AI's True Energy Price-You Pay The Bill
The short version: PJM's record $16.4B capacity auction shows AI data centers driving $6.3B in added costs, a bill that lands squarely on ratepayers across 13 states.
Our Take
The Shortman On Tech lays out the math plainly: PJM's latest capacity auction just posted its worst result ever, and the grid's own market monitor pins $6.3 billion of that on data center demand. That's not a projection or a worst-case scenario — it's the price tag for 2028/2029 delivery, already locked in. Four straight auctions of this trend line means one thing: the AI buildout isn't paying for its own grid impact, and the gap is getting billed to 67 million people who never asked for a seat at the table.
What gets lost in the megawatt numbers is who actually signs off on these contracts and interconnection queues. Utilities and regulators are approving gigawatt-scale hookups for AI training campuses on timelines that outrun any real generation buildout, then letting capacity auctions socialize the cost across every ratepayer in the PJM footprint — data center customer or not. That's the real story here: a transfer of financial risk from hyperscalers to households, dressed up as market mechanics.
Credit to The Shortman On Tech for tracing the auction data back to its source instead of just repeating the topline number. If you want to see which facilities in your state are drawing down this same grid capacity, check our facility map, and if you're looking for ways to push back on these deals before they're finalized, our take-action resources are a good place to start.
This is GridWatch the USA’s original commentary. The video above is the work of The Shortman On Tech, published on YouTube — full credit to the creator.