James Watt, the Scottish inventor and mechanical engineer, created the concept of horsepower in the late 1780s as a way to sell steam engines to mine operators and mill owners. The unit was not born from pure scientific curiosity but from a very practical problem: Watt needed a way to show potential customers exactly how many horses his engine could replace. His solution was to watch horses working at a coal mine, estimate how much work they could do in a given time, and then express his engine’s output in those terms. The number he landed on, and the story behind it, turned out to be one of the most durable units of measurement in history.
Why Watt Needed a New Unit
By the 1760s, steam engines were already in use at British coal mines, pumping water out of shafts. The dominant design was Thomas Newcomen’s atmospheric engine, which had been around since 1712. Watt dramatically improved on this design by adding a separate condenser, which made his engines far more fuel-efficient. But when he and his business partner Matthew Boulton began selling these improved engines in the 1770s and 1780s, they ran into a communication problem. Mine owners and brewers understood horses. They employed teams of horses to turn mills, haul coal, and drive machinery. Telling a customer that an engine produced a certain number of foot-pounds per minute meant nothing to them. Telling them it could replace ten horses meant everything.
Watt was not the first person to compare mechanical output to animal labor. Earlier engineers had made rough comparisons, and the idea of equating engine work to horse work floated around the engineering circles of the day. But Watt was the first to formalize it into a specific, repeatable number and use it systematically as a commercial standard. That distinction matters, because the concept only became useful once it was pinned down precisely enough that buyers and sellers could agree on what they were trading.
How Watt Measured the Horses
The exact details of Watt’s experiments have been debated by historians, and some of the popular retellings are embellished, but the general method is well documented. Watt observed draft horses working at a brewery or a mine (accounts differ on the location) that were harnessed to a capstan or mill wheel, walking in circles to raise loads of coal or turn grinding equipment. He measured how much weight a horse could lift over a set distance in a set time.
His observations suggested that a strong horse could push with a force of roughly 180 pounds while walking at about 181 feet per minute. Multiplying those together gave approximately 32,580 foot-pounds per minute. Watt rounded this up to 33,000 foot-pounds per minute, and that became the definition of one horsepower. In other words, one horsepower is the power needed to lift 33,000 pounds by one foot in one minute, or equivalently, 550 pounds by one foot in one second.
Whether Watt rounded up deliberately to give customers a generous impression of what they were buying is a question historians enjoy speculating about. A slightly inflated number meant the engine would always seem to overperform relative to the horses it replaced, which is good salesmanship. Some accounts suggest Watt tested only a few horses and chose a figure on the high side of typical sustained output. Whatever his exact reasoning, the number stuck.
Can a Real Horse Produce One Horsepower?
This is one of the most common follow-up questions, and the answer is surprisingly nuanced. A healthy draft horse can sustain about 0.7 horsepower over a full working day. For short bursts, a horse can produce well over one horsepower, with peak outputs estimated at roughly 10 to 15 horsepower during a brief sprint or explosive effort. So Watt’s figure of one horsepower actually represents something close to the maximum sustained output of a particularly strong horse working steadily, not the average output of an ordinary horse over a full shift.
This mismatch is not a flaw in the unit. Watt was trying to capture what a horse could do at peak sustained effort, because that was the relevant comparison for mine operators. If your horse team could haul a certain amount of coal per hour at steady work, you wanted to know that the engine could match or beat that rate. The fact that the same horse would tire and slow down over eight hours while the engine would not was part of the sales pitch, not a problem with the measurement.
Horsepower as a Marketing Tool
Boulton and Watt’s business model depended heavily on the horsepower concept. Rather than selling engines outright at first, they often charged customers a royalty based on how much fuel the Watt engine saved compared to a Newcomen engine doing the same work. Having a standardized way to express an engine’s output made these contracts possible. If both parties agreed that the engine produced, say, twenty horsepower, they could calculate its value relative to the cost of keeping twenty horses.
The unit spread rapidly through British industry in the late 18th and early 19th centuries precisely because it solved a real problem. Engineers adopted it because it was intuitive. Factory owners adopted it because it let them compare machines from different manufacturers on equal terms. By the time the industrial revolution was in full swing, horsepower had become the lingua franca of mechanical power, used for everything from textile mills to locomotives to steamships.
The Formal Definition and Its Variants
Watt’s original figure of 33,000 foot-pounds per minute became known as mechanical horsepower or imperial horsepower. As the unit traveled across borders, it picked up regional variations. In continental Europe, a slightly different version emerged based on the metric system. Metric horsepower, sometimes abbreviated as PS (from the German Pferdestärke, literally “horse strength”), is defined as the power needed to lift 75 kilograms by one meter in one second. That works out to about 735.5 watts, compared to 745.7 watts for mechanical horsepower. The difference is small, roughly 1.4 percent, but it can cause confusion when comparing specifications across countries.
Other specialized versions exist too. Boiler horsepower, used for rating steam boilers, equals about 9,810 watts, a much larger figure that reflects the total thermal energy a boiler can produce rather than mechanical shaft output. Electrical horsepower is defined as exactly 746 watts, which is a convenient rounding of the mechanical figure. These variants arose because different industries needed the concept adapted to their specific equipment and measurement conventions.
From Horsepower to the Watt
The irony of James Watt’s legacy is that the unit that eventually replaced horsepower in scientific and engineering contexts was named after him. The watt, as a unit of power, was formally adopted in the second half of the nineteenth century as part of a broader effort to standardize electrical units. The watt and the joule were part of a group of practical electrical units named and defined during that period, honoring scientists and engineers whose work underpinned the understanding of energy and power.1Substantia. Watt’s in a name? Units of power and energy One watt equals one joule per second, and one mechanical horsepower equals roughly 746 watts.
In most scientific, engineering, and regulatory contexts worldwide, the watt and its multiples (kilowatt, megawatt) have replaced horsepower. The International System of Units uses the watt exclusively. European car manufacturers list engine output in kilowatts, and electrical equipment everywhere is rated in watts. Yet horsepower persists stubbornly in everyday language, especially in the automotive world, which raises an obvious question.
Why Horsepower Refuses to Die
Despite being a unit invented for 18th-century mine operators, horsepower remains the dominant way most people in the United States, United Kingdom, and several other countries talk about engine power. Car advertisements, lawnmower specs, and outboard motor ratings still use it. Even in countries that officially use metric horsepower or kilowatts, consumers and journalists often convert back to horsepower because it feels more tangible.
Part of the reason is cultural momentum. Generations of car buyers have built an intuitive sense of what 200 horsepower “feels like” versus 400 horsepower. That intuition does not transfer easily to kilowatts, even though the conversion is straightforward. Telling someone a car has 150 kilowatts is mathematically equivalent to saying it has about 201 horsepower, but the kilowatt figure does not trigger the same gut-level understanding for most people.
Another reason is marketing. The automotive and marine industries discovered long ago that bigger horsepower numbers sell products. A figure like “707 horsepower” on a muscle car sounds dramatically more impressive than “527 kilowatts,” even though they describe the same engine. Manufacturers in horsepower-dominant markets have little incentive to switch, and consumers have little reason to demand the change.
Brake Horsepower, Wheel Horsepower, and Other Measurement Points
If you have ever looked at car specifications closely, you may have noticed terms like brake horsepower (bhp), wheel horsepower (whp), or shaft horsepower (shp). These are not different units. They all use the same underlying definition of horsepower. The difference is where in the drivetrain the measurement is taken.
Brake horsepower is measured at the engine’s crankshaft, typically using a device called a dynamometer (historically a friction brake, hence the name). It tells you how much power the engine itself produces before any of it is lost to the transmission, differential, axles, and other drivetrain components. Wheel horsepower is measured at the wheels, after all those losses. The gap between the two is usually 10 to 20 percent for most passenger cars, which means a car rated at 300 brake horsepower might deliver 250 to 270 horsepower to the road. Shaft horsepower is the equivalent measurement for turbine engines, propeller shafts on ships, and similar applications.
The distinction matters because manufacturers sometimes advertise the most flattering number. Historically, American automakers measured horsepower using gross ratings, which tested the engine on a stand with no accessories attached: no alternator, no power steering pump, no air conditioning compressor, no exhaust restrictions. This made engines look more powerful on paper. In 1972, the industry switched to net horsepower ratings, which include all standard accessories and a full exhaust system. The change caused advertised figures to drop overnight, sometimes by 20 percent or more, even though the engines themselves had not changed. If you ever compare a 1970 muscle car’s rated horsepower with a modern car’s, keep in mind that the older number was measured under artificially favorable conditions.
Horsepower in the Age of Electric Motors
Electric vehicles have introduced a new wrinkle in how people think about horsepower. Electric motors deliver their full torque from zero speed, which means the driving experience of an electric car with 300 horsepower feels very different from a gasoline car with 300 horsepower. The electric car tends to feel faster off the line because it is not waiting for the engine to reach its peak power band. This has led some enthusiasts to argue that horsepower comparisons between electric and gasoline vehicles are misleading, though the underlying physics is the same: a watt is a watt regardless of what generates it.
Electric vehicle manufacturers generally list output in both horsepower and kilowatts, depending on the market. Tesla, for instance, tends to emphasize acceleration times and torque figures rather than raw horsepower, partly because the numbers are so high they sound implausible to people calibrated on gasoline engines. A modern high-performance electric sedan can produce over 1,000 horsepower, a figure that was exotic supercar territory just a decade ago. Whether the unit still serves a useful comparative purpose when the underlying technology has changed so dramatically is a fair question, but for now, horsepower remains the common language.
Predecessors and Parallel Inventions
Watt often gets sole credit for inventing horsepower, but the idea of comparing mechanical work to animal work existed before him. Thomas Savery, who patented an early steam pump in 1698, advertised his device’s capability in terms of horses it could replace, though he never formalized a specific unit. John Smeaton, an engineer working in the mid-1700s, also made calculations comparing engine output to horse labor. What Watt added was precision and standardization: a single number, derived from measurement, that anyone in the industry could use.
It is also worth noting that Watt did not work in isolation. His partnership with Matthew Boulton was critical. Boulton handled much of the commercial strategy, and the decision to market engines explicitly in terms of horsepower was as much a business decision as a scientific one. Watt was the engineer who did the measuring; Boulton was the businessman who recognized how powerful the concept would be as a sales tool. The combination of rigorous measurement and sharp marketing instinct is what made horsepower a lasting standard rather than just another engineer’s estimate.
Common Misconceptions About the Origin Story
Several popular versions of the horsepower story contain errors that have been repeated so often they feel true. One claim is that Watt deliberately overestimated horse power to make his engines look good. While his 33,000 foot-pounds figure does exceed the average sustained output of most horses, historical evidence suggests Watt was genuinely trying to measure peak sustained effort, not cheat. The rounding was likely a practical simplification rather than a calculated deception.
Another common myth is that Watt measured pit ponies, the small horses used underground in mines. Most historical accounts indicate he observed full-sized draft horses working above ground at a mill or brewery. Pit ponies were common in mines, but they were much smaller and weaker than the heavy draft breeds Watt reportedly studied. Using pit ponies would have produced a much lower figure and made his engines look even more impressive, which somewhat undermines the “deliberate inflation” theory.
A third misconception is that horsepower was Watt’s most important contribution to engineering. It was not. His separate condenser, his double-acting engine design, his governor mechanism for speed regulation, and his work on the compound engine were all far more significant in engineering terms. Horsepower was a clever measurement and marketing tool, but it was the least technically innovative part of his career. It just happens to be the part most people remember.
How Horsepower Shapes Regulation and Taxation
In several countries, engine power ratings directly affect what you pay in taxes or insurance. France, for example, has a system of fiscal horsepower that determines annual registration taxes. This fiscal figure is calculated from engine displacement and other parameters, not from actual dynamometer output, which means it only loosely correlates with real-world power. The system dates back to the early 20th century, when governments needed a simple proxy for vehicle capability to set tax brackets.
Britain had a similar system under the RAC horsepower formula, introduced in 1910. That formula was based only on cylinder bore and number of cylinders, ignoring stroke length entirely. British car manufacturers responded by building engines with long strokes and narrow bores to minimize their tax rating while maximizing real output, a design quirk that shaped an entire generation of British engines. The formula was finally abandoned after World War II, but its engineering legacy lingered for decades in the form of cars whose engine architecture had been warped by tax policy rather than engineering optimization. When a unit of measurement gets embedded in law, it can shape technology in ways its inventor never anticipated.