Editorial review · 260614-011
How ORA’s piece on Who pays when the data centre arrives scored.
Read the article →Solid reporting. Some issues but credible overall. The reader is well-served.
Accuracy
Headline figures (130B, 75 projects, 833 groups, 49 states, 156B for 2025) are attributed to Data Center Watch via Ars Technica and are post-cutoff but properly sourced. The Loudoun County figures (53M square feet, near-half of budget) are plausible but the budget share is asserted without a primary citation (-5). The New York moratorium claim is dated specifically (June 4, awaiting Hochul's signature) and attributed, so no deduction.
Balance
The piece has a clear point of view but engages the strongest pro-buildout case (Loudoun) seriously rather than strawmanning it, and concedes the mobility argument has merit. Industry voices and hyperscaler rationale are summarised but not quoted directly, a minor source-diversity gap on a contested policy topic (-8). Language stays measured; 'civic veto' and 'being notified' are pointed but not pejorative beyond editorial justification (-5).
Concerns (4)
- minoraccuracy
“data-centre property taxes fund close to half its budget”
Specific fiscal claim about Loudoun cited only via Ars Technica aggregation.
Evidence: A primary Loudoun County budget source would be appropriate for a load-bearing figure.
- minoraccuracy
“75 US data-centre projects, worth roughly 130 billion dollars”
Post-cutoff, source attributed to Data Center Watch via Ars Technica.
Evidence: Recorded for transparency; no deduction under the post-cutoff rule.
- minorbalance
“(source set)”
No direct industry, hyperscaler, or utility voice on a contested policy topic.
Evidence: Sources are Ars Technica, Food & Water Watch, and NY Post; opposing camp paraphrased, not quoted.
- minorbalance
“It is being notified.”
Pointed framing in favour of the opposition camp without symmetric rhetorical treatment.
Evidence: The pro-buildout case is engaged substantively but never gets a comparably sharp line.
Reproducibility
How this review works: read the methodology. Each published Dispatch is scored by a single primary reviewer (Claude Opus 4.7) against the public rubric. A second model (Gemini 2.5 Pro with Google Search) runs the same prompt as a variance signal and is shown above only when the two scores diverge by more than ten points.