AI Isn’t Draining the Grid - New Study Says We Got It Wrong

⚡ KEY TAKEAWAYS

A new study finds AI’s overall energy use is far smaller than widely believed.
Researchers say AI barely registers at the national level, despite local data-center strains.
The findings suggest AI could become a climate solution rather than a climate burden.

For years, artificial intelligence (AI) was framed as a looming climate disaster, a runaway energy hoarder threatening to overwhelm power grids. Yet new research suggests the story is far more surprising.

Scientists now say AI’s actual energy footprint is much smaller than feared, raising big questions about whether we misjudged one of the world’s fastest-growing technologies. The findings reveal a reality that challenges a major tech myth and potentially reshapes how we think about AI’s future impact on the planet.

New Study Shows AI’s Energy Impact Barely Registers Nationally

Researchers from the University of Waterloo and Georgia Tech merged U.S. economic data with AI adoption estimates across industries to understand how much energy AI actually consumes. Their results point to a dramatic mismatch between public fear and measurable reality.

Even with AI expanding rapidly, its total U.S. electricity use is comparable to the annual consumption of Iceland. At national and global levels, that amount is so small it barely registers.

Dr. Juan Moreno-Cruz, Canada Research Chair in Energy Transitions, says the nuance matters. From a local perspective, communities hosting data centers may see major strain, and in some cases, even doubled electricity output. But on a national scale, AI’s climate effect remains minimal. In his words:

“If you look at that energy from the local perspective, that’s a big deal because some places could see double the amount of electricity output and emissions. But at a larger scale, AI’s use of energy won’t be noticeable.”

He explained that the real risk isn’t the technology itself, but where and how AI-driven demand clusters. The bigger picture shows that AI’s energy footprint today remains far too small to shift national emissions in any meaningful way. As Dr. Moreno-Cruz pointed out:

“For people who believe that the use of AI will be a major problem for the climate and think we should avoid it, we’re offering a different perspective. (…) The effects on climate are not that significant, and we can use AI to develop green technologies or to improve existing ones.”

Why AI Could Become a Climate Tool, Not a Climate Threat

The study goes further by suggesting that fears about AI harming environmental progress may overlook an important upside. If AI adoption continues at similar levels, its carbon impact remains low enough that its benefits could outweigh its footprint.

Moreno-Cruz notes that AI can accelerate green technology development, optimize energy systems, and even reduce emissions by improving efficiency across high-impact industries. Instead of draining the grid, AI could help modernize it.

Computational model designed after the human brain that can detect faults in power lines before wildfires start. Source: NREL

The researchers also analyzed which jobs and sectors AI might meaningfully affect, finding that AI could support economic productivity without imposing heavy climate tradeoffs. Their next phase is applying the same model to other countries to see if the pattern holds internationally.