AI Data Center Water Use in California Is Negligible Compared to Total Consumption

A new analysis by Jay Lund, Emeritus Distinguished Professor at UC Davis, estimates that artificial intelligence data centers in California consume between 32,000 and 290,000 acre-feet of water annually, representing just 0.08% to 0.7% of the state's total human water use. The calculations, based on fundamental physics and cross-checked with four AI models, show that even the highest estimate (290,000 acre-feet) is less than 1% of California's 40 million acre-feet annual consumption. Lund criticizes media and advocacy groups for exaggerating AI's water footprint, noting that beer production in Central Arizona uses more water than local data centers. He argues that preliminary quantitative estimates are often omitted from policy discussions, and that AI itself can help produce such estimates transparently. The analysis highlights that data center water use is modest in California but may be larger in other states with less developed water infrastructure. Lund concludes that public discourse on water issues is often driven by fear rather than evidence, and that AI provides tools for more reasoned debate.