Large Load Tariff Revenue: Reading RMI's 300 MW Model
RMI modeled large load tariff revenue for one 300 MW customer under eight utility tariffs in a Colorado case. What the model shows and what it does not.
How much would a 300 MW data center pay a utility in a year? In a Colorado rate case, RMI answered that question for eight different utility tariffs, and the answers ranged from about $145 million to about $303 million. That spread is the headline people will repeat. It is also the easiest number in the filing to misread.
The large load tariff revenue comparison was filed August 19, 2026 in Colorado PUC Proceeding 26AL-0137E, Public Service Company of Colorado’s large load tariff case, as Hearing Exhibit 701, Attachment ASMI-2. The Office of Utility Consumer Advocate filed it, and the attachment says RMI did the modeling at the Advocate’s invitation. It responds to the Commission’s request in Decision No. C26-0430-I for an analysis of the revenues a peer group of large load tariffs would collect “from a generic large new load customer with a demand of approximately 300 MW operating at a 100 percent load factor.”
What RMI’s Large Load Tariff Revenue Model Assumes
The scenario is deliberately artificial, and every assumption shapes the totals:
- One customer, about 300 MW, running flat out all year. A 100 percent load factor means no hour below peak. Real loads ramp, so a real bill would differ.
- 100 percent power factor, so charges based on kVAR are left out.
- Transmission-voltage service where the tariff offers it, with the most similar class used as a proxy where it does not.
- Standard riders included; optional ones excluded. Clean energy and interruptible-service riders are out, as is sales tax.
- One rider rate applied to the whole year. For riders that change during the year, such as fuel and purchased power, RMI applied the most recent period’s rate to all twelve months. The fuel cost riders used are from June, July or August 2026, except FPL, which updates its factors annually.
RMI started from the utility peer group in Ann E. Bulkley’s testimony in Proceeding No. 25AL-0494E, then dropped utilities without a PUC-approved large load tariff that has published energy and demand rates for a 300 MW customer, and utilities without rate-based generation. Where a parent company had several qualifying subsidiaries, RMI analyzed one representative tariff.
The Eight Results
| Tariff | Modeled annual revenue | From demand and energy charges | From other riders and adjustments |
|---|---|---|---|
| PSCo Schedule TL (proposed) | $144,802,992 | $50,043,120 | $94,759,872 |
| Evergy Kansas LLPS | $184,797,008 | $103,626,800 | $81,170,208 |
| Ameren Missouri LLCS | $198,475,393 | $169,471,417 | $29,003,976 |
| Florida Power & Light LLCS-1 | $204,247,188 | $114,536,268 | $89,710,920 |
| Consumers Energy LGS GPD | $207,931,204 | $159,630,041 | $48,301,164 |
| AEP Kentucky Power IGS | $226,754,474 | $173,565,036 | $53,189,438 |
| Dominion GS-5 | $241,115,039 | $84,760,883 | $156,354,156 |
| Portland General Schedule 96 | $303,140,136 | $107,735,520 | $195,404,616 |
Source: RMI, Hearing Exhibit 701, Attachment ASMI-2, Summary of Results, Colorado PUC Proceeding 26AL-0137E. Rows reordered from lowest to highest total.
Look at the right-hand column before the totals. For PSCo’s proposed Schedule TL, Dominion and Portland General, riders and adjustments make up roughly two thirds of modeled revenue. For Ameren Missouri they are about 15 percent. A tariff that looks expensive may simply pass through more fuel and other costs in riders, and those riders were frozen at one summer month for this exercise.
PSCo’s Number Moved Before the Hearing
The PSCo row uses the Schedule TL rates as originally proposed, which were based on PSCo’s Schedule TG. RMI notes that Schedule TG rates changed effective July 1, 2026. Recalculated with the new TG values, the modeled Schedule TL total rises from $144,802,992 to $152,701,246, about 5 percent. The generation demand charges fell 18 percent in that update while the energy charge rose, so base charges went up about 15 percent and rider charges barely moved.
That is two different PSCo figures in one exhibit, and neither is an approved rate. As of a check of the Colorado PUC e-filing docket on November 8, 2026, Proceeding 26AL-0137E was listed as suspended. Hearing exhibits were filed in late October, and the docket list showed no Commission decision on the merits of Schedule TL.
Why Large Load Tariff Revenue Is Not an Equipment Budget
It is tempting to treat $145 million to $303 million a year as money a utility will spend on substations and transformers. The filing does not support that.
- Revenue is not capital. These are modeled collections under a tariff, not profit, not net benefit to other ratepayers and not a capital plan.
- Much of it is riders. Where riders and adjustments are two thirds of the total, most of the modeled revenue is collected outside base demand and energy charges, including fuel cost adjustments. RMI’s breakdown does not say what any rider funds.
- The customer is hypothetical. RMI modeled a generic 300 MW load because the Commission asked for one. Scaling it to a gigawatt or to a real project’s load shape is a new calculation, not a reading of this one.
- The rates are a snapshot. Summer rider rates applied to a full year will not match any actual year.
For a municipal utility, cooperative or distributor, the useful part of this exhibit is the method. When a utility or developer cites what a large customer “will pay,” ask for the load factor, the service voltage, which riders are in, which month’s rider rates were used and whether the tariff is approved. Then look separately for the approved project scope and equipment orders before reading the revenue as a procurement signal. Our pipeline vs contracted explainer applies the same discipline to load forecasts, and our read on Kentucky Power’s large load contract terms shows what a signed agreement adds.
Related Reading
- Pennsylvania Large Load Tariff: ‘But For’ Cost Allocation
- Large Load Contract Terms: Kentucky Power’s 1 GW Deal
- Data Center Load Pipeline vs Contracted: How to Discount It
All revenue figures, assumptions and the Schedule TG recalculation are from RMI’s Hearing Exhibit 701, Attachment ASMI-2, filed August 19, 2026 in Colorado PUC Proceeding 26AL-0137E, opened November 8, 2026. They are RMI’s modeled outputs, not utility bills. Proceeding status is as of the Colorado PUC e-filing docket list on November 8, 2026.
Tariff filings like this one move every week. The Feeder is our free weekly digest of the regulatory and supply signals that change what utility buyers specify.
Frequently Asked Questions
What did RMI model in the Colorado large load tariff case?
RMI calculated the annual revenue each of eight utility large load tariffs would collect from one generic customer of about 300 MW running at a 100 percent load factor, as the Colorado PUC requested in Decision No. C26-0430-I. RMI also assumed a 100 percent power factor. The comparison was filed August 19, 2026 as Hearing Exhibit 701, Attachment ASMI-2, in Proceeding 26AL-0137E.
Has Colorado approved Public Service Company's Schedule TL?
Not as of a check of the Colorado PUC e-filing docket on November 8, 2026. The proceeding was listed as suspended, with hearing exhibits filed in late October and no Commission decision on the merits of the tariff in the docket list. RMI's analysis compares PSCo's proposed Schedule TL rates.
Can tariff revenue figures be used as a utility's equipment budget?
No. RMI's figures are modeled annual revenue under a controlled scenario. Most of some totals are riders such as fuel cost recovery, and revenue is not profit, net benefit to other customers or a capital plan. Equipment spending needs its own evidence: approved projects, scope and orders.
Free RFQ templates. Weekly digest. Yours.
Free Member tier. Pick your topics. Get a weekly digest filtered to what you actually buy.