The Forgotten History of the Air Conditioner: How Cooling Reshaped Where People Live

Willis Carrier did not set out to invent the air conditioner. He set out, in 1902, to solve a humidity problem for a Brooklyn printing company whose color registration was being thrown off by the way paper expanded and contracted with summer humidity. His solution, the first device that mechanically controlled both temperature and humidity together, was the prototype for what would become one of the most consequential technologies of the twentieth century, and one of the most invisible.

The story of air conditioning is mostly the story of consequences. The invention itself was the product of a specific industrial demand, but the technology spread far beyond that demand once the manufacturing economics caught up, and the spreading reshaped human geography in ways that the textbook history of technology rarely treats as a primary effect.

The industrial era

For the first thirty years after Carrier's 1902 device, air conditioning was an industrial technology. The customers were printing plants (humidity for paper handling), textile mills (humidity for fiber processing), tobacco warehouses (humidity for leaf curing), pharmaceuticals (temperature and humidity for chemical stability), and a few prestige commercial buildings (department stores wanted summer customers, movie theaters wanted summer audiences). The 1925 installation in the Rivoli Theater in New York was a marketing event; the New York Times reported on the line of summer customers who came specifically to experience the cooled air, regardless of what movie was playing.

The technology in this era was bulky and expensive. The compressors were industrial-scale; the refrigerants were toxic ammonia, methyl chloride, or sulfur dioxide; the installation required dedicated mechanical rooms and skilled operators. A household air conditioner was technically possible but commercially unviable; the units were too large and the operating cost too high for residential use.

The Freon era

The transformation was the 1928 invention of Freon by Thomas Midgley Jr at General Motors's Frigidaire subsidiary. Freon (chlorofluorocarbon CFC-12, specifically) had the properties needed for residential refrigeration: non-toxic, non-flammable, efficient at residential-scale temperatures and pressures, and producible at industrial scale. Midgley's demonstration of safety involved inhaling Freon and exhaling it onto a candle to show it was both non-toxic and non-flammable, a performance that would not pass modern safety review but was effective theater for the time.

The downstream consequences of Freon were enormous. Within a decade, household refrigerators went from a rare luxury (under 10 percent of American households in 1930) to standard equipment (50 percent by 1940, 90 percent by 1950). The same chemistry enabled residential air conditioning; window units appeared in the late 1930s, central air systems in the 1950s. The technology that had been industrial-only became consumer-grade in a working generation.

Midgley's worst-luck pairing also produced tetraethyl lead, the gasoline additive that prevented engine knock and produced two generations of lead poisoning. Between Freon and leaded gasoline, Midgley contributed two of the largest environmental disasters of the twentieth century, both with substantial public-health consequences that the original development did not anticipate. The 1974 Molina-Rowland paper on stratospheric ozone depletion identified CFCs as the cause; the 1987 Montreal Protocol phased them out, and the modern refrigeration industry has cycled through HFCs and HFOs and is now returning toward natural refrigerants like CO2 and propane that were used before Freon.

The geographic transformation

The 1950s-1970s diffusion of residential air conditioning produced a population shift that is one of the most significant demographic events in American history, and one that is rarely framed as a consequence of HVAC technology. Before air conditioning, the population of the United States was concentrated in the Northeast, Midwest, and California. The South and Southwest were sparsely populated relative to their land area, partly because their summer climates were genuinely brutal in ways that limited what kinds of economic activity were viable.

From 1950 to 2000, the Sun Belt states (Florida, Texas, Arizona, Nevada, Georgia, North Carolina, South Carolina, and the rest of the southern crescent) grew by hundreds of percent while the Northeast and Midwest grew modestly or shrank. Houston went from 600,000 people in 1950 to 2 million by 2000 and 2.3 million by 2020. Phoenix went from 100,000 in 1950 to 1.6 million by 2020. Atlanta went from 300,000 to 500,000 (city proper) but the metro grew to 6 million.

Air conditioning was not the only cause of the Sun Belt migration; cheap land, post-war federal investment, defense industry concentration, the interstate highway system, and changing labor markets all contributed. But air conditioning was a necessary condition. Without reliable residential and commercial cooling, the kinds of economic activity that drove Sun Belt growth (corporate headquarters, electronics manufacturing, software development, retiree communities) would not have been viable at the latitudes where they happened. Phoenix in particular is a city that essentially could not exist as a major metropolitan area without air conditioning; the summer temperatures regularly exceed 110 degrees Fahrenheit, and the population growth tracks the residential air conditioning saturation curve almost exactly.

The energy and emissions story

Air conditioning is now the largest single category of electricity consumption in many warm-climate countries. Cooling accounts for roughly 10 percent of global electricity use; in the United States it accounts for roughly 15 percent of residential electricity. The cost of summer peak demand in air-conditioned regions shapes grid investment, fuel mix decisions, and emissions profiles in ways that the technology's history did not anticipate when it was just a humidity control device for printing plants.

The future trajectory is upward. Air conditioning saturation in India, Southeast Asia, sub-Saharan Africa, and Latin America is currently low, and rising as those regions urbanize and grow wealthier. Climate change is increasing the cooling demand even in regions already saturated; warmer summers in northern Europe are producing residential air conditioning adoption in countries where it was historically rare. The International Energy Agency projects that global air conditioning electricity demand will roughly triple by 2050.

The mitigations under research include heat-pump efficiency improvements (modern variable-speed units are 2-3x more efficient than 1980s fixed-speed units), district cooling (centralized chillers serving neighborhood loops), passive design (better insulation and shading reducing baseline cooling demand), and absorption refrigeration powered by waste heat. None of these is currently displacing the standard vapor-compression cycle at scale, and the technology that Carrier prototyped in 1902 is still the dominant residential and commercial cooling architecture in 2026.

Three observations

First, the chronology from industrial invention to consumer ubiquity is a working generation. The 1902 invention reached middle-class American homes in roughly 1955; the 1955 home reached global middle-class homes in roughly 2010 in the regions where it has saturated, and is still expanding in the regions where it has not. The 50-60 year diffusion timeline is typical for household-scale technologies; television, automobiles, and refrigerators all show similar curves.

Second, the geographic consequences of household-scale technologies often dwarf the technical consequences. Air conditioning is most often discussed as a comfort and productivity technology, but its largest single effect was probably the population shift it enabled. The same pattern recurs in other technologies; the automobile is discussed as transportation, but its largest single effect was probably the geographic dispersion of the suburban-and-exurban housing pattern that depends on it.

Third, the environmental consequences of household-scale technologies are often delayed by decades from the original invention. The Freon-as-ozone-destroyer story was fifty years after Freon was introduced; the air-conditioning-as-emissions-driver story is now playing out roughly 75 years after the technology became consumer-grade. The pattern of environmental consequences emerging on multi-decade timelines is one of the underappreciated regularities of industrial technology, and one of the reasons that present-day environmental decisions about new technologies are unusually high-stakes.

The deeper observation

The deeper observation is that some inventions become more consequential the more invisible they become. Air conditioning is now so embedded in modern infrastructure that its absence in a major metropolitan area during a heat wave is treated as a public-health emergency, but its presence is treated as background. The buildings, the supply chains, the population distributions, and the energy systems of the warm-climate world are all built around the assumption of reliable mechanical cooling. The 1902 humidity-control device for a Brooklyn printing plant is, in this sense, one of the load-bearing technologies of the modern world, and one of the ones that the modern world is least likely to acknowledge.