Rebuilding a fleet charging business case without the credit

Before you calculate the damage, check which number the proforma carried. The commercial Section 30C credit had a six percent base rate. Thirty percent applied only where prevailing wage and apprenticeship requirements were met on the installation, with the compliance documentation to prove it.A great many fleet models carried thirty percent because that is the figure the summaries repeated. If yours did, and your installation was not going to meet those requirements, then part of the loss you are now absorbing was never real — you are correcting an error at the same time as absorbing a change in law. That is an uncomfortable conversation and a much better one than discovering it during an audit.Either way the credit is gone for property placed in service after June 30, 2026. The correction is not optional and it is not a judgement call.

Simple payback is net capital divided by annual net saving. Remove a credit worth a fraction of capital and payback multiplies by one divided by one minus that fraction. Nothing else in the model has to change for this to happen, which is why a proforma that only edits the incentive line and leaves the payback cell alone is still wrong.The table below is the whole effect in one place. It assumes the annual saving is unchanged, which it is — the credit was capital, not operating.Notice how much less the six percent case moves. For a fleet that was honestly modelling the base rate, this change is a nuisance rather than a crisis, and the project that cleared a hurdle at 4.0 years and now clears at 4.3 has not fundamentally changed. For a fleet that was modelling thirty percent, a project at 4.0 years is now at 5.7, and that frequently crosses an approval threshold.

The clean way to do this is to rebuild the capital stack from zero incentives and then add back only what you have verified with a sponsor and a date. It takes an afternoon and it produces a model that survives review, which the alternative does not.Two things to watch while you do it. First, the vehicles. The clean vehicle credits under Sections 30D and 25E ended for deliveries after September 30, 2025, so if your model still nets those against vehicle capex, that line is stale too and it is usually larger than the charging line. Second, anything calculated downstream of the netted figure: cost per port, cost per vehicle, internal rate of return, and the year-by-year cash flow all inherit the error.The operating side is untouched by the change in law, which is good news, because that is where the recovery has to come from.

Demand management is first because it attacks the largest operating line and it costs a controller rather than a construction project. Charging is billed on the highest short interval of the month, so a depot that plugs everything in at shift end and lets it run pays for a peak it did not need. Spreading the same kilowatt-hours across the dwell window can cut the effective cost of every kilowatt-hour the site buys, and it does that every month for the life of the depot.Phasing is second because it moves capital out into later years without giving anything up, provided the underground work is done once. Trench and conduit for the full plan, energise the ports the fleet actually needs, and the capital profile changes shape while the per-port cost of later phases stays low. Building the make-ready once rather than retrofitting is reported to cut per-port cost forty to sixty percent.The rate schedule is third only because it is not always available, not because it is small. Moving an account to a tariff designed for overnight or off-peak load can change operating cost by more than any equipment decision in the project, and it is an application rather than a build. Ask the utility what the account qualifies for at the same time you ask about make-ready.What is not on the list: hardware discounts. Chargers are a minority of a depot bill — commercial hardware commonly runs $700 to $3,500 per port against a total in the thousands to low tens of thousands per port — and negotiating hard on the smallest line is the most common way a project team spends its energy without moving the answer.

A payback number without a named tariff is decoration. The demand rate and the energy rate at your depot are the two inputs that move the operating saving most, and they are utility-specific and rate-schedule-specific. Two depots forty miles apart can produce paybacks that differ by years on identical vehicles and identical routes.So write the tariff into the model header. Utility, rate schedule, energy rate, demand rate, and the date you read them. When somebody challenges the payback a year later — and they will — the argument is then about whether the tariff changed rather than about whether the model was invented.The same discipline applies to residual values, and it is worth flagging that the market context has moved. Battery electric vehicles were about six percent of new US vehicle sales in the first half of 2026, down from about seven percent a year earlier, according to the Energy Information Administration. That is not an argument against electrifying a depot, which is decided by your routes and your tariff. It is a reason to be careful with resale assumptions that were written when volumes were rising, and to make the case stand up without them.

Shrink it before you kill it. A depot sized for the whole fleet may not clear a hurdle rate while a depot sized for the routes with the best duty cycle does. Electrify the vehicles that sit on the yard longest and drive the most predictable miles, keep the rest on their existing powertrain, and let the second phase be decided by the first phase's actual invoices rather than by a forecast.Consider transferring the capital instead of spending it. Charging as a service converts the build into an operating expense and moves demand-charge and utility-delay risk onto a counterparty, at a price. Whether that is a good trade depends almost entirely on how the contract allocates the risks, which is worth reading carefully rather than assuming.And build the model so it can be re-run. The federal line went to zero this year with a single amendment. The thing that made that survivable for well-run fleets was that their models were built from tariffs and duty cycles rather than from a headline incentive, so the change moved one cell instead of invalidating the case.