Researchers who previously worked on sustainability at Microsoft have found that artificial intelligence could substantially accelerate fossil fuel extraction and production, potentially offsetting climate benefits AI brings to renewable energy sectors. Their economic modelling, based on documented AI gains reported by oil and gas companies, suggests that AI-driven productivity improvements across the fossil fuel industry—from resource discovery through refining—could increase global energy-related emissions between 1.2 and 4.8 percent annually, a scale comparable to the entire emissions output of countries like Mexico or Russia.

The study highlights a critical measurement blind spot: while technology companies track their own operational emissions and supply chains, they do not quantify what researchers call 'enabled emissions'—the additional greenhouse gases their software and systems help fossil fuel companies produce. This gap is evident in recent commercial partnerships, such as Chevron's arrangement with Microsoft to build a gas-fired power plant for data centres, with Chevron explicitly planning to use the resulting computational capacity to power AI within its own operations, demonstrating how the technology and fossil fuel sectors have become financially and strategically merged.