Draft v0.1 is open for comment until January 31, 2027.Tell us what we missed.
A rating system for data centers · Draft v0.1

Place.People.Purpose.

This draft starts from the community. A data center earns its place when it is clean, when local people share in what it creates, and when the work it does helps them. We score all three, the way LEED scores buildings, so neighbors, councils and engineers can read the same scorecard.

Rebecca Butler and Christina Liebner · November 2026 · For discussion

Three overlapping circles: Place, People and Purpose. Their overlaps are fair footprint, clean useful compute and shared prosperity. Common ground sits in the center. Place People Purpose Place People Purpose Fair footprint Clean, useful compute Shared prosperity Common ground
Select a circle to see its credits.
Community first

The useful question is how a community decides which data centers it welcomes, and which it turns away.

Most stories about data centers start from fear: bills that rise, wells that run low, jobs that go elsewhere, decisions made far away. Those fears are reasonable. A large campus can draw as much power as a small city, use water where it is scarce, and send most of its value somewhere else.

The fears also reach past any one building. People worry that AI will take work without giving any back, that their data will leave town, and that the costs of a changing energy system will land on the households with the least say. A data center is where those worries become concrete. It has an address, a zoning hearing and a fence, so it is where opposition gathers.

This started when we asked ourselves what a data center would look like if it met our communities' needs and values, and what standards we would hold it to. We didn't find anything that covered all our concerns, so we started our own rubric.

It begins with the community. What would residents need to see to welcome a site? Clean power and water they can trust, a fair share of what it earns, a real voice in what happens next, and work that helps the people who live there. Each of those can be measured, and data centers differ widely on all of them.

People need a common way to tell these apart. LEED gave buildings a shared language for "green." Place, People, Purpose aims to do the same for data centers: a public scorecard that a council member, a utility engineer and a neighbor at the farmers market can all read.

Our bias is toward small, clean, community-owned sites. The rating does not require them, but it is built so they can reach the top, and so that large projects earn their way up with verified commitments.

How the rating works

100 points, four tiers, and a balance rule

9
Prerequisites

Disclosure and basic protections. Miss one and the project is not eligible, whatever its points.

30·30·30
Three pillars, equal weight

Place, People and Purpose each carry 30 points. Ten more sit in the center, for work that ties all three together.

≥12
Balance rule

Higher tiers need a minimum score in every pillar. A project cannot reach Included on clean power alone.

Not rated0–39
Informed40+
Consulted50+
Included60+
Empowered80+
040506080
TierTotal pointsMinimum in each pillarWhat it signals
Informed40–49noneThe community knows what is being built and how it performs. Meets every prerequisite and does real work in at least one area.
Consulted50–596 of 30The community's input shapes the project. Solid in one or two pillars, present in all three.
Included60–7912 of 30The community shares in the benefits and the decisions. Strong overall, with no pillar left behind.
Empowered80–10018 of 30The community has real power over the site and what it does. A leader in all three circles. No credit is mandatory, so leaders get here without perfect scores.
Where the circles meet

Each overlap is a promise a community can check

Fair footprint

Place and People. Clean, quiet, flexible on the grid, and paying its own way, so neighbors are not stuck with the costs.

Clean, useful compute

Place and Purpose. Energy and water go into work that matters, measured as carbon and water per unit of useful work.

Shared prosperity

People and Purpose. Local ownership, local services and local data, so the value and the decisions stay close to home.

Common ground

All three. Ten center points reward integrated design, an open playbook, and an independent survey of what residents actually think.

The credits

What each pillar asks for

Plain-language intent first, then the measurable requirement. Thresholds are a conservative first draft for testing, not final values.

Three illustrative projects

How the balance rule sorts projects

Hypothetical scores to test the system, not assessments of real projects. Credit-by-credit scores are in the companion spreadsheet.

Try it

Move the pillars and watch the tier

24
19
10
4
Result
57 / 100
Silver

Design principles

Conservative by design

  1. Measured, not modeled.Final ratings use 12 months of operating data. Ratings at the design stage are provisional.
  2. Hourly where it matters.Clean energy and flexibility are counted hour by hour, because annual totals hide the nights and still days.
  3. Balance is required.Minimum scores in every pillar keep a project from buying its way up with one strength.
  4. Disclosure before points.Prerequisites cover transparency and basic protections. Fail one and nothing else counts.
  5. Built on existing standards.ISO/IEC 30134 for PUE, WUE and heat reuse, the GHG Protocol for emissions, and established security and recycling certifications.
  6. Fair to small and large sites.Credits use shares and ratios, so a library-sized site and a large campus are judged on the same terms.
  7. Ratings expire.Sites recertify every three years with fresh data, and points can be lost.
  8. Climate-fair efficiency.Efficiency is judged against a climate-zone baseline, and carbon and water per unit of useful work lead, so sites can be built where people live.
What gets rated

The site, as it operates

The rating covers one site: the facility, its owner and operator, and the work it hosts. A green building with poor data practices scores well in Place and poorly in Purpose, and the balance rule keeps it out of the top tiers.

A separate label for AI tools could come later. Like a seafood buying guide with green, yellow and red picks, it would help cities choose applications. That is a companion project, not part of this rating.