They are the same machine
An air conditioner takes heat out of the house and dumps it outside. It does this with a compressor, a refrigerant loop, an indoor coil and an outdoor coil. A heat pump has the same compressor, the same refrigerant, the same two coils — and one extra component, a reversing valve, which swaps which coil is which. Run it one way and heat leaves the house. Run it the other way and heat comes in.
That is the whole of the difference in hardware. It is why every heat pump sold is also an air conditioner, why they are rated on the same SEER scale for cooling, and why nobody sells a “heat pump for cooling” as a separate product. If you buy a heat pump you have bought an air conditioner that does something else as well.
The difference is entirely in winter
At the US average rate, on each appliance’s own stated assumption. The cooling column is identical on both rows and cancels — which is the finding, not a rounding artefact.
| What is installed | Cooling | Heating | A year |
|---|---|---|---|
| Heat pump, doing both jobs | $421 | $423 | $844 |
| Air conditioner plus electric resistance heat | $421 | $1,653 | $2,074 |
| Difference | $0 | $1,230 | $1,230 |
The cooling column is the same on both rows because it is the same machine doing the same job. Everything that separates the two options happens when the outside temperature drops, and at that point the air conditioner does nothing at all while the heat pump becomes the cheapest electric heating there is.
So “heat pump vs air conditioner” is not really a choice between two machines. It is a choice between a machine that cools, and the same machine with the winter half switched on. The question worth asking is the one underneath it: what heats the house today, and what would the heat pump be replacing?
What it is actually replacing decides everything
Against electric resistance heating — a furnace with strip heat, baseboards, a fan heater in every room — the case is overwhelming, and the figure above is the size of it. Resistance heating converts electricity to heat at 100%, which is its ceiling; a heat pump delivers two to four units of heat per unit of electricity because it moves heat rather than making it.
Against a gas furnace the arithmetic is different and this site cannot finish it, because half of it is a gas bill and every figure here is electricity. Switching raises the electricity bill and removes the gas one, and which way the total moves depends on the price of both fuels where you live. Anyone quoting you a single national answer to that has not done the sum.
Against nothing at all — a house with no cooling and no central heating — you are buying two things at once, and the comparison to make is against installing an air conditioner alone and leaving the heating as it is.
One more source of confusion is worth naming, because search engines suggest it in the same breath: a mini split is not a third option. A ductless mini split is a heat pump — the same compressor, refrigerant and reversing valve — packaged as one or more wall units instead of feeding ductwork. It is a different way of delivering the air, chosen when a house has no ducts or when individual rooms need separate control, and it changes the installation rather than the physics. Everything above applies to it unchanged.
Why the ratings look different when the machines are not
A heat pump and an air conditioner sitting beside each other in a showroom carry different-looking labels, and that is where most of the confusion starts. Both are rated for cooling on the same SEER2 scale, so those numbers are directly comparable and a heat pump has no inherent advantage or penalty there. The heat pump then carries a second number, HSPF2, which describes its heating season and which an air conditioner has no equivalent of because it does not heat.
Two units with the same SEER2 cool the same. If one is quoted as cheaper to run in summer than the other, the difference is in that rating or in how the system was sized and installed — a unit too large for the house short-cycles and performs below its label whatever is printed on it. None of that is a property of being a heat pump.
The same caution applies in the other direction. A high HSPF2 is measured under a standard heating season that may look nothing like yours: the rating assumes a particular distribution of outdoor temperatures, and a house in Minnesota and one in Georgia are not sampling the same distribution. It is a fair way to rank two heat pumps against each other and a poor way to predict your own bill.
What the salesman is not wrong about
Two caveats are real rather than sales resistance. The first is the balance point: below some outdoor temperature the heat pump cannot keep up and a backup resistance element takes over, at which point the running cost converges on the furnace it replaced. A cold-climate model and an auxiliary-heat lockout set correctly are what keep the saving real, and a thermostat calling for emergency heat on every recovery erases it.
The second is that a heat pump costs more to install than an air conditioner, because it is an air conditioner plus a reversing valve plus, usually, a bigger electrical service and a control upgrade. Whether that pays is your quote divided by the saving above, and it is the one number this site refuses to invent.
The full head-to-head against resistance heating, with a payback calculator that takes your own quote and your own rate, is at heat pump vs electric furnace.
Questions
Does a heat pump cool as well as an air conditioner?
Should I buy an air conditioner if I already have gas heat?
Is a heat pump worth it just for the cooling?
All guides · How the figures are computed · Break down your own bill