Water capacity is now a siting gatekeeper for AI data centers. A Virginia campus requested 2M gallons per day; Georgia told a 6 MGD project 'we just don't have the water.'
A Virginia AI data center campus requested 2 million gallons per day of water for its initial deployment. The utility filing acknowledged demand 'exceeded existing water and wastewater planning assumptions.'
Key facts
- Virginia campus: 2 MGD initial, up to 8 MGD future demand
- Georgia water authority: 'We just don't have the water' for 6 MGD
- Texas projects $174B in water infrastructure over 50 years
- Texas water supply may decline 10% by 2080
- UC Riverside: Nearly all server energy converts to heat
For two years, the AI infrastructure race focused on electricity — grid interconnection queues, substation lock-ups, gas capacity. Now water is the new siting gatekeeper.
Key Takeaways
- Water capacity is now a siting gatekeeper for AI data centers.
- A Virginia campus requested 2M gallons per day; Georgia told a 6 MGD project 'we just don't have the water.'.
The 2 MGD Problem
A proposed Virginia campus requested 2 million gallons per day (MGD) for initial deployment, with future demand reaching 8 MGD, according to a utility-services agreement [The Breaking Points]. The filing required 'continuous evaporative cooling to protect sensitive equipment required for essential operations' — a specification that locks in water-intensive cooling regardless of local supply.
In Newton County, Georgia, a water authority representative told a proposed 6 MGD project: 'We just don't have the water.' The remark captures a broader reality: securing electricity does not guarantee sufficient cooling water, wastewater capacity, or municipal support.
Texas: $174B for 50 Years
The draft 2027 Texas State Water Plan projects existing supplies could decline 10% by 2080 while population rises 50%. The state estimates $174 billion in water infrastructure projects may be required over 50 years to meet growing AI demand [The Breaking Points]. Notably, the plan does not model AI-related data center demand as its own planning category — a blind spot that will grow as hyperscale campuses proliferate.
Cooling Physics vs. Municipal Reality
'Nearly all the server energy is converted into heat, which must then be removed from the data center server room to avoid overheating,' UC Riverside researchers wrote. That heat rejection now strains municipal systems built for slower, steadier growth.
Google, which committed to 24/7 carbon-free energy by 2030 and is funding a $5B+ Texas data center for Anthropic, faces this tension directly. The company's TPU venture with Blackstone, announced May 21, 2026, targets AI infrastructure financing — but water constraints may cap where that capital can be deployed.
The unique take: Water, not power, will determine the next wave of AI campus locations. Grid capacity can be expanded with transmission lines and gas peakers; water requires watersheds, treatment plants, and drought planning at a geological timescale. The Newton County rejection is a leading indicator of a structural bottleneck that will reshape AI infrastructure geography.
What to watch
Watch for the Texas Water Development Board's 2027 State Water Plan update to explicitly model AI data center demand. Also track municipal approval timelines for Google's $5B+ Texas campus — if water permits lag, it signals a broader bottleneck.
Originally published on gentic.news
Top comments (0)