From closing window to overshoot: The retreat on climate action

The Emissions Gap Report's suggestion that the world can cross the 1.5°C global warming threshold, but still limit, adapt to and return from higher temperatures, is utterly misleading

For the past several years, the United Nations Environment Programme (UNEP) has been issuing annual Emissions Gap Reports, tracking the progress being made in mitigation of greenhouse gas emissions pursuant to the Paris Agreement of 2015. The titles of the reports over the past five years tell a depressing story of fading ambition and systematic retreat from the solemn commitments made in Paris:  2022, The Closing Window; 2023, Broken Record; 2024, No More Hot Air Please; 2025, Off Target; and 2026, The Point of Return.

 

The window for limiting global warming to 1.5°C now lies closed; it has already been exceeded. The goal has now shifted to managing the overshoot. The latest report describes the “overshoot, peak and decline pathway as the best remaining option to minimise the magnitude and duration of this overshoot and try to return to below 1.5°C as soon as possible.”

 

But “as soon as possible” may be decades. “The overshoot duration can vary but typically last for at least multiple decades.”

 

The Report envisages an eventual return to 1.5°C of global warming only by 2100, but shies away from spelling out the several catastrophic and irreversible impacts in the intervening years, not only on the climate but on planetary ecology as a whole. 

 

If this is the best remaining option, are we not accelerating our headlong plunge towards a planetary catastrophe? Instead of stating the obvious, the Emissions Gap Report offers the comforting but utterly misleading statement: “The world set to cross 1.5°C global warming, but can still limit, adapt to and return from higher temperatures.”

 

The question is not whether we could return to a lower temperature in the future but what happens to the earth’s ecology as a whole in the meantime. A lower temperature in a ravaged world would not be of much consequence.

 

The much-celebrated Paris Agreement was considered by the then French President, François Hollande, to be more path-breaking than the French Revolution of 1789. In reality, the outcome was modest. It committed the world to keeping global warming to within 2°C by 2050 and preferably within 1.5°C, recognising that the latter limit would be “a significantly safer defence line against the worst impacts of a changing climate”.

 

A Special Report of the Inter-Governmental Panel on Climate Change (IPCC) issued in 2018 stated that keeping to the 1.5°C limit would “dramatically reduce the most catastrophic, irreversible impacts” of climate change. To keep to this limit, emissions would need to be reduced by 45 per cent by 2030, compared to the level pertaining to 2010. Based on the updated pledges by countries party to the Paris Agreement, we may end up with 58.5 per cent higher emissions by 2030, against the 2010 baseline. Paris did inaugurate a revolution but one of diminishing expectations. The historic Rio Convention of 1992 having been eviscerated by its most powerful parties, the Paris Agreement, too, has been hollowed out unceremoniously.

 

In order to enable “overshoot, peak and decline”, massive quantities of legacy CO₂ would need to be directly removed from the atmosphere, “creating a state of net negative emissions”. Over the course of this century, UNEP estimates that roughly 320 billion tonnes of CO2 will have to be sucked out from the air. This is equivalent to eight years of total CO₂ emissions at current levels. The Report refers to several technological fixes that may lead to the anthropogenic removal of atmospheric carbon dioxide. There is the Carbon Dioxide Removal (CDR) either through Direct Air Capture (DAC) or bio-energy with storage (BECCS). These are not yet technologically or economically viable. There are ethical issues involved too, for example, in the deployment of geo-engineering. Preserving forests, undertaking afforestation and cleaning up the oceans offer tried-and-tested natural alternatives, as forests and oceans are natural carbon sinks. Forests and oceans, are each responsible for absorbing 30 per cent of global carbon emissions. But they are being degraded instead at an accelerating pace.

 

What about artificial intelligence (AI), which is touted as the panacea for all the formidable challenges humanity is confronted with?

 

Whatever be its promise for the future, AI is currently worsening the climate crisis. AI operations depend on data centres, which store and process enormous volumes of data almost instantaneously. They require a continuous supply of power, with reliable back-up, to run the large array of computers and vast volumes of water to provide cooling. This demand for power and water is increasing at 15 per cent per annum but is likely to accelerate significantly with new, more powerful versions of AI being launched.

 

In 2025-26, power consumption at data centres worldwide was 485 terawatt-hours (Twh). If it were a country, it would be the fifth-largest consumer of power in the world, ranking somewhere between Russia and Japan. It is estimated that by 2030, the demand for power for AI would be 945-960 Twh and this would constitute 3 per cent of all global power demand.

 

The direct onsite consumption of water by data centres was 222 billion litres in 2024-25 and is estimated to reach 644 billion litres by 2030. A single 100Mw hyperscaler data Centre consumes roughly 2 million litres of water a day.

 

These enormous demands for power have led to fossil fuel-based power generation to be revived, although the possibility of using small modular nuclear reactors is being explored. This is for the future.

 

China has an advantage in its ambitious pursuit of green energy and nuclear power, even while coal is currently its main source of energy. If India wishes to be a leader in AI, it must resolve the energy-water constraint it currently confronts.  This would inevitably put climate change action on the back burner.

 

The “overshoot, peak and decline pathway” is no pathway at all and AI does not offer any credible solution to the climate crisis. Technological hubris blinds humanity to its insignificance in the vastness of the universe and the overwhelming power that Nature can unleash to snuff out existence in an instant. The irony is that this sensibility is deeply ingrained in Indian culture and intellectual tradition but is ignored. It needs to be resuscitated.

 

The writer is a former foreign secretary and has served as India’s chief negotiator for climate change