Afry, a global energy consulting firm, forecasts Europe’s data center demand to increase from 75-85 TWh in 2025 to 150 TWh by 2030 and 330 TWh by 2050. Subsequently, in this DCD>Talk, Carlos Pérez Linkenheil, head of reports at Afry Management Consulting, sits down with DCD’s Alex Dickins to explore the evolving role of data centers in the wider energy ecosystem.

This includes the need for grid flexibility, the importance of procurement strategies and Power Purchase Agreements (PPAs) to manage rising energy costs, as well as the need for data centers to either partner with energy companies, or partake in their own onsite energy production to help address grid constraints, and provision for sustainable energy solutions.

As independent advisors, Afry works “with utilities, governments, regulators, developers, banks and also large industrial off takers,” Linkenheil explains. While the narrative is commonly about demand for energy and power increasing, he suggests that there has been some stagnation:

“What we’ve observed, specifically in Europe, is that for years, we have seen stagnating demand because on the one hand, you had increasing economy, but you also had increasing efficiency,” he notes. As a result there has been a leveling out of demand expectations, which he says has also led to “a strong assumption that demand will increase due to the fact that decarbonization is one of the key pillars going forward.”

This drive toward decarbonization is largely spearheaded by the electrification of other sectors, such as transport and heat. Linkenheil suggests that political decision-makers and regulators have slept on the role data centers would play in this scenario. Combined with demand from electric vehicles and heat pumps, data centers are one of the key drivers for future demand, although exactly where the demand will come from could vary significantly.

Data center demand

“Location by location, there could be a very small country such as Ireland, where a lot of its demand will be actually coming from data centers. Whereas in other geographies, the impact of data centers will be to a lesser extent,” he clarifies.

With electricity demand projected to reach 330 terawatt hours by 2050, in Europe, this equates to a four-fold increase in just 24-25 years. However, Linkenheil believes this is a conservative view. Nobody, he suggests, has completely understood the impact of artificial intelligence (AI), and the overall increase could be much stronger.

This means that it is critical to understand demand across the wider ecosystem – including in other industries. In part this is because the demand from data centers is a bit different, he suggests, from other sources of energy and power demand growth. This is because data centers typically have a load factor of more than 80 percent. In other words, they run all of the time during the day.

“This is very much unlike electric vehicles, heat pumps, residential and commercial demand, which are way more peaky and more flexible,” he stresses. With data centers, he highlights the need for firm round-the-clock power to supply them.

Price volatility

Another challenge is price volatility, which data centers must consider within their long-term planning. He adds: “We actually ran a study for the government in Finland, and we quantified that adding load from additional data centers pushed peak prices up very quickly, from 500 euro, up to 900 euro per megawatt hour,” indicating that flexibility is becoming a core strategic asset for operators.

Another consideration is the growth in renewable penetration, with Linkenheil highlighting the fact wholesale markets are producing more hours of exceptionally low cost energy due to the fact that wind and sun are abundant.

The key to getting a handle on the volatile nature of energy is flexibility. “Meaningful share of data center load is not time critical. It can be shifted away from peak price hours, which can really reduce energy costs, and operators who can credibly offer load flexibility will also automatically be more attractive to grid operators and regulators, because they can be part of the solution.”

Long-term procurement strategies

Subsequently, long-term procurement strategies – including nuclear PPAs – are no longer optional. They also need to cover renewables and gas.

“There have been lots of stories lately on nuclear PPAs, but structuring these contracts well requires deep understanding of forward power markets, market design, and regulatory trajectories, because you’re locking in a price for up to 20 years. Obviously, you don’t want to purchase electricity at a much higher rate than you could expect from alternative routes.”

As the adage goes, time is money, and unfortunately, infrastructure takes time to build. For example, in Northern Virginia, grid connections for projects over 100 megawatts can take more than seven years to build, and in the UK it can take 10 years or more. In the Republic of Ireland, he says that “5.8 billion data center projects are stranded; they have land and permits but no grid capacity, and across the globe, hundreds of gigawatts, if not terawatts, of grid queues exist.”

Lengthy queues for grid connectivity reveal that “there are lots of different technologies and demand sources competing for the same bottleneck, which is appearing more and more to be the grid,” Linkenheil explains. Within this context, he suggests it’s important to consider the need for data centers to “offer flexibility to reduce stress on the grid.”

And while a data center can be built in 18 months, an electricity substation can take five years to construct. This is driving a fundamental shift in procurement architecture from backup power to a primary energy strategy. Linkenheil adds that gas turbines, fuel cells, battery storage, renewable hybrids, are all being deployed at scale next to the sites where land is available.

He adds: “The industry is also diversifying away from established hubs and to markets where grid headroom might still exist. This is possible – at least for some. This is where we currently see the market going as well.”

Partner, or self procurement?

With data center growth and partnerships with the grid and power providers strengthening, data centers could become part of the solution rather than just a source of pressure on the grid. However, data centers aren’t energy companies. So there is a need to bridge the gap between the two.

Linkenheil says they have to decide to either go into a strong partnership with energy players who can solve their problems or become an energy company in their own right. He therefore concludes by suggesting that many of these companies will move towards having a role in energy to deliver the power they need as a key driver of success going forward – and even contribute to the grid.

To find out more about the data center’s role in the energy ecosystem watch the full DCD>Talk with Afry’s Carlos Pérez Linkenheil here.