“An Extremely Energy-Needy AI Industry Turns to Natural Gas Out of Necessity”, By THOMAS J SHEPSTONE
“The power needs of artificial intelligence and cloud computing are growing so large individual data center campuses could soon use more electricity than some cities, and even entire U.S. states.”
An Extremely Energy-Needy AI Industry Turns to Natural Gas Out of Necessity
NOV 25
The AI revolution is here.
Tech companies are scrambling for energy and CNBC has put out a fascinating article summing up the situation. It makes several salient points:
The power needs of artificial intelligence and cloud computing are growing so large that individual data center campuses could soon use more electricity than some cities, and even entire U.S. states.
Renewable energy alone won’t be sufficient to meet their power needs. Natural gas will have to play a role.
Data centers are now at a scale where they have started “tapping out against the existing utility infrastructure.”
Available land in this country that’s industrially zoned and can fit the data center use case is becoming more and more constrained.
Power demand is expected to surge, while supply is falling due to the retirement of coal- and some natural gas-powered plants.One data center operating at 85% capacity over the course of a year will consume nearly as much energy as 710,000 U.S. households or 1.8 million people.
A data center campus that size would use more power in one year than retail electric sales in Alaska, Rhode Island, or Vermont.
A gigawatt-size data center campus running at even the lower end of peak demand is still roughly comparable to about 330,000 households, or a city of more than 800,000 people — about the population of San Francisco.
The average size of individual data centers operated by the major tech companies is currently around 40 megawatts, but a growing pipeline of campuses of 250 megawatts or more is coming, according to data from the Boston Consulting Group.
The U.S. is expected see a growing number of data center campuses of 500 megawatts or more, equivalent to half a gigawatt, in the 2030s through mid-2040s.
The average size of the data centers is increasing at a rapid pace.
A 1,000-acre campus in Abilene is expected to open in the first quarter of 2025 with 250 megawatts of power that will ramp up to 1.2 gigawatts in 2026.
The minimum power requirement for certain data center customers is now a gigawatt, and future plans involve scaling them up to between three and five gigawatts.
Developers prefer carbon-free renewable energy, but they also see solar and wind alone as unable to meet current demand due to their reliance on changing weather conditions.
The major technology companies are some of the largest purchasers of renewable power in the U.S., but they are increasingly turning to nuclear in search of more reliable sources of electricity.
Building new nuclear reactors is expensive and fraught with delays.
In the short run, natural as it is the main, short-term power source providing the reliability these facilities require.
Investments could be made in new gas generation that adds carbon capture and battery storage technology over time to mitigate the environmental impact.
The industry hopes gas demand will taper off as renewables expand, battery storage costs come down and AI helps data centers operate more efficiently, but in the near term, there’s no question that data center expansion is disrupting technology companies’ emissions targets.
As the last several points indicate, natural gas will inevitably be called upon to power the AI revolution. All but the true believers are already acknowledging solar and wind will never get it done, and the risks associated with unreliability are too great anyway. Nuclear is coming, but it's still costly, overregulated, and unlikely to come at the speed desired. That leaves exactly one option, and it's the most economical. It's natural gas, and we have humongous amounts of it.
Yes, however much Big Tech and states such as New York don't want to admit it, their futures are inextricably linked to natural gas. It's undeniable, and more so every single day.
#NaturalGas #Nuclear #AI #Climate #PowerDemands #Utilities #Reliability