Data center guzzled 30 million gallons of water and nobody noticed for months

A major data center in Georgia inadvertently consumed nearly 30 million gallons of water without being billed for several months. This oversight occurred while nearby drought-stricken residents were urged to conserve water, leading to frustration and questions about local utility monitoring. An investigation revealed the facility had unmonitored water hookups, one installed without the utility's knowledge and another not linked to the company's account.

QTS, the data center operator, eventually paid approximately $150,000 for the water. However, there were no penalties for exceeding consumption limits, a decision that angered residents. The county cited self-blame and a desire to maintain a good relationship with their largest customer as reasons for not imposing fines. The issue was exacerbated by the county's transition to a new smart water meter system and understaffing in their monitoring department.

QTS maintains that all water usage followed regulations and denies accusations of improper consumption, stating they paid all charges once the billing issue was flagged. While residents complained of decreased water pressure, county officials stated that QTS does not draw from wells, which were the source of the residents' issues. The county has since confirmed improved monitoring for QTS's water hookups, and QTS anticipates a drastic reduction in water needs post-construction.

This incident in Georgia sheds light on the growing water demands of the AI industry. The increasing reliance on data centers and semiconductor factories, many located in water-stressed regions, is projected to more than double AI-associated water usage in the next 25 years. Cooling power plants for hyperscale data centers further contributes to this escalating demand.

To address this challenge, some AI firms are investing in technology to mitigate water loss. Companies like Microsoft are implementing AI-powered leak detection systems to quickly identify and fix leaks in aging water infrastructure, potentially recovering a significant portion of lost water. QTS is also exploring alternative solutions, such as stormwater capture.

While the EPA is working to strengthen water systems, particularly in rural areas, the broader water sector is still cautiously evaluating the full impact and potential of AI in managing national water resources.