Burning to cool: COP31 has a cooling challenge

Chart of the month – November 2026. 35 by 35 is the mantra at COP31: Raise electricity’s share of global final energy from just over 20% today to 35% by 2035. Cooling electricity demand, meanwhile, is growing more than 60% over the next decade.

By 2035, cooling alone will be 13% of all electricity demand – double the aggregate demand for either EV charging or AI data centres. And, unlike EVs, cooling does not drive a transition away from fossil fuels. Even the energy demand for AI data centres, where the rush-to-power is favouring gas-fired generation in the US, at least comes with the thesis, if not the promise, of efficiencies, transformative productivity and added economic value. Cooling represents something different; increasingly it is adaptation capital, electricity consumed to protect health, productivity and economic activity against rising temperatures. Thus, an alarmingly large share of future electricity demand will be devoted not to making societies more prosperous, but to preventing societies from becoming poorer as temperatures rise.  

The challenge is compounded by the fact that we are burning fuels to cool ourselves and warming the planet as we do so. As the world adds billions of air-conditioner units, the electricity they draw will not be fully decarbonized for decades. Indeed, by 2035, emissions from cooling electricity alone reach 1.1 Gt CO₂ a year, roughly equivalent to the entire annual CO₂ emissions of Japan today.  

Cooling demand surges in emerging economies 

Global cooling demand is already at 2,800 TWh and grows 80% by 2040, compared with 55% over the last 15 years. Rising incomes and expanding commercial and residential floor space are important drivers. But so is a warming climate.  

The regional pattern for where cooling is growing is noteworthy. High income regions already have a high saturation of air-conditioning, limiting further demand growth. In North America, for example, cooling demand increases by just over 5% over the next 15 years, with data centre cooling accounting for much of the additional load. By contrast, cooling demand triples in the Indian Subcontinent and quadruples in South East Asia.  

Cooling is becoming essential infrastructure in some of the regions least able to afford its rapid expansion. During extreme heat, access to cooling protects health and economic continuity. Yet rapidly rising cooling loads also puts additional pressure on the power system, particularly when temperatures are highest. The challenge is not only the volume of electricity required but when it is required: cooling demand can surge across millions of buildings simultaneously during periods of extreme heat, precisely when power systems are under greatest strain.

Cleaner grids reduce cooling emissions, but not fast enough 

Launched at COP28 in 2023, the Global Cooling Pledge calls for cooling-related emissions to fall by at least 68% from 2022 levels by 2050, while rapidly expanding access to sustainable cooling.  

Our recently launched Energy Transition Outlook, a forecast to 2060, finds that CO₂ emissions from electricity used for space cooling decline substantially towards mid-century, but only by around 40% from 2022 levels Although this is a narrower emissions boundary than the Pledge’s 68% target (which includes HFCs) it shows how difficult the task is: even substantial grid decarbonization fails to outrun rapidly rising cooling demand.   

The Cooling Pledge 

The Third Annual Cooling Ministerial on 16 November is particularly consequential. Cooling sits at the intersection of two of COP31's headline ambitions: faster electrification and lower building-energy intensity. Effective and sustainable cooling is one of the greatest energy challenges of a warming world: keeping people cool without making the planet hotter. 

At DNV, we will intensify our research into this issue. Cooling is already a significant business concern. It is a major and rapidly growing source of electricity demand that generators, grid operators and distribution companies need to anticipate and plan for. But more fundamentally, it is a societal challenge. Access to cooling is increasingly a question of health, affordability and inequality. For low-income households in the world’s hottest regions, the economics are especially unforgiving. Those able to afford cooling at all may be pushed towards cheaper, less inefficient equipment, locking them into higher costs because they cannot afford the capital outlay for better technology.

Our analysis shows why this matters. Rising incomes are the single strongest driver of cooling-energy demand between 2025 and 2040, while improvements in equipment and controls offset only a small part of that growth. The Cooling Pledge seeks to change that trajectory, in part by raising the global average efficiency of new air conditioners by 50% by 2030 from a 2022 installed baseline. But an efficiency ambition is not an efficiency outcome.  

Governments will need to translate the Pledge into standards that change what is actually installed, including by progressively removing the least-efficient equipment from the market. Real progress is not about chasing an abstract global average but ensuring increasingly efficient cooling becomes the affordable default, especially in countries where demand is growing fastest. Without that, we risk meeting the world’s rapidly growing need for cooling by adding still more heat to the problem we are trying to solve.