Voters Block AI Data Centers as Water Usage Hits 10 Million Gallons Daily

Voters Push Back Against AI Data Centers Over Resource Concerns

Voter concerns over AI data centers reached a tipping point this week, with 40 of 41 proposed ballot measures passing to restrict more than 2,000 new facilities planned across the nation1

1

. Water usage and power consumption topped the list of ratepayer grievances, yet measuring water consumption at individual sites remains elusive. Operators typically report usage figures companywide instead of site by site, leaving communities struggling to understand the true impact on local resources.

Giant 1-gigawatt AI data centers from Meta, Google, Amazon, and SpaceX can consume up to 10 million gallons of water per day, according to Shaolei Ren, a data center researcher and professor at University of California, Riverside1

1

. The exact number varies dramatically depending on hardware type, location, and advanced cooling methods employed. This complexity has created room for misinformation from both builders and opponents, making transparent reporting critical.

How Liquid Cooling Technology Transforms Water Consumption

AI data centers generate enormous heat that demands sophisticated cooling infrastructure. Legacy data centers rely on air cooling with large fans, but modern facilities packed with high-power GPUs require different approaches. These GPUs run so hot they can "boil your eggs," Ren noted1

1

. Water proves about 3,000 times more efficient for cooling than air, according to Equinix sustainability head Christopher Wellise1

1

.

Nvidia's latest rack-scale systems, Vera Rubin and AMD's Helios, rely on liquid cooling to manage heat generation1

1

. Most large AI data centers use closed-loop systems that continuously circulate coolant through hardware, using minimal water on a daily basis. At an Equinix data center in San Jose, California, closed-loop liquid cooling keeps eight Nvidia Grace Blackwell racks running for AI marketing startup Alembic1

1

. Despite this efficiency, Equinix operations consumed 1.2 billion gallons of water in 2025, up 10% from the previous year1

1

.

The Hidden Water Cost: External Cooling Systems

While closed-loop systems minimize internal water use, the AI data center water crisis persists because of external cooling requirements. "How do we reject the heat from the closed loop to the ambient environment? That's where the water usage comes in," Ren explained1

1

. This external process removes heat from the separate internal system, particularly on the hottest days.

Meta's 1-gigawatt data center in Lebanon, Indiana, illustrates this challenge. The company claims its "closed-loop, liquid-cooled system that recirculates the same water and will use zero water for a majority of the year"1

1

. However, water allocation for the site could reach 8 million gallons per day by 2031, according to Meta's law firm1

1

. Meta projects this peak usage only on the hottest days, anticipating "only a few days per year" at maximum levels1

1

.

Power Usage Versus Water: The Cooling Trade-Off

Two primary methods handle external cooling: dry cooling and evaporative cooling. Dry cooling uses far less water but demands significantly more power usage, especially in hotter climates where many facilities are being built. "There's no sort of silver bullet," Wellise said. "At some point you need to eject that heat out to the atmosphere. And that's where those trade-offs occur"1

1

.

Dry cooling relies on fans blowing air across heat exchangers, but effectiveness drops when outside temperatures climb. Data centers then activate refrigeration units to cool air first, creating a power-intensive process1

1

. Evaporative cooling reduces power demands but increases water consumption, forcing operators to balance infrastructure requirements against local resource availability.

Nvidia's 45°C Breakthrough and Future Solutions

Nvidia's next-generation architecture introduces specialized liquid cooling that could reshape sustainable AI data center operations. The company's warmer coolant technology enables systems to operate at 45°C, potentially reducing reliance on energy-intensive chillers and water-consuming cooling towers2

2

. This 100% liquid cooling approach could bring facility-level cooling water consumption close to zero under suitable conditions2

2

.

Yet liquid cooling cannot eliminate water consumption across the entire AI infrastructure supply chain, including electricity generation and semiconductor manufacturing2

2

. High installation costs, infrastructure requirements, climate conditions, and maintenance challenges influence adoption speed2

2

. Watch for how quickly operators implement these advanced cooling methods and whether transparency in measuring water consumption improves as voter pressure mounts.