Data centers powering AI have become contentious. Are concerns warranted?
The topic:
AirTalk
Background: Data centers as a concept aren’t new, and the internet has long relied on buildings built specially to house networking and data storage equipment. The advent of commercial cloud computing in the mid-2000s allowed data centers to not just store data, but also do things with the data with large, rentable computers.
What’s new: AI training and use has created new, unprecedented demand for data centers with large amounts of computer power. Power monitoring service Cleanview estimates that between 2016 and 2025, the average data center size more than doubled from 41 megawatts to 91 megawatts, and that the average could exceed 600 megawatts by 2030. For context, estimates vary widely, but one megawatt can be expected to power between 670 and 1,000 homes, depending on the time of day.
Public opinion: Recent public opinion polling from Gallup has shown that roughly seven in ten Americans oppose local construction of AI data centers. Water use, electricity use, as well as concerns around noise and land use were cited as reasons from respondents for being against these centers.
Yes, but: Actual water use from data centers in California is largely unknown, and public opinion polling on pre-AI, lower megawatt datacenters is scant.
Local ban: Monterey Park made their city the first in the nation to permanently ban new data centers, with more than 86% voting yes on the ballot measure. Meanwhile, a private company is pitching small data centers at fairgrounds up and down the state, including the Antelope Valley Fair in Lancaster…