Solar glossary

Demand Charges: Why a Business Electric Bill Includes kW

A demand charge is a utility charge based on a customer's measured or tariff-defined power demand, commonly expressed in kilowatts.

Updated November 12, 2026 · 10 min read

A demand charge is a utility charge based on a customer’s measured or tariff-defined power demand, commonly expressed in kilowatts. It differs from an energy charge based on total kilowatt-hours consumed. A business can reduce monthly energy use while still incurring a substantial demand charge if its highest qualifying demand remains unchanged.

Energy and demand measure different things

Energy is the quantity used over time. Demand describes the rate of use over the utility’s specified measurement interval. The Department of Energy’s rate guidance explains energy charges, demand charges, and fixed charges as separate bill components.

The tariff determines the interval, billing period, time windows, and any historical-demand or minimum-billing rules. Do not assume every utility uses the same interval or that the bill always reflects only the current month’s single highest reading.

A simple billing example

Suppose a hypothetical tariff charges $18 per kW of billing demand. If the applicable billing demand is 40 kW, that demand component is 40 × $18 = $720. Energy charges and other bill items are additional.

If an operational change reduces qualifying billing demand to 35 kW under the same simplified tariff, that component becomes $630. The $90 difference assumes no ratchet, minimum, other demand component, or tariff change. These are illustrative figures, not a Connecticut, Massachusetts, or Rhode Island rate quote.

Why solar may not eliminate the charge

Solar can reduce the grid power a site needs while it is generating. But the business’s billing peak may occur during low solar output, after sunset, or during a short cloudy period. The actual interval data and tariff determine the effect.

A calculation that multiplies annual solar production by a blended bill rate can overstate savings if it assumes demand charges fall in proportion to energy consumption. Model the demand component separately.

How storage and operations can help

A battery may supply power during selected demand periods, a strategy often called peak shaving. Success depends on available energy, output, controls, forecasting, and the duration and timing of the peaks. A depleted battery cannot reduce a later peak simply because it reduced an earlier one.

Operational changes can also matter: staggering equipment starts, changing schedules, or controlling discretionary loads may reduce coincident demand. Compare those options with storage rather than treating a battery as the only available tool.

How demand is measured

A demand measurement describes power over the interval defined by the tariff and metering system. It is not necessarily the instantaneous reading an employee sees when a machine starts. The billing method may use a maximum interval value, a time-specific maximum, a historical reference, or another defined rule.

That distinction matters when investigating a bill. A brief equipment event may have a different effect from a load sustained through the relevant interval. The actual tariff and interval data are needed to connect operational behavior with billing demand. Do not assume a generic fifteen-minute example applies to every account.

Request the rate schedule associated with the meter, including any applicable riders or demand provisions. The name of the electric company alone does not identify the customer’s billing method. Different account classes and service arrangements can have materially different cost structures.

A kWh-versus-kW example

Imagine two hypothetical businesses that each use 100 kWh during a defined operating period. One draws a steady 10 kW for ten hours. The other draws a steady 50 kW for two hours. Their energy use is the same, but their demand profiles are different.

Under a simplified tariff that bills those sustained maxima, the second business would have greater billing demand. This does not establish either business’s complete bill because actual tariffs can contain time windows, ratchets, minimums, and other components. It illustrates why equal monthly kWh does not imply equal utility cost.

A solar or battery proposal should therefore examine both the amount and timing of consumption. Annual energy use is a useful input, but it cannot identify every demand-saving opportunity or prove that a proposed system will reduce the qualifying peak.

Different demand provisions need different strategies

A non-time-specific demand provision may consider the highest qualifying demand during the billing period. A time-of-use demand provision may focus on specified hours. A ratchet can link billed demand to a previous peak or another historical condition. The tariff defines the details; these descriptions are categories, not claims that a particular utility uses all of them.

A strategy that addresses one provision may leave another unchanged. Reducing an afternoon peak may help a time-window charge while a higher off-hours event still determines a different component. Similarly, reducing this month’s measured peak may not immediately lower a bill governed by an applicable historical minimum.

Map every demand-related line item before modeling savings. Ask the analyst to identify the measurement basis, rate, window, and any carryover condition. This prevents the entire demand portion of a bill from being treated as one simple charge with one simple remedy.

Read the interval data alongside the operating schedule

Interval data can show when high demand occurs, but the operating schedule helps explain why. Compare the timestamps with production starts, heating and cooling schedules, refrigeration cycles, equipment tests, vehicle charging, and other major activities relevant to the site.

Look for repeatable patterns and unusual events separately. A routine daily peak may suggest a different control opportunity from an isolated event associated with maintenance or an exceptional operating day. Avoid removing an unusual peak from the analysis merely because it is inconvenient; if it can recur under the tariff, it may still matter financially.

A useful review produces a short list of candidate causes supported by data. It does not diagnose equipment solely from a graph. Facility personnel, electricians, controls specialists, and the energy designer may need to combine their knowledge to establish which changes are practical and safe.

Peak shaving with a battery

Peak shaving uses battery output to reduce the grid demand seen under the applicable billing method. The battery must have enough power to reduce the peak and enough energy to sustain the response over the relevant period. Controls must also dispatch it at the right time.

As a hypothetical example, reducing a sustained 20 kW portion of a peak for two hours would require 40 kWh delivered at the relevant load boundary, before any additional allowances not already reflected in that boundary. A battery with sufficient energy but inadequate output would not meet that target. A battery with sufficient output but too little energy could stop helping before the peak period ends.

The example is not a storage recommendation or a guaranteed demand reduction. The actual profile may have multiple peaks, variable duration, different billing windows, and recharge constraints. Model the full sequence rather than size storage around one attractive point on a chart.

Why controls and forecasting matter

A battery used too early can be depleted before the interval that determines the bill. A battery reserved too conservatively may fail to reduce an avoidable peak. The control strategy therefore matters alongside the hardware ratings.

Ask how the proposed controls identify the target, respond to changing demand, and preserve energy for later conditions. Also ask how unexpected operating changes are handled. A business that adds a shift, new machinery, or vehicle charging can change the load profile enough to require a new strategy.

A credible savings model should explain its control assumptions and test less favorable cases. It should not assume perfect knowledge of every future peak unless that assumption is explicitly identified as a theoretical upper case rather than a dependable operating forecast.

Solar’s contribution depends on coincidence

Solar generation can reduce grid demand when it coincides with the site’s relevant load. A daytime business may have useful overlap, but clouds, seasonal changes, and peak timing can limit the effect. A single low-generation interval can matter when the tariff uses the maximum qualifying demand.

This is why a large annual solar offset does not automatically translate into a similarly large demand-charge reduction. The energy and demand components should be modeled separately using the same interval framework where appropriate. Avoid applying the full blended electricity bill rate to every generated kWh without examining what charges are actually avoided.

Solar and storage can be evaluated together, but their benefits must not be counted twice. If the solar model already reduces the peak, the storage model should address the remaining grid profile rather than claim the same original reduction again.

Operational changes may be part of the answer

Some sites can change when discretionary equipment operates, stagger starts, or adjust supported control schedules. Others have production, food-safety, comfort, or process requirements that limit flexibility. The energy analysis must respect those constraints rather than assume every load can be shifted without consequence.

Compare practical operational measures with equipment investments. A low-cost scheduling change may address a recurring coincidence, while another site may need storage or a different technical solution. The appropriate choice depends on the data and business requirements, not a universal preference for one technology.

Any electrical or control modification should be reviewed and implemented by qualified personnel. A billing opportunity is not a reason to bypass protective devices, disrupt necessary processes, or change equipment operation outside its supported requirements.

Demand response payments and demand-charge savings differ

A demand-response program compensates defined participation or performance under its own rules. A tariff demand charge is part of the electricity bill. The event that determines program compensation may not be the same interval that determines the site’s billing demand.

A battery can potentially be evaluated for both purposes where permitted, but the operating assumptions must fit together. Energy dispatched for one objective may be unavailable for another. The model should identify the priority and any restrictions on simultaneous participation rather than add two independent maximum-benefit forecasts.

Keep program revenue, avoided energy charges, and avoided demand charges in separate lines. That makes it easier to see which benefit depends on which assumption and to assess the result if one opportunity changes.

How to judge a commercial proposal

Request the baseline tariff and load data, the proposed post-project profile, the calculation for each bill component, and the operating assumptions. Ask whether the model includes changes in rate class, minimum charges, or other provisions triggered by the project where relevant.

Compare modeled results over representative periods rather than one favorable month. Seasonal production and business activity can change both energy and demand outcomes. A reasonable sensitivity analysis should show how the forecast responds to different peak timing, storage availability, or operating schedules.

After installation, use an agreed measurement approach to compare actual performance with the modeled assumptions. Differences should be investigated rather than immediately attributed to faulty equipment or dismissed as normal. The quality of the original records determines how useful that comparison will be.

What to give a commercial solar designer

Provide the exact tariff, recent bills, interval data, operating schedule, and planned equipment changes. Identify unusual seasonal loads and any separate meters. A warehouse, restaurant, office, and manufacturing site can have very different demand patterns.

Ask the proposal to show energy savings and demand savings separately, with the model’s assumptions and constraints. A demand-response program payment is also distinct from avoiding a tariff demand charge; prevent double-counting or incompatible operating assumptions.

Frequently asked questions

Can a business reduce energy use while its demand bill stays high?

Yes. If the qualifying peak or applicable billing minimum remains unchanged, lower total kWh may reduce the energy component without reducing the demand component. Review the actual tariff and interval profile to understand the result.

Does a short peak always set the whole month’s charge?

Not necessarily. The measurement interval and billing rules control. A sustained peak, a time-window maximum, or a historical ratchet can have different effects. Avoid applying a common rule of thumb without checking the account’s rate schedule.

Should every business install a battery?

No. Storage needs a supported technical and financial case. Some sites have little addressable demand value, limited flexibility, or a more practical operational alternative. Evaluate the actual load shape, tariff, control strategy, installation cost, and competing uses of the battery before deciding.

Do all commercial accounts have demand charges?

No universal rule applies. Check the actual rate schedule and bill.

Is the demand charge the same as total electricity use?

No. Demand is a power measure; energy use is accumulated over time.

Where can Teamsun help?

A commercial solar consultation can begin with the tariff and load profile. Where storage is relevant, the analysis should explain how it addresses the site’s measured peaks and what could limit the projected result.

Related topics