EV Towing Range Loss: What a Roof Box Really Costs

Chevrolet's 2026 Silverado EV page leads with two figures side by side: Up to 478 miles GM-estimated range and Up to 12,500 lbs. Max available towing (checked 5 September 2026). Both are true. They are not true at the same time, and nothing on the page connects them, because nothing could. The range figure describes a truck. The towing figure describes a limit. What happens in between depends on an object Chevrolet has never seen: your trailer.

That gap is where trips go wrong. Not because towing is mysterious, but because the number people substitute for the missing one is usually somebody else's number, measured on somebody else's rig.

A tow rating is a stability limit, not an energy limit

The rating in your manual answers a question about brakes, frame loads, cooling and control at speed. It says nothing about kilowatt-hours. Treat it as a hard ceiling on mass and then stop asking it anything else.

Two things about that ceiling surprise people. The first is that it is a property of a specific build, not of a model. Rivian's maximum towing capacity page lists a Gen 1 R1T (2022 to 2024) at 11,000 pounds across every battery and drivetrain combination, and a Gen 2 R1T (2025 and later) at 7,700 pounds with the Standard or Large pack, 11,000 only with Max. Same badge, same bed, different number.

The second is that the hardware on the ball changes the rating outright. The same page states that "a weight distributing hitch is required to achieve the maximum towing capability of R1S and R1T," and that without one the capacity is 5,000 pounds (read 5 September 2026). That is not a safety suggestion. It is the number that applies to you.

Payload usually bites first on a family trip. Tongue weight, the download the trailer puts on the ball, counts against payload rather than against the tow rating, and it shares that budget with every human and duffel bag aboard. Rivian's maximum payload page is worth reading for one line: "Check your vehicle's door sticker to see its specific payload value." It also shows the trade nobody advertises. A Gen 2 R1T Dual Large is rated at 2,172 pounds of payload; the Max pack version, with more range, is rated at 1,764 (both read 5 September 2026).

And some EVs barely acknowledge the question. Kia's 2026 EV9 page advertises "an available towing capacity of 5,000 lbs.", while Chevrolet's 2027 Equinox EV page does not contain the words tow, trailer, hitch or haul anywhere on it (both checked 5 September 2026). A hitch vendor will happily sell you a receiver for a car the manufacturer never rated. The receiver is not the rating.

Frontal area is the number nobody prints

On a flat highway at cruise, the dominant cost of a trailer is not its weight. It is the hole it punches in the air, and drag rises with the square of speed. fueleconomy.gov puts it plainly on its energy-flow page for electric cars: a vehicle expends energy moving air out of the way, "less energy at lower speeds and more as speed increases," and "this resistance is directly related to the vehicle's shape and frontal area."

Which explains the ranking that trips up anyone reasoning from weight alone. A 3,000-pound flat utility trailer riding in the tow vehicle's wake can cost less than a 1,500-pound box trailer standing proud of the roofline. Height and width beat the scale ticket. A rooftop cargo box weighs almost nothing, sits in the fastest clean air the vehicle meets, and adds frontal area exactly where it hurts. A hitch-mounted tray of identical mass, tucked into the wake, barely registers.

Weight has its own bill, but it comes due elsewhere: on grades, where you are paying potential energy, and in stop-and-go, where you are paying to accelerate mass. Those are the same two places the margin arithmetic for a route already told you to compute. This article is about what a trailer does to the flat parts in between.

The federal test everyone is quoting, and why you cannot borrow its number

The published percentages everyone cites trace back to one place. fueleconomy.gov's driving-efficiently page says a large, blunt roof-top cargo box "can reduce fuel economy by around 2% to 8% in city driving, 6% to 17% on the highway, and 10% to 25% at Interstate speeds (65 mph to 75 mph)," while rear-mount boxes or trays cost "only 1% or 2% in city driving and 1% to 5% on the highway" (read 5 September 2026).

Follow the footnote and you reach Oak Ridge National Laboratory's SAE 2014-01-1614, and its abstract is worth reading in full. A compact sedan and an SUV were put through SAE J2263 coastdown procedures with each accessory fitted, the coastdowns were used to derive new dynamometer coefficients, and the vehicles were then run on standard cycles and at steady highway speeds. The rooftop box penalty on the sedan reached 25 to 27 percent at higher speeds. The SUV pulling a 3,500 pound enclosed cargo trailer showed penalties "ranging from 30%, for the city cycle, to 50% at 80 mph."

Now read the rest of that sentence, because it is the reason you cannot carry the 50 percent across. At 80 mph, the abstract notes, "significant CO generation indicated protective enrichment due to high load." Part of that penalty is a gasoline engine dumping extra fuel to protect itself. An electric drivetrain has no such behavior, and it carries far less thermal loss to begin with: fueleconomy.gov's energy-flow charts put 71 to 73 percent of an EV's energy at the wheels in highway driving, against 22 to 30 percent for a gasoline car on the same kind of driving. Change the powertrain and the same added drag arrives as a different percentage of a different total.

The government's own caveat on the roof box is narrower still. The footnote says the results come from a single roof-top cargo box measuring 20 by 40 by 50 inches, and that boxes with other dimensions or shapes may have a different effect. One box. Two gasoline vehicles. Twelve years ago.

Forum and video figures are weaker than that, not stronger: one trip, the trailer's weight reported and its height and width not reported at all, at an unstated speed, in unstated weather, with the pack's state of health unknown. The physics transfers. The percentage does not.

One finding from this literature does travel, because it is about behavior rather than about a percentage. Yuche Chen of NREL and Alan Meier of Lawrence Berkeley National Laboratory modeled what roof racks cost the national fleet for Energy Policy in 2016, and put it at 0.08 percent of US light-duty fuel use in 2015, roughly 100 million gallons of gasoline a year. The sensitivity analysis is the part to keep: the result was "most sensitive to the fraction of vehicles with installed roof racks but carrying no equipment." An empty rack is still a bluff body in fast air.

An hour on a familiar road gives you a number that is actually yours

Pick a flat stretch of divided highway you can drive out and back, ten miles or more each way. Out-and-back is the whole trick: it cancels the grade and most of the wind, which is why a one-way run tells you nothing.

1. Set cruise at the speed you will actually tow at, not the speed limit. Note the outside temperature.
2. Reset a trip meter at the start of each direction and record the kWh per mile the car reports. If yours shows miles per kWh, invert it.
3. Run it unloaded. Then hitch up, or fit the box, and run the identical route at the identical speed within the next hour or two. Same tire pressures.
4. Average each pair of directions and divide the loaded figure by the unloaded one.

That ratio is your multiplier, and it is valid at that speed and roughly that temperature only. If you expect to drive at two different speeds, do it twice. Because the aerodynamic share grows with the square of speed, the multiplier grows with speed too, which is why slowing down repays more when towing than it ever does unloaded.

Read the trip meter, not the range estimate. Many vehicles do adjust that estimate once a trailer profile is registered, and Rivian lists "the amount that you're towing" and gear "that might affect aerodynamics or weight like the Cargo Crossbars or a bike rack" among the things affecting a vehicle's range. But the estimate is a model. The kWh per mile is a measurement.

None of this appears on the window sticker, and it never will. EPA figures come off a dynamometer where, as fueleconomy.gov's how vehicles are tested page describes it, "engineers adjust the amount of energy required to move the rollers to account for wind resistance and the vehicle's weight." The vehicle's. Not the vehicle plus a 22-foot travel trailer.

From a multiplier to a stop count

The multiplier only becomes a plan when it turns into stops. Say a 600-mile day, a 90 kWh usable pack, a working window from 10 percent to 80 percent, and a measured baseline of 0.32 kWh/mi at your cruise speed. Leave home at 100 percent.

Every input in that table is a placeholder except the arithmetic. The 1.15 and the 1.80 are round numbers picked to show the shape of the result. They are not measurements, and they are emphatically not the Oak Ridge gasoline percentages converted into EV figures — no such conversion has been published, which is the entire reason for the hour on the highway above. Substitute your pack, your floors and your own two measured multipliers before any total here means anything.

Two things fall out of it that a single range percentage never shows. The first is how cheaply a stop gets added. A 1.15 multiplier sounds like rounding error, but the three-leg plan that covered 647 miles bare reaches only 562 with a box on the roof, and 562 is 38 miles short of where you were going. The second is that stops are not linear in the multiplier. The trailer turned two into five, and each of those costs far more than its charging minutes once you count exiting, queuing and rejoining traffic.

That last row is also where per-minute billing stops being an abstraction. Five sessions with a warm pack near the top of the curve is the worst case for a station that bills by the minute rather than the kilowatt-hour. And if the trip is a winter one, the multipliers stack on top of each other rather than sitting side by side, so run the trailer number through the cold-weather split before believing any of it.

The stall is a separate problem from the range

You can arrive with plenty of charge and still be unable to use the charger.

Whether a station can take a rig without unhitching is published data, not luck. The federal disclosure rule that makes connector data in the apps worth trusting carries a trailer field as well. Item (vi) in the port-level list at 23 CFR 680.116(c)(8) is "Accessibility by vehicle with trailer (pull-through stall) by port (yes/no)" — current eCFR text, retrieved 5 September 2026, and the annual CFR edition on govinfo reads identically. That value exists for every port at a federally funded site. If the app you use does not surface it, that is a fact about the app.

Operators hedge, and the hedging is informative. Rivian's answer on whether its Adventure Network sites can handle a vehicle pulling a trailer reads: "Yes, many will... While not every site can accommodate a trailer, we do factor it into our selection criteria" (read 5 September 2026). Many is not all, and factoring it into site selection is not a pull-through at the one you need.

So plan the unhitch as the default rather than the disaster: know before you pull in where the trailer would go, whether the surface is level enough to leave it, and whether the chocks are within reach rather than under everything else. An idle fee clock starts when the session ends, not when you finish re-hitching.

Before the next trip, get three numbers onto a card in the glovebox: your measured kWh per mile hitched at the speed you actually drive, the usable kilowatt-hours between your two state-of-charge floors, and the name of one station on the route with a confirmed pull-through port. The first takes an hour on a road you already know. The other two take ten minutes. All three outlast the trip.