AI and accelerating digitalisation are driving a new wave of data centre development across Europe, while power, connectivity and increasingly water availability are reshaping where projects can be built. Peter Junker, Associate Partner at Drees & Sommer, explains why development is moving beyond traditional hotspots and how rising temperatures are changing the requirements for new data centres.
What are the key trends currently shaping the European data centre market?
Everybody is talking about AI, and it will certainly be a megatrend for data centre development across Europe. I expect perhaps 10 to 15 very large AI data centres in individual European countries. We will not necessarily need hundreds of these huge facilities everywhere, as we might see in the US or APAC markets.
At the same time, there is another major trend. Digitalisation and AI are increasing demand for local data centres close to users. More and more applications need information to be processed within one or two milliseconds. Telemedicine, autonomous driving and virtual reality are examples of services where low latency is critical.
So we are essentially seeing two parallel trends: very large AI data centres of 100 MW, 400 MW or even more, and hundreds of smaller edge data centres located close to major cities and users.
Where are you seeing the strongest development activity in Europe?
If you had asked me two or three years ago, the obvious answer would have been the traditional FLAP-D markets: Frankfurt, London, Amsterdam, Paris and Dublin. These locations remain extremely important for internet traffic and will continue to see development.
However, AI is changing the map. We are increasingly seeing large projects outside the traditional hotspots. For example, there are major developments around Linz in Austria and in the Nordics, often far away from cities such as Stockholm, Oslo or Copenhagen. In France, Lille has also emerged as an important location outside Paris.
In my view, there is no longer a rule saying that a large AI data centre has to be developed in one of Europe's established hubs. Frankfurt, London and Paris will remain important, but rural areas that nobody talked about before can become new data centre hotspots.
Why are developers increasingly looking beyond major cities?
There are three main reasons: power, connectivity and land.
Power availability is obviously critical. At the same time, connectivity has improved dramatically because carriers have invested heavily in fibre networks across Europe. Today, you can achieve very low latency even in locations outside the traditional hubs.
The third factor is land. It is generally cheaper and more readily available in these areas. Together, these factors are making locations outside the traditional FLAP-D markets increasingly attractive.
Europe is also experiencing hotter summers. How are rising temperatures affecting data centre development and operation?
This has become a much bigger issue, particularly following the heatwaves and dry conditions we have experienced. In the future, selecting a site will not only be about power, connectivity and land. Developers will increasingly have to ask whether the location is exposed to extreme heat, whether water is available and whether it is in a water-stressed area.
There is a fundamental trade-off when it comes to cooling. Dry cooling requires significantly more power to cool the same load, while water-based cooling can be much more efficient but obviously requires access to water.
During extremely hot summers, dry cooling systems can reach their limits. That is why operators are increasingly having to reconsider how they approach cooling. Even though the industry has made commitments to reduce water consumption, we are seeing operators consider water-based solutions again where local conditions allow them.
Are there ways for data centres to reduce their impact on local water resources?
There can be opportunities if developers look at the wider infrastructure around a site. We recently examined an industrial location at a chemical park where groundwater already had to be pumped out and discharged into a river because of environmental requirements. Our question was simple: if that water is already being extracted, why not use it to cool a data centre before discharging it?
Technically, it made sense, but it proved very difficult politically. Water use by data centres is a sensitive topic, and public perception can become a major obstacle even where the proposed solution does not actually require additional water extraction.
This shows why water has to be considered from the very beginning of a project. Developers traditionally start by checking power and connectivity. Today, water should effectively be the third question. If you only address it at the end, it can become a major permitting issue.
