Why do AI data centers use so many resources?

Here is a 199-word summary of the news article: The rapid growth of AI has led to a surge in the construction of new data centers, which are placing a significant burden on the environment and local communities. These data centers consume massive amounts of energy and water due to the resource-intensive hardware they use, particularly graphics processing units (GPUs) which are essential for running AI models. The article explains that GPUs are much more power-hungry than traditional CPUs, generating significant heat that requires intensive cooling systems. Data centers are estimated to have more than doubled their energy usage from 2014 to 2023, now accounting for nearly 4.4% of the US's total electricity consumption, with projections of up to 12% by 2030. Additionally, data centers have a huge water footprint, consuming billions of liters annually, often drawing from stressed water sources. The article explores potential solutions, including liquid cooling systems, free cooling using natural resources, and renewable energy sources like geothermal power. However, the article emphasizes the need for greater transparency and more efficient hardware and software design to address the sustainability challenges posed by the rapid growth of AI.