Majority of US’s new AI datacenters to be built on drought-hit land

Editor's take

New AI data centers are disproportionately slated for construction in regions experiencing severe drought, according to an analysis of planned facilities. This trend raises significant concerns about the immense water demands of AI infrastructure, particularly as these facilities require vast quantities of water for cooling.

The implications are far-reaching, impacting communities already struggling with water scarcity and exacerbating existing environmental pressures. This development highlights a critical tension between the rapid expansion of AI compute power, epitomized by models like OpenAI's GPT-4 and Google's Gemini, and the sustainability of the resources they consume. The burgeoning AI industry’s thirst for water is now a tangible environmental challenge.

Future developments will likely involve increased scrutiny of data center siting and water usage disclosures. The debate will shift towards regulatory interventions, the adoption of more water-efficient cooling technologies, and potential mandates for AI companies to offset their water footprints, especially in arid states like Arizona and Nevada where much of this expansion is occurring.