Historical brief · 2026-11-11

AI Datacentres Plan Massive On-Site Gas Power Stations Despite Climate Risks

Developers of AI datacentres across the US, UK, and Europe are planning to build on-site gas power stations on a huge scale, aiming to generate 181 gigawatts of fossil-fuel capacity. Driven by the urgent need for power, this massive rollout risks locking in high carbon emissions that could equal the annual pollution of millions of cars.

Source image · The Guardian AI

Background

As the artificial intelligence boom drives up electricity demand, data centre developers face severe delays in connecting to national power grids. To bypass these bottlenecks and achieve faster power delivery, companies in the US, the UK, and mainland Europe are turning to fossil fuels.

Key Details

According to research by BloombergNEF, planned data centres across these regions propose building on-site gas-fired capacity totaling 181 gigawatts. About 90% of these projects are planned in the US, where they could emit roughly 345 million metric tonnes of carbon dioxide equivalent annually, matching the emissions of about 72.8 million cars. By 2035, global AI data centre electricity demand is predicted to nearly quadruple to 2,354 terawatt-hours, accounting for 7% of global demand.

Impact

While developers claim these gas generators are temporary measures while waiting for grid access, experts warn that economic incentives once they are online could keep them running for years. In the UK, the government has even greenlit temporary on-site fossil fuel generation for fast-tracked data centres, sparking strong criticism from environmental campaigners over long-term climate commitments.

View selection rationale
Media reportStage not established

Importance 79/100

Signal85Structure80Consequence70Confidence85Momentum75
About this brief
Content type
News summary
Source type
Publisher report
Original source
The Guardian AI ↗
Source published
Process
AI-assisted discovery, translation, summarisation, classification and scoring; standards and corrections are maintained by the editorial desk.
Verification
The original URL and publication metadata are retained. This brief is not independently fact-checked unless explicitly stated.
Published by
Polyemergence Editorial Desk · PolyEmergence Pte Ltd.
Conflict disclosure
No financial or commercial conflict declared.

Editorial standards → · Corrections →

POLYEMERGENCE / EXPLORE

Keep exploring

Follow the topic, not just the headline.

Explore the Artificial Intelligence archive →