Data Centre Planning 2026: Energy Efficiency Starts with the Feasibility Study

Sound data centre planning starts by identifying the digital services the facility must reliably deliver. These define requirements for IT capacity, availability, location and operations. The feasibility study combines them into a basis for decision-making: which solution meets the need, what dependencies exist, and which assumptions require confirmation before investment? Energy efficiency belongs in that assessment from the outset.

In this article

What a data centre feasibility study should deliver

A suitable location and sufficient space are important prerequisites. For a sound decision, however, we recommend assessing IT, building services and operating arrangements together. A building may offer suitable space while its grid connection, service routes or necessary expansion remain unresolved. Conversely, a technically attractive new build may involve unnecessary effort compared with targeted modernisation.

The outcome should therefore be more than a preferred option. A robust feasibility study documents the starting position, comparable solutions, clear exclusion criteria and a reasoned recommendation. Outstanding matters remain visible: a preliminary grid enquiry is not treated as a binding connection commitment, and a potential heat recipient is not counted as heat reuse already in operation.

This provides management and project leaders with a basis for decisions that can be reviewed and updated. Requirements for the subsequent DC-Planning are also recorded so they can be carried through into design development, tendering and implementation.

Distinguish IT load, connection capacity and energy demand

Rack count initially describes a space requirement. Electrical and thermal design also need a load profile. Two equally sized IT areas can have very different requirements if one runs conventional business applications and the other intensive AI workloads. A standard power allowance per rack may be a useful starting hypothesis, but should explicitly be identified as an assumption.

We recommend three separate starting figures: expected IT power at initial operation, planned power in later phases and the timing of actual use. Add peak loads, reserves and maintenance scenarios. This makes clear what capacity is needed immediately and which expansion options should remain open.

Kilowatts and kilowatt-hours must not be confused. Power describes an instantaneous demand; energy accumulates over time. A high available connection capacity does not mean it will be fully used throughout every operating hour. Annual assessment requires the assumed operating states and their duration.

Worked example: what a PUE comparison tells you

PUE is the ratio of total data centre energy use to IT energy use within a defined measurement boundary. An initial scenario can use an assumed constant IT load. The following example is a model calculation, not a forecast for a specific project.

Assume average IT power of 100 kW and 8,760 operating hours a year. This gives 876 MWh of IT energy. At an assumed annual PUE of 1.50, total energy use is 1,314 MWh. At an assumed annual PUE of 1.20, it would be 1,051.2 MWh. The calculated difference is 262.8 MWh, or 20 per cent of the first option’s total energy use.

This result applies only under the stated assumptions: identical IT work, identical IT energy and comparable system boundaries. It neither proves that a particular installation will achieve these annual values nor that every environmental impact falls proportionally. Investment decisions need project-specific assessment of load profiles, cooling and auxiliary energy.

Our data centre energy calculator lets you vary the starting values. Use it for orientation and to prepare questions for the design team. Results do not replace technical design or operational measurement evidence.

Assess the site and expansion together

We recommend comparing sites against the same criteria: available electrical power, building suitability, services, access and expansion opportunities. Assess current demand and a credible growth path separately. A solution designed exclusively for the final build-out may behave differently during initial operation.

For a modular approach, the key question is what else must expand when a module is added. Is existing supply sufficient? Are interfaces provided? Can work proceed without compromising agreed operating objectives? The BLUE DC concept is assessed against these requirements during project planning; the appropriate implementation follows from the specific project.

For an existing facility, compare documentation with the actual condition. Are reserves verified or were they merely originally intended? Which components need replacement? Where could work affect ongoing operations? Address these questions before comparing refurbishment and new build as equivalent alternatives.

Turn availability goals into testable requirements

A general goal such as “maximum availability” leaves many design questions unanswered. Describe the operating situations that must be managed instead. Which applications may be interrupted, and for how long? Which maintenance tasks must be possible during operation? Which supply paths or components may fail simultaneously without breaching the defined objective?

We recommend developing these requirements with IT and operations. Then assess how each technical option meets them and what organisational arrangements it requires. Spares, access rules, maintenance windows and recovery procedures become part of the decision. Additional equipment alone is not a complete operating concept.

If assessment or certification against a standard is planned, specify its edition, relevant parts and target scope in the project brief. A general reference to EN 50600 does not replace this definition. Technical design and the subsequent evidence process can then be aligned.

Energy efficiency and regulatory requirements in 2026

Germany’s Energy Efficiency Act includes specific provisions for data centres. The definition in Section 3 refers, among other factors, to non-redundant rated electrical connection power of at least 300 kW. Sections 11 to 13 address energy efficiency, waste heat, management systems and information obligations. Applicability must be checked against the facility, operator, commissioning date and exemptions. Source: German Energy Efficiency Act, current statutory text.

Data centre performance reporting is also relevant at European level. The European Commission describes a reporting system for energy and water indicators and refers to Delegated Regulation EU/2024/1364. Its topic page presents further European measures as being prepared. Planning should monitor these initiatives and distinguish them from requirements already in force. Source: European Commission, Energy performance of data centres.

In practice, maintain a requirements register with sources, responsibilities and intended evidence. Update it at defined project milestones. This shows whether a statement is based on legislation, a voluntary criteria catalogue or a project’s own objective. An environmental label and a legal obligation are assessed separately.

Include heat reuse and cooling in the options early

For possible heat reuse, consider the data centre’s and recipient’s demand profiles together. Temperature, availability over time, distribution routes and transfer equipment help determine whether an idea can become a workable solution. Explicitly identify required agreements and unconfirmed conditions in the decision paper.

Apply the same approach to cooling: compare concepts under the same site and load assumptions, including expected part-load operation and relevant exceptional conditions. Projects seeking the Blue Angel label can find further preparation guidance in our technical article on the July 2026 criteria .

Documents to prepare at the start

  • Demand: Applications, current IT load, expected development and significant peaks.
  • Operations: Availability targets, maintenance requirements and responsibilities.
  • Location: Plans, floor areas, existing infrastructure and reliable connection information.
  • Measurements: Available energy and power data, including period and measurement boundary.
  • Project objective: Target opening date, expansion phases and intended assessments.

Missing documents do not prevent you from starting. Record them as open tasks and distinguish reliable information from assumptions and claims awaiting verification. A good feasibility study makes uncertainty manageable instead of hiding it behind apparently precise numbers.

From feasibility study to delivery

High Knowledge connects IT, building services and operations. We develop options together, assess dependencies and define the next planning steps. Learn more about our approach and services in data centre feasibility studies and DC-Planning. To discuss a specific project, share your starting position through our contact page .

Editorial review: 8 September 2026. Calculations explicitly use model assumptions. Legal requirements must be determined for the specific project using the applicable version.