AI carbon footprint in US South Central - Texas: 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 South Central - Texas?
A 1210-token query to a class B model (gpt-4o) served from US South Central - Texas emits about 0.043 g CO₂e, and a million tokens emit 0.0312 kg CO₂e.
Rank 12 of 16 for class B, lowest carbon first (full table). 4.4x the level of EU North - Sweden, Asia Pacific - Singapore is 1.3x higher.
Which grid supplies the electricity in US South Central - Texas?
US South Central - Texas maps to ERCOT All, a US eGRID subregion, at 375.3 g CO₂e per kWh for 2023. That is 1.6 times the EU-27 average (236 g) and 0.8 times the global average (483.2 g). The source is eGRID2023 ERCT 736.6 lb/MWh CO2e total output x 1.1232 (Ember lifecycle basis).
Cloud regions on this grid, 2 in all (1 Azure, 1 Google Cloud).
| Provider | Region id | Region name | Location | Mapping confidence | Mapping note |
|---|---|---|---|---|---|
| Azure | southcentralus | South Central US | Texas | high | Provider states state only; Texas spans ERCT/SPSO/SRMV/AZNM. San Antonio-area site assumed; polygon test at San Antonio (29.42,-98.49) = ERCT. CONFIRMED 2026-10-14 by Data Center Map (datacentermap.com, operator pages for Microsoft and Amazon AWS, read 14 October 2026): Microsoft SAT campus, San Antonio, and Castroville campus, Castroville TX (24 listings); polygon test at San Antonio (29.42,-98.49) and Castroville (29.36,-98.88) = ERCT; agrees with the assumed site, confidence raised from medium to high. |
| Google Cloud | us-south1 | us-south1 | Dallas, Texas | high | Subregion settled in paper (anchor). Polygon test at Dallas (32.78,-96.80) = ERCT. |
Which grids have an intensity close to ERCOT All?
| Difference g | Grid | Code | g CO₂e per kWh | Ratio |
|---|---|---|---|---|
| -2.5 | Tanzania, the United Republic of | TZA | 372.8 | 0.993 |
| 2.6 | SERC Mississippi Valley | SRMV | 377.9 | 1.007 |
| -11.7 | Germany | DEU | 363.6 | 0.969 |
| -15.4 | Belarus | BLR | 359.9 | 0.959 |
| -15.5 | WECC Southwest | AZNM | 359.8 | 0.959 |
| 17.1 | Zimbabwe | ZWE | 392.4 | 1.046 |
| 17.6 | United States of America | USA | 392.9 | 1.047 |
| 18.8 | Reunion | REU | 394.1 | 1.050 |
| 20.2 | Estonia | EST | 395.5 | 1.054 |
| -20.7 | Rwanda | RWA | 354.6 | 0.945 |
Emission factors for US South Central - Texas 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) | 0.00817 | 0.00027 | 0.00014 | 0.00060 | 0.009 | 0.013 | 0.032 | 0.045 |
| Class B (mid-size) | 0.02675 | 0.00101 | 0.00046 | 0.00293 | 0.031 | 0.043 | 0.106 | 0.15 |
| Class C (frontier) | 0.09962 | 0.00388 | 0.00169 | 0.01284 | 0.118 | 0.16 | 0.393 | 0.54 |
| Model class | Central | Class range low | Class range high | High case | B200 | Development |
|---|---|---|---|---|---|---|
| Class A | 0.009181 | 0.005713 | 0.01326 | 0.01463 | 0.008066 | 0.01114 |
| Class B | 0.03115 | 0.02028 | 0.03766 | 0.04902 | 0.02205 | 0.03648 |
| Class C | 0.118 | 0.04801 | 0.1753 | 0.1846 | 0.1213 | 0.1359 |
How much does siting matter for a model served from US South Central - Texas?
Electricity is 85.9% of the class B total here (class A 89%, class C 84.4%), with hardware at 3.2%, construction 1.5% and training 9.4%. Electricity carries most of the 0.0312 kg, so the grid is the biggest lever.
The example company (2,323 million class B tokens a year) emits 72 kg CO₂e from US South Central - Texas. Moving to EU Central - Germany would save 1.9 kg (2.7%) or US West - Oregon would save 8.5 kg (12%).
| Region | Grid g per kWh | Class A | Class B | Class C |
|---|---|---|---|---|
| EU Central - Germany | 363.6 | 0.591 | 1.94 | 7.21 |
| US West - Oregon | 323.7 | 2.61 | 8.55 | 31.8 |
How much water does AI usage consume in US South Central - Texas?
A class B query of 1210 tokens from US South Central - Texas consumes about 0.15 mL of water and a million tokens 0.106 litres, or 0.1 litres without the water attributed to hydropower.
| Input | Value |
|---|---|
| On-site WUE, L per kWh | 0.292 |
| WUE denominator | total site electricity |
| WUE basis as published | withdrawal |
| Withdrawal to consumption | 0.6667 |
| EWIF, L per kWh | 1.287 |
| EWIF without hydro, L per kWh | 1.213 |
| AWARE factor | 13.56 |
| Confidence | derived-site |
| Source | Microsoft; San Antonio (TX) (site); FY25 |
| Model class | On site | Grid electricity | Total | Total without hydro | AWARE-weighted on site |
|---|---|---|---|---|---|
| Class A | 0.00423 | 0.028 | 0.0322 | 0.0306 | 0.0574 |
| Class B | 0.0139 | 0.0917 | 0.106 | 0.1 | 0.188 |
| Class C | 0.0516 | 0.342 | 0.393 | 0.374 | 0.7 |
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 "0.0312 kg CO₂e per million tokens, class B, US South Central - Texas, SAEF, 19 October 2026" and link somaai.earth/ai-footprint/factors/us-texas.