AI Computing Surge Could Boost US Power Sector Emissions by 42%
A report by the Sierra Club reveals that surging power demands from AI data centers could increase US power sector carbon emissions by 42% by 2030. This projected spike would effectively wipe out 15 years of greenhouse gas emission reductions. This trend highlights a major conflict between the rapid expansion of artificial intelligence infrastructure and global environmental sustainability goals. The heavy reliance on fossil fuels to power these facilities threatens to reverse long-standing climate progress. If all planned AI computing projects are realized, power generation carbon dioxide emissions could rise by 618 million tons by the end of 2030. This additional pollution is roughly equivalent to the annual electricity-related carbon emissions of 128.8 million households.
## BACKGROUND
The United States has experienced a decline in net greenhouse gas emissions from a peak in 2007 down to 2025. However, the explosive growth of energy-intensive data centers, combined with government policies supporting fossil fuel infrastructure and easing pollution controls, is driving power sector emissions back upward.