Every electrification plan starts with a share. If hot water is 5 percent of a building's energy, replacing the water heater is a maintenance decision. If it is 30 percent, it is the largest single measure available without entering an apartment.
The U.S. Energy Information Administration publishes the figures that establish which case a building is in, by housing type in the Residential Energy Consumption Survey and by building activity in the Commercial Buildings Energy Consumption Survey. The figures below convert those survey values into kWh thermal, gallons and heat pump electricity without treating thermal kWh as electrical kWh.
EIA Table CE3.1 reports that apartments in buildings with five or more units used 10.1 million Btu per household for water heating in 2020, 29.9 percent of their 33.7 million Btu of total site energy, against 18.3 percent for all U.S. homes. EIA Table E1 reports that water heating was 111 trillion Btu of lodging's 598 trillion Btu in 2018, or 18.6 percent, and 343 trillion Btu of all commercial building energy of 6,787 trillion Btu, or 5.1 percent.
Ten million Btu of gas at an assumed 80 percent efficiency is 2,370 kWh thermal of hot water per apartment per year, or about 950 kWh electrical at an assumed heat pump COP of 2.5.
Key Takeaways
In this article
Where the numbers come from and what they measure
RECS is a household survey. EIA's 2020 edition reports site energy consumption by end use in Table CE3.1, “Annual household site end-use consumption in the United States, totals and averages, 2020.” The table reports trillion Btu for the housing stock and million Btu per household.
Site energy is what crosses the meter: the gas burned in the boiler or the kWh drawn by the element, not the heat that reached the water. The end-use split is an EIA estimate built from survey and billing data.
CBECS is a building survey. The 2018 edition reports major fuels consumption by end use in Table E1 as trillion Btu by principal building activity. The EIA lodging summary page reports end-use intensities per square foot.
Lodging is the commercial category closest to residential occupancy, which is why its water heating share is much higher than the commercial average.
Neither survey reports a multifamily building as a single entity. RECS reports per household, so a 100-unit building represents 100 households plus whatever the central plant loses in the recirculation loop and tank. The per-household estimate does not separate those losses.
Multifamily: hot water is the largest end use
RECS 2020 Table CE3.1 reports the following values for apartments in buildings with five or more units, representing 22.84 million housing units:
Apartments compared with all U.S. homes
For all 123.53 million U.S. homes, RECS 2020 Table CE3.1 gives 9,481 trillion Btu and 76.8 million Btu per household. Space heating was 4,026 trillion Btu, or 42.5 percent, and water heating was 1,737 trillion Btu, or 18.3 percent.
An apartment uses 28 percent less water heating energy per household than the average home and 74 percent less space heating energy. Hot water therefore moves from the second end use in the average home to the first end use in an apartment building.
Shares are computed from the trillion Btu stock totals. EIA's per-household averages for some end uses are averaged over the households that have the end use, so those averages do not sum to the total.
Fuel-specific Northeast averages
RECS 2020 Table CE4.7, the Northeast averages by fuel, reports water heating per household for apartments in buildings with five or more units as follows:
Why the fuel changes the reported site energy
The electric figure is lower because a resistance element converts all of its input to heat in the tank, while a fuel-fired heater loses part of its input up the flue. The 5.2 million Btu electric value is 1,524 kWh electrical per apartment per year.
For all Northeast homes, Table CE4.7 reports 8.3 million Btu per household for electric water heating, 19.9 million Btu for natural gas and 20.0 million Btu for fuel oil.
Converting a RECS figure into hot water
The survey reports fuel input. Converting it to hot water delivered requires an efficiency and a temperature rise. Both are assumptions.
For a gas-fired system at an assumed 80 percent efficiency:
Derived gallons per day
At an assumed 70 F temperature rise, from 55 F inlet water to 125 F delivery water:
What the gallon figure includes
The gallon figure is derived, not surveyed. It absorbs the recirculation and standby losses that the survey folded into fuel use, so fixture draw is somewhat below the derived value.
A building with a 24-hour recirculation loop should expect the loop to be a large part of the total. The heat loss article works the loop at 100 W per apartment, or about 2.4 kWh thermal per apartment per day. That is more than a third of the 6.5 kWh thermal per day implied by the RECS figure.
100-unit building conversion
For a 100-unit building:
Storage context
The daily figure is the quantity a storage bank is sized to carry across a charging window. The 700-gallon Thermal Tank module stores 108 kWh thermal at a 35 C, or 63 F, delta according to the TEHQ specification sheet. At that delta, six modules hold a full day and two hold a morning peak.
The sizing method for peak hot water demand converts a building's own draw profile into module count, and the sizing calculator performs the arithmetic.
What a heat pump draws for the same hot water
The electricity a heat pump water heater consumes is the thermal output divided by its coefficient of performance. The Statewide Electric Emerging Technologies Program field study of five California multifamily buildings of 8 to 50 units reported average system COPs from 1.9 to 3.2.
Applied to the per-apartment hot water figure:
Comparison with Northeast resistance water heating
Against the 1,524 kWh electrical per apartment reported by RECS 2020 Table CE4.7 for resistance water heating in Northeast apartments, a heat pump at COP 2.5 uses 38 percent less electricity and a heat pump at COP 3.2 uses 51 percent less.
The comparison is between a survey average and a calculated value and is indicative only. Thermal kWh is not electrical kWh at any point in this arithmetic; COP is the divisor every time.
System COP and storage
The system COP in the California study is the whole-plant figure, including recirculation losses and resistance backup where present. The study notes that at sites with resistance backup, the element barely ran, but its standby losses lowered system efficiency. It also found that recirculation return temperature control allowed the lower-cost piping arrangement to perform as well as the alternative.
Those findings establish the design reason for storage: a tank lets the heat pump run at steady output into a cold return rather than cycling against a warm one. The single-pass versus multi-pass article explains that operating condition.
Commercial buildings: lodging is the outlier
CBECS 2018 Table E1 reports the following major-fuels consumption by end use:
Lodging energy intensity
The EIA lodging summary reports a mean of 85.7 kBtu per square foot across all end uses. Water heating is the highest single end-use intensity at 16.5 kBtu per square foot. Office equipment is the lowest at 0.4 kBtu per square foot.
EIA reports these values in “MBtu” per square foot, which EIA uses to mean thousand Btu. This article writes the values as kBtu per square foot.
EIA describes water heating and space heating as each about 20 percent of lodging end-use energy. Lighting, cooling, refrigeration, computing and office equipment are each 10 percent or less. Lodging used 342 trillion Btu of electricity and 224 trillion Btu of natural gas in 2018.
100,000-square-foot hotel conversion
For a 100,000-square-foot hotel, 16.5 kBtu per square foot is 1.65 billion Btu of water heating fuel per year. At an assumed 80 percent efficiency, that is 1.32 billion Btu delivered, or 387,000 kWh thermal per year and about 1,060 kWh thermal per day.
The hotel design article covers the laundry and guestroom peaks that shape the daily load. The solar hot water article covers preheat for the same load.
Why the commercial average differs
The commercial average of 5.1 percent is why offices, retail and warehouses rarely electrify hot water first. In building types where the share is 7 to 19 percent, the water heating plant is a measure engineers reach for before the central plant.
Restaurants are a special case. A 7.1 percent water heating share of a high energy intensity is still a large absolute load, and the commercial kitchen article explains why the health department worksheet, not the energy survey, sizes it.
From share to electrical load
The share determines priority. The daily kWh thermal determines the plant.
The PNNL-34990 Rev. 2 technical brief proposes a service water heating performance path for ASHRAE 90.1-2022 and tabulates recovery and storage values by building type for its target systems. These are proposed code-overlay sizing values, not measured consumption and not an adopted standard.
Recovery rate compared with daily load
A 100-unit building's recovery-rated plant at 5 gallons per hour per unit heats 500 gallons per hour. At a 70 F temperature rise, that is 291,900 Btu per hour, or 85.5 kW thermal, whether or not anyone is showering.
The same building's daily load from the RECS figure is 649 kWh thermal. Spread over a 16-hour charging window, the average required output is 40.6 kW thermal.
A heat pump sized to the daily load with storage is roughly half the capacity of one sized to the recovery rate. The electrical consequence is worked in the companion article on heat pump water heater sizing, and the capacity-versus-storage curve is addressed there.
Using the share under a building performance standard
Building performance standards express targets as energy use intensity or greenhouse gas intensity per square foot. A multifamily building where hot water is 29.9 percent of site energy can reach a 13 percent site EUI reduction from the water heating measure alone if the heat pump cuts that end use by more than 43 percent on a site energy basis.
At an 80 percent boiler baseline and a COP of 2.5, the calculation is:
Building type changes the measure priority
An office at a 5.1 percent water heating share cannot reach the same building-level result from hot water alone. Hot water is therefore a headline item in a multifamily plan and a line item in an office plan.
The Washington Clean Buildings article works the arithmetic under one specific program. A companion article addresses New York's Local Law 97. The applicable building performance standard and regulatory requirements must be evaluated separately.
Caveats that change the share
Frequently Asked Questions
What percent of a multifamily building's energy goes to hot water?
EIA RECS 2020 Table CE3.1 reports 10.1 million Btu per household for water heating in apartments in buildings with five or more units, or 29.9 percent of 33.7 million Btu of total site energy. For all U.S. homes, water heating is 14.1 of 76.8 million Btu per household, or 18.3 percent.
How much energy does a hotel use for hot water?
EIA CBECS 2018 Table E1 reports 111 trillion Btu of water heating out of 598 trillion Btu for lodging, or 18.6 percent. The EIA lodging summary gives a water heating intensity of 16.5 kBtu per square foot against 85.7 kBtu per square foot of total energy.
How many kWh does hot water use per apartment per year?
Converting the RECS figure at an assumed 80 percent boiler efficiency gives 2,368 kWh thermal per apartment per year. A heat pump at an assumed COP of 2.5 supplies that thermal output for about 947 kWh electrical. RECS 2020 Table CE4.7 reports 5.2 million Btu, or 1,524 kWh electrical, for resistance water heating in Northeast apartments.
How many gallons of hot water does an apartment use per day?
Derived from the RECS energy figure at an assumed 80 percent efficiency and a 70 F temperature rise, the result is about 38 gallons per day. This includes distribution and standby losses. The survey does not measure gallons.
What share of hospital energy is water heating?
EIA CBECS 2018 Table E1 reports 33 trillion Btu of water heating out of 436 trillion Btu for inpatient health care, or 7.6 percent. The hospital central plant article covers where storage fits in that load.
Is hot water a bigger share in apartments than in houses?
Yes. Apartments use 28 percent less water heating energy per household than the U.S. average but 74 percent less space heating energy. Water heating therefore rises from 18.3 percent of total energy in all U.S. homes to 29.9 percent in apartments in buildings with five or more units.
What is the water heating energy intensity for commercial buildings?
EIA reports 343 trillion Btu of water heating across 6,787 trillion Btu of commercial major-fuels consumption in 2018, or 5.1 percent. Lodging is 18.6 percent, food service is 7.1 percent and education is 5.6 percent. EIA publishes intensities per square foot by building type on its CBECS pages.
The bottom line
Hot water is 29.9 percent of site energy in a U.S. apartment building and 18.6 percent in a hotel, against 5.1 percent across commercial buildings as a whole, according to EIA's current surveys.
Converted, the apartment figure is about 2,370 kWh thermal per unit per year, 650 kWh thermal per day for 100 units, and roughly 950 kWh electrical per unit per year through a heat pump at an assumed COP of 2.5.
Those values rank the measure, establish the storage load and set the electrical load. They come from national averages; the building's own meter data replaces them at the first opportunity.
