Artificial intelligence is helping fossil fuel companies extract oil and gas more efficiently, potentially adding as much greenhouse gas to the atmosphere as entire countries emit annually, according to new research published in npj Climate Action. While tech companies focus heavily on measuring their own data center emissions, they're largely ignoring a far larger climate threat: the way AI tools amplify fossil fuel production itself.

The research, conducted by Will and Holly Alpine, former sustainability workers at Microsoft, models AI as a productivity enhancer across the oil and gas industry, from extraction to refining to electricity generation. Their findings suggest that AI-enabled improvements in fossil fuel operations could increase global energy-related emissions between 1.2 to 4.8 percent annually. At the low end, that's equivalent to Mexico's yearly emissions; at the high end, it rivals Russia's total annual greenhouse gas output, making Russia the world's fourth-largest emitter.

Why Are Tech Companies Missing This Climate Impact?

The disconnect stems from how the tech industry measures environmental responsibility. Most major technology companies, including Google and Microsoft, publicly disclose their operational emissions and work to reduce their data center energy consumption. However, they rarely account for what the Alpines call "enabled emissions," the pollution created when their tools help other industries operate more efficiently.

"Sustainability measures within tech companies are very much focused on operational emissions rather than enabled emissions," explained Holly Alpine, one of the study's authors. "The new paper argues that while accounting for that pollution is important, it ignores emissions that do much more damage to the climate."

Holly Alpine, Coauthor, npj Climate Action Study

This oversight is particularly significant because the benefits AI provides to renewable energy and clean technology development pale in comparison to the emissions boost it gives the fossil fuel sector. The Alpines' modeling found that AI's climate benefits for solar, wind, and other clean technologies are substantially outweighed by the productivity gains it delivers to oil and gas companies.

What Does the Chevron-Microsoft Deal Reveal About This Problem?

A recent partnership between Chevron and Microsoft illustrates the self-reinforcing cycle between AI development and fossil fuel expansion. Chevron is building a large natural gas power plant in Texas specifically to supply electricity to Microsoft's data centers. But the arrangement goes both ways: Chevron plans to use some of the computing power generated by that facility to run AI tools within its own operations.

During a call with analysts in June, Jeff Gustavson, president of Chevron's New Energies division, confirmed that Chevron would "use some of that compute" to "power AI inside of our company." This arrangement exemplifies what researchers describe as a "self-reinforcing effect between supply and demand," where technology companies and fossil fuel producers reinforce each other's growth.

"It's perfectly illustrative of the relationship between AI and fossil fuels," said Will Alpine, referring to the Chevron-Microsoft deal. "One of the key insights of our paper is that you cannot treat them independently. They are two sides of the same coin."

Will Alpine, Coauthor, npj Climate Action Study

How Does This Compare to Data Center Emissions?

The scale of enabled emissions dwarfs current projections about AI data center energy use. Researchers have raised concerns about the rapid construction of data centers straining electricity grids and requiring new power plants. Some projections suggest that gas-fired power plants being developed in the United States to supply data centers could eventually produce emissions comparable to an entire country like Australia.

Yet the Alpines' modeling indicates that AI's impact on fossil fuel productivity is significantly larger than these data center projections. This means that even as tech companies invest in more efficient cooling systems and renewable energy sources for their facilities, the net climate effect of AI deployment could still be substantially negative.

Steps to Understanding AI's Full Climate Footprint

• Operational Emissions: The direct energy consumption of data centers and computing infrastructure, which tech companies currently measure and report publicly.
• Enabled Emissions: The indirect greenhouse gas pollution created when AI tools help other industries, particularly fossil fuels, operate more efficiently and produce more output.
• Supply Chain Emissions: The energy and raw materials required to manufacture advanced chips, servers, and other hardware components used in AI systems.
• Water and Infrastructure Impact: The enormous quantities of water required for data center cooling and the environmental costs of constructing new power generation facilities.

What Do Energy Experts Say About These Findings?

Jon Koomey, an energy researcher who was not involved in the study, validated the Alpines' approach and conclusions. He noted that while machine learning can improve data center cooling efficiency by 30 to 40 percent, it can simultaneously make fossil fuel extraction "much cheaper and faster".

"Machine learning can make data center cooling 30 to 40 percent more efficient, but could also make fossil fuel extraction much cheaper and faster. How that nets out nobody yet knows for sure, but this new research is a credible attempt to answer that question using a macroeconomic model," said Jon Koomey, energy researcher.

Jon Koomey, Energy Researcher

Koomey also criticized what he called "hand-waving arguments" from AI boosters who claim the technology will solve climate problems without examining its effects across all industries. "Such arguments ignore the effects that AI will have on ALL industries, not just renewable energy and efficiency," he explained.

Why Are Major AI Companies Avoiding Climate Disclosure?

Unlike established tech giants such as Google and Microsoft, many leading AI firms, including OpenAI, Anthropic, and SpaceX, have disclosed relatively little about their environmental impact. These companies do not publicly release detailed greenhouse gas emissions data, make clear commitments to carbon neutrality, or publish comprehensive sustainability reports.

The lack of transparency reflects a broader shift in corporate climate accountability. Investors had previously expressed growing concern about AI companies' environmental footprint, but amid a backlash against corporate climate policies, many investors and businesses have become less willing to raise environmental concerns. Simultaneously, U.S. regulators have eased pressure on companies to address climate-related risks.

The scale of these companies' operations now places them among the most powerful players in the global technology sector, with enormous investments in computing infrastructure, data centers, and energy supply. Yet their environmental accountability remains minimal compared to older, more established technology firms.

The Alpines' research suggests that addressing AI's climate impact requires a fundamental shift in how the technology industry measures and reports its environmental footprint. As long as tech companies focus exclusively on their own operational emissions while ignoring the enabled emissions their tools create in other sectors, the true climate cost of AI will remain hidden from public view and regulatory scrutiny.