The idea of building giant computer facilities in space has moved from the pages of science fiction into the boardrooms of some of the world’s biggest technology companies. But economists and engineers remain deeply divided over whether it can ever be made to work at a cost that justifies the enormous sums now being thrown at it.
The backdrop to all of this is the extraordinary stock market debut of Elon Musk’s rocket company, SpaceX, earlier this month. The listing raised $85.7 billion and valued the firm at several trillion dollars, instantly making Mr Musk the first person in history to be worth more than a trillion dollars on paper. But with that kind of valuation comes pressure. Some investors argue that for SpaceX to justify such a number, it must eventually make a success of something far more ambitious than launching satellites and resupplying the International Space Station. They believe the long game is this: clusters of computer servers, orbiting the Earth, running on sunlight, and doing the heavy lifting that is currently done by vast, power-hungry facilities on the ground.
“The company comes down to data centers in space,” Duncan Davidson, a partner at the venture capital firm Bullpen Capital, said recently on the American broadcaster CNBC, which first examined whether the economics of the idea hold up. “That is the big, long-term play.” Mr Davidson’s own firm has no direct stake in SpaceX, though it does have an indirect interest in a space start-up called Starcloud. He acknowledged that economically the idea is currently “marginal,” even as he argued the technical obstacles are gradually being overcome.
The logic behind the idea is straightforward enough on its face. On Earth, the facilities that house enormous stacks of computer servers require three things in vast quantities: land, water and electricity. Communities increasingly push back against them. Land in suitable locations is becoming harder to find. Water, needed in enormous volumes to cool the equipment, is a growing political flashpoint in many regions. Electricity demand from these facilities is set to rival the entire national consumption of a country like Japan by the end of this decade, according to the International Energy Agency. In orbit, none of those constraints apply. The sun shines without interruption. There are no neighbours to complain, no rivers to drain and no electricity grid to overload.
Mr Musk has said he expects space-based facilities to become more cost-effective than those on the ground within two to three years. He has also said that in orbit, adding more power becomes cheaper and easier over time, while on Earth the opposite is happening. In January, SpaceX filed a proposal with American regulators for a constellation of up to one million satellites that would form the backbone of exactly such a system. The satellites involved, referred to internally as AI1, would carry far more processing hardware than the existing Starlink communications satellites and would require staggering quantities of computer chips. To meet that demand, SpaceX, Tesla and the chipmaker Intel have joined forces to build a manufacturing facility in Austin, Texas, covering ten million square feet and potentially costing up to $119 billion to complete. It is due to open in 2029.
Not everyone shares Mr Musk’s confidence in his own timeline. Jeff Bezos, whose rocket company Blue Origin is pursuing its own version of the same idea, told CNBC he considers space-based facilities “very realistic” in principle, but that a two-to-three-year window is “a little ambitious.” His company plans to launch more than five thousand satellites by late 2027 as part of a constellation called TeraWave, designed partly with this kind of work in mind.
Mark Weinzierl, an economist at Harvard University who has spent years studying the finances of space-based businesses, puts the central problem plainly. “One of the biggest questions is, are you sure that we can’t just do that cheaper on Earth?” he said. Having reviewed the business models of companies currently working in this space, he found none that could demonstrate genuine cost competitiveness with ground-based alternatives right now. That said, he does see a point in the future where the two lines on the graph might cross. Building and running facilities on Earth will, he believes, get steadily more expensive as land, water and power become more contested. Building in space, meanwhile, should get cheaper as rockets become more reusable and launch costs fall. At around five hundred dollars per kilogram to space, according to one industry estimate, space-based computing begins to match what it costs on the ground. Below that, it becomes cheaper.
There are real technical hurdles too, beyond the money. Computer hardware in orbit faces radiation, extreme swings between heat and cold, and a vacuum that makes cooling it an entirely different engineering challenge from what is done on Earth. A current SpaceX design for one of its compute satellites is estimated to be between one hundred and one thousand times less powerful than the servers sitting in today’s ground-based facilities. Anything requiring a fast response from users like financial transactions, interactive services also suffers from the time it takes signals to travel between the ground and space and back again. The first workloads that make sense to run in space, researchers suggest, may be those that are not time-sensitive and are already connected to space operations, such as processing data gathered by Earth observation satellites.
Whether SpaceX can prove the doubters wrong, and on something close to Mr Musk’s timetable, remains the defining question for investors who have put big faith in a company that has yet to show what an orbital computer facility, at commercial scale, actually looks like when it is switched on.
About the Author
Marcel Chidozie is a tech analyst and writer covering foreign news, fintech, and emerging technologies at TechRegard. Based in Nigeria, He's passionate about translating complex tech developments into compelling, accessible stories for diverse audiences. His work focuses on how technology shapes innovation across Africa and globally.
