Dashboard
Key Takeaways by Leitmotiv on Dutch Data Centers
Executive Summary
Our analysis of legally required reporting by data centers in the Netherlands - as well as estimates for missing data centers - reveals critical transparency gaps and raises questions about the industry's energy use, emissions and capacity. We identify five key concerns:
1. Severe lack of transparency
While 104 data center buildings reported to RVO, we estimate that xx data centers should have reported to RVO. Of the 104 data center buildings that reported, only 77 disclosed their electricity consumption. 24 of the 27 data center buildings that did not share their energy use with RVO are ultimately owned by US-headquartered companies.
2. Policy implications of missing data
Without accurate figures, policymakers cannot reliably assess grid pressure or climate impacts now and the in the future. This data vacuum enables industry interests to shape policy narratives, particularly regarding future growth scenarios.
3. Actual electricity use
Reported data center buildings consumed only 0.91 TWh in 2024. Our estimates - including non-reporting facilities - reach 4.45 TWh for 2024, in line with estimates by CBS and PBL. That is significant, and equals the annual electricity of 2.4 million Dutch households. It is also important to have a firm factual basis for current electricity use in light of what appeas as extreme growth projections for the near future.
4. Greenhouse gas emissions
Estimated greenhouse gas emissions related to data center electricity use in 2024 is around 1.1 million tons of CO2-equivalent. If you add up data center emissions related to servers installed and use of back-up generators, total greenhouse gas emissions for Dutch data centers were at 2.1 million tons of CO2-equivalent emissions.
5. Low capacity utilization exposes antisocial aspects of data center power grab
Reporting data centers operate at just 32.2% electrical capacity utilization, meaning two-thirds of their electrical capacity remains unused while blocking grid access for other users, such as households and other businesses. Even with higher utilization estimates for non-reporting data centers, data centers can more than double their server count before reaching full capacity.
Background
Via the Leitmotiv datacenter dashboard, we attempt to make accessible basic information on data centers in the Netherlands such as their total number, how much energy they consume, and how many greenhouse gases they emit.
The evidentiary basis for the dashboard is actual information reported by data center operators to the Dutch authorities (RVO), a requirement under European and national legislation. Since information for a significant number of data centers has not been reported to the RVO, we have added estimates for missing data center buildings. How we arrived at those estimates is described in the methodology section of our dashboard.
Our dashboard highlights the lack of transparency in the data center sector in the Netherlands. It also shows the lack of enforcement of existing transparency regulations. 104 data center buildings in the Netherlands with a capacity of 500 kW or more (the legal floor) reported to the RVO. But by our count, there are [XXXX] data center buildings in the Netherlands, meaning [XXX] data centers are missing altogether from the RVO data altogether.
Perhaps more importantly, key performance indicators are missing for a number of the reported 104 data center buildings For example, of the 104 data center buildings that were reported to the RVO, electricity use was only shared for 77 of those 104 data centers. What is also striking is that of the 27 data center buildings that did not report their energy use to the RVO, 24 were ultimately owned by companies headquartered in the United States. To be clear: this severe lack of transparency is not normal and would not be accepted if this concerned another sector of the economy.
Electricity Consumption and Greenhouse Gas Emissions
Electricity use is a crucial figure for datacenters because it shows how much of the total electricity produced and distributed in a country goes to the data center sector at present. It is also important because data centers are mostly powered via the electricity grid in the Netherlands, meaning that most of its so-called 'scope 2-greenhouse gas emissions' relate to its grid electricity use. Based on the 77 data center buildings that reported their electricity use to the RVO, data centers in the Netherlands consumed 0.91 TWh of electricity in 2024.
Energy Consumption Comparison
Comparison of energy consumption estimates from various sources
That figure of 0.91 TWh for 2024 is very low compared to other estimates. According to the Dutch national statistics office (CBS), data centers consumed around 3.73 TWh in 2021. CBS looked at the actual client files of grid operators, which is likely to have led to a fairly accurate estimate. For 2023, the Dutch Environmental Assessment Agency (PBL) estimated total electricity consumption by data centers in the Netherlands at 4.72 TWh. It is only when we add estimates for missing data center buildings to the figure of 0.91 TWh of actually reported electricity use that we arrive at a total estimated data center electricity use for 2024 of 4.45 TWh. That is more or less in line with estimates by CBS and PBL for 2021 and 2023, respectively. Using the figure of 4.45 TWh of electricity consumption of data centers in the Netherlands in 2024, we arrive at total greenhouse gas emissions from that consumption (so-called 'scope 2 emissions') of 1.1 million tons of CO2-eq. This figure rises to 2.1 million tons of CO2-eq when we add emissions related to server production and back-up generators.
The Policy Relevance of Accurate Data
Why is it relevant to have solid figures on key performance indicators such as data center electricity use? For one because without such solid figures, policymakers are not certain of the exact pressure of data centers on the electricity grid or their contribution to climate change at present. More importantly, in the absence of hard data, special interests can manipulate policymakers and public opinion more easily, including on future trajectories for this sector.
Grid operator TenneT, for example, assumes that data centers consumed 5.6 TWh in electricity in 2023 and that this may reach around 10 TWh as soon as 2030. Netbeheer Nederland, the association of all gas and electricity grid operators in the Netherlands, estimates that data centers use around 10 TWh of electricity in 2025 and assesses it may grow to around 25 TWh in 2030. It seems likely that those are inflated figures, relying (at least in part) on estimates from the data center industry itself. Given where data center electricity use figures are likely at right now (around 5 TWh), we think it is highly unlikely that electricity use will be five times higher less than 5 years from now.
Capacity Utilization and Grid Allocation
This relates to another matter that our dashboard shines a light on: electrical capacity utilization rates. Among the data centers that reported relevant data to the RVO, the current capacity utilization rate stands at 32.2%. In short, this means that data centers that reported are on average only using 1/3rd of their electrical capacity. That means, generously speaking, that these reportingg data centers are at least 2/3rd empty, yet taking up space on the electricity grid that can't be used by other users, such as new houses, electrical vehicles or other businesses. While we estimate that capacity utilization rates are higher for non-reporting data centers, our estimate is still that data centers in the Netherlands have a capacity utilization rate below 50%.
It should be stressed that maximum electrical capacity is not the same as actual use of electricity. IT equipment inside data centers does not run at full capacity all the time. This only underlines that there is a disconnect between grid capacity reserved by data centers and actual need at present.
Power Capacity
Comparison of power capacity and utilization metrics
The foregoing relates to the public debate on data centers in two ways. First, it is often claimed by the data center industry in the Netherlands that more data centers need to be built and given priority on the grid because we are rapidly running out of data center capacity. Our data shows that this is not the case in the near future in the Netherlands. Even if we assume that non-reporting data centers have higher utilization rates than the average for data centers reporting to RVO, we only arrive at an average utilization rate for all data centers in the Netherlands of around 32.2%. In that scenario, the number of servers in data centers in the Netherlands can grow by more than 100% before data centers are at full capacity. For context: if we go by the growth in electricity use by data centers, the market grew by around 27% between 2021 and 2023.
Another reason why electrical capacity figures are relevant is because of the pressure data centers put on the Dutch grid. When a data center is built is basically reserves space on the electricity grid that can no longer be used by other companies and households. What is currently happening can be summarized with the following analogy: if grid operators in the Netherlands have built a highway for data centers, that highway is – only occasionally – filled only for about 50% of its capacity. Meanwhile, the other 50% remains unused until it may be used in the future. In other words, there is a huge gap between the electricity that can be used by all data centers in the Netherlands if they were 100% capacity and the actual electricity use of data centers at present.
Growth Scenario Analysis
Analysis of capacity considering IaaS growth and contracted power