Sustainability and local impact

Choose the evidence branch

Start with the type of claim. Ambition is a target or slogan. A planned solution is a design that is not yet run. A measured outcome is a dated reading with a boundary. The next page depends on the decision: compare PUE, test a heat customer, assess water and neighbours, or understand reporting duties.

Keep the Swedish project example attached to its source. Meta’s Luleå page, GleSYS’s Falkenberg heat-delivery notice, atNorth SWE01 and Conapto describe different projects and evidence states. None can fill the missing metric in another project.

• Ambition — a target year or slogan.
• Planned solution — a design that is not yet run.
• Measured outcome — a dated reading with a boundary.
• Open the leaf page that matches the decision, not the slogan.

How Swedish institutions frame these claims

Energy claims split into mix, connection and hall efficiency. Energimyndigheten describes the national electricity system. That mix is not a local connection offer. The network owner answers the property. Svenska kraftnät’s indicative transmission map does not replace that answer. Bidding areas SE1 Luleå, SE2 Sundsvall, SE3 Stockholm and SE4 Malmö show wholesale-price constraints. They do not allocate spare megawatts and they do not score a PUE. Skatteverket’s energy-tax pages are a tax file.

Water, noise and land sit with the municipality, Boverket’s Planning and Building Act guidance, and Naturvårdsverket’s Environmental Code guidance. Heat reuse sits with a named district-heat company and a contract state. Jobs sit with whoever published the estimate, usually the operator. Inspektionen för strategiska produkter screens certain investments; it does not certify sustainability copy. Keep each letterhead on its own paper.

Named examples already in the register

Meta’s Luleå page describes outdoor-air cooling and hydropower and cites an operator PUE of 1.09. Current megawatts are not on that page. GleSYS published a 1 October 2020 start of heat delivery to Falkenberg Energi, with 3 MW utility power in that release and 5 MW on a later operator page. atNorth SWE01 states design PUE 1.2, liquid cooling and heat recovery with Stockholm Exergi. Conapto’s south-campus materials mention recovery toward Stockholm Exergi.

EcoDataCenter’s Falun site page states PUE 1.2 according to ISO standards and 80 MW available power, and describes recovery with Falu Energi & Vatten. The 2025 annual report’s 41 MW and 65 MW sentences stay in separate rows. Microsoft’s Sweden community PDF discusses local projects and employment figures as company estimates. evroc’s 2025 notice designed for up to 16,000 GPUs and did not state megawatts. None of those sentences is a national heat-market total or a PUE league table.

What this hub does not prove

This hub does not rank Swedish halls on sustainability, does not add operator PUE figures, and does not invent a national heat or jobs total. A renewable-contract sentence is not a measured PUE. A prepared heat pipe is not delivered heat. A construction peak is not a permanent payroll. Read the leaf pages, keep the three states visible, and leave empty cells empty.

Classify the claim before you repeat it

Sustainability language becomes decision-useful when its claim type is visible. An ambition is a target or intention. A planned solution is a design that may not yet be built or operated. A measured outcome is a dated result with a stated boundary and method. These are not stages on one automatic ladder. A renewable-electricity statement, a design PUE, a prepared heat connection and a measured water reading answer different questions.

Start with the decision. If the question is efficiency, go to the PUE or energy page. If it is heat, identify the named customer, contract state and delivery evidence. If it is water or land, open the cooling and local-impact material. If it is reporting, read the applicable authority or EU context and check the reporting period. Do not let a broad sustainability headline choose the evidence branch for you.

The public source matters. Operator pages can be primary sources for what an operator says about its design or ambition. They are not automatically independent measurements, permits or municipal assessments. The register preserves the source’s wording and date, then states the limit. That lets a reader use useful information without treating marketing language as a verified outcome.

• Ambition — target or intention.
• Planned solution — design not yet proven in operation.
• Measured outcome — dated value, boundary and method.
• Source role and independence kept visible.

Read energy, water and heat with their boundaries

A metric cannot travel safely without its system boundary. PUE may describe a design target or an observed period, and it may cover a facility boundary that is not comparable with another operator’s. Water use can mean make-up water, cooling consumption, discharge or a different accounting boundary. Heat reuse can mean technical potential, a prepared connection, a contract or delivered energy. Write the boundary beside the number or keep the comparison qualitative.

The register’s examples illustrate why. Meta’s Luleå material describes outdoor-air cooling and an operator PUE. atNorth SWE01 describes a design PUE, liquid cooling and heat recovery toward Stockholm Exergi. GleSYS has a dated heat-delivery start in Falkenberg and separate 3 MW and later 5 MW capacity wording. EcoDataCenter describes recovery with Falu Energi & Vatten and keeps Falun’s capacity rows separate. Those are useful project observations, not one league table.

A missing period is material. A value from a launch date cannot silently become the current value. A design target cannot silently become a measured outcome. A city objective cannot silently become a hall’s delivered heat. Ask for the period, source version, measurement method, boundary and responsible counterparty before computing a comparison.

1. Metric name and unit.
2. Facility, phase, campus or external-system boundary.
3. Design, reported, contracted or measured basis.
4. Period, method, source and named counterpart.

Put local impact beside the claimed benefit

A sustainability file should include effects that are not captured by a green label. Land use, construction traffic, water demand, heat, noise, backup fuel, visual change and competing uses can matter to the municipality and neighbours. Describe the expected effect, its phase, who published it and which authority or agreement addresses it. Keep construction jobs separate from permanent roles and potential heat separate from delivered heat.

Local impact is not a verdict that DataCenterSweden can issue. A municipal or operator source can document a stated objective, project design or local programme. It does not establish a measured national total. A positive community story and an unresolved environmental question can coexist. The right response is to keep both visible and identify the next document.

The same discipline protects comparisons between places. Stockholm heat-reuse language, Falkenberg’s dated delivery start and Falun’s recovery description belong to different local systems. SE1 or SE4 belongs to electricity-market context, not a sustainability score. Do not add local benefits across municipalities unless the sources define a common period, metric and boundary.

• Land, power, water, noise, fuel, traffic and alternative uses.
• Construction phase, operation and permanent employment separated.
• Local source, counterparty, authority and date.
• Benefits and unresolved impacts shown together.

Treat reporting as dated evidence work

Reporting questions have their own scope and version. The European Commission’s data-center energy material describes reporting work and aggregated information; it should not be read as proof that every individual Swedish facility publishes the same fields. Check the applicable period, entity, facility boundary and reporting method. A report can be a source for a stated value without making every unreported value zero.

When a source changes, preserve the earlier observation and record what changed. A new design can alter cooling, water, energy, heat and local-impact assumptions. A new annual report can update a phase without changing an older event. A later operator page can state capacity with different wording. Do not overwrite the prior row or change every “last updated” field when only one claim was reviewed.

Before publishing a sustainability comparison, ask four questions: are the measures comparable, are the periods comparable, are the boundaries comparable and is the result measured rather than promised? If any answer is no, show the limitation. The honest output may be a table of differently bounded observations and an open request for better data.

1. Applicable source version and reporting period.
2. Entity, facility, phase and metric boundary.
3. Change history instead of silent replacement.
4. Comparability and measured-versus-promised check.