Air Conditioning Could Add Equivalent of U.S. Annual Emissions to Warming by 2050

The same device that protects people from lethal heat is accelerating the warming that makes that protection necessary. Air conditioning is, in this sense, both response and cause - a feedback loop that researchers have long recognized but rarely quantified with the precision now published in Nature Communications.

An international team led by the University of Birmingham combined climate science, energy modeling, and inequality analysis to project how air conditioning demand will evolve across five different global development pathways through 2050, and how much additional warming the resulting emissions would contribute in each scenario.

The numbers in 2050

Current air conditioning use is already substantial, and the trajectory runs sharply upward. The modeling projects that AC use will more than double by mid-century. Electricity demand for cooling could reach 4,493 terawatt-hours under mid-range scenarios - substantially more under high-emissions pathways.

Emissions from air conditioning could reach 8.5 billion tonnes of CO2-equivalent per year in worst-case scenarios. For comparison, current annual emissions of the United States total approximately 5.9 billion tonnes. The temperature contribution from AC emissions specifically is estimated at 0.03 to 0.07 degrees Celsius of additional warming by 2050. The team translates this as equivalent to 74 to 183 billion transatlantic return flights.

Professor Yuli Shan, the corresponding author, described the dynamic directly: "Global warming is raising temperatures and causing more heatwaves, and as economic growth in some of the worst-affected countries means more people can offset extreme heat with air conditioning. As global temperatures rise, we risk being locked into an 'arms race' where defending ourselves against extreme heat is causing the issue to get worse."

What is driving the growth

Rising temperatures are a contributing factor, but they are not the dominant driver of AC demand growth. "Most of this extra warming is caused by income-enabled growth in cooling consumption, more households adopting and using air conditioning, and not just rising temperatures," the Birmingham team found.

This distinction matters for policy. If temperature rise alone drove adoption, then slower climate change would automatically reduce cooling demand. But if economic development and rising incomes are the primary drivers, then even successful mitigation of warming would leave a large fraction of the demand increase intact.

The inequality problem

The analysis surfaces a profound mismatch between cooling need and cooling access. South Asia and Africa face among the highest heat stress levels globally. These regions also have among the lowest rates of air conditioning ownership and use, driven by lower incomes and unreliable electricity supply.

Closing this gap would have climate costs. If all low-income regions gained the same air conditioning access as high-income regions, the modeling estimates an additional 94 million units at medium income levels, 150 million at high-income levels, and over 220 million at the highest-income levels. The additional warming from equitable cooling access would reach 0.05 degrees Celsius even in the most climate-friendly scenario analyzed.

This is not an argument against equitable access to cooling. Extreme heat kills people, reduces economic productivity, and disproportionately affects the elderly and the poor. It is an argument for urgency in transitioning the cooling sector itself to cleaner technology before the global demand surge arrives.

Paths forward

The research team outlines several parallel interventions. Rapid decarbonization of electricity grids is the most impactful single change. Transition to refrigerants with lower global warming potential is a second priority. Better building design, including insulation, shading, and passive cooling features, can reduce how much active cooling a building requires.

"The world must transition quickly to cleaner, more efficient cooling technologies - while ensuring fair access to cooling, especially for vulnerable populations," Shan said. The projections carry uncertainty inherent in long-range scenario modeling - the five SSP-RCP scenario brackets are intended to capture a range of plausible futures rather than predict a specific outcome.