How Elevators Safely Carry Millions of People

TL;DR. Riding an elevator forty floors up is, on paper, one of the stranger things people do without a second thought: stand in a metal box, suspended over open space by steel ropes, and trust it completely. That trust is earned by a system with almost no single point of failure. A traction elevator is balanced by a counterweight so the motor never lifts the car's full weight, held up by multiple ropes each strong enough to carry the load alone, and backed by a governor and safety brake that can stop a runaway car by clamping onto the guide rails even if every rope failed at once. That last part isn't marketing. It's a legally required, physically tested mechanism, and it dates to a single, deliberately staged public demonstration in New York in 1854. The elevator that resulted didn't just make tall buildings more convenient. It made them possible.

Key takeaways

  • A traction elevator's motor doesn't lift the car's full weight. A counterweight, roughly equal to the car plus 40 to 50 percent of its rated load, hangs on the other end of the same ropes over a grooved wheel, so the motor only has to overcome the difference between the two sides.
  • The safety mechanism that matters most, a spring-loaded brake that clamps onto the guide rails if the car overspeeds, was invented and demonstrated by Elisha Otis in 1854, decades before it was ever needed in an emergency. It's still required by law on every traction elevator built today.
  • Elevators run on multiple independent hoist ropes, never just one, and each rope alone is rated to hold the full load with a wide safety margin. Codes require at least three ropes on a traction elevator and monthly visual inspection of every one of them.
  • A specialized, often unionized trade (elevator constructors and mechanics) installs and services elevators, and a separate licensed inspector, independent of the installing company, has to certify most elevators safe on a recurring schedule, commonly annually.
  • Elevators are, by the numbers, one of the safer ways to move through a building: roughly 0.015 reported accidents per elevator per year, against about 0.221 per escalator, and stairs cause far more injuries and deaths than either one.
  • The U.S. and Canada together run somewhere close to a million elevators, collectively making an estimated 18 billion passenger trips a year, a number large enough that dense urban life, in its current form, would not function without them.

The moment nobody thinks about

You step into a metal box a little larger than a closet, press a button with a number on it, and stand still while the box accelerates upward at a rate that would alarm you in almost any other context. Cables you cannot see are holding you over open space, sometimes twenty, sometimes eighty, in a handful of buildings well over a hundred floors, above solid ground. Nobody grips the handrail. Nobody checks the inspection certificate posted near the door, if they even notice it's there. The doors open on the requested floor a few seconds or the better part of a minute later, and the whole event leaves no impression at all.

Described in the abstract, an elevator ride sounds like exactly the kind of thing a cautious person should fear: an enclosed box, dangling on ropes, with a floor that could theoretically give way beneath you. Almost nobody experiences it that way, and the reason isn't complacency. It's that the system underneath the button has been engineered, since the 1850s, around one specific design question: what happens if something fails, and how do we make sure the answer is never "the car falls."

The mechanism: what's actually holding the car up

Most elevators in mid-rise and high-rise buildings are traction elevators. The car hangs from a set of steel wire ropes (in everyday speech people call them cables, though in the industry a cable specifically means an electrical or communication line, not a load-bearing rope) that run up from the top of the car, over a grooved wheel called a sheave, and back down the other side of the shaft to a counterweight, a solid block of iron or concrete that rides in its own set of guide rails.

The counterweight is the part almost nobody notices, and it's doing most of the actual work. It's sized to roughly match the empty car's weight plus somewhere around 40 to 50 percent of the car's rated (maximum legal) load, according to elevator industry engineering references. That specific fraction isn't arbitrary: it's the point at which, averaged across a lifetime of rides at every possible passenger load, the motor has to do the least total work, because the heaviest lift (a fully loaded car) and the lightest lift (an empty car, which makes the counterweight the heavier side) end up costing roughly the same effort in opposite directions. An electric traction motor turns the sheave, and friction between the grooves in the sheave and the ropes is what actually transmits force, not a spool winding rope onto a drum. The motor's job, on a well-balanced system, is only ever to overcome the difference between the car's side and the counterweight's side, plus friction and acceleration, never the building's full multi-ton static weight.

Hydraulic elevators work on a completely different principle and show up mostly in low-rise buildings, roughly two to seven or eight stories. There's no counterweight and, in most designs, no rope holding the car up at all. Instead, a piston sits beneath (or beside) the car, and an electric pump forces hydraulic fluid, usually a type of oil, into the cylinder housing that piston, physically pushing the car upward. To descend, a valve simply releases fluid from the cylinder and gravity lowers the car back down. Hydraulic systems only draw significant power going up, which makes them efficient for short buildings, but the piston's own length limits how high they can practically reach, which is why they essentially disappear from anything taller than a parking garage or a small office building.

Both designs share the two features that make falling, as most people imagine it, essentially a non-event: multiple independent supports, and a brake that doesn't depend on the thing it's protecting against.

For a traction elevator, codes (adapted from state regulations that mirror national safety standards) require a minimum of three separate hoist ropes, each one individually rated, with a wide safety margin, to hold the car's full load by itself. California's elevator safety regulations, for instance, specify that the ratio of a rope's breaking strength to the maximum load it carries has to clear a set minimum, commonly cited around 12 to 1 when three ropes share the load, and mechanics are required to visually inspect the ropes at least once every 30 days for fraying, corrosion, or broken wires. The result is that "all the cables failing at once" isn't really a statistical near-miss that inspection happens to catch early. It would require several independently rated, independently inspected structural members to fail simultaneously, which almost never happens, and even if one rope did fail, the others are each individually strong enough to hold the car on their own.

Don't be confused: the ropes aren't the safety system, they're the normal operating system. The genuinely dangerous scenario an elevator has to survive isn't one worn rope snapping; it's every rope failing at once, or the car losing its ability to slow itself in a way the motor and brake don't catch. That's a different mechanism entirely, described below, and it's the one that actually mattered enough to change the course of architecture.

The part that's designed to catch you: governor and safety brakes

Every traction elevator carries a second, entirely independent system whose only job is to physically stop the car if it ever moves faster than it's supposed to, regardless of what the ropes are doing. That system has two parts.

The governor is a small, separate device, usually mounted in the machine room, connected to the car by its own thin rope that moves at the same speed as the car. Inside it, spinning weights respond to centrifugal force exactly the way a mechanical engine governor does: as the car's speed increases, the weights swing outward, and once they swing far enough, they trip a switch. Industry engineering references commonly describe a two-stage response: an electrical switch trips around 110 percent of rated speed, cutting power to the drive motor and applying the normal operating brake, and if the car is still overspeeding after that, around 125 percent of rated speed, the governor mechanically engages a second, more aggressive device.

That second device is the safety brake (also called safety gear), a set of steel wedges or jaws mounted directly to the car frame, positioned against the vertical guide rails that run the full height of the shaft. When the governor rope yanks the tripping mechanism, spring force or a cam action forces those wedges to bite directly into the rails, and the harder the car tries to keep moving, the tighter the wedges grip, in most designs, bringing the car to a controlled stop over a short distance rather than an instant, jarring one. This is the exact mechanism Elisha Otis demonstrated in 1854, described below, and every conventional traction elevator built since is required to have some version of it. It does not depend on the hoist ropes, the motor, or building power. A car with every single hoist rope cut still has a governor-triggered brake standing between it and the bottom of the shaft, which is why "the elevator cable snapped and it plummeted to the ground" is close to the least likely thing to actually happen in a modern, maintained building, however often it appears in fiction.

The complete journey: design, installation, and the system that decides which car comes

An elevator begins as an engineering calculation specific to one building. Vertical transportation consultants and engineers model expected traffic (how many people, on which floors, at what times of day, most critically the morning "up-peak" when an entire office tower tries to arrive within the same twenty minutes) and size the number of cars, their speed, and their capacity around that projected demand, not around a generic standard.

Once ordered, an elevator's core mechanical parts (the car, hoist ropes, motor, sheave, controller, guide rails, and doors) are manufactured, largely off-site, by companies like Otis, KONE, Schindler, and TK Elevator among others, then assembled and adjusted inside the building's shaft by specialized mechanics. This installation work, aligning guide rails to fractions of an inch over dozens of stories, reeving the ropes correctly around the sheave, wiring the controller, and calibrating the safety systems, is skilled trade labor, typically performed by union elevator constructors in North America, and a single elevator installation in a new high-rise can take weeks per car.

The controller is the part that decides how cars behave once installed, and it has changed more than any other part of the system since the 1900s. Early elevators needed a human operator physically working a lever. Automatic push-button control (press a button, wait for whichever car answers) replaced attendants through the mid-20th century, and modern buildings, especially taller ones, increasingly use destination dispatch: instead of pressing an up or down button and then choosing a floor once inside the car, a passenger enters their destination floor at a terminal in the lobby before boarding at all. The building's control system then groups everyone headed to nearby floors into the same car and assigns that car before the doors even open, which cuts down on the number of intermediate stops any single car has to make and can meaningfully shorten both wait times and travel times in buildings with heavy traffic.

Don't be confused: destination dispatch isn't just a fancier call button. In a conventional elevator, the building doesn't know where you're going until you're already inside and have pressed a floor number, so assignment happens almost blind. In a destination dispatch system, that information exists before you board, which is what lets the building's software group passengers by floor and route entire carloads of people efficiently, something that isn't possible once passengers are already scattered across multiple cars pressing their own buttons.

The last major design split is where the motor physically lives. Traditional elevators use a machine room, a dedicated space (commonly at the top of the shaft) housing the motor, sheave, and controller. Machine-room-less (MRL) elevators, made possible by compact gearless motors small enough to mount directly inside the hoistway itself, remove that separate room entirely. KONE introduced the first commercially significant MRL elevator in the Netherlands in 1996, and the design has since become common in mid-rise construction because it saves both construction cost and usable building space, at the cost of somewhat harder access for major repairs and certain rescue scenarios, since technicians and rescuers now have to reach components from inside the shaft rather than a separate room.

Who keeps it running

Elevator mechanics (sometimes called elevator constructors) install, adjust, and service the equipment, and in the United States and Canada, most belong to the International Union of Elevator Constructors (IUEC), which represents more than 27,000 members. Becoming a mechanic isn't a short process: the IUEC's apprenticeship, run jointly with the National Elevator Industry Educational Program (NEIEP), typically runs four to five years, combining roughly 2,000 hours of supervised, paid on-the-job work per year with 100 to 200 hours of classroom instruction, and candidates have to pass an aptitude test and an interview just to be accepted.

A separate, independent role is the elevator inspector. In most U.S. states, an elevator can't legally keep operating without a periodic inspection, commonly annual, performed and certified by someone holding a state or municipal license, and many jurisdictions specifically require the national Qualified Elevator Inspector (QEI) credential, issued by the National Association of Elevator Safety Authorities, on top of any state paperwork. The point of keeping inspection separate from maintenance is straightforward: the company servicing an elevator has some incentive to sign off on its own work, and an inspector who works for neither the building nor the maintenance contractor is a structural check against that.

Beyond mechanics and inspectors, vertical transportation engineers design traffic and equipment plans specific to individual buildings before anything is installed, and every modern elevator's cab includes a two-way emergency communication system, required by code, connected to a 24/7 dispatch or monitoring service that can talk a stopped passenger through what's happening and summon a mechanic or the fire department, whichever the situation calls for.

Where this came from

Hoists of some kind long predate anything resembling a modern elevator. Simple rope-and-pulley lifts, worked by hand, animal, or later water power, appear as far back as antiquity and were common in mines and monasteries by the Middle Ages. By the early 1800s, steam-powered hoists were standard equipment in European and American mines and factories, lifting ore, coal, and timber. What none of these systems solved was the one failure that mattered most: the rope itself was the only thing between the load and the bottom of the shaft. Hemp and early wire ropes wore, frayed, and occasionally snapped without warning, and a hoist built for freight simply wasn't trusted for people. Some German states banned rope-driven passenger lifts outright in 1859 over exactly this risk.

Elisha Otis, a mechanic working at a bed-frame factory in Yonkers, New York, built a solution: a wagon-spring-loaded ratchet mounted to the platform's frame that would spring outward and bite into a toothed guide rail the instant tension on the hoisting rope was released, whether that release came from the rope failing or simply going slack. In 1854, Otis staged what became the industry's founding demonstration, at the Crystal Palace exhibition in New York (a fair modeled on London's own, and often referred to in this context as the New York World's Fair). He rode a platform up in an open shaft in front of a crowd, then had an assistant cut the only rope holding it with an axe. The platform dropped a few inches, the safety catch engaged the rails, and it held. Otis is widely quoted as having announced, "All safe, gentlemen, all safe." Three years later, in 1857, he installed what's generally credited as the first passenger safety elevator in commercial service, in the E.V. Haughwout & Co. store in New York.

The demonstration mattered because it solved the specific problem holding back tall buildings, not general vertical transportation. Freight hoists already existed. What didn't exist, before Otis, was a passenger elevator anyone would trust with their life on a daily commute up and down a building, and without that trust, there was no point building higher than people could reasonably be expected to climb by stairs, somewhere around five or six stories in practice. Once a safe passenger elevator existed, architects had a genuine reason to build upward, and structural engineering (cast iron and, later, steel-frame construction) advanced in tandem with it over the following decades. The elevator didn't follow the skyscraper. It's closer to the reverse: the skyscraper became a realistic building type once the elevator solved the one problem that had made height impractical.

Standards and coordination

The primary safety standard covering elevators in the United States and Canada is ASME A17.1, formally the Safety Code for Elevators and Escalators, jointly developed with Canada's equivalent, CSA B44, since the two are published as a single harmonized document. It covers design, construction, installation, operation, inspection, testing, and maintenance, and it's revised on a regular cycle. The most recent edition at the time of writing, from 2025, is the code's twenty-fourth revision and adds requirements addressing cybersecurity and remote operation, on top of the older mechanical safety provisions. States and municipalities adopt A17.1 (or an older edition of it) into their own building codes, and enforcement, licensing, and inspection rules are then set locally: some states run their own elevator inspector licensing programs, others rely on the national QEI credential, and mechanics may need a state license in addition to union certification.

Accessibility requirements come from the Americans with Disabilities Act (ADA), which sets specific, measurable requirements for elevator cabs and controls in public and commercial buildings: a minimum cab size (commonly cited as roughly 51 inches deep by 68 inches wide, enough for a large wheelchair to enter and turn around), minimum door widths, control buttons mounted within a reachable height range and no smaller than a set diameter, raised lettering and Braille beside every button, and audible or visual signals confirming which floor the car is on. None of this is optional styling; it's federal civil rights law applied to a specific piece of equipment.

Keeping it working

Almost every commercial elevator operates under a recurring preventive maintenance contract, typically with the manufacturer or an independent elevator service company, under which a technician visits on a set schedule to inspect and adjust ropes, brakes, door operators, the motor, and the control system, catching wear before it becomes a malfunction.

On top of routine servicing, codes require periodic formal testing at two different intensities, commonly known in the industry as Category 1 and Category 5 tests. A Category 1 test, required annually, checks the safety devices without loading the car to its rated capacity or full speed. A Category 5 test, required roughly every five years, is far more demanding: the car is loaded to its full rated capacity and run at full rated speed while inspectors verify that the governor trips at the correct speed, the safety brake actually engages and holds under full load, and the shock absorbers ("buffers") at the bottom of the shaft function correctly. It's the closest thing to intentionally recreating a worst-case scenario under controlled conditions, on purpose, on a schedule, rather than waiting to find out during an actual emergency.

Hoist ropes themselves aren't replaced on a fixed calendar interval. Instead, mechanics inspect them on a recurring schedule (commonly monthly) for broken wires, corrosion, and wear, and codes set specific thresholds, such as a maximum number of visibly broken wires within a given length of rope, that trigger mandatory replacement regardless of the rope's age. A rope that passes inspection keeps running; one that doesn't gets replaced immediately, which is a more conservative approach than a blanket "replace every ten years" rule would be, since actual wear depends heavily on how hard a specific elevator is used.

When it breaks

The scenario most people picture when they imagine an elevator failing, a sudden, uncontrolled plunge to the bottom of the shaft, is extraordinarily rare precisely because of the redundant systems described above: multiple ropes each individually capable of holding the full load, plus a governor and safety brake that don't depend on the ropes at all. When news reports describe an elevator that "plunged" or went into "free fall," the more common reality, when investigated, is a car that dropped a short distance before the safety brake engaged, or a hard stop caused by a brake malfunction rather than an unchecked fall the full height of the shaft.

The statistics back this up. Analysis based on U.S. Consumer Product Safety Commission data, compiled in a widely cited safety report, put the accident rate at roughly 0.015 per elevator per year, compared with roughly 0.221 per escalator, meaning escalators produce well over ten times as many reported incidents per unit. Elevator-related injuries run around 17,000 a year in the U.S., with roughly 30 deaths annually, and a meaningful share of those deaths involve maintenance workers in shafts rather than riders in cars. Stairs, by comparison, cause more than a million injuries and roughly 2,000 deaths a year. By the numbers, riding the elevator is the safer choice, not merely the more convenient one.

The far more common failure mode is the entrapment: a car stops between floors, usually from a power interruption, a door sensor fault, or the elevator's own control system detecting an irregularity and deliberately halting rather than continuing to move. Fire departments generally distinguish between an "incident" (occupants are safe, uninjured, and not panicking) and an "emergency" (illness, injury, panic, or fire danger), and respond differently to each. A typical incident rescue involves confirming that an elevator mechanic has been called, cutting power to the car, opening the hoistway door on the floor above the stalled car, and lowering a rescuer, secured by a harness and rope, to help the occupant climb out through the car's emergency access panel into the shaft and up to the open landing. It is deliberate, procedural work, not an improvised climb.

The clearest large-scale test of these systems in practice was the Northeast blackout of 2003, which cut power across the northeastern United States and Ontario, affecting an estimated 50 million people. In New York City alone, hundreds of people found themselves trapped in elevators the moment power failed, and the city's fire department worked through the evening clearing stalled cars in an estimated 800 Manhattan high-rise office and residential buildings. It was, by any measure, a mass entrapment event, and it's also a useful illustration of what the safety systems are actually for: cars stopped safely in place rather than falling, occupants were rescued methodically rather than left in danger, and no elevator-related deaths from the blackout itself were reported. The system did exactly what it was designed to do under the worst plausible stress test: hold still.

The scale of it

The National Elevator Industry, Inc., the U.S. trade association for elevator manufacturers and service companies, estimates that the United States and Canada together operate somewhere in the neighborhood of 900,000 to a million elevators, and that U.S. elevators alone make approximately 18 billion passenger trips a year, a figure that works out to roughly 49 million rides a day. Framed differently, the average American who rides elevators at all takes something like four elevator trips on a typical working day, and a single elevator, on average, carries on the order of 20,000 people over the course of a year.

Those numbers put elevators in the same rough order of magnitude as major public transit systems, except the entire "network" exists inside individual buildings, is largely invisible as infrastructure, and moves people over distances measured in floors rather than miles. Escalators, worth mentioning for scale even though this chapter is about elevators, carry an even larger raw number of riders in North America (over 100 billion rides a year across a much smaller fleet of roughly 56,000 units), simply because they load and unload continuously rather than waiting for a car to arrive. Dense urban life, in its current form, of stacking tens of thousands of people into a single tower and expecting them to move between floors dozens of times a day without a second thought, is not really possible without this volume of elevator capacity quietly running underneath it.

Trade-offs and what's next

Destination dispatch, already common in new high-rises, is increasingly paired with predictive software that learns a building's actual traffic patterns (which floors fill up at 8:45 a.m., which empty out at noon) and pre-positions cars before demand spikes rather than reacting to it, along with mobile apps and badge readers that can call a car before a passenger even reaches the lobby. The tradeoff is added complexity and a control system with a lot more to go wrong, weighed against measurably shorter wait times in buildings tall enough for it to matter.

Regenerative drives address a different kind of waste. A traditional elevator motor burns off the energy of a heavily loaded descending car (or a nearly empty ascending one, since the counterweight is then the heavier side) as heat through resistors. A regenerative drive instead converts that excess mechanical energy back into electricity and feeds it into the building's electrical system, where it can power lighting, HVAC, or another elevator that happens to be working against its counterweight at that moment. It doesn't eliminate the elevator's power draw, but it meaningfully reduces the net energy the building has to buy for a machine that, in a busy tower, almost never stops moving.

The least settled trade-off is around fire evacuation. For decades, the near-universal rule, still posted on signage in most buildings, has been simple: never use the elevator during a fire, take the stairs. That advice exists because a hoistway can act like a chimney, drawing smoke, and because a car might open its doors directly onto the floor where the fire is. But stairs aren't a realistic evacuation option for many wheelchair users or people with mobility impairments, especially from high floors, a gap that became a legal and ethical problem once accessibility law required equal access to those same tall buildings. Since roughly the late 2000s, building codes have begun incorporating Occupant Evacuation Elevators (OEE) and Fire Service Access Elevators (FSAE), systems built with smokeproof, pressurized shaft enclosures, dedicated emergency power, and protection against water intrusion at the top of the hoistway, specifically so that elevators can be used to evacuate people, under controlled conditions, during an actual fire. These remain far from universal. Most existing buildings, including the vast majority built before this shift, still rely entirely on stairs, and the debate over how quickly to require retrofits, and how to protect people who simply cannot use stairs in the meantime, is still active in fire and building code circles.

Back to the elevator

The next time the doors close and the car starts moving, almost nothing about that ordinary thirty seconds is left to chance. Multiple ropes, each strong enough to carry you alone, are doing the lifting, balanced against a counterweight sized to make the motor's job as small as possible. A governor is quietly watching the car's speed the entire ride, connected to a brake that has one job and does not care whether the ropes are still intact. The whole arrangement traces back to a single deliberate demonstration in 1854, a man standing on a platform while someone else cut the rope holding it, done specifically to prove that a fall wasn't actually possible anymore. It's still true. That's why nobody thinks about it.

The leap: what it replaced, and the work behind it

Before a car anyone would trust, height was a penalty. Every floor above the ground meant another flight of stairs to climb with groceries, with a child, at the end of a shift, and rents tracked that reality straight down the building: the top floor was the cheapest, the least desirable, the place a landlord put the people who could pay the least. In ancient Roman apartment blocks the poor lived up under the roof; centuries later the pattern held, and a five- or six-story walk-up was about as tall as ordinary living got, because past six flights people simply would not climb. The word "penthouse" did not carry its modern meaning until the 1920s, and it originally named the servants' shed on the roof, next to the water tank. New York only legalized penthouse apartments in 1925, once wealthy residents noticed their servants had the best light and views in the building.

The safety brake Elisha Otis demonstrated in 1854 did not just make tall buildings convenient. It inverted the whole value of vertical space. Once a fall was no longer the thing you feared, the top floor became the quiet floor, the floor with the view and the clean air above the street, and the rent climbed with the elevation instead of falling. That inversion is now written into law: New York's 1929 Multiple Dwelling Law requires an elevator in any residential building taller than six stories, roughly the height people had always refused to walk. The safety record behind that trust holds up under counting. Analysis of U.S. injury data puts the elevator accident rate near 0.015 per elevator per year, well under a tenth the rate for escalators, and stairs, the thing an elevator replaces, injure more than a million Americans a year. The machine that feels like the risky choice is the safe one.

You feel the whole arrangement only when it stops. In June 2026, residents of a Miami-Dade condominium told reporters they had been stranded for more than a month by a broken elevator, elderly neighbors unable to manage the stairs relying on family to carry up food. During heat waves, senior high-rises have had every car fail at once, leaving people who cannot climb stuck outside in the heat or trapped upstairs, choosing between a sixteen-flight climb and not going home. That is the ordinary morning turned inside out: a doctor's appointment missed, groceries that cannot come up, a wheelchair user cut off from the street. On a normal day none of it registers, because the doors open, the car is there, and the top floor is the good floor.

The reason the good floor is up top, and the reason you never think about the stairs you are not climbing, is a governor watching the car's speed every second and a brake that does not care whether the ropes are intact.

Real-world examples and recent developments

The four major manufacturers named earlier keep pushing the same two problems, going faster at extreme height and cutting the machine room out of a building entirely, into new territory.

  • Mitsubishi Electric (2016): installed the elevators inside Shanghai Tower, China's tallest building, that Guinness World Records certified as the fastest elevator, the tallest elevator run, and the fastest double-deck elevator on Earth, reaching 20.5 meters per second (73.8 km/h) on the ride from the second basement to the 119th-floor observation deck. CNN, Shanghai Tower in China has world's fastest elevator
  • KONE: the Finland-based manufacturer's UltraRope, a carbon-fiber hoist rope roughly 80 percent lighter than steel rope of the same length and rated for far more flex cycles, lets a single elevator run higher without its own rope weight working against the motor, a limit that gets worse the taller a building is. Media OutReach, KONE launches High-Rise MiniSpace DX elevator in Southeast Asia
  • TK Elevator: developed MULTI, the world's first rope-free elevator system, replacing hoist ropes with linear motors so multiple cars can travel vertically and horizontally through the same shaft; it has been running as a live test installation since 2017 inside the company's 246-meter test tower in Rottweil, Germany. TK Elevator, MULTI, A new era of mobility
  • Otis: the company Elisha Otis founded now runs Otis ONE, a cloud-connected monitoring platform that uses sensors and machine learning to flag developing elevator problems before they cause a breakdown; by the end of 2025, Otis reported roughly 1.1 million of its elevators worldwide connected to the platform. Otis Worldwide Corporation Annual Report 2025

Recent developments

Glossary

Traction elevator. An elevator design in which steel ropes run over a motor-driven grooved wheel (the sheave), with the car on one end and a counterweight on the other.

Hydraulic elevator. An elevator raised by a piston pushed with pressurized fluid rather than ropes, common in low-rise buildings up to roughly seven or eight stories.

Counterweight. A weighted block, roughly equal to the car plus 40 to 50 percent of its rated load, that balances the car on the opposite end of the hoist ropes to reduce the motor's workload.

Sheave. The grooved wheel, turned by the traction motor, over which the hoist ropes pass to move the car and counterweight.

Governor. A speed-sensing device connected to the car that triggers the safety brake if the car exceeds a set percentage of its rated speed.

Safety brake (safety gear). Spring-actuated wedges mounted to the car frame that clamp onto the guide rails to stop the car if the governor trips, independent of the hoist ropes or building power.

Hoistway. The vertical shaft that a car and its counterweight travel through.

Machine-room-less (MRL) elevator. A design that places the drive motor inside the hoistway itself, eliminating the separate machine room used by older traction elevator designs.

Destination dispatch. A control system in which passengers enter their destination floor before boarding, letting the building group riders headed to nearby floors into the same car.

Category 1 / Category 5 test. Required periodic safety tests under ASME A17.1: Category 1 is an annual check of safety devices without full load; Category 5, required roughly every five years, tests the same devices under the car's full rated load and speed.

Qualified Elevator Inspector (QEI). A national certification, issued by the National Association of Elevator Safety Authorities, held by many licensed elevator inspectors independent of the maintenance company.

ASME A17.1. The Safety Code for Elevators and Escalators, the primary U.S. and Canadian safety standard covering design, installation, testing, and maintenance, harmonized with Canada's CSA B44.

Regenerative drive. An elevator drive system that converts the surplus mechanical energy of a heavily loaded descending (or lightly loaded ascending) car back into electricity for the building, instead of releasing it as heat.

Occupant Evacuation Elevator (OEE). An elevator built to a higher fire and smoke protection standard specifically so it can be used, under controlled conditions, to evacuate building occupants during a fire.

Sources and notes

Open questions

  • The widely repeated NEII figures of roughly 900,000 to a million North American elevators and 18 billion annual U.S. passenger trips trace in part to data compiled around 2007; treat them as order-of-magnitude estimates rather than an exact current count.
  • Regenerative drive energy savings are commonly cited by manufacturers in broad terms, but the actual savings depend heavily on a specific building's traffic pattern, and no single independently audited industry-wide savings percentage exists.
  • How quickly existing (not newly built) high-rises will adopt Occupant Evacuation Elevators, and how that timeline intersects with accessibility requirements for evacuating people with disabilities, was still an actively evolving area of fire and building code policy at the time of writing.

This closes the "Roads, traffic, and buildings" part of the book. From here, the book turns to a different kind of vertical and horizontal movement: getting people and packages across distances no building, road, or elevator shaft can cover on its own. Next: How an aircraft flies and how airports support it 👉