AI Gas Power Plants Could Emit 287M Tons Annually in Texas
A new report from the Environmental Integrity Project warns of a massive surge in greenhouse gas emissions from AI infrastructure.
The Environmental Integrity Project (EIP) has released a comprehensive report revealing that a wave of proposed natural gas power plants built specifically for AI data centers could derail climate goals. In Texas alone, 32 proposed gas-fired plants could collectively emit more than 287 million tons of greenhouse gases each year.
Key details
The 44-page report, titled "Environmental Cost of AI's Energy Use: Carbon, Water and Land Footprints," identifies Texas as the national leader in proposed fossil fuel infrastructure for artificial intelligence. Researchers estimate the 32 projects in Texas could emit thousands of tons of particulate matter and nitrogen oxides alongside their carbon footprint.
Specific findings include:
- Project Matador: Located near Amarillo, Texas, this facility alone is projected to release more than 40 million tons of greenhouse gases per year.
- Pennsylvania Impact: At least seven proposed gas plants in Pennsylvania could emit 68 million tons of carbon dioxide equivalent annually, representing a 24% increase in the state's total emissions.
- National Trends: Nearly 90% of proposed gas plants nationwide are located in counties with life expectancies below the national average, and 40% of residents living within three miles of the projects are in low-income households.
Why this matters
The EIP report highlights a direct conflict between the technology industry’s rapid infrastructure growth and global decarbonization efforts. As AI companies race to secure power for high-density compute clusters, many are turning to behind-the-meter natural gas generation to bypass grid interconnection delays that can stretch for years. This shift risks locking in decades of fossil fuel reliance at a time when climate scientists warn of the urgent need to phase out carbon emissions.
Context
This finding follows a series of state-level responses to the AI data center boom. While states like Pennsylvania have recently moved to repeal tax incentives and implement stricter grid standards, the EIP report suggests that federal and state permitting for off-grid power is moving ahead with limited oversight. The scale of the projected emissions in Texas — 287 million tons — rivals the entire domestic carbon footprint of some medium-sized nations, illustrating the sheer intensity of AI's physical footprint.
Risks and open questions
The report raises significant health concerns for communities situated near these planned facilities. With at least 188 schools located within three miles of proposed projects, the impact of nitrogen oxides and particulate matter on local air quality is a primary concern for advocacy groups. A major open question remains whether the industry's significant investments in renewable energy and "next-generation" cooling can actually scale fast enough to offset these massive fossil fuel commitments.
What happens next
The findings are expected to fuel calls for more stringent federal oversight of data center environmental impacts. Following the report's release, PJM Interconnection stakeholders have already begun advancing a "reliability backstop" procurement plan to address near-term power shortfalls. Regulatory bodies, including the Federal Energy Regulatory Commission (FERC), are likely to face increased pressure to mandate transparent resource reporting and environmental impact assessments for large-scale AI power projects.
Source: KERA News Published on AI Usage Global, author: AUG Bot
