The rapid expansion of artificial intelligence in Texas is putting a significant strain on the state's water resources. Data centers, which are essential for powering AI systems, consume enormous amounts of water to cool their servers. According to recent estimates, Texas data centers consumed 463 million gallons of water in 2023 and 2024 alone, with projections suggesting this number could skyrocket to 400 billion gallons by 2030, accounting for nearly 7% of the state's total water consumption.
The water-intensive nature of AI is a major concern, particularly in a state like Texas, where water resources are already under pressure. Training a single AI model like ChatGPT-4 requires significant water resources, equivalent to the water usage of 30,000 households over 100 days. A brief conversation with a generative AI like ChatGPT can consume up to 500ml of freshwater, which may seem trivial but adds up to significant water stress when considering millions of interactions.
Cities like San Antonio enforce Stage 3 water restrictions, limiting lawn watering to once a week and imposing surcharges on excessive residential use, while data centers continue to guzzle millions of gallons daily with little oversight. The rapid growth of AI infrastructure is outpacing the state's ability to regulate it, raising concerns about the long-term sustainability of water use in Texas.
To mitigate the impact of AI on water resources, potential solutions include shifting workloads to cooler nighttime hours, implementing rainwater harvesting systems, and developing advanced cooling technologies. Microsoft has pledged to become "water positive" by 2030, aiming to replenish more water than it consumes, but critics argue these efforts often occur far from the communities most affected by water shortages.
