The data center market has been enjoying a big boom for a few years now, and the introduction of artificial intelligence (AI) to the mass market will continue to fuel the data center growth for years to come. With that said, there are a lot of changes that are happening in the data center market, especially in Europe.
Many of these changes are provoked by new challenges that are arising. In Europe there are some unique situations that are specific to the region. Thus, the European data center market shapes up to be quite dynamic and interesting for the foreseeable future. Let’s explore a bit more, shall we?
First, let’s get to know the state of the data center market in Europe
According to Statista, as of 2022 there are 2,904 active data centers in Europe. The majority of them are in Germany – 487, the United Kingdom – 456, the Netherlands – 281, and France – 264. Unsurprisingly, there are several hotbeds where there is a high concentration of data center investments. Those are generally Frankfurt, London, Amsterdam, Paris, and Dublin, they’ve even been nicknamed as the FLAPD. It’s up to personal interpretations if that is a lot of data centers or not, butwhat is for sure is that they aren’t enough and the market needs more. And that’s a problem.
This is because of the rise of AI use and demand is sucking up all the available data center capacity. But at the same time the demand for other data center services, like cloud, hosting, data storage and colocation, is not easing. On the contrary, there’s an increase in all areas. According to a report by CBRE, colocation was in high demand across Europe in 2023.
The report says that 601 MW of space was leased in the 14 biggest European markets last year. Meanwhile only 561 MW of new space was opened. What’s more, CBRE’s initial forecast for 2023 was 480 MW of take-up and 524 MW of new space. So, the market is growing much, much faster than expected.
This is the second time in five years that the demand outpaces the supply. In 2022 there was a similar picture – the demand grew by 10% while the supply by only 2%. You can see where this is going. At this rate we are soon going to have “waiting lines” for data center space.
The CBRE report shows that FLAPD are the main locations fueling the imbalance. In Q4 of 2023, there was a 41% jump of take-up to 252 MW. This was mainly due to pre-let capacity in Dublin, London, and Paris.
The bubbling water of challenges
So, the solution is obvious and simple, right? Just build more data centers and build them fast. Ah, but sadly, it’s not that easy. The demand continues to surge and highlights local challenges like not enough power for all data centers in a said region or simply not enough physical land to build them.
“Hyperscalers have sought increasingly larger facilities that are tailored to their needs for some time. As such, the average size of new wholesale facilities in Europe has grown dramatically over the past two years and will only accelerate. However, there are fewer deals being struck by data centre providers with their largest customers,” Kevin Restivo, head of European data centre research at CBRE says to PropertyEU.
And it’s not like the demand will ease any time soon. “Demand in Europe will only accelerate given that artificial intelligence (AI) workloads are expected to generate higher demand for capacity, as organisations look to enhance their operations by leveraging machine learning algorithms. For the meantime, the challenge remains the lack of available power and AI-appropriate facilities across Europe,” Andrew Jay, head of data centre solutions at CBRE adds.
“There’s still an insatiable demand for these FLAPD locations. The cloud has generally gone to the most mature markets because they’re either capitals, or financial hubs, or just big metros,” said Ben Stirk, co-head of data center advisory at Knight Frank, during a DataCenterDynamics panel.
Ash Evans, Google’s EMEA lead of energy and location strategy, theorizes that we are at a point of major divergence, where data center sub-asset classes are becoming even more distinct. “I think we’re seeing the emergence of a new asset class in terms of AI and ML data centers in particular, which are on another scale again, and arguably have a different set of specifications. Are they less reliable? Do they even need to have the same reliability? Do they have the same latency requirement? This is then where we have the divergence of two different machine learning style data centers,” says Evans.
JLL also has data that shows the take-up of data center space and resources is far larger than the new capacity being added. And it also shows more of the same – data center demand is increasing for all uses and verticals. Specific data centers, edge, etc., are all in high demand and it doesn’t seem like we are even close to the peak.
So, how will the market adapt?
The challenges are specific to each location and there can’t be a universal solution. One idea is to use more Edge data centers. But the thing is Europe is already very dense when it comes to data centers, and most of them are already built close to the hotbeds for specific industries and goals.
There aren’t a lot of locations in Europe which are both in need of high-powered computing and remote enough to warrant an Edge solution. Also, the Edge is for smaller deployments, while the demand in Europe – and especially FLAPD – is for hyperscale capacities. The cloud, colocation, and AI all need big resources and hyperscale is the answer for them.
So, the solution is different. “I think we will start to see more and more concepts whereby large-scale renewable developments are sited or colocated next to hyperscale AI or machine learning data centers,” said Evans.
Then we move to the question of land. Real estate in the FLAPD areas is already very expensive, and a hyperscaler demands a lot of land which just drives the price even higher. Thus, we can expect that part of the focus will shift to different regions and the Nordics are the prime candidates.
“The Nordics are the obvious place for these big training centers. They have cheap land, cheap renewable power, and free air cooling, all of which provide a great opportunity for these massive deployments,” says Stirk. But not so fast. They have their own share of specific issues. For example, they are less mature markets for such developments and thus there’s not enough skilled workers. It will take time to train local talent up to the desired levels and relocating people from other areas might not be so easy. That’s an approach only big players can afford. And it’s something that big data center building contractors are already applying.
“They take their teams and settle them near the project, and then build the data centers. That’s a model we’ve seen in FLAPD, and we see that now happening elsewhere. Whether that is sustainable is to be seen but, for now, it is working. With the amount of data centers required in these areas, one project can be finished, and the same team can easily stay on and develop the next one,” says George Demetriou, director of mission-critical businesses for Arup.
But it’s not so easy when it comes to finding the people who will work in the data centers long-term. “While Google has a large ground-up talent pool development scheme, we can still do better at attracting talent and engaging earlier. A great example of a sister industry that’s been successful at this is the oil and gas industry. They attract talent straight out of universities and colleges and provide very clear career paths and progressions. As of today, I certainly haven’t seen a lot of that in the data center sector,” says Evans.
Fight for the prize
Then there’s the question of job security. With the development of AI and further data center automation being inevitable, there will be less and less need for actual workers. Such will always be needed, but in smaller numbers. So, companies must solve the puzzle of convincing workers to develop their skill to the desired levels for data centers, to work at remote locations and to set up lives there, with the notion that one day sooner rather than later, but without an exact forecast when, their services might no longer be needed.
And some countries already have realized that and are using it to their benefit. Germany for example doesn’t want to lose that FLAPD gravy and the potential business revenue that comes with it. On the contrary, the country wants to further expand its role in the European data center market and it’s working towards that goal. It wants to show data center operators that it has a good central location, reliable access to power, good overall infrastructure, and conditions for all types of data centers. Oh, and it also has a lot of land for them to be built.
The country is hoping to continue to attract big investments in the segment and it recently secured a major contract with Microsoft. The software giant announced it will invest 3.3 billion Euros in Germany to double the capacity of its AI and data center infrastructure. In addition to that, the company will also help train up to 1.2 million people and improve their AI skills to get them ready for the AI age. Microsoft also plans to use the investment to get closer to major companies in other industries and offer them tailored data center services. Thus, we see how other economic factors and possibilities can also influence data center operators’ decisions on where to build new capacity. And Germany certainly seems to have a good grip on that strategy and it’s already paying off.
Sofia Data Center
Sofia Data Center (SDC) is the brand under which the global telecom operator Neterra provides services in its 4 data centers SDC 1, SDC 2, SDC Stolnik and SDC Ruse.
They are ISO and PCI DSS certified contributing to its customers with a strategic location in Southeastern Europe, security, quality, competitive pricing, low latency and 24*7 professional technical support.