Secret AI Energy Deals?! How Big Tech Offloads Costs to You And Keeps Profits

Our Take

Good Revenue lays out something GridWatch readers already sense in their bills: the AI buildout isn't happening in a vacuum, it's happening on the grid you also depend on, and the pricing arrangements between utilities and hyperscalers are often structured so regulators — let alone the public — can't easily see what's actually being charged. When Microsoft, Amazon, Google, and OpenAI each cut different kinds of power deals, the common thread is the same: speed and secrecy get prioritized over transparency, and the interconnection queues and grid stress that follow get treated as somebody else's problem.

That "somebody else" is usually the residential ratepayer sitting behind the data center in line for capacity, or the household whose utility now has to justify emergency generation buildout to a regulator that's negotiating half-blind. PJM's backlog isn't an abstract engineering footnote — it's the difference between a stable rate case and a surprise surcharge showing up next winter. Credit to Good Revenue for tracing the mechanics of this rather than just gesturing at "AI is expensive."

We'd add the piece that often gets left out of these energy-only breakdowns: the same negotiating opacity shows up in water allocations, land-use approvals, and tax abatements that local governments sign off on with just as little public scrutiny. If a facility near you is part of this pattern, check our facility map, and if your utility or county is about to vote on something that affects your rates or your water table, see our take-action resources before the deal is already done.