Unprecedented Water Consumption by AI Data Centers
The number of data centers (DCs) is growing exponentially due to the boom in AI technologies. These AI-driven data centers demand vast amounts of water, and recent data indicates that data centers globally consume approximately as much water as the entire human population. However, there are two crucial aspects to clarify. Firstly, this figure pertains specifically to AI data centers, not all data centers. Secondly, the water consumption refers solely to bottled water, yet the figures remain staggering: estimates range between 312.5 billion to 764.6 billion liters this year alone.

Gigantic Energy Consumption
The electricity equation is equally intriguing. Sources report a consumption of 23 gigawatts (GW), surpassing the total power infrastructure of Bitcoin mining from last year. These calculations have emerged through research led by Alex de Vries-Gao from the Amsterdam Institute for Environmental Studies. To derive these numbers, de Vries-Gao utilized analyst estimates and matched them with financial result conference calls and other publicly accessible information to extrapolate the volume of AI equipment deployed by these companies and their energy usage. Recent updates from the research community affirm that AI data centers are now one of the fastest-growing electricity consumers globally.
Massive Carbon Footprint
On the basis of these data points, it has been calculated that AI systems release between 32.6 million to 79.7 million metric tons of carbon dioxide annually. Comparatively, Singapore’s entire greenhouse gas emissions in 2022 totaled only 53 million metric tons. This stark contrast highlights an escalating environmental challenge posed by the rapid expansion of AI technologies. Many environmental experts emphasize the urgent need for sustainable solutions, pointing to innovative strategies employed by some tech giants to mitigate these emissions.
It is noteworthy that the construction of new data centers is often constrained by a shortage of electricity, and occasionally, in some regions, a lack of water availability. Just yesterday, we reported that a significant proportion of the world’s data centers are situated in regions particularly unfavorable regarding temperature.