Category Archives: FPL

Datacenter Discussion @ Columbia County Commission 2026-07-23

“Data Center Discussion” is the first item on the agenda for the Columbia County Commission meeting,

5:30 PM, Thursday, July 23, 2026, at the School Board Administrative Complex, 372 West Duval Street, Lake City, FL 32055

The board packet includes 215 pages of related materials, all from Datacenter supporters, except for one by 1000 Friends of Florida. Some of them try to conflate hyper-scale AI datacenters with search and cloud datacenters, while others argue that the growth in AI is inevitable so just go along to get along.

Hint: it’s not inevitable. It’s a bubble that will pop, more likely sooner than later.

[Datacenter Discussion @ Columbia County Commission, 5:30 PM, Thur4sday, 2026-07-23, 372 W Duval Street, Lake City, FL]
Datacenter Discussion @ Columbia County Commission, 5:30 PM, Thur4sday, 2026-07-23, 372 W Duval Street, Lake City, FL

There are several growth projection graphs, all of them wrong.

This one is clearly wrong at first glance, because its forward projection is linear: a straight line.

[Graph: Electricity Consumption at U.S. Data Centers is Expected to More Than Double by 2030 Total electricity consumption by U.S. data centers (terawatt-hours), 2026-01-01 --1000 Friends of Florida]
Graph: Electricity Consumption at U.S. Data Centers is Expected to More Than Double by 2030 Total electricity consumption by U.S. data centers (terawatt-hours), 2026-01-01 –1000 Friends of Florida
PDF

This graph is wrong because it projects a smooth curve. Sure, an exponential curve could be closer to right. See growth curves of the Internet, or of adoption of solar power, smartphones, or PCs.

But in all those historical examples, the exponential growth eventually looked more like an S-curve.

Also note that even if the amount of data produced does grow exponentially, that doesn’t mean that power use or datacenter numbers or sizes must grow like that.

The bubble will pop, sooner, or later, for one or both of two reasons:

  1. Investors will wise up that AI companies have grossly oversold the benefits of their products, and they are failing to produce profits
  2. Someone, probably in China, will leapfrog the current crude stack-computers-in-a-warehouse approach with faster, cheaper, and smaller methods.
Continue reading