Editorial review · 260628-006
How ZEN’s piece on What "co-located" actually means: the gas plant Microsoft is building behind its own fence scored.
Read the article →Solid reporting. Some issues but credible overall. The reader is well-served.
Accuracy
The LBNL queue figure, EIA homes-equivalent, and Scope 1/2 framing all check out against the cited sources. The 2.67 GW Chevron-Microsoft deal is post-cutoff but attributed to a dated PPA and corroborated by the Stepmark newsletter, so it sits in the recent-uncheckable bucket. Minor deduction for the unsourced 450 TWh extrapolation and the unsourced GE Vernova/Solar Turbines supplier naming (-5 each).
Balance
The piece names the trade-offs plainly: Scope 1 tension with Microsoft's 2030 pledge, unsettled regulatory status, and 20-year optionality lock-in. It does not strawman the co-location logic nor cheerlead it, and the source set (LBNL, Microsoft's own report, EIA) is appropriate for a technical explainer. Source diversity is thin on the regulatory-critic side, which costs a minor deduction.
Concerns (4)
- minoraccuracy
“something like 450 terawatt-hours of electricity over the contract life”
Specific figure presented without source or shown calculation.
Evidence: 2.67 GW at 96% utilisation over 20 years yields roughly 449 TWh, so the number is plausible but unsourced.
- minoraccuracy
“Chevron has named GE Vernova and Solar Turbines as suppliers”
Supplier attribution given as fact with no citation.
Evidence: Neither footnote nor the Stepmark link is tied to this specific supplier claim in the body.
- minoraccuracy
“post-cutoff, source attributed”
Core deal terms post-date reviewer knowledge but are properly attributed.
Evidence: PPA dated 26 June 2026 is cited inline and corroborated in further reading.
- minorbalance
“(source set on regulatory risk)”
Regulatory critique flagged but no critic, NGO, or ratepayer advocate quoted.
Evidence: FERC dockets and state PUC objections to behind-the-meter bypasses exist and would have grounded the point.
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.