Not a systematic survey. Gaps reflect research capacity, not absence of organizing. Counties not shown may have active fights we haven't documented.

Status current as of May 2026. Moratoria expire; legal challenges may reverse wins. Check county government websites for current status.

Proactive zoning varies widely in strength — some ordinances are binding, others advisory. Panel text notes where details are limited.

Legislative status changes rapidly. Bills listed as "active" may have died or advanced since last update.

MW figures are approximate. Operational status may have changed. Not a complete census of US data centers — focuses on counties where community organizing is active or likely.

Connecticut uses electoral districts, not counties — displayed at approximate district boundaries. Alaska uses borough equivalents.

This map covers only counties already in the resistance dataset (~60 counties), not all 3,142 US counties. Full county data is available directly from March of Dimes PeriStats. Classification is based on 2022 data.

Same coverage limitation as maternity care — resistance counties only, not full national coverage.

County-level data does not exist in public form. CDC suppresses county-level maternal mortality data for counties with fewer than 10 deaths (confidentiality) and fewer than 20 deaths (statistical reliability). Most counties have 0–5 maternal deaths per year. State rate applies to all counties shown within that state — a meaningful proxy because state policy (Medicaid, OB licensing, hospital regulation) largely determines outcomes. Data suppressed for 11 states with fewer than 20 deaths total. Racial disparities within states are documented but not mapped — Black maternal mortality nationally is 3–4× white maternal mortality.

State-level medians mask significant within-state variation, particularly urban/rural divides. Does not capture informal care work, which is substantial and disproportionately performed by immigrant women. BLS OEWS excludes self-employed workers. Annual figure assumes 2,080 hours (40hr/week, 52 weeks) — many care workers are part-time or have irregular hours.

Federal policy is actively threatening expansion: work requirements for expansion states are scheduled to begin January 1, 2027. Oklahoma Republicans have proposed reversal mechanisms. This layer may be significantly outdated by late 2026 or 2027. Wisconsin's BadgerCare+ covers some low-income adults but does not use ACA federal matching funds — classified here as not expanded, per KFF convention.

This is a documented sample, not a systematic EJ screening. Counties shown have active documented fights with named organizations and federal legal tools. Many additional counties bearing data center-related environmental burdens are not shown. EPA EJScreen and CDC PLACES offer national screening tools for identifying additional communities.

This map shows 6 of at least 106 documented sites. Honor the Earth's Data Center Tracker documents proposed and operating data centers near or on Native lands nationally. This map shows depth where research is actionable — named tribes, specific water rights instruments, federal legal hooks. It is not a census. See Honor the Earth Data Center Tracker for the broader picture.

This layer requires frequent updating. DHS Secretary Mullin paused new warehouse purchases in April 2026 pending review of contracts signed under Secretary Noem. Some "purchased" sites may face legal delays or cancellation. Some "planned" sites may have been abandoned. Romulus MI and Williamsport MD renovation work is on hold by court order as of May 2026. Merrillville IN warehouse owner stated it is not negotiating with federal officials. Status accurate as of May 2026 — verify current status before using in organizing contexts.

Shown for approximately 70 counties on this map relevant to the resistance dataset — not full national coverage. Full county data for all 3,142 US counties available at census.gov/programs-surveys/saipe. Racial disparities in child poverty are substantial (Black child poverty ~26%, Native American ~28%) but are not broken out at county level here.

This is not a peer-reviewed dataset. It is a synthesis of public data, news reporting, and community organization intelligence. Errors are possible and corrections are welcomed. The selection of what to show reflects a political economy framework: where extractive infrastructure is arriving, care infrastructure is already missing or being stripped away — and the technology built from that extraction is being turned on the same communities. That framing is intentional, not neutral.

Care infrastructure data updates on varying cycles. BLS OEWS releases annually (May). March of Dimes releases annually (September). KFF Medicaid tracker updates as state decisions are made. CDC maternal mortality data lags 2–3 years.

We describe grassroots participants and residents by role rather than by name, to avoid exposing private individuals in a persistent national tool — except where someone has clearly taken a public role (a named plaintiff, or an organizer who has stepped onto a public national platform).