Oracle is pushing ahead with an AI data center project in rural New Mexico called Project Jupiter, but residents in the desert region are raising alarms about what it means for their already strained water supply.
According to Tom's Hardware, the project's initial fill alone requires 11 million gallons of water — a significant figure for a desert community where water scarcity is an ongoing reality. Oracle, however, is pushing back on the framing, characterizing the water demand as "negligible."
The tension highlights a collision between two urgent priorities: the insatiable thirst of the AI industry for computing power, and the finite natural resources of the communities where that infrastructure lands. Data centers use water in large quantities to cool their servers, and as AI workloads grow more intensive, so does the cooling demand.
New Mexico is no stranger to water stress. The state sits in an arid climate where aquifers and river systems face pressure from agriculture, population growth, and climate change. Placing a large-scale AI facility in this environment puts the question of corporate water stewardship front and center.
Oracle's "negligible" characterization refers specifically to a one-time fill figure, according to Tom's Hardware — though ongoing operational water use in data centers is typically a separate and recurring concern not addressed by that framing.
The story matters because it's a preview of fights communities across the American West and other water-scarce regions will increasingly face as tech giants race to build AI infrastructure wherever land and power are available — often without enough scrutiny of what else those places can afford to give up.