Orbital Data Centers Seen Eclipsing Ground Capacity
Musk projects orbital AI computing will surpass ground-based capacity by 2029, while Google flags launch costs as a major obstacle.
Updated on Nov. 2, 2026 in Data Centers
Elon Musk forecast that AI and cloud computing capacity deployed in near-Earth orbit will exceed ground-based capacity by 2029. Google said launching computing hardware into orbit could cost as much as 100 times the cost of building equivalent capacity on the ground.
Why it matters
The competing forecasts put the economics of space-based computing at the center of the race to meet surging AI demand. Morgan Stanley's projections of rapid global data-center growth frame how much capacity the industry will need to build either on the ground or in orbit.
Morgan Stanley projects global data-center IT capacity will climb from about 136 GW at the end of 2026 to 316 GW in 2030, a 23% compound annual growth rate adding roughly 180 GW. Google's estimate that orbital launches could cost up to 100 times ground construction remains unexplained in its methodology.
The players
Elon Musk
The billionaire entrepreneur behind SpaceX, Tesla and xAI, who has repeatedly argued that space offers a path to scaling AI computing.
Alphabet's technology company, which operates one of the world's largest networks of ground-based data centers and has assessed the economics of orbital computing.
Morgan Stanley
The global investment bank whose research projects rapid growth in worldwide data-center IT capacity through 2030.
The details
Musk's forecast compares capacity deployed in orbit against ground-based capacity, with orbital AI and cloud computing expected to overtake terrestrial capacity by 2029. Google's estimate frames the central cost barrier as the expense of launching hardware into orbit relative to building the same capacity on Earth.
Timeline
End of 2026: global data-center IT capacity at roughly 136 GW, per Morgan Stanley.
2029: Musk forecasts orbital AI and cloud capacity exceeding ground-based capacity.
2030: Morgan Stanley projects global data-center capacity reaching 316 GW.
Big Picture
Morgan Stanley's projection that global data-center IT capacity reaches 316 GW by 2030 sets the ground-based baseline that orbital computing must either supplement or surpass. Musk's 2029 forecast and Google's launch-cost caveat test whether space can absorb a meaningful share of that expansion.
For cloud customers and AI developers, the cost and location of computing capacity shape the price and availability of hosted services over the coming years. Ground-based expansion remains the primary path for near-term supply, while orbital options remain largely speculative given the launch-cost estimates cited.
The takeaway
Capacity forecasts for AI infrastructure vary widely depending on whether orbit is included, so readers should weigh launch economics alongside projected growth rates. Ground data-center expansion is the more tangible baseline for now.
Further reading
Explore more on Data Centers to see how ground-based capacity is expanding alongside the orbital proposals.
Source note: This article includes information reported by TokenPost.
