On a summer afternoon in an Arab city, no economic report is needed to understand the meaning of demand for cooling. A few minutes outdoors, followed by a single step into an air-conditioned lobby, are enough to show that cool air is no longer a luxury in this part of the world. It has become a condition of daily life, alongside water and electricity.
Yet behind the cool air filling a home or office lies an investment story far larger than the unit mounted on the wall. When millions of air conditioners switch on at roughly the same time, loads surge across power stations, transmission lines and transformers. What looks like a simple individual decision—lowering the temperature in a room—becomes, when aggregated across a city or country, a multibillion-dollar challenge.
The world is currently preoccupied with the energy that AI data centres will require. Understandably so: the new servers are power-hungry, the announced investments are enormous, and the companies leading them are among the world’s largest. Yet another source of demand is growing away from the noise, with a much broader geographical and social base: cooling.
Data centres are concentrated mainly in the United States and China, alongside a limited number of other digital hubs. The need for cooling, by contrast, reaches every hot city—from a family buying its first air conditioner in India or Indonesia to a tower in the Arabian Gulf; from a pharmaceutical warehouse in Africa to a lorry carrying food across the Arabian Peninsula.
This is how what might be called the ‘cooling economy’ is taking shape. It is not just a fancy name for the air-conditioning industry, but a system that begins with power stations and extends to the design of buildings and cities. Between the two lie solar power, batteries and thermal storage; district-cooling networks; food and pharmaceutical cold chains; smart control systems; and the metals, refrigerants and technical skills that will be needed in ever greater quantities as the planet grows warmer.
Need vs purchasing power
Around 3.5 billion people live in regions exposed to high temperatures, yet only 15% of them own an air conditioner, according to the International Energy Agency. This deep gulf between need and purchasing power represents the largest reservoir of future demand.
Heat alone does not explain the spread of air conditioning. Income matters just as much. Once a household crosses a certain purchasing-power threshold, an air conditioner quickly jumps from the list of luxuries to the top tier of necessities. A family may postpone buying a car or new furniture but sleeping and working in oppressive heat make it harder to defer air conditioning with each passing year.
That shift is already underway across large parts of Asia and will repeat as incomes rise and urbanisation accelerates in Africa and the Middle East. The IEA expects emerging and developing economies to account for more than 80% of the projected increase in electricity demand for cooling through to 2050.
In South-East Asia, the stock of air conditioners could grow ninefold between 2020 and 2040 under current policy settings. In Indonesia, the share of people who own one could rise from 14% in 2023 to 85% by mid-century. In India, air-conditioner ownership is projected to increase tenfold by 2050, while peak demand in buildings rises sixfold.
The impact goes well beyond the price paid at the shop. Every air conditioner installed in a home creates years of electricity demand and requires generation capacity, a grid able to bear the load, maintenance technicians, spare parts and refrigerants. A cheap machine at the point of purchase may prove the most expensive over the long term if it is voracious in its use of electricity.
Not every need, however, becomes a market. According to the 2025 Chilling Prospects report by Sustainable Energy for All, just over 1 billion people in 77 countries face high risks because they lack essential cooling. That number could reach 1.05 billion by 2030.
Alongside them are 2.83 billion people in a middle category: their incomes allow them to consider buying some form of cooling, but they cannot always find an efficient appliance at an affordable price, reliable electricity or finance to cover the upfront cost. This vast population is the market waiting to be opened up; the danger is that it will be flooded with cheap, energy-hungry machines, leaving households to pay many times the difference through their electricity bills.
Beyond the socket plug
Cooling currently consumes around 10% of the world’s electricity and close to one-fifth of the electricity used in buildings. Its effect on the power system is greater than those shares suggest, because consumption is not distributed evenly across the year. It is concentrated on the hottest days, usually at much the same time.
The IEA says cooling can account for about 30% of global peak electricity demand on average. In some hot regions, including the Middle East, it can exceed 70% of peak residential demand on the hottest days.
This is where the most expensive part of the cooling economy begins. A power company cannot build its network around average annual consumption. It must prepare for the hours when loads reach their highest levels, or risk outages at the very moment air conditioning becomes most essential.
That may mean building a power station, transmission line or transformer that is used at full capacity for only a limited number of summer hours. Societies pay for those investments throughout the year, even though their peak need is seasonal.
It is therefore difficult to measure the cooling economy by the number of appliances sold. Most of the spending takes place behind the plug socket: in power stations, networks, batteries and smart meters; in programmes that encourage consumers to reduce or defer consumption at peak times; and in buildings designed to require less cooling in the first place.
The World Bank estimates that the cooling market in developing economies could grow from about $300bn at the start of this decade to $600bn a year or more by 2050. That includes residential and commercial space cooling, industrial refrigeration, cold storage and refrigerated transport, but does not necessarily capture all the indirect investment in power systems, property and minerals.
Rising consumption
The figures reveal part of the difference. Data centres consumed about 415 terawatt-hours of electricity in 2024, equivalent to roughly 1.5% of global consumption. According to the IEA’s latest estimates, their consumption rose to about 485 terawatt-hours in 2025 and is projected to approach 950 terawatt-hours by 2030—around 3% of global demand.
This is an enormous increase in a short period, with AI the fastest-growing driver. Even so, the IEA says the growth in data-centre consumption will account for less than 10% of the global increase in electricity demand between 2024 and 2030. Industry, electric vehicles and air conditioning will remain among the larger drivers.
This does not mean the market value of cooling companies will overtake that of the technology giants, or that investors will turn away from AI. The difference lies in the pattern of diffusion. Much of the digital economy’s value is concentrated in chips, software and cloud platforms, whereas the gains from cooling are spread across electricity companies, renewable energy, property, construction, transport, agriculture, retail, metals industries and the banks financing the projects.
A technology company can choose a data centre site based on the availability of power, land, and connectivity. Cooling, however, must reach the places where people live, work and preserve their food and medicines. Its economy therefore appears less concentrated and more closely tied to daily life and the physical economy.
The separation between the two is not complete. AI needs to cool its servers, while it can also forecast loads, regulate air conditioners, detect faults before they occur and improve the operation of buildings and plants. At the same time, the two economies compete for electricity, transformers, copper, land and capital.
Natural partner
Solar power appears to be an ideal partner for cooling. Its output rises during the day, when temperatures climb, and offices, factories and shopping centres are operating. A large share of that electricity can be used directly to power air conditioners and chillers.
The match is not perfect, however. In hot cities, high residential demand continues after sunset, when people return home and solar output declines. That creates a need for batteries, demand management and more flexible grids—as well as a less familiar solution: storing the cooling itself.
A central plant can chill water or make ice when solar power is abundant, or tariffs are low, then draw on that stored thermal energy during peak hours. In some applications, storing cooling in a chilled or frozen material is cheaper than storing electricity in a battery and later using it to run cooling equipment.
The cheapest cooling energy, however, is the energy that never needs to be produced. Insulation, shading, reflective glazing, cool roofs and good ventilation reduce the heat entering a building before the air conditioner is switched on. The United Nations Environment Programme estimates that passive-cooling measures can lower indoor temperatures by between 0.5°C and 8°C and, in many cases, pay for themselves within two to eight years.
Hybrid systems combining fans and air conditioners can reduce energy use by about 30%. Architects, glass and insulation manufacturers, and software companies managing buildings are thus entering a market that was long seen as the preserve of appliance makers.
Buying cooling
One of the most important investment shifts may come from a simple idea: rather than buying an air conditioner, customers buy the cooling they need. Under a ‘cooling as a service’ model, a specialist company finances, owns, operates and maintains the equipment, while the customer pays according to the amount of cooling used or the level of service provided. Instead of choosing the cheapest machine at the point of purchase, the operator has an incentive to install a more efficient and longer-lasting unit, because savings on electricity and maintenance become profit.
The model removes the upfront cost from the building owner and turns cooling equipment into an asset that generates long-term cash flows, making it financeable by banks, sovereign funds and infrastructure investors.
District cooling is its most prominent application. Instead of installing a separate system in every building, a central plant produces chilled water and distributes it through a network of pipes. This allows larger, more efficient equipment and central storage of cooling.
It is not a magic solution for every location. A plant needs high urban density to be viable, substantial initial investment and an accurate estimate of demand. It also requires clear regulation, because a customer connected to a single network cannot easily switch providers if prices rise or service deteriorates.
