How Many kWh Does an EV Need Overnight

Every wall box is sold on speed. Miles of range per hour, kilowatts, minutes to fill a pack from empty — the specification sheet picks whichever of those flatters the unit, and the showroom conversation follows. But a car sitting in a driveway overnight is not trying to finish quickly. It is trying to put back a particular number of kWh before you leave in the morning, and it usually has eleven or twelve hours in which to do it, which is a far easier assignment than the brochure is answering.

Put a real commute through it. Twenty-three miles to work and twenty-three home, in a 2026 Chevrolet Equinox EV AWD that EPA rates at 32.6 kWh per 100 miles, comes to 15 kWh to replace. An ordinary 120-volt outlet at 12 amps — about 1.44 kW — moves that in ten and a half hours, so a car plugged in at seven in the evening is done before six. A 240-volt circuit at 32 amps finishes the same job in one hour and fifty-seven minutes and then sits there for another nine hours with nothing to do. Both cars leave full. Only one of them needed a permit, a conductor run and a licensed electrician to get that way.

Which means the purchase is decided by two quantities the salesman has no access to: how much energy your driving actually removes, and how many hours the car stands still afterwards. Neither is a property of the equipment. Both take about a fortnight to establish, and the rest of this page is how to establish them.

The miles you drive, not the range you bought

Everything downstream of this is a multiplication, so an error here scales into every other number on the page. National figures are worth knowing mostly as a sanity check on instincts that will be wrong in a predictable direction. Oak Ridge National Laboratory's tabulation of the 2022 National Household Travel Survey, published by the Alternative Fuels Data Center and last updated January 2026, puts 51.7 percent of all US vehicle trips under six miles and only 6.9 percent over thirty. DOE's own read of the same survey puts the "average daily mileage" of the single vehicle in a one-vehicle household at about 50 miles, and the first vehicle in a two-vehicle household at nearly 60. A second DOE page reports that electric vehicles in the same survey covered 12,400 miles a year, which is 34 miles for every day on the calendar. Those two figures are not in conflict and they are not measuring the same thing — 50 miles a day every day would be 18,000 a year — but neither page spells out its denominator, so do not size a circuit on either.

None of those is your number. Get yours the boring way: reset a trip meter, drive normally for fourteen days including both weekends, then divide the total by fourteen. Two weeks, because one week can be dominated by a single unusual Saturday and you want that Saturday diluted rather than excluded.

One correction catches nearly everyone. Your commute is not your daily mileage. Groceries, the school run, the hardware store on the way home — those are the trips under six miles that make up half of everything, and they are exactly the ones you leave out when estimating. The odometer counts them. You will not.

Your car's number is somewhere between 24 and 44

Guessing at this one gets expensive, because the spread across current EVs is nearly two to one, and nothing about a car's size or price tells you reliably where in that range it sits. EPA publishes combined electricity consumption in kWh per 100 miles for every model, and fueleconomy.gov exposes it as a field called combE through its public web service. I pulled these on 21 August 2026, all 2026 model year:

All figures in kWh per 100 miles. The same 46-mile day is 11.2 kWh in the Model 3 and 20.4 kWh in the R1T. Same commute, same driveway, nearly double the energy — which is why an article telling you that most people need about 10 kWh a night is telling you about a car that may not be yours.

Now look at the last two columns, because they invert what gasoline taught you. Highway consumption is higher than city consumption for every vehicle in that table. Aerodynamic drag rises with the square of speed and there is nothing to recover through regenerative braking on an open interstate, so the stop-and-go commute is the cheap one. The OPTIQ spends 26 percent more energy per mile on the highway cycle than in the city.

Which means: if your commute is thirty minutes of 70 mph each way, use the highway column, not the combined one. If it is surface streets and traffic lights, city is closer. Combined is a weighted blend, and it is the right default only when your driving is a genuine mix.

Two more places to find your figure. The window sticker on a new car carries it directly, in a box labelled in kWh per 100 miles. And your car's own trip computer will give you a lifetime average in mi/kWh or kWh/100 mi — the most accurate number available to you, with one caveat that deserves its own section.

Wall kilowatt-hours and battery kilowatt-hours are not the same quantity

This is the step where people apply a correction that has already been applied, and overbuild by fifteen percent for their trouble.

Charging is not free of loss. Energy is converted from AC to DC and some of it leaves as heat, in the cable and in the car's onboard charger. The useful question is which of your two available efficiency numbers already accounts for that.

EPA answers it plainly. On its fuel economy and EV range testing page (read 21 August 2026), it states that MPGe values "include charging losses," that they "assume level 2, alternating current (AC) charging and account for losses from the charging cable ... and the on-board vehicle charger," and that this "moves the measurement from the vehicle to the outlet in the wall." The kWh/100 mi figures in the table above are wall-side numbers. Do not divide them by an efficiency factor.

Your dashboard readout is the opposite. It measures energy leaving the battery, so it is systematically optimistic about what your meter will register. If you are working from the car's own mi/kWh, divide by something in the region of 0.85 to 0.90 to get back to the wall.

You can see the size of that gap inside EPA's own data. The 2026 LEAF is listed with a 75 kWh pack in its model name, an EPA range of 288 miles, and consumption of 30.0 kWh/100 mi. Covering the full range therefore draws 288 × 0.30 = 86.4 kWh from the wall to move 75 kWh of nominal pack capacity. Roughly 11 kWh, or 13 percent, never became stored energy — part charging loss, part the difference between nominal and usable capacity, and this arithmetic cannot separate the two. Treat 13 percent as a ceiling on the loss rather than a measurement of it.

There is also a smaller, steadier drain that neither number captures. The car wakes up, runs pumps, conditions the pack, phones home. Cabin preconditioning on a timer draws from the wall as well when the car is plugged in, which is usually what you want it to do. All of that lands on your meter and none of it appears in your consumption estimate.

Divide by the hours the car is parked, not by the hours in the night

The two quantities from the sections above finally meet here, and this is the division that turns them into a breaker size.

Required power (kW) = overnight kWh ÷ available hours

Fifteen kilowatt-hours over a thirteen-hour dwell from 6 p.m. to 7 a.m. is 1.15 kW. That is less than a hair dryer, and less than a 120-volt cordset delivers. Fifteen kilowatt-hours over a six-hour window is 2.5 kW, which needs 240 volts but only about 10 amps of it.

The six-hour case is the one worth thinking about, because for a lot of people the real dwell window is not the night. It is the off-peak block on a time-of-use rate — often something like midnight to 6 a.m. — and if you have set a charging timer to stay inside it, you have voluntarily cut your available hours roughly in half. Every hour removed raises the power required in direct proportion. Read the actual off-peak hours off your utility's tariff sheet before you decide how many hours you have.

From required power to a circuit is two short steps. Amps at 240 volts is kW × 1000 ÷ 240. And the NEC treats vehicle charging as a continuous load, so the branch circuit gets sized at 125 percent of that — the reason a 48-amp unit needs a 60-amp circuit, which the wall box article works through in detail. Whether your panel has room for the result is a separate calculation you can run from your own utility bill under NEC 220.87.

Be ready for the load calculation to ignore your modest commute entirely. Recent NEC editions include a provision requiring EV supply equipment to be counted at a fixed minimum — 7,200 VA in the editions that introduced it — or the equipment's nameplate rating, whichever is larger. Section numbers and the exact treatment move between editions, and energy management systems change the picture again. That is a rule about what the paperwork must assume, not about what your car will draw, and the two can disagree by a wide margin. Ask your permit office which edition they enforce today and write the date beside the answer.

I am not an electrician and none of this is an instruction to install anything. Conductor sizing, breaker selection and the final word belong to your licensed installer and the authority having jurisdiction.

What the 80 percent ceiling costs you, in miles

Most EVs ship with a settable daily charge limit, and a great deal of advice says to leave it somewhere below 100 percent for routine charging. Fine. The question here is narrower: does that ceiling change the kWh figure you just calculated? For ordinary commuting, no.

A 20-to-80 percent window is 60 percent of the pack. Convert it to miles by multiplying EPA range by 0.6:

• Tesla Model 3 Standard RWD, 321 miles → about 193 miles inside the window
• Chevrolet Equinox EV AWD, 307 miles → about 184 miles
• Nissan LEAF 75 kWh, 288 miles → about 173 miles
• Rivian R1T Quad Max, 374 miles → about 224 miles

Below those distances the ceiling is invisible to you, because daily charging replaces what you spent rather than filling from empty. You leave at 80 percent, come home at 65, go back to 80 overnight. The 20 percent floor never comes into play either.

It starts to matter when one day's driving approaches the width of the window, or when two heavy days arrive back to back without a full night between them. That is also the point at which the honest answer is to raise the limit for that trip and put it back afterwards, which every car offering the setting lets you do.

Your manual is the authority on where the limit belongs for your pack, and chemistry changes the answer — LFP packs are commonly treated differently from nickel-based ones, and some manufacturers ask for a periodic full charge to keep the state-of-charge estimate honest. That calibration behaviour is worth understanding alongside how to read your battery's state of health.

The Saturday that breaks it, and the one that only looks like it does

Two hundred miles on a Saturday, in a car rated at 30 kWh/100 mi, is 60 kWh. At 7.7 kW that is seven hours and forty-eight minutes — one night, comfortably, on a circuit sized for a commute. People assume the long day is what forces a big charger. Usually it does not.

What forces it is the sequence. Two hundred miles Saturday, two hundred Sunday, and Monday morning you are short. The other case is a genuinely long commute in a thirsty vehicle: 120 miles a day in something drinking 44 kWh/100 mi is 53 kWh every night, which is a real load on a small circuit.

Winter changes the multiplier rather than the method. Cold raises kWh per mile substantially, and the split between cabin heating and the battery itself decides by how much — the numbers behind winter range loss are worth running before you size anything in November. Practically: take your two-week average in the season you find hardest, or apply your worst month's consumption to your average distance.

And there is a pressure valve a home circuit does not have to duplicate. A DC fast charger, ten minutes out of your way, four or five times a year, covers the outlier days for far less than a permanently larger circuit. Sizing a house for the rarest week of the year is how people end up paying for a 60-amp circuit to answer a 40-mile day that a 120-volt outlet already covered.

Commute distance, hours needed, and the circuit each one implies

At 30 kWh per 100 miles — near the middle of the table above; substitute your own figure and the whole column scales:

Power assumed: 1.44 kW, 3.84 kW and 7.68 kW respectively. At 125 percent, a 16-amp unit points at a 20-amp circuit and a 24-amp unit at a 30-amp one.

The last column is the finding. A 160-mile day — more than four times the 34 miles a day implied by DOE's annual EV figure — is covered by 24 amps. The 40-amp and 48-amp units that dominate every comparison chart appear nowhere in that table, because nothing in it needs them. They still make sense for two cars sharing one supply, for a vehicle that genuinely accepts 19.2 kW, or when the trench is open once in your life and the larger conductor costs almost nothing extra. They rarely make sense as an answer to "how far do I drive."

Start the trip meter tonight. That is the whole preparation: fourteen days of ordinary driving, weekends included, divided by fourteen, multiplied by your car's kWh per 100 miles from whichever column matches how you actually drive, then divided by the hours between the plug going in and you pulling out of the driveway. It is a minute of arithmetic sitting on top of a fortnight of doing nothing differently, and it produces a kilowatt figure that no brochure can argue with because it came off your own odometer.

Bring that figure to the quote rather than a preference. Ask for the circuit it points at and the next size up priced on separate lines, and the conversation stops being about how fast the unit could go and starts being about what the larger conductor costs — which is a question with an answer on it in dollars, and usually a smaller one than the upsell implies.