Most articles about demand charges explain what they are. This one is about the rate that protects you from them, because that protection often comes with transition rules, and in several major territories the trigger that reduces it is your own growth.
If you are modeling a charging site on today's utility rate, you may be modeling one year of a rate schedule that is scheduled to change. That does not show up in a pro forma built from a single tariff sheet.
First, why the relief exists at all
A demand charge bills your highest power draw in the month rather than the energy you sold. A fast charger sets a high peak the first time it runs, and if the same charger reaches approximately the same maximum kW each month, its demand-charge component can stay similar across very different session volumes. Energy charges and other volumetric components still rise with use, so the total bill is not fixed. But the demand component lands hardest on new and underused sites, which is every site at the beginning.
The size of the effect was measured rather than estimated. In 2017, Rocky Mountain Institute analyzed every charging session from 2016 across all 230 of EVgo's DC fast chargers in California and published the results as the eLab EVgo Fleet and Tariff Analysis. Its headline finding:
"With today's EV market penetration and current public DCFC utilization rates, demand charges can be responsible for over 90% of electricity costs, which are as high as $1.96/kWh at some locations during summer months."
At that reported extreme, the utility cost was nearly seven times RMI's 2017 gasoline-equivalent benchmark of $0.29/kWh.
RMI also described the mechanism operators talk about among themselves and that rarely makes it into print, though note the scope carefully. Under the particular high-demand, low-energy tariffs it studied, "tariffs with high demand charges and low energy charges show minimal variation in the total bill across a wide range of DCFC utilization." That is a finding about those tariffs, not a universal rule.
Two things about this evidence matter. It is a decade old, and the tariffs it names are 2016 general commercial schedules rather than EV rates, because in 2016 EV-specific commercial rates barely existed. That gap is exactly what the relief rates were created to close. But the exposure has not gone away: Paren's Q2 2026 report puts average US fast-charger utilization at 15.76%, ranging from 29.7% in the District of Columbia down to 2.0% in Alaska. A large share of the country is still in the band where this bites.
Four designs, and why the difference matters more than the rate
Many large utilities have introduced EV-specific rates or other mechanisms addressing charging economics, but they did not converge on one answer. Four distinct designs appear in the examples below, and the difference between them is worth more to your model than a few cents per kWh.
Utilization-dependent relief. The benefit shrinks or ends as your load factor rises.
A dated discount that steps down. Relief shrinks on a published calendar regardless of how the site performs.
Volumetric conversion. The demand charge is replaced by a per-kWh charge, removing the low-utilization penalty structurally rather than temporarily.
A structurally lower demand charge. The demand charge remains but is set well below the standard commercial rate.
The first two change over time by design. The second two have no stated sunset, which is not the same as permanent. These four illustrate important designs; they do not establish which design is nationally dominant, and no comprehensive national tariff inventory is cited here.
Relief that narrows as the site matures
Florida Power & Light runs a version of this today. Its GSD-1EV and GSLD-1EV demand limiter tariffs apply to qualifying public fast-charging sites and, in FPL's framing, "appropriately set demand charges based on utilization." In a 2025 filing to the Florida Public Service Commission, FPL reported the result:
"Since the introduction of these tariffs in 2021, 34 out of the total 76 customers (45%) who initially signed up have successfully transitioned to regular rates."
Forty-two customers remained enrolled at the end of 2024. FPL's own annual report describes the mechanism plainly: sites whose load factors rose above 10% stopped receiving the demand limiter benefit and were moved to the applicable standard rate.
FPL frames this as success, and from a ratemaking view it is, because the rate did its job and the site graduated. From the operator's seat it is a cost increase triggered by the volume growth the project was underwritten on. Worth noting for anyone modeling Florida: FPL is seeking approval to make these pilot tariffs permanent, so the mechanism looks likely to persist rather than lapse.
Con Edison offers an EV Phase-In Rate for qualifying commercial customers who would otherwise take service under SC9 rates 1 and 2. In the utility's own words it is "designed to mitigate the impact of demand charges by phasing them in as site utilization increases." Con Edison separately runs incentive programs for EV charging, and those are a different thing from a tariff. An expired rebate and a rate schedule are operationally and legally distinct, and you should model them separately.
In California, SCE's commercial EV rates are currently energy-only. Rate schedules TOU-EV-7, TOU-EV-8 and TOU-EV-9 have no demand charges today. SCE's July 2025 rate fact sheet states that demand charges for TOU-EV-8 and TOU-EV-9 "are currently not applicable until determined and authorized by the Commission."
An Electric Vehicle Rate Design Settlement Agreement in SCE's pending rate case, application A.24-03-019, would maintain that energy-only structure and directs post-2030 EV rate design to a separate proceeding. It also describes a temporary, load-factor-based option intended to protect qualifying low-load-factor sites once demand charges begin phasing in after 2030. For that specific option, the settlement describes eligibility rules worth reading closely:
"participation is limited to three years from the commencement of energization. SCE will monitor the participating facilities' load factors on a monthly basis to ensure continued eligibility. The settlement agreement indicates that a facility that exceeds the load factor upper threshold for three consecutive months will be moved to the then default EV rate option and cannot return to the rate option even if its load factor declines below the threshold."
Two cautions. Those rules govern that temporary option for qualifying sites, not every customer on a future default EV tariff. And as of this writing the settlement appears in a proposed decision rather than a document we have confirmed as finally adopted, so treat it as the direction of travel and let current filed tariffs control near-term modeling.
These Florida, New York and future SCE examples share a broad principle, greater cost recovery as sites mature, but their triggers and timelines differ materially and they should not be treated as one policy.
The calendar version: SDG&E
SDG&E's EV-HP rate works on a schedule instead. Its filed tariff states that Schedule EV-HP "replaces the traditional maximum demand charge with a subscription charge billed monthly based on the customer's subscription level," alongside time-varying energy charges.
The transition was set by the California Public Utilities Commission in Decision 20-12-023, issued in December 2020:
"Subscription and energy charges would remain the same for the first three years and then would linearly phase in recovery of applicable allocated EPMC distribution [revenues] over the remaining seven years. Customers would pay the full EV-HP rate, reflecting their full EPMC-scaled cost of service, beginning in Year 11."
With the rate opening in 2022, that means three flat years through 2024, then seven annual steps, then full cost recovery from 2032. Filed rates confirm the flat period: the subscription charge was $76.71 per 25 kW block in both January 2022 and January 2024.
SDG&E publishes an illustrative chart of the resulting discount, declining from roughly 56% in 2024 to zero in 2032. Treat those figures with care. SDG&E's own footnote describes them as illustrative, based on an example load profile, and reflecting rates as of February 1, 2025. They are a whole-bill estimate rather than a tariff parameter, and your site's curve depends on your own load shape.
One structural point is often stated wrongly. EV-HP does not become SDG&E's general commercial rate in 2032. The decision says customers pay "the full EV-HP rate." The schedule survives; the discount is what disappears.
EV-HP also contains a utilization-linked escalator of its own. Under the tariff's overage provisions, if a customer's maximum demand exceeds their subscription level for three consecutive months, SDG&E increases the subscription level to match the highest maximum demand in that period, and the customer must remain at the higher level for an additional three months. Growth raises the floor.
Two designs with no stated end
Commonwealth Edison took the structural route. An eligible EV-charging delivery point can take service under the Watt-Hour Delivery Class, where the distribution facilities charge is billed per kWh rather than per kW. That removes the low-utilization penalty rather than suspending it. There are commitments attached: an EV-charging delivery point must stay in the class for at least 12 monthly billing periods, and a customer who leaves cannot return for at least 12 months.
Georgia Power kept a demand charge and set it low. Schedule TOU-EVC-5, effective with bills rendered for June 2026, applies to non-residential premises dedicated to EV charging behind a dedicated meter and charges $5.34 per kW on maximum demand, defined as the highest 30-minute demand in the month, with no separate on-peak demand charge. The basic service charge is $154.48 per month.
Neither has a published sunset, and neither is permanent. Georgia Power's tariff carries a one-year contract term and states that the bill is subject to changes approved or amended by the Georgia Public Service Commission. The accurate phrase for both is "no stated sunset," and the gap between that and "permanent" is the gap between a modeling input and a caveat.
What this does to a price you set today
A cost-plus framework is one common starting point for pricing: establish a cost floor, add margin, then check the result against competition, utilization goals, roaming arrangements, contract requirements and what customers will actually pay. The method is reasonable. The assumption underneath it is the fragile part, because cost-plus treats the cost floor as a number you can look up.
If your site sits on any of the mechanisms above, you know this year's version of your cost floor.
Two escalators can operate at once, and they should be modeled separately rather than blended.
The first is ordinary rate inflation. As a broad reference, EIA's May 2026 Short-Term Energy Outlook forecast the average residential price to rise nearly 5% in 2026 and about 2% in 2027, reaching 18.2 cents per kWh in 2026. Those are residential averages and should not be treated as forecasts for commercial EV tariffs, whose customer, demand, delivery and energy components can move differently. EIA's commercial figures in that outlook are sales growth, which is a demand forecast, not a price forecast.
The second is the relief phase-out, and it can be larger. For one modeled California DCFC load profile, RMI estimated in 2017 that the then-proposed SCE and SDG&E tariffs would reduce the bill by roughly 50% to 80% initially. Those results were utility-specific, proposal-specific and load-profile-specific, and they are not a general estimate of today's phase-in exposure. Even in RMI's own example the fully phased SCE tariff still produced a bill about 25% below the old TOU-EV-4 rate, so a phase-out is not always a return to the starting point.
Whether either escalator can be absorbed depends on your margin, utilization trajectory, contract terms, repricing rights and the specific tariff. Some sites can reprice, shift load, add storage, or grow utilization enough to offset higher fixed costs. Others cannot, and the ones with the least room are those locked into multi-year host agreements or network contracts at a fixed retail price while the cost floor moves underneath them.
What to actually do about it
Model the post-transition year, not just year one. Run the pro forma at the rate you expect in year five or year seven alongside today's. If the project only works during the introductory-rate period, its long-term economics depend on continued relief, repricing, higher utilization or another cost-control strategy, and you should know which one you are relying on. The commercial ROI snapshot takes a demand charge and tariff type directly.
Find out which design you are on. This is one question to your utility account representative, and the answer moves your model more than most equipment decisions. Ask specifically whether the relief ends on a date, narrows with load factor, or has no stated end, and get the trigger in writing.
If the trigger is a load factor, find the threshold. A utilization-linked rate means the year your site performs best may also be the year your costs step up. That is not a reason to avoid the rate. It is a reason to know where the line sits and to model your ramp against it.
Keep the right to reprice. On a host agreement or network contract, a fixed retail price over a long term against a cost floor with scheduled transitions is a margin risk. A tariff-change or change-in-law provision costs nothing at signing.
Do not read "no stated sunset" as permanent. ComEd's and Georgia Power's designs are more durable than a dated waiver, and both remain tariffs a commission can revise.
None of this is a trap being sprung on operators. These rates exist because regulators concluded that fast charging struggles to pencil without them, and most were designed to phase toward full cost recovery as the market matures. That is defensible policy. It becomes a problem only when a long-lived asset is underwritten on a rate whose transition rules were never read.
- How to Read Your Commercial EV Tariff Sheet, for the anatomy of the rate document itself
- Pricing Your Charging, for the cost-floor-plus-margin method this article stress-tests
- DC Fast Charging ROI, where demand charges dominate the math