On the surface, the idea of putting data centers in space, partly in a bid to reduce carbon emissions, among other objectives, sounds somewhat far-fetched.

After all, blasting a rocket into space isn’t exactly an environmentally friendly process. It takes more than 400,000 liters of rocket fuel to get SpaceX’s Falcon 9 into space, for example, and getting an engineer out to fix something isn’t quite as straightforward as sending them to Slough or downtown Dallas.

Then there’s the issue of power, whereby the entire facility will need to be powered by solar. Moreover, with geosynchronous orbit some 22,200 miles away, latency will also be palpable while all communications will be subject to atmospheric interference.

That’s not all, there’s also a greater risk of electromagnetic interference from solar flares, not to mention a high-speed collision with the increasing amount of space junk cluttering up Earth’s orbit.

And, finally, there’s the risk that at some point it will fall to Earth.

None of these issues are news to Yves Durand, director of technology at Thales Alenia Space, one of the world’s biggest manufacturers of satellites.

But as part of a 16-month European Union-funded project, called Ascend [Advanced Space Cloud for European Net zero emission and Data sovereignty], the task for Durand and his team is to work out the feasibility of running data centers in space, and whether they could be consistent with the EU’s Green Deal plan to make the continent carbon neutral by 2050.

“In essence, it’s a feasibility study to try to see whether putting some data center capacity in space might help reduce the carbon footprint of data centers. [We’re asking] whether putting them in space and capturing the energy from space, and using the natural cooling of space, could help?” Durand said during a recent broadcast on DCD’s Edge Computing Channel.

So, while it might take a gargantuan quantity of rocket fuel to get it into orbit, there could well be plenty of power and cooling savings that might make it a worthwhile consideration for particular data center applications – and it’s Durand’s job to find out with the €2 million ($2.1m) project, part of the EU’s Horizon Europe science research program.

“Space is becoming increasingly important. We capture lots of data, for example, of the environment, satellites taking pictures of the planet, and we also have to observe the Earth for potential environmental problems, such as fires or Earthquakes, and have to react quickly.

“So to have processing and data storage of all that in space makes a lot of sense. It’s an ambitious project, but will give us a perspective on what we should do in terms of manufacturing servers and electronics that can withstand the environment in space. And having data centers in space should allow us to integrate space assets in a more global cloud environment.”

We have already gone from standalone satellites to constellations, which “can now communicate with each other with inter-satellite links and laser communication – it’s an information network in the sky,” says Durand.

The next step, coming soon, is the installation of a small server on the International Space Station (ISS) as a joint experiment that also brings Microsoft into the fold. Microsoft will implement it, integrate into the cloud, and will have cameras set up so specialists on the ground can see what they can do with this capacity in space – without having to go up to the ISS in person.

That follows previous tests by HPE and Skycorp, which both sent hardware to the ISS to test out what the impact of radiation has on hardware. Companies like JSAT and NTT are also evaluating putting data centers in space.

But what about setting it up and maintenance?

According to Durand, this is where robotics will come in. In addition to having Ariane Group on board, the biggest satellite launch company in Europe, and Airbus, Germany’s Institute of Robotics and Mechatronics (DLR) is also part of the team, examining the feasibility of using robots to set up the data center once it has been parked in orbit, and to provide ongoing maintenance.

So what about the issue of power – both the 400,000-liter question of getting there in the first place, as well as whether it’s truly possible to run a data center of any significant capacity on solar power alone?

Durand admits that the biggest issue is simply working out whether it makes sense. It may be technically feasible and the research may be valuable, but is it possible to do in terms of the EU’s net-zero ambitions?

“By 2050, it’s likely that we will not have unlimited renewable energy capacity. So the first challenge is to analyze how data center needs will evolve because there’s an explosion in the use of data centers, and with people starting to use ChatGPT instead of Google as a search engine for example, the need for data center processing will be huge,” says Durand.

So data center energy demands will continue to increase and renewable power sources on Earth may struggle to keep up.

“When I started the project, I thought it would be a simple lifecycle analysis, but it’s become very complex and involved several types of analysis. It also involves energy prediction, which is huge. But we’re talking to more and more experts in the field who are interested in the project to make sure that it makes sense.

“This might even come as a great push for both the space industry and the technology industry in Europe,” says Durand.

In terms of sustainability, a large launch rocket like the SpaceX Falcon 9 – or the 2050 equivalent – might be able to bring economies of scale to bear, not just in terms of price, but also usage of rocket fuel.

Moreover, by 2050, there may well be a few more space stations orbiting Earth, not to mention Moon bases and possibly even a base on Mars, if SpaceX CEO Elon Musk gets his way.

And each and every one of them will require data center facilities to support them as we expand beyond our planet.