Energy Demand of Data Centers Increases Elevenfold Due to AI

According to projections by the Öko-Institut commissioned by Greenpeace Germany, total electricity use for centralized data processing — including all data centers — will reach around 1,400 billion kilowatt-hours by 2030. This will be accompanied by an increase in greenhouse gas emissions from data centers, rising from 212 million tons in 2023 to 355 million tons.

Experts also point to additional burdens from the demand for water used in cooling systems, which will nearly quadruple to 664 billion liters over the same period, as well as up to five million tons of additional electronic waste resulting from the expansion of data centers and the increased capacity needed to meet growing AI usage.

Use of Fossil Fuels

According to experts, AI is already being used to accelerate environmentally harmful business practices — for instance, to speed up and optimize the exploration of new fossil fuel sources. This promotes monocultures and further encourages private consumption. Negative environmental impacts may also occur unintentionally due to flawed data.

“Even in the years ahead, data centers will continue to rely on fossil fuels such as natural gas and coal — with correspondingly high environmental costs,” says Jens Gröger, research coordinator for sustainable digital infrastructures at the Öko-Institut.