Heated driveway cost dataset

The load, in kilowatts and amps

At 40 watts per square foot — the trade’s figure for a cold-climate driveway, in a band of 36 to 50 — a whole slab wants a 167 amp breaker, sized at 1.25× the running current because a load that runs for hours is continuous. ⚠️ That is more than an entire 100 amp service. It is why almost every heated driveway is two strips, and it is the reason this dataset prices the contactor panel and the copper feeder by the load rather than flat: modelling them flat made a 33 amp system look dear and a 167 amp one look cheap, which is backwards from what decides this. The design standards are ASHRAE’s and article 426 of the National Electrical Code; both are sold by their publishers and neither is reproduced here.

What a season costs, in both fuels

For an 800 square foot slab at national energy prices — 18.11 cents per kilowatt-hour and $15.34 per thousand cubic feet of gas. ⚠️ The two money columns are comparable and the two consumption columns are not: kWh and Mcf are different units and adding them would be adding pears. The model runs the system at 70 per cent of full output while it is on, because a snowmelt system brings the slab up to setpoint and then modulates — leaving that factor out, which is what the trade does when it multiplies kilowatts by hours, overstates the bill by more than forty per cent. Sources are EIA, Electric Power Monthly, table 5.6.B — residential price by state (May 2026, year to date) and EIA, Natural Gas — average residential price by state (2025, annual), converted with EIA, Energy conversion calculators: “1 cubic foot=1,036 Btu (based on U.S. average for natural gas delivered to consumers in 2026; estimate)”

The same driveway, priced by where it is

Identical equipment, two and a half times the bill — and the spread across all fifty states is wider still, from 11.92 cents to 44.61. ⚠️ It is the single largest variable on this page and it has nothing to do with the driveway. How many hours the system runs is the other half, and that comes from the climate rather than the ZIP code: snow varies far more inside a state than between states, so this calculator asks rather than guesses. NOAA NCEI, U.S. Climate Normals is where to look yours up — “A “normal” is the 30-year average of a particular variable’s measurements, calculated for a uniform time period.” and “Of the stations reporting precipitation, more than 5,700 have adequate observations to report snowfall and snow depth normals.”

By winter, and against not having one

⚠️ Three of these four cells are not numbers, and that is the point. A whole heated slab costs more in electricity each winter than a ploughing contract does, so no number of seasons closes it and the cell says “never”. Where a payback does exist but lands beyond the 40 years the buried element lasts, the cell says that instead — a payback longer than the thing’s own life is not a payback, and this table published “3,029 seasons” before that rule went in. And the model returns nothing rather than a negative number when hydronic wins outright, because a cell reading “−9 seasons” is the kind of thing that gets quoted on its own. The same discipline flips the electric against hydronic row: on a whole slab hydronic is cheaper to install and cheaper to run once the service upgrade the electric system forces is counted, so there is no crossover to publish. The costs of clearing snow another way are our estimates and labelled as such — there is no federal series for a plough contract. What there is: National Weather Service, Winter Storms — during a storm — “Avoid overexertion such as shoveling heavy snow, pushing a car or walking in deep snow if you are not in good health. The strain from the cold and the hard labor may cause a heart attack.”

How long each part lasts

⚠️ The first row is not really a lifespan, it is a constraint, and it is the whole argument for how to spend the money: what is buried cannot be repaired without breaking the driveway, and everything else can be reached. Which is also why there is no such thing as retrofitting one of these into a driveway that already exists — the driveway comes out.

Line items

80% of the reference total comes from estimated rather than official figures, which is high and honest: there is no published price series for heating mats or snowmelt controls, and the official half is the labour. The reference job is two heated tyre tracks on a 40 by 20 foot driveway in new concrete, electric mats on automatic control, a 60 foot feeder run: 160 heated square feet, 6.4 kW at 240 volts, 20.1 trade hours. Options that are not selected sit at zero quantity, so the table below is all three systems rather than one job.

The reasoning behind each figure

Pre-spaced heating matsestimated, ±28%

• Cable woven into mesh at the spacing that gives about 40 watts per square foot, which is what the trade uses for a driveway in a cold climate. The band runs from 36 to 50 depending on the surface and the winter.
• ⚠️ Priced per square foot of heated surface, not of driveway. Two tyre tracks are about a fifth of a driveway, and that is the single biggest lever on this page.

Loose heating cable and spacing stripestimated, ±30%

• Cable on a reel plus the plastic strip that holds the spacing. Cheaper than mats to buy and slower to lay, and the only electric option that survives being paved over with hot asphalt.

Rolling out and tying down the mats

• Modelled at one hour per 55 square feet, plus four hours of setting out, splicing and cold-lead terminations. Mats are the quick way: the spacing is already right and the work is rolling and tying.
• ⚠️ The hours are ours; the wage is official BLS data for electricians, adjusted to your metropolitan area.

Laying and clipping loose cable

• One hour per 32 square feet plus five hours of setting out and terminations — slower than mats because every loop is placed and clipped by hand, and the spacing has to be right first time.

Contactor enclosure and controls wiringestimated, ±41%

• ⚠️ The line people do not expect: a system this size does not plug into a spare breaker in the house panel, it gets its own contactor panel. This is the part of it that does not change with the load — the box, the mounting, the control wiring.

Contactors and breakers, per amp of switched loadestimated, ±45%

• ⚠️ Priced per amp on purpose, and it is a correction rather than a refinement. The first version charged one flat band whatever the load, so a 33 amp two-track system and a 167 amp full slab got the same switchgear — and the line's own note admitted they are not the same equipment. A band that wide is a modelled difference in hiding.
• Sized at 1.25× the running current, because a load that runs for hours is a continuous load and the trade sizes it that way.

Running the feeder from the panel to the driveway

• Six hours plus an hour per twenty feet of run, for conduit, conductors, terminations and the ground-fault protection this has to have.
• ⚠️ Multiplied by 1.35 on a retrofit. Pulling a feeder heavy enough for this through a finished house is a different job from roughing it in while the walls are open, and that difference is real money rather than a rounding.

Buried conduit, fittings and ground-fault protectionestimated, ±42%

• Per foot of run from the panel to the slab: conduit, fittings, the trench and the ground-fault protection this has to have outdoors. It barely changes with the size of the load.

Copper, priced by how much of it the load needsestimated, ±42%

• ⚠️ The unit is deliberate: one unit is a foot of conductor sized for 100 amps, so the quantity is the length of the run multiplied by the amps and divided by a hundred. Copper is sold by weight and a conductor for 167 amps is several times the metal of one for 33.
• Modelling it flat per foot — which is what the first version did — made the two-track system look dear and the full slab look cheap, which is exactly backwards from what makes this decision.

Oxygen-barrier PEX tubing and clipsestimated, ±35%

• Half-inch oxygen-barrier tubing at nine-inch centres, with clips or ties to the mesh. Cheap material — the tubing is not why hydronic costs what it does.

Laying and pressure-testing the tubing

• One hour per 42 square feet plus six hours for the loops, the sleeves through the slab edge and the pressure test that has to be holding while the concrete goes in.
• ⚠️ The pressure test is not a formality here. A tubing fault found after the pour is found by breaking the driveway.

Manifold, mixing valve, circulators and controlestimated, ±46%

• A snowmelt manifold with balancing, a mixing assembly to keep the slab supply well below boiler temperature, circulators and the snow-sensing control.
• ⚠️ The mixing valve is not optional and it is what separates this from a heating manifold: sending boiler-temperature water into a frozen slab is how slabs crack.

Propylene glycol chargeestimated, ±42%

• ⚠️ A snowmelt loop is outdoors and full of liquid, so it runs on glycol rather than water. It is priced per square foot because the system volume follows the tubing.
• It is also a maintenance item nobody mentions at the quote: glycol degrades and its inhibitors are used up, so it gets tested and eventually changed. That is on the exclusions below rather than hidden in this line.

Snowmelt boiler, sized to the slabestimated, ±33%

• Priced per thousand Btu per hour of output, because that is how a boiler is bought. The slab asks for 136 Btu/h per square foot, and the boiler is sized above that for the 10 per cent it loses.
• ⚠️ This is the line that makes hydronic expensive, and it is also the honest answer to why you cannot usually just tie into the boiler you already have: a driveway asks for as much heat as a whole house, and a house boiler was sized for a house.

Setting, piping and venting the boiler

• Fourteen hours plus an hour per 25 thousand Btu of output: gas, flue, condensate, near-boiler piping, filling and purging the glycol, and commissioning the control.

Snow and moisture sensor with automatic controlestimated, ±53%

• A sensor that reads both moisture and temperature — in the slab or on a mast — plus the controller that reads it. It is the dearest control and the cheapest to run, because it is the only one that never heats the driveway for nothing.
• ⚠️ On a system this size the controller pays for itself in one winter against leaving it switched on, and the calculator shows by how much.

Manual switch and timerestimated, ±52%

• A switch and a run-on timer so it cannot be left on for a week by accident. ⚠️ Cheapest to fit and the one that costs most in the hours it runs, because the slab has to be started before the snow rather than when you see it.

Slab sensor and setpoint control for idle modeestimated, ±50%

• A slab temperature sensor and a control that holds it a few degrees above freezing all winter so the system responds instantly.
• ⚠️ Cheap to fit and the most expensive decision on this page by a wide margin. What it costs is not here, it is in the running-cost table.

⚠️ Second paving pass over the heating cableestimated, ±39%

• ⚠️ Asphalt goes down near 300 °F, so the cable cannot simply be buried in a single lift. It sits on the base course and a second, thinner lift goes over it — an extra mobilisation of the paving crew and an extra pass of material.
• It is priced over the whole driveway rather than the heated part, because the paver lays the whole thing.

Working around the heating layer during the pour

• One hour per 130 square feet of heated area, plus three. Concrete cannot be barrowed or raked over a heating element the way it is over bare mesh, and somebody watches the element while it is placed.
• ⚠️ This is a small line that protects a large one: an element damaged during the pour is found the following winter, under a driveway.

Electrical or mechanical permit and inspectionestimated, ±56%

• A new circuit of this size, or a new gas appliance, is permitted and inspected work almost everywhere. Set locally and varies more than this band suggests.

Not included in this dataset

The driveway itself

⚠️ Deliberately not priced here, and it is usually the larger half. Concrete and asphalt driveways are modelled whole on their own pages — including demolishing an old one — and this page shows those figures alongside rather than inventing a second set. One job, one model, one number, and it is also why the two pages cannot drift apart.

A service upgrade, if the amps are not there

⚠️ Not priced here either, and for a full slab it is the likeliest reason a project stops. This page publishes what the system draws; our electrical panel upgrade page prices the upgrade whole. What it does not do is guess how much spare capacity your house has, because that is a load calculation an electrician does with the code in hand.

Snow removal you will still be paying for

A heated driveway does not clear the path, the steps or the ridge the town plough leaves across the end of it. The comparison on this page is against clearing the driveway, and that is honest only if you read it as the driveway.

Replacing the glycol, and what happens if you do not

⚠️ Not a cost on this page and it is the one that shortens a hydronic system. Propylene glycol degrades and its corrosion inhibitors are consumed; left long enough it turns acidic and attacks the very components it is there to protect. It is tested every few years and changed eventually, and a system that has never had it looked at is a system running on unknown fluid.

ASHRAE and NEC design requirements

The design standard for snow melting is ASHRAE’s, and fixed outdoor electric de-icing and snow-melting equipment is covered by article 426 of the National Electrical Code. Both are copyrighted documents sold by their publishers, and neither is reproduced or derived here. What this page publishes is arithmetic of our own, and your local code is what decides.