It was only 10am, but the searing heat was already burning intensely against my skin. I don’t remember mornings like this growing up.

Back then, the sun was something we ran under, not something we fled from. We played konda-kondi and rounders in open fields—long before anyone spoke about UV rays or heatwaves.

Each morning, as I left for school, my family’s only reminder would be: “Keep your uniform clean.”

But by the time I crept back home that uniform would be stained with mud and sweat. I would try to slip in unnoticed, but would fail miserably.

Today, children leave for school in crisp uniforms, and return with them in the same condition. There are no grass stains, no stories written on fabric. Our children’s playgrounds having shifted from fields to screens.

Perhaps this is where the story begins to turn, as we rarely connect our digital habits to climate change.

The more we retreat into our phones, the more heat our technology emits, and this is where a quiet cycle begins.

The infrastructure powering our online lives, from social media to artificial intelligence, is energy-intensive.

Behind every search, every scroll, and every AI-generated response lies a network of data centres working relentlessly.

And those systems are, quite literally, thirsty.

According to the US-based Environmental and Energy Study Institute, large data centres can use up nearly 19 million litres, of water per day—equivalent to the consumption of a town with a population of between 10,000 and 50,000 people.

Meanwhile, the International Energy Agency reports that large data centres typically operate in the 10 to 100 megawatt range. Every 10 megawatts can power about 8,000 to 10,000 homes.

But technology is not the villain. It has connected us, informed us, and, in many ways, improved our lives. But it will continue to have an impact on the environment.

So the question is not whether humans are “killing” themselves, but whether we have adequate policies to keep our future cool.

Cybersecurity and AI expert Selvakumar Manickam told me recently that AI could account for about 3% of global energy demand by 2030, with consumption set to rise.

Drawing parallels with cryptocurrency mining, he said there is little transparency over how much power and water data centres consume, with minimal monitoring in place.

He said policymakers must set limits on energy and water consumption and require companies to disclose usage levels.

Selvakumar said energy-intensive processes should be shifted to off-peak hours, such as late nights or weekends, to ease pressure on the national grid.

He said greater investment is needed in research to develop more efficient algorithms that require less power.

“If there are no limits, companies will continue to consume freely.

“With caps, they will be forced to optimise operations or invest in more efficient infrastructure,” he said, adding that early intervention is crucial to avoid repeating mistakes seen in other high-growth sectors.

The last thing we want is for intervention to take place only when it is too late.

By then, the impact would be irreversible, and the heat would no longer just be something we feel.

It would be something we failed to act on.

The writer is a senior journalist at FMT’s English Desk.

This article represents the writer’s opinion and does not necessarily reflect FMT’s position.