Editorial Standards
DataCenterFeed covers the data center and AI infrastructure buildout accurately and plainly, for people who have to act on the news — and for the communities living with it. This page sets out how we report, source, and check what we publish, and how we use AI.
What we report
We report on-the-record facts: votes and vote tallies, filings, permits, docket orders, statements made at public meetings, and company announcements. We describe what people said and did, not why we think they did it — we do not characterize anyone's motives. We name public officials, companies, and named spokespeople; we do not name private residents, including those who speak at public meetings.
Attribution of project figures
Data center projects are described in megawatts, acres, square feet, dollars, gallons, and jobs — and those numbers often change between an announcement and a final approval. Every such figure in our coverage is attributed to the source that stated it ("according to the rezoning application", "the company said in a news release"). Where sources disagree, we say so and cite each.
Sourcing
Every article cites at least three sources, listed openly in a "Sources" section at the foot of the piece. For project, zoning, power, deals, and policy coverage, at least one is a primary record: a county or city agenda, minutes, or permit file; a state utility commission docket; a FERC or grid-operator filing; an SEC filing; or the company's own announcement. Trade press alone is never enough, and we do not rewrite a single outlet's story.
Not investment advice
We report capital spending, transactions, and financing as disclosed. Nothing on DataCenterFeed is investment, legal, or financial advice, and we do not make price predictions or recommendations.
Fact-checking
Before publication, every article passes an automated review that confirms its sourcing requirements are met and verifies that each cited link resolves to a live page — so readers never hit dead citations. Factual claims are grounded in the cited material; where a figure or date is uncertain, we say so rather than guess. When new findings supersede old reporting, we update.
How we use AI
We are transparent about this: DataCenterFeed articles are drafted with AI assistance, working from sources gathered and read for each story. Bylines on this site are AI-assisted desks, and every byline says so. Every article is checked against its cited sources and reviewed by the publisher before release — nothing goes live unreviewed. We do not mass-publish unsourced content.
Our desks
Articles carry a desk byline, each with a defined beat and a published description:
Corrections
Project plans change quickly — sizes are revised, hearings are postponed, applications are withdrawn. We aim to be clear about what is approved versus proposed, and we correct mistakes promptly when they are brought to our attention. To report an error, email [email protected] with the article and the correction.
Independence
Editorial decisions are made independently of any advertising on this site. Advertising relationships, where they exist, never influence what we cover or how we cover it.
Independence from the companies we cover
DataCenterFeed is an independent publication and is not affiliated with any government body, utility, grid operator, developer, or hyperscaler. Coverage is for general information; for official project status, consult the relevant public record. For questions, corrections, or partnerships, reach us at [email protected].
