Next Meeting: June 16, 2026
✉️ Subscribe ☕ Support
Source:Aterio + Pew / DCM + USGS
Last update:Wed May 20 2026
Civic water dashboard

Thirst Trap Map

AI looks hot. The water bill is hotter.

Mapping the water risk behind the U.S. data center boom — from AI infrastructure to the rivers, aquifers, watersheds, and communities carrying the cost.

Don’t fall for the thirst trap. AI may look sleek, seamless, and inevitable. But behind the glow are real rivers, aquifers, watersheds, and communities absorbing the cost.

The surface: AI looks hot, magical, frictionless, futuristic.

The trap: It hides extraction, water demand, cooling infrastructure, energy strain, land use, and local risk.

Active data centers 1,974 Currently operating
Announced 3,424 Proposed / planned growth
Under construction 791 Coming online soon
Extreme-stress projects — Based on projects with available metadata
Project stage:
Active
UC
Announced
Water context:
Water-impact
River

Submit a tip for review

Submissions do not appear automatically. Tips, updates, and organizing notes are reviewed before they are added to the public map.

Submit a data center for review Submit an organizing note Submit an update
Filters are first-pass civic review helpers.

Selected project

DFW3 Dallas Data Center

CyrusOne · Active

Dallas, TX

Water context

Water source —
Nearby river —
Basin / watershed —
Stress level —
Flow percentile Needs gauge match
Water-impact metadata —
Est. water use —

Community, organizing & verification

Community & organizing

Organizing group Not yet added
Local notes Not yet added
Summit priority TBD

Sources & verification

Source links —
Metadata completeness
0%
Submit an organizing noteSubmit an update
Teardrop pins mark projects with water-source data. Click any pin to inspect verified details.

Video Playlist + Field Context

A lightweight media layer for explainers, public meetings, campaign clips, or community documentation connected to data centers and water impact.

Media resource
Companion video context

Thirst Trap Map video context

Open the playlist on YouTube to watch explainers, public meeting clips, campaign material, and field context related to data centers and water impact.

Highest data center concentration by state

StateActivePlannedTotal
Virginia398287685
Texas296170466
California277150427
Ohio166250416
Illinois120123243
Georgia100141241
New York14880228
Arizona8086166

Active = currently operating. Planned = under construction + announced + land banked. Source: Pew Research / Data Center Map, Feb 2026.

Needs Research workflow queue

  • Water source missing—
  • Stress level missing—
  • River / basin missing—
  • Organizer unknown—
  • Source link needed—
  • USGS gauge ID needed—

Counts are directional and based on curated + locally added projects with available metadata.

River Stress Signals

Hybrid signals combine WRI-style baseline stress with USGS 90-day streamflow where a gauge is available. Click any river to highlight matching pins on the map above.

Source: USGS Water Services · 90-day daily mean discharge · Stress badges from local WRI Aqueduct-style baseline lookup

Water Contamination + Public Trust

A companion context section for water contamination, water access, infrastructure failure, criminalization, and environmental accountability stories. This is a resource layer, not project-level verified data.

Offshoot resource

Jennifer Combs / Trinidad, Texas

Resident arrested after posting about local water concerns; useful public-trust context for Trinity River-area water anxiety.

Open source

PFAS contamination map

National reference layer for forever-chemical contamination and affected water systems.

Open EWG map

Military bases + PFAS

State-by-state resource for military-base water contamination concerns.

Open list

Historic water crisis references

Sand Branch, Flint, Love Canal, Cuyahoga River fires, Elk River spill, and Jackson, Mississippi remain reference points for civic review.

Context only

Public layer —Thirst Trap Map is part art project, part civic data experiment. Public submissions help identify missing data, local context, and organizing efforts, but all submitted information is reviewed before appearing on the map.

V1 note —This is a V1 public layer. It shows curated and verified information while inviting public tips through a review process. Future versions may include a private admin layer, moderated submissions, and a live database.

Trust note — Data is compiled from public sources (Aterio, Pew Research / Data Center Map, USGS Water Services, US Census, AP) and may be incomplete or inconsistent across trackers. Individual pin coordinates outside the curated set are statistical approximations within known metro markets, not source-verified addresses. Water-source relationships should be confirmed with local water authorities, permits, utility districts, environmental reports, or USGS gauge data before use in advocacy or reporting. Water stress signals are directional and should not be treated as final hydrological analysis. Spanish translation should be reviewed by a native speaker before public launch.

THE CLOUD IS THIRSTY · EVERY DATA CENTER HAS A WATERSHED · RIVERS ARE NOT CHECKBOOKS · PUBLIC DATA, PUBLIC WATER · DON’T FALL FOR THE THIRST TRAP · AI LOOKS HOT. THE WATER BILL IS HOTTER. · WATER RISK IS COMMUNITY RISK · FIELD NOTES FROM THE CLOUD’S WATERSHED · MAP THE THIRST · PROTECT THE FLOW ·