We’re running out of space…
The growth of cloud computing, streaming platforms, social media, and online storage has created an enormous demand for data processing and storage capacity. Technologies like AI, the Internet of Things (IoT), and big data analytics also require powerful infrastructure to handle vast amounts of information in real time. In addition, businesses and governments are increasingly digitising their operations, further driving demand and use. As more people and devices connect to the internet worldwide, companies continue building new data centres to ensure fast, reliable, and secure access to digital services.
At the same time, there has been exponential growth in the number of satellites being launched into space for global mobile communications, earth observations, emergency response, crop monitoring etc. These services require those satellites to possess data processing functionality to send information through the satellite network and back to Earth to be processed at terrestrial data centres on the ground.
The physical resources required for continuing to deploy cloud computing infrastructure on earth is significant. Vast centres to house, power and maintaining such technologies is, in some locales vying for priority against other, very human needs such as housing or agriculture. As such, our reliance upon cloud-based tech is becoming rapidly unsustainable.
Becoming virtually sustainable
Beyond physical space, the rapid growth of data centres has significant environmental and social implications too. Environmentally, they consume vast amounts of electricity, often relying on non-renewable energy sources, which contributes to greenhouse gas emissions and climate change. They also require large volumes of water for cooling, putting pressure on local water resources, especially in already dry regions.
Socially, while data centres can create jobs and support digital economies, they may also strain local infrastructure, raise land use concerns and provide relatively few long-term employment opportunities compared to their physical footprint.
Quite simply our digital world is swelling beyond what our physical world can hope to cope with. Finding a means to balancing technological growth with sustainable practices and targets is becoming increasingly important. And demand for both is continuing to rise.
To address the challenge, the solution may lie in where we house our next datacentres – not on the ground but instead up in the sky.
Shall we then build datacentres in Space to cater to the growing data processing demand on earth and Space?
Demonstrations show AI inference and training in space is technically feasible, but so far is still small-scale. Start-ups like Starcloud have already launched graphics processing unit (GPU)-equipped satellites and even run AI models in orbit for exactly this purpose. The company has, in its short lifespan, has been valued at over $1bn, Major players are also now exploring projects including SpaceX, by proposing massive constellations of compute satellites, Google’s “Project Suncatcher” which is testing satellites, and Blue Origin’s plans for large solar-powered satellite constellations. The thinking is that space-based data centres will have access to a significant amount of solar energy, and their operation will not require the consumption of earthly resources.
Weighing up the benefits and pitfalls
In this context, it is important to understand the concept of latency. Latency refers to the time delay (measured in milliseconds) between when a data packet is sent from a user’s device and when it receives a response from the network. It is often referred to as “ping time” or round-trip time (RTT), representing the total journey from the user, up to the satellite, down to a ground station, and back again.
As satellite signals must travel to outer space and back—a distance of tens of thousands of miles—latency is significantly higher in satellite systems than in terrestrial fibre or cable connections. Research from the Bells University of Technology shows that in the absence of the proposed space-based cloud computing architecture, the transmission of data to the ground station via a geostationary communications satellite increases the latency by 44.8% on average. When satellites in the low earth orbit communicate with assets in the geostationary orbit, the use of space-based data centres instead of ground stations reduces the latency by 45.7% on average. Not only are space-based datacentres can prove to be useful from a sustainability perpective, they can potentially provide a user and service advantage too.
However, space-based data centres face several major challenges that limit their practicality. One of the biggest barriers is the extremely high launch cost – not all start-ups will secure the investment Starcloud has secured, and not all major players will want to invest the sums required for their own projects until ROI can be fully demonstrated.
Such data centres also need a large number of solar panels which will have its own carbon footprint of production and logistics. They would also require data equipment as well as infrastructure to protect, power, and cool them, thereby adding weight and complexity. Not to mention cost. Cooling the data centres will be problematic as conventional cooling systems will not work in zero gravity. So some ingenuity and innovation are still required.
And then there’s ensuring their upkeep. Maintenance and upgrades are difficult, as hardware failures cannot easily be repaired, raising the risk of systems becoming effectively non-functional as they age. The harsh environment of space – characterised by radiation, microgravity, and debris – can damage electronics, resulting in shorter lifespans, and, as a recent report in the BBC shares, would require specialised equipment to fix. In-orbit servicing capabilities need to be devised. In addition, latency limitations mean signals from orbit can be slower, making these systems better suited to non-real-time tasks.
Environmental trade-offs further complicate the issue, since launching the data centres into orbit will produce significant levels of emissions.
Finally, economic uncertainty remains high, as cheaper Earth-based solutions—such as renewable energy, nuclear power, and improved cooling technologies—may ultimately prove more viable, especially if demand does not justify the immense costs.
One small step for man
The near-term options may be Space based data centres used for interlinking satellites and satellite data processing, and supporting earth-based data centres. These could specialise in services that relate directly to Earth’s continuous demand for Space based communications providing supportive services for spacecraft and other space facilities, through more efficient space-to-space data transfers. Not only does this provide a direct industry usage but provides a test bed for refining the tech and logistics required to make the next step in their adoption easier to reach.
But progress will be slow. As it stands, there are likely to only be a few such data centres because of the high launch, operating and maintenance costs attached. As a result, their economic and environmental viability does not yet seem promising. Significant technical challenges also need to be overcome.
But what is clear is that we need to find a solution soon. Whilst earth-based data-centres are reliable, accessible and less costly, their expansion and requirements are putting the Earth under a near-unbearable weight.
Where better than space to alleviate the strain?
Dr. Atanu Chaudhuri
Dr. Atanu Chaudhuri is a Professor of Operations and Technology Management and Director of Business Partnership and Engagement for the Department of Management and Marketing at Durham University Business School, UK. He is also the Co-Director of Durham University’s Multi-disciplinary Space Research Centre.
