AI will do more to grow fossil fuel output than clean energy, new study warns

Key points

• A new study modelled 64 scenarios and found AI causes a net rise in yearly carbon emissions of 0.47 to 1.8 gigatonnes, roughly 1 to 5 percent of the global energy sector's annual output.
• AI could increase recoverable oil and gas reserves by 5 percent by making drilling and extraction more efficient.
• Deepwater drilling projects, which operate on the ocean floor and are among the most expensive energy operations on Earth, could see costs fall by 10 percent with AI assistance.
• The study found AI's emissions-cutting benefits for wind and solar don't offset the extra fossil fuel production it enables.

The promise was simple: artificial intelligence would help speed up the clean energy transition. Smarter grids, faster wind farm planning, cheaper solar. The numbers tell a messier story.

A new study, reported by The Guardian, compared what AI can do for renewable energy against what it can do for fossil fuels. Researchers ran 64 separate scenarios, adjusting assumptions about adoption speed, productivity gains and sectoral uptake. Every single scenario pointed the same direction. Net carbon emissions went up.

So what exactly did they find?

AI raises net yearly carbon pollution by between 0.47 and 1.8 gigatonnes. The top end of that range is roughly equivalent to adding another Germany-sized economy's annual energy emissions to the atmosphere.

The reason isn't complicated. Fossil fuel companies have enormous incentives to squeeze more out of every well and every mine, and AI gives them a powerful new tool. The study found it could boost recoverable oil and gas reserves by 5 percent, helping geologists find resources that would otherwise stay in the ground.

Deepwater projects, the offshore operations that sit miles beneath the ocean surface and cost billions to run, could get 10 percent cheaper with AI assistance. Lower costs mean more projects become financially worth doing. We reported in August on Amazon's Texas gas plant permitted to emit more CO2 than any coal facility in America, a sign that AI's energy appetite is already reshaping the fossil fuel calculus before any study's warnings have landed.

Does AI not help renewables too?

It does, but not enough. AI genuinely improves wind output forecasting, solar panel placement and grid management. The study doesn't dispute that. The fossil fuel industry is larger and has more to gain from efficiency improvements in the short term, because it operates at a scale renewables haven't yet reached.

The clean energy sector is growing fast from a smaller base. AI helps both sides. It just helps the bigger, dirtier side more, right now.

What does this mean for ordinary people?

For anyone tracking energy bills or climate targets, this study is a signal that AI alone isn't a climate fix. It's a tool, and its impact depends entirely on who uses it and for what.

Policymakers who assume AI will automatically tilt the energy system toward cleaner sources may need to reconsider. Without rules directing AI investment toward renewables and away from new fossil fuel extraction, hitting climate goals gets harder.

The study doesn't argue for banning AI in the energy sector. It argues for paying much closer attention to where the benefits land. That distinction matters, and nobody in power should be allowed to blur it.