Amazon isn’t framing this publicly as an orbital compute play. The SEC filing talks about “accelerating a new era of Global Direct-to-Device Connectivity.”
But Amazon didn’t announce AWS as a cloud computing empire either. It filed it under “web services” and built it quietly for years before anyone understood what was happening.
And what the company is building now – a massive, licensed, Apple-anchored LEO constellation under its own control – is exactly the platform you’d need if you planned to run the world’s first commercial orbital data center network.
Meanwhile, SpaceX is preparing to go public, confidentially filing for an IPO targeting a $1.75 trillion valuation. When it does, Musk will have billions in new capital and every incentive to start launching compute nodes into the Starlink constellation at scale.
The race for orbital compute is coming whether Wall Street has named it yet or not. Amazon just made sure it shows up to that race with the right hardware.
The Bigger Picture: Where the Money Actually Goes
The commercial space economy – the one that’s been “five years away” for as long as we can remember – is finally coming into focus.
Amazon just paid $11.6 billion for satellite spectrum and LEO infrastructure. Microsoft has Azure Space partnerships with Starlink baked into enterprise cloud contracts. Google is investing in satellite data pipelines. Apple is building satellite features into every iPhone. And SpaceX, which started as a rocket company, is now one of the most strategically valuable infrastructure businesses on Earth – and soon in orbit.
The trillion-dollar question is, who are the picks-and-shovels suppliers for all of this?
Because here’s the thing about the space economy: it’s not one market but several overlapping markets, each with its own investment cycle. And those in the earliest stages are the most interesting precisely because they’re the most overlooked.
The Picks-and-Shovels Layer In Space
The compute owners – Amazon, Microsoft, Google, Meta (META) – will capture enormous value. But they’re already worth trillions. The real alpha is in the suppliers:
- Rockets
- Orbital network operators
- Space-grade semiconductors and radiation-hardened chips
- Power systems for satellites
- And lasers – specifically, the inter-satellite laser link manufacturers who build the data routing layer without which orbital compute networks can’t function at speed
In fact, this last bucket may be the single most underappreciated opportunity in the entire space stack. Without high-throughput laser inter-links, an orbital data center is inoperable.
Starlink already knows this, which is why it’s deploying laser inter-satellite links at scale. The rest of the industry will follow. The companies building those lasers are worth watching very carefully.
The Bottom Line
We view Amazon’s acquisition of Globalstar as a declaration that the commercial space era has begun – and that the world’s largest companies are now competing for orbital infrastructure the same way they did for cloud infrastructure 20 years ago.
Space connectivity and spectrum are a massive growth market over the next two to three years. And orbital compute is a massive growth market over the next five to 10.
Two stories, one conclusion: the space economy is the biggest emerging investment theme of this decade.
There’s a pattern to how this plays out.
The infrastructure gets funded first. It has to – nothing else works without it.
But over time, the center of gravity shifts – not to the hardware, but to the layer that decides how a technology actually gets used.
That shift isn’t always obvious in the moment. For a while, it’s easy to miss. Then it becomes difficult to ignore.
Cloud followed that pattern. So did mobile.
There’s a strong case that AI is doing the same thing now.
Because while Amazon and SpaceX are racing to control where compute lives, a different question is starting to matter more: Who controls what runs on top of it?
Right now, one company sits closer to that position than anything else.
OpenAI.
The challenge is that it’s still private. Which means most investors will only encounter it later, when access is easy and expectations are already set.
We’ve spent some time looking at how that transition could play out – and what it means to .
.
