Winter EV Range Loss: Heat vs Battery

Two drivers with the same car post the same complaint every January, and they are describing different problems. One lost about a tenth of their range. The other lost close to half. Both were parked outside overnight in the same weather.

The difference is not the battery. It is the thermostat.

Cold does two separate things to an electric car, and they are not the same size. It reduces what the pack can deliver, and it hands the cabin heater a job that a gasoline car gets for free from engine waste heat. Almost everything written about winter range mashes those two into one alarming percentage. Pulling them apart tells you which of them you can actually do something about.

Two AAA numbers, and the gap between them

The most useful winter range test ever published is also one of the oldest, because of how it was structured rather than how recent it is. In February 2019, AAA ran five battery electric cars on a chassis dynamometer in a temperature-controlled cell, following SAE J1634, at 20F, 75F and 95F. The part that matters: at 20F they ran each car twice, once with the HVAC system switched off and once with it holding the cabin at 72F.

With the heater off at 20F, combined range fell an average of 12 percent against the 75F baseline. With the heater on, it fell 41 percent. Both figures are in the report's own key findings, and the full methodology is in AAA's Electric Vehicle Range Testing report (PDF, read 20 August 2026).

So the split, in round terms, is 29 points of cabin heat sitting on top of 12 points of everything else. Heating is roughly seven-tenths of the problem.

Two things keep that from being a universal law. First, both figures are five-car averages. AAA published the per-vehicle spread as a chart rather than a table, so the defensible reading is that individual cars landed on both sides of 12 and 41, not that every car loses 41 percent. Second, the fleet was a 2018 BMW i3s, a 2018 Chevrolet Bolt, a 2018 Nissan Leaf, a 2017 Tesla Model S 75D and a 2017 Volkswagen e-Golf. Eight- and nine-year-old cars as of 2026, in other words. The direction has held up in every study since. The magnitudes belong to those five.

The other honest caveat is about the 12 percent itself. It is not a pure battery number. With the cabin HVAC off, a cold car still has to push through air about 11 percent denser than at 75F, which is just the ideal gas law comparing 266 K against 297 K. Gear oil and grease are thicker. Tire pressure has dropped a few PSI with the temperature. The pack's own thermal management may still be drawing power. Battery chemistry is inside that 12 percent. It is not all of it.

Heat is a kilowatt draw, so your speed decides the percentage

Here is the mechanism that makes winter range loss so inconsistent between drivers.

Cabin heat is measured in kilowatts and is largely indifferent to how fast you are going. Argonne National Laboratory's instrumented testing, published in the Department of Energy Vehicle Technologies Office's September 2024 program record on cold ambient temperature and BEV performance (PDF, record dated 12 September 2024, read 20 August 2026) put three model year 2019-2020 BEVs on the dynamometer at Argonne's Advanced Mobility Technology Laboratory across 0F to 95F ambient, cabin thermostatically held at 72F. HVAC power draw, battery thermal management included, came in between 1.8 and 3.7 kW at 0F and 20F depending on the heating hardware. AAA's report puts resistive cabin heating in the same neighborhood from the other direction, at typically 2 to 4 kW.

Propulsion, meanwhile, scales hard with speed. Nothing in the table below is a measurement. It is arithmetic on two invented round numbers, a car using 280 Wh per mile in mild weather and a flat 3 kW heater sitting on top of it, picked because they divide cleanly and not because any tested car posted them:

Those are round numbers holding per-mile consumption constant across speeds, which no real car does. A real highway figure is higher than a real city one. Treat the table as the shape of the effect, not as your car's spec sheet. And note the conversion people get wrong: a 43 percent increase in consumption is a 30 percent loss of range, not 43. Range is the reciprocal.

The measured data lands where that arithmetic says it should. Argonne's testing at 0F found range loss averaging about 50 percent, with a maximum of 59 percent on the urban cycle with ample idle time and a minimum of 39 percent on the highway cycle with none. DOE also notes that above roughly 40 mph, range impacts across different ambient temperatures begin to converge.

Which is why the two drivers complaining in January are both right. The commuter doing eight stop-start miles to a train station in traffic is paying the worst possible version of this bill. The person driving 70 miles of interstate is paying close to the best one.

What the battery itself actually costs you

The battery side of the ledger is real but smaller, and it shows up in three places rather than one.

Usable energy shrinks. Cold slows the chemical reactions in the cells, which lowers output voltage and available capacity. DOE describes this as the secondary reason for BEV winter range loss, with HVAC as the primary one, and notes it becomes pronounced at 0F and below rather than at 30F.

Regeneration gets limited. A cold pack cannot accept high charge current, and regenerative braking is charge current. Until the pack warms, some or all of the energy you would normally recover on a downhill or a long deceleration goes into the friction brakes as heat instead. Most cars display this: a dashed line on the power meter, a grayed-out battery icon, a message about reduced regenerative braking. AAA's report reproduces the line straight out of one test car's owner's manual: "Regenerative power may be limited when the high voltage battery is near full charge or cold. The regen battery icon will appear gray when limited." Watch for it on the first drive of a cold morning. It usually clears after a few miles.

Some of the heating you are paying for is not for you. Modern packs have their own heaters, and warming several hundred kilograms of cells is a real energy expense that lands on your consumption figure if it happens after you unplug. Run it from the wall while the car is still plugged in and it costs the pack nothing, which is the whole argument for preconditioning before you leave, and the reason Argonne measured a 9 to 20 percent cut in energy consumption on a 7.5-mile regulatory city cycle at 20F when the trip started preconditioned. The same warm-up buys you something quite different at a fast charger, where the binding constraint is how much current cold cells will accept rather than how far the energy goes, and the arithmetic there is about minutes rather than miles.

Heat pumps buy you the middle of winter, not the bottom of it

The DOE record measured a 38 percent reduction in total HVAC power draw at 20F for a car with a heat pump versus one with resistance heating only, and puts heat pumps at three to four times the efficiency of resistance heat in general terms. Tesla, Ford, BMW, Hyundai and Kia are all named as adopters.

Real-world data agrees on the direction and is vaguer on the size. Recurrent's winter range analysis (updated 19 November 2025, read 20 August 2026) draws its 2025/2026 winter study from more than 30,000 vehicles and reports that across 34 popular models, EVs retain an average of 78 percent of their maximum range near freezing and 70 percent at 20F. Its heat pump claim, though, is one sentence long: heat pump technology extends range by about 10 percent at 32F. No sample size, no model list, no method is published for that particular comparison, only an illustration of a single model with and without one. Those 30,000 vehicles belong to the 78 percent figure. They do not belong to the 10.

Now the awkward part. Two of AAA's five cars, the i3s and the Leaf, ran heat pumps backed by auxiliary resistive heaters, and the report went in explicitly hypothesizing that they would lose significantly less range than the three resistive-only cars. They did not. The conclusion is worth reading in the report's own words: "It was noted that the cabin heating mechanism did not significantly affect vehicle performance in terms of energy consumption, driving range and equivalent fuel economy. Heat pumps are largely ineffective within extremely cold environments. At 20F, auxiliary resistive heating was likely utilized to maintain cabin temperature."

Both findings are true, and the reconciliation is the coefficient of performance. A heat pump moves ambient heat rather than creating it, so as the outside air gets colder there is less of it to move. Around freezing, DOE puts it at three to four times better than resistive heating. Recurrent describes the far end of the band the way DOE does: the benefit diminishes as temperatures fall toward 0F, at which point a heat pump is about as efficient as a resistive heater and the car quietly switches over. Heat pump hardware has also improved considerably between 2018 and 2026, which is likely most of why the newer fleet data is kinder to it than AAA's cars were.

The practical version: if your winters sit mostly in the twenties and thirties, a heat pump is worth shopping for. If your winters are regularly below zero, budget your range as though you do not have one.

Splitting your own car's number in one evening

Published averages cannot tell you your split, because they cannot know your commute speed, your cabin setpoint, or where the car sleeps. Measuring it takes about an hour of doing very little.

Measure the heater by itself. Park the car, note the state of charge or the trip energy counter, set the climate control the way you actually drive with it, and sit there for 30 minutes without moving. Note the energy used and double it. That is your heating draw in kW at that outside temperature. Write the temperature down next to it, because the number is meaningless without it. Compare against the 1.8 to 3.7 kW band from the Argonne data.

Convert it to a per-mile figure. Divide those watts by your typical average speed, not your typical speed limit. A 3 kW heater over a commute that averages 30 mph is 100 Wh per mile added. If your car uses 280 Wh per mile in October, that is a 36 percent consumption increase and a 26 percent range hit from heat alone.

Then you subtract. Drive the same route you drove in the fall, note the winter Wh per mile, and take off the heater share you just worked out. What is left over is battery, air density, tires and drivetrain drag all rolled together, and on most cars it lands a lot closer to AAA's 12 percent than to the number in the headline that brought you here.

The last step is the one worth repeating each winter, and it is the only one that tests a decision rather than measuring a condition: drive that same route twice in the same weather, once preconditioned on grid power and once not. Whatever separates those two numbers is what the feature is worth on your commute, in your climate, with your driving. Not a five-car average from 2019. Not a fleet median. Yours.

One trap to avoid while you do this: do not compare a February dashboard estimate against the EPA range on the window sticker and call the difference degradation. The EPA figure comes out of a 75F test cell, cold range comes back in April, and mixing the two is the most common way owners talk themselves into believing their pack is failing. To actually check the pack, read state of health the way it is meant to be read on a warm, settled battery.

All of this changes what you should carry as reserve on a long winter drive, since the loss is worst in exactly the stop-and-go conditions that a delayed trip creates, which is the multiplier your route planner is quietly assuming away. It also shows up on the receipt. More kilowatt-hours per mile means more kilowatt-hours bought, and if the station bills by the minute rather than by energy delivered, a cold day charges you twice for the same weather, which is why the pricing model on the screen matters before you plug in.

The single most useful thing to do before the first hard freeze is that 30-minute parked test. One outside temperature, one number in kilowatts, written down. Everything else in winter range planning is arithmetic on top of that one measurement.