Elon Musk is no longer flirting with the idea of orbital data centers — he’s all in.
When SpaceX quietly filed documents with the FCC last Friday proposing a million-satellite data center network, many assumed it was another Musk-style moonshot. One week later, it’s clear this isn’t speculation — it’s strategy.
The biggest signal came Monday with the formal merger of SpaceX and xAI, bringing Musk’s space and artificial intelligence ambitions under one roof. The move makes far more sense if orbital data centers are the connective tissue between rockets and AI models.
FCC Approval Signals Momentum
On Wednesday, the Federal Communications Commission accepted SpaceX’s filing and opened it for public comment. While this is usually routine, FCC Chairman Brendan Carr took the unusual step of publicly sharing the proposal on X — a move that suggests political tailwinds for Musk’s plan.
As long as Musk remains in favor with the current administration, industry watchers believe regulatory approval for orbital AI infrastructure will move swiftly.
Why Elon Musk Thinks AI Belongs in Space
Musk recently expanded on his vision during an appearance on Stripe co-founder Patrick Collison’s podcast Cheeky Pint, alongside guest Dwarkesh Patel.
According to Musk, the economics of orbital data centers hinge on energy efficiency.
“It’s harder to scale on the ground than it is to scale in space,” Musk explained.
“Any given solar panel gives you about five times more power in space than on Earth.”
In Musk’s view, abundant solar energy could dramatically reduce one of the highest operating costs of AI data centers — electricity.
The Skepticism: Power Isn’t Everything
Not everyone on the podcast was convinced.
While solar efficiency in orbit is real, Patel pointed out that power is just one piece of the cost puzzle. Cooling systems, maintenance, hardware failures, and GPU servicing during AI training remain major unanswered challenges for orbital data centers.
Fixing a faulty GPU on Earth is difficult enough. Doing it hundreds of kilometers above the planet? That’s a different problem entirely.
2028: The Turning Point for Orbital Data Centers?
Despite the skepticism, Musk remains confident — and specific.
He marked 2028 as the tipping point year when orbital data centers become economically superior, leading infrastructure for AI workloads.
“In about 30 to 36 months, the most economically compelling place to put AI will be space,” Musk said.
He went even further, predicting that within five years, more AI computing capacity will be launched into orbit annually than the total currently operating on Earth.
Why This Matters for the Global Tech Industry
By 2030, global data center capacity is expected to exceed 200 gigawatts, representing nearly $1 trillion in infrastructure investment. If even a fraction of that shifts toward orbital data centers, the ripple effects could reshape AI economics, cloud computing, and space commercialization.
And conveniently, SpaceX makes its money launching payloads into orbit — now with an AI company attached.
With a SpaceX–xAI IPO reportedly just months away, expect orbital data centers to dominate Musk’s narrative — and investor decks — throughout the year.
One thing is certain: the future of AI infrastructure may not stay Earth-bound for long.
Don’t miss out on our latest news—follow us for the latest AI news, breakthroughs, and insights that matter.