The rapid expansion of AI data centers is driving a significant increase in energy consumption, primarily fueled by fossil fuels like natural gas, which exacerbates climate change and air pollution. Additionally, these centers’ heavy use of freshwater for cooling strains scarce water resources, posing serious public health risks and environmental challenges, especially in drought-affected regions.
The McKinsey report recently highlighted a concerning projection regarding the rapid expansion of data centers and supercomputers for AI development. It estimates that within the next five years, the global energy demand for these facilities will increase by an amount equivalent to half to 1.2 times the current annual energy consumption of the UK. This surge in energy use is expected to be predominantly met by fossil fuels, particularly natural gas, which was acknowledged by Sam Altman during a recent Senate hearing. Despite natural gas being considered a relatively cleaner fossil fuel, the environmental impact remains significant.
In fact, the reliance on fossil fuels for powering data centers is already causing setbacks in climate goals. Reports indicate that coal plants, which were scheduled for retirement, are now being kept operational explicitly to supply energy for data center growth. A notable example is Elon Musk’s XAI supercomputer, Colossus, located in Memphis, Tennessee. This facility is powered by approximately 35 unlicensed methane gas turbines, which emit thousands of toxic air pollutants, adversely affecting the local community’s air quality and public health.
Beyond the climate implications, the expansion of AI data centers is exacerbating public health crises, particularly concerning access to clean air and water. One critical but often overlooked issue is the massive freshwater consumption required to cool these data centers. To prevent equipment corrosion and bacterial growth, these centers predominantly use fresh public drinking water. This demand places additional strain on communities already struggling with water scarcity, as the existing infrastructure is typically designed to supply water to residents and businesses alike.
The situation is especially dire in regions facing severe droughts and water shortages. For instance, in Montevideo, Uruguay, a historic drought led to a shortage of potable water, forcing the government to mix toxic wastewater into the public supply to ensure some water availability. Vulnerable populations, including the poor who could not afford bottled water, suffered serious health consequences such as increased miscarriage rates and worsening chronic illnesses. Amid this crisis, Google proposed building a new data center that would further consume precious drinking water resources, raising ethical and environmental concerns.
Overall, the rapid growth of AI data centers presents a multifaceted ecological disaster. It accelerates climate change through increased fossil fuel consumption, worsens air pollution, and strains limited freshwater supplies, thereby impacting public health and community well-being. These challenges highlight the urgent need for sustainable energy solutions and responsible resource management in the AI industry to mitigate its environmental footprint.