Multiple studies and environmental assessments confirm that artificial intelligence infrastructure, including data centers and semiconductor manufacturing, consumes substantial amounts of water primarily for cooling and electricity generation.
The retrieved literature consistently supports the claim that artificial intelligence infrastructure has a substantial water footprint, driven by the cooling needs of data centers, power generation, and hardware manufacturing. Papers [0], [2], [3], [4], [5], [6], [8], and [10] explicitly detail these water-intensive operations and environmental impacts, while papers [1], [7], [9], and [11] are off-topic or tangential.