AI Tools Boost Oil Output, Adding Up to 1.8 Gigatonnes CO2 Annually

Artificial intelligence applications deployed by oil and gas companies are generating climate emissions far larger than the carbon footprint of AI data centers themselves, according to new research published in Nature.

The study, conducted by Will and Holly Alpine—former Microsoft employees who founded the nonprofit Enabled Emissions Campaign—quantifies what they call "enabled emissions": the additional greenhouse gases released when AI makes fossil fuel extraction faster, cheaper, and more profitable. Their analysis found these tools increase global emissions by 0.47 to 1.8 gigatonnes of carbon dioxide annually, representing 1.2 to 4.8 percent of worldwide energy-related emissions in 2024.

That figure is up to 13 times larger than the International Energy Agency's 2025 estimate for data center emissions.

How AI accelerates fossil fuel production

The researchers used economic modeling to simulate AI's impact across multiple adoption scenarios, from minimal to heavy use. They examined how AI affects both fossil fuel and renewable energy sectors, converting those effects into "productivity shocks"—measures of how much AI speeds up each industry.

The practical applications are already visible across the oil and gas sector. AI has compressed seismic data analysis from roughly one year to approximately two weeks, according to Boston Consulting Group research cited in the study. This acceleration makes identifying promising drill sites dramatically faster and less expensive.

At the CERAWeek conference in March 2026, Chevron CEO Mike Wirth noted that energy companies are using AI to streamline operations through creative partnerships with tech firms. Consultancy Wood Mackenzie estimates AI tools could unlock access to an additional 470 billion barrels of oil from existing fields.

The asymmetric effect

Will Alpine, the study's lead author, emphasized that AI "acts as an economic lever that reinforces the viability and dominance of fossil fuels." The research revealed a troubling asymmetry: to merely offset AI's boost to fossil fuel production and keep emissions flat, AI would need to improve renewable energy productivity four to five times more than it improves oil and gas.

Clara Vondrich, senior policy counsel at Public Citizen, characterized the findings as exposing tech companies as "lead accomplice to the fossil fuel industry," noting that AI developers are selling proprietary tools specifically designed to accelerate oil production.

Why it matters

This research challenges the dominant narrative around AI and climate impact. While much attention focuses on data center energy consumption and AI's potential to optimize clean energy systems, the study reveals a larger, largely unmeasured problem: AI is making fossil fuel extraction economically viable at scales that would otherwise be unprofitable. Until these "enabled emissions" are measured and regulated, policymakers and companies are addressing only a fraction of AI's true climate footprint—potentially undermining broader decarbonization goals even as tech companies tout renewable energy commitments for their own operations.

Microsoft, whose former employees conducted the research, stated that its approach must evolve with AI's changing context as it pursues carbon-negative goals. The company said it matched all global electricity consumption with renewable energy in 2025 and is prioritizing carbon-free electricity for grids where it operates.

The findings were first reported by Inside Climate News.