Calories Burned Cycling, by Intensity

Watts are a measurement; everything else on the bike is a model#

For a 70 kg rider, an hour of cycling costs roughly 280 calories at a gentle 8 mph, about 560 at a steady 13 mph, and around 840 at a brisk 17 mph — gross figures, before you subtract the energy you would have spent sitting still. Those come from the standard catalogue of activity energy costs, and they are population averages with all the usual slack in them.

But cycling is the one endurance sport where you do not have to stop there. A bike fitted with a power meter puts a strain gauge in the drivetrain and records the mechanical work you actually performed, in kilojoules. No wrist device measures work; it infers energy from motion and pulse. That difference is why a cyclist can get closer to a real number than a runner or a swimmer ever can — and why the last step, turning measured work into calories, is still the part that goes wrong.

What an hour costs, by speed#

Every "calories burned cycling" calculator online is running on the same catalogue: the 2024 Adult Compendium of Physical Activities, which assigns each activity a MET value — its energy cost as a multiple of a standardized resting rate of 3.5 mL of oxygen per kilogram per minute¹. Convert at the usual approximation of 1 MET ≈ 1 kcal per kilogram per hour and you get the middle column below. The right-hand column is ours: the hourly rate divided by the speed, which the Compendium does not tabulate.

Riding speed	MET	Gross kcal/hour at 70 kg	Gross kcal/mile
Under 10 mph, leisure	4.0	~280	~35
10–11.9 mph, light	6.8	~476	~43
12–13.9 mph, moderate	8.0	~560	~43
14–15.9 mph, vigorous	10.0	~700	~47
16–19 mph, racing pace	12.0	~840	~48
Over 20 mph, not drafting	16.8	~1,176	~56

MET values are the Compendium's; the two right-hand columns are our conversion at representative speeds within each band (8, 11, 13, 15, 17.5 and 21 mph), scaled to a 70 kg rider. Multiply by your own weight in kilograms and divide by 70 to move them to you.

The Compendium's own authors are blunt about the limits here: the catalogue "does not reflect precise individual EE values," and they cite work in which women with obesity spent 8 to 15 percent more energy than lean women moving at identical speeds¹. Note too that every figure above is gross — it bills your ride for the calories you would have spent breathing on the sofa. Subtracting that resting share is a step almost no calculator takes, and what a running mile actually costs works through why it matters.

Why a cycled mile gets more expensive the faster you ride#

Now look down the right-hand column, because that is where cycling parts company with running. A running mile costs about the same at any pace — the price is paid per unit of distance, so it barely moves. A cycled mile does not behave that way at all: it climbs about 60 percent from a pootle to a racing effort, from roughly 35 calories to roughly 56.

The reason is that a runner is mostly paying to move their own body, while a cyclist is mostly paying to move air out of the way. Aerodynamic drag rises with the cube of speed, so it overtakes everything else surprisingly early. Above 14 metres per second — about 31 mph, sprint territory — roughly 90 percent of a rider's mechanical power output goes into overcoming aerodynamic drag alone⁵. Long before you reach that speed, drag is already the dominant term, which is why a headwind or a tuck changes your calorie total in a way that a runner's posture never could.

This has a practical consequence that runs opposite to the usual advice. On a bike, riding faster genuinely does buy you more calories per mile, not merely more per minute. It also means no single "calories per mile cycling" figure can survive contact with your actual ride — a flat spin in still air and the same distance into a 15 mph headwind are different events, and no calculator that asks only for distance can tell them apart.

The kilojoule shortcut, and the 24 percent hiding inside it#

If you ride with power, you already know the rule of thumb: kilojoules of work equal calories burned, one for one. It is a genuinely useful shortcut and it is not arbitrary. A kilojoule is 0.239 kilocalories, and the human body converts roughly a quarter of the chemical energy it spends into mechanical work at the pedals. Divide 0.239 by 0.24 and you land almost exactly on 1.0 — which is the entire basis of the rule.

The assumption doing all the work there is that 24 percent. Measured gross efficiency in cyclists tends to come in lower. When 69 male cyclists spanning a wide range of aerobic capacity were tested at 165 watts, gross efficiency averaged 18.6, 18.8 and 17.9 percent across low, medium and high fitness groups — and the authors found no relationship at all between efficiency and VO₂peak, concluding there are "no differences in efficiency and economy between elite cyclists and recreational level cyclists"³. A separate validation in nine national-team women found gross efficiency of 18.4 to 19.1 percent and recommended, in plain terms, that total mechanical work in kilojoules be multiplied by 5.3 — a figure they derive from a gross efficiency of 19 percent².

Run that multiplier out and the shortcut moves. A 1,000 kJ ride at 19 percent efficiency costs 5,300 kJ of metabolic energy, which is about 1,265 calories, not 1,000 (our arithmetic on Haakonssen's own recommendation). At the 18.5 percent that Moseley measured, it is closer to 1,290. The rule of thumb is not wildly wrong, but where it errs it errs downward — an unusual direction in this field, where nearly every consumer calorie readout flatters the user.

A power meter is the only instrument in consumer fitness that measures work rather than inferring it. It still cannot give you calories without assuming how efficiently you turn food into watts.

Two honest caveats keep this from being a correction factor you can bank. Gross efficiency is not a constant: it rises with work rate, because the fixed baseline cost of being alive is a shrinking fraction of the total as intensity climbs, and about 91 percent of the variance in energy expenditure during cycling is explained by work rate alone⁴. So a hard ride sits nearer the 1:1 rule than an easy one. And even with a calibrated meter, Haakonssen's group estimates landed within about 1 percent of indirect calorimetry while individual errors ran to about 11 percent — the population is well described; you are not.

What a heart-rate strap does instead#

Without a power meter, your computer or watch has no work signal at all, so it falls back on pulse. That method has a real physiological basis and a well-mapped failure mode. When 18 active adults cycled and ran at 60, 75 and 85 percent of VO₂max with indirect calorimetry running alongside, a heart-rate algorithm using group calibration underestimated cycling energy expenditure by 3 to 12 percent, with agreement rated poor to moderate (ICC 0.64 to 0.73). Calibrating the same algorithm to each individual lifted agreement to 0.79 to 0.88⁶.

That gap between a group equation and a personal one is the whole story of wrist-based burn estimates, and it has its own article in how heart rate is turned into a calorie number; the broader accuracy record of the devices themselves is in how accurate fitness-tracker calories are. For cycling specifically, the ranking is clear enough: a calibrated power meter beats a heart-rate estimate, a heart-rate estimate beats a speed-and-distance calculator, and none of them is precise enough to spend.

Riding a number you can actually trust#

If you have power, log kilojoules and treat calories as a range. Somewhere between your kJ total and about 1.3 times it, depending on how hard you rode. Reporting "800 kJ" is a measurement; reporting "800 calories" is a measurement plus an efficiency guess.
If you do not have power, expect the wider band. Speed-based tables cannot see wind, gradient, drafting or your position on the bike — the four things that most change what a mile costs.
Don't build the deficit out of the ride. The burn is real but partly reclaimed, and the estimate is loose; the case is laid out in does exercise burn as many calories as you think. Set the target from intake, where the arithmetic leaks less — see how a calorie deficit drives weight loss.
Compare rides, not readouts. The same rider, the same bike, the same route: the error largely repeats, so this week against last week carries information the absolute number does not.

Cycling is the cheapest mile in endurance sport and the best-instrumented one. Both of those are worth knowing before you decide what a two-hour ride has bought you.

FAQ#

How many calories does an hour of cycling burn?#

For a 70 kg rider, roughly 280 calories an hour at under 10 mph, about 560 at 12–13.9 mph, and about 840 at 16–19 mph, gross. Those follow from the 2024 Compendium's MET values of 4.0, 8.0 and 12.0 for those speed bands. Scale them to your own weight in kilograms, and treat them as population averages — the Compendium's authors note that body composition alone can shift the cost of identical movement by 8 to 15 percent.

Does 1,000 kilojoules on a power meter really equal 1,000 calories?#

Roughly, but it slightly undercounts. The one-to-one rule assumes you convert about 24 percent of the energy you spend into work at the pedals. Measured gross efficiency in cyclists is commonly 18 to 19 percent at moderate power, which puts 1,000 kJ nearer 1,250 to 1,290 calories. Efficiency rises with intensity, so a hard ride sits closer to 1:1 than an easy one.

Is a power meter more accurate than a heart-rate monitor for counting calories?#

Yes, though neither is exact. A power meter measures mechanical work directly; heart rate only correlates with it. In a laboratory comparison against indirect calorimetry, a group-calibrated heart-rate algorithm underestimated cycling energy expenditure by 3 to 12 percent, while calibrated power-meter estimates landed within about 1 percent on average — with individual errors of about 11 percent even then.

Sources#

This article was researched and drafted with AI assistance and reviewed for accuracy by the BurnWeek team. It is general information, not medical advice. How we research and correct our articles →