Commercial Refrigeration Energy Consumption: U.S. Statistics and Benchmarks

These are modeled estimates, not meter readings. CBECS does not directly submeter the refrigeration end use. EIA builds engineering models of each end use from surveyed building characteristics, imputes missing consumption, and calibrates the modeled end uses to each sampled building’s reported or imputed whole-building energy total. EIA documents this openly. Describe these as modeled national estimates, not measured readings.

The reference year is 2018, not 2026. CBECS is a periodic federal survey with separate building and energy-supplier collection phases. As of our verification date, EIA still presents 2018 as the latest completed national consumption cycle.

Site energy, not source energy. All Btu figures here are site energy — what arrives at the building. Source energy includes generation and transmission losses and uses a different accounting basis. Mixing the two produces a result that isn’t comparable to anything, and it is exactly why the 4% and 7% figures look like they contradict this page’s 5.4%.

“Major fuels” is not “all fuels.” EIA’s 6,787-trillion-Btu total covers electricity, natural gas, fuel oil, and district heat. That is the denominator behind the 5.4% share.

Equipment categories overlap and cannot be added. One restaurant can hold a walk-in, an ice machine, a reach-in, and a vending machine, and CBECS counts it in all four rows.

We did not summarize a secondary source. Every published quantitative claim on this page is identified as a primary-source value, a documented calculation, or an unresolved provenance finding. Derived values use the formulas shown here.

Refrigeration share of a category's electricity
    = category refrigeration kWh ÷ category total kWh

Refrigeration share of a category's major-fuels site energy
    = category refrigeration TBtu ÷ category total major-fuels TBtu

Category share of national refrigeration electricity
    = category refrigeration kWh ÷ 108 billion kWh

Category share of national floorspace
    = category floorspace ÷ 96,423 million sq ft

Gross refrigeration intensity (kWh/sq ft)
    = category refrigeration billion kWh × 1,000 ÷ category floorspace in million sq ft

Refrigeration Load Concentration Index
    = category share of national refrigeration electricity
      ÷ category share of national floorspace

Two shares, one end use. The reason they differ is arithmetic, not disagreement: refrigeration is essentially all electric, so it takes a bigger bite out of the electricity pie than out of the major-fuels pie, which also contains a great deal of natural gas burned for space and water heating.

Which one should you quote? Use 9.0% of commercial-building electricity when the story is about electricity, load, or the grid. Use 5.4% of major-fuels site energy when the story is about total building energy. Say which. A reader who sees “refrigeration is 9% of commercial buildings” with no unit attached has been handed a number they cannot check.

Ranking building types by total refrigeration energy tells you where the load sits. It doesn’t tell you where the load is concentrated, because a category with a lot of floorspace will accumulate a lot of everything.

So we built an index for it. The Refrigeration Load Concentration Index (RLCI) divides a building type’s share of national commercial-refrigeration electricity by its share of national commercial floorspace. An RLCI of 1.0 means a category’s refrigeration share matches its floorspace share. Above 1.0 means more concentrated; below means less.

This is our own derived metric, not an EIA statistic, and it is calculated from EIA’s rounded published values.

Food-sales buildings’ share of national commercial-refrigeration electricity was about 23 times their share of commercial floorspace. Food service was about 10 times. Nothing else in the commercial building stock is close — third place, enclosed and strip malls, sits at 2.6.

How precisely should you quote it? Not to the decimal. RLCI is calculated from rounded survey estimates, so quote food sales as “about 23 times” or “more than twenty times its floorspace share.” A defensible confidence interval would require the joint covariance of the two survey estimates or replicate-weight analysis, neither of which is available from the published tables — so we are not publishing one, and you should be suspicious of anyone who does.

One structural note: EIA’s classification rule places some grocery-store and restaurant load in the mercantile row, because stores and restaurants inside strip shopping centers are classified as mercantile rather than food sales or food service. CBECS does not quantify how much of the mercantile figure that rule explains.

This distinction sits in a footnote, and it is why two people can quote two different “warehouse refrigeration benchmarks” and both be correct.

EIA’s definition: each column is electricity consumption for the end use divided by the floorspace in buildings that use electricity for that particular end use. Buildings without the end use are excluded from the denominator entirely. Note that it is a ratio of aggregate consumption to qualifying floorspace — not a median, and not a typical building.

A useful identity. Divide the gross figure by the conditional figure and you get the share of a category’s floorspace that EIA’s conditional denominator retains. For warehouses that is 0.57 ÷ 1.0, or roughly 57%; for food sales it is close to 100%. So the two columns together tell you something neither tells you alone: how much of a category has electric refrigeration at all. Both inputs are rounded, so read these as approximations, not precise shares.

Saying “food sales is 38% of refrigeration” is wrong in a way that sounds right. Refrigeration is 38% of food sales. Food sales is 24% of refrigeration.

On the 87.1 figure. EIA reports food-sales refrigeration energy intensity as 87.1 thousand Btu per square foot. This is a conditional building-level intensity — refrigeration major-fuels site energy divided by floorspace in food-sales buildings that use energy for refrigeration. It is not the consumption of an individual refrigerator, freezer, or display case. Reproduce it as 87.1 kBtu/sq ft and the ambiguity disappears.

Why food sales is not the same as supermarkets. EIA’s food-sales category held about 163,000 buildings in 2018. By building count, roughly three-quarters were convenience stores; grocery stores and food markets carried far more floorspace and energy per building and accounted for about 63% of the category’s energy. Add the strip-center classification rule and a meaningful share of American grocery retail sits in the mercantile row instead.

Within food-sales buildings, EIA reports refrigerated cases or cabinets in 83% and walk-in refrigerators or freezers in 70%. Food-sales buildings used 53.3 kWh per square foot of electricity in 2018 against 12.6 for the average commercial building — more than four times the national figure before you isolate refrigeration at all.

What the equipment table shows once you read it correctly

Compare the first and last rows. EIA’s conditional total-electricity intensity is 13.5 kWh per square foot for buildings with any refrigeration equipment and 8.5 for buildings with none — a gap of 5.0. The reported refrigeration end use for the first group is 1.4. So refrigeration itself accounts for under a third of the difference between the two groups. These rows describe different building populations, and Table E6 does not establish what causes the remainder.

On large cold storage. These 69,000 buildings have the highest conditional refrigeration intensity among the equipment-characteristic rows in Table E6, at 6.2 kWh per square foot. Treat the count carefully: EIA publishes a 17.2% relative standard error, which puts a normal approximation at roughly 46,000 to 92,000 buildings.

CBECS is not a cold-storage census. It excludes buildings whose principal activity is industrial or manufacturing, including food-processing plants, while including commercial warehouse-and-storage buildings that meet the survey definition. The 2018 CBECS tables used here do not isolate a national cold-storage-only energy total, and the warehouse-and-storage category combines refrigerated and unrefrigerated facilities.

First, CBECS cannot answer it. EIA reports building-level energy, not equipment-level energy, and says so in its table notes.

Second, DOE has published per-unit estimates — just not recently. Its 2009 report gives unit energy consumption for the seven major equipment classes, estimated for the 2008 installed base.

The end-use estimates are modeled. EIA constructs engineering models of each end use from building characteristics, imputes missing consumption, and calibrates the modeled end uses to each sampled building’s reported or imputed whole-building total. For high-refrigeration categories such as food sales, the modeling leans on end-use intensity estimates per square foot by building type.

The data year is 2018. Equipment stock, refrigerants, store formats, and standards have all moved since. Anyone using these figures should give the data year.

Several estimates carry high relative standard errors. EIA publishes them alongside the tables. The consequences are real:

Treat 233 trillion Btu for food sales as an estimate with a wide band, not a precise total. Treat 69,000 cold-storage buildings the same way.

Scope exclusions matter. CBECS covers buildings larger than 1,000 square feet whose principal activity is nonresidential, nonagricultural, and nonindustrial. Food-processing refrigeration and industrial facilities fall outside it. Grocery stores and restaurants inside strip centers are classified as mercantile.

Derived percentages come from rounded published values and will differ slightly from calculations run on the unrounded public use microdata. Where the underlying value rounds to 1 or 2 billion kWh, the derived percentage is coarse.

This dataset was compiled by the editorial team at Columbus Commercial Refrigeration. Every published quantitative claim is identified as a primary-source value, a documented calculation, or an unresolved provenance finding, and each carries its source, data year, and denominator. Derived values use the formulas shown in the methodology section, and the arithmetic was checked against EIA’s three published gross major-fuels intensities and the total-electricity Btu conversion. Where a widely cited figure could not be traced to a primary source, it is reported as unresolved rather than repeated.

The Columbus Commercial Refrigeration research section is a research and reference resource on commercial refrigeration data, standards, and regulation. Columbus Commercial Refrigeration is a commercial refrigeration service-request website; this research page contains no service recommendation, provider placement, lead routing, or commercial offer.

Next scheduled review: July 2027, or on release of a new CBECS consumption cycle, whichever comes first.