AI's Insatiable Appetite for Power

­We’ve all heard that crypto is a ravenous energy hog (about 70 TWh per year in the U.S., according to the EIA). But what about everyone’s favorite talking point/ prognosticator of doom, AI?

Artificial intelligence is slowly infiltrating every corner of our lives, from search engines to generative tools and even the final frontier – fully-autonomous vehicles. It also has a bottomless appetite for the juice.

Of course, if we’re talking about AI energy usage, it’s worth pointing out that one of AI’s major selling points is energy efficiency – that is, using machine learning to analyze massive reams of data to foresee potential outages and equipment failures and find new oil stores, plus AI-driven smart home solutions, AI-powered robots in the energy sector, and countless other applications.   

As the IEA points out, “The technology is uniquely placed to support the simultaneous growth of smart grids and the massive quantities of data they generate. Smart meters produce and send several thousand times more data points to utilities than their analogue predecessors.”

But that roundabout energy savings is at least partially offset by AI’s high electricity demands.

As the MIT Tech Review notes, producing one generative AI image can draw enough juice to power your phone, and 1,000 images produces as much carbon dioxide as about four miles in a gas-powered car.

And while it’s tough to isolate AI’s specific energy demands, apparently, electricity consumption from data centers, AI, and cryptocurrency, together, amounted to about 2% of global electricity demand in 2022, and those numbers will spike as those three technologies take center stage.

Some are even saying that AI could derail climate goals.

“In 2020, we unveiled what we called our carbon moonshot. That was before the explosion in artificial intelligence,” said Brad Smith, Microsoft’s president.

The key will be finding that careful equilibrium between energy savings and energy consumption.