AI data centre resource strain

The rapid expansion of AI‑focused data centres is prompting concerns over critical resource supplies. In late July, reports highlighted India's water‑stressed megacities—Mumbai, Hyderabad, Chennai, Bengaluru and Pune—facing potential shortages as liquid‑cooled GPU racks could consume up to 37.5 billion litres of water annually. At the same time, New Zealand debated the electricity impact of a large AI facility near Invercargill, with the Green Party urging stricter rules to manage grid stress.

A few days later, Italian opinion pieces warned that similarly sized AI data centres could overload power systems, citing a scenario in Lake Tahoe where tens of thousands of households might lose electricity. The articles framed the issue as part of a broader technocratic push by Silicon Valley and comparable practices in China, linking the resource demands of AI infrastructure to threats to human wellbeing and personal freedom.

Together, these snapshots trace a widening narrative: from specific regional water and grid challenges in India and New Zealand to global apprehensions about AI data centres’ capacity to strain electricity supplies and amplify technocratic control.