Why I Conducted This
Buildings are consistently cited as one of the EU's largest sources of energy consumption and emissions, yet they receive less analytical attention than heavy industry or transport in most sustainability work. I wanted to understand two things with actual data: first, is total building sector energy consumption declining in a meaningful way, and second, does the per-capita picture reveal anything about where inefficiency is structurally concentrated?
The analysis uses three Eurostat datasets: final energy consumption by sector, disaggregated household end-use (space heating and cooling specifically), and population. Together they build a cross-country picture grounded in reported figures. The cooling data turned out to be the most striking finding: a 60% increase in residential cooling demand between 2015 and 2023, running directly counter to the headline downward trend in heating.
This matters beyond energy efficiency. Under CSRD and the EU Taxonomy, real estate companies must now report and align on energy performance. The data here shows the scale of what that alignment task actually involves across different member states, and how unevenly distributed the challenge is.
What the Data Shows
Combined household and commercial energy consumption dropped from 373,793 ktoe in 2015 to 345,857 ktoe in 2023. But a rebound in 2021, when a colder winter pushed residential heating up 7.6% in a single year, reversed several years of apparent progress. Much of the headline decline is driven by mild winters and fuel switching, not structural improvements to the building stock itself.
Finland sits at 27.0 GJ per person; Malta at 1.9 GJ. Climate explains part of this gap. But the divergence within similar climate zones is equally striking: Germany at 17.5 GJ versus the Netherlands at 10.5 GJ, both in Northwestern Europe with comparable winters. That gap points to differences in building stock age and insulation quality, not just geography.
Space cooling went from 37,956 TJ in 2015 to 60,759 TJ in 2023. This is a sustained multi-year increase that accelerated after 2020. While cooling remains far smaller than heating in absolute terms, its trajectory is the structurally significant story: as hotter summers compound the demand, buildings not designed for summer heat face a new energy burden on top of the existing renovation challenge.
High per-capita heating markets such as Germany, Austria, Belgium, Estonia, and Czechia face the largest exposure to incoming EPBD minimum energy performance standards and EU Taxonomy alignment requirements. Southern European markets face a different but equally urgent challenge: a cooling demand surge in building stock that was not designed for it. CSRD reporting will make both of these gaps visible at the portfolio level for the first time.
Total Consumption: A Declining but Volatile Trend
The chart below shows EU-27 final energy consumption in buildings: households and commercial/public services separately, across 2015 to 2023, in thousand tonnes of oil equivalent (ktoe). Both series come directly from Eurostat's energy balance dataset (nrg_bal_c).
The steepest drop (2022–2023) coincides with the energy price shock and a mild 2023 winter. Both are temporary drivers. Without physical improvements to the building stock, the trend is highly reversible the moment conditions change.
One colder-than-average winter pushed EU-27 household energy back to 2017 levels. Buildings that have genuinely improved their thermal envelope would not show this sensitivity. The spike is a proxy for how little structural renovation has actually occurred.
Commercial and public services energy fell 7.9% over the period versus 7.3% for households. This likely reflects structural changes in office utilisation post-COVID and more active energy management in larger professionally managed assets, where energy costs are a more visible line item. Households are structurally harder to reach: renovation decisions are fragmented across millions of individual property owners, each facing different financing constraints, tenure arrangements, and incentive structures. This fragmentation is precisely what makes the renovation wave target of 3% annual deep renovation so difficult to operationalise in practice.
The Heating Gap: Where Building Stock Inefficiency Shows Up
Per-capita residential space heating energy (GJ per person) controls for country size and gives a more meaningful cross-country comparison than absolute consumption. The values below are derived from Eurostat's disaggregated household energy dataset (nrg_d_hhq, space heating in TJ) divided by population (demo_pjan), both for 2023. Orange bars indicate countries above 12 GJ per capita; red bars above 17 GJ.
Both countries share broadly similar winter conditions. The gap likely reflects Germany's larger share of pre-1980 building stock and slower deep renovation rates compared to the Netherlands, which has pursued more active insulation and district heating programmes.
Nordic and Baltic countries face genuine climatic demand, but also older building stock. Much of Central and Eastern Europe's residential stock dates from the Soviet era, with poor insulation standards. High heating demand here combines climate exposure with a structural renovation backlog that is large relative to national GDP.
Spain (4.7 GJ), Portugal (3.9 GJ), Greece (8.0 GJ), and Bulgaria (6.0 GJ) show low per-capita heating demand, primarily reflecting warmer climates rather than efficient buildings. But these are precisely the markets where cooling demand is rising fastest and where building stock was designed around heating, not summer heat management. The policy implication is structurally different to Northern Europe: less about adding insulation for winter, more about redesigning how buildings handle a climate that is shifting faster than their thermal envelopes were built to accommodate.
Cooling Is Rising Fast, Running Counter to the Heating Trend
While total building energy and space heating have broadly declined, residential cooling demand has moved in the opposite direction. The chart below shows EU-27 residential space heating (PJ, left axis) and residential space cooling (TJ, right axis) across the same period. Note the scale: heating is still orders of magnitude larger than cooling in absolute terms, but the cooling trajectory is the structurally significant story for the next decade.
Every TJ saved through better insulation risks being offset by rising air conditioning demand in summer. Buildings designed only for heating efficiency may perform well in winter but drive electricity consumption up in summer. Net progress is smaller than the headline heating number suggests.
Italy, Spain, Greece, and France account for the majority of EU-27 cooling demand growth. These are markets where building stock was historically designed around heating, not cooling. This creates a structural mismatch between the thermal performance of existing buildings and the climate they now face.
Most EU building renovation frameworks, including the Renovation Wave and the revised EPBD, were designed with heating efficiency as the primary objective: better insulation, better windows, better heating systems. The cooling data shows this framing is becoming insufficient. A building that performs well in winter but becomes unliveable in summer without air conditioning has not solved the decarbonisation problem. It has shifted it to electricity demand. For Southern European markets in particular, deep renovation strategies need to incorporate passive cooling design as a primary objective: appropriate shading, thermal mass, orientation, and natural ventilation. This also carries implications for EU Taxonomy technical screening criteria, which remain primarily calibrated around heating energy performance and will need to evolve to capture cooling-related efficiency as a core criterion.
How to Apply This Analysis
The three data patterns above: volatile total consumption, wide per-capita heating divergence, and rising cooling demand, translate into four concrete implications for real estate companies, investors, and ESG practitioners working in the EU built environment.
The 7.5% decline in EU-27 buildings energy since 2015 is real but fragile. The 2021 rebound demonstrates that a significant portion of apparent improvement is weather-driven, not structural. For real estate portfolio managers and ESG analysts, this means that raw energy consumption figures cannot be taken at face value as evidence of decarbonisation progress. CSRD reporting under ESRS E1 will increasingly require the distinction between weather-driven and structural improvement to be made explicit, a data discipline that most organisations have not yet operationalised.
Markets with the highest per-capita heating demand, including Germany, Austria, Belgium, Estonia, and Czechia, are precisely the markets where incoming EPBD minimum energy performance standards will have the most disruptive impact. Buildings in the lowest EPC categories in these markets face a direct pathway to reduced lettability and mortgageability. Real estate investors with concentrated exposure to these geographies who have not yet mapped their portfolio's EPC distribution are carrying unquantified regulatory transition risk that is now on a defined legislative timeline.
The 60% growth in EU-27 residential cooling demand since 2015 is concentrated in Southern and Central European markets. Buildings in these markets that lack passive cooling design face rising air conditioning loads that erode any efficiency gains made on heating. This translates into a forward-looking energy cost and comfort risk that is not yet fully reflected in asset valuations or EPC ratings, both of which remain primarily calibrated around heating performance. For real estate strategy, positioning assets in Southern Europe requires factoring in cooling-related capital expenditure that was not part of the renovation calculus five years ago.
The Eurostat data used in this analysis is country-level aggregate, useful for understanding sector-wide patterns but insufficient for portfolio-level compliance. Under CSRD (ESRS E1) and EU Taxonomy technical screening criteria, companies need building-level data: EPC ratings, actual consumption per m², heating system type, and renovation history. Most real estate portfolios, particularly those with older stock in German, French, or Central European markets, lack this data at the asset level. Closing this data gap is now an operational compliance requirement, not an optional ESG initiative, and it represents one of the most significant near-term data infrastructure challenges in European real estate.
Where your assets are located shapes your decarbonization cost more than most portfolio reviews acknowledge. A building in Finland consumes 14 times more energy for heating than a comparable one in Malta — and that gap translates directly into retrofit investment requirements, energy cost exposure, and EU Taxonomy compliance obligations. Two funds with similar asset sizes but different geographic concentrations face fundamentally different financial burdens on the path to alignment. Investors who haven't mapped their portfolio against country-level energy intensity are likely working with an incomplete picture of where their largest costs are coming from.
Cooling demand across EU buildings has risen 60% since 2015 — a shift that most building assessments designed before 2020 did not account for. A building that met energy performance benchmarks at the time of design may already be underperforming against current consumption patterns, particularly in Southern and Central Europe where cooling growth has been sharpest. For developers planning new projects or major retrofits, this means energy modelling needs to reflect where demand is actually going, not where it was when the standards were written.
CSRD and EU Taxonomy reporting will make country-level energy performance visible to investors and regulators in a way it has never been before. Companies with real estate exposure across multiple member states will need to explain not just their aggregate energy consumption, but why it looks the way it does — and what they are doing about the markets where performance is weakest. The 14-fold variation in per-capita heating intensity across the EU means that explanation will be very different depending on where assets sit. Companies that understand this geography now will be better positioned when the questions start coming.
Methodology
Eurostat Complete Energy Balances, filtered to household energy use and commercial and public services energy use. Unit: ktoe. EU-27 aggregate and individual member states, 2015–2023. Downloaded April 2026.
Eurostat Disaggregated Final Energy Consumption in Households, filtered to space heating and space cooling. Unit: terajoules (TJ). EU-27 and member states, 2014–2023. Several member states report zero for some years due to incomplete national reporting; these are excluded from per-capita calculations. Downloaded April 2026.
Eurostat Population on 1 January, total population, all ages, both sexes. Unit: persons. EU-27 member states, 2016–2025. 2023 figures used for per-capita heating calculations to match the heating data year. Downloaded April 2026.
GJ per capita = (space heating in TJ × 1,000) divided by population. The multiplication by 1,000 converts terajoules to gigajoules. Countries where the 2023 heating figure is zero or missing are excluded from the country comparison chart.
This analysis covers energy consumption, not CO₂ emissions. Converting to emissions requires country-specific emission factors by fuel type, which vary significantly and are outside the scope of this dataset. Cooling data likely undercounts actual demand in countries with incomplete Eurostat reporting. All figures reflect reported national statistics and are subject to Eurostat revision cycles.