Data Center Expansion in Virginia: Closing Critical Gaps for Informed Water Planning and Permitting

By: Ana Pinheiro Privette*, Ximing Cai*, Hari David^#, Zach Sugg^@, and Peter Colohan^@

*Center for Secure Water, The Grainger College of Engineering, University of Illinois Urbana-Champaign, ^#The Grainger College of Engineering, University of Illinois Urbana-Champaign, ^@ Lincoln Institute of Land Policy

Header image: Hugh Kenny/The Piedmont Environmental Council

Executive Summary

The rapid expansion of AI, cloud computing, and digital services has positioned Virginia, particularly Northern Virginia, as the world’s largest data center hub, delivering substantial economic value and critical digital infrastructure while creating increasingly concentrated demands on energy and water systems. Although data centers account for a relatively small share of total statewide water withdrawals, especially compared with fossil fuel and nuclear power plants and the largest municipal systems, their impacts are highly localized, temporally variable, and cumulative, particularly where facilities are geographically clustered and depend on shared municipal infrastructure. When indirect water use associated with electricity generation is considered, the sector’s total water footprint becomes substantially larger and more difficult to quantify.

Most existing and planned data centers are located within already water-constrained watersheds—including the Potomac, Rappahannock, North Anna, and James—where rapid expansion adds cumulative pressure to systems facing increasing supply stress. Evidence from Northern Virginia utilities indicates that data center water use increased by approximately 86% between 2019 and 2023, with a small number of facilities accounting for a disproportionate share of demand, a trend the Interstate Commission on the Potomac River Basin projects will continue.

Virginia’s water governance framework, shaped by a history of relative abundance, riparian principles, basin-based planning, and distributed permitting, relies heavily on robust data, monitoring, and modeling to function effectively. Yet data center water use remains among the least transparent components of the system: facility-level withdrawals, timing, consumptive use, return flows, and indirect water use from electricity generation are largely absent, inconsistently reported, or disclosed only in aggregated form. These transparency gaps hinder cumulative impact assessment, weaken proactive planning, erode public trust, and increase the risk of fragmented, permit-by-permit decision-making that detects stress only after critical thresholds are crossed—particularly as hydroclimatic variability and drought risk intensify.

Ensuring sustainable data center growth therefore requires a systems-level approach that integrates water availability, infrastructure capacity, climate variability, and evolving demand, supported by robust historical and present-day water mass-balance analysis and credible scenario-based futures. The principal constraints are not data scarcity, but uncertainty in climate–hydrology translation, assumptions about future infrastructure and operational growth, and limited, non-standardized disclosure of data center water use. The paper concludes with actionable recommendations centered on standardized facility-level reporting, multi-scale assessment, integrated water–energy–AI data systems, stronger institutional coordination, and forward-looking scenario analysis to move Virginia toward more anticipatory, transparent, and resilient water governance while safeguarding long-term water security and public trust.

1. Introduction

The rapid adoption of artificial intelligence (AI), cloud computing, and digital services is driving unprecedented growth in data center infrastructure. Modern hyperscale facilities supporting AI workloads are far more computing- and energy-intensive than earlier generations, with new centers built at power capacities of roughly 100–1,000 MW, comparable to the electricity demand of 80,000 to 800,000 homes (Lawson et al, 2026). U.S. data centers consumed an estimated 183 terawatt-hours (TWh) of electricity in 2024, exceeding 4% of total U.S. demand, and consumption is projected to rise by 133% to approximately 426 TWh by 2030 (IEA, 2025). The resulting heat loads require continuous cooling, and many widely deployed technologies, such as evaporative cooling, depend on substantial freshwater withdrawals, directly linking digital infrastructure growth to local water resources and intensifying pressure on already stressed supplies.

The water footprint of data centers includes both direct and indirect water withdrawals and consumptive use. Direct water use (Scope 1) is primarily associated with onsite cooling systems and can involve large withdrawals, a portion of which is consumptively used through evaporation; the remainder may be returned to the system as warmer effluent, depending on cooling design, water quality, and discharge conditions. Scope 1 impacts vary widely with server density, cooling technology, energy efficiency, and local climate. Indirect water use includes water withdrawn and consumed in electricity generation (Scope 2) and water embedded in upstream supply chains (Scope 3). Importantly, indirect impacts are often dominated by consumptive losses. According to the International Energy Agency (IEA, 2025), Scope 1 and Scope 2 together account for the majority of data center water use globally: on average, roughly two-thirds of total consumptive water use is attributable to electricity generation, while approximately one-quarter is associated with onsite cooling, underscoring the importance of accounting for both withdrawal volumes and net consumption when assessing cumulative water impacts.

Despite their growing role as critical infrastructure, data centers remain among the least transparent industrial water users. Reporting is typically partial, aggregated, and non-standardized, with little distinction between direct withdrawals that stress local supplies and indirect water use embedded in electricity generation, obscuring cumulative impacts and the full water footprint (WEF, 2024; EPA, 2024). As the U.S. data center market is projected to grow by ~23% annually through 2030 (Upwind, 2024), understanding water impacts is increasingly urgent. Although data centers account for a small share of national withdrawals—about 0.3% of the public water supply (Ren & Luers, 2025)—their impacts are highly localized: demand is concentrated in specific watersheds, varies seasonally and diurnally, and often peaks during heat waves and droughts. Where facilities cluster, large data centers can impose cumulative demands comparable to small towns, placing disproportionate stress on shared water, wastewater, and supporting infrastructure (Shehabi et al., 2024).

Beyond their energy and water demands, data centers drive significant land-use change through large facilities, transmission infrastructure, and roads, increasing impervious surfaces and fragmenting ecosystems. On-site diesel backup generators add localized air pollution, greenhouse gas emissions, and noise. Together—energy consumption, water withdrawals, emissions, air quality effects, and land transformation—these interconnected pressures underscore the need for integrated environmental assessments that consider cumulative and cross-sectoral impacts, particularly in regions where facilities are clustered.

Virginia has emerged as the U.S. epicenter of digital infrastructure, with Northern Virginia’s “Data Center Alley” hosting over 300 facilities—roughly 13% of global operational capacity and 25% of capacity in the Americas (JLARC, 2024). This concentration is enabled by dense fiber connectivity, reliable low-cost power, available land, proximity to major customers, and long-standing tax incentives. This unprecedented concentration has significant energy implications, requiring rapidly increasing amounts of electricity. Data center electricity demand in Virginia has more than doubled in recent years (Mulkey, 2024), with facilities accounting for ~21% of utility load in early 2023 and projections indicating that peak demand could double by 2038 (JLARC, 2024). High energy use drives substantial water demands for cooling, often peaking during hot, dry periods when water systems are stressed. While individual facilities may rely on municipal systems rather than direct withdrawals, their water use ultimately traces back to shared rivers, reservoirs, and aquifers, linking private operational decisions to public water resources.  In Virginia, most existing and planned data centers are situated within already water-constrained watersheds—including the Potomac, Rappahannock, North Anna, and James—where continued expansion compounds growing supply risks.

Despite their growing role as critical infrastructure, data centers remain among the least transparent industrial water users. Reporting is typically partial, aggregated, and non-standardized, with little distinction between direct withdrawals that stress local supplies and indirect water use embedded in electricity generation, obscuring cumulative impacts and the full water footprint

The economic and societal importance of digital infrastructure is unquestioned. The real question is not whether it should grow, but how it can expand sustainably—protecting communities, ecosystems, and the long-term reliability of the systems themselves. Achieving this requires integrated socio-environmental assessments that account for cumulative, cross-sectoral impacts, which in turn depend on detailed, facility-level data on operations and resource use—data that remain limited, fragmented, and inconsistently reported. This document focuses on the water-related dimensions of data center expansion, analyzing how growth may affect water availability, system reliability, and long-term sustainability, with a particular focus on conditions in the Commonwealth of Virginia.

2. Current Evidence on Data Center Water Use in Virginia

Most data centers in Virginia obtain water from local public utilities, which draw from rivers and aquifers regulated by the Virginia Department of Environmental Quality (VA DEQ) and, for the Potomac River, the Interstate Commission on the Potomac River Basin. Direct withdrawals outside a community water system (CWS) require permits (DEQ, 2026a). While DEQ oversees large withdrawals through permitting and reporting, data center water use is often aggregated within utility reports rather than disclosed at the facility level. Consequently, there is limited visibility into how rapidly growing data center demand affects local allocations and cumulative water stress. No centralized, publicly accessible database tracks water withdrawals or consumption for all Virginia data centers, and reporting methods are not standardized. Most operators do not disclose facility-level water use, and utilities are often constrained by confidentiality provisions.

Data center water withdrawal and consumptive use can be partially inferred through existing water and energy data systems, but only indirectly and in aggregate. Sources include: (1) public water utility withdrawal data, capturing demand indirectly for facilities served by CWSs; (2) permitted withdrawals outside CWSs, where facilities self-supply from surface or groundwater; and (3) estimates derived from electricity consumption combined with grid-level water intensity assumptions. These data allow broad inferences about withdrawal magnitude, spatial concentration, and seasonal patterns but rarely quantify direct consumptive use—the portion of water evaporated or lost during cooling—which must be estimated using assumptions about cooling technology and operational practices.

At the statewide level, water use data are compiled in Virginia’s Annual Water Resources Report (DEQ, 2024), which aggregates withdrawals by sector (e.g., public supply, commercial, industrial). Data centers are not reported as a separate category and are instead embedded within broader commercial and industrial groupings. Although DEQ’s reporting systems contain the raw data needed to estimate withdrawals, isolating data center–specific use requires supplemental utility data, modeling assumptions, and bespoke analysis.

Using this approach, the Joint Legislative Audit and Review Commission (JLARC, 2024) produced the most comprehensive statewide assessment of data center water use to date, based on a subset of major utilities, primarily in Northern Virginia. JLARC estimated that Virginia data centers consumed approximately 2.1 billion gallons of water in 2023, an 86% increase since 2019 (~1.13 billion gallons), reflecting the sector’s rapid growth. While total use remains under 0.5% of statewide withdrawals—masking impacts at the state scale—the distribution is highly uneven. Most facilities consume volumes comparable to large commercial buildings (~6.7 million gallons per year), yet a small number of sites dominate demand: eleven buildings exceeded 50 million gallons annually, and one facility alone used ~243 million gallons, roughly 10% of total data center water use in 2023.

The spatial concentration of impacts is particularly pronounced in Northern Virginia. Utilities in Loudoun County reported roughly 899 million gallons of potable water used by data centers in 2023, representing a roughly 250% increase over earlier years (Benforado et al., 2025). Water use peaks during summer months, coinciding with both maximum cooling demand and heightened drought risk. Importantly, JLARC found that just over one-third of total data center water use statewide was supplied through reclaimed water rather than new freshwater withdrawals, illustrating both the mitigating role of reuse and its limits as demand continues to grow (JLARC, 2024). Utility-provided analyses in Fairfax and Prince William counties further indicate substantial variability among facilities: average daily use may appear modest (~18,000–38,000 gallons per day per building), but peak-day demands can approach 88,000 gallons per day, stressing local systems during critical periods (County of Fairfax, 2024).

Regional analyses, including the 2025 Washington Metropolitan Area Water Supply Study (Ahmed et al., 2025) indicate that existing demand estimation methods may underestimate data center consumptive water use as the sector scales. The study projects that data center consumptive use in the Washington metro area could rise substantially under growth scenarios — from about 4.0 million gallons per day (MGD) in 2025 to roughly 22.2 MGD by 2050 on average, with peak-day use increasing from about 14.3 MGD to 80.5 MGD in the same period, illustrating how data center demand could become a significant and variable component of total water use. These projections highlight concerns that as the sector grows, especially with widespread adoption of water-cooled systems, data center water demand could significantly affect watershed availability and stress regional supplies during droughts or peak periods (Ahmed et al., 2025). The analysis also underscores substantial uncertainty stemming from the lack of mandatory, facility-level reporting, a concern echoed by JLARC and regional planners, reinforcing calls for explicit disclosure and accounting of data center water use within permitting frameworks. The study further warns that drought-related supply risks are likely to intensify as regional hydroclimatic conditions continue to change.

3. Virginia’s Water Governance Framework

Virginia has historically been characterized as a relatively water-rich state, shaped by its humid Mid-Atlantic climate, abundant surface waters, and substantial groundwater resources. Annual precipitation typically exceeds 40 inches (DEQ, 2010), supporting extensive river systems and aquifers and underpinning a riparian water-rights framework that emphasizes shared “reasonable use” rather than fixed volumetric allocations. However, water availability in Virginia is far from uniform. Pronounced spatial variability exists between surface-water-rich river basins and groundwater-dependent regions, particularly east of the Fall Line and along the coast, where aquifers are subject to long-term overdraft and saltwater intrusion (USGS, 2026). These localized vulnerabilities are often obscured by statewide averages, masking areas of emerging stress.

Water supply infrastructure in Virginia, consists of a diverse network of surface water reservoirs, rivers, groundwater aquifers, treatment facilities, and distribution systems that support municipal, industrial, agricultural, and ecological needs. Major river basins such as the James, Potomac, Rappahannock, and Roanoke underpin regional water supplies, with large reservoirs (e.g., Smith Mountain Lake and Lake Gaston) providing storage, regulation, and drought resilience. Urban areas rely on centralized treatment and distribution systems, while rural regions often depend on smaller utilities or groundwater wells. Increasing demands from population growth, climate variability, and environmental flow requirements have made integrated water resources management and infrastructure modernization key priorities across the state.

Because Virginia’s riparian system lacks automatic allocation or curtailment, effective governance depends heavily on robust data, monitoring, and basin-scale modeling; without this visibility, stresses on water supplies, ecosystems, and interconnected infrastructure may only become evident after critical thresholds are crossed

Virginia’s water governance (DQE, 2026a) is overseen primarily by the Department of Environmental Quality (DEQ). Surface-water withdrawals require permits under the Virginia Water Protection (VWP) program for non-agricultural withdrawals exceeding 10,000 gallons per day (gpd; nontidal) or 2 million gpd (tidal), and for agricultural withdrawals over 1 million gallons per month (gpm; nontidal) or 60 million gpm (tidal), with exemptions for small-scale residential uses, firefighting, and pre-1989 withdrawals. Groundwater permits are required only within designated Groundwater Management Areas (GWMAs)—mainly Eastern Virginia and the Eastern Shore—for withdrawals of 300,000 gallons per month or more. Groundwater withdrawals outside these areas are largely unpermitted and lightly regulated, even when they rely on sole-source or locally critical aquifers. However, DEQ still has oversight through water withdrawal reporting, especially if withdrawals exceed thresholds. Even without permits, users must report annual withdrawals to DEQ if thresholds are exceeded (>10,000 gpd daily or >1 million gallons per month for irrigation). Permitting involves multi-agency review, defined timelines, and fixed terms, making the system dependent on accurate reporting and oversight to manage cumulative impacts. Water supply planning is carried out by local and regional utilities and basin-based Regional Planning Units under state regulations (9VAC25-780).

Although this framework supports basin-scale planning, water allocation and cumulative impact assessment remain fragmented across agencies and jurisdictions, with no integrated statewide system to proactively evaluate the demands of emerging high-intensity users such as data centers. Because Virginia’s riparian system lacks automatic allocation or curtailment, effective governance depends heavily on robust data, monitoring, and basin-scale modeling; without this visibility, stresses on water supplies, ecosystems, and interconnected infrastructure may only become evident after critical thresholds are crossed. In the absence of formal allocation priorities, transparent facility-level disclosure and integrated systems-level analysis are critical for anticipating stress, guiding permitting, and aligning data center growth with long-term water security.

Rising climate variability is undermining Virginia’s water governance and challenging long-standing assumptions about water reliability, increasing the risks posed by high-intensity users such as data centers on already stressed water systems. While total annual precipitation is projected to increase, hydroclimatic variability is intensifying, with more extreme rainfall events, higher evapotranspiration rates, rapid soil drying, and a growing prevalence of episodic and “flash” droughts (NCA, 2023). In recent years, the Mid-Atlantic has experienced recurring drought and soil-moisture deficits, including severe conditions during 2024–2025 (Ahmed et al., 2025). Loudoun County—home to the world’s largest concentration of data centers—faced month’s-long drought conditions in 2024 that triggered mandatory water-use restrictions (Styer, 2024). As of January 2026, unusually dry conditions and drought persist across much of Virginia, following sustained precipitation deficits since late 2025 (Woolsey, 2026). These trends underscore that even historically water-abundant states face growing uncertainty in water reliability, particularly during peak demand periods when data center cooling needs are highest.

Together, these conditions position Virginia as a critical case study: a state with historically abundant water resources, increasing hydroclimatic variability, rapid expansion of digital infrastructure, and governance structures that depend heavily on data transparency to function effectively.

4. What We Do Not Know Can Hurt Us

Existing data limitations highlight a fundamental challenge for water planning in Virginia: reliance on aggregated reporting and annual averages obscures localized and peak-period impacts that actually drive water stress. Water scarcity emerges from short-term, site-specific demand interacting with local hydrology, infrastructure capacity, and regulatory constraints, not from statewide totals or annual means. The continued expansion of data centers may generate unintended social, economic, and environmental consequences, driven by persistent uncertainties and critical knowledge gaps.

Because data center water disclosure in Virginia remains partial and uneven, this introduces a large uncertainty into the equation. While statewide averages can be estimated, the absence of standardized, facility-level reporting and a centralized public database limits the ability of regulators and communities to assess cumulative impacts, drought vulnerability, and long-term water balance reliability (JLARC, 2024). Differences in cooling technologies, operating practices, and reuse availability mean that facilities with similar IT loads can have vastly different water footprints, a variability masked when planners rely on average Water Usage Effectiveness (WUE) values or sector-wide assumptions.

Another critical and often overlooked gap concerns indirect water use embedded in electricity consumption. Because a large share of data center water impacts occurs offsite through thermoelectric power generation, focusing solely on onsite cooling underestimates total water demand and misidentifies where stress occurs. In Virginia, where data centers are projected to consume roughly 25 percent of statewide electricity by 2025, indirect water use is both spatially displaced and temporally aligned with peak heat and low-flow conditions, when ecological and infrastructure constraints are most binding. Yet no single public dataset links data center load profiles, grid mix, power-plant cooling technologies, and associated water use at sufficient temporal resolution, requiring analysts to piece together information (Galanter et al., 2023) and custom modeling (Siddik et al., 2021; Chini et al., 2020).  Tracking the indirect water use associated with energy production in the state is challenging. Water use by energy utilities in Virginia are often excluded from key published reporting, despite its large withdrawals. For example, withdrawals or diversions of water for hydroelectric power generation are considered nearly all non-consumptive and are exempt from the annual water withdrawal reporting requirements (LIS, 2026a)

These transparency gaps have tangible governance and resilience implications. Limited access to high-resolution, time-sensitive data constrains proactive planning, increases the risk of permit-by-permit decision-making, and raises the likelihood that emerging stresses will only become visible after hydrologic or infrastructure thresholds are crossed. Without improved transparency and cumulative assessment, high-demand users risk pushing local systems toward stress thresholds before impacts are fully understood, undermining both long-term water resilience and the reliability of the digital infrastructure that increasingly underpins Virginia’s economy. Lack of transparency in how data centers use water also erodes public trust, fuels speculation and opposition, and weakens confidence in regulators. When facility-level information on withdrawals, timing, sources, and return flows is not disclosed, it creates the perception that decisions favor large corporate users over households and the environment, and that impacts and costs may be shifted onto the public.

The sector’s limited transparency raises critical questions: Can existing water resources sustain the rapid expansion of digital infrastructure? How might future water conditions constrain operations, reliability, and growth? And could continued development compromise the structural and ecological integrity of the Commonwealth’s water systems? Addressing these challenges will require basin-scale, climate-sensitive water planning.

5. Ensuring Sustainable Growth of Data Centers

Ensuring sustainable data center expansion in Virginia requires a systems-level assessment of the full water balance that considers: how much water is physically and legally available, how it is allocated, how supply–demand dynamics may shift under climate change and economic growth, the hazards and risks of water supply failure (particularly for data center cooling), and the policy and infrastructure innovations needed to manage these risks. Determining whether water use is “reasonable” requires detailed knowledge of withdrawal volumes, timing, and locations, and how these interact cumulatively with other users and environmental needs (EPA, 2024). This context makes cumulative, basin-scale modeling essential.

Effective assessment must capture seasonal and extreme-event variability, surface–groundwater interactions, and the linkages between water, energy, and infrastructure. Integrating climate and socioeconomic scenarios with data center growth trajectories within a mass-balance framework allows future water availability to be evaluated as the combined outcome of changing hydrology and demand. On the supply side, climate change alters precipitation, evapotranspiration, streamflow, groundwater recharge, and drought frequency; on the demand side, water use evolves with population growth, power-sector transformation and cooling technology choices, agricultural shifts, and the expansion of other water-intensive industries.

Virginia has substantial hydrologic and water use datasets and has produced state planning documents with many components of a water budget (DEQ, 2024; DEQ, 2026b, LIS, 2006b), but a comprehensive, published water mass-balance assessment that links supply, demand, infrastructure, return flows, and future scenarios in an integrated model does not yet exist. Multiple tested modeling frameworks (Cai et al., 2006; Chen et al., 2022; Luukkonen et al., 2023; Zheng  at al., 2025)and applied at large basin scales (e.g., Vora & Cai, 2026), can support such analysis and have demonstrated the feasibility of integrated, scenario-based assessments that explicitly link infrastructure growth with hydrologic constraints.

An assessment of publicly available data layers indicates that Virginia has sufficient information to conduct robust historical and present-day water mass-balance analyses and to provide a credible foundation for scenario-based future planning. The primary constraints are not gaps in hydrologic or climate data, but rather: (i) uncertainty in translating projected climate signals into hydrologic responses, (ii) assumptions about the evolution of infrastructure, operations, and demand, and (iii) limited, non-standardized transparency around data center water use. Addressing these gaps—particularly through improved facility-level reporting and integrated scenario analysis—would significantly enhance the Commonwealth’s capacity to evaluate cumulative water risk and guide the expansion of critical digital infrastructure in alignment with long-term water resilience. Applying these approaches at spatial scales aligned with Virginia’s water governance framework (from HUC‑12 watersheds to Regional Planning Areas under 9VAC25‑780) and temporal scales capturing peak cooling demand and drought conditions would enable identification of localized cumulative stress and assessment of effective supply portfolios.

Understanding Virginia’s water landscape can guide strategic allocation for data center growth through three potential pathways: (1) developing new supply, (2) reallocating existing supply, and (3) substituting withdrawals with non-traditional sources. New supply involves expanding surface-water or groundwater infrastructure, limited by hydrology, drought risk, environmental flows, and sustainability. Reallocation uses voluntary transfers or agreements with existing users, subject to legal, ecological, and transactional constraints. Non-traditional sources, such as reclaimed water, industrial recycling, and circular-economy approaches can reduce freshwater withdrawals and improve local reliability. Feasibility varies across the state, as constraints often stem from deliverability during droughts, permitting, infrastructure capacity, and local acceptance. Planning should therefore treat water supply as a portfolio, evaluating reliability, regulatory compliance, costs, community equity, and adaptive measures like efficient cooling, onsite reuse, and drought-contingency operations.

6. Recommendation for a Transparent and Robust Path Forward

As data center development accelerates across Virginia, traditional water planning approaches—designed for more predictable and diffuse demand—are increasingly strained. Managing the cumulative and localized water impacts of digital infrastructure requires a shift toward more transparent, integrated, and forward-looking planning frameworks that reflect the tight coupling between water, energy, and computing systems. The following pathways outline actionable strategies to strengthen water planning and governance, reduce uncertainty, and ensure that data center expansion proceeds in alignment with long-term water availability, infrastructure capacity, and climate resilience.

1. Improve transparency. Standardized, facility-level reporting of water use is critical to understanding the sector’s cumulative impacts. This should include both direct consumptive and withdrawal use for onsite cooling (Scope 1) and energy-related information—such as electricity consumption, sourcing, and generation mix—that can be used to estimate water embedded in electricity production (Scope 2). Transparent reporting enables regulators, planners, and utilities to assess peak and seasonal demand, evaluate cumulative stress on water systems, and anticipate potential conflicts among users.
2. Adopt multi-scale assessment. Water availability and demand must be analyzed across watershed, basin, and regional scales to align operational management with long-term planning. Integrating assessments at multiple scales ensures that localized stress—often obscured by statewide averages—is captured and that management strategies are coherent across jurisdictions.
3. Evaluate data center siting at the regional level. Given the spatial distribution (or heterogeneity) of water availability for existing and emerging data center cooling demands, including variability in natural renewable supplies (streamflow and groundwater recharge) and water supply infrastructures (storage, diversion channels, and treatment facilities), data center siting should be evaluated at a regional scale. Such planning must integrate water availability with land-use constraints, environmental impacts, infrastructure capacity, and community acceptance to avoid localized water stress and unintended cumulative effects.
4. Integrate water–energy–AI data systems. The interplay between electricity use, AI-related IT workloads, cooling technology, and water withdrawals requires linked datasets and modeling frameworks. High-resolution integration of grid operations, server load profiles, cooling system types, and water-use data allows more accurate allocation of offsite water impacts and supports real-time operational adjustments to reduce stress on water systems.
5. Strengthen governance coordination. Effective oversight requires alignment across state, regional, and local water planning and management processes. Coordinated permitting, standardized reporting, and robust inter-agency communication help reduce fragmentation, enable basin-scale cumulative impact assessments, and support adaptive responses to emerging water demands. In addition, fostering collaboration between public agencies and private-sector actors—including data center operators, utilities, and industry associations—can further enhance resource planning, optimize allocations, and facilitate proactive investment in efficiency, reuse, and infrastructure resilience.
6. Support forward-looking scenario analysis. Scenario-driven planning incorporating projected digital infrastructure growth, drought conditions, and climate stress enhances resilience. Integrating hydrologic projections, socio-economic pathways, and water demand scenarios enables planners to anticipate vulnerabilities and risks, evaluate adaptive strategies including water re-allocation through market-based approaches, and prioritize investments in infrastructure and technology, efficiency, or reuse systems.
7. Encourage adoption of water-free/low water cooling technology: Encouraging widespread adoption of water‑free and low‑water cooling technologies offers one of the most direct pathways to reducing data center water footprints. These technologies, ranging from closed‑loop systems and immersion cooling to dry air cooling and hybrid approaches, can enable data centers to remain operationally efficient while significantly lowering pressure on local freshwater resources. Integrating these approaches into permitting, planning, and investment decisions can align data center expansion with sustainable water management and resilience goals.

7. Pre- and Post-Permit Water Use Disclosures

To support a coherent framework for managing data center water demand in Virginia and other rapidly growing regions, and to enable effective regional water-resources planning, resilience analysis, and cumulative-impact assessment, data center operators should disclose a standardized set of facility-level information as part of both the permit application process and ongoing operations. Disclosures should extend beyond annual averages and be structured to support watershed-scale mass-balance analysis, peak-stress evaluation, and long-term planning under climate variability.

During permit review, operators should submit standardized information that enables regulators and utilities to assess both near-term and cumulative impacts across the full facility lifecycle. This includes a clear project description and phased IT load disclosures at initial operation and full build-out, allowing evaluation of cooling demand, electricity use, and indirect water use associated with electricity demand and generation mix before long-term water resources are committed. Applicants should provide estimates of average and peak withdrawals, consumptive use, seasonal demand profiles, and expected annual losses, along with detailed identification of water sources, permitted or contracted supply volumes, and any seasonal or drought-related constraints.

Managing the cumulative and localized water impacts of digital infrastructure requires a shift toward more transparent, integrated, and forward-looking planning frameworks that reflect the tight coupling between water, energy, and computing systems.

Facility-specific information on cooling technologies, water-use efficiencies, and water-use effectiveness (both annual and peak), sensitivity to temperature extremes, and planned efficiency upgrades is essential, as facilities with similar IT loads can have markedly different water footprints. Equally important is accounting for return flows and discharges, including their timing, location, receiving systems, and potential water-quality effects. This information is critical for characterizing system-wide environmental impacts and evaluating opportunities for safe and effective water reuse. Applicants should also document water reuse plans and alternative supply strategies, as well as planned growth trajectories.

Water permitting and planning should evaluate facilities in the context of other major users within the same watershed or aquifer, the cumulative demand associated with existing and proposed data centers, and the share of local demand attributable to individual facilities. This framework should explicitly avoid permit-by-permit decision-making that obscures systemic risk. It should also incorporate equity and community considerations such as proximity to water-stressed or disadvantaged communities, impacts on municipal demand and rate structures, and priority of service relative to residential users during scarcity. Meaningful engagement with local communities and water utilities throughout the permitting process is essential to identify cumulative risks early and address them proactively, shifting water governance from reactive response toward anticipatory, preventive, and integrated water management. See Appendix A for more detailed information on recommended disclosures prior to permits being issued.

Once operational, all water appropriation permit holders in Virginia should be subject to a unified, enforceable post-permit monitoring and reporting framework designed to support adaptive management and long-term tracking of cumulative impacts. Permittees should be required to measure and report monthly water withdrawals at the individual facility level, using certified metering technologies with accuracy within 10 percent. Standardized annual water-use reports, employing harmonized metrics across sectors, would enable longitudinal trend analysis, inter-facility comparison, and regulatory oversight. Public access to non-sensitive water-use data will strengthen transparency, accountability, and public trust, while provisions for independent third-party verification (where appropriate) will enhance data integrity and regulatory confidence. See Appendix B for more detailed information on recommended disclosures once data centers are running operationally.

8. Final Remarks

Virginia’s ability to support data center growth hinges on transparency and regional-scale planning. Under the Commonwealth’s riparian water governance framework—which relies on data, monitoring, and basin-based analysis rather than priority-based allocation—consistent, facility-level disclosure of water withdrawals, timing, and sources is essential to assess basin-scale water balances, drought risk, and tradeoffs among economic development, community needs, and environmental protection. Equally important is managing the siting of data centers at a regional level, aligning new facilities with spatially variable water availability, infrastructure capacity, and ecological constraints rather than evaluating projects in isolation. Integrating transparent water-use data into system-level and regional planning enables regulators, utilities, and communities to move from reactive, permit-by-permit decisions toward anticipatory management—ensuring that digital infrastructure growth proceeds in a manner that is reliable, equitable, and environmentally sustainable.

Fundamental questions remain about whether Virginia’s existing water infrastructure can reliably support additional, concentrated demand from data centers without compromising service, resilience, or equity. Most water systems were designed for historical demand, conservative growth, and long-lived assets; rapid, clustered increases, often coinciding with heat waves and low-flow conditions, can strain treatment, conveyance, storage, and drought-response capacity, even where annual water availability seems sufficient. Aging infrastructure compounds these risks: in some Virginia systems, particularly in rural areas, studies have shown that 35–50 % of treated water is lost to leaks and metering issues (Mcgill, 2026; Atkinson, 2017). Such non-revenue water (NRW) increases stress on supply and treatment, particularly during droughts or peak-demand periods.

Virginia’s basin-based planning framework—organized around river basins, Regional Planning Areas, and finer-scale watershed units (e.g., HUC-8 and HUC-12)—provides a strong conceptual foundation to answer these questions through an integrated assessment. However, without early, facility-level disclosure of projected withdrawals, consumptive use, seasonal demand, and growth trajectories, emerging demands such as clustered data centers are difficult to incorporate into water budgets, drought scenarios, or infrastructure planning in a timely way. This stands in contrast to states such as Minnesota, where House File 16 requires large prospective water users, including data centers, to disclose expected water use prior to permit submission, enabling upfront assessment of sustainability risks and cumulative impacts (Minnesota Legislature, 2025).

These challenges are amplified by uncertainty in the pace, scale, and persistence of data center growth and their demand on local water resources. While technological innovation may reduce water intensity per unit of compute, experience across other sectors highlights the risk of Jevons’ paradox, in which efficiency gains lower costs and ultimately increase total resource consumption. Addressing this uncertainty requires planning approaches that move beyond static demand projections toward adaptive frameworks capable of managing variability in scale, timing, and operational behavior.

Equally critical are issues of cost allocation, risk, and long-term accountability. Expanding water and wastewater infrastructure to serve data centers can impose substantial capital and operating costs on utilities and communities, raising questions about who bears these costs and under what conditions. If systems are expanded to meet projected growth that later fails to materialize, due to market shifts, technological change, or project cancellation, communities may be left with stranded assets and long-term ratepayer burdens. Conversely, if facilities are decommissioned or relocated, few mechanisms currently ensure recovery of infrastructure investments or responsible site transition. Addressing these risks requires clear guardrails: transparent disclosure of water demand and growth trajectories; phased and conditional permitting tied to demonstrated need; financial assurance mechanisms to cover infrastructure expansion and decommissioning; and shared-risk frameworks between operators and utilities. Transparency underpins each of these tools—without consistent, facility-level data on water use, growth plans, and operational flexibility, Virginia cannot assess cumulative impacts, plan infrastructure prudently, or balance the imperative to support critical digital infrastructure with the obligation to safeguard long-term water security and public trust.

Acknowledgements

This work was supported by funding from the Lincoln Institute of Land Policy. We thank Chris Miller from the Piedmont Environmental Council of Virginia for his valuable review and feedback on the manuscript. His insights greatly improved the clarity and relevance of this study.

---

---