Years ago as a young-ish analyst I learned an important thing about building meaningful information products: they require both a supply push and a demand pull in order to really work. If there is supply and no demand, no one will buy; if there is demand and no supply, there is nothing to sell.
US electricity demand is growing at a rate not seen in three decades. You know this, of course, if you read this blog, but you also know it from any given week of coverage of the US energy sector. News flow has conditioned us to expect that much of this demand is coming from data centers, which are certainly a highly concentrated, highly visible manifestation of electricity consumption in many state power markets.
But, it is worth remembering that data centers are not the only source of growing electricity demand in the US — far from it. We are in the midst of a manufacturing boom which has also not been seen in decades, with more than $230 billion dollars of manufacturing construction spending at an annualized rate as of March 2025. Current trade policy uncertainty means that construction spending has pulled back slightly from its peak a few months earlier, but investment is still very elevated from levels of even five years ago.
Energizing that $230 billion of manufacturing means providing it with thermal energy inputs, such as natural gas or even coal. It also means providing facilities with electricity — and a lot of it. But, just how much, per facility, is not something that companies necessarily publicize or advertise, and in some cases, it might be more akin to a trade secret than a statistic to celebrate. As with many of the data points we talk about here at Machine Readable, these are numbers that are at once very big, and all-but-invisible to most people, including even those who would really like to know.
So I was pleased this week to locate some solid data points on US industrial electricity consumption, and from a high-profile project: the BlueOval SK battery park, a manufacturing joint venture between Ford Motor Company and Korea’s SK. The plant website describes its cost ($5.6 billion), its physical footprint (1,500 acres) and its battery manufacturing potential (80 gigawatt-hours a year of production capacity). Unfortunately, it does not disclose the energy required for this very large project.
However, the Kentucky Public Service Commission does! In a response to a request for information from the Sierra Club, Louisville Gas and Electric and the Kentucky Utilities Company stated that the BlueOval SK plant will have a 260 megawatt load in summer, and a 225 megawatt load in winter, built out in two phases of 140 and 120 megawatts.
Importantly, the utility response also provides some useful contingency: the second phase will be dependent on the growth of the US electric vehicle market (and also, presumably, on the demand for the specific products manufactured at the plant).
For context, 260 megawatts of load is the equivalent of a large (though not extremely large) data center today, and prior to the current data center expansions, would be one of the biggest sources of new demand being built in Kentucky, or its neighboring states.
Of course, there are data centers in play too. In February, the Kentucky utilities provided their latest resource assessment, in which they described “about 2,000 megawatts of economic development load” that will be added to their system by 2032. Of that, all but the BlueOval SK plant will be data centers.
But the BlueOval SK plant’s quarter-gigawatt of power demand is not nothing. It is a reminder that it’s not just data centers that will determine the shape of US power demand. And, it's also a reminder that it is not just the information contained within press releases that will shape it, either: that shape will come from dockets and filings, from intervenor testimony, from requests for information, and from proceedings as obscure as they are important.
This is the approach we’ve taken with sourcing the data that powers our recently announced US gas power plant tracker, and based on the interest we received (thank you to those of you who have already reached out; email dataproducts@halcyon.io if you’d like to learn more), it’s resonating with many energy professionals. If you have ideas about other types of data we can liberate — or what we can build with that data — get in touch.
Comments or questions? We’d love to hear from you - sayhi@halcyon.io, or find us on LinkedIn and Twitter