A new edition of the Blue Angel criteria for data centres is available: DE-UZ 228, July 2026. For operators, public buyers and project teams, this is a concrete reason to review design documents and evidence. Projects seeking the label should consider the criteria early. This article explains the published changes and how to prepare for new builds, modernisation and colocation.
In this article
- Blue Angel 2026: what has changed?
- First define responsibilities
- Measurement data: a dashboard does not replace traceable measurement
- Plan heat reuse as a shared infrastructure project
- Assess cooling and water demand together
- Bring IT into the planning process
- Checklist for initial project preparation
- Common questions in data centre planning
- Turn the criteria into a sound design
Blue Angel 2026: what has changed?
The Environmental Label Jury approved the revision in June 2026. Its official announcement identifies more precise water-demand information, stricter requirements for heat reuse and renewable energy, treatment of on-site fossil energy generation, consideration of small data centres in efficiency metrics and a new IT utilisation indicator. The update therefore covers both building infrastructure and how IT is used. Source: Blue Angel announcement of 17 June 2026.
The new edition is valid until 31 December 2029. The complete criteria, including exemptions and required evidence, are decisive. For example, the fossil-fuel provision includes an exception for standby generation used to maintain availability. A shortened claim such as “backup power is no longer permitted” would therefore be wrong. Source: DE-UZ 228, July 2026 edition, version 1, Section 3.2.4 and footnote 6.
For existing projects, carry out a documented comparison: which edition underlies the design, which criteria apply to the intended application, and which documents are current? Publication alone does not establish contractual or transitional arrangements for an already labelled site. Clarify these against the specific label-use contract.
First define responsibilities
The award criteria distinguish requirements for all applicants, data centre operators, and colocation customers or operators of their own IT. Measurement and evidence arrangements must match the role. An independent auditor appointed for the procedure reviews the application’s energy-efficiency report. RAL gGmbH assesses label-use applications and concludes the corresponding contracts. Source: DE-UZ 228, Sections 1.5, 2 and 3.1.2.
Our planning recommendation: record on one page who is responsible for the building, power supply, cooling, servers, storage and measurement data. Add a contact and the location of reliable documents for each area. Especially in mixed operating models, this prevents building management, IT and landlords from referring responsibility back and forth. It also supports coordination with procurement, energy management and external providers.
Measurement data: a dashboard does not replace traceable measurement
To prepare the project, check data along the energy flow. Start with existing electrical and cooling schematics. Map each relevant value to its meter, unit, recording period and data owner. Then select a month and check that values remain traceable from source systems to analysis.
A common pitfall is a changed measurement boundary: assigning a consumer to the building first and later to IT can improve an indicator without reducing energy use. Missing periods, unexplained meter replacements and inconsistent timestamps are equally problematic. Document their causes and treatment. A plausible number with no traceable origin is less useful for investment decisions than a clearly explained data gap.
This work is useful even before an application is planned. It establishes comparable operating data for later assessment of technical measures. New sensors should meet a defined information need; extra measurement points alone improve neither data quality nor operations.
Plan heat reuse as a shared infrastructure project
When preparing heat reuse, start with the possible recipient. Is heat needed throughout the operating year or mainly in winter? At what temperature? Who would operate a heat pump? Where are the transfer points and how far must a pipeline run?
Consider a data centre next to a building with high winter heating demand. Proximity is only a starting point. A workable solution needs aligned load profiles, a technical connection and an operating agreement. If summer demand is absent, you cannot simply count all annual waste heat as usable. The heat-rejection strategy should also account for maintenance and outages at the recipient.
The decision paper should distinguish technically available heat, expected heat actually taken and outstanding prerequisites. A planned interface only becomes demonstrable reuse once the connection is implemented and operated.
Assess cooling and water demand together
One favourable metric should not determine the whole cooling decision. Use the same IT-load, site and operating-hour assumptions for every option. Then assess electrical energy, water demand, maintenance and behaviour at high outdoor temperatures together.
This gives procurement concrete questions: which figures apply at expected part load, which are calculations and which measurements, what conditions does the manufacturer assume, and what changes with later expansion? Include the answers in the technical assessment before specifying a system for procurement.
For modular data centres, assess every expansion phase. A concept that makes sense at full build-out should also work coherently during initial operation. Where IT demand grows slowly, start-up operation and expansion strategy become significant design tasks.
Bring IT into the planning process
For operators of their own IT, connect infrastructure planning with the applications actually needed. Which services require continuous availability? Which tasks can move in time? Which systems can be consolidated without compromising agreed performance and availability?
Local AI adds further questions. A knowledge assistant handling short queries has a different use profile from extensive document processing. Expansion decisions should use measured workloads and planned peaks. Our article on sizing on-premises AI explains how to turn these into testable requirements.
Checklist for initial project preparation
The following checklist is our preparation recommendation. It does not replace the complete award criteria, but makes the next project steps concrete.
- Project and edition: Document the site, operating model, existing contractual position and criteria edition.
- Responsibility: Assign responsibility for IT, building services, energy, procurement and evidence.
- Data: Compile measurement schematics, available time series and identifiable gaps.
- Technical options: Compare cooling, power supply and possible heat reuse using the same assumptions.
- Actions: For each gap, define a concrete task, an owner and a realistic deadline.
- Procedure: Coordinate review steps, required documents and timing dependencies for the application.
Use a simple action register with four statuses: evidenced, in progress, clarification needed, and not applicable with reasons. It gives a more useful view of progress than a blanket statement that the data centre is already “ready”.
Common questions in data centre planning
Is a good PUE sufficient for the Blue Angel?
No. The published revision priorities already show that several areas must be considered together. Prepare a complete requirements register. PUE remains an important input to technical comparisons, but cannot replace that register.
When should preparation begin for a new build?
Ideally during the feasibility study, while the site, measurement concept and technical options can still be developed together. Reviewing these decisions only after procurement may require further coordination or changes to commissioned work. Include the required evidence periods in the programme from the outset.
Does BLUE DC automatically carry the Blue Angel label?
BLUE DC is High Knowledge’s modular data centre concept. Any claim about a label must refer to the specifically assessed project and its valid evidence. The product family alone does not provide that evidence. Objectives, implementation and the intended procedure are therefore agreed separately for each project.
Turn the criteria into a sound design
High Knowledge helps turn technical requirements, responsibilities and evidence into one coordinated project. Your site, IT and intended operations are the starting point. Learn more about planning for the Blue Angel, our data centre feasibility studies and the modular BLUE DC.
Editorial review: 8 September 2026. Linked primary sources were checked on that date. Practical suggestions support project preparation; binding requirements derive from the applicable award criteria and the specific procedure.
