A recent analysis by the Sierra Club reveals that nearly 15 years of hard-fought pollution reduction in the United States could be undone by a new wave of data centers powered by gas plants. The analysis predicts that if current AI developments proceed as planned, the US energy grid will emit 2.1 billion metric tons of carbon dioxide by 2030.
The analysis, reported by Bloomberg, indicates that the AI industry and its data center developers could see a 42% rise in emissions over the next five years. This would translate to an increase of 618 million tons compared to 2025, which was already a near-record year for US carbon emissions.
"The numbers continue to sort of boggle our minds," Jeremy Fisher, principal advisor for climate and energy at Sierra Club, told Bloomberg. This dramatic surge in energy demand is driven largely by the immense computational needs of AI models, which require large, power-hungry data centers.
A primary concern is that to meet this growing demand, utility companies are turning to building and running new natural gas plants, a major source of carbon dioxide emissions. This could undermine the progress made in renewable energy sources like solar and wind.
The report highlights the tension between technological advancement and environmental goals. While AI has the potential to solve some of the world's biggest problems, its carbon footprint presents a serious challenge that requires innovative solutions in energy efficiency and clean power sources.
Environmental advocates are calling for technology companies to take greater responsibility for their environmental impact and to invest more heavily in clean energy to power their data centers.
Source: Sierra Club, Bloomberg
🌐
