Professor: New power generation should accompany data center construction

Large data centers planned for Estonia could end up consuming a significant share of the country's electricity, says Mart Landsberg, a professor at Tallinn University of Technology. The most important question is how much added value can be created with that energy.
Professor Landsberg pointed to two data centers planned in Estonia, one near Tallinn in Hüüru and another in Risti. Both are designed to have a capacity of roughly 200 megawatts.
For comparison, Estonia's peak electricity demand during this past winter's highest-demand hour was about 1,700 megawatts. A single 200-megawatt data center would therefore account for more than 10 percent of that figure. In terms of annual electricity consumption, such a facility could use approximately 1.3 to 1.6 terawatt-hours of electricity, equivalent to roughly 20-25 percent of Estonia's annual electricity consumption.
"What matters about energy is not where it comes from or how much it costs. What matters is how much added value we can create with that energy," Landsberg said.
According to him, even cheap and clean energy may bring limited benefits to society if it is used for low-value-added activities. Conversely, more expensive energy can make economic sense if it generates sufficient wealth and raises people's incomes.
Landsberg said data centers differ from many current industrial and household consumers because their electricity consumption does not depend as heavily on whether conditions are warm or cold. As a result, they can be valuable customers for the electricity system because their demand is more stable.
At the same time, he noted that a growing share of electricity system costs consists of fixed costs, while the cost of the electricity itself, such as the wholesale market price, makes up an increasingly smaller share of total expenses.

"Energy itself no longer costs very much, but capacity and access to energy are worth their weight in gold," Landsberg said.
The problem, according to Landsberg, is that electricity consumption among households and commercial customers in Estonia is growing while industrial electricity consumption is declining. This means there are fewer consumers helping to share fixed network costs, potentially shifting a larger burden onto households and businesses.
For that reason, he said the impact of data centers on electricity costs for Estonian consumers cannot be viewed as purely positive or negative. The overall cost of energy supply must be considered, along with the broader economic benefits data centers bring to society.
Important factors include the jobs created, tax revenue generated and the extent to which data centers use goods and services produced in Estonia. Landsberg cited local industry as an example of a sector that could benefit from new opportunities created by data center development. In his view, the presence of local data centers could also support product development by speeding up innovation processes and creating competitive advantages for companies.
Data center connection requests exceed Estonia's current electricity demand
Asked whether Estonia's electricity system is prepared for the addition of several large data centers, Landsberg said everything depends on how quickly and on what scale those projects are implemented.
He referred to data from Elering showing that Estonia's current peak demand is around 1,700 megawatts, while connection requests from data centers and other large consumers total approximately 3,000 megawatts.
If all of those projects were realized simultaneously, the resulting change would be enormous and could not immediately be accommodated by the existing electricity system. However, connection requests do not automatically translate into completed projects. Landsberg cited Finland as an example where, according to him, there are far more interested developers than signed connection agreements.
In addition, data centers are typically built in phases rather than all at once. One module is completed first, followed by the next. For that reason, development of the electricity system and the regulatory environment must keep pace with project development.
Landsberg argued that investments in electricity networks required for data center projects should not be borne solely by other electricity consumers. Data centers themselves should shoulder a larger share of those costs.
New power generation should accompany data center development
One of the biggest questions, Landsberg said, is how to avoid a situation in which new data centers push up electricity market prices by introducing major new demand that competes for existing generation capacity.
That effect could be reduced by requiring data center developers to ensure additional electricity generation capacity alongside their projects. This could take the form of building their own power plants or entering into long-term agreements that support the development of new generating capacity.
Landsberg pointed to a regulatory change currently being discussed in the Finnish parliament that would require data centers seeking grid connections to ensure additional electricity generation capacity on a long-term basis.
"If done intelligently, data centers could certainly help drive the construction of new power plants," he said.
In his view, it is not enough to rely on market forces alone and assume new generating capacity will automatically appear in response to increased demand. Appropriate mechanisms need to be designed in advance.
As one example, he highlighted plans for the Risti data center, where the project is being developed alongside solar and wind generation as well as battery storage. Power purchase agreements and other obligations may also form part of the solution.
According to Landsberg, the goal should be to ensure that existing electricity consumers benefit rather than lose out from the arrival of data centers.
He also noted that the investment horizons of data centers and power plants differ significantly. Power plants are long-term investments. Wind farms typically have a lifespan of around 20 years, while nuclear power plants can operate for about 60 years.
Data centers, by contrast, have shorter investment cycles because the computers they use require relatively frequent replacement. Landsberg estimated that the average lifespan of computing equipment may be around five years or slightly longer.
This means that if energy costs rise significantly, a data center could relatively easily decide to relocate elsewhere. If the project also includes investments in power generation or energy infrastructure, however, lasting value remains in Estonia even if the data center eventually leaves.
Data centers require a continuous electricity supply, which means their operations cannot rely directly on the variable generation of wind and solar energy. According to Landsberg, one solution is battery storage, which can store energy from wind and solar power and release it when data centers need it. Batteries can help manage fluctuations within and between days.
The unresolved issue remains seasonal fluctuations. Estonia receives little sunlight during winter, raising the question of how electricity generated from solar energy in summer can be stored for winter use.
"At the moment, we do not have a better solution than fossil-fuel power plants, such as gas-fired plants, which can cover the shortfall during winter peak demand periods," Landsberg said.
Over the longer term, he sees an important role for hydrogen, both in hydrogen production and in converting hydrogen back into electricity.
Landsberg said Estonia's strengths include favorable offshore wind conditions and a relatively cool climate that provides efficient cooling conditions for data centers. Combined with hydrogen technology, these advantages could form one of Estonia's future growth engines.

Waste heat use depends on nearby demand
Data centers also produce waste heat that can potentially be reused. According to Landsberg, successful utilization depends largely on whether a sufficiently large heat consumer is located nearby.
He compared two planned Estonian projects: the data center in Hüüru and the planned facility in Risti. Because of its proximity to Tallinn, it will likely be easier to find a major heat user for the Hüüru project.
At the same time, utilizing waste heat requires additional investment, creates additional risk and lengthens project timelines for developers. This can increase costs and delay completion.
As a result, developers must determine whether potential revenues from selling waste heat outweigh the added costs and risks.
Landsberg said that while waste heat utilization may make perfect sense from society's perspective, what is optimal for society is not necessarily optimal for developers. In his view, the state can help by designing a regulatory framework that ultimately delivers the most cost-effective outcome for consumers.
Estonia must learn to capitalize on the electric age
More broadly, Landsberg argued that the world has entered the electric age.
According to him, the International Energy Agency has highlighted the growing shift toward electricity as the dominant form of energy consumption. In the coming decades, key technologies will increasingly be those powered by or connected to electricity.
Estonia's challenge, he said, is determining how to extract as much economic benefit as possible from this transformation. If the country fails to keep up, it may find itself among the losers. If it uses new opportunities wisely, however, it can increase both national and societal prosperity.
Landsberg compared current technological developments to the Industrial Revolution.
"If the first Industrial Revolution centered on transferring physical labor to machines, then the next Industrial Revolution now underway is about transferring intellectual labor to machines. The question for us is whether we want to be among the winners or the losers in this revolution. Energy exists to do work on behalf of people," Landsberg said.
--
Editor: Argo Ideon











