MISO carves out data centers as a new load class

Data center demand is becoming its own tariff category, and MISO's latest proposal draws the line. In a filing reported by Utility Dive, the grid operator would impose reliability requirements on large loads while carving out a separate computational load classification for demand that includes at least 25 MW of information technology equipment, a distinction designed to let MISO aim requirements specifically at data centers.

Under the proposal, large loads are those above 50 GW, and a computational load is a large load whose IT equipment draws at least 25 MW—a threshold that separates computational load from the rest of the large-load category and lets MISO write data-center rules without applying them to every industrial customer on the system. The bar is high enough that the bucket should capture only the largest facilities or aggregated campus developments; MISO is not reclassifying the entire data center fleet, only the concentrated demand points that can move the grid when they ramp.

This publication has argued that grid access is becoming an asset class, and a formal category for computational load is the regulatory framework catching up to that reality. Data center demand stops being just another large customer in MISO's footprint; it becomes a distinct reliability risk the operator wants to condition separately.

The separate classification is also a more precise tool than a blanket large-load rule, letting MISO treat data centers as a distinct risk class while other large industrial loads remain on the existing framework. The structure creates a new line in the grid-access ledger: computational load now has a defined shape, a defined threshold, and requirements that can be adjusted without revisiting the entire large-load docket.

Utility Dive's report does not detail the requirements themselves, but the structure points to where the costs will land. The filing gives developers and their backers a new variable to model—queue position, power price, and now the reliability terms attached to computational load as its own class. For private capital, the proposal is a reminder that the price of grid access is set by rules like this as much as by power markets. A facility tripping the 25 MW threshold would face obligations other large loads do not, and that asymmetry would flow into the contracts and capital structures behind grid-connected data centers.