Demand Charges: Why EV Chargers Can Wreck a Commercial Electric Bill

Read a commercial electric bill and you will find a consumption line, priced per kilowatt-hour, and a demand line, priced per kilowatt. The first measures energy over the month. The second measures power at a single moment — typically the highest fifteen-minute average interval the meter recorded during the billing period.

The demand line has no relationship to how much energy you used. A site that drew a large amount of power for one quarter of an hour and almost nothing for the rest of the month pays a full demand charge. This is not a penalty; the utility sizes transformers, conductors and generation capacity for the peak you might draw, and the demand charge is how that capacity is billed.

For most commercial buildings the demand peak is set by chillers, elevators or the mid-afternoon HVAC load, and it sits at a familiar level year after year. Chargers arrive and add a new load that is large, that starts on a human schedule rather than a thermostat, and that frequently coincides with the peak that already existed.

The numbers below are nameplate continuous output — what the equipment draws when it is running flat out — and they are the figures that matter for demand, because demand is about instantaneous power, not about how long anything runs.

The pattern that hurts is simultaneity. Four Level 2 ports are not four small loads if all four start at 5:30 pm when the shift ends; they are one thirty-eight kilowatt load arriving at once. That is why an unmanaged bank of Level 2 ports can move a demand peak that individual ports never would.

DC fast charging is a different order of problem. A single fast charger can exceed the entire existing demand of a small commercial building, and one session in the wrong fifteen minutes sets a peak the site will pay for all month — and, under a ratchet, for longer.

Many commercial tariffs include a ratchet: a provision that sets the billed demand at the higher of this month's measured peak or some percentage of the highest peak recorded over the preceding several months. The exact percentage and lookback period are tariff-specific and you have to read yours.

The consequence is that a single bad fifteen minutes is not a one-month event. A fast-charging session, a maintenance test that ran everything at once, or a hot afternoon when the chargers and the chillers peaked together can raise the floor under your demand charge for the rest of the ratchet window.

This is the mechanism behind the most common version of the question that brings people to this page: the bill went up, and it stayed up, and the chargers do not appear to be busy enough to explain it. They are not explaining it through energy. They explained it once, through power, and the tariff is repeating the answer.

Every rate in the table below is invented to show the arithmetic. Demand rates vary enormously between utilities and between rate classes within a utility, and there is no national figure worth quoting. Pull your own tariff, find the demand rate and the ratchet language, and substitute.

The example runs a site with four Level 2 ports under three operating regimes: unmanaged and coincident with the building peak, unmanaged but shifted off the peak, and managed so the bank never exceeds a set ceiling.

Notice that the energy sold is identical in all three rows. The vehicles received the same charge. Only the timing and the ceiling changed.

The mitigations are not equal, and the market tends to promote them in roughly the reverse of their usefulness on a typical property.

First, schedule. If the vehicles have a long dwell window and the building has a known peak period, simply not charging during that window removes most of the problem at no capital cost. On an overnight residential or depot site this is close to free, because a car that sits for ten hours does not care which of those hours it charges in.

Second, power sharing. Load management across a bank of ports caps the total the group can draw, sharing a fixed allocation among however many vehicles are connected. It converts a hard peak into a ceiling you choose, and it is usually cheaper than the service upgrade it replaces as well as cheaper than the demand it avoids.

Third, rate selection. Some utilities offer tariffs designed for EV charging loads, occasionally with reduced or deferred demand charges. Whether one exists for your class and territory is a utility-specific fact worth an explicit phone call, and it is free to ask.

A distant fourth, storage. Batteries can shave a peak, and on the right tariff with the right load shape they pay. They are also a substantial capital cost with their own space, safety and maintenance requirements, and they are frequently proposed for sites whose problem would have been solved by a schedule. Treat storage as the answer only after the first three have been ruled out with your actual interval data in hand.

All of this is far cheaper to solve at specification than at retrofit. Power sharing has to be supported by the equipment and configured at commissioning. Scheduling needs a networked charger or, at minimum, a control input. Both are ordinary features to ask for and awkward to add to a bank of pedestals already set in concrete.

Ask the installer for the connected load of the proposed system and the maximum simultaneous draw after any management is applied — those are two different numbers, and the second is the one your demand charge responds to. Then take both to your utility account representative before the order goes in.

One thing that will not help with any of this: a federal tax credit. Section 30C terminated for property placed in service after June 30, 2026 under Public Law 119-21. Utility make-ready programs sometimes do cover load-management equipment, which is worth asking about specifically, because it is the rare case where the sponsor's money and the sponsor's interest point the same direction.