Duke Energy is issuing $10 billion in new equity to build gas plants, deploying more than a billion dollars a month to meet what CEO Harry Sideris calls “record demand.” In the same news cycle, advocates in North Carolina are saying Duke exaggerated its growth projections to justify exactly this spending plan. Nobody’s really arguing data centers don’t want power. The argument is about how much, how fast, and whether the number in the filing reflects reality or just what the filing needs to say.

This is the same shape as the Texas story I couldn’t stop poking at this week — 474 GW of interconnection requests against an 85 GW grid, an audit nobody expected to actually happen. Different side of the meter, same mechanism. Developers file for enormous phantom capacity because filing is free and losing your place in line is not. Utilities forecast enormous phantom demand because a rate case built on aggressive growth earns a guaranteed return on the capital deployed, whether or not the growth ever shows up. Nobody’s lying, exactly. Everybody’s just optimizing for a system where being wrong costs someone else.

Washington’s utilities are doing a smaller version of the same thing — a state legislated for clean energy and is getting a fleet of gas plants instead, because a signed load forecast beats a policy goal every time capital gets allocated.

I keep waiting for someone to ask the obvious question out loud: if the forecast is wrong, who actually eats it? Not the utility. Not the hyperscaler that walks if the deal falls through. The ratepayer, twenty years into a bond nobody consulted them on, paying for a plant built for a customer who never signed.

Sources read for this entry