When the business itself is taxed: five states, three new cost seams

ForgeAsset's coverage now includes Washington, Ohio, Utah, New Hampshire, and Kentucky — seven filed tariffs across the five states, bringing the library to fifty-three tariffs in thirty-three states. This wave was different from the ones before it. Every earlier state fit the model's existing cost structure. These five did not: each carried a tax the model had no honest way to express, so before any rate was seeded, the underlying engine gained three new cost seams. Nothing here is advice; it is a description of what the filed rules say and how the model now prices them.

Three costs the model used to leave out#

The engine's tax math was built around California, where the taxes that touch a Supercharger site are the ones on the owner's personal return plus a few fixed entity fees. Most states fit that shape. These five broke it in three distinct ways:

• Sales tax on installation labor. California exempts separately stated install labor, so the model taxed only the hardware. But Washington's custom-construction rule taxes the full contract — labor, materials, permits, profit — and its EV-charger labor exemption expired in mid-2025. Ohio's business-fixture doctrine reaches the same result. On a site with $600,000 of installation, at Seattle's 10.55% that is about $63,000 of sales tax the hardware-only math never saw. The engine now carries an install-labor share of the sales-tax base — zero for California, one for Washington and Ohio.
• A percent-of-revenue charging tax. Utah levies 12.5% on the retail charging sale itself, in force since 2024 — a cost that scales with revenue, not with kWh or with a fixed fee. Under the reading where the operator absorbs it, that is about 11% of gross charging revenue, the largest single revenue-side line in the researched set. Washington's business & occupation tax is the same shape at a much smaller rate. A fixed-dollar estimate misstates the moment a site's price or utilization changes, so the engine now carries a percent-of-revenue line with its own effective-from year.
• A state income tax on the entity itself. The model's parity rule leaves state income tax to the owner's return — correct, because in most states an LLC's profit passes through to the owner. New Hampshire's Business Profits Tax does not: it taxes the entity at 7.5% regardless of pass-through. Kentucky's local occupational net-profits taxes (Louisville 2.2%, Lexington 2.75%) do the same at the city level. The engine now carries a state entity income-tax line that bills a share of positive annual operating cash — loss years pay nothing, and no loss carryforward is assumed, so the estimate errs toward caution.

Each seam defaults to zero, so every state already in the library is unchanged. They switch on only where a state's own rules turn them on.

The rate corrections the verification pass caught#

As in every wave, each rate was re-derived against the currently filed sheet before seeding, and the pass moved real numbers:

• Washington produced the wave's largest correction. Puget Sound Energy's Schedule 26 had been captured at its base rate only; folding in the full stack of power-cost, carbon, conservation, and decoupling riders moved the modeled energy rate up 51%, while the decoupling schedule turned out to carry a credit that lowered the demand charge. The two moved in opposite directions and both had been missing.
• New Hampshire's default-supply rate is a trailing-twelve-month average, and the newest six-month strip had just published — a winter block that prices December and January above 18 cents per kWh. Rolling it into the average lifted the supply fold by nearly three cents.
• Utah was seeded while a rate-case settlement is still pending before the commission — the settlement moves both the price level and the split between energy and demand — so the row carries current rates with a re-derive trigger for the order, and a sales-tax rate that a cross-check corrected from 7.75% to 8.45%.
• Ohio and Kentucky each re-derived with no rate corrections at all — Ohio's whole stack reproduced to the digit, and Kentucky's roughly sixty rate cells and two fuel-adjustment series matched exactly. The catches there were on the tax side: a Kentucky school-tax increase that had been voided for defective public notice, and a settlement mechanism that was actually denied rather than approved.

Two heavy demand charges, and a nested-window rule#

Kentucky's LG&E and KU rates bill demand as the sum of three nested time-window components — a rate shape the model accepts because a charging site draws its peak inside every window, so the components simply add. Utah's Schedule 8 raised a new variant: its winter demand window is two disjoint blocks, a morning and an evening, billed as one charge on one measured maximum. That reads as a weaker version of the single-window case the model already accepts — the peak only has to land in either block — so it folds the same way, erring conservative where the coincidence ever fails. Both states' base rates are unusually stable: Kentucky's are frozen by settlement through August 2028.

Each state's guide covers its tariffs, tax stack, and programs: Washington, Ohio, Utah, New Hampshire, and Kentucky. The methodology page documents the derivation conventions, and every figure traces to a filed source.

Reminder: descriptive language only (design doc §18.3) — the model computes, the report shows; no recommendations, no guarantees.