# Rate Case The proceeding before a [[Public Utility Commission]] in which a regulated utility proposes new rates and the commission, after adversarial hearings, decides what the utility may charge. It is the central ritual of American utility regulation and the reason electricity prices are a legal outcome rather than a market one. ## The arithmetic A rate case resolves into one equation and one division. **The revenue requirement** — the total the utility is permitted to collect in a year: > revenue requirement = operating expenses + depreciation + taxes + (rate base × allowed rate of return) **The rate base** is the depreciated value of the capital the utility has invested and the commission has found prudent. **The allowed rate of return** is the percentage the commission lets the utility earn on that capital, and the equity portion of it is the utility's profit. The structural consequence follows immediately and explains almost everything about utility behavior: **a regulated utility earns a return on capital it invests, not on power it sells.** Building things is how it makes money. Operating expenses are pass-through. This is the Averch-Johnson incentive, and it means a utility greeting a new data center with enthusiasm is behaving exactly as the rate formula tells it to. **The division** is [[Cost Allocation]]: the revenue requirement is split among customer classes — residential, commercial, industrial, and now data centers — and then turned into rate designs within each class. The allocation study that performs this split is contested, methodologically underdetermined, and where the real money moves. ## Who is in the room The utility, with a full-time regulatory department and outside counsel. Large industrial customers, who intervene because a percentage point of allocation is worth millions. Sometimes a state consumer advocate, usually with a fraction of the utility's resources. Increasingly, hyperscalers with their own counsel. Residential customers are, in the aggregate, the largest single party affected and are represented — when they are represented — by one office with a small staff. The asymmetry is not a scandal; it is the ordinary condition, and it is why outcomes drift in a predictable direction over decades. See [[Regulatory Capture]]. ## Why the AI buildout strains the format Three features of a rate case fit a growing but stable system and fit the 2020s badly: **It is backward-looking.** Rates are set from a historical or near-term test year. A load that will triple in four years is a forecast problem, and the format handles forecasts by litigating them. **It assumes cost causation can be traced.** The allocation methods were built for a world where classes grew together. A single customer requiring dedicated new generation breaks the premise that shared costs are shared benefits. **It is slow and it is public.** A general rate case runs the better part of a year. That is why so much 2025–2026 data-center regulation arrived through other instruments — tariff frameworks, settlements, executive orders, and, in TVA's case, a single board vote — rather than through general rate cases. ## The minimum-take answer The device that state commissions converged on is not a new rate level but a new contract term: a floor on billed demand whether or not power is consumed, plus a long commitment and collateral. Ohio's AEP settlement used 85% of contracted capacity; Virginia's GS-5 class 85% of distribution and transmission demand and 60% of generation; Indiana at least 80% with a minimum 12-year term. Pennsylvania's PUC framework required deposits sized to cover the infrastructure if the project never materializes. All of these are answers to the same question, and it is a question about risk rather than price: if the grid is built for a customer that then disappears, who holds the asset? See [[Large Load Tariff]]. ## See Also - [[Public Utility Commission]] — the body that decides - [[Cost Allocation]] — the division inside the revenue requirement - [[Large Load Tariff]] — the current instrument of choice