AI carbon footprint in Netherlands: emissions per token and per query
Factor version soma-ai-ef-2026.10-dataset-v3.0.0. Last updated . Location-based grid averages, not renewable contracts. Estimates from published benchmarks.
Part of the SOMA AI Emission Factors cluster.
How much CO₂ does an AI query emit when served from Netherlands?
A 1210-token query to a class B model (gpt-4o) served from Netherlands emits about 0.032 g CO₂e, and a million tokens emit 0.0235 kg CO₂e.
Rank 7 of 16 for class B, lowest carbon first (full table). 3.3x the level of EU North - Sweden, Asia Pacific - Singapore is 1.7x higher.
Which grid supplies the electricity in Netherlands?
Netherlands maps to Netherlands, a national grid, at 268.2 g CO₂e per kWh for 2023. That is 1.1 times the EU-27 average (236 g) and 0.6 times the global average (483.2 g). The source is Ember yearly electricity data, CO2 intensity 2023, 268.17 g/kWh (lifecycle, IPCC AR5).
Cloud regions on this grid, 2 in all (1 Azure, 1 Google Cloud).
| Provider | Region id | Region name | Location | Mapping confidence | Mapping note |
|---|---|---|---|---|---|
| Azure | westeurope | West Europe | Netherlands | high | |
| Google Cloud | europe-west4 | europe-west4 | Eemshaven, Netherlands | high |
Which grids have an intensity close to Netherlands?
| Difference g | Grid | Code | g CO₂e per kWh | Ratio |
|---|---|---|---|---|
| -2.0 | ASCC Miscellaneous | AKMS | 266.2 | 0.993 |
| -2.0 | Panama | PAN | 266.2 | 0.993 |
| -5.8 | Armenia | ARM | 262.4 | 0.978 |
| 8.5 | NPCC New England | NEWE | 276.7 | 1.032 |
| -9.8 | Colombia | COL | 258.4 | 0.963 |
| 10.1 | Fiji | FJI | 278.3 | 1.038 |
| 14.3 | Ireland | IRL | 282.5 | 1.053 |
| -18.2 | Cook Islands (the) | COK | 250 | 0.932 |
| -23.3 | French Guiana | GUF | 244.9 | 0.913 |
| 23.6 | Peru | PER | 291.8 | 1.088 |
Emission factors for Netherlands by model class
| Model class | Electricity kg | Hardware kg | Construction kg | Training kg | Total kg CO₂e | g per query | Water L | mL per query |
|---|---|---|---|---|---|---|---|---|
| Class A (small) | 0.00584 | 0.00027 | 0.00014 | 0.00060 | 0.007 | 0.0094 | 0.076 | 0.1 |
| Class B (mid-size) | 0.01912 | 0.00101 | 0.00046 | 0.00293 | 0.024 | 0.032 | 0.247 | 0.34 |
| Class C (frontier) | 0.07119 | 0.00388 | 0.00169 | 0.01284 | 0.090 | 0.12 | 0.921 | 1.3 |
| Model class | Central | Class range low | Class range high | High case | B200 | Development |
|---|---|---|---|---|---|---|
| Class A | 0.006849 | 0.00432 | 0.009822 | 0.01077 | 0.006046 | 0.007961 |
| Class B | 0.02352 | 0.01557 | 0.02821 | 0.03641 | 0.01697 | 0.02607 |
| Class C | 0.08961 | 0.03858 | 0.1313 | 0.1376 | 0.09197 | 0.09708 |
How much does siting matter for a model served from Netherlands?
Electricity is 81.3% of the class B total here (class A 85.2%, class C 79.4%), with hardware at 4.3%, construction 1.9% and training 12%. Electricity carries most of the 0.0235 kg, so the grid is the biggest lever.
The example company (2,323 million class B tokens a year) emits 55 kg CO₂e from Netherlands. Moving to EU West - UK (London) would save 5.4 kg (9.9%) or EU South - Spain would save 16 kg (30%).
| Region | Grid g per kWh | Class A | Class B | Class C |
|---|---|---|---|---|
| EU West - UK (London) | 235.6 | 1.65 | 5.39 | 20.1 |
| EU South - Spain | 169.7 | 4.98 | 16.3 | 60.7 |
How much water does AI usage consume in Netherlands?
A class B query of 1210 tokens from Netherlands consumes about 0.34 mL of water and a million tokens 0.247 litres, or 0.247 litres without the water attributed to hydropower.
| Input | Value |
|---|---|
| On-site WUE, L per kWh | 0.0356 |
| WUE denominator | total site electricity |
| WUE basis as published | withdrawal |
| Withdrawal to consumption | 0.6667 |
| EWIF, L per kWh | 3.445 |
| EWIF without hydro, L per kWh | 3.445 |
| AWARE factor | 1.89 |
| Confidence | derived-site |
| Source | Microsoft; Hollands Kroon, Netherlands (site); FY25 |
| Model class | On site | Grid electricity | Total | Total without hydro | AWARE-weighted on site |
|---|---|---|---|---|---|
| Class A | 0.000517 | 0.075 | 0.0755 | 0.0755 | 0.000976 |
| Class B | 0.00169 | 0.246 | 0.247 | 0.247 | 0.0032 |
| Class C | 0.0063 | 0.915 | 0.921 | 0.921 | 0.0119 |
Where do these numbers come from and how do I cite them?
Figures per query are one plain exchange, before the Tier 2b reasoning uplift. Electricity term. GPU energy per token from the ML.ENERGY Leaderboard v3, a facility overhead of 1.20, and grid carbon intensity from EPA eGRID 2023 and Ember 2023. Embodied hardware, data-centre construction and amortised training terms are documented on the methodology page. Water combines on-site cooling water (Li et al. 2025) with the water embedded in grid electricity (Reig et al. 2020). The modelled range of the models inside a class, around the class value, is −41% to +48% for class A, −39% to +24% for class B and −67% to +55% for class C.
Dataset: AI Inference Emission and Resource Factors for Corporate GHG Inventories (SOMA AI Emission Factors (SAEF)), Llopis-Garcia, G., SOMA AI, 19 October 2026, CC BY 4.0, concept DOI 10.5281/zenodo.20443585, which resolves to the latest version (this version 10.5281/zenodo.23371447, earlier version 2 values at 10.5281/zenodo.22767475). Method: Accounting for Purchased AI Inference in Corporate GHG Inventories: A Token-Based Method, arXiv:2606.10660, under peer review.
Partners can fetch any row from the AI factors API, and the whole table is a CSV download.
Cite as "0.0235 kg CO₂e per million tokens, class B, Netherlands, SAEF, 19 October 2026" and link somaai.earth/ai-footprint/factors/eu-netherlands.