Sustainable Data Centers: Metrics and Trade-offs

Sustainable Data Centers: Measuring and Reducing Environmental Impact

Sustainability in data centers is no longer just a branding question. It has become an engineering, operations, and reporting problem shaped by power constraints, water constraints, and rising demand for computing. The International Energy Agency (IEA) projects global electricity consumption for data centres will roughly double to around 945 TWh by 2030 in its Base Case, with rapid growth from 2024–2030.

At the same time, requests for comparable reporting are increasing. In the European Union, Commission Delegated Regulation (EU) 2024/1364 sets a reporting mechanism and measurement methodology for data centres above a threshold (installed IT power demand of at least 500 kW), along with specified reporting deadlines and key performance indicators.

This article explains what “sustainable data centers” means in practice, which metrics are used (and what they miss), and how design and operational choices affect outcomes. It also provides a structured way to evaluate sustainability claims without relying on a single number. This is informational content only, not legal or engineering advice.

Explain: What “Sustainable” Means in a Data Center

Why sustainability pressure has increased

Data centers have always required careful power and cooling design, but the sustainability debate has sharpened because demand is rising and constraints are more visible to utilities, regulators, and communities. The IEA’s analysis links growth in data centre electricity use to both AI adoption and broader digital services, with strong growth projected through 2030.

In the United States, the Energy Act of 2020 requested an updated public estimate of data center energy use, leading to a major Lawrence Berkeley National Laboratory (LBNL) update that extends historical estimates and frames future scenarios.

Four impact areas to measure

A practical definition of a sustainable data center usually includes:

• Energy: total electricity use and how much overhead is required to deliver IT services

• Carbon: emissions associated with purchased electricity (and any on-site fuels), plus lifecycle impacts where tracked

• Water: water used directly for cooling/humidification, interpreted in local context

• Materials: embodied impacts and circular handling of IT equipment and construction materials

Many public discussions reduce sustainability to a single metric (often PUE). That metric matters, but it does not represent carbon intensity or water stress on its own.

Key metrics and what each one can and can’t answer

Power Usage Effectiveness (PUE)

PUE compares total data center energy to IT equipment energy. It is widely used and has detailed measurement guidance, including notes on metering and boundaries.

What PUE answers well: how much facility overhead (cooling, power conversion, lighting, and other non-IT loads) exists relative to the IT load.

What PUE does not answer: how carbon-intensive the electricity is, or how much water the site uses.

Water Usage Effectiveness (WUE)

WUE was developed to quantify water used by a data center relative to IT energy, highlighting that water can be a siting and operations constraint.

WUE helps teams track water impacts consistently, but interpretation still depends on context (climate, cooling design, water source type, and what is included in “water use”).

Scope 2 emissions reporting for purchased electricity

Most data centers are electricity-intensive, so Scope 2 accounting is central. The GHG Protocol Scope 2 Guidance defines two reporting methods:

• Location-based: reflects the average emissions intensity of the grids where consumption occurs

• Market-based: reflects emissions associated with electricity that an organization has chosen through contractual instruments (when they meet defined quality criteria)

The same site can report different Scope 2 totals depending on method and procurement choices. The Guidance also describes “dual reporting” and notes that companies in markets with supplier/product-specific data “shall report scope 2 emissions in two ways.”

Reporting and transparency trends

In the EU, Commission Delegated Regulation (EU) 2024/1364 describes a reporting mechanism that collects information and key performance indicators, and it lists KPI themes that include energy consumption, power utilisation, temperature set points, waste heat utilisation, water usage, and renewables.

Inform: Design and Operations Levers That Change Outcomes

Start with the IT load because it shapes everything downstream

Facility efficiency improvements help, but IT energy remains the core driver of total demand. Public best-practice guidance often starts with IT measures because they affect cooling and electrical sizing.

Common IT-side levers include:

• Removing unused servers and storage (“decommissioning”)

• Consolidating workloads to raise utilization where appropriate

• Selecting more efficient hardware generations for performance-per-watt gains

• Using power management settings and scheduling to reduce idle draw

These changes usually require coordination between IT and facilities teams, because a shift in rack density changes cooling requirements and sometimes changes the preferred cooling technology.

Cooling strategy and environmental control

Cooling often represents a large share of non-IT energy, and it can also be a major driver of water use depending on design. The U.S. DOE best-practices guide covers air management, cooling systems, electrical systems, and heat recovery, linking design choices to energy and water outcomes.

Air management and setpoints

Good airflow management (containing hot/cold air streams and limiting bypass air) can reduce fan energy and allow higher supply temperatures. Temperature and humidity strategies should stay within recognized equipment environmental guidance.

Liquid cooling and higher-density computing

As densities rise, liquid cooling options become more common. ASHRAE TC 9.9 has published materials that discuss the growing role of liquid cooling in data centers and related design considerations.

Liquid cooling can reduce the need to move large volumes of air, but it introduces different operational practices (leak management, maintenance procedures, water chemistry where applicable, and vendor interoperability).

Electrical system efficiency and reliability trade-offs

Power conversion and distribution losses add up at scale. Design guidance highlights UPS, transformer, and distribution choices, as well as how redundancy strategies affect operating efficiency.

A sustainability evaluation typically needs to treat resilience choices (redundancy level, operating load factor, maintenance approach) as part of the overall impact picture, not as an afterthought.

Electricity sourcing and the timing problem

Reducing facility energy is only one side of carbon impact. The other is the carbon intensity of electricity at the hours when the site consumes power.

Annual matching vs hourly matching

Many organizations match annual electricity use with renewable procurement instruments. That can support clean generation, but annual matching does not prove hourly alignment between consumption and carbon-free supply.

What “24/7 carbon-free energy” tries to measure

The UN-aligned definition of 24/7 carbon-free energy describes matching every kilowatt-hour of electricity consumption with carbon-free electricity sources every hour of every day.

This concept increases transparency about time and location, but it also increases data requirements and may shift strategy toward storage, flexible demand, and more granular procurement.

Water stewardship and local constraints

Water impact is highly site-specific. Some cooling designs reduce electricity use by using evaporation, which can increase direct water use. WUE provides a consistent way to measure, but interpretation still needs local context.

A responsible evaluation asks:

• What water source is used (and under what permits/constraints)?

• How does seasonal water availability line up with peak cooling demand?

• Are there alternatives that reduce dependence on water during scarcity periods?

Waste heat reuse and systems thinking

Waste heat reuse can improve outcomes beyond the data center boundary when there is a nearby heat user and workable infrastructure. The EU framework explicitly includes “waste heat utilisation” among the KPI themes used to assess sustainability elements.

Whether heat reuse is practical depends on geography, temperature level, demand stability, and commercial agreements with the receiving system.

Materials and equipment lifecycle

Operational electricity often dominates environmental impact, but materials still matter—especially with rapid build cycles and frequent hardware refresh.

The Global E-waste Monitor 2024 describes the scale of the e-waste challenge and why circular handling of electronics remains a major issue globally.

For data centers, lifecycle practices typically include:

• Extending useful life where reliability and efficiency targets allow

• Reuse/refurbishment programs for parts and equipment

• Documented end-of-life handling with qualified recyclers

Practical Insight: How to Evaluate a Sustainability Plan

Step 1: Define boundaries before comparing numbers

Start with three boundary decisions:

• Asset boundary: one building, a campus, or a portfolio

• Energy boundary: what counts as “total facility energy” and where IT energy is metered (important for PUE)

• Time boundary: monthly vs annual reporting, and whether hourly electricity matching is in scope (for 24/7 approaches)

Without boundaries, comparisons often turn into “metric shopping,” where different sites use different methods and still claim to be comparable.

Step 2: Use a small, balanced metric set

A workable starting set for many operators includes:

• PUE (facility overhead indicator)

• WUE (water intensity indicator)

• Scope 2 emissions reported using both location-based and market-based methods (where applicable)

• Lifecycle indicator (documented reuse/refurbishment approach, or an LCA scope for major builds where used)

This set prevents one metric from hiding trade-offs in another category.

Step 3: Build an improvement roadmap with two lanes

The DOE best-practices guidance is helpful for separating actions into:

Operational tuning (lower disruption)

• airflow management improvements

• setpoint review within recognized equipment guidelines

• improved metering and monitoring coverage

• maintenance planning that avoids unnecessary simultaneous peaks

Retrofit/new-build measures (higher change)

• cooling system changes and economization choices

• power distribution improvements

• liquid cooling support for higher-density computing where appropriate

• heat reuse feasibility assessment

Step 4: Ask vendors and colocation providers for evidence, not slogans

If you rely on a colocation provider, request:

• PUE measurement method and metering points (and reporting cadence)

• WUE boundary and how water is measured

• Scope 2 reporting method(s) and what contractual instruments are used for market-based claims

• Whether an energy management system is in place (ISO 50001 is one recognized framework)

Step 5: Know the direction of reporting requirements

In the EU, Delegated Regulation (EU) 2024/1364 sets a reporting schedule (including deadlines listed in the regulation) and applies to operators of data centres meeting the installed IT power demand threshold.

Even outside Europe, the general direction is toward clearer, more standardized disclosures and better-defined measurement boundaries.

Skills and careers: where students and early-career readers can contribute

Sustainable data centers involve multiple roles:

• Facilities operations (monitoring, maintenance, incident response)

• Electrical engineering (power distribution, redundancy, metering strategy)

• Mechanical engineering (cooling, heat rejection, water systems)

• Sustainability reporting (Scope 2 methods, documentation quality)

• Data and measurement support (dashboards, anomaly checks, meter reconciliation)

A good entry-level portfolio project is a mock monthly sustainability report for a hypothetical site that shows: a PUE calculation boundary, a WUE boundary, and both Scope 2 methods with clear assumptions.

Outcomes and Limitations: Reading Claims Clearly

What stronger sustainability practice typically changes

When measurement is consistent and improvements are prioritized, teams often see:

• fewer unknowns about where energy and water are going (better metering and accountability)

• clearer conversations with utilities and stakeholders about growth and constraints

• better comparability over time because boundaries stay stable

Common limitations and trade-offs

• PUE does not represent carbon or water. It is an efficiency ratio, not a full sustainability score.

• Water outcomes depend on local context. WUE improves tracking, but climate and water availability still matter.

• Scope 2 totals depend on method. Location-based and market-based totals answer different questions and should be labeled clearly.

• Hourly matching is harder than annual matching. 24/7 approaches add precision, but raise complexity and data needs.

• Materials impacts are harder to compare. Hardware refresh cycles and end-of-life practices vary widely, and global e-waste handling remains uneven.

A quick checklist for evaluating sustainability claims

Look for:

1. The boundary (site vs portfolio; calendar year vs hourly)

2. The metric definitions and measurement points (especially for PUE and WUE)

3. Scope 2 method labeling (location-based and market-based) and instrument quality criteria if market-based claims are used

4. Acknowledged trade-offs (energy–water, density–cooling choice, reliability–losses)

5. Evidence of continuous improvement practices (energy management approach and governance)

Conclusion

Sustainable data centers are defined by measurement boundaries and trade-offs, not by a single headline metric. PUE helps quantify facility overhead, WUE helps quantify direct water intensity, and Scope 2 reporting methods help clarify how electricity-related emissions are calculated and reported.

Design and operations choices—IT efficiency, cooling approach, power distribution design, and electricity sourcing strategy—shape outcomes differently depending on climate, grid conditions, and workload density. For readers comparing sites or providers, the most reliable approach is to ask for consistent boundaries, clearly labeled methods, and evidence that metrics are tracked over time.

FAQs

What is the simplest working definition of a sustainable data center?

A data center that measures and reduces impacts across energy use, electricity-related emissions, water use, and materials lifecycle using clear boundaries and repeatable methods.

Is a lower PUE always better?

Lower PUE usually means less facility overhead per unit of IT energy, but it does not indicate the carbon intensity of electricity or the site’s water use.

How is WUE defined in a standard way?

WUE is defined as a data centre KPI in ISO/IEC 30134-9 and is also described in The Green Grid’s WUE guidance.

Why do Scope 2 emissions totals differ between organizations?

Because Scope 2 can be reported using location-based and market-based methods, and market-based totals depend on the contractual instruments and quality criteria used.

What does EU data center sustainability reporting cover at a high level?

Delegated Regulation (EU) 2024/1364 describes a reporting mechanism and KPI themes that include energy consumption, power utilisation, temperature set points, waste heat utilisation, water usage, and renewables for qualifying sites.

Reference

• International Energy Agency (IEA). Energy and AI: Energy demand from AI (web report section).

• European Commission. Commission Delegated Regulation (EU) 2024/1364 (Official Journal PDF).

• GHG Protocol (WRI/WBCSD). Scope 2 Guidance (location-based and market-based methods; dual reporting).

• The Green Grid. WP#49 PUE: A Comprehensive Examination of the Metric (measurement considerations).

• ENERGY STAR. Recommendations for Measuring and Reporting Overall Data Center Efficiency (Version 2) (PUE reporting guidance).

• U.S. Department of Energy (FEMP). Best Practices Guide for Energy-Efficient Data Center Design (PDF).

• Lawrence Berkeley National Laboratory. 2024 United States Data Center Energy Usage Report (publication and PDF).

• ASHRAE TC 9.9. Thermal Guidelines for Data Processing Environments (Reference Card).

• The Green Grid. WP#35 Water Usage Effectiveness (WUE) Usage Guidelines (PDF).

• ISO. ISO/IEC 30134-9: Water usage effectiveness (WUE) (standard abstract).

• ISO. ISO 50001 — Energy management (energy management system overview).

• ITU. The Global E-waste Monitor 2024 (publication page and PDF).

• United Nations. 24/7 Carbon-Free Energy Compact (definition).