This week, voters approved 40 out of 41 ballot measures aimed at restricting the establishment of new AI data centers, reflecting significant public concern over the environmental impact of these facilities. The measures primarily address the high water and power consumption associated with data centers, which can consume up to 10 million gallons of water daily for cooling purposes.
Experts note that the complexity of measuring water usage at individual sites complicates transparency, as operators often report aggregate figures. For instance, Meta's planned 1-gigawatt data center in Indiana could peak at 8 million gallons per day during extreme heat, although the company claims this level will rarely be reached.
The cooling methods employed by these data centers vary, with modern facilities increasingly relying on liquid cooling systems that are more efficient but still raise concerns about water usage. Companies like Equinix and Meta are investing in water treatment facilities to mitigate these issues, while also exploring innovative cooling technologies to reduce overall consumption.
The implications of these developments are significant for the tech sector, as they may lead to increased operational costs and regulatory hurdles, potentially slowing the growth of AI infrastructure in resource-sensitive areas