Where the data comes from

Every project on this site is built from public records and credible reporting. Our primary sources include:

  • County and municipal governments: comprehensive plan amendments, rezoning applications, site plans, commission meeting agendas, and vote records
  • Water management districts: consumptive use permit applications and board meeting records
  • State legislatures: bill text, committee analyses, and voting records
  • Public utility commissions: utility tariff filings, rate cases, and interconnection agreements
  • Federal agencies: EIA power plant capacity data, FERC transmission filings, Census population estimates
  • Local and national news: coverage from NPR affiliates, regional papers, state political outlets, and national outlets including The Washington Post, Reuters, and Bloomberg
  • Trade publications: Data Center Dynamics, Data Center Knowledge, and project tracking databases
  • Community reports: tips submitted by residents through our contact form, verified against public records before publishing

Every project entry on the site links to its sources. If a source goes offline, we note it.

How the data-center risk score works

The address check adapts the four-factor structure of our Florida risk calculator to a worldwide, comparative regional index. It adds power infrastructure (up to 30 points), water conditions (15), land cover (15), and current project exposure (40). The sum is the displayed score out of 100. It is an editorial model of conditions that may attract development, not a prediction about the searched parcel.

  • Power, 6–30 points: the World Bank global grid map shows observed and modeled power lines from 2020. We count roughly 5.6 km map cells containing observed grid lines within 50 km. Counts of at least 220, 150, 100, 60, and 25 score 30, 26, 22, 18, and 12 points. Lower observed counts fall back to 6 points, or 10–16 when the modeled map has broader coverage. This measures mapped regional network density, not spare grid capacity, voltage, or an available connection.
  • Water, 6–15 points: WRI Aqueduct 4.0 baseline annual basin stress runs from 0 to 5. Points equal 15 minus 9 times the stress score divided by 5, rounded and bounded at 6–15. A missing or unrated basin receives 8 neutral points and is identified in the full score. This does not measure water rights, permits, or site supply.
  • Land, 2–15 points: NASA MODIS 2021 land cover is sampled across roughly 15 km around the address. Barren land receives 15 points; shrubland, savanna, and grassland 13; cropland or mixed cropland 10; forest 7; built-up land 4; wetlands, snow, and mapped water 2. Missing coverage receives 7 neutral points. These broad classes are a regional proxy, not a survey of vacant parcels or zoning.
  • Known activity, 4–40 points: tracked campuses within 30 miles add 7 points each to a 12-point starting value, up to four campuses. Nearby colocation sites add 2 points each, up to five. A proposed, approved, or under-construction site adds 4 points, capped at 40 overall. If there are only colocation sites within 30 miles, the score starts at 8 plus 4 points per site, capped at 28. With no site within 30 miles, a campus within 100 km receives 18 points; another tracked site within 150 km receives 10; otherwise the baseline is 4. The address score reads the same project pins as the homepage globe. Unknown status still counts; withdrawn, cancelled, suspended, and blocked projects do not add activity points. Tracker pins may be approximate and can represent a campus, city, or hosted service rather than a separate physical facility.

The address result shows the mapped or tracked finding behind each factor: nearby tracker pins, power-line map distance, regional water stress, and surrounding land cover. When a map fails to load, the result says so and uses the stated baseline. Addresses are located with Geoapify Geocoding. When the exact address cannot be confirmed, the visitor chooses a suggested location before seeing its area score. A tracker pin may be approximate or represent a campus or cloud region rather than a separate physical facility. Electricity service, water rights, available acreage, permits, and future proposals all require site-specific verification.

What we cover

We track hyperscale and AI data center proposals that meet any of these thresholds:

  • 50 megawatts or larger
  • 200 acres or larger with stated data center intent
  • Any project drawing significant local opposition, county commission hearings, or news coverage

Smaller traditional data centers and colocation facilities (under 25 MW) generally fall outside our scope since they do not trigger the same infrastructure or community impact concerns.

How we handle corrections

If you find an error in a project status, risk score, commissioner name, or anything else, report it through the contact form or email us. We verify against primary sources before making a correction. When we correct something significant, we add a dated note on the affected page so the change is transparent. See our corrections policy for details.

Limitations

Data center proposals move fast. A project status on this site reflects the last public record we could verify, not necessarily what is true at the moment you are reading it. Some proposals are filed confidentially under state economic development exemptions and may not appear in public records until the developer applies for land use changes. We update as developments surface, but there will always be a lag between what is happening and what we can report.