AI Data Centers Are Increasing Pressure on Water Resources Across US Cities

Artificial intelligence is reshaping the global economy, but the physical infrastructure supporting the AI boom is creating a growing environmental challenge across the United States: water consumption.

As hyperscale AI data centers expand rapidly in states such as Texas, Arizona, Virginia, Georgia, and Iowa, concerns are intensifying over how much water these facilities consume for cooling and electricity generation. Researchers, utilities, environmental organizations, and local communities are increasingly warning that the accelerating buildout of AI infrastructure could place substantial pressure on already stressed regional water systems.

Recent reports from Lawrence Berkeley National Laboratory (LBNL), the International Energy Agency (IEA), Ceres, Brookings Institution, Bloomberg, the Environmental and Energy Study Institute (EESI), and other organizations suggest that the AI infrastructure boom is becoming not only an electricity challenge, but also a major water resource issue.

The rapid growth of artificial intelligence is now forcing policymakers and infrastructure planners to confront an uncomfortable reality: the future of digital infrastructure may depend as much on water availability as on computing power.

Why Do AI Data Centers Consume So Much Water?

AI data centers require enormous amounts of cooling because advanced computing hardware generates intense heat during operation.

Modern AI workloads, especially large language models and generative AI systems, rely heavily on graphics processing units (GPUs) and high-density server clusters. These systems consume significant electricity and produce substantial thermal loads that must be continuously managed to avoid equipment failure.

Water is used in two major ways:

Direct Water Consumption

Data centers often use water-based cooling systems to remove heat from servers and cooling equipment.

This includes:

• evaporative cooling systems
• cooling towers
• chilled water systems
• hybrid cooling technologies

Indirect Water Consumption

Electricity generation itself also requires water.

Power plants supplying electricity to data centers frequently consume water for:

• steam generation
• cooling processes
• thermal regulation

As AI infrastructure expands, both direct and indirect water usage increase significantly.

What Do the Latest National Water Consumption Figures Show?

Recent federal and institutional research shows that data center water consumption in the United States has increased dramatically over the past decade.

According to the “2024 United States Data Center Energy Usage Report” by Lawrence Berkeley National Laboratory (LBNL), authored by Arman Shehabi and colleagues, US data centers directly consumed approximately:

• 66 billion liters of water
• roughly 17.4 billion gallons

during 2023 alone.

That represents a major increase compared with approximately:

• 21.2 billion liters
  used in 2014.

The report also estimated indirect water consumption from electricity generation reached approximately:

• 800 billion liters

during 2023.

Importantly, hyperscale facilities — many increasingly tied to AI infrastructure — accounted for approximately 84% of direct water use.

Researchers project that water demand could continue rising substantially through 2028 as AI adoption accelerates.

How Is Artificial Intelligence Amplifying Water Demand?

Artificial intelligence is increasing data center water use because AI workloads consume far more computational power than many earlier digital services.

The International Energy Agency’s (IEA) 2025 report “Energy and AI” emphasized that AI-focused data centers are among the fastest-growing segments of global electricity demand.

The IEA estimated a typical 100-megawatt data center can consume water equivalent to approximately:

• 6,500 households

when both direct and indirect water usage are included.

As AI models become larger and more computationally intensive, cooling requirements rise alongside electricity demand.

This is especially important because:

• AI training clusters often run continuously
• inference workloads scale across millions of users
• hyperscale campuses increasingly operate at industrial energy levels

The result is a rapidly growing infrastructure ecosystem with substantial water requirements.

Why Are Phoenix and Arizona Becoming Major Water Stress Flashpoints?

Phoenix has emerged as one of the most closely watched examples of the growing tension between AI infrastructure expansion and regional water stress.

The nonprofit organization Ceres released a major report in September 2025 titled:
“Drained by Data: The Cumulative Impact of Data Centers on Regional Water Stress.”

The report found that data centers in the Phoenix metropolitan area currently consume approximately:

• 385 million gallons of water annually
  for direct cooling alone.

However, planned facilities could increase this figure by approximately:

• 870%

raising annual direct consumption to:

• approximately 3.7 billion gallons

According to Ceres, that amount would be sufficient to supply a city the size of Flagstaff, Arizona, for nearly:

• 1.75 years.

The report also estimated electricity-related water consumption tied to these facilities could exceed:

• 14.5 billion gallons annually.

Arizona is particularly vulnerable because:

• it already faces long-term drought conditions
• Colorado River supplies remain strained
• urban growth continues accelerating
• temperatures increase cooling requirements

As AI infrastructure expands in the Southwest, water resource management is becoming increasingly controversial.

➡️ Read the related Post: The AI Infrastructure Boom in the US Is Shifting Data Center Growth Beyond Silicon Valley

Why Is Texas Emerging as a Major Water Consumption Concern?

Texas is rapidly becoming one of the largest AI infrastructure markets in the United States, but researchers warn this growth could significantly increase pressure on regional water systems.

A 2025 report from the Houston Advanced Research Center (HARC) and the University of Houston titled:
“Powering Texas’ Digital Economy: Data Centers and the Future of the Grid”
estimated Texas data centers could consume approximately:

• 49 billion gallons of water during 2025.

The report projected water usage could rise dramatically to:

• approximately 399 billion gallons annually by 2030.

Researchers noted that amount could theoretically lower Lake Mead by more than:

• 16 feet per year.

Texas faces unique challenges because of:

• rapid AI infrastructure growth
• extreme heat conditions
• population expansion
• industrial demand
• recurring drought cycles

Additional analysis from the University of Texas at Austin’s Bureau of Economic Geology in May 2026 suggested data centers could eventually account for:

• 3% to 9% of total Texas water use by 2040,
  up from less than:
• 1% today.

These projections highlight how AI infrastructure may become a major long-term factor in state-level water planning.

How Are Local Communities Being Affected?

The water demands of hyperscale AI facilities are increasingly becoming local political and environmental issues.

In Newton County, Georgia, a Meta data center reportedly consumes approximately:

• 500,000 gallons of water per day,
  according to reporting by The New York Times during 2025.

That single facility reportedly accounts for nearly:

• 10% of the county’s total water usage.

Proposed additional facilities in the region could potentially consume:

• up to 6 million gallons per day.

As more AI infrastructure projects move into suburban and rural areas, local communities are beginning to question:

• whether water systems can support future growth
• how utilities will prioritize allocation
• whether residents may face higher costs
• how environmental sustainability will be maintained

In several regions, residents have expressed concerns about:

• groundwater depletion
• drought resilience
• agricultural competition
• long-term sustainability

The issue is becoming especially sensitive in water-stressed regions across the Southwest and parts of the Southeast.

How Much Water Do Individual Data Centers Use?

Water consumption varies significantly depending on:

• facility size
• climate conditions
• cooling technology
• energy source
• operational intensity

Brookings Institution reported in November 2025 that:

• a typical data center may use approximately 300,000 gallons of water daily
• larger hyperscale facilities can consume up to 5 million gallons per day

That range is equivalent to the daily water needs of approximately:

• 1,000 to 50,000 households.

Google’s environmental disclosures provide some of the most transparent site-level data in the industry.

Google reported its global operations used:

• between 5 and 8.1 billion gallons of water
  in recent years.

Its Council Bluffs, Iowa data center alone reportedly consumed approximately:

• 1.3 billion gallons during 2024,
  equal to roughly:
• 3.7 million gallons per day.

These figures illustrate how individual facilities can become major regional water consumers.

Why Is Northern Virginia Under Growing Pressure?

Northern Virginia, often called “Data Center Alley,” remains the world’s largest data center concentration.

However, its rapid expansion is increasing pressure on local utilities and water infrastructure.

The Environmental and Energy Study Institute (EESI) estimated Northern Virginia data centers consumed approximately:

• 2 billion gallons of water during 2023,
  representing a:
• 63% increase compared with 2019.

Virginia’s growth is being driven by:

• hyperscale cloud expansion
• AI infrastructure demand
• federal digital infrastructure
• strong fiber connectivity

As the region continues expanding, utility planners are facing difficult questions regarding:

• long-term water availability
• infrastructure upgrades
• sustainability standards
• environmental tradeoffs

Why Are Indirect Water Impacts Often Overlooked?

Many discussions about AI infrastructure focus only on direct cooling water consumption, but indirect water usage tied to electricity generation may be even larger.

According to LBNL estimates, indirect water use linked to powering US data centers reached approximately:

• 800 billion liters during 2023.

Electricity generation from:

• natural gas
• coal
• nuclear facilities
• some thermal plants

often requires large quantities of cooling water.

As AI electricity demand accelerates, indirect water demand may expand substantially unless grids increasingly shift toward:

• wind power
• solar power
• less water-intensive generation technologies

This creates a critical connection between:

• AI infrastructure
• energy policy
• water sustainability

Are Technology Companies Trying to Reduce Water Usage?

Major technology companies are increasingly investing in water-efficiency strategies as public scrutiny grows.

Several firms are experimenting with:

• closed-loop cooling systems
• liquid immersion cooling
• air cooling technologies
• recycled water usage
• wastewater recovery
• water-positive initiatives

Microsoft, Google, and other hyperscalers have announced efforts to reduce operational water intensity in future facilities.

Some companies are shifting certain facilities toward:

• air-assisted cooling
• non-potable water systems
• advanced thermal management technologies

However, experts caution that total water demand may still rise because AI infrastructure growth is expanding so rapidly.

Efficiency improvements may reduce water intensity per computation, but overall infrastructure scale continues increasing.

Why Is Transparency Becoming a Major Issue?

One of the biggest challenges in understanding AI infrastructure’s environmental impact is limited disclosure.

Researchers note that:

• reporting standards vary widely
• many companies provide limited site-level water data
• indirect water usage is difficult to calculate
• regional utility impacts are often unclear

Ceres, Brookings, Bloomberg, and other organizations have called for:

• improved transparency
• standardized reporting
• stronger sustainability disclosure requirements

Without consistent reporting, policymakers may struggle to evaluate cumulative impacts on regional water systems.

Can Water Infrastructure Keep Up With AI Growth?

The long-term sustainability of AI infrastructure expansion may increasingly depend on regional water planning.

States competing aggressively for hyperscale investment now face difficult tradeoffs involving:

• economic growth
• electricity demand
• water sustainability
• climate resilience
• infrastructure investment

In regions already facing drought conditions, future data center growth could intensify political and environmental tensions.

At the same time, many analysts emphasize that:

• agriculture remains the dominant water consumer in most US regions
• data centers represent a relatively smaller share overall
• efficiency technologies continue improving

Still, the unprecedented scale of projected AI infrastructure expansion means water consumption is becoming an increasingly important public policy issue.

The AI Economy Is Becoming a Water Infrastructure Story

The rapid growth of artificial intelligence is transforming not only digital infrastructure, but also the physical systems supporting modern economies.

Research from LBNL, the IEA, Ceres, Brookings Institution, EESI, Bloomberg, HARC, and other organizations increasingly shows that AI infrastructure growth is placing rising pressure on water resources across multiple US regions.

Texas, Arizona, Georgia, Virginia, Iowa, and other emerging data center markets are now confronting the environmental implications of hyperscale AI expansion.

The challenge is becoming especially important because many fast-growing AI infrastructure hubs are located in areas already facing:

• drought stress
• rapid urbanization
• rising temperatures
• infrastructure limitations

As AI adoption accelerates, the conversation around digital infrastructure is expanding beyond semiconductors and electricity grids.

Increasingly, the future of artificial intelligence may also depend on one of the world’s oldest and most essential resources: water.

Check for more information: Core Insights Review

Core Insights Review contributors publish research-based analysis and editorial insights on commercial real estate, PropTech, smart infrastructure, sustainable construction, industrial real estate, and emerging technologies shaping the future of the built environment.