AI Data Centers: The Battle for Power is Just Beginning

Behind the magic of Artificial Intelligence hides a less glamorous reality: a colossal demand for energy and water. The data centers hosting AI models are true resource guzzlers, often located in places where water is already an issue.

Imagine an entire city lighting up just to run your favorite chatbots. That's pretty much it. Every single request to an AI model requires complex computations, meaning scorching hot servers to cool and an amount of electricity that would make a small industry blush. In 2023, the energy demand for AI data centers grew immensely.

A single large data center can consume the equivalent of tens of thousands of homes, and to cool its servers, it can use millions of liters of water per day. For instance, Google revealed its water consumption in 2022 increased by 20% over the previous year, reaching 21.3 billion liters, largely due to its data center expansion. Quite a gulp, isn't it?

Local communities are tired of seeing their environment and resources plundered to fuel the latest tech fad. The construction of new data centers is meeting growing resistance, with citizens and activists saying "no, thank you" to these thirsty giants.

It's not just about ruined landscapes or noise. It's a battle for drinking water, for energy that could serve schools or hospitals, and for a less polluted environment. Protests are multiplying, especially in areas already stressed by climate or with limited infrastructure.

In Oregon, for example, residents successfully blocked the expansion of an Amazon Web Services data center in Umatilla County for years, citing water usage concerns. This sends a clear signal: ordinary people are no longer willing to sacrifice everything on the altar of AI progress.

The fight against AI data centers has just begun and promises to become a central issue for the development of Artificial Intelligence. Governments and companies will have to contend with popular opposition and seek more sustainable solutions, or risk hindering their own progress.

We thought AI was just algorithms and code; turns out it's also bricks, concrete, and water pipes. Companies will need to invest in more efficient cooling technologies or renewable energy sources, but these aren't quick fixes. The pressure is intense.

For the first time, the environmental and social impact of AI infrastructure is becoming a public discussion topic, with The New York Times dedicating several articles to the subject throughout 2023. This pushes Big Tech to reconsider where and how they build their AI temples.