Amazon plans highly polluting power plant for Texas data center
Amazon is reportedly planning to build what could become the most polluting power plant in U.S. history in West Texas to power a new data center. This move contradicts the company's 2019 Climate Pledge for carbon neutrality, with a spokesperson stating that "the world looks different now" due to the energy demands of AI. The natural gas power plant is projected to emit 33 million tons of CO2 annually.
Amazon, a co-founder of The Climate Pledge in 2019 which aimed for carbon neutrality by 2040, is now reportedly moving forward with plans to construct the most polluting power plant in U.S. history. This natural gas-fired facility in a remote part of Pecos County, West Texas, is intended to power a massive Amazon data center, driven by the significant electricity demands of commercial AI and large language models.
The proposed power plant, featuring 35 giant natural gas turbines, is projected to generate up to 7.65 gigawatts and emit a maximum of 33 million tons of carbon dioxide annually. This output would more than double the emissions of the current most polluting power plant in the U.S., Alabama's James H. Miller Jr. Electric Generating Plant. Despite the dramatic increase in emissions, Amazon spokesperson Margaret Callahan stated that the company's commitment to carbon neutrality "hasn't changed," citing that "the world looks different now" as a rationale. This sentiment is echoed by other major tech companies like Microsoft, who are also reconsidering sustainability pledges due to AI's energy consumption.
The Pecos County site, located near the heart of Texas's natural gas industry and away from population centers, is strategically chosen to minimize public resistance. Amazon claims the on-site generation "won't raise electricity costs for Texas families." The article highlights that on-site gas plants are favored by the data center industry for their speed and lower initial cost compared to connecting to existing, often strained, power grids. Beyond CO2 emissions, the article raises concerns about massive waste heat, potential localized temperature increases, and the rapid wear of these turbines under AI's demands.