Water Risk Directory
Water availability is the second-largest risk factor for AI datacenter site selection after power. GLRI.io's Water Stress Score (WSS) tracks drought risk, municipal supply constraints, and permit friction across 40+ metros. A high WSS can add 6-18 months to a datacenter deployment timeline.
Water Stress Overview
4
Metros with minimal water risk
Chicago, Ashburn, Seattle
6
Metros requiring planning
Dallas, Houston, Austin
4
Metros with significant risk
1
Critical constraints
WSS Data by Metro
| Metro | WSS Score ↓ | Stress Level | Permit Friction | Notes |
|---|---|---|---|---|
| Las Vegas | 85 | Extreme | 8/10 | Critical water constraints |
| Phoenix | 78 | High | 7/10 | Significant water scarcity concerns |
| San Francisco | 75 | High | 7/10 | Drought vulnerability |
| Los Angeles | 72 | High | 6/10 | Municipal supply constraints |
| Denver | 65 | High | 6/10 | Mountain snowpack dependent |
| Austin | 58 | Medium | 5/10 | Growing demand pressures |
| Miami | 55 | Medium | 5/10 | Climate adaptation needed |
| Houston | 52 | Medium | 5/10 | Gulf coast water access |
| New York | 48 | Medium | 5/10 | Complex permitting |
| Dallas | 45 | Medium | 4/10 | Adequate water supply |
| Atlanta | 42 | Medium | 4/10 | Southeastern water resources |
| Ashburn | 35 | Low | 3/10 | Adequate regional supply |
| Chicago | 28 | Low | 3/10 | Great Lakes water access |
| Seattle | 25 | Low | 3/10 | Pacific NW water abundant |
| Minneapolis | 22 | Low | 2/10 | Abundant freshwater |
WSS methodology: composite of USGS watershed stress, EPA permit data, and municipal water authority reports
Water Usage Calculators
Calculate your facility's projected water consumption based on cooling type, rack density, and PUE target.
Water Risk in AI Datacenter Siting: What $200M Projects Get Wrong
Water risk is the most underestimated factor in AI datacenter site selection. While operators obsess over power availability and interconnection timelines, water constraints can silently derail projects that have cleared every other hurdle.
The math is stark: a 100MW datacenter using evaporative cooling can consume 25-50 million gallons of water per day. In regions facing drought conditions or municipal supply constraints, this volume of water draw can trigger public backlash, regulatory scrutiny, and permit delays that add 6-18 months to project timelines.
Phoenix and Las Vegas exemplify the water risk challenge. Both markets offer abundant power and excellent fiber connectivity, but WSS scores of 78 and 85 respectively indicate critical water constraints. Projects in these metros should plan for closed-loop cooling systems, water recycling infrastructure, or alternative cooling technologies like direct liquid cooling (DLC) that reduce water consumption by 80-95%.
Chicago and Minneapolis represent the opposite end of the spectrum — both offer abundant Great Lakes water access with WSS scores under 30. These markets should be prioritized for water-intensive cooling deployments, though cold-weather considerations and winter operational complexities must be factored into the analysis.