Abishek D Praphullalumar
The global AI boom is accelerating demand for massive data centers, but their hidden environmental cost is now under scrutiny. A recent Bloomberg investigation reveals that these sprawling server farms, essential for powering AI models, consume staggering amounts of water often in regions already grappling with severe drought.
Data centers rely on water-intensive cooling systems to keep their heat-generating servers operational. In drought-prone U.S. states like Arizona, Texas, and Nevada, over two-thirds of new data centers built since 2022 are in high water-stress areas. A single 100-megawatt facility can consume up to 2 million liters of water daily enough for 6,500 households.
The trend extends globally, with Saudi Arabia, the UAE, China, and India rapidly expanding AI data infrastructure in already water-scarce regions. Local communities and environmental advocates have voiced growing concerns. In Chile and Uruguay, public protests and legal challenges have erupted against new data center projects over fears of worsening water shortages.
While some tech giants like Microsoft are piloting closed-loop cooling systems to minimize water waste, most data centers still depend on evaporative cooling, which loses up to 80% of water to the atmosphere.
Adding to the problem is a lack of transparency, as companies frequently withhold water consumption data. In Oregon, Google fought a 13-month legal battle to block the release of its water usage records.
Experts warn that without urgent reforms in cooling technology, data center siting policies, and public reporting requirements, the AI-driven digital expansion risks exacerbating water scarcity in already vulnerable regions.
