How GridWatcher grades a data center

GridWatcher assigns every US data-center project a single letter grade — A through F — for how transparent and accountable it is. The grade is built only from verified public records, computed by a deterministic, published methodology, and it is un-buyable: we take zero revenue from the projects we grade.

The principle: opacity is a grade, never a gap to fake

Most "sustainability" or "impact" scores reward whoever discloses the most polished narrative. GridWatcher does the opposite. When a project discloses nothing on a given dimension, we do not guess, average, or fill the hole with an industry assumption. We mark that pillar Insufficient Disclosure and say so plainly. Silence is information — and it is graded as silence, never laundered into a number.

The pillars we grade

Each project is scored across weighted pillars. The weights reflect real-world impact on communities and the grid:

• Grid strain — heaviest weight. Statewide large-load and interconnection context: how much new data-center load a state's grid is absorbing, drawn from RTO/ISO interconnection queues (ERCOT large-load, NYISO) and public utility filings. Labeled a state-context signal, never a fabricated per-project megawatt figure.
• Energy. A proxy built from the project's state air permit, which exists because of on-site backup generation. Permit tier signals the scale of generation and the depth of public monitoring — higher means more disclosure and oversight, not "greener."
• Water. The Clean Water Act record from the EPA's ECHO system — NPDES discharge permits, compliance status, and any effluent exceedances.
• Environment. The Clean Air Act compliance record — permit status, violations, and penalties — plus measured emissions inventories where a state publishes them.
• Economic. The public rezoning and zoning docket — a named, filed land-use case is a sign of an accountable, on-the-record process.
• Governance. The public planning process itself: planning-commission review, magisterial district, and overlay-district disclosure.
• Disclosure & completeness. How much of the picture the project actually put on the public record.

How a grade is computed

For each project, GridWatcher pulls every public record it can find, scores each pillar that has real data on a 0–100 scale, and aggregates only over the pillars that are actually disclosed. A completeness figure records what share of the pillars carried real data.

Then the opacity caps apply — this is the heart of the method:

• Below 25% completeness, a grade is capped at C+.
• Below 50% completeness, a grade is capped at B−.

In plain terms: a project cannot earn an A by staying quiet. Thin disclosure mathematically cannot reach the top of the scale, no matter how clean the one record it did publish looks. The more a project discloses — and the better that disclosed record is — the higher it can climb.

Insufficient Disclosure, never a fake number

If a project has no Clean Water Act record, its Water pillar reads Insufficient Disclosure. If it has never filed a public zoning case, its Governance pillar reads the same. A project with zero disclosure across the board stays ungraded — we never assign a letter to a project we cannot source. Every pillar score links back to the specific public record it came from, so anyone can check our work.

Why the grade is un-buyable

Credit-rating agencies are paid by the companies they rate — the structural conflict at the center of the 2008 financial crisis. GridWatcher takes no money from developers, operators, or the projects it grades. There is no premium tier that lifts a score, no "respond to improve your grade" upsell. The methodology is deterministic and published (current version 2.5.0); the same inputs always produce the same grade. The grade cannot be negotiated because there is no one to negotiate with.

Common questions

Does a low grade mean a project is "bad"?

Not necessarily. A low or capped grade most often means a project has not disclosed enough for the public to judge it. The grade measures transparency and accountability on the public record — not a private verdict on the facility's engineering.

Can a developer pay to change its grade?

No. Every score is computed from public records by published rules. The only way to raise a grade is to disclose more, verifiable, public information — which is exactly the outcome the grade is designed to encourage.

What if the public records are wrong?

We grade what the record says and link to the source. If an authoritative public record is corrected, the grade recomputes from the corrected record on the next run. We never override the public record with a private claim.