Beyond power: Why water efficiency is becoming the next big data center metric
For years, sustainability in data centers has been defined primarily by energy efficiency. Power Usage Effectiveness (PUE) emerged as the industry’s standard metric, shaping how facilities were designed, operated and evaluated. Improvements in PUE became synonymous with progress toward greener and more efficient infrastructure.
However, this energy‑centric approach has overlooked another essential resource that is critical to data center operations—water. Used extensively for cooling and thermal management, water has enabled higher compute densities and operational stability, yet it has remained largely absent from mainstream efficiency frameworks.
As artificial intelligence (AI) workloads scale rapidly and global water stress intensifies, this imbalance is becoming increasingly evident. Data centers are entering a phase where water efficiency must be measured, managed and optimized with the same rigor as energy efficiency.
A sustainability blind spot in data center operations
The industry’s focus on PUE has delivered measurable benefits. Optimized airflow, advanced cooling systems and intelligent controls have reduced energy waste and improved operational efficiency. These advancements were both necessary and impactful.
At the same time, many energy‑efficient cooling strategies—particularly evaporative cooling—depend heavily on water. In several cases, improvements in PUE have been achieved by increasing water consumption, effectively shifting environmental impact rather than reducing it.
For years, this trade‑off received limited attention. Water was often assumed to be readily available and low risk. That assumption is no longer valid.
How AI is accelerating the water imperative
The rapid adoption of AI has fundamentally altered the operational profile of data centers. AI training and inference workloads generate significantly more heat than traditional enterprise applications. High‑density GPU clusters and racks exceeding 40 kW are becoming increasingly common, placing unprecedented demands on cooling infrastructure.
To maintain performance and reliability, many facilities rely on water‑intensive cooling methods. As AI infrastructure scales, this reliance is translating into materially higher water consumption.
According to a Morgan Stanley analysis, the expansion of AI‑optimized data centers could drive annual water consumption to over 1,000 billion liters by 2028, representing a multi‑fold increase from current levels as facilities scale cooling capacity and electricity generation to support AI workloads1. The report highlights water availability as an emerging constraint for data center growth, particularly in regions already experiencing water stress.
As AI becomes integral to enterprise and cloud strategies, water availability and efficiency are emerging as material considerations for scalability, resilience and long‑term sustainability.
Understanding Water Usage Effectiveness (WUE)
To manage water responsibly, it must first be measured. Water Usage Effectiveness (WUE) provides a standardized way to assess water consumption in relation to IT output.
WUE is defined as:
WUE = Annual site water usage (liters) ÷ Annual IT energy consumption (kWh)
While PUE measures how efficiently a data center uses electricity, WUE measures how efficiently it uses water. Together, these metrics provide a more complete view of operational sustainability.
WUE enables organizations to:
- Understand the water impact of cooling strategies
- Compare facilities across regions with different water risks
- Identify hidden trade‑offs between energy and water efficiency
- Support informed decision‑making for infrastructure design and workload placement
In high‑density, AI‑enabled environments, this visibility is increasingly important.
Why WUE is gaining strategic importance
Several converging factors are elevating WUE from a secondary consideration to a strategic priority:
- Rising water scarcity: Many current and emerging data center hubs—including India, Southeast Asia, the Middle East and parts of North America—are classified as water‑stressed regions. As data center capacity expands, water availability can directly affect operational continuity and future growth.
- Increased cooling demands from AI: AI workloads intensify thermal output, often leading to higher water consumption when traditional cooling methods are used. Without WUE measurement, organizations risk underestimating long‑term sustainability exposure.
- Evolving regulatory expectations: Regulatory frameworks are expanding beyond energy and carbon to include water usage and conservation. In several regions, water reporting and efficiency targets are becoming formal compliance requirements.
- Investor and ESG scrutiny: Environmental, social and governance (ESG) frameworks increasingly emphasize water stewardship. Enterprises operating water‑intensive infrastructure face growing expectations for transparency and responsible management.
- Industry leadership signals: Major cloud and technology providers are publishing water metrics, committing to water replenishment goals and piloting low‑ or zero‑water cooling designs. These actions indicate where industry benchmarks are heading.
Collectively, these trends suggest that WUE will soon carry comparable weight to PUE in evaluating data center sustainability.
Technologies supporting water‑efficient data centers
Advances in infrastructure design and operations are making it possible to reduce water usage without compromising performance:
- Liquid cooling technologies, including direct‑to‑chip and immersion cooling, reduce reliance on evaporative cooling
- Closed‑loop cooling systems minimize freshwater intake and enable water reuse
- AI‑driven cooling optimization dynamically adjusts cooling based on real‑time conditions
- Waste heat reuse reduces overall cooling demand by repurposing excess heat
These approaches are increasingly central to the design of AI‑ready, resource‑efficient facilities.
With the rise of AI-ready, high-density computing, these insights are increasingly critical.
Enabling water‑efficient computing through hybrid cloud
As organizations modernize their infrastructure and adopt hybrid cloud models, water efficiency must be integrated into planning and operations. HCLTech supports enterprises in this transition by addressing water efficiency across the infrastructure lifecycle:
- Assessment and benchmarking of PUE, WUE and regional water risk
- Cooling optimization and modernization aligned with AI workloads
- Sustainability‑aligned architecture and workload placement strategies
- ESG reporting and analytics for water, energy and carbon metrics
- AI‑led operational optimization to reduce overconsumption
By embedding water considerations into cloud and data center strategies, organizations can strengthen resilience while advancing sustainability objectives.
Redefining efficiency for the AI era
As data centers evolve to support AI‑driven workloads, sustainability can no longer be evaluated through a single metric. While Power Usage Effectiveness remains a critical measure of energy efficiency, it does not fully capture the environmental impact of modern, high‑density infrastructure.
Water Usage Effectiveness provides the missing perspective. It brings visibility to cooling trade‑offs, regional water risk and long‑term operational resilience—factors that are increasingly material as AI adoption accelerates and water scarcity intensifies.
Organizations that incorporate WUE into infrastructure planning and cloud strategies will be better positioned to manage risk, meet emerging regulatory expectations and build sustainable digital platforms at scale. In the AI era, responsible computing will be defined not only by how efficiently power is used, but by how carefully water is managed.
