Data centers are emerging as a new environmental flashpoint for the tech industry. Public concern is growing over their hunger for power and thirst for water. What’s the tech industry’s reaction to the problem? Put the data centers in space.
Space is a great place for data centers because it solves one of the biggest problems with locating data centers on Earth: power, argues Google’s Senior Director of Paradigms of Intelligence, Travis Beals.
“The Sun is the ultimate energy source in our solar system, emitting more power than 100 trillion times humanity’s total electricity production,” he explained in a company blog. “In the right orbit, a solar panel can be up to eight times more productive than on Earth, and produce power nearly continuously, reducing the need for batteries.”
“In the future, space may be the best place to scale AI compute,” he asserted. “Working backwards from there, our new research moonshot, Project Suncatcher, envisions compact constellations of solar-powered satellites, carrying Google TPUs [Tensor Processing Units] and connected by free-space optical links. This approach would have tremendous potential for scale and also minimizes impact on terrestrial resources.”
SpaceX is also on board with the idea of data centers in space. Last month, it filed a request with the Federal Communications Commission to launch a constellation of up to one million solar-powered satellites that it said will serve as data centers for artificial intelligence.
The filing maintains that its satellite scheme is “the most efficient way to meet the accelerating demand for AI computing power” and frames them as “a first step towards becoming a Kardashev II-level civilization — one that can harness the Sun’s full power” while also “ensuring humanity’s multi-planetary future amongst the stars.”
The Kardashev Scale is a framework proposed in 1964 by Soviet astronomer Nikolai Kardashev to classify civilizations by their energy consumption. A level I civilization uses all the energy on its home planet; a level II civilization uses all the energy output of its star; and a level III civilization uses all the energy of its galaxy.
Unlimited Power
Industry analysts say the energy argument is driving much of the early momentum behind orbital data center proposals.
“Space offers essentially unlimited access to power at marginal costs approaching zero dollars,” asserted Gordon Bell, a principal for strategy and execution at EY-Parthenon, in Boston, the global strategy consulting arm of Ernst & Young.
“Space offers a potential solution to the largest and most persistent bottleneck currently facing the data center industry,” Bell told TechNewsWorld. “The data center industry is plagued by power constraints that are not going to be resolved any time soon.”
He explained that construction of new generation and transmission assets are multi-year endeavors for a multitude of reasons, including misaligned incentives between utilities and the data center industry, supply chain, regulatory and labor constraints, NIMBYism, and more.
Providing energy to support the operations of the IT and cooling equipment within a data center is also the largest operating expense, he added.
“Most space-based data centers are envisioned to be placed in sun-synchronous orbit, facing the sun 24/7,” he noted. “This would allow for the generation of solar power at very high efficiency, as well as a reduced need for any cooling or back-up equipment.”
“Traditional cooling equipment would be mostly eliminated, as it is very cold in space and heat would be ejected into space via radiators,” he continued. “Back-up equipment could be removed, as the power generation source — the sun — would always be shining and generating energy.”
Orbital Promise vs Practical Risks
Ellis Scherer, the policy analyst for broadband, spectrum, and space policy at the Information Technology & Innovation Foundation, a research and public policy organization in Washington, D.C., agreed that orbital data centers could address some of the environmental and power grid concerns of building more data centers on Earth.
“Data centers in space can access solar power 24/7 in certain ‘sun-synchronous’ orbits, giving them all the power they need to operate without putting immense strain on power grids here on Earth,” Scherer told TechNewsWorld. “This would alleviate concerns about consumers having to bear the costs of higher energy use.”
“There is also less risk of running out of real estate in space, no complex permitting requirements, and no community pushback to new data centers being built in people’s backyards,” he added.
However, he pointed out that there are challenges to building orbital data centers, especially at the scale proposed by SpaceX in its FCC filing.
A million-satellite data center could greatly increase the risk of collisions with other objects, he observed.
“A constellation of one million satellites is magnitudes larger than anything that currently exists,” he said. “In December 2025, China made headlines by proposing two new constellations, totaling 200,000 satellites, which is the largest constellation filing in ITU [International Telecommunication Union] history.”
“Space is a harsh and unpredictable environment, and if something goes wrong with a data center, it would be nearly impossible to fix it once it’s in orbit,” he continued.
“Finally,” he added, “the costs associated with building and launching data centers into space at the pace necessary to keep stride with the AI boom may be prohibitive.”
Hardware Costs Could Outweigh Energy Gains
Andrew Yoon, a software engineer with CivAI, a non-profit organization focused on public education about AI capabilities and risks, acknowledged that data centers are causing a short-term energy supply crunch, causing energy prices to go up where data centers are operating, but he maintained that energy is not the primary cost driver of data centers.
“By some estimates, energy and land costs are only around 25% of the total cost for a data center,” Yoon told TechNewsWorld. “AI hardware is the real cost driver, and shifting to space only makes that hardware more expensive.”
“Hardware cannot be repaired or upgraded at scale in space,” he explained. “Maintaining satellites is extremely hard, especially if you have hundreds of thousands of them. Maintaining a traditional data center is extremely easy.”
He added that for data centers in space to be feasible, they have to be cheaper to build in space than on Earth. “If you look at the math right now, it’s vastly more expensive to do anything in space for a variety of reasons than it is to do it on Earth,” he said.
“If you assume that the cost of putting stuff into space continues to drop, maybe at some point in the future, the cost equation becomes a little more competitive,” he continued. “But this requires making quite a few assumptions that are, in my opinion, quite optimistic and unlikely to pan out anytime soon.”
In the most optimistic case, it will take 10 years to put a data center in space, he noted. “That means you have to assume that those energy supply crunches will continue to exist for the next 10 years,” he said. “I just don’t think that’s a very realistic scenario.”
Need for Data Processing in Space
While skeptical of floating terrestrial-style data centers in space, Jim Dunstan, general counsel for TechFreedom, a technology advocacy group in Washington, D.C., argued that there’s a critical need for data processing in space.
“About 85% of all the data that is collected by Earth remote sensing satellites is dumped and is never downlinked,” he told TechNewsWorld. “They haven’t got the bandwidth to downlink it, nor do they have the onboard processing power to throw AI and other stuff at the data, to triage it and determine what’s worth keeping.”
“What they end up doing is geofencing the data and saying, okay, let’s dump everything from out over the ocean because it’s just the ocean, even though there could be some really interesting stuff in there,” he explained.
“So it makes a lot of sense to have data centers to do the data mining and the crunching on all of this remote sensing data that is essentially being dumped into the ether because it’s not downlinked,” he added.
Alvin Nguyen, a senior analyst for infrastructure outsourcing, data center services, and semiconductor research at Forrester Research, a multinational market research company headquartered in Cambridge, Mass., agreed.
“There will be a need for computation in space in the future to support exploration, mining, and other purposes,” he told TechNewsWorld. “Having computation near data originating in space will provide real advantages, especially in reducing the latency of having data processed back on Earth and provide more effective use of bandwidth in space networks.”
Hybrid Landscape
Early forecasts suggest orbital computing will evolve gradually rather than replace terrestrial data centers overnight.
The data center in space scene will initially be hybrid, predicted Mark N. Vena, president and principal analyst at SmartTech Research, a technology advisory firm in Las Vegas.
“The first killer apps will be in orbit workloads like satellite data processing and space network backhaul, while general-purpose cloud stays mostly terrestrial until costs and debris governance improve,” he told TechNewsWorld.
Vena added that small prototypes show up first, not a full hyperscale clone in the sky. “Google has discussed prototype work around 2027, and that fits a realistic arc where early orbital compute appears in the late 2020s, with the first commercially meaningful clusters more likely in the early 2030s,” he said.
Starcloud, a space infrastructure startup, told Time magazine that by the early 2030s, it envisions having a four-km by four-km (2.5-mi. by 2.5-mi.) mega-structure in orbit composed of hundreds of individual satellites docked to a truss and a massive array of solar wings. That one facility, it predicted, could replace about 50 data centers on Earth.
As society’s appetite for data expands exponentially, fueled by artificial intelligence, it may want to look beyond Earth’s atmosphere for its data center needs, suggested Rob Enderle, president and principal analyst at the Enderle Group, an advisory services firm in Bend, Ore.
“I think it would be far better to place them on the moon, if you can address the data latency problem,” he told TechNewsWorld. “You’d have a far more efficient way to dump the heat, and the location would be far less risky both for the data center and for Earth should the data center fall or break up.”