The Consequence of AI: When AI's Demand Overrides the Earth's Needs

By 2028, it is predicted that artificial intelligence (AI) data centers will exhaust more than 352 terawatt-hours of electricity per year, in turn depleting more than 150–450 billion gallons of water for cooling (Law, 2025). AI data centers threaten the planet's fresh water supply, as their energy is harnessed through electrical components reliant on extreme amounts of water for cooling (Yañez-Barnuevo, 2025). Whilst data centers use massive amounts of water to operate, it is possible for AI to become both sustainably powered and conserve water through increasing the flexibility and efficiency of the energy grid by involving the United Nations (UN) in passing renewable energy initiatives.

Of the 400 to 700 million weekly users of OpenAI in 2025, over 73% of their prompts were non-work related, translating to AI transforming from a corporate luxury to a mundane resource (Law, 2025). Concerningly, data centers use one bottle of water for every 20 prompts, which equalled 49 billion gallons of water in 2025 in Texas alone (Gorey, 2025). A single ChatGPT prompt "consumes 10 times the electricity of a Google Search," rendering it essential that consumers understand where AI's power comes from in order to realize the impact their choice of browser or AI has on their own environment (UNEP, 2025).. From an ecological perspective, "all water is local," meaning that how water is managed within a region can affect everything it interacts with later down the line (Gorey, 2025). Data centers impact surrounding communities by using enormous amounts of this shared resource, even evaporating or polluting the water with lead and mercury (United Nations Environment Programme). Two-thirds of data centers have been built in hot, dry climates where water is scarce, prompting many electricity bills to double in 2025 as water levels dwindle (Gorey, 2025). Some argue that the solution to the data centers' excessive water usage is to locate them in water-saturated areas, yet in London where the River Thames supplies energy to more than 80% of the U.K.'s data centers, usage peaks during times of drought (Gorey, 2025). 

While AI data centers threaten the planet's fresh water supply through employing an excessive amount of water to cool electrical components, there is a possible solution – one focused on energy conservation and sustainability. Energy grid flexibility could be a real long-term solution to the excess water used by data centers, by conserving surplus energy and creating "[stability] during times of peak demand" (Gorey, 2025). The U.K. serves as a prime candidate for flexible grids – its original grids are "nearly maxed out," requiring the use of polluting off-grid generators (Gorey, 2025). Flexible grids make it possible to "reap the benefits [and economic growth] of AI," while maintaining sustainable water and energy practices (Gorey, 2025). 

António Guterres, the UN Secretary-General, advocates that "every major tech firm [must] power all data centers with 100% renewables by 2030," endorsing flexible grids' implementation (Guterres, 2025). Most power grids are owned by private investors who prioritize their profit over sustainability, as "electric utilities are seen as valuable long-term investments that earn around 10% returns," meaning that investors are less likely to change the energy system for fear of losing money. While private investors own portions of the grid, they are ultimately overseen by federal and state regulators. The goal of "100% renewable" energy is achievable through federal regulators who can help enforce the UN's energy initiatives for a more flexible grid (Guterres, 2025). With the help of governmental bodies, flexible grid initiatives can make sustainable energy a reality.

The impact of data centers on surrounding environments and water resources currently inhibit a sustainable future. However, action can be taken through promoting flexible energy grids within the UN's initiatives and enforced by state regulators. AI usage will assuredly grow, yet through implementing a sustainable and flexible energy grid, a negative environmental impact does not have to expand with it. 

Sources

https://theconversation.com/ai-chatbots-are-becoming-everyday-tools-for-mundane-tasks-use-data-shows-266670 

https://www.lincolninst.edu/publications/land-lines-magazine/articles/land-water-impacts-data-centers/ 

https://www.eesi.org/articles/view/data-centers-and-water-consumption#:~:text=During%20immersive%20cooling%2C%20water%20or,irrigation%20and%20flooding%20in%20agriculture

https://www.unep.org/news-and-stories/story/ai-has-environmental-problem-heres-what-world-can-do-about 

https://www.un.org/sg/en/content/sg/statements/2025-07-22/secretary-generals-remarks-climate-action-moment-of-opportunity-supercharging-the-clean-energy-age-delivered-scroll-down-for-all-french 

https://iee.psu.edu/news/blog/why-ai-uses-so-much-energy-and-what-we-can-do-about-it#:~:text=Smaller%20institutions%20with%20limited%20GPU,contributing%20to%20greenhouse%20gas%20emissions

https://www.unep.org/news-and-stories/story/ai-has-environmental-problem-heres-what-world-can-do-about

Previous
Previous

Coffee

Next
Next

Diving Into Design: 8th Grade NuVu Experience