StackPath Ireland

Try a question

Answers are quoted from the site's own pages, with the section and its sources. Nothing is generated.

Dashboard · Our coverage · data to 12 October 2026

Data coverage: how deep we go, country by country

Depth of our data

Ireland: tier 4, 1st of 216. That is the highest.

tier 4
Ireland, 1st of 216
2026-10-12
tier 2
Europe-31 (weighted)
reference
tier 3
World (weighted)
reference
216 / 216
countries with a figure
inference
131111
0 · none1 · headline2 · national profile3 · regional or site4 · every projectno dataIrelandtier · classes: one per value

How to read this. 0 none, 1 headline measures, 2 national profile, 3 regional or site rows, 4 every project. Auto-detected, then the register (research/datasets/data_coverage.json) confirms or overrides with evidence. Tier 4: at least two project-level datasets with planning or grid status above their row floors. Tier 3: a sub-national split (5+ regions) or a site list of 10+ data centres for the country, in the data-centre or power layers. Tier 2: a profile page and at least 50% of the scope's dashboard measures. Tier 1: any dashboard measure. Each tier needs the one below it. Higher is better.

Why it matters. How far down we go in each country: project, regional, national profile, headline or none.

Source: StackPath data-coverage analysis (country_metrics/dashboards, eu_scorecard, world_power (Epoch AI), the data manifest, scripts/refresh.py, research/datasets), data for 2026-10-12, pulled 2026-10-12 · inference · 216 of 216 countries with a figure · source · Compiled by StackPath

Read the analysis →

We hold 91% of the depth the market calls for; 23 countries are below target

Σ weight × min(depth, target) ÷ Σ weight × target: how much of the depth the market weights call for we hold. Below target in Europe: France, Netherlands, United Kingdom. Ireland: depth 4 of 4, in 73 datasets.

How deep we go: 1 country to the project, 3 to the region or site, 28 to a national profile

Each country sits on one rung of this ladder. The target rung depends on how much the market matters: Ireland is the home market, the largest markets get regional detail, and a smaller country needs a national profile, not council-level sites.

DepthWhat it meansExampleCountries
4 · ProjectEvery project or site, with planning, grid and statusIreland (gap table, councils, campuses)1
3 · Regional / siteSub-national rows: regions, a site register, a state-level splitGermany's data-centre register, Italy's regional grid queue3
2 · National profileA country page and most national measuresthe 31 European countries28
1 · HeadlineA few national measures from global datasetsmost of the world182
0 · NoneNothing beyond the map outline2

Get this data next: the 15 biggest gaps, weighted by market size

Priority = the market's weight × the rungs still to climb. The sources named are public; we have not pulled them yet.

  1. France depth 2 → 3 priority 3.79 · weight 3.8%

    Have: Depth 2 (National profile): profile /europe/countries/france; 19 of 19 measures; 19 of 19 Europe dashboard measures; Data centres' share of electricity (published figure); EU EED data-centre reports; Data-centre grid queue (eu_grid_queues); Partial site rows: eu_neoclouds (8); 27 datasets

  2. China depth 1 → 2 priority 3.4 · weight 3.4%

    Have: Depth 1 (Headline): 12/14 measures, no profile, global datasets only; 12 of 14 world dashboard measures; Topic pages: /global/china-power-build; 3 frontier AI site(s) tracked by Epoch; Partial site rows: world_power (3); 9 datasets; none for Capital, Policy and rules

  3. United Kingdom depth 2 → 3 priority 3.19 · weight 3.2%

    Have: Depth 2 (National profile): profile /europe/countries/united-kingdom; 13 of 19 measures; 13 of 19 Europe dashboard measures; Data centres' share of electricity (published figure); Data-centre grid queue (eu_grid_queues); Partial site rows: eu_neoclouds (5); 23 datasets

  4. Australia depth 1 → 2 priority 1.62 · weight 1.6%

    Have: Depth 1 (Headline): 13/14 measures, no profile, global datasets only; 13 of 14 world dashboard measures; 5 datasets; none for Policy and rules

    • Data-centre electricity use and forecast by NEM region (state) — AEMO, 2026 ESOO data-centre forecasting overview (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  5. United Arab Emirates depth 1 → 2 priority 1.6 · weight 1.6%

    Have: Depth 1 (Headline): 13/14 measures, no profile, global datasets only; 13 of 14 world dashboard measures; 1 frontier AI site(s) tracked by Epoch; 5 datasets; none for Capital, Policy and rules

  6. Malaysia depth 1 → 2 priority 1.55 · weight 1.5%

    Have: Depth 1 (Headline): 10/14 measures, no profile, global datasets only; 10 of 14 world dashboard measures; 2 frontier AI site(s) tracked by Epoch; 3 datasets; none for AI compute, Capital, Policy and rules

    • Data-centre projects, contracted GW and load in TNB's system, quarterly (with the Johor share) — Tenaga Nasional Berhad, quarterly analyst briefing (2QFY2026) (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  7. Japan depth 1 → 2 priority 1.3 · weight 1.3%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 6 datasets; none for Capital, Policy and rules

    • Data-centre and chip-fab additions in the 10-year national and area demand forecast — OCCTO, demand forecast (需要想定) page (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  8. Netherlands depth 2 → 3 priority 1.29 · weight 1.3%

    Have: Depth 2 (National profile): profile /europe/countries/netherlands; 19 of 19 measures; 19 of 19 Europe dashboard measures; Data centres' share of electricity (published figure); EU EED data-centre reports; Data-centre grid queue (eu_grid_queues); 23 datasets

  9. Canada depth 1 → 2 priority 0.98 · weight 1%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 5 datasets; none for Capital, Policy and rules

    • Data-centre large-load connection requests in Alberta (provincial, project count and MW) — AESO, Large Load Integration (opened 2026-10-12)
    • Ontario demand outlook including data centres — IESO, Annual Planning Outlook (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  10. Russia depth 1 → 2 priority 0.88 · weight 0.9%

    Have: Depth 1 (Headline): 10/14 measures, no profile, global datasets only; 10 of 14 world dashboard measures; 2 datasets; none for AI use and models, Capital, Policy and rules

  11. India depth 1 → 2 priority 0.83 · weight 0.8%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 5 datasets; none for Capital, Policy and rules

    • National demand forecast with data-centre load (the Electric Power Survey; newer edition expected to add data centres) — Central Electricity Authority (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  12. Singapore depth 1 → 2 priority 0.62 · weight 0.6%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 4 datasets; none for Capital, Policy and rules

  13. Brazil depth 1 → 2 priority 0.56 · weight 0.6%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 4 datasets; none for Capital, Policy and rules

    • Data-centre access requests to the transmission grid (Rede Básica) and the load forecast revision — ONS / EPE / CCEE load forecast revisions; EPE publications (opened 2026-10-12)
    • Press report of ONS's data-centre demand outlook (lead to the primary document) — Agência iNFRA (opened 2026-10-12)
    • A country profile page from the measures already held plus the national source(s) below — StackPath (a /global/countries/<slug> page family) (opened internal)

  14. South Africa depth 1 → 2 priority 0.49 · weight 0.5%

    Have: Depth 1 (Headline): 9/14 measures, no profile, global datasets only; 9 of 14 world dashboard measures; 3 datasets; none for Capital, Policy and rules

  15. Indonesia depth 1 → 2 priority 0.48 · weight 0.5%

    Have: Depth 1 (Headline): 11/14 measures, no profile, global datasets only; 11 of 14 world dashboard measures; 3 datasets; none for Capital, Policy and rules

What we hold, layer by layer, for the 30 largest markets

Number of datasets with rows for each country, per layer of the stack. Empty cells are the gaps.

CountryWeightDepthData centresPowerAI computeAI use and modelsCapitalPolicy and rules
United States63.2%347795
France3.8%2 → 34814653
China3.4%1 → 22344
United Kingdom3.2%2 → 3499454
Germany2.8%351014644
Australia1.6%1 → 223231
United Arab Emirates1.6%1 → 22333
Malaysia1.5%1 → 2232
Japan1.3%1 → 22433
Netherlands1.3%2 → 34812543
Ireland1%42426125214
Canada1%1 → 22234
Russia0.9%1 → 2222
India0.8%1 → 22333
Italy0.7%34710541
Singapore0.6%1 → 22233
Brazil0.6%1 → 22233
South Africa0.5%1 → 22232
Indonesia0.5%1 → 22322
Finland0.5%24810543
Belgium0.5%24610542
Spain0.5%24912543
Portugal0.5%246954
Argentina0.4%1 → 22222
Sweden0.4%2499543
Mexico0.4%1 → 2222
Switzerland0.4%244754
New Zealand0.3%1 → 2222
Poland0.3%2461054
Norway0.3%2497543

Pipeline health: 46 of 93 datasets are within their refresh cadence; 0 are due or stale

Each dataset's newest data against how often its source publishes. "Due" means past its cadence; "stale" means past one and a half cadences. 47 have no stated cadence yet.

Cross-checks: 26 of 58 figures agree with a second source

Where two sources measure the same thing, we set them side by side. A gap is reported, not smoothed over; the note says why the two could differ.

What is checkedSource ASource BGapResult
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Austria
Reporting sites are 179% of mapped sites; the Commission estimates 47 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
14 sites
OpenStreetMap via country_metrics (2026-10-11)
25 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
44%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Belgium
Reporting sites are 48% of mapped sites; the Commission estimates 48 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
40 sites
OpenStreetMap via country_metrics (2026-10-11)
19 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
111%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Bulgaria
Reporting sites are 43% of mapped sites; the Commission estimates 30 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
7 sites
OpenStreetMap via country_metrics (2026-10-11)
3 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
133%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Cyprus
OpenStreetMap maps no site; 2 report to the EU; the Commission estimates 18 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
0 sites
OpenStreetMap via country_metrics (2026-10-11)
2 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
100%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Czechia
Reporting sites are 19% of mapped sites; the Commission estimates 50 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
21 sites
OpenStreetMap via country_metrics (2026-10-11)
4 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
425%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Germany
Reporting sites are 189% of mapped sites; the Commission estimates 456 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
240 sites
OpenStreetMap via country_metrics (2026-10-11)
454 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
47%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Estonia
Reporting sites are 300% of mapped sites; the Commission estimates 12 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
1 sites
OpenStreetMap via country_metrics (2026-10-11)
3 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
67%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, France
Reporting sites are 47% of mapped sites; the Commission estimates 264 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
324 sites
OpenStreetMap via country_metrics (2026-10-11)
152 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
113%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Greece
Reporting sites are 62% of mapped sites; the Commission estimates 20 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
8 sites
OpenStreetMap via country_metrics (2026-10-11)
5 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
60%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Croatia
Reporting sites are 8% of mapped sites; the Commission estimates 19 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
13 sites
OpenStreetMap via country_metrics (2026-10-11)
1 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
1,200%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Hungary
Reporting sites are 33% of mapped sites; the Commission estimates 19 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
21 sites
OpenStreetMap via country_metrics (2026-10-11)
7 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
200%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Ireland
Reporting sites are 20% of mapped sites; the Commission estimates 123 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
88 sites
OpenStreetMap via country_metrics (2026-10-11)
18 sites
EU EED first report, 2023 (via eu_dc_capacity)
389%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Italy
Reporting sites are 69% of mapped sites; the Commission estimates 178 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
61 sites
OpenStreetMap via country_metrics (2026-10-11)
42 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
45%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Lithuania
Reporting sites are 60% of mapped sites; the Commission estimates 18 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
5 sites
OpenStreetMap via country_metrics (2026-10-11)
3 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
67%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Latvia
Reporting sites are 25% of mapped sites; the Commission estimates 25 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
4 sites
OpenStreetMap via country_metrics (2026-10-11)
1 sites
EU EED data-centre dashboard, 2023 (via eu_dc_capacity)
300%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Malta
OpenStreetMap maps no site; 1 report to the EU; the Commission estimates 8 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
0 sites
OpenStreetMap via country_metrics (2026-10-11)
1 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
100%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Netherlands
Reporting sites are 73% of mapped sites; the Commission estimates 192 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
110 sites
OpenStreetMap via country_metrics (2026-10-11)
80 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
38%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Poland
Reporting sites are 163% of mapped sites; the Commission estimates 87 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
27 sites
OpenStreetMap via country_metrics (2026-10-11)
44 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
39%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Portugal
Reporting sites are 53% of mapped sites; the Commission estimates 42 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
15 sites
OpenStreetMap via country_metrics (2026-10-11)
8 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
88%
within 25%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Sweden
Reporting sites are 229% of mapped sites; the Commission estimates 103 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
31 sites
OpenStreetMap via country_metrics (2026-10-11)
71 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
56%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Belgium
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
4 %
Published share via eu_scorecard (2024): BCG, Belgium: the power of compute (via The Brussels Times); estimate; share of national e
2.32 %
Our arithmetic: EED-reported 1.99 TWh (2025) / Ember demand 85.5 TWh
72%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Germany
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
3.9 %
Published share via eu_scorecard (2024): Bitkom/Borderstep, estimate (BDEW 2024 total); estimate; share of gross national electrici
1.33 %
Our arithmetic: EED-reported 6.86 TWh (2025) / Ember demand 516.5 TWh
193%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Spain
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
2.4 %
Published share via eu_scorecard (2024): DNV, via Energías Renovables (REE 2024 total); estimate; share of national demand: over 6
0.23 %
Our arithmetic: EED-reported 0.62 TWh (2024) / Ember demand 275.2 TWh
944%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, France
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
1 %
Published share via eu_scorecard (2024): SDES, French ministry statistics service; measured; share of national electricity consumpt
0.66 %
Our arithmetic: EED-reported 3.17 TWh (2024) / Ember demand 480.6 TWh
52%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Ireland
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
23.2 %
Published share via eu_scorecard (2025): CSO (Ireland), MEC02; measured; share of all metered electricity consumption
3.78 %
Our arithmetic: EED-reported 1.41 TWh (2023) / Ember demand 37.4 TWh
514%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Italy
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
1.9 %
Published share via eu_scorecard (2024): Politecnico di Milano Energy & Strategy, via QualEnergia; estimate; share of national cons
0.28 %
Our arithmetic: EED-reported 0.86 TWh (2025) / Ember demand 312.6 TWh
579%
within 25%?
✗ differ
Data centres' share of electricity: published vs EU-reported, Poland
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
1 %
Published share via eu_scorecard (2025): Warsaw Business Journal (article on the Polish Data Center Association and PSE's plan); es
0.37 %
Our arithmetic: EED-reported 0.64 TWh (2025) / Ember demand 173.7 TWh
170%
within 25%?
✗ differ
Ireland's data-centre electricity: EU-reported vs metered, Ireland
Ireland's EED reports covered 22% of metered data-centre electricity in 2023; Ireland is absent from the EED dashboard for later periods. Not a data error: a reporting gap.
1,412 GWh
EU EED first report, 2023: 18 sites reporting
6,339 GWh
CSO MEC02 metered data-centre electricity, 2023
78%
within 25%?
✗ differ
Germany: data centres mapped vs registered, Germany
The register covers every data centre of 300 kW or more by law; OSM counts what volunteers mapped. The gap is a direct measure of how far OSM under-counts one large European market (it feeds the market weight).
240 sites
OpenStreetMap via country_metrics (2026-10-11)
485 data centres
German data-centre register (RZReg), 2025, via de_dc_register
51%
within 25%?
✗ differ
US frontier AI data-centre power operating now (Epoch AI), United States
One source (Epoch AI's frontier data-centre hub), two columns: world_power sums each site's timeline row dated on or before the pull date; global_context sums Epoch's own 'Current power' field. Both pages show a US figure; they should state which one, or read one shared module (review point 6).
15,179 MW
world_power: Epoch timelines, latest row on or before 2026-10-11
11,851 MW
global_context: Epoch 'Current power (MW)' column (2026-10-11)
28%
within 5%?
✗ differ
Ireland: electricity price for the largest users (band IG, excl. VAT), Ireland
eu_power_prices leads with 2025-S2 because the 2026-S1 release covers only 18 countries (its 2026-S1 figure for Ireland is 0.1654). Different half-years, so two pages show two prices for Ireland; each states its period.
0.17 €/kWh
country_metrics via eu_scorecard: Eurostat band IG (2026-S1)
0.16 €/kWh
eu_power_prices headline (2025-S2)
4%
within 2%?
✗ differ
Top 10 data-centre states, United States
6 of the Census top 10 are in ours (CA, TX, VA, OH, NJ, GA). NAICS 518210 counts business establishments in computing infrastructure, data processing and web hosting, which follow population and offices more than data-centre buildings. Agree = at least 7 of 10.
6 of 10 states
Our OpenStreetMap count (2026-10-11): VA, TX, CA, OR, OH, AZ, WA, NJ, GA, IA
10 of 10 states
Census Bureau, County Business Patterns 2023, establishments in NAICS 518210: CA, TX, FL, NY, PA, VA, OH, IL, NJ, GA
-40%
within 30%?
✗ differ
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Denmark
Reporting sites are 118% of mapped sites; the Commission estimates 58 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
22 sites
OpenStreetMap via country_metrics (2026-10-11)
26 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
15%
within 25%?
✓ agree
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Spain
Reporting sites are 95% of mapped sites; the Commission estimates 165 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
40 sites
OpenStreetMap via country_metrics (2026-10-11)
38 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
5%
within 25%?
✓ agree
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Finland
Reporting sites are 122% of mapped sites; the Commission estimates 72 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
23 sites
OpenStreetMap via country_metrics (2026-10-11)
28 sites
EU EED data-centre dashboard, 2024 (via eu_dc_capacity)
18%
within 25%?
✓ agree
Data-centre sites: mapped in OpenStreetMap vs reporting to the EU, Luxembourg
Reporting sites are 100% of mapped sites; the Commission estimates 14 sites in scope. OSM counts volunteer-mapped buildings of any size; the EED covers sites of 500 kW or more that reported. More reporters than mapped sites means OSM under-maps the country; far fewer means reporting is incomplete.
9 sites
OpenStreetMap via country_metrics (2026-10-11)
9 sites
EU EED data-centre dashboard, 2025 (via eu_dc_capacity)
0%
within 25%?
✓ agree
Data centres' share of electricity: published vs EU-reported, Denmark
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
2.8 %
Published share via eu_scorecard (2024): Danish Energy Agency, Klimastatus og -fremskrivning 2026 (consultation draft); estimate; s
2.81 %
Our arithmetic: EED-reported 1.18 TWh (2025) / Ember demand 41.9 TWh
0%
within 25%?
✓ agree
Data centres' share of electricity: published vs EU-reported, Finland
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
1.9 %
Published share via eu_scorecard (2024): Finnish Government, national data-centre roadmap (rapporteur's report 2025:94); estimate;
1.57 %
Our arithmetic: EED-reported 1.38 TWh (2024) / Ember demand 87.8 TWh
21%
within 25%?
✓ agree
Data centres' share of electricity: published vs EU-reported, Netherlands
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
4.6 %
Published share via eu_scorecard (2024): CBS (Statistics Netherlands); measured; share of total electricity consumption; provisiona
4.69 %
Our arithmetic: EED-reported 5.68 TWh (2025) / Ember demand 121.0 TWh
2%
within 25%?
✓ agree
Data centres' share of electricity: published vs EU-reported, Sweden
The EED figure covers only sites that reported, so it is a floor: well below the published share means incomplete reporting, above it would question the published share. Years and bases can differ.
2.1 %
Published share via eu_scorecard (2022): Swedish Energy Agency (RISE), ER 2023:04; estimate; share of demand: midpoint of 2.8-3.2 T
2.15 %
Our arithmetic: EED-reported 2.93 TWh (2025) / Ember demand 136.5 TWh
2%
within 25%?
✓ agree
Ireland: data centres' share of electricity, Ireland
Different years and bases: the CSO divides metered data-centre use by all metered consumption; Ember/ICIS estimate a share of national demand. Growth between the two years explains part of the gap.
23.2 %
CSO, Data Centres Metered Electricity Consumption 2025 (2025)
19 %
Ember, Grids for data centres in Europe (2024)
22%
within 25%?
✓ agree
Ireland: data centres mapped vs our campus directory, Ireland
Different units: OSM counts mapped buildings or sites (a campus of several named halls counts several times); the directory chains planning, EPA, valuation and OSM points into campuses, including proposed and lapsed ones. This calibrates how far OSM counts can be trusted where we have no directory.
88 sites
OpenStreetMap via country_metrics (2026-10-11)
79 campuses
StackPath directory (2026-10-10): 55 operating
11%
within 25%?
✓ agree
Germany: data centres in the register vs reporting to the EU, Germany
Germany transposed the EED through its Energy Efficiency Act: the same operators file both, so the counts should be close.
485 data centres
RZReg 2025 (de_dc_register)
454 sites
EU EED dashboard 2025 (eu_dc_capacity)
7%
within 10%?
✓ agree
Germany: data-centre electricity, register vs EU-reported, Germany
Same operators, two filings: a small gap is timing (late filers) or rounding in the register's published classes.
7.2 TWh
RZReg 2025 (de_dc_register)
6.86 TWh
EU EED dashboard 2025 (eu_dc_capacity)
5%
within 10%?
✓ agree
TSMC monthly revenue, Taiwan
45 months in both datasets; 0 differ by more than 0.5%. Same company figure, two pipelines: one should own the source (review point 6).
511.9 NT$ bn
ai_supply_chain: TSMC 6-K filings at SEC (2026-09)
511.9 NT$ bn
taiwan_monthly: MOPS (2026-09)
0%
within 1%?
✓ agree
Taiwan server makers' monthly revenue (sum of the companies in both), Taiwan
270 company-months in both datasets; 0 differ by more than 0.5%. Two pipelines pull the same exchange filings (review point 6).
2,199.1 NT$ bn
ai_compute_added: TWSE/MOPS (2026-09)
2,199.1 NT$ bn
taiwan_monthly: MOPS (2026-09)
0%
within 1%?
✓ agree
Ireland: electricity price for the largest users (band IG, excl. VAT), Ireland
Same Eurostat table (nrg_pc_205) read by two pipelines; they should match exactly when both hold the same half-year.
0.17 €/kWh
country_metrics via eu_scorecard: Eurostat band IG (2026-S1)
0.17 €/kWh
eurostat_prices (2026-S1, MWH_GE150000)
0%
within 2%?
✓ agree
Data centres mapped (OpenStreetMap), United States
Same OpenStreetMap run and duplicate rule; a gap would mean US sites falling outside every state outline (none did).
1,612 sites
Sum of our 51 state counts (us-atlas outlines)
1,612 sites
country_metrics USA count (Natural Earth outline)
0%
within 1%?
✓ agree
Frontier AI data-centre power operating now, United States
Same Epoch pull and summing rule; small gaps come from per-site rounding; every US site has a state.
15,179 MW
Sum of our state totals (Epoch sites placed in states)
15,179 MW
world_power, Epoch by country (pulled 2026-10-11)
0%
within 1%?
✓ agree
Frontier AI data-centre power planned, United States
67 sites placed from Epoch's address field, 3 from the site name, 4 from the sources Epoch lists, 3 from a source Epoch cites (checked by hand, research/datasets/us_states.json).
43,412 MW
Sum of our state totals
43,412 MW
world_power, Epoch by country (pulled 2026-10-11)
0%
within 1%?
✓ agree
Population, United States
An exact match means the World Bank uses the Census estimate for the same year: a check that both are the same vintage, not two independent counts. Census excludes Puerto Rico and other territories.
341,784,857 people
Census Bureau, Vintage 2025, sum of 50 states + DC
341,784,857 people
World Bank SP.POP.TOTL (via country_metrics)
0%
within 1%?
✓ agree
Top 10 data-centre states, United States
8 of EPRI's top 10 are in our top 10 (VA, TX, CA, OR, AZ, IA, GA, WA). EPRI ranks electricity use in 2023, we rank mapped sites in 2026: a state with a few huge campuses ranks higher on power than on site count, and OpenStreetMap misses some hyperscale sites. Agree = at least 7 of 10.
8 of 10 states
Our OpenStreetMap count (2026-10-11): VA, TX, CA, OR, OH, AZ, WA, NJ, GA, IA
10 of 10 states
EPRI, Powering Intelligence (May 2024), states ranked by 2023 data-centre electricity use: VA, TX, CA, IL, OR, AZ, IA, GA, WA, PA
-20%
within 30%?
✓ agree
Industrial electricity price, United States
Both read the same EIA table; a gap would mean the caches are from different releases.
9.03 US¢/kWh
This pack: EIA EPM 5.6.B U.S. Total, January–July 2026
9.03 US¢/kWh
world_power / us_power: EIA industrial average, July 2026 YTD (commercial 13.97¢)
0%
within 1%?
✓ agree
Industrial electricity price, United States
Two EIA series for different periods (this year to date against last full year); prices rose in between.
9.03 US¢/kWh
EIA Electric Power Monthly 5.6.B, January–July 2026
8.51 US¢/kWh
EIA-861 annual, 2025 (revenue ÷ sales)
6%
within 15%?
✓ agree
Electricity generated, 2025, United States
Ember's release (2026-08-19) came before EIA's final 2025 data (September 2026). Solar explains 89 of the 95 TWh gap (Ember 389 TWh against EIA's utility-scale 300 TWh), consistent with Ember adding small-scale (rooftop) solar, which this EIA table leaves out (our reading).
4,519.8 TWh
Ember US electricity data (last updated 2026-08-19)
4,425 TWh
EIA, net generation by state, Total Electric Power Industry (released September 2026)
2%
within 3%?
✓ agree
Generating capacity under construction (2026-08), United States
Same EIA file and status rule; units outside the 50 states and DC are in the US total but not in a state.
98,731 MW
Sum of our 51 state totals (EIA-860M Planned sheet, Plant State)
98,731 MW
us_power, EIA-860M US total under construction
0%
within 1%?
✓ agree
Data centres' share of electricity, Virginia
Different denominators: EPRI estimates all data-centre use in the state; Dominion reports one utility's sales. Dominion's later 10-Ks give 26% for 2024 and 28% for 2025.
25.6 %
EPRI, Powering Intelligence (May 2024), 2023 estimate, % of Virginia's electricity consumption
24 %
Dominion Energy 10-K for 2023: data centres' share of Virginia Power's electricity sales
7%
within 25%?
✓ agree
US data-centre electricity use, 2023, United States
Two independent estimates with different methods. EPRI's Feb 2026 update puts 2024 at 184 TWh.
152.1 TWh
EPRI, Powering Intelligence (May 2024): about 4% of US load
176 TWh
LBNL for DOE (Dec 2024): about 4.4% of US electricity
-14%
within 25%?
✓ agree

More dashboards

Dashboard

Europe

31 countries: data centres, power, AI chips and AI use

Dashboard

Ireland by council

31 councils: grid draw, campuses, planning decisions, on-site power and rates

Dashboard

World

Every country: data centres, AI compute, power and its price

Dashboard

US states

50 states and DC, from public and third-party sources, cross-checked

Every view keeps its measure and layout in the address bar: copy the link and the reader opens the same view. Data build 2026-10-10. How we know · corrections.