AI Data Centers: The True Cost of Compute
The short version: A YouTube explainer tallies AI's trillion-dollar buildout and staggering power/water appetite — the numbers are big, but the local accountability question is bigger.
Our Take
Credit to The Pleopod Times for putting real numbers on a story we cover from the ground up: over $1 trillion spent on AI infrastructure since 2023, another $745 billion expected in 2026, and global data center electricity demand potentially doubling in just four years. Those are the macro figures. What they don't capture is which substation gets upgraded on your dime, which aquifer gets drawn down for cooling, or which county commission approved a 20-year tax abatement in a closed-door session.
The video's '100,000 households' comparison for a single large AI facility is the kind of stat that should stop readers cold — because that's not an abstraction, it's a specific grid interconnection queue, a specific rate case, and often a specific community that finds out after the deal is signed. Multiply that by the hundreds of facilities now in planning or construction across the country and you get the real story: a buildout financed by hyperscalers but increasingly backstopped by ratepayers and local infrastructure.
If this kind of spending and energy math is happening near you, don't wait for the utility filing to find out. Check our facility map to see what's already been sited close to home, and visit our take-action page for ways to get into the public record before the next deal closes.
This is GridWatch the USA’s original commentary. The video above is the work of The Pleopod Times, published on YouTube — full credit to the creator.