The global data centre expansion

A global backlash against AI data centers intensified through 2026, with local communities from Sterling, Virginia to Lane Cove, Australia to Brick Lane, London reporting invasive noise, diesel fumes from backup gas turbines, and pressure on water and electricity supplies from facilities built near residential areas. In Virginia, resident Lindsay Shaw described a constant hum and blazing security lights disrupting her sleep, dismissing President Trump's suggestion that opposition to data centers was driven by Chinese propaganda: 'It's not China telling us anything... It's truly living next to it is what it is.' In Australia, which has 162 data centers with 90 more planned and roughly A$155 billion in investment lined up, some facilities were found to use up to 40 million liters of water daily -- equivalent to 80,000 households -- with Sydney Water projecting data centers could consume up to 25% of Sydney's drinking water by 2035; Prime Minister Anthony Albanese announced new obligations in July requiring large data centers to underwrite power supplies and limit water use, though only for future projects.

In Malaysia, regulators moved to set efficiency requirements rather than an outright ban, while in the US at least 75 data center projects worth roughly $130 billion were delayed or blocked by local opposition in early 2026 alone, and New York became the first state to impose a temporary ban on large new facilities over energy and water concerns, with more than a dozen other states considering similar moratoria. Polling by Gallup found 71% of Americans opposed having a data center in their area -- more unpopular than a nuclear power plant (53% opposed) -- with opposition cutting across both political parties and helping fuel gubernatorial wins for Democrats in New Jersey and Virginia. Even pro-business Texas saw Republican Senate candidate Ken Paxton campaign against data-center tax breaks, while critics noted a functioning data center typically employs only dozens to a few hundred permanent workers despite consuming massive resources, comparing the energy cost of casual AI use like generating images to running household appliances for hours.