Alliant is pushing a custom, heavily redacted power contract for Meta’s Beaver Dam data center, with unknown load size, term length, and exit-fee math, which makes it hard to judge who really pays for the infrastructure.
The utility claims Meta will fully cover its costs and even pay system-offset charges, but ratepayer and environmental groups warn there’s a real risk existing customers end up subsidizing grid upgrades for a single hyperscale tenant.
Critics want standardized data center tariffs instead of bespoke deals, plus tougher early-termination penalties and more transparency on load forecasts, because AI data center demand locks in long-lived capacity and fuel decisions.
We Energies is trying a more generic data center rate with vocal backing from Microsoft and OpenAI, yet local customers still worry about cost shifts and increased dependence on fossil generation to feed GPU-heavy campuses.
Xcel is more explicit about the scale, talking about 2 GW already signed and 6 GW of data center capacity by 2027, and planning large-load tariffs across multiple states to align revenue, buildouts, and AI-driven demand.
The throughline is clear: AI data centers are now big enough to reshape utility planning, but the financial and energy mix impacts are being negotiated in partially closed-door processes that communities can’t fully see.
Worth a read if you care how GPU campus economics, grid buildouts, and rate design are quietly locking in the next decade of AI infrastructure.
Source: Alliant Energy proposes custom electric rates for Meta’s data center campus – WPR
