The EU Has Proposed a New Data Centre Rating System—Which Metering and Audit Gaps Need to Be Identified?

Introduction

• On 21 September 2026, the European Commission presented the Data Centre Energy Efficiency Package.
• The initiative includes a proposal for a common EU data centre sustainability rating system.
• Comparability depends on the ability to measure energy and water use, renewable energy and waste heat reuse.
• Operators therefore need to review not only their KPIs but also the traceability of their metering and data-management processes.

The European Commission’s package of 21 September 2026 includes a common EU sustainability rating system, an assessment of data centre information reported in 2024, and a consultation on possible future minimum performance requirements. From an engineering perspective, the most important message is that the value of energy-efficiency claims depends directly on the reliability of the underlying measurements.

This is not simply a reporting issue. If submetering is incomplete, measurement boundaries are unclear or the data cannot be reproduced, upgrade decisions, site comparisons and post-investment evaluations may also become uncertain.

Who is this for, and what is the technical focus?

This article is intended for data centre operators, facility managers, IT infrastructure managers, project owners, technical procurement teams and consultants. Its focus is not legal interpretation, but an engineering assessment of how complete, comparable and auditable a data centre’s energy and water-use data are.

The issue is particularly important in operational facilities, where monitoring systems may have been installed during different technology cycles, the load structure may have changed over time, and upgrades must be completed without compromising operational resilience.

What did the European Commission announce?

The Data Centre Energy Efficiency Package proposes the development of a common EU sustainability rating system. According to the available Commission information, the aim is to make data centre performance more comparable based on factors including energy use, water use, renewable energy and waste heat reuse.

The package also builds on the assessment of data reported in 2024 and provides for consultation on possible future minimum performance requirements. This does not replace a separate assessment of the requirements applicable to each operator, but it indicates a clear technical direction: sustainability indicators need to be supported by a verifiable measurement chain.

What should an energy-efficiency audit identify?

The first task of a targeted data centre energy efficiency audit is to map the metering architecture. Listing the meters is not enough. The audit should document which load groups they cover, where they are located on the single-line diagram, which system collects the data, and how frequently readings are sampled.

The audit should consider verifying:

• the metering point for the facility’s total energy input;
• the measurement boundary assigned to the IT load;
• the relationship between UPS-, PDU- and rack-level data;
• the separation of cooling, pumping and air-handling energy use;
• data associated with generators, energy storage or on-site generation;
• water-metering points and the boundaries of individual process circuits;
• the completeness, time synchronisation and retention of time-series data;
• meter identifiers, condition and available verification documentation.

The result should be a meter-point register and system diagram that allow an independent technical reviewer to understand how a reported indicator was produced.

PUE, WUE and other indicators: assumptions behind the numbers

To interpret PUE, the facility’s total energy consumption and the energy consumption assigned to IT equipment must be clearly defined. The audit should also record whether the calculation is based on an instantaneous reading, a monthly aggregate or a longer reporting period.

For WUE, the system boundary for water use is equally important. The result may vary depending on which process circuits, water-treatment losses or other site uses are included in the data. Two apparently identical indicators may therefore represent different conditions if their measurement boundaries or reporting periods differ.

Assessing renewable energy and waste heat reuse also requires energy flows to be separated. From an engineering perspective, it is necessary to demonstrate which physical metering points or documented data sources support the reported quantities and whether double counting can be avoided.

What does this mean for operations?

Reliable metering must not compromise the availability of critical infrastructure. The installation of new meters, current transformers, communications gateways or water meters should therefore be planned with electrical and mechanical redundancy, maintenance routes and scheduled shutdown windows in mind.

A phased approach is often appropriate in an operational data centre. The process may begin with a review of existing data and documentation, followed by temporary metering and then the installation of permanent submetering points. Every change requires a rollback plan, formal change management and verification of the affected monitoring alarms.

Better data quality provides direct operational benefits by making part-load losses, cooling-system deviations and unused capacity more visible. However, the order of interventions should always be determined through a combined assessment of potential savings, operational risk, maintainability and achievable shutdown windows.

Common mistake: KPIs without context

A common problem is comparing data centres with different loads, designs or climatic conditions using a single PUE or WUE value. The number alone does not reveal measurement uncertainty, data gaps or changes in operating conditions.

Automatically combining BMS, DCIM, UPS and utility metering data is also risky. If timestamps, units of measurement or sampling intervals differ, the aggregated KPI may be technically calculable without necessarily being comparable. The audit should therefore also identify data transformations and manual corrections.

Recommended next step

Preparation should begin with a limited-scope metering and data-quality assessment:

1. Define the site-level and system-level boundaries of the assessment. 2. Create an inventory of electrical, cooling and water-side metering points. 3. Compare the metering points with single-line diagrams and mechanical schematics. 4. Review the data flow for at least one representative period, including gaps and anomalous values. 5. Document the calculation methods used for PUE, WUE and other indicators. 6. Rank metering gaps by risk and decision-making value. 7. Prepare a phased improvement plan covering cost, shutdown and redundancy impacts.

In Digital Technologies’ engineering approach, the purpose of an audit is not to produce an isolated score. The key outcome is a traceable baseline that can support maintenance, upgrade and capacity-expansion decisions. This may include an integrated assessment of critical power supply, UPS systems, distribution, cooling, the rack environment and associated monitoring.

Conclusion

The proposed EU rating system increases the importance of auditable data centre operations. The key question is not whether an energy-efficiency indicator is available, but whether all its input data, measurement boundaries and calculation steps can be verified.

Operators that identify metering gaps, standardise data sources and establish a documented baseline now will be better positioned to prioritise future interventions. Improvements must nevertheless always be implemented with due consideration for operational resilience, redundancy and the constraints of a live environment.