New natural gas projects linked to major tech companies could generate greenhouse gas emissions exceeding those of some small nations, according to air permit documents reviewed by Wired.
A review of air permits reveals that just 11 US data center campuses—connected to OpenAI, Meta, Microsoft, and other tech firms—are poised to emit over 129 million tons of greenhouse gases annually through new gas-powered projects.
For context, this emission level rivals the total carbon output of numerous countries. The scale underscores the environmental footprint of AI infrastructure, which requires enormous computational power and energy consumption.
The permits detail plans for natural gas generation at these facilities, a choice made despite growing pressure on tech companies to adopt renewable energy sources. The findings highlight a tension between the industry's expansion and climate commitments.
These projects represent significant investments in data center capacity as companies race to build AI infrastructure. However, the reliance on fossil fuels contradicts sustainability pledges many of these organizations have publicly announced.
The revelation comes amid broader scrutiny of tech's environmental impact and questions about whether current renewable energy availability can meet AI's surging power demands.
The UK government has dismissed Green party leader Zack Polanski's proposal for a moratorium on large-scale AI datacentre construction, citing economic concerns over the energy and water-intensive projects.
ChangXin Memory Technologies (CXMT), China's leading memory chip producer, filed a lawsuit against the Pentagon on Friday challenging its designation as a Chinese military company. CXMT denies military affiliation and claims its chips are for civilian use only.
A US appeals court has ruled against Kalshi's attempt to block Nevada from regulating its platform, determining that the prediction market platform cannot escape state gambling laws by rebranding bets as 'swaps.'
Vijay Pande has left Andreessen Horowitz's $4 billion biotech practice to launch VZVC, a smaller AI-native venture fund focused on transforming drug discovery through engineering principles rather than traditional discovery methods.