Orbital data centers powered by solar arrays in space could help ease AI's crushing demand for electricity, but the economics remain deeply uncertain and deployment timelines stretch years into the future. While prototype satellites carrying single AI accelerator chips already orbit Earth, full-scale constellations of thousands or even millions of satellites remain largely theoretical, with costs currently running roughly $51.1 billion for a 1-gigawatt facility compared to $15.9 billion for an equivalent ground-based data center.

Why Are Tech Companies Considering Putting Data Centers in Space?

The appeal is straightforward: terrestrial data centers face mounting obstacles. Global data center electricity consumption surged 17 percent in 2025 to 485 terawatt-hours, while AI-focused facilities consumed 50 percent more power than the previous year, according to the International Energy Agency. The IEA projects consumption will roughly double to 950 terawatt-hours by 2030. Building new ground-based facilities has become contentious, with communities opposing projects and regulators imposing lengthy permitting processes that can stretch years.

Satellites in sun-synchronous orbits, by contrast, receive near-continuous sunlight and can generate power via solar arrays without competing for grid capacity or local water supplies. They also bypass the multiyear local permitting process that now delays many terrestrial builds. As a result, investment in orbital compute is accelerating despite the underlying economics remaining unresolved.

Which Major Companies Are Already Building Orbital Data Centers?

Several companies with substantial capital backing have moved from concept to hardware deployment:

• Blue Origin: Amazon founder Jeff Bezos' space venture filed with the Federal Communications Commission in April 2026 for "Project Sunrise," a constellation of up to 51,600 satellites to operate as a data center in space in sun-synchronous orbit at altitudes between 500 and 1,800 kilometers.
• Google: The company announced Project Suncatcher in November 2025, calling it a "research moonshot" similar to its work on autonomous vehicles and quantum computing. Google plans to launch two prototype satellites in early 2027 to test Tensor Processing Units on satellites flying in tight formation.
• SpaceX: The company filed with the FCC in January 2026 for a constellation of up to one million satellites for orbital AI compute, targeting deployment "as early as 2028" and working toward a long-term goal of 100 gigawatts of orbital compute capacity per year.
• Lonestar Data Holdings: The company has already launched four data centers into space, including two to the moon, with its highest-profile mission reaching the lunar surface in February 2025 aboard Intuitive Machines' Athena lander.

What Are the Main Cost and Technical Barriers?

Cost remains the primary obstacle. According to an interactive cost model maintained by Andrew McCalip, head of research and development at Varda Space Industries, orbital power currently costs about $14,700 per kilowatt per year versus $570 to $3,000 for a U.S.-based data center. However, if launch costs fall to $200 per kilogram, a level researchers call plausible by the mid-2030s, orbital power costs would drop to around $810 per kilowatt per year, bringing them within the range of terrestrial data center power costs.

Beyond cost, deployment speed may ultimately matter more than absolute price. "SpaceX does not even need to be able to deploy cheaper than on Earth. It just needs to be faster," said Patrick Bowen, CEO of optical processing unit developer Neurophos. With generator lead times now approaching a decade, speed itself has become the binding constraint. "We're nearing the point where even if GPUs were free, you can't turn them on," Bowen explained.

Technical challenges are equally daunting. Hardware must survive launch vibration, radiation exposure, temperature swings, and long periods with limited physical access. Once a satellite is in orbit, hardware failures cannot be addressed. "That raises the bar for reliability, maintenance and lifecycle planning, because swapping out a CPU, GPU, or SSD is much harder when you can't walk over to the server," noted Sebastien Jean, Chief Technology Officer at NAND storage provider Phison Electronics.

How to Evaluate Orbital Data Centers as a Viable Solution

• Cost Trajectory: Monitor whether launch costs decline toward the $200-per-kilogram threshold that would make orbital power economically competitive with terrestrial facilities within the next five to ten years.
• Chip Obsolescence Risk: Consider whether satellite lifecycles can keep pace with AI chip development cycles, which move far faster than the typical lifespan of orbital hardware, creating potential stranding of expensive infrastructure.
• Data Movement Constraints: Assess whether bulk data movement to and from orbital infrastructure can be solved at scale, as this remains one of the hardest technical and logistical challenges for space-based compute.
• Regulatory and Orbital Congestion: Track whether orbital traffic management can accommodate the thousands or millions of satellites proposed by major companies, given that more than 17,000 satellites already orbit Earth.

Daniel Thorpe, director and head of EMEA data center research at JLL, a commercial real estate agency, summarized the challenge: "Relocating general AI workloads into space depends on a long list of factors that haven't fallen into place yet: launch costs, the inability to service hardware once it's up there and chips becoming obsolete faster than the satellites carrying them".

Despite these obstacles, the sheer scale of AI's power demands is driving serious investment. SpaceX's regulatory filing lists the orbital compute initiative among its principal risk factors, cautioning it "may not achieve commercial viability". Yet the company and its competitors continue to pursue the vision, betting that the convergence of cheaper launches, improved satellite reliability, and desperate terrestrial power constraints will eventually make space-based data centers not just viable, but necessary.