Optimizing Data Center Siting
Considering Local Infrastructure, Environmental, and Community Constraints
With the increase in artificial intelligence (AI) use and cloud computing systems, demand for computational resources provided by data centers is growing. The United States has more than 4,000 data centers, one-third of which are in Virginia, Texas, and California. Yet as the sector grows, the locations considered for data centers are expanding rapidly. Powering and cooling these systems impose substantial demands on the energy grid and water supply. Potential impacts to water and energy infrastructure, as well as other community resources, make the proper siting of data centers essential.
Skip ahead to tool ↓Why Use This Tool?
This project provides local and state government decisionmakers and other industry stakeholders with data and visualizations to compare trade-offs among locations for regional data centers. The Data Center Siting Model and Suitability Mapper for Community Planners helps identify suitable sites and align development with other community and industry needs. By integrating results from our resource optimization model with other local factors, the tool supports balanced decisions on power, water, broadband, and land use for sustainable data center siting and planning.
Government stakeholders may find this tool helpful to evaluate data center permit applications and to prioritize where to offer incentives. Industry stakeholders may find it useful to better understand community priorities, such as utility affordability or environmental pollution concerns.
Background
What Are Data Centers?
Data centers are critical infrastructure that house computer servers and the systems needed to power, cool, and secure those servers. Data centers support essential digital services, such as online banking, social media, cloud storage, and communications.
Data Center Types and Sizes
| Type | Size |
|---|---|
| Small | Fewer than 150 square feet: typically integrated within commercial buildings to support localized computing needs |
| Midsize | Between 2,700 and 6,900 square feet: often stand-alone facilities that provide data storage and general computing services; commonly owned by businesses or telecommunications providers |
| Co-located | Around 11,000 square feet: computing capacity leased to multiple clients, offering shared infrastructure and operational efficiencies |
| Hyperscale | Approximately 30,000 square feet or more: often part of large campuses, designed for high-capacity processing to support cloud computing and large-scale digital operations |
| AI-specialized | A subset of hyperscale centers built to train and operate large language models (LLMs) and other advanced AI systems |
What Factors Determine a Good Location for a Data Center?
Resource Input Factors
| Factor | Description |
|---|---|
| Land use | Low population density and flat terrain indicate areas with available space and lower land costs suitable for data centers. |
| Energy impact | Hyperscale data centers need at least 100 MW of electricity. Existing electricity capacity helps gauge how much new generation may be required. |
| Water stress | Data centers rely on water for cooling. Water stress measures demand versus supply, indicating resource availability for operations. |
| Broadband coverage | High-speed fiber networks (upload speeds >100 Mbps) are essential for data transfer and optimal data center performance. |
Additional Governance, Community, and Environmental Factors
Local factors relevant to assess the benefits and impacts of data centers include
- natural disaster risk
- social vulnerability
- electricity and water utility affordability
- air and water pollution
- government finance metrics.
Explore the Tool
Explore the interactive maps in the tool (the Data Center Siting Model and Suitability Mapper for Community Planners) below to evaluate individual factors representative of data center benefits and impacts. The tool allows users to identify each county’s resource suitability rating based on model results and compare locations side by side using tables.
The tool displays two categories of factors:
- data center resource inputs: land availability, power capacity, water access, and broadband infrastructure
- local impact factors: governance, community, and environment.
By reviewing these maps and evaluation criteria, decisionmakers can pinpoint areas best suited for data center development, assess potential trade-offs, and prioritize investments that align with local planning and development goals.
National Assessment Mapper
Use this visualization to assess the resource suitability of a location—how well a location meets the infrastructure needs of a data center without overstraining local resources. Larger orange dots indicate higher suitability.
Quick Start Instructions
In the right panel of the visualization:
- Step 1: Under Filter Controls, select the state and counties to show.
- Step 2: Under Filter Controls, select the factor to display on the base map.
- Step 3: Under Model Priorities, set stakeholder priorities among resource inputs. Rank resource inputs from 0 (least important) to 4 (most important).
Please see our technical guidance for details on model parameters and formulation.
Local Assessment Value Finder
Use this visualization to conduct a detailed comparison of two counties and to support completing the Data Center Siting for Local Decisionmakers Worksheet included in Appendix B of the technical guidance [1]. Values shown are the same as the metrics available for viewing in the National Assessment Mapper. A blank cell indicates missing data for that county.
Local Assessment Mapper
Use this visualization to explore additional datasets that may be helpful for local planning purposes.
Learn More
Get more detailed instructions on the tool and learn more about our methods in the technical guidanceFact Sheets: The Needs and Impacts of Data Centers
Additional Information
This tool was funded through internal resources and conducted in the RAND Center for Climate and Energy Futures (CCEF). The center is a collaboration between RAND Global and Emerging Risks and RAND Education, Employment, and Infrastructure. For more information, visit www.rand.org/ccef or email climate@rand.org.
Acknowledgments
We thank Ben Preston for championing this work as part of CCEF’s portfolio.
We thank Tom LaTourette, Scott Stephenson, and Andrew Star for their support, guidance, thoughtful reviews, and feedback as we developed and executed the model and tool described here. That we received help and insights from those acknowledged here does not imply they agree with the views and perspectives expressed in this tool.
Note
- [1] Sara Stullken, Jonathan Cohen, Rebecca Tisherman, Devin Tierney, Emily Hoch, Andrew Lauland, Jason Thomas Barnosky, and Zohan Hasan Tariq, Data Center Siting Model and Suitability Mapper for Community Planners: Overview and Technical Documentation, RAND Corporation, TL-A4617-2, 2026, https://www.rand.org/pubs/tools/TLA4617-2.html