AI Adoption in Energy Sector: A Carbon Catastrophe in the Making?

A stark warning has emerged from a new report published in the prestigious Nature journal this week: the widespread adoption of Artificial Intelligence (AI) in the energy sector is poised to significantly escalate global emissions, with AI-assisted oil and gas drilling leading to a substantial increase that could overshadow any environmental benefits gained from green energy applications. This revelation comes as the planet grapples with escalating heatwaves and wildfires, painting a grim picture for the future.

The report unequivocally states that “AI-driven productivity gains enable more emissions than they avoid—reinforcing fossil incumbency rather than displacing it.” This conclusion points to a critical flaw in current assumptions about AI’s role in the energy transition.

According to the research, AI could contribute an additional 1.8 billion tonnes of carbon dioxide to the atmosphere annually. To put this into perspective, this figure is equivalent to nearly 5 percent of the total carbon output produced by all energy companies in 2024. Alarmingly, the carbon footprint of AI-assisted oil and gas drilling is projected to be even larger than that of the climate-damaging data centres required to power the technology. Data centres, themselves significant emitters, produced approximately 180 million tonnes of carbon in 2025, a figure expected to surge to 500 million tonnes within the next decade.

The report highlights that a substantial 44 percent of fossil fuel companies are already leveraging AI, with a further 45 percent planning to integrate it within the next three years. The study’s methodology goes beyond direct emissions, focusing on “enabled” versus “avoided” emissions, providing a more comprehensive understanding of AI’s environmental impact.

While AI offers promising applications in the renewables sector, such as optimizing weather pattern tracking and maintenance schedules for batteries, these benefits are projected to be dwarfed by the emissions generated through its application in fossil fuels. The report indicates that AI in renewables would only avoid 500 million tonnes of carbon emissions – a mere fraction compared to the potential emissions unleashed by the fossil fuel industry’s adoption of the technology.

“Renewables productivity gains must exceed fossil fuel gains by approximately 4-5 times for net emissions to break even,” the report emphasizes, further warning that “small fossil fuel gains are sufficient to outweigh much larger improvements in renewables generation.”

Crucially, the report’s projections are based on an assumption of “parallel adoption,” meaning AI would be utilized equally across both renewable and fossil fuel sectors. However, this assumption may be overly optimistic. Global energy supply disruptions, exacerbated by conflicts in Iran and Ukraine, have led to a surge in fossil fuel investments and a significant decline in renewable investments. This imbalance suggests that AI technology is likely to be deployed far more extensively in the fossil fuel industry, leading to significantly fewer “avoided emissions” than the report’s projections.

The report concludes that “AI does not inherently decarbonise or intensify emissions–its productivity gains reinforce the economic and institutional structures in which it operates.” This statement underscores the critical need for policy intervention and societal shifts to ensure that AI’s powerful capabilities are directed towards sustainable goals rather than exacerbating environmental crises.

Local Resistance Against Data Centres Gains Momentum

Concurrently, local resistance against the burgeoning growth of data centres is becoming a significant flashpoint in the broader climate change battle across Britain. These campaigns are uniting diverse groups, addressing concerns ranging from working-class housing to climate catastrophe and warfare.

A recent sold-out meeting in east London saw over 150 people gather to protest the planned development of a £500 million data centre at the site of the Truman Brewery on Brick Lane. Despite the Labour government fast-tracking data centre developments, classifying them as “critical national infrastructure,” local residents hold a different view.

Eileen, an attendee at Monday’s meeting, told Socialist Worker that the proposed data centre would be “absolutely disastrous.” She highlighted the stark contrast between the expedited development of data centres and the delays in housing projects, noting that “Tower Hamlets has one of the highest waiting lists for secure housing at over 24,000.”

Eileen sees immense potential in this burgeoning campaign. “Councillors, trade unions, long-time campaigners and residents all attended the meeting. There could be real pull and influence from this campaign,” she stated, advocating for a national movement against data centres in Britain.

She further explained that the Truman Brewery data centre alone would consume as much water as 18,000 homes, at a time when 20 million homes in Britain are facing water restrictions. “We could get this overturned, but we need numbers and we need objectives,” Eileen asserted. A sentiment echoed by an activist at the meeting: “‘Every time they put a brick up we need to take a brick down.’ This is a really popular response locally and this is what we need.”

The convergence of these concerns—the environmental impact of AI in energy and the localized battle against data centres—underscores a growing imperative for a just transition. This includes an end to new oil and gas projects and the repurposing of workers’ skills, without loss of pay, for sustainable industries.