Pull up your last commercial electric bill. Find the line that says "demand charge" or "peak kW" or something similar. It's probably a dollar figure that looks bigger than you'd expect — and on most commercial bills, it's quietly making up 30% to 50% of the total. If you don't know exactly how that number got calculated, you're not alone. Demand charges are the least-understood line item on the commercial utility bill, and they're worth understanding because they're also one of the most controllable.

What a Demand Charge Actually Is

Every utility bill has two kinds of charges: energy and demand. Energy is what you'd expect — it's how much electricity you used, measured in kilowatt-hours (kWh). Run a 100-watt lightbulb for 10 hours, you've used 1 kWh. Multiply by millions for a commercial facility, and you've got your energy charge.

Demand is different. Demand measures how fast you used electricity at your peak moment during the billing cycle. It's not about total consumption — it's about the highest single 15-minute (or sometimes 30-minute) interval of the month. That peak, measured in kilowatts (kW), gets multiplied by a utility-specific demand rate, and that's your demand charge.

Think of it like a highway. Energy charges are the tolls — you pay per mile driven. Demand charges are the lane reservations — you're paying for the width of road the utility had to build to handle your peak rush hour. Even if you only used that capacity for 15 minutes once, the utility had to provision for it.

Why Demand Charges Exist

Utilities have to build their grid to handle the highest demand on their system, not the average. A factory that runs at 500 kW all day requires less infrastructure than a factory that sits at 200 kW most of the time but spikes to 800 kW for one hour a day. Both might consume similar total kWh over the month, but they cost the utility very different amounts to serve.

Demand charges recover the cost of that capacity. They exist so that customers whose peaks drive infrastructure costs pay a proportional share. This isn't unfair — it's economically rational. The problem is that it creates a cost dynamic most customers never think about and can dramatically reduce with a few operational changes.

How the Calculation Actually Works

Here's the basic formula for most commercial tariffs:

Monthly Demand Charge = Peak Demand (kW) × Demand Rate ($/kW)

Your peak demand is the highest 15-minute average of kW usage during the month. The demand rate depends on your utility and your tariff class — it ranges from about $3/kW on some residential-adjacent commercial tariffs up to $25+/kW on large industrial tariffs.

Some tariffs add a wrinkle called a demand ratchet: instead of charging you for the current month's peak, they charge you for the highest peak over the last 11 or 12 months (or a fraction of it). So a single bad hour in August can inflate your demand charges through the following July. That's not a typo — it's how some ratchet structures work.

In capacity-market states (PJM, ISO-NE, NYISO), there's also a capacity charge that's calculated similarly but on a different basis — your usage during a small number of grid-wide peak hours over the course of a year. That's a separate topic (we've written about capacity elsewhere), but it operates on the same principle: peak usage drives future cost.

Where Demand Peaks Come From

The operational reality is that most commercial demand peaks come from a handful of predictable events:

• Startup loads in the morning. All your equipment turning on at once. HVAC ramping, compressors starting, production lines firing up. This is the classic 7-9 AM peak at most facilities.
• Summer afternoons. Air conditioning plus everything else running at full tilt. For most commercial buildings, the annual peak lives somewhere between 2-5 PM on a July or August weekday.
• Simultaneous equipment operation. The welder turning on at the same time as the compressor at the same time as the wash cycle. A scheduling accident that creates a peak that doesn't reflect typical operation.
• One-off events. Testing a backup generator, starting up after a shutdown, ramping up for a production campaign. These create peaks that are atypical but still expensive.

Understanding where your peaks actually happen is the first step in doing anything about them. Most facilities have never looked. Pull 15-minute interval data from the utility (most utilities provide it on request or through an online portal) and graph it. Your peaks will become obvious.

How to Reduce Demand Charges

The good news: demand charges are among the most controllable line items on a commercial bill. A few operational changes can move them materially.

• Stagger startup. Instead of everything turning on at 7 AM, stagger equipment over 30-60 minutes. The total daily energy is the same, but the peak is lower. Cheapest intervention available — often just a scheduling change.
• Pre-cool or pre-heat. For buildings with thermal mass, running HVAC harder in the early morning and easing off in the afternoon can shift the peak off the expensive hours. Requires a programmable BMS but no capex.
• Peak shaving with storage. Battery storage systems are designed for this. Charge overnight when demand is low, discharge during your peak hours. The ROI has gotten dramatically better as battery costs have fallen — for facilities with high demand charges, paybacks of 3-5 years are increasingly common.
• Demand response enrollment. In some markets, the utility or ISO will pay you to curtail demand when the grid is stressed. Different from reducing your own peak, but related — the monitoring and curtailment protocols overlap.
• Load scheduling. Don't run the large non-urgent loads during peak hours. Move washdown cycles to nights. Run batch processing overnight. Every kW you move off-peak is a kW you don't pay a demand charge on.
• Tariff optimization. Different tariffs have different demand structures. For some load profiles, a time-of-use tariff with higher energy charges but lower demand charges works out better. Your advisor should be modeling this against your actual usage.

The Ratchet Problem

If your tariff has a demand ratchet, the arithmetic is different. Reducing this month's peak doesn't help — you're still paying based on the ratchet from six months ago. To escape a bad ratchet, you have to wait out the window (11-12 months, depending on the tariff) while keeping peaks controlled in the meantime.

Once the ratchet rolls off, you've reset to the new lower baseline. Then you have to be careful not to spike again, because that new spike becomes the ratchet for the next year. This is why a single bad hour can be so expensive — it sets the floor for months.

Bill auditors regularly find demand ratchets that were set by anomalous events years ago and are still inflating current bills. If you're on a ratcheted tariff, this is worth checking.

The ROI Case

For a typical commercial facility paying $15/kW in demand charges with a 500 kW monthly peak, that's $7,500 a month — $90,000 a year — just in demand. Reducing the peak by 20% through operational changes saves $18K a year. Installing a battery that shaves 100 kW off the peak saves $18K a year on demand charges alone.

For larger industrial facilities, the numbers get big fast. A manufacturing site with a 3,000 kW peak at $12/kW is paying $432K a year in demand charges. A 25% reduction is $108K. Every year. Forever.

This is why demand management shows up on serious industrial energy programs — the ROI is typically better than supply-side procurement, and the savings compound over time.