Google and SpaceX are discussing plans to place data centers in orbit, potentially solving cooling and power challenges while raising questions about space debris, costs, and practicality. The proposal represents an ambitious but unproven approach to hyperscale computing infrastructure - and I'm skeptical until I see working prototypes.
According to TechCrunch, the two companies are exploring orbital infrastructure as a solution to the growing computational demands of AI workloads. The pitch is that space offers unique advantages: abundant solar power, vacuum cooling, and freedom from terrestrial power grid constraints.
The physics might actually work. Space is cold (in shadow), solar power is available 24/7 without weather interruptions, and the vacuum environment eliminates atmospheric cooling challenges that plague ground-based data centers. Latency to ground stations via laser links could theoretically be competitive with transcontinental fiber. On paper, it's not crazy.
But the economics are a different story. Launch costs have dropped dramatically thanks to SpaceX's reusable rockets, but they're still far higher than building a data center in Iowa. A single Falcon 9 launch costs around $67 million and can carry about 22,800 kg to low Earth orbit. Compare that to the cost of shipping server racks via truck - essentially free per kilogram.
Then there's maintenance. When a server dies in a ground data center, a technician swaps it out in minutes. When a server dies in orbit, you need a service mission or you write off that compute capacity. Hardware lifespans in space are typically shorter due to radiation exposure, meaning higher replacement rates. Unless you're planning regular Starship missions with maintenance crews, orbital data centers would have much higher total cost of ownership.
The space debris concerns are real too. A single data center satellite failure could create a debris cloud that threatens other satellites in the same orbital shell. The more hardware you put in orbit, the higher the collision risk. SpaceX's Starlink constellation already faces criticism for contributing to orbital congestion - adding massive data center satellites would magnify those concerns.
