The power-hungry AI boom is making power grids dirtier and less reliable

The rapid development of artificial intelligence is causing a sharp increase in demand for electricity, which may pose a challenge to networks and the companies behind this technology.

From simple ChatGPT queries – which alone can use as much electricity as a 60-watt light bulb in 10 minutes – to more complex images and video creations, to rapidly expanding enterprise deployments and hardware integrations, the AI ​​boom is increasing demand for power on an already taut net.

As Bloomberg notes, in 2023, data centers have already consumed more energy than entire countries such as Italy and Taiwan, and their energy demand has increased more than sevenfold since 2008, even as chips have become more energy efficient.

Artificial intelligence’s energy demand will only increase, according to a report by Boston Consulting Group, which puts current data center energy consumption at 2.5% and is projected to triple to 7.5% by 2030. This increase in demand may not be sustained. supplemented by existing generation capacity, and certainly not by renewable sources.

For example, in Texas, where devastating power outages during a winter storm killed more than 240 people in 2021, a significant increase in energy demand could mean less reliable power for everyone.

“Cryptocurrency miners and data centers will account for more than 50% of the additional growth (relative to energy demand). We need to take a closer look at these two industries,” Texas Lieutenant Governor Dan Patrick said in a June 12 post to X. “We want data centers, but it can’t be the Wild West of data centers and cryptocurrency miners. it breaks our network and turns off the lights.”

And while big tech companies may be required to use green energy, the Washington Post reports that because they operate on the same grid as typical energy consumers, these data centers typically consume most of the available renewable energy, leaving energy providers with dirty energy to fill in the gaps.