The expansion of AI is increasing data centers’ demand for water, both to cool larger facilities and to manage the heat produced by energy-intensive processors. A hyperscale center can use 1 million to 5 million gallons daily - comparable to a city of 10,000 to 50,000 people. Companies’ estimates of water use per chatbot query often omit the wider demands of powering and cooling these sites, while facilities commonly rely on clean public water because treating other supplies is costly. The result is competition with local residents, often in drought-stricken areas, even as developers promise jobs and investment that may largely amount to temporary construction work.
In Chile, residents of Cerrillos organized to scrutinize Google’s proposed facility, which projected water use of 169 liters per second - more in a day than the town’s 80,000 residents drink in a year. A local referendum opposed the project by 49% to 38%, and Google later proposed water-free air cooling. In Uruguay, activists and researchers sought disclosure of Google’s plan to use 2 million gallons daily amid a historic drought and fears over drinking-water rights. In Brazil, several data centers, including a major TikTok-serving project, are planned in a drought-prone region. These cases show how communities are challenging companies’ claims to be good neighbors and demanding transparency about impacts on essential resources.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.