Energy demands of AI – some info!
September 28th, 2025
There’s been a LOT of talk about the energy use of data centers amongst the people, websites, and pages challenging data center proposals, there are MANY of them, at least 19, in Minnesota, in various stages from whispers to active to withdrawn:
- Rosemount – 4 active projects: Dakota East, Rich Valley East, Rosemount Industrial, and U-More
- Farmington – Farmington West and Farmington Tech Park
- Becker – Xcel – WITHDRAWN
- Chaska – no documentation yet
- Cannon Falls Industrial (the one I’m most familiar with)
- North Mankato – subject of MCEA lawsuit
- Hampton Industrial – Tech Park
- Lakeville Industrial – also subject of MCEA lawsuit
- Apple Valley – Tech Campus? Rockport Data Center?
- Eagan – thompson Reuters – ?? add to existing small one?
- Northfield – old and ? inactive
- Pine Island – Skyway
- Hermantown
- Faribault – Archer Data Center – Not AUAR – EAW?
- Monticello Tech/Monticello Industrial Park
- Monticello – Scannell Technology Park
- Others I’m missing? Let me know in comments!!
What bothers me is the histrionics, with zero specific information. The AUARs that I’ve seen say nothing about energy use, and those I’ve asked about it, particularly those spouting off about energy use, do not have any details. AAARGH. And the DOE says, “This load growth is not historically unprecedented and presents an economic opportunity.” So deep breathe???
I have found some info on tract’s page showing some in the ground and operating projects, tract data centers showing acreage and MW, which is helpful. Here’s the page, and here’s one example:
And the “project detail” page: https://www.tract.com/project/medina-technology-pa
But yesterday a friend sent an article with a couple of studies linked. First the article:
Researchers Just Found Something Extremely Alarming About AI’s Power Usage
And now the studies & reports:
And an EPRI Report (these have been so helpful over the years, particularly in coal gasification!):
And from that EPRI report, this map that is circulating:
Followed by Berkeley Lab:
From the Dept. of Energy:
- How AI Can Help Clean Energy Meet Growing Electricity Demand:
- Reports across the United States highlight a rapid increase in projected electricity demand, partly driven by growth from artificial intelligence (AI) and data center energy use. This load growth is not historically unprecedented and presents an economic opportunity. In fact, DOE has more solutions and technologies available today than it did during previous decades of load growth to help meet growing demand. However, deploying these solutions in the near-term is incumbent on our ability to permit and build clean energy infrastructure.
- AI FOR ENERGY: OPPORTUNITIES FOR A MODERN GRID AND CLEAN ENERGY ECONOMY
A Congressional report with a LOT of links:
Data Centers and Their Energy Consumption: Frequently Asked Questions
Speaking of rabbit holes, this Congressional report has this most strange paragraph:
Footnote 69: Notice of Agreement, Tex. Blockchain Council v. Dep’t of Energy, No. 6:24-cv-99 (W.D. Tex. Mar. 1, 2024), ECF No. 24
The agreement — just WOW (is it a surprise there’s so little information available?):
And back to energy use by data centers, another article and links there:
We did the math on AI’s energy footprint. Here’s the story you haven’t heard.
There’s an ICF report that I can’t get at… GOT IT!!
A couple of blurbs from our friends at Goldman Sachs:
For this one, note the little downward arrows — click to expand:
How AI Is Transforming Data Centers and Ramping Up Power Demand
And…




Leave a Reply