Key Question
How much water does AI infrastructure actually require, and is water use the real issue—or simply the most visible one?
Questions Worth Exploring
- How much water does AI infrastructure actually consume?
- Why do water demand estimates vary so widely?
- How do cooling technologies influence water use?
- What role does climate play in water demand?
- Should water use be evaluated independently from energy use?
- Can municipal wastewater become a sustainable water source?
- What happens to cooling tower blowdown?
- How should communities evaluate water-related tradeoffs?
- What technologies exist to reduce water demand?
- Are we asking the right questions?
What We Are Learning
Water has become one of the most visible topics in discussions surrounding AI infrastructure and data centers. Communities, policymakers, developers, utilities, and researchers are all seeking to understand how growing computational demand may affect local water resources.
One of the most important lessons emerging from these discussions is that water demand is not a simple metric.
Water use is influenced by a combination of factors including:
- Cooling architecture
- Climate conditions
- Facility design
- Computing density
- Operational strategy
- Water source availability
- Energy infrastructure
- Regional resource constraints
As a result, two facilities with similar computing loads may have significantly different water footprints.
Understanding water demand therefore requires a broader perspective than simply asking how many gallons a facility consumes.
Observations
Water Use Is Not a Single Number
Discussions often focus on annual water consumption.
However, understanding water demand requires consideration of:
- Water withdrawals
- Water consumption
- Water discharge
- Seasonal demand variations
- Peak demand periods
- Source water availability
Different stakeholders may be discussing different metrics while using the same terminology.
Cooling Architecture Matters
Cooling systems represent one of the largest drivers of water demand.
Different architectures can produce significantly different outcomes:
- Air Cooling
- Evaporative Cooling
- Direct-to-Chip Cooling
- Immersion Cooling
- Hybrid Systems
Each involves tradeoffs related to:
- Water use
- Energy consumption
- Reliability
- Cost
- Scalability
Water and Energy Are Connected
One recurring observation is that water and energy cannot be evaluated independently.
In some cases:
- Reducing water use increases energy consumption.
- Reducing energy use increases water consumption.
The challenge is often not minimizing a single metric, but optimizing outcomes across the broader ecosystem.
Wastewater May Become Part of the Solution
Growing interest exists around:
- Municipal wastewater reuse
- Industrial water reuse
- Closed-loop systems
- Water recovery technologies
These approaches shift the discussion from:
“How much water is consumed?”
to
“What water sources are available and how should they be utilized?”
Community Concerns Extend Beyond Water Quantity
Many public discussions focus on gallons consumed.
However, stakeholders are increasingly raising questions about:
- Water source sustainability
- Wastewater impacts
- Treatment requirements
- Cooling tower blowdown
- Infrastructure capacity
- Long-term resilience
These considerations highlight the importance of evaluating the entire water system rather than a single metric.
TIIM Perspective
A data center does not simply consume water.
It interacts with:
- Water infrastructure
- Energy infrastructure
- Wastewater systems
- Municipal planning
- Technology selection
- Community priorities
- Environmental conditions
The question therefore becomes:
How should water demand be evaluated within the broader ecosystem rather than as an isolated facility metric?
TIIM seeks to move the conversation beyond water consumption alone and toward a more complete understanding of infrastructure interactions, tradeoffs, and opportunities.
Supporting Diagrams
Traditional Water Evaluation
Potable Water
↓
Data Center Cooling
↓
Cooling Tower
↓
Blowdown
↓
Wastewater Treatment
Key Question:
How much water is consumed?
Ecosystem Water Evaluation
Water Sources
↓
Infrastructure System
↓
Cooling Architecture
↓
Water Use
↓
Wastewater Generation
↓
Resource Recovery Opportunities
↓
Community & Infrastructure Outcomes
Key Question:
How does the infrastructure system interact with the ecosystem?
Water-Energy Tradeoff Framework
Cooling Technology
↓
Water Consumption
↔
Energy Consumption
↓
Community & Infrastructure Impacts
Key Question:
What tradeoffs create the best overall outcome?
Supporting Research
University of Texas at Austin
Research examining AI infrastructure growth, water demand, power requirements, cooling technologies, and infrastructure constraints.
Related Topics
- Understanding Data Center Water Demand
- Cooling Technology Tradeoffs
- Community Impact Assessment
Links
- Data Centers Are Growing in Texas, But Big Questions Remain About Water Use (UT Austin)
- Water Use Requirements for Data Centers in Texas (UT Bureau of Economic Geology)
Lawrence Berkeley National Laboratory
Research evaluating water use intensity and operational characteristics of data center workloads.
Related Topics
- Water Demand
- Infrastructure Performance
- Resource Management
Links
- The Water Use of Data Center Workloads: A Review and Assessment of Key Determinants
- The Environmental Footprint of Data Centers in the United States
- 2024 United States Data Center Energy Usage Report
Municipal Wastewater Reuse Research
Studies examining the use of treated municipal wastewater as an alternative water source for industrial and infrastructure applications.
Related Topics
- Municipal Wastewater Reuse
- Water Recovery Technologies
- Community Impact Assessment
Links
- Use of Treated Municipal Wastewater as Power Plant Cooling Water
- Evaluation of the Combined Cost of Makeup Water Treatment and Reuse in Thermoelectric Power Plants (DOE/OSTI)
- Utilization of Municipal Wastewater for Cooling in Thermoelectric Power Plants
- Reclaimed Water Could Help Power Plants Run More Efficiently
Additional Reference
For broader context on data center water consumption and measurement methodologies: