Haley Zaremba

Haley Zaremba is a writer and journalist based in Mexico City. She has extensive experience writing and editing environmental features, travel pieces, local news in the…

More Info

Premium Content

• The integration of AI is causing a significant increase in global energy consumption, leading to a resurgence of fossil fuel investments and higher energy bills, while shifting policymaker priorities away from clean energy.
• Experts express skepticism about the reliability of current AI energy projections due to vague data, unproven claims about AI's energy efficiency benefits, and uncertainty about its future omnipresence.
• Despite massive investments, there are growing indications that the AI bubble may be bursting, with large language models showing limited improvement, profits slow to materialize, and financial experts projecting significant capital misallocation.

Artificial intelligence is changing our energy landscape on a global level. Runaway AI integration is suddenly causing huge growth trends in energy consumption in developed nations where growth had previously flatlined for years. It is causing a resurgence of fossil fuel investments and raising energy bills for consumers, and it’s pushing policymakers to change their priorities away from a focus on clean energy production to producing as many new energy projects as quickly as possible to meet soaring demand projections. But are these projections reliable? Experts aren’t so sure.

A recent report from the MIT Technology Review highlights three key ways in which the scope of AI’s current and future energy impact is still fuzzy at best: vague and incomplete data, unfounded claims about the role of large language models in increasing energy efficiency in other sectors, and whether AI is ever going to grow as omnipresent as current projections assume. 

There are no regulations requiring AI companies to disclose their energy usage or environmental impact, so the vast majority don’t release any official numbers. While companies like OpenAI and Google have started to become a little bit more transparent about the energy footprint of an AI query through their platforms, these numbers are not official and lack clear grounding.

Plus, individual queries are not what’s driving up AI’s energy footprint. While your chats with large language models are having some impact on AI’s energy consumption, it’s a drop in the ocean compared to the rampant integration of AI into a huge number of systems and processes both within and outside of the tech sector. 

While training and operating large language models eat up an enormous amount of energy and other resources, proponents are quick to point out that AI will be instrumental in making a huge breadth of industries more energy-efficient, with the potential to more than make up for its own energy footprint. But MIT challenges these claims, highlighting that such efficiency gains have not yet come to fruition. And while numbers on AI’s efficiency gains are lacking, new data centers are being planned at a breakneck pace. 

“AI’s integration into almost everything from customer service calls to algorithmic “bosses” to warfare is fueling enormous demand,” reports the Washington Post. “Despite dramatic efficiency improvements, pouring those gains back into bigger, hungrier models powered by fossil fuels will create the energy monster we imagine.”

But some experts wonder if these alarmist claims are warranted. In fact, there is some indication that the AI bubble might already be bursting. Despite disconcertingly enormous levels of investment, large language models aren’t getting much better at what they do. OpenAI’s launch of GPT-5 last month “was largely considered a flop, even by the company itself,” reports MIT. And it’s looking unlikely that investors are going to see a return on the billions they’ve poured into AI and data centers, with profits slow to materialize. 

Praetorian Capital CIO Harris Kupperman projects that the "AI datacenters to be built in 2025 will suffer $40 billion of annual depreciation, while generating somewhere between $15 and $20 billion of revenue." While he told Futurism that he recognizes the utility of AI, he went on to add: "I also recognize massive capital misallocation when I see it. I recognize an insanity bubble, and I recognize hubris."

When you look at the scale of AI’s influence on political, financial, and environmental decisionmaking, it’s quite concerning that a peek behind the curtain reveals not solid and consensus-led projections but pervasive fuzziness and mathematical acrobatics. And we’re already suffering the consequences of these potentially rash decisions. Energy poverty is creeping higher while consumers are footing the bill for the tech sector’s aggressive AI integration push. But the problem goes way beyond punishing utility bills – the whole economy could be on the line.

“Over the past few years, the tech industry's plans for artificial intelligence have grown from ambitious to outright treacherous, with the amount of money invested in the space so high it now poses a serious risk to the broader economy,” writes Futurism. 

By Haley Zaremba for Oilprice.com

Download The Free Oilprice App Today