AI data centers’ thirst for energy is so severe it can wreck gear on and off site, including the dirty workhorses powering the artificial intelligence boom — thermal turbine generators.
The problem begins with fluctuating computational loads that gulp down energy for artificial intelligence operations, such as training machine learning models. Changes in normal grid demand tend to be gradual.
Gas turbines have been at the heart of some of the largest AI hyperscaler data center energy deals, including names like Microsoft (MSFT), Meta Platforms (META), Chevron (CVX) and Amazon (AMZN). Life expectancy for the complex, pricey equipment often runs to a decade or more.
But they are designed to run steadily for days, weeks or months at a time. And, by its nature, AI training demand comes in short-term surges. This cycling on and off appears to be sharply reducing the life of the turbines, leading to downtime and costly repairs or, worse yet, the need for new equipment as the backlog for new turbine deliveries runs to the end of the decade.
The AI power loads “have the potential to change their demand almost instantly,” FERC Chair Laura Swett said. “This rapid fluctuation causes voltage stability issues that threaten grid reliability.”
The stability risks posed by AI and other sources of large loads are profound enough that Tesla is funding research and America’s grid watchdogs have begun to clamp down.
The Federal Energy Regulatory Commission on July 16 ordered its nonprofit counterpart, the North American Electric Reliability Corporation, to set enforceable reliability standards for volatile energy pits such as data centers and crypto mining operations. Or else.
Read the full article on Investors.com
