Most people use AI and online services without thinking about where all of it comes from at all. I didn’t really think about it either. But behind the scenes are data centers, which are large buildings filled with computers and servers. These servers need a lot of electricity to keep running and also get insanely hot, which means they need to be cooled down. As more data centers are being built, especially because of AI, they are starting to use a lot more resources.
Water is a huge concern, and because of the heat created by the equipment, it needs to be cooled by cooling systems, which use water. According to Lawrence Berkeley National Laboratory, data centers in the United States used around 66 billion liters of water in 2023.
Another problem is the electricity used by the data centers. The International Energy Agency states that data centers used around 485 terwatt-hours of electricity worldwide in 2025. By 2030, that number could reach 950 terawatt-hours in that year.
The effects can also reach people unrelated to the data centers. PJM, an organization that manages electricity for around 67 million people in the U.S., proposed a plan where data centers could have to use backup power during emergencies. PJM is facing a huge amount of 6.8-gigawatt reliability shortfall. Reuters states that when this can’t be fixed, millions of homes and businesses could face a high risk of blackouts, and many problems can occur.
Data centers are very important as they provide AI and online services people use all the time, but I believe we have to pay closer attention to how much electricity and water they use and also how that can affect the people living around them.
The data sources are solid and the key statistics are valid. Only I would note that the 66 billion liters is for water use, and PJM’s 6.8 GW figure is a forecast reliability shortfall, not a current deficit. Otherwise, the argument is solid.