Summary
- Average EU non-household electricity prices declined 5.4% year on year to €0.18/kWh in the first half of 2026.
- Ireland recorded €0.27/kWh in the measured consumption band, compared with €0.09/kWh in Finland.
- The figures cover annual consumption of 500–2,000MWh and cannot establish electricity prices paid by large data centre operators.
Eurostat has reported substantial differences in non-household electricity prices across the European Union, with Ireland recording approximately three times Finland’s average price in the consumption category covered by its latest statistics.
Average EU prices fell to €0.18 per kilowatt-hour during the first half of 2026, declining 5.4% compared with the corresponding period in 2025 and 2.0% from the second half of last year.
Ireland recorded the highest average at €0.27/kWh, followed by Cyprus and Italy at €0.23/kWh. Finland reported the lowest price at €0.09/kWh, while Sweden recorded €0.12/kWh. Denmark, Spain, and Portugal were each at approximately €0.13/kWh.
The figures relate to non-household consumers using between 500MWh and 2,000MWh annually. This is a defined statistical consumption band rather than a comparison of tariffs available to hyperscale data centres, whose electricity consumption can be substantially greater.
Across the EU, 19 countries recorded year-on-year price declines and eight experienced increases. Luxembourg reported the largest reduction, at 23.8%, followed by Czechia at 11.9% and Greece at 11.8%.
Prices increased by 18.8% in Sweden and 18.7% in Finland, even though both remained among the lower-priced countries in the measured category. Bulgaria recorded a rise of 14.2%.
Electricity prices and data centre economics
Electricity represents a substantial operating expenditure for computing facilities, particularly those running continuously at high utilisation. Nevertheless, the price a data centre pays depends on its individual supply arrangements and cannot be derived reliably from a national average covering smaller industrial and commercial customers.
Large operators may procure power through bilateral contracts, power purchase agreements, wholesale-linked arrangements, or a combination of contractual structures. The resulting costs depend on energy prices, network charges, balancing arrangements, taxes, and other applicable terms.
Network access creates a further distinction. A facility seeking substantial import capacity may need a new connection, additional substations, or wider network reinforcement before the requested power becomes available.
These investments can introduce costs and delays that are not captured by a published electricity price per kilowatt-hour. Two sites operating in the same national market may consequently face different development costs and electricity procurement conditions.
Existing facilities with secured electrical capacity may also have a different commercial position from projects still awaiting a connection. A location with relatively inexpensive electricity does not automatically provide a faster route to energising a new data centre.
The relationship between electricity prices and competitiveness is further affected by operating efficiency. Equipment utilisation, cooling design, power distribution losses, and the surrounding climate influence the quantity of electricity a facility consumes for a given computing workload.
Power usage effectiveness describes the relationship between total facility energy use and the energy consumed by IT equipment. It can provide information about infrastructure overhead, but it does not establish the unit price of electricity or the total operating cost of a computing installation.
National differences need careful interpretation
The prices reported by Eurostat provide a consistent statistical basis for comparing a particular class of non-household customers. Their usefulness for evaluating data centre investment lies in showing broader market differences rather than providing a direct quotation for a new hyperscale campus.
Ireland illustrates the importance of separating electricity cost from infrastructure availability. Major data centre developments there also encounter questions about network access, generation adequacy, and the terms under which additional large electricity loads can be connected.
Finland’s lower measured electricity price may contribute to its attractiveness for energy-intensive investment, but developers still require sufficient connection capacity, suitable land, technical infrastructure, and acceptable commercial terms.
Sweden’s year-on-year increase demonstrates that a country with comparatively low average prices can nevertheless experience substantial price movement. Longer-term investment assessments need to consider the duration and structure of electricity contracts rather than assuming one reporting period will determine future operating costs.
Local conditions can differ from national averages, particularly where transmission capacity, network congestion, or specific industrial supply arrangements affect electricity procurement. Currency movements can also influence comparisons expressed in euros for countries outside the euro area.
High-density computing introduces further operational considerations. Accelerators and associated networking equipment can increase electrical and cooling demand, while the facility’s support systems must maintain acceptable operating conditions under changing equipment loads.
Absolute electricity consumption depends on both the computing workload and the efficiency of the infrastructure supporting it. A low unit tariff does not establish low total expenditure where the facility consumes substantial amounts of energy.
The latest Eurostat statistics confirm that electricity prices remain uneven across Europe despite the overall reduction during the first half of 2026. Assessing a data centre investment still requires project-specific evidence about electricity supply, grid connection costs, contracted capacity, technical design, and the final delivered price.
