The $800 Billion Problem Nobody's Talking About
The short version: Good Revenue's breakdown of AI data centers' power and water appetite is a useful primer on the math behind the buildout — and the $800B revenue gap it's betting against.
Our Take
Credit to Good Revenue for laying out the numbers plainly: AI-specific data centers aren't your old enterprise server rooms — they run nonstop, pack in high-density GPUs, and need cooling systems that guzzle water most communities never signed up to provide. When the video pegs AI's share of U.S. power demand climbing from 15% toward 35-50% by 2030, that's not an abstract industry trend — it's your utility's next rate case.
The part that should worry every ratepayer isn't just the electricity math, it's the financing logic underneath it. Over a trillion dollars committed, an $800 billion revenue gap still unaccounted for, and the video's own nod to "AI bubble" risk — that's the kind of setup where the buildout gets financed today on grid upgrades and tax breaks, and the bill for water, land, and power gets settled by local communities regardless of whether the revenue ever shows up. Virginia hosting 666 sites isn't an accident; it's what happens when tax incentives and permitting stay generous long after the math gets shaky.
We track exactly these deals — who approved the substation, who signed the water agreement, who's on the hook if the projected revenue never materializes. Check our facility map to see what's landing near you, and if your utility or county is quietly fast-tracking approvals, our take-action page has what you need to start asking questions.
This is GridWatch the USA’s original commentary. The video above is the work of Good Revenue, published on YouTube — full credit to the creator.