Natural Gas To Revolutionize AI Data Centers
The largest challenge of building AI data centers in the US is not demand. Companies, from Microsoft to OpenAI, know they will need massive amounts of electricity to power them. This means an improved grid and new sources of energy. Typically, these sources are described as a combination of solar, wind, and nuclear energy. Natural gas is usually left off the list until now.
According to Axios, Chevron New Energies president Jeff Gustavson said, “There’s no reason the U.S., with its large, abundant natural resources, natural gas in particular, can’t win this [AI] race.” He added that the new capacity Chevron could bring online would also service residential and legacy business customers.
If Chevron is only partially successful, it will start to solve the colossal AI challenge. AI server farms could account for double digits of total US electricity production in two to three years. AI development would need to slow, or residential customers would need to be metered. It is also likely that residential rates would rise. At some point, the conflicting needs mean this becomes a political football.
Back To Fossil Fuels?
The other alternative is a return to the use of fossil fuels, particularly cheap coal. This undermines US goal warming goals, which are already in tatters.
Large tech companies have begun to develop their own AI energy sources. Some of these will run outside major grides. One energy source solution is small modular reactors, but they may not be fully tested for half a decade.
If natural gas works as an electricity generation source, AI companies will be winners, as will be the case with Chevron.
More from ClimateCrisis 247
- Your Electricity Bill Is About To Jump 50%
- Is Meta America’s Greenest Company?
- America’s Electricity Famine
- AI Energy Use Is Dirty
