Bezos’ Blue Origin Wants 51,600 “Data Center” Satellites, A Direct Shot at SpaceX’s Next Frontier

Jeff Bezos’ Blue Origin is pitching a moonshot that would put a chunk of the world’s computing power in orbit, literally. The company has filed paperwork with the Federal Communications Commission seeking approval for “Project Sunrise,” a plan to launch up to 51,600 satellites designed to run computing workloads in space.

The sales pitch is blunt: AI is driving a data-center boom that’s colliding with real-world limits, electricity, water for cooling, land, and local opposition. Blue Origin says moving compute off-planet could ease the strain, using near-constant solar power in orbit instead of pulling from crowded power grids on the ground.

The filing doesn’t prove Blue Origin can pull it off. But it does put the company formally into a fast-forming race, one where SpaceX, startups, and even Big Tech are floating plans to build “orbital cloud” infrastructure above Earth.

A major FCC filing signals Blue Origin is serious

Project Sunrise isn’t a glossy concept video. Blue Origin went to the FCC, the U.S. agency that regulates satellite communications, to request authority to operate a massive constellation and use Ka-band spectrum for telemetry, tracking, and control.

In the documents, Blue Origin describes Sunrise as an orbital computing platform made up of tens of thousands of satellites. The company hasn’t publicly detailed how much processing power each satellite would carry or what specific workloads it would target first, but the regulatory move is a key step: you don’t ask for this kind of orbital real estate unless you’re trying to build something real.

Blue Origin says the satellites would fly in sun-synchronous circular orbits, paths that keep spacecraft in consistent sunlight, at altitudes ranging from about 311 to 1,118 miles above Earth (500 to 1,800 kilometers). The company argues that steady access to solar energy is central to the concept.

The “space data center” pitch: less power-grid pain, less water for cooling

Blue Origin is leaning hard on a problem Americans are already seeing play out from Northern Virginia’s “Data Center Alley” to fast-growing markets in Texas and the Southwest: data centers are hungry, and communities are pushing back. AI training and inference are pushing power demand higher, and cooling systems can consume enormous amounts of water, especially in hot, dry regions.

Sunrise reframes the bottleneck. Instead of fighting for grid hookups and water rights, Blue Origin argues, orbital compute could run on solar power and avoid many of the land-use and local-permitting battles that slow projects on Earth.

Bezos has also framed the idea as a long game. Speaking at a tech conference in Italy, he called orbital data centers a “next step” in moving heavy industry into space, and suggested space-based compute could beat the cost of Earth-based data centers “in the coming decades.” That timeline matters: this is positioned as an industrial bet, not a quick revenue play.

TeraWave: the networking layer meant to make Sunrise work

Project Sunrise is tied to another Blue Origin concept called TeraWave, a separate constellation the company has pegged at 5,408 satellites. The idea is that TeraWave would provide the high-throughput connectivity, between satellites, to ground stations, and to customers, that Sunrise would need to function as a distributed computing system.

Blue Origin has pointed to optical links, laser-based connections, as part of that architecture. That’s a logical choice if you’re trying to move large volumes of data quickly in space, where traditional radio links can become a constraint.

The company’s target customers sound familiar: large enterprises, government agencies, and defense-related users. It’s a playbook the satellite industry knows well, build a network that supports your own flagship service while also selling connectivity to outside customers to help justify the cost.

New Glenn is the linchpin, if it can deliver

Launching 51,600 satellites would require an assembly-line approach to space. Blue Origin is pointing to its heavy-lift rocket, New Glenn, as the engine that could make that scale possible. The company has said New Glenn’s first stage is designed to be reusable for at least 25 flights.

This is where the SpaceX comparison becomes unavoidable. SpaceX didn’t just build a rocket; it built a machine for launching frequently and cheaply, then used that advantage to deploy Starlink at scale. Blue Origin is signaling it wants a similar vertically integrated model, control the launch cadence to support a sprawling orbital services business.

But a powerful rocket isn’t the whole equation. To make Sunrise real, Blue Origin would need high-rate satellite manufacturing, streamlined integration and testing, and a reliable operations pipeline, plus a deployment schedule it hasn’t shared publicly.

A crowded field: SpaceX, a Seattle-area startup, and even Google

Blue Origin is jumping into a category where the numbers are already getting absurd. SpaceX has floated an orbital data-center concept that could involve “up to” 1 million satellites. Against that, Sunrise’s 51,600 satellites almost sounds restrained, until you remember that it would still be one of the largest space systems ever attempted.

There are smaller players, too. Starcloud, a startup based in Redmond, Washington (across Lake Washington from Seattle), has been cited as pursuing a similar approach, with plans that have circulated around roughly 60,000 satellites.

And Big Tech is sniffing around the idea. Google has been linked to a concept called “Project Suncatcher,” with reported plans for demonstrations through a partner, Planet Labs. The broader mix, space companies, cloud giants, and chipmakers, signals that orbital compute is being treated less like sci-fi and more like a strategic option.

The hard part: congestion, spectrum fights, and what “compute in orbit” is actually good for

Even if space-based computing helps dodge some Earth-bound constraints, it creates new ones. Low Earth orbit is getting more crowded, radio spectrum coordination is becoming more contentious, and regulators are under pressure to address debris and long-term sustainability.

There’s also a basic technical question Blue Origin hasn’t answered publicly: what workloads make sense to run hundreds of miles up, and how do you move data back and forth efficiently enough for customers to care? Latency, bandwidth, and the cost of launching and maintaining hardware in orbit could erase the advantages of “free” solar power for many use cases.

Still, the FCC filing makes one thing clear: Blue Origin wants a seat at the table in the next infrastructure fight, one where the competition isn’t just about rockets anymore, but about who controls the computing backbone of the AI era, whether it’s on the ground or circling overhead.