Welcome to Locally Sourced, a biweekly Covering Climate Now newsletter for journalists working to localize the climate story. Share this newsletter with colleagues and journalism students interested in localizing the climate story.
Story Spark: AI and Data Centers
AI is everywhere these days — with chatbots providing helpful answers to questions about climate change (and sometimes promoting false climate denial claims), helping scientists track icebergs, taking your drive-through orders, analyzing medical scans, and powering more accurate weather forecasts. AI’s strongest supporters say that its possibilities are endless, with some believing that it could even help save the planet. The reality is that AI-driven technology, and the vast amounts of energy powering it, are currently doing the opposite.
Earlier this year, ChatGPT became the fifth-most visited website in the world. As more of the general public, industries, and journalists use AI tools, more data centers powering this technology must be added. Over the past decade, the number of data centers has doubled with the US now having over 5,400 — 10 times more than any other country in the world. These data centers consume incredible amounts of energy, especially the ones used for AI’s power-hungry activities. It’s estimated that more than half of the electricity going to data centers will be used for AI by 2028 with AI consuming as much electricity annually as 22% of all US households.
It’s not just the sheer amount of energy, but also where it’s sourced. Given that AI data centers require power 24/7, they rely primarily on electricity produced using fossil fuels instead of cleaner, but more intermittent, renewable sources such as solar and wind. Responding to AI’s increasing energy demands, some areas are keeping coal-powered electricity plants online, undermining clean energy goals, while some big tech companies are considering bringing old nuclear power plants back online to fuel these energy-hungry data centers.
Researchers say that it’s difficult to fully predict how energy demands will change as AI becomes more prevalent. To get started reporting this story, learn where hundreds of new data centers are being planned across the US, particularly in the Mid-Atlantic, and investigate who is paying for this infrastructure — as subsidies paid to data centers may raise the utility rates paid by local consumers. Also explore how communities, often disadvantaged and especially vulnerable, near data centers face health and economic challenges as AI infrastructure raises pollution levels and chugs fresh water to cool the supercomputers powering your chat prompts.
Expert Tips
Dan Gearino, an Inside Climate News reporter, offers tips to help journalists report on the electricity needs of AI and data centers. In addition to covering the business and policy of renewable energy and utilities at Inside Climate News, he is also the main author of their Inside Clean Energy newsletter.
Identify the size and type(s) of data centers in your area. There are various types of data centers and it’s important to specify which types are opening in your market. Is a data center mainly focusing on cloud computing, a super-size AI data center, or some other type? This is a simple question that a developer or local economic development official should be able to answer.
The size of a data center, and the aggregate size of all the data centers in a market, have big implications for power and water use. Real estate companies such as CBRE and Cushman & Wakefield have free reports you can use to provide estimates of current and projected data center capacity in your area.
Use comparisons, not figures, to convey the size of data centers. Stats for gigawatts of demand aren’t meaningful to most of your audience. Contact your local utility and ask what last year’s peak demand was for the largest metro area in the utility’s service territory, and compare the data center’s demand to the figure for the metro area.
Ask follow-up questions. The rise of this new demand for electricity can lead to numerous story ideas including: “Where is this new electricity coming from?” “How does this affect consumer rates?” “How does rising electricity demand affect air quality in our community or the state?” The urgency of these questions will only grow as data center energy demands increase.
Stories We Like
- DW News breaks down the complicated question, “Is AI a climate solution or a problem?”
- In South Memphis, a predominantly Black community battles against Musk’s xAI as nearby data centers powered by gas turbines cause a public health emergency, TIME reports.
- Ben Tracy with Climate Central explores how artificial intelligence and balloon launches may be used to create better weather forecasts.
- In Georgia, the Columbia Ledger Enquirer highlights how rapidly expanding data centers are forcing residents to leave their “forever homes.”
- In Malawi, a small-scale farmer who lost everything during Cyclone Freddy seeks support from an AI chatbot to adapt his farm for a climate changed future, the Associated Press reports.
- Not only are AI data centers bad for the climate, they’re bad for your wallet. The New York Times examines how soaring demand for electricity is driving up your utility bills.
Resources
- Research data center locations in your area on Data Center Map’s interactive website.
- Explore the findings of the latest US Data Center Energy Usage Report, which includes estimates of future energy consumption through 2028.
- Dive into the International Energy Agency’s detailed outlook report on AI use and future projections.
- Read MIT’s analysis of how much energy the AI industry uses and how much it could use as AI barrels towards billions of daily uses.
Experts
- Ann Bostrom, chair, National Academies of Science’s Roundtable on AI and Climate Change
- Yuan Yao, associate professor, Yale School of the Environment
- Vijay Gadepally, senior staff, Lincoln Laboratory Supercomputing Center at MIT
Before We Go…
The next Locally Sourced will highlight how to cover scientific articles. Have you reported on a recent scientific study that goes beyond just re-stating their findings? Send them to us at local[at]coveringclimatenow[dot]org. We’d love to consider them for the next edition of Locally Sourced and our media trainings and social platforms.
The Climate Newsroom is a free-of-cost training program from CCNow that equips local US journalists to cover climate news more effectively. For inquiries, please email Elena González at elena[at]coveringclimatenow[dot]org. Or apply here.
Want more story ideas? Check out the Locally Sourced archive for more topics to explore, including resilient agriculture, emergency alerts, climate anxiety, and more.
Know someone who might be interested in this newsletter? Forward Locally Sourced to a colleague!
Support Covering Climate Now
We’re working to help journalists worldwide improve and expand their climate coverage. Meet our staff and learn more about CCNow.
