AI Could Boost Oil Emissions More Than Clean Energy: Study
Get the Energy newsletter
Daily energy & climate — solar, EVs, oil, the policy fights and tech bets shaping the transition. Free.
- Will and Holly Alpine, former Microsoft sustainability workers, published research in npj Climate Action finding AI could raise global energy emissions by 1.2–4.8% by enhancing productivity across the fossil fuel industry, from extraction to refining to electricity generation.
- The research modeled AI-driven emissions from oil and gas at a range equivalent to Mexico's annual output at the low end and Russia's at the high end, finding that increase outweighs AI's benefits to solar, wind, and other clean technologies and significantly exceeds projected data center buildout emissions.
- The Alpines quit Microsoft in early 2024 over the company's continued work with the oil and gas industry and coined the term "enabled emissions" for the AI-supported greenhouse gas pollution that current tech-sector sustainability metrics do not capture.
- Chevron will build a behind-the-meter gas plant in Texas to power Microsoft data centers, with Chevron New Energies president Jeff Gustavson telling analysts the oil company will "use some of that compute" to power AI inside Chevron itself.
- Independent energy researcher Jon Koomey, who was not involved in the study, called the conclusions solid, noting that while machine learning can make data center cooling 30–40% more efficient, it can also make fossil fuel extraction much cheaper and faster.
- The modeling used oil and gas industry reports on demonstrated AI productivity gains and applied them across extraction, refining, and electricity generation to estimate how the productivity boost translates into increased global emissions.
Why it matters: This research reframes the AI-emissions debate: the carbon footprint of AI's own compute is small compared to the emissions unlocked when AI helps the fossil fuel industry extract and burn more oil and gas. Tech companies like Microsoft have direct financial stakes in enabling that production — Chevron's Texas gas plant will both power and consume Microsoft data-center compute, making the AI–fossil fuel loop a closed business arrangement rather than an abstract externality.
Ask SkimNews
