Why does it matter where we build data centres?

Yes, demand for them is growing at pace, driven by more and more data creation; demand for capacity is expected to increase at a 23% compound annual growth rate through 2030. But there is method to where they’re built, and why. Not least in terms of connectivity, climate impact, political stability, and proximity to innovation hubs.

There are planning regulations to think about too, of course. In the UK, the new Labour government is working to streamline the process for building data centres by removing planning barriers and encouraging construction on brownfield sites near cities. It’s a push that’s largely driven by the rapid expansion of AI technologies, and the government’s ambition to use them to spur innovation and economic growth.

However, the environmental costs of AI are another a big part of the equation. It requires vast computing power and data storage, which can have significant ecological consequences, particularly in terms of energy demand. In many regions, the energy grids powering these data centres still rely heavily on fossil fuels due to limited access to renewable energy sources.

So, as demand for cutting-edge data centres increases, it prompts a crucial question: why not locate them in areas with abundant renewable energy or in less congested regions?

Put simply, not all data centres are the same. Several geographic and locational factors are important to consider, including the type of data centre, which holds significant influence. For example, hyperscale facilities that support AI deployments or cloud services differ from data centres that are purely used for data storage.

Network connectivity is also paramount; data centres need to be where large data networks exist, such as in Slough in the UK or Virginia in the US. The reason being that these areas host extensive cloud services and offer dense network connectivity, which ensures redundancy (i.e. a backup system should a primary component fail), reliability, and low latency; all of which are critical for data centre operations.

For the most part, it’s driven by customers’ demands and requirements for their data. Whether that’s data centre capacity to support their AI deployments or simple data storage, or where they need to access their data. We then build facilities in locations where our customers want to expand. That’s a big driver behind the growth we’re seeing in secondary markets, in fact. While traditional hubs in the West continue to grow, there's increasing investment in the likes of Milan, Madrid, and Berlin, driven by local demand for digital services and infrastructure.

There are myriad different factors which impact the choice of location, not least the availability of affordable and reliable power, which is essential for operating data centres efficiently, as the demands of AI and other data-intensive applications increase energy requirements.

The investment required for building data centres is substantial too, and many regions lack the network infrastructure to support these investments, which is why investment continues to be highest in the West. But while established hubs in the West remain dominant due to their infrastructure and connectivity, as mentioned, new markets are becoming increasingly attractive for data centre investments, particularly those with thriving economies like Japan.

Ultimately, a good metric comparison for data centre growth is a country’s GDP. If that’s growing, then it’s likely data centre construction is too.

AI adds another dimension that’s also shifting the data centre landscape in the short, medium and long term.

AI models, particularly large language models like ChatGPT, require significant computing power - both for model training, and to be applied in real-time (known as “inference”). Right now, we’re mostly at the training stage, which doesn’t typically require low-latency connectivity. This means data centres located further outside of metropolitan areas, like the Nordics, are becoming attractive for training AI models due to their abundance of this cheap, reliable power as well as their skilled workforces.

However, when these models reach the application stage (inference), proximity becomes more important; these models need to be hosted in data centres that are in dense, metropolitan areas like London, to improve connectivity and reduce latency.

In the medium term, as more of these models reach maturity and move to the inferencing stage, investment into data centres in existing hubs is likely to grow substantially, pulling the focus back to major economic hubs, especially hotspots like Virginia, Dallas, Frankfurt, London, and Paris.

Digital adoption is essential for economic growth. AI is already proving that it will drive innovation and productivity across industries, so if a region lacks access to AI, its businesses will miss out big time. It’s not just about consumer use cases like ChatGPT, but vital tools that drive business growth and efficiency.

Think about it: without AI, companies can't effectively harness the cloud services or advanced data analytics that enable them to innovate, improve their customer experiences, or stay competitive. The result? Economic stagnation and a widening gap between regions with robust digital infrastructure and those without.

In today's world, every business is a digital business. The digital economy is also the backbone of modern growth, and regions without access to AI are falling off the map. So yes, data centres are needed to support AI deployment, and what’s more, the location of those data centres, does matter. AI isn’t just the future—it’s the key to staying competitive. Missing out on it means missing out on success.

That’s why the location of data centres plays such a crucial role to the digital economy. Proximity to major business centres is essential for real-time AI applications; having the right infrastructure will be a key factor in helping organisations – and nations – compete on the global stage.

Author

Séamus Dunne is the UK&I managing director for Digital Realty. He leads the firm in the UK and Ireland, with a focus on delivering growth and enabling businesses to manage digital transformation. He works closely with customers to navigate the evolving digital landscape and the solutions co location can offer.