Why Cooling Costs Keep Rising, and How to Separate the Causes

Key takeaways

• Electricity prices are rising: EIA reports the average U.S. commercial price was 14.53 cents per kWh in July 2026, up 3.4% from a year earlier.
• Demand charges typically make up 30% to 70% of a commercial customer's electric bill, according to an NREL and Clean Energy Group analysis, and cooling often drives the peak.
• Weather, building use and equipment efficiency change kWh; rates and tariffs change the price of each kWh and kW.
• Splitting a year-over-year increase into rate, weather and unexplained portions shows how much is within your control.
• The unexplained portion is where equipment decline, controls drift and schedule creep hide, and where action pays off.

When the cooling bill goes up, everyone has a theory. Finance blames the utility. The utility blames the weather. Operations blames the tenants who prop doors open. Usually all of them are partly right, and the useful question is how much each cause contributes, because only some causes are within your control. This page lays out the main drivers of rising cooling costs and a step-by-step method to separate them using data you already have.

The seven drivers

1. The price per kWh

Retail electricity prices have been rising. The U.S. Energy Information Administration (EIA) reports that the average revenue per kWh for the commercial sector was 14.53 cents in July 2026, up 3.4% from July 2025 (the latest month published as of October 2026). Your rate depends on your utility or retail provider, your tariff and, in areas with retail electric competition, such as much of the ERCOT region in Texas, your supply contract. Riders and pass-through charges can change mid-contract.

2. Demand charges

Many commercial tariffs bill not only for energy (kWh) but for the highest rate of use (kW) during the billing period, often measured over a 15- or 30-minute interval. A 2017 analysis by the National Renewable Energy Laboratory and Clean Energy Group found that demand charges typically account for 30% to 70% of a commercial customer's electric bill, and that nearly 5 million of the country's 18 million commercial customers could subscribe to tariffs with demand charges above $15 per kW. Cooling is often the largest contributor to the summer peak, so a cooling system that works harder at 4 p.m. on the hottest day can raise a whole month's demand charge.

3. Weather

More and hotter days mean more cooling. Cooling degree days (CDDs), which EIA defines using a 65°F base temperature, are the standard way to measure it. EIA's modeling of a warmer-weather scenario projected 4% more purchased electricity in buildings by 2050 from added space cooling, and electricity prices 8% higher than the reference case, partly because of that summer demand. Weather is real, but it is measurable, which means you can correct for it. For the Gulf Coast context, see hotter summers and cooling load.

4. Load changes

More occupants, longer hours, new servers, new kitchen equipment, a tenant who runs a gym, or a change from offices to lab space all add cooling load. These are business decisions, but they need to be separated from efficiency problems so equipment does not get blamed for them, or vice versa.

5. Equipment efficiency decline

Equipment delivers less cooling per kWh as it ages, through coil fouling, refrigerant charge drift, airflow problems, internal oil films and wear. A common planning model from NREL assumes about 1% efficiency loss per year with annual professional maintenance and 2% to 3% per year without it. See why HVAC efficiency declines with age.

6. Controls drift

Schedules overridden "temporarily" and never restored, setpoints nudged down, simultaneous heating and cooling, sensors out of calibration and stuck economizers all add hours or lift. A 2004 ACEEE summary of rooftop unit field studies reported a 64% economizer failure rate. These problems are often cheap to fix once found.

7. Tariff and rate structure changes

Moving to a time-of-use tariff, a new ratchet clause (where a high summer peak sets a minimum demand charge for months afterward), or new riders can change the bill even if usage does not change at all. Read the tariff sheet, not just the bill total.

What you control and what you don't

A four-step method to separate the causes

1. Gather 24 to 36 months of bills. For each month, record kWh, peak kW, the energy charge, the demand charge, fixed charges and riders. Interval data from your utility or smart meter portal is even better.
2. Calculate the effective rates. Divide energy charges by kWh and demand charges by kW for each month. This shows how much of the increase is price alone.
3. Weather-normalize consumption. Get monthly CDDs for your nearest weather station. Fit a simple line of monthly kWh against CDD for a baseline year: the intercept is the weather-independent base load, and the slope is the extra kWh per CDD. Use that line to predict what the current year "should" have used given its weather.
4. Attribute the change. Price effect = last year's kWh times the change in effective rate. Weather effect = predicted change in kWh times this year's rate. Whatever is left is the unexplained portion: load changes, equipment decline and controls drift.

This is a simplified version of the baseline-and-adjustment approach used in formal measurement and verification. For more, see how HVAC energy savings are measured.

Worked example

This example is illustrative. A building's annual electricity bill (energy charges only, for simplicity) rose from $135,000 to $151,200.

A baseline fit from Year 1 monthly data gives a base load of 600,000 kWh a year plus about 142.9 kWh per CDD. With Year 2's weather, the building would be expected to use 600,000 + (142.9 × 2,912) ≈ 1,016,000 kWh. It actually used 1,080,000 kWh.

In this example, the largest share of the increase is not the utility and not the weather. About 6% more electricity was used than the weather can explain. The next step is to check whether load changed; if it did not, equipment and controls are the prime suspects.

Common mistakes when comparing bills

• Comparing bills, not periods. Billing cycles vary from about 28 to 33 days and rarely match calendar months. Convert to daily averages, or line up degree days with the actual meter read dates.
• Ignoring estimated reads. An estimated bill followed by a catch-up bill can make one month look terrible and the next look great.
• Mixing tariffs. If the account moved to a new rate or supply contract during the period, separate the two before calculating trends.
• Forgetting one-off events. A failed chiller running on a rental unit, a construction project or a data closet added mid-year will show up as "unexplained" unless you note it.
• Looking only at totals. A building with gas heat, electric cooling and a large plug load needs the cooling-season months examined on their own; annual totals can hide a summer problem behind a mild winter.

Don't forget the demand side

Equipment that has lost capacity tends to run flat out through the afternoon peak, and multiple units that once cycled independently start running at the same time. That can raise the monthly peak even when total kWh barely moves. As an illustration only: at $12 per kW, a 20 kW rise in monthly peak adds $240 a month, and under a ratchet clause it can keep costing money well after summer ends. Check whether your peak kW per CDD on the hottest days is trending upward.

Acting on the unexplained portion

• Controls first. Review schedules, setpoints, economizer operation and simultaneous heating and cooling. The DOE Federal Energy Management Program's O&M guide cites estimates that O&M programs targeting energy efficiency can save 5% to 20% on energy bills without significant capital investment.
• Maintenance second. Clean coils, verify charge, fix airflow and, for chillers, check approach temperatures. See commercial HVAC maintenance for efficiency.
• Then the inside of the equipment. If equipment still underperforms, internal causes such as oil fouling or compressor wear are next. CryogenX4 describes its one-time treatment as removing insulating oil from internal coil surfaces and improving compressor lubricity; the company reports energy savings of up to 30% and typical payback of 12 to 36 months, with results varying by equipment condition. See what is oil fouling.
• Replacement last, and deliberately. Some equipment is genuinely at the end of its life. Make that call with data, not because of one expensive summer.

Next step

Run the four-step method on your last two summers of bills. If the unexplained share is large, use the HVAC savings calculator to size what an efficiency improvement would be worth at your current rate.

Frequently asked questions

Where can I find cooling degree days for my location?

NOAA and many weather data services publish daily and monthly degree days by station, and some utility portals and energy management tools include them. Use a station near your building and a 65°F base unless your analysis shows a different base fits better.

What is a demand ratchet?

A tariff clause that sets your billed demand for a number of months to at least a percentage of the highest peak recorded in a prior period, often summer. One high peak can therefore raise demand charges for months. Check your tariff sheet for the details.

Is the price per kWh the biggest driver of higher cooling bills?

Sometimes, but often not. Separating the bill into rate, weather and unexplained portions frequently shows that consumption changes, including equipment decline and controls drift, are a large share of the increase.

Does fixing equipment efficiency reduce demand charges?

It can. Equipment that delivers the same cooling with less power, or holds setpoint without running every compressor at once, can lower the peak. The size of the effect depends on when your peak occurs and what drives it, which interval data can show.

Sources

1. Electric Power Monthly: Electricity Monthly Update (July 2026 data, released September 2026) — U.S. Energy Information Administration
2. Identifying Potential Markets for Behind-the-Meter Battery Energy Storage: A Survey of U.S. Demand Charges (NREL and Clean Energy Group, 2017) — OurEnergyPolicy
3. Degree days explained — U.S. Energy Information Administration
4. EIA explores effects of weather projections on energy consumption in buildings — U.S. Energy Information Administration, Today in Energy
5. Best Practice: SEER, EER, HSPF and AFUE Degradation (applies NREL Building America Performance Analysis Procedures) — Texas Department of Housing and Community Affairs
6. Upstream Solutions to Downstream Problems: Improving Field Performance of Small Commercial Rooftop Units (ACEEE Summer Study 2004) — Jacobs, Higgins & Shwom, ACEEE
7. O&M Best Practices: A Guide to Achieving Operational Efficiency, Release 2.0 (PNNL-14788) — Pacific Northwest National Laboratory for DOE Federal Energy Management Program