Artificial intelligence (AI) has created a new kind of infrastructure race. It’s no longer just about who has the best chips, models or software. Increasingly, it’s about who can secure enough power, land and cooling to keep the whole machine running.
That’s why one of the more futuristic solutions is starting to attract serious attention – data centres in space.
At first glance, it sounds like science fiction. But the logic isn’t as wild as it appears. Data centres on Earth are expensive, power hungry and often unpopular with local communities. Space is a different starting point. Solar power is abundant, land constraints disappear, and there aren’t any local planning objections in low Earth orbit.
The catch is obvious. Getting the kit into space is also expensive, maintaining it is difficult, and the economics still have a long way to go. But the technical case is becoming harder to dismiss. Smaller compute satellites, connected in a network, could be more realistic than launching a single giant space-based data centre.
That makes this an early-stage theme to watch, rather than a near-term replacement for the data centres being built on Earth today.
This article isn’t personal advice. If you’re not sure an investment is right for you, seek advice. Investments and any income from them will rise and fall in value, so you could get back less than you invest.
Why space is getting a serious look
The big attraction is power.
In orbit, satellites can draw on solar energy far more directly than most sites on Earth. Some orbital paths can keep satellites in sunlight pretty much 24/7, which is helpful for workloads that need steady power. For AI, where electricity has become one of the biggest constraints on growth, that’s a powerful idea.
Cooling is another key piece of the puzzle. On Earth, data centres use air or liquid cooling to stop chips from overheating. In space, there’s no air, so heat has to be radiated away. That sounds like a major hurdle, and it is, but it doesn’t look impossible. A smaller satellite with a sensible amount of onboard compute may be able to manage heat through its body and added radiators.
The most realistic version isn’t a single huge facility floating above the Earth. It’s a cluster of smaller satellites, each carrying chips, solar panels, communications equipment and cooling systems. Those satellites could be linked together, giving the overall network more resilience if one unit fails.
The cost problem has not gone away
The biggest issue isn’t whether a small data-centre satellite can be built. It’s whether it can make financial sense.
Launch costs are the main barrier. A satellite may get access to solar power once it’s in orbit, but it still has to be built, launched, operated and eventually replaced. At today’s launch costs, the economics look challenging. The value of the electricity saved on Earth may not yet be enough to offset the cost of putting the hardware into orbit.
That could change if two things happen.
First, launch costs fall sharply. Reusable rockets are already pushing the industry in that direction, but a much lower cost per kilogram would be needed to make orbital data centres broadly competitive. Second, the hardware becomes more power-dense, meaning more computing power per tonne launched.
There’s also an issue of scale. A serious orbital computing network would need thousands of satellites. That’s possible in theory, but it would be a huge logistical and capital challenge.
Space solves some problems, but it creates others
Hardware in orbit has to cope with radiation, temperature swings and limited access for repairs. If something breaks, an engineer can’t simply walk into the building and replace a component. That puts a premium on reliability, redundancy and careful design.
Data movement is another issue. Some workloads may be well-suited to space, especially where the data is already generated in orbit. Others may be less practical if large volumes of information need to move between Earth and satellite networks. Latency, bandwidth and communications costs all matter.
There are also environmental and regulatory questions. Low Earth orbit is already busy, and adding large numbers of compute satellites could increase congestion and collision risk. Rocket launches have their own environmental footprint, too, so the sustainability case isn’t as simple as saying solar power in space is free and clean.
The full picture depends on launch emissions, satellite lifetimes, replacement cycles and what happens when satellites re-enter the atmosphere.
The companies moving first
The listed company angle isn’t straightforward, since some of the most important names are private.
SpaceX is impossible to avoid. It dominates commercial rocket launches and has been vocal about the potential for orbital AI compute. Cheaper, more frequent launches are among the biggest requirements, so it’s a key enabler for others, even if, for now, it remains a private company.
Amazon is one of the more obvious established listed names to watch through Project Kuiper (now Amazon Leo), its planned low Earth orbit broadband network designed to challenge SpaceX’s Starlink. Alphabet’s Project Suncatcher is exploring the idea of solar-powered satellites carrying its own AI chips, with early prototypes planned.
That does not mean either project will be dial movers anytime soon, but it does show some of the largest AI companies are taking the concept seriously.
More broadly, the potential benefits would not stop with the companies launching satellites. The entire chip supply chain could feel the lift. More data centres mean more demand for advanced processors, memory, networking chips, power management, packaging, testing equipment and the manufacturing tools needed to make it all. In that sense, orbital compute would be less about one winner taking all and more about another wave of demand rolling through the semiconductor ecosystem.
The investment takeaway
Data centres in space aren’t about to replace the facilities being built on Earth. The economics are still uncertain, the engineering is complex, and the capital requirements could be huge.
But the idea is no longer just a futuristic talking point. Falling launch costs, more efficient chips and rising pressure on power grids are making orbital compute more credible. For investors, the best way to think about the theme is as another layer in the AI infrastructure story.
The opportunity is exciting, but it’s very early. The stars may be the next frontier for AI infrastructure, but investors should keep both feet firmly on the ground.
This article is original Hargreaves Lansdown content, published by Hargreaves Lansdown. It was correct as at the date of publication, and our views may have changed since then. Unless otherwise stated estimates, including prospective yields, are a consensus of analyst forecasts provided by LSEG. These estimates are not a reliable indicator of future performance. Past performance is not a guide to the future. Investments rise and fall in value so investors could make a loss. Yields are variable and not guaranteed.
This article is not advice or a recommendation to buy, sell or hold any investment. No view is given on the present or future value or price of any investment, and investors should form their own view on any proposed investment.