Methodology

Everything on this site comes from public sources. The headline rule: if we can't link to a primary source, we don't publish the claim.

Facility records

The spine of the database is the EIA's annual generator-level report (EIA-860 / EIA-861) for every U.S. facility > 1 MW, joined to EPA's Facility Registry Service (FRS) for cross-program identifiers. We add facilities that fall below EIA's threshold via news monitoring (GDELT), county zoning filings, and community submissions that pass moderation.

LLC unmasking

Operators frequently file local permits under throwaway LLCs with codename project names ("Project Hercules", "Sterling Park Holdings IV"). We unmask these via OpenCorporates' jurisdiction-aware chain search, joined to SEC EDGAR's full-text filings for public parents, and human-verified when the trail is ambiguous. Aliases are stored separately so you can search by codename and find the real parent.

Tax abatements & subsidies

Subsidies come from Good Jobs First's Subsidy Tracker and the GASB 77 disclosures every public jurisdiction has filed since 2017. When a county discloses an aggregated abatement number, we attribute the dollar value to specific facilities only when there is a direct documentary link (resolution number, application ID, named LLC).

Permits, violations, and water

EPA permits and violations come from EPA ECHO (live API) and ICIS-NPDES for water-discharge specifics. Water-withdrawal context comes from USGS county-level use and groundwater-well APIs. We compute distance from each facility to known EPA-permitted dischargers; "EPA violation within 1 mile" is computed and cached, not editorial judgment.

Public hearings & votes

We scrape county and city government portals (Granicus, CivicPlus, Tyler, Accela, Laserfiche) for agendas, minutes, and video. Video transcripts are produced with Whisper-distil. Council votes are keyed to specific facility decisions where the matter description names the facility or its LLC.

Lobbying disclosures

Federal lobbying comes from Senate LDA filings. State lobbying comes from 35 state ethics-commission databases (Virginia VPAP is the gold standard; we have parity with most state systems for filings since 2018). A facility-to-lobbying link is asserted only when the disclosure's "specific issues" text names the facility, the operator, or a clearly identifiable codename.

Personalized bill impact

Our impact calculator estimates the share of your monthly utility bill attributable to data-center load in your balancing authority (PJM, ERCOT, MISO, etc.) using EIA hourly demand series and FERC capacity market clearing prices. The number is an estimate — explicitly labeled as such — based on published regulatory filings. We show our work on every estimate.

Confidence levels

Every facility, abatement, and link has an explicit high / medium / low confidence value. "High" means a primary public source independently asserts it. "Medium" means strong indirect evidence (e.g. an LLC name matched in a state filing). "Low" means triangulated but not confirmed. Low-confidence records are visible but visually marked.

Corrections

Mistakes get fixed and logged. The corrections log is public. Email us with documentation and we'll respond within 72 hours.