• 2 min read
AI datacenters are driving a gas power boom
The Register’s latest Kettle podcast looks at how AI datacenter expansion is colliding with climate goals, local air quality, and noise complaints.
Image: The Register
Wildfire smoke over the US Great Lakes set the stage for this week’s episode of The Register’s Kettle podcast, which focused on a sharper question: whether the AI datacenter buildout is accelerating the climate and public-health problems people are already living through.
Host Brandon Vigliarolo was joined by The Register editor-in-chief Matt Rosoff and CNBC technology and climate reporter Lora Kolodny, whose recent reporting examined xAI’s Colossus datacenter operations in Memphis, Tennessee.
Kolodny said xAI, Elon Musk’s AI company now part of SpaceX, moved into a former Electrolux factory in early 2024 and built out the site in 122 days. She said the project required more power than local utilities could immediately provide, leading the company to deploy large natural gas or methane-burning turbines while it worked toward grid interconnection and additional support from Tesla Megapack battery systems and solar.
Recommended reading
Infinity lands $15M to make AI chips easier to use
That temporary solution, she said, came with significant consequences: emissions, PM2.5 particulate matter, NOx, and persistent noise for nearby residents.
“It sounds like living right next door to an airport; jet engines, very loud.”
According to the discussion, public records indicate up to 59 turbines were being used, even though residents say they were given little visibility into the project early on. Kolodny said thermal-camera footage showed 30-plus running at once at one site. A lawsuit backed by the NAACP argues the developments in historically majority-Black and lower-wealth neighborhoods raise environmental justice concerns and did not go through the proper Clean Air Act permitting process.
The panel also pointed to a stalled promise to help build a gray water recycling facility for Memphis Light Gas and Water, intended to reduce pressure on local water supplies.
How AI power demand changed the conversation
Rosoff noted how dramatically datacenter scale has shifted. Facilities that once consumed 50 to 100 megawatts were considered huge; now, he said, projects discussed in places like Texas and Louisiana are measured in gigawatts. He cited Colossus at around a gigawatt, alongside reports of two-gigawatt and even two- to four-gigawatt builds.
The broader criticism from all three speakers was that major tech companies have talked up nuclear and clean-energy plans while leaning heavily on gas in the near term. Vigliarolo pointed to a 20-year deal in Texas tied to a datacenter intended to run entirely on gas turbines. Rosoff said companies including Amazon, Microsoft, Meta, and Google have moved quickly away from the environmental commitments they emphasized just a few years ago.
The podcast’s conclusion was blunt: with political pressure mounting to “win” on AI, climate promises are being pushed behind immediate demand for power — and communities near these massive facilities are absorbing the cost first.
Enterprise Editor
Marcus follows the money. He covers enterprise software, cloud architecture, and the tectonic shifts in Big Tech strategy. He translates dense earnings calls and complex M&A activity into actionable insights about where the industry is actually heading. If a tech giant makes a silent pivot, Marcus is usually the first to notice.
via The Register
