Oregon moved first.
The Oregon Public Utility Commission reportedly adopted Schedule 96 on May 12, establishing a dedicated customer class for large-scale data centers in the Portland General Electric service area, according to Renewable Energy World’s coverage. If Oregon is indeed the first state utility commission to create a billing class specifically designed for AI data center operators, this is a regulatory template, not just a local tariff.
The framework’s core mechanism is a minimum capacity payment. Data centers classified under Schedule 96 reportedly pay for at least 85% of their subscribed energy capacity regardless of actual consumption. That’s a departure from standard utility billing, where customers pay for what they use. The 85% floor ensures the grid infrastructure built to serve peak AI workloads gets funded even during periods of lower utilization — a direct response to the lumpy, high-peak demand profile of large-scale AI inference and training operations.
Oregon Data Center Billing: Before and After Schedule 96
A “Peak Growth Modifier” mechanism reportedly allocates grid infrastructure costs to the fastest-growing customer classes. In practice, that means data center operators — not residential or standard commercial ratepayers — bear the cost of grid upgrades their growth requires. Schedule 96 reportedly also includes provisions for special clean energy contracts, allowing data centers to directly support new clean energy infrastructure as part of their service agreement.
The math on the 85% floor matters. A data center subscribing to 100 megawatts of capacity pays for 85 megawatts minimum in every billing period, regardless of actual draw. For operators running variable AI workloads — inference demand spikes during peak usage windows, drops during off-peak — that minimum billing creates a predictable fixed cost that must be absorbed into the economics of every deployment decision in the PGE service area.
The real story is the cost-allocation logic, not just the tariff. The national policy conversation around AI infrastructure and grid capacity has centered largely on whether data centers should pay for the grid upgrades they require, or whether those costs get socialized across all ratepayers. Congress held a subcommittee hearing on AI power demand, permitting reform, and ratepayer protection earlier this year. Oregon’s Schedule 96 is the state-level answer to that question: a regulatory mechanism that keeps the cost with the customer driving the demand.
Who This Affects
What to watch
The OPUC official order and docket filing — the primary source that would confirm Schedule 96’s exact applicability threshold, effective date, and mechanism details. Other state utility commissions with high data center concentration (Virginia, Texas, Georgia, Arizona) should be monitored for similar proceedings. The White House ratepayer protection framework and hyperscaler capital infrastructure dynamics provide the federal backdrop against which Oregon’s state action sits.
Don’t bet on this staying a local Oregon story. The policy logic is exportable. Any state utility commission facing the same grid capacity pressure has the same regulatory toolkit available. Schedule 96 gives those commissions a named framework to cite.