Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers

Mongird, Kendall; Burleyson, Casey; Akdemir, Kerem Ziya; Thurber, Travis; Rice, Jennie

doi:10.57931/2572012

Title: Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers

Dataset
Other Related Research

Abstract

Projections of the need for new data centers to support Artificial Intelligence (AI) are large but highly uncertain. Recent projections indicate up to a 15% annual growth rate in data center electricity demand within the next 5-10 years. Given that most electric utilities are required to have a reserve margin of roughly the same magnitude as the projected growth in demand, these new data center loads could soon threaten resource adequacy and reliability unless data centers build their own generation, interruptible loads are negotiated, commensurate new capacity and/or transmission is built, or some combination of these options. Similarly, depending on the cooling technology and geographic location of new data centers, they could threaten water adequacy in water scarce regions. This presentation highlights the grid and water implications of new data centers to support AI.

Authors:

; Burleyson, Casey; Akdemir, Kerem Ziya; Thurber, Travis; Rice, Jennie

Publication Date:: Mon Jun 16 00:00:00 EDT 2025

DOE Contract Number:: AC05-76RL01830

Research Org.:: Pacific Northwest National Lab (United States)

Sponsoring Org.:: USDOE Office of Science (SC), Biological and Environmental Research (BER)

OSTI Identifier:: 2572012

DOI:: https://doi.org/10.57931/2572012

Citation Formats


                    Mongird, Kendall, Burleyson, Casey, Akdemir, Kerem Ziya, Thurber, Travis, and Rice, Jennie. Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers.  United States: N. p., 2025. 
        Web.  doi:10.57931/2572012.

Copy to clipboard


                    Mongird, Kendall, Burleyson, Casey, Akdemir, Kerem Ziya, Thurber, Travis, & Rice, Jennie. Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers.  United States.  doi:https://doi.org/10.57931/2572012

Copy to clipboard


                    Mongird, Kendall, Burleyson, Casey, Akdemir, Kerem Ziya, Thurber, Travis, and Rice, Jennie. 2025.  
"Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers".  United States.  doi:https://doi.org/10.57931/2572012.  https://www.osti.gov/servlets/purl/2572012. Pub date:Mon Jun 16 00:00:00 EDT 2025

Copy to clipboard


                    
@article{osti_2572012,

  title        = {Energy and AI: Evaluating Future Grid and Water Stress Due to Data Centers},

  author       = {Mongird, Kendall and Burleyson, Casey and Akdemir, Kerem Ziya and Thurber, Travis and Rice, Jennie},

  abstractNote = {Projections of the need for new data centers to support Artificial Intelligence (AI) are large but highly uncertain. Recent projections indicate up to a 15% annual growth rate in data center electricity demand within the next 5-10 years. Given that most electric utilities are required to have a reserve margin of roughly the same magnitude as the projected growth in demand, these new data center loads could soon threaten resource adequacy and reliability unless data centers build their own generation, interruptible loads are negotiated, commensurate new capacity and/or transmission is built, or some combination of these options. Similarly, depending on the cooling technology and geographic location of new data centers, they could threaten water adequacy in water scarce regions. This presentation highlights the grid and water implications of new data centers to support AI.},

  doi          = {10.57931/2572012},

  journal      = {},

  number       = ,

  volume       = ,

  place        = {United States},

  year         = {Mon Jun 16 00:00:00 EDT 2025},

  month        = {Mon Jun 16 00:00:00 EDT 2025}

}

Copy to clipboard

Dataset:

View Dataset

DOI: https://doi.org/10.57931/2572012

Save / Share:

Export Metadata

Save to My Library

Similar records in DOE Data Explorer and OSTI.GOV collections:

Implications of AI and Data Center Service Demand on Energy, Water, and Economic Systems

Dataset Morris, Stephanie ; Wild, Thomas B. ; Zhao, Mengqi ; ...

This talk focuses on the challenges posed by the rapid growth of data centers (DCs), which rely heavily on electricity, water, land, and mineral resources, and have significant interactions with economic growth. The U.S. leads in DC deployment, but understanding how resource demands and constraints impact DC expansion requires strategic multisector analysis. To address this, the talk introduces a new conceptual framework and modeling capability within GCAM-USA, allowing integrated analysis of DC service demand, electricity, and direct and indirect water needs under varied scenarios. Insights include state-level water and electricity trends for DCs, uncertainty bounds, and the broader implications formore » resources and economics.« less
Energy-Water Dynamics of Data Centers and Their Consequences for the Electricity Grid and Water Availability

Dataset Burleyson, Casey ; Akdemir, Kerem Ziya ; Mongird, Kendall ; ...

Briefing slide deck on IM3's work on the energy and water impacts of data centers. Briefing was given to BER program managers on 19-Mar and 26-Mar 2025.
IM3 Data Center Driven Grid Stress Dataset for the U.S. Western Interconnection

Dataset Akdemir, Kerem Ziya ; Burleyson, Casey ; Mongird, Kendall ; ...

This dataset provides projected grid stress and reliability results (including all model inputs and outputs from an open-source grid operations modeling framework - GO), for the Integrated Multisector, Multiscale Modeling (IM3) project, under varying levels of data center demand growth between 2025 and 2035 in the U.S. Western Interconnection. The scenarios and sensitivity experiments are combinations of different data center demand growth rates and energy, weather, population and economic pathways. Data center demand growth projections were sourced from the Electric Power Research Institute (EPRI). The data center demand growth projection names are: Low (3.71% annual data center demand growth) Moderate (5% annual data center demand growth) High (10% annual data center demand growth) Higher (15% annual data center demand growth) Energy, weather, population and economic pathways are informed by two Shared Socioeconomic Pathways (SSP3 and SSP5) and two Representative Concentration Pathways (RCP4.5 and RCP8.5) following the hotter general circulation model (GCM) forcing group from a set of perturbed thermodynamics simulations. The resulting pathway names are: rcp45hotter_ssp3 rcp45hotter_ssp5 rcp85hotter_ssp3 rcp85hotter_ssp5 The main scenarios and sensitivity experiments are detailed below. Reference scenario: The projected grid stress and reliability results for the U.S. Western Interconnection from a previous study. This scenario does not consider data center demand growth explicitly. Data center scenario: Building on the reference scenario, this scenario considers various data center growth rates and how they impact the U.S. Western Interconnection. Data center loads are modeled as flat 8760-hr profiles. This scenario does not consider new generation and transmission capacities specifically designed to meet the new data center demands. The related folder is named "flat". Delayed generator retirements sensitivity experiment: Building on the data center scenario, this experiment explores the impact of different levels of natural gas and nuclear generator retirement delays. The resulting scenario names are: (1) postponing 100% nuclear retirements; (2) postponing 100% nuclear and 25% natural gas retirements; (3) postponing only 50% natural gas retirements; (4) postponing 100% nuclear and 50% natural gas retirements; (5) postponing 100% nuclear and 75% natural gas retirements; and (6) postponing 100% nuclear and 100% natural gas retirements. The related folder names are: no_gen_retire_0_gas, no_gen_retire_25_gas, no_gen_retire_50_gas, no_gen_retire_50_gas_only, no_gen_retire_75_gas, and no_gen_retire_100_gas. Demand response through curtailment sensitivity experiment: Building on the data center scenario, this experiment explores the impact of different participation and compensation levels of data center demand response. The resulting scenario names are: (1) 5% demand available for curtailment with 750more »« less
Data for "Evaluating Advanced Nuclear Fission Technologies for Future Decarbonized Power Grids"

Dataset Cano Renteria, Emilio ; Schwartz, Jacob ; Jenkins, Jesse

This dataset includes input and output data for the GenX runs, as well as GenX postprocessing and plotting scripts, for the paper "Evaluating Advanced Nuclear Fission Technologies for Future Decarbonized Power Grids".
MSD CoP Webinar: Energy and AI

Dataset Burleyson, Casey ; Morris, Stephanie T. ; Mongird, Kendall

Context: This webinar was hosted by the MultiSector Dynamics Community of Practice (MSD CoP; https://multisectordynamics.org). Abstract: Projections of the need for new data centers to support Artificial Intelligence (AI) are large but highly uncertain. Recent projections indicate up to a 15% annual growth rate in data center electricity demand within the next 5-10 years. Given that most electric utilities are required to have a reserve margin of roughly the same magnitude as the projected growth in demand, these new data center loads could soon threaten resource adequacy and reliability unless data centers build their own generation, interruptible loads aremore »« less

Similar Records