Heat pumps at −30°C · Made with makr.ai

The short answer

Yes, with conditions. A modern cold-climate heat pump keeps running and still beats electric baseboard well below −25°C, and in a 22-home U.S. and Canadian field trial the median efficiency was 1.9× resistance heat around −16°C.¹ But at −30°C output falls, and some popular models shut themselves off at −28°C to −30°C,¹⁰¹¹ so you need a backup that can carry the house on its own. Whether it saves money depends on what you burn now: against oil and electric resistance it usually wins; against cheap natural gas it usually doesn't.

1.9

Median field COP at −18 to −15°C, 22 homes¹

66%

Output left at −30°C vs +8°C, one 3-ton unit's spec sheet⁹

−28°C

Where a best-selling ductless model locks out¹¹

4.5

COP needed to tie a gas furnace on Minnesota prices (worked example)

01How cold changes a heat pump

A heat pump moves heat instead of making it. The colder the outdoor air, the harder that job gets, and two things fall at once: how much heat it can deliver (capacity) and how much heat you get per unit of electricity (COP).

A COP of 1 is what electric baseboard gives you: one unit of heat per unit of electricity. A COP of 2 means half the electricity for the same heat. The Pacific Northwest National Laboratory puts the problem with older equipment bluntly: a conventional heat pump's capacity falls below its rating once it drops under about 4°C (40°F), and only a fraction is left by the time temperatures reach the teens °F (roughly −7 to −10°C).¹ A literature review cited by the same team found existing air-source heat pumps averaging a COP of about 1.5 between −15°C and −30°C, with a range of 1.0 to 2.5.⁴

What “cold-climate” is supposed to mean

Three yardsticks are in use. Each one tests at −15°C (5°F). None of them tests at −30°C.

NEEP Cold Climate ASHP Specification v4.0 (effective 1 Jan 2023): a variable-capacity compressor, COP of at least 1.75 at −15°C running at maximum capacity, and published performance data down to the lowest temperature the manufacturer catalogues.²
ENERGY STAR Cold Climate: COP of at least 1.75 at −15°C, and capacity at −15°C of at least 70% of capacity at +8°C (47°F), checked with the unit's own controls.³
U.S. DOE Cold Climate Heat Pump Challenge (prototypes, 2021–2024): COP of at least 2.4 for systems up to 4 tons and 100% of rated capacity at −15°C, with an optional tier requiring operation down to −26°C.⁴

In practice, “cold-climate” means “proven at −15°C.” At −30°C you have to read the manufacturer's extended data yourself.

Efficiency (COP) vs. outdoor temperature

One manufacturer's published curve at full output, against efficiencies measured in real homes and test houses

Spec sheet, 3-ton ducted unit at max outputField measurementReference lines

Sources: spec-sheet curve from the Mitsubishi Zuba-Central PVA-A36AA4/PUZ-HA36NHA5 maximum heating performance table, no auxiliary heat⁹. Field points: DOE CCHP Challenge median, 0 to 5°F bin¹; NRCan Canadian Centre for Housing Technology ducted⁵ and mini-split⁶; Minnesota CEE ductless sites 6 and 8, 0 to −13°F daily averages plotted at the bin midpoint⁷; ORNL prototype in Fairbanks⁸. Threshold line: NEEP / ENERGY STAR²³. Field points come from different machines, homes and methods and are not directly comparable to each other.

Show the numbers

What real homes measured

The best recent evidence is DOE's Challenge field validation: 22 ducted prototypes from eight manufacturers, installed in homes and lab houses across the U.S. and Canada and monitored from late 2022 to late 2024.¹ Running on the heat pump alone, median COPs below −1°C (30°F) ranged from 1.6 to 2.7, and the median in the −18 to −15°C bin was 1.9, above the 1.75 cold-climate threshold. Homeowners' average heating satisfaction rose from 3.61 to 3.88 out of 5 after the switch.¹

Canadian government testing lines up. At NRCan's Canadian Centre for Housing Technology, a ducted cold-climate unit held a COP of 1.5 at −21.1°C (2013), and a mini-split measured about 2.3 at −15°C and close to 2.0 at −25°C (2014).⁵⁶

02What actually happens at −30°C

Three things happen together: output shrinks, efficiency drops toward resistance heat, and on some machines the compressor stops completely. You design around all three.

1. Output shrinks while your heat loss peaks

The published table for one 3-ton cold-climate ducted unit shows full output of 38,000 Btu/h down to −15°C, then 29,000 Btu/h at −25°C and 26,500 Btu/h at −30°C. That's about 66% of its 40,000 Btu/h output at +8°C.⁹ Your house needs the most heat at exactly that moment, so somewhere in between the lines cross. That crossing is the balance point, and below it something else has to supply the rest.

Heat-pump output vs. what a house needs

Maximum output of the same 3-ton unit (Btu/h) against an illustrative house heat-loss line

Unit's maximum output (spec sheet)Hypothetical house heat loss (assumption)

Source: Mitsubishi Zuba-Central PUZ-HA36NHA5 maximum heating performance table⁹. The house line is not data. It is a straight line from 0 at 18°C to 36,000 Btu/h at −30°C, drawn only to show how a balance point works. Your own line comes from a room-by-room heat-loss calculation. Lockout markers: newer PUZ-HA36NKA submittal (−30°C)¹⁰; MUZ-FS12NAH ductless (−28°C)¹¹.

2. Efficiency approaches (but stays above) resistance heat

At full output that same unit is rated at a COP of 1.26 at −30°C. At minimum output the same table shows COP 2.01, because a heat pump running gently is more efficient.⁹ Field evidence from very cold places:

Fairbanks, Alaska: an Oak Ridge National Laboratory prototype with two vapor-injection compressors, tested at the Cold Climate Housing Research Center, ran reliably down to about −34°C (−30°F), holding about 75% of rated capacity with no auxiliary heat. Average system COP, including defrost, was about 2.0 at −25°C and 1.8 at −33°C.⁸ It's a research prototype in a test facility, not something you can buy today, but it shows the engineering is solvable.
Minnesota: in a 2018 state-funded study of older-generation equipment, one ductless unit held a COP above 1.0 at −28°C (−19°F), and two ductless sites averaged daily COPs of 1.7 and 1.3 between −18°C and −25°C.⁷

3. The compressor may simply stop

Every heat pump has a low-temperature cut-out. On Mitsubishi's newer Zuba-Central submittal it's −30°C, and the unit doesn't restart until −25°C.¹⁰ The popular MUZ-FS12NAH ductless unit is rated down to −25°C and locks out at −28°C.¹¹ Even within one brand, published cut-outs range from −36.7°C to −10°C.²³ If your design temperature is −30°C, assume that on the coldest nights the backup may be carrying the whole house.

What the evidence doesn't cover. Data at true −30°C is thin. In the DOE trial only 8 of 22 homes logged 10 or more hours below −18°C (0°F), because winters were milder than normal.¹ In its coldest bins (−23 to −18°C), auxiliary heat was on about 25% of the time.¹ We found no large, independent field study of commercial units running unassisted for long stretches at −30°C. Treat anything below −25°C as spec-sheet territory, plus a few test houses and prototypes.

03What it costs to run

The fair comparison is the cost of one unit of heat actually delivered into the house. Here it is for one real place and season. Swap in your own prices.

Worked example · assumptions Place and season: Minnesota, 2025–26 heating season, EIA data. Electricity 16.25¢/kWh, the residential average for Jan–Jul 2026.¹⁴ Natural gas $10.65 per thousand cubic feet, the simple average of Oct 2025–Mar 2026 monthly residential prices.¹⁵ Propane $1.95/gal and heating oil $3.63/gal, averages of 26 weekly residential prices from 6 Oct 2025 to 30 Mar 2026.¹⁶¹⁷ Energy content: 3,412 Btu/kWh, 1,036 Btu/ft³ gas, 91,452 Btu/gal propane, 138,500 Btu/gal oil.¹⁸ Efficiencies: gas and propane furnaces at 97% (ENERGY STAR U.S. North), oil furnace at 87% (ENERGY STAR),¹⁹ resistance at 100%, and a heat pump at a seasonal COP of 1.9 (low end of Minnesota modeled annual heat-pump COPs) or 2.5 (just under measured annual COPs of 2.51–2.78).⁷

Formula: cost per MMBtu delivered = price ÷ MMBtu per unit ÷ efficiency. This excludes fixed monthly charges, equipment cost, maintenance and cooling.

Cost of 1 million Btu of delivered heat

Minnesota residential prices, 2025–26 heating season (US$ per MMBtu). Shorter is cheaper.

Calculated from EIA prices¹⁴¹⁵¹⁶¹⁷ and EIA heat contents¹⁸, with the efficiencies listed above⁷¹⁹. This is a worked example, not a forecast. Inputs and formula are in data.json.

Show the numbers

What the example shows. Divide the price of electricity per MMBtu by a competitor's delivered cost and you get the COP a heat pump needs to break even. On these prices that's 1.6 against oil, 2.2 against propane and 4.5 against natural gas. Real cold-climate units clear the first easily, clear the second in a good installation, and don't reach the third on a seasonal basis. That matches NRCan's national modelling, which found cold-climate heat pumps cheaper to run than electric resistance or oil furnaces “in all parts of Canada,” while in its scenario that kept gas alongside the heat pump, modelled bills rose by $100–$500 a year in Ontario, Manitoba, Alberta and colder parts of B.C.¹³

The biggest single lever is your electricity rate. Ask your utility whether it has a heat-pump, off-peak or dual-fuel rate, and if it does, rerun the arithmetic with it.

04The installation decides the outcome

The same machine can be a success or a disappointment depending on five decisions.

Sizing: to the heat load, not the air-conditioning load

NRCan's sizing guide describes four approaches, from cooling-first to sizing on the full design heating load. It notes that “for most climate zones in Canada” a cold-climate unit is needed when the heat pump is the main heat source.¹² It strongly recommends checking the manufacturer's extended low-temperature data, not just the headline rating.¹² Insist on a room-by-room heat-loss calculation (Manual J in the U.S.; all 22 DOE Challenge sites had one¹).

Ducted vs. ductless

Ducted units replace a furnace and heat every room. But NRCan warns that existing furnace ducts often can't move enough air for a heat pump sized to the full load unless the envelope has been improved.¹² Ductless heads are efficient and flexible but heat zones, not whole houses. In the Minnesota study, the ductless units kept running down to −25°C while the ducted ones had been set to hand off to the furnace at −12°C (10°F).⁷

Backup strategy

If the heat pump can shut off above your design temperature, NRCan's guide says the backup must deliver 100% of the design heat load on its own.¹² Two common setups:

Keep the furnace (dual fuel). Set the switchover by cost, or let the furnace take over below the heat pump's cut-out. In Minnesota this cut propane use by 64%, from about 967 to 346 gallons a year.⁷
Electric resistance. Staged strips or baseboards. The DOE prototypes used three or four stages, and most only ever needed the first (about 5 kW).¹

Defrost and snow

Outdoor coils frost up and have to be defrosted. In the DOE trial, defrost took a median 5–8% of energy between −15°C and +2°C.¹ In Minnesota, removing defrost would have improved annual savings by 8.3%.⁷ Mount the outdoor unit above your snow line, make sure it drains, and ask whether the model has a base-pan heater (the MUZ-FS12NAH does¹¹).

The envelope

Every bit of heat loss you remove moves the balance point colder and shrinks the backup. NRCan lists new efficient homes and deep retrofits among the cases where a heat pump can be the main heat source.¹² Air sealing and attic insulation are usually the cheapest way to make a heat pump work at −30°C.

05Incentives: check before you sign

Programs changed a lot in 2025–26. As of this report:

U.S. federal 25C tax credit: the IRS says the Energy Efficient Home Improvement Credit applied to property placed in service “before December 31, 2025” (IRS page reviewed 28 April 2026). A heat pump installed in 2026 doesn't qualify.²⁰
Canada Greener Homes Grant and Loan: both closed. The Oil to Heat Pump Affordability program took its last applications on 31 July 2026 and is moving to Environment and Climate Change Canada. NRCan points low- and median-income households to the Canada Greener Homes Affordability Program (page modified 4 Aug 2026).²¹
State, provincial and utility rebates still exist in many places and change often. In the U.S., search DSIRE, the incentives database run by N.C. State University.²² In Canada, check your province and your electric utility. Many programs require a NEEP-listed or ENERGY STAR Cold Climate model²³ and a pre-approval before installation.

We haven't listed dollar amounts because they vary by address and change often. Get them in writing from the program.

06Decision checklist

Take this to every quote.

Know your design temperature and your heat loss.Get a room-by-room heat-loss calculation at your local design temperature (−30°C here). No calculation, no quote.
Demand extended data to −30°C.Ask for capacity and COP at −25°C and −30°C from the manufacturer's table, not the brochure. Confirm the model is on the NEEP list or ENERGY STAR Cold Climate certified.
Find the cut-out temperature.Ask where the compressor stops and where it restarts. If it stops above −30°C, plan for a backup sized for 100% of the load.
Mark the balance point on a graph.Ask the installer to plot the unit's output against your heat loss. Below the crossing, how many kW of backup cover the gap?
Choose the backup on purpose.Keep an existing furnace (dual fuel) or add staged electric heat. Check your electrical panel can take the heat pump plus backup.
Run your own cost numbers.Use your actual rates in the formula above. Break-even COP = electricity cost per MMBtu ÷ current fuel's delivered cost. Ask your utility about heat-pump or off-peak rates.
Fix the cheap heat loss first.Air sealing and insulation lower the balance point and the size (and cost) of everything else.
Plan for snow and defrost.Outdoor unit on a stand or wall bracket above the snow line, clear drainage, base-pan heater, out of roof drip lines.
Confirm incentives before signing.Written pre-approval where required. Don't count on the U.S. 25C credit or the closed Canadian grants.
Set the thermostat's auxiliary-heat settings.The DOE team saw switchover temperatures vary between homes with identical equipment, most likely because of thermostat setup and occupant behavior.¹ Ask the installer to show you the settings before they leave.

07Sources

We opened every source listed here while preparing this report (September 2026). Where a source didn't give a number, the report says so rather than estimate one.

Mendon V. et al., Performance Results from DOE Cold Climate Heat Pump Challenge Field Validation, PNNL-37127, Pacific Northwest National Laboratory, Jan 2025. pnnl.gov/…/PNNL-37127.pdf
Northeast Energy Efficiency Partnerships, Cold Climate Air Source Heat Pump Specification (Version 4.0), effective 1 Jan 2023. neep.org/…/version_4.0_final.pdf
ENERGY STAR, Heat Pump Equipment Key Product Criteria (Cold Climate). energystar.gov/products/air_source_heat_pumps/key-product-criteria
Mendon V. et al., “Rising Up to the Challenge: Cold Climate Heat Pumps in the Field,” ACEEE Summer Study on Energy Efficiency in Buildings, 2024. aceee.org/…/ssb24/…Cold Climate Heat Pumps in the Field.pdf
Natural Resources Canada, Cooling and Heating Season Performance Assessment of a cold climate air source heat pump at the Canadian Centre for Housing Technology, Sep 2013. natural-resources.canada.ca/…/cold-climate-air-source-heat-pump-canadian-centre-housing-technology
Natural Resources Canada, … Performance Assessment of a mini-split cold climate air source heat pump at the Canadian Centre for Housing Technology, Sep 2014. natural-resources.canada.ca/…/mini-split-cold-climate-…-housing-technology
Schoenbauer B., Bohac D., Haynor A. et al., Cold Climate Air Source Heat Pump, Center for Energy and Environment, Minnesota Dept. of Commerce CARD final report, 2018. mncee.org/…/86417-Cold-Climate-Air-Source-Heat-Pump-(CARD-Final-Report-2018).pdf
Hu Y., Shen B., Chen Y. (Oak Ridge National Laboratory), “Experimental Evaluation of a High-Performance Cold-Climate Heat Pump Using Tandem Vapor Injection Compressors,” ACEEE Summer Study, 2026. aceee.org/…/experimental-evaluation-of-a-high-performance-cold.pdf
Mitsubishi Electric Sales Canada, Zuba-Central Heating Performance, PVA-A36AA4 / PUZ-HA36NHA5 (maximum and minimum tables, without auxiliary heat). mitsubishitechinfo.ca/…/Zuba_36_Heating_Performance.pdf
Mitsubishi Electric Sales Canada, Submittal Data: PVA-A36AA7 & PUZ-HA36NKA, 2024. mitsubishitechinfo.ca/…/SB_PVA-A36AA7_PUZ-HA36NKA_202401.pdf
Mitsubishi Electric Sales Canada, Submittal Data: MSZ-FS12NA-U1 & MUZ-FS12NAH-U1, 2024. mitsubishitechinfo.ca/…/SB_MSZ-FS12NA-U1_MUZ-FS12NAH-U1_202401.pdf
Natural Resources Canada / CanmetENERGY, Air-Source Heat Pump Sizing and Selection Guide, Version 1.0, 21 Dec 2020. natural-resources.canada.ca/…/ASHP Sizing and Selection Guide (EN).pdf
Natural Resources Canada, Cold-climate air source heat pumps: Assessing cost-effectiveness, energy savings and greenhouse gas emissions reductions in Canadian homes. natural-resources.canada.ca/…/cold-climate-air-source-heat-pumps-assessing-cost-effectiveness…
U.S. EIA, Electric Power Monthly, Table 5.6.B, residential average price, year-to-date through July 2026 (released 24 Sep 2026). eia.gov/electricity/monthly/…t=epmt_5_6_b
U.S. EIA, Minnesota Price of Natural Gas Delivered to Residential Consumers, monthly. eia.gov/dnav/ng/hist/n3010mn3m.htm
U.S. EIA, Minnesota Propane Residential Price, weekly (heating season 2025–26). eia.gov/dnav/pet/hist/…W_EPLLPA_PRS_SMN_DPG
U.S. EIA, Minnesota No. 2 Heating Oil Residential Price, weekly (heating season 2025–26). eia.gov/dnav/pet/hist/…W_EPD2F_PRS_SMN_DPG
U.S. EIA, British thermal units (Btu), Energy Explained. eia.gov/energyexplained/units-and-calculators/british-thermal-units.php
ENERGY STAR, Furnaces Key Product Criteria. energystar.gov/products/furnaces/key_product_criteria
Internal Revenue Service, Energy Efficient Home Improvement Credit (page reviewed 28 Apr 2026). irs.gov/credits-deductions/energy-efficient-home-improvement-credit
Natural Resources Canada, Canada Greener Homes Initiative (modified 4 Aug 2026). natural-resources.canada.ca/…/canada-greener-homes-initiative
N.C. Clean Energy Technology Center, DSIRE: Database of State Incentives for Renewables & Efficiency. dsireusa.org
Mitsubishi Electric Sales Canada, Hyper-Heat (H2i) Cold Climate Heat Pumps. mitsubishielectric.ca/en/hvac/home-owners/zuba