AI carbon footprint in US West - Oregon: 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 US West - Oregon?
A 1210-token query to a class B model (gpt-4o) served from US West - Oregon emits about 0.038 g CO₂e, and a million tokens emit 0.0275 kg CO₂e.
Rank 10 of 16 for class B, lowest carbon first (full table). 3.8x the level of EU North - Sweden, Asia Pacific - Singapore is 1.5x higher.
Which grid supplies the electricity in US West - Oregon?
US West - Oregon maps to WECC Northwest, a US eGRID subregion, at 323.7 g CO₂e per kWh for 2023. That is 1.4 times the EU-27 average (236 g) and 0.7 times the global average (483.2 g). The source is eGRID2023 NWPP 635.3 lb/MWh CO2e total output x 1.1232 (Ember lifecycle basis).
Cloud regions on this grid, 4 in all (1 AWS, 1 Azure, 2 Google Cloud).
| Provider | Region id | Region name | Location | Mapping confidence | Mapping note |
|---|---|---|---|---|---|
| AWS | us-west-2 | US West (Oregon) | Oregon | high | Subregion settled in paper (anchor). Polygon test at Boardman and Hillsboro OR also NWPP. |
| Azure | westus2 | West US 2 | Washington | high | Provider states state only; Quincy/Moses Lake area assumed; polygon test at Quincy (47.23,-119.85) = NWPP. CONFIRMED 2026-10-14 by Data Center Map (datacentermap.com, operator pages for Microsoft and Amazon AWS, read 14 October 2026): Microsoft Columbia campus, Quincy WA (16 listings) and Wenatchee WA (11); polygon test at Quincy (47.23,-119.85) and Wenatchee (47.42,-120.31) = NWPP; agrees with the assumed site, confidence raised from medium to high. |
| Google Cloud | us-west1 | us-west1 | The Dalles, Oregon | high | Polygon test at The Dalles (45.60,-121.18) = NWPP. |
| Google Cloud | us-west3 | us-west3 | Salt Lake City, Utah | high | Polygon test at Salt Lake City (40.76,-111.89) = NWPP (Utah is WECC Northwest in eGRID2023). |
The note for AWS us-west-2 reads "Subregion settled in paper (anchor). Polygon test at Boardman and Hillsboro OR also NWPP.".
Which grids have an intensity close to WECC Northwest?
| Difference g | Grid | Code | g CO₂e per kWh | Ratio |
|---|---|---|---|---|
| -0.4 | Italy | ITA | 323.3 | 0.999 |
| 4.6 | Honduras | HND | 328.3 | 1.014 |
| -5.8 | Suriname | SUR | 317.9 | 0.982 |
| 12.9 | Greece | GRC | 336.6 | 1.040 |
| -13.8 | Guatemala | GTM | 309.9 | 0.957 |
| 14.6 | Bulgaria | BGR | 338.3 | 1.045 |
| -18.4 | RFC East | RFCE | 305.3 | 0.943 |
| -18.9 | Chile | CHL | 304.8 | 0.942 |
| -19.9 | SERC Virginia/Carolina | SRVC | 303.8 | 0.939 |
| 22.5 | Korea (the Democratic People's Republic of) | PRK | 346.2 | 1.070 |
Emission factors for US West - Oregon 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) | 0.00705 | 0.00027 | 0.00014 | 0.00060 | 0.008 | 0.011 | 0.208 | 0.29 |
| Class B (mid-size) | 0.02307 | 0.00101 | 0.00046 | 0.00293 | 0.027 | 0.038 | 0.682 | 0.94 |
| Class C (frontier) | 0.08592 | 0.00388 | 0.00169 | 0.01284 | 0.104 | 0.14 | 2.540 | 3.5 |
| Model class | Central | Class range low | Class range high | High case | B200 | Development |
|---|---|---|---|---|---|---|
| Class A | 0.008057 | 0.005042 | 0.0116 | 0.01277 | 0.007093 | 0.009608 |
| Class B | 0.02747 | 0.01801 | 0.03311 | 0.04294 | 0.01961 | 0.03146 |
| Class C | 0.1043 | 0.04347 | 0.1541 | 0.162 | 0.1072 | 0.1172 |
How much does siting matter for a model served from US West - Oregon?
Electricity is 84% of the class B total here (class A 87.4%, class C 82.3%), with hardware at 3.7%, construction 1.7% and training 11%. Electricity carries most of the 0.0275 kg, so the grid is the biggest lever.
The example company (2,323 million class B tokens a year) emits 64 kg CO₂e from US West - Oregon. Moving to US East - N. Virginia would save 3.3 kg (5.2%) or EU West - Ireland would save 6.8 kg (11%).
| Region | Grid g per kWh | Class A | Class B | Class C |
|---|---|---|---|---|
| US East - N. Virginia | 303.8 | 1 | 3.29 | 12.3 |
| EU West - Ireland | 282.5 | 2.08 | 6.82 | 25.4 |
How much water does AI usage consume in US West - Oregon?
A class B query of 1210 tokens from US West - Oregon consumes about 0.94 mL of water and a million tokens 0.682 litres, or 0.046 litres without the water attributed to hydropower.
| Input | Value |
|---|---|
| On-site WUE, L per kWh | 0.12 |
| WUE denominator | IT load |
| WUE basis as published | withdrawal |
| Withdrawal to consumption | 0.6667 |
| EWIF, L per kWh | 9.500 |
| EWIF without hydro, L per kWh | 0.579 |
| AWARE factor | 3.91 |
| Confidence | published-site |
| Source | AWS; AWS Region U.S. West (Oregon) us-west-2; 2025 |
| Model class | On site | Grid electricity | Total | Total without hydro | AWARE-weighted on site |
|---|---|---|---|---|---|
| Class A | 0.00145 | 0.207 | 0.208 | 0.0141 | 0.00568 |
| Class B | 0.00475 | 0.677 | 0.682 | 0.046 | 0.0186 |
| Class C | 0.0177 | 2.52 | 2.54 | 0.171 | 0.0692 |
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 "0.0275 kg CO₂e per million tokens, class B, US West - Oregon, SAEF, 19 October 2026" and link somaai.earth/ai-footprint/factors/us-west.