Orbital said its long-term vision is to create a distributed network of space-based data centres capable of providing AI computing power from orbit, using the unique advantages of the space environment, including continuous solar energy availability and passive heat rejection into space.
The company’s proposed constellation would represent a dramatic expansion of commercial space infrastructure, moving beyond traditional communications and Earth observation satellites towards a new category: orbital computing platforms designed specifically for artificial intelligence workloads.
Orbital argued that the rapid growth of artificial intelligence is increasingly constrained by terrestrial infrastructure, particularly electricity generation, grid capacity, land availability and cooling requirements for large-scale data centres.
The company’s solution is to relocate portions of AI computing infrastructure into low-Earth orbit, where satellites can access uninterrupted solar energy and use the vacuum of space as a natural thermal environment.
Unlike conventional data centres built around large facilities on Earth, Orbital’s approach relies on a distributed architecture made up of many smaller satellites operating as individual compute nodes. The company said this design allows the system to scale progressively, with additional satellites increasing overall computing capacity.
Orbital’s development roadmap begins with a technology demonstration mission planned for 2027, known as Pathfinder, which will test AI computing hardware in orbit aboard a hosted payload. The mission is intended to validate key technologies, including radiation tolerance, thermal management and data transmission from space.
This will be followed by Orbital-1, a purpose-built satellite designed to demonstrate the company’s first dedicated orbital AI computing platform.
The company said production satellites are being designed around approximately 100 kilowatts of computing capability per spacecraft, with the ultimate goal of deploying a constellation exceeding 100,000 satellites capable of delivering more than 10 gigawatts of orbital computing power.
Orbital’s proposal reflects a broader shift in the commercial space sector, where falling launch costs, reusable rockets and mass satellite manufacturing are enabling concepts that would have previously been considered impractical.
The company is developing its own satellite manufacturing capability, including a planned production facility in Los Angeles to support high-volume spacecraft construction.
Orbital said its systems will focus primarily on AI inference, the process of running trained artificial intelligence models, rather than the massive data-processing requirements associated with training the largest AI models. The company argued that inference workloads are better suited to distributed satellite architectures because individual tasks can be processed independently across many nodes.
While Orbital’s vision represents a potential transformation of the space and artificial intelligence industries, deploying such a large constellation would require overcoming major engineering, regulatory and economic challenges.
A 100,000-satellite network would require unprecedented launch capacity, automated spacecraft production, long-term orbital management and robust systems for avoiding congestion in increasingly busy low-Earth orbit environments.
The company’s plans come as the space sector enters an era of increasingly large commercial constellations, with operators exploring new applications ranging from global connectivity to Earth intelligence and now, space-based computing.
Orbital believes that moving computing infrastructure beyond Earth could become a critical component of the next generation of artificial intelligence systems.
The company’s founders argued that space offers a fundamentally different infrastructure model, one where energy generation and cooling limitations faced by terrestrial data centres can be reduced through orbital solar power and the natural environment of space.
If successful, the project would mark a significant milestone in the commercialisation of space, shifting satellites from being primarily communications and sensing platforms into becoming active computing infrastructure supporting the global AI economy.
Want to see more stories from trusted news sources?Make Space Connect a preferred news source on Google.