Data Centre In Space (Part 1): Why The AI Industry Wants To Take The Cloud Into Orbit
As AI drives data centre growth, companies are exploring orbital computing while experts debate its technical feasibility, environmental benefits and commercial viability.
Published : July 29, 2026 at 11:40 AM IST
There are more than 10,700 data centres worldwide, powering the internet and running everything from web searches and email to video streaming and artificial intelligence (AI) chatbots. Nearly 4,000 more facilities are currently under construction, and a good chunk of this expansion can be credited to the AI race that keeps demanding an increase in computing power.
The surge has resulted in an exponential increase in demand for GPUs, memory chips, and other components to the point of straining global supply chains and driving up manufacturing costs for smartphones, laptops and other devices that rely on many of the same components.
Up until this point, there have been more than enough reports highlighting the strain that data centres place on the planet. These facilities gobble up vast amounts of water, consume enormous quantities of electricity, and generate significant greenhouse gas emissions, setting tech giants back on their net-zero and carbon-neutral targets. The problem is only expected to worsen as the rapid expansion of data centre capacity places even greater pressure on natural resources and energy infrastructure. Click here to read more.
Humans living near these data centres often bear the brunt of their environmental impact, which is why new projects are increasingly facing resistance from local communities. The growing opposition in parts of the US is one of the factors pushing big tech companies to look beyond traditional locations for their AI infrastructure. However, shifting a data centre outside one country does not eliminate the problem since—breaking news—all countries exist on the same planet. But what if the data centre is moved beyond Earth altogether, or even submerged beneath the ocean?
The industry has started to look for data centre solutions, which would hopefully lessen the burden on the power grid and water table. One of those ideas is to build data centres in space, where sunlight can provide unlimited solar energy and the freezing vacuum of space can provide infinite ‘radiative’ cooling. The orbital data centre concept is already moving beyond theory.
Starcloud has launched a satellite carrying an NVIDIA H100 GPU to test AI workloads in orbit, while Axiom Space is evaluating orbital edge computing aboard the ISS. Google and Planet Labs plan to launch prototype satellites in 2027 to test space-based data centre technologies. Ramon.Space is developing space-hardened processors for in-orbit data processing. In India, TakeMe2Space and Agnikul (+ NeevCloud) aim to launch India's first AI data centre prototype to demonstrate edge computing and radiative cooling in space.
Discussing the orbital data centre concept with ETV Bharat, Ronak Kumar Samantray, founder and CEO of Hyderabad-based Indian space tech company TakeMe2Space, says, “The scepticism is understandable, because an orbital data centre cannot be thought of as a terrestrial data centre simply placed in space. The science, economics, and use cases all have to be evaluated differently.”
However, the entire thing is facing scepticism about its feasibility due to the operational challenges, with some critics going so far as to call it nothing more than science fiction. Google itself, in its 18-page paper, acknowledges four key hurdles that must be addressed before orbital data centres become practical: thermal management, on-orbit reliability and maintenance, high-bandwidth space-to-ground communications, and viable launch economics.
Contrary to popular belief, cooling is not "free" in space. In the vacuum of space, heat cannot be carried away by air or water and must instead be dissipated through thermal radiation. Likewise, repairing or replacing faulty hardware in orbit is far more complex than in terrestrial data centres, while maintaining high-bandwidth ground communication remains another major technical challenge.
According to Abhishek Agarwal, President, Judge India Solutions, space- and underwater-based data centres may make for compelling headlines, but neither is likely to solve the industry's challenges this decade. Instead, he advocates practical solutions that rely on existing technologies and can be deployed today, without waiting for major technological breakthroughs.
Manoj Dhanda, Founder and CEO, Utho Cloud, shares a similar view, describing orbital data centres as a distraction from the more pressing work that needs to be done on Earth. He argues that expanding renewable energy use, improving operational efficiency, and choosing locations with abundant clean energy would have a far greater impact on reducing the environmental burden of data centres.
Ronak Kumar Samantray of TakeMe2Space also discusses these challenges and explains how TakeMe2Space plans to overcome them. Additionally, he proposes a different set of initial use cases for orbital data centres, suggesting they should function as specialised in-orbit computing platforms for latency-sensitive, data-intensive space applications rather than as direct replacements for terrestrial data centres.
Now that we have set the stage for orbital data centres, the next parts of this four-part series will dig deep into the concept. The second part discusses the engineering challenges of orbital data centres, the bottlenecks that still await solutions, and whether the economics of orbital computing make sense.
| READ ABOUT OTHER ASPECTS OF ORBITAL DATA CENTRES |
| [You are here] 👉 Data Centre In Space (Part 1): Why The AI Industry Wants To Take The Cloud Into Orbit |
| Data Centre In Space (Part 2): The Engineering Bottlenecks & Economics Behind Orbital Computing |
| Data Centre In Space (Part 3): Why The Real Solution To AI Infra Load May Still Be On Earth |
| Data Centre In Space (Part 4): Why India's TakeMe2Space Is Betting On Orbital Computing |
