AI Data Centres and Water Crisis

Imagine telling someone ten years ago that a conversation with a robot would need water.

Lots of it! They would probably laugh.

But, here we are.

If you've opened social media recently, you've probably seen people arguing about the water crisis.

Mumbai, with depleted reservoirs leaving residential areas without daily running water. Just a few hours away, Pune has cracked down on non-essential consumption by cutting off supply to public swimming pools and construction sites. And, Bengaluru's over-reliance on rapidly depleting underground aquifers has forced suburban residents to depend entirely on water tankers.

Some blame climate change. Some blame rapid urbanisation. Some point to disappearing lakes and unchecked construction.

And then there is a newer theory making its way through comment sections and WhatsApp groups:

AI. The infrastructure behind it. The giant data centres. 

In February 2026, New Delhi hosted the India AI Impact Summit, bringing together world leaders, policymakers, technology companies, researchers and investors. 

And it shows that the country wants to actively shape it, major investments are already proving that. 

Microsoft is expanding its AI and cloud infrastructure across the country. Amazon continues to grow its presence. Google is moving ahead with major projects, including a large data-centre development near Visakhapatnam. Reliance has unveiled plans for a gigawatt-scale AI hub in Jamnagar. Adani has committed tens of billions of dollars towards data infrastructure across multiple states. AirTrunk recently announced a $30 billion investment push into India's data-centre market.

Behind Every AI Prompt

While using an AI assistant to summarize documents, translate text, or generate images- it feels weightless. But, it’s not. 

The reality is much more physical. 

It lives inside giant, boring buildings packed with rows of computer servers running day and night. Every single prompt we send makes these machines work hard, and that generates a ton of heat. 

If they get too hot, they crash.

To prevent that, they have to put massive effort into cooling, mostly by using water. A typical 100-megawatt facility can easily gulp through two million liters of water every single day, and the next generation of AI hubs will be even larger.

Don’t worry, AI isn't about to drain water supply overnight. But it does mean water is now a huge part of the AI conversation. 

Location Matters More Than You Think

More than half of India's data centres are located in water-stressed regions.

According to analysis from WRI India, roughly 75% of the country's data centres are concentrated across just five states- Maharashtra, Tamil Nadu, Karnataka, Telangana and Uttar Pradesh.

Those are not just random locations.

They are home to some of India's largest tech hubs.

• Mumbai

• Bengaluru 

• Chennai

• Hyderabad

• Delhi NCR

  
  

They are also places that have repeatedly appeared in news reports about water stress, groundwater depletion, drought concerns and growing urban demand.

But, India's water challenges did not begin with AI.

• Mumbai's reservoirs were under pressure long before ChatGPT existed.

• Bengaluru's groundwater issues were already making headlines years before the current AI boom.

• Chennai's famous "Day Zero" crisis happened in 2019.

-population growth, 

-rapid urbanisation, 

-disappearing lakes, 

-changing rainfall patterns and years of groundwater extraction,

-poor water conversation planning

all contributed to those problems.

So, these fancy computer brains did not create them. But yes, it’s showing up at a time when tons of cities are already struggling to share the water they have.

The Numbers Are Getting Bigger

Government figures show data-centre capacity growing from around 0.4 gigawatts in 2020 to approximately 1.5 gigawatts by 2025. Industry forecasts suggest that figure could reach between 8 and 10 gigawatts by the end of this decade.

Researchers estimate Indian data centres consumed around 150 billion litres of water in 2024-25. 

By 2030, that figure could rise to roughly 358 billion litres annually if current growth trends continue.

This massive spike is happening just as India rolls out one of the largest digital infrastructure programs in the world. Through the IndiaAI Mission, the country is adding tens of thousands of powerful computer chips. 

State governments are actively competing for these projects, while tech companies are racing to lock down land, power, and fast internet connections.

Right now, the entire focus is on speed:

• Who can build first?

• Who can expand fastest?

• Who can pull in the next major investment?

Because everyone is trying to move as fast as possible, water rarely gets the same attention.

How Other Countries Are Responding

India is far from the only country staring down this reality. Governments worldwide have already realised that the entire AI boom is being sold as a software story, but it is actually a massive infrastructure story. 

And infrastructure always has a physical cost. You can’t run the world’s most powerful algorithms without a heavy reliance on real-world resources.

We are already seeing the global landscape change because of this. 

Take the UK, for example, where planners are mapping out new "AI Growth Zones," where they are looking hard at power grids, land access, and local water resources before a single brick is laid. 

Over in Singapore, the government has essentially put its foot down, linking any future data center expansions directly to strict green standards.

Meanwhile, parts of Europe are pushing for full transparency, demanding that tech companies pull back the curtain on the exact environmental footprint of their digital networks.

Even in the US, massive data center projects that used to be welcomed with open arms are now facing real pushback from nearby communities worried about what these facilities will do to local electricity grids and water tables.

What is fascinating is that this isn't driven by tech-hating critics who want to stop progress. It is coming from ordinary residents who are just asking practical questions:

• How much water is this facility going to pull from our system every day?

• Where exactly is that water going to come from?

• What is the backup plan when a drought year hits?

• If water supplies actually run low, who gets priority- the local community or the servers?

  
  

These aren't anti-AI questions at all. Instead…smart, necessary planning questions that we probably should have been asking all along.

Taken together, these point to an important shift.  Now, attention is turning towards how that infrastructure should be planned, where it should be located and what information operators should disclose about the resources they consume.

India is not alone in facing these questions.

As money pours into this new infrastructure all over the world, governments, regulators, and industry groups are all trying to figure out what responsible growth actually looks like in practice.

Right now, there is no single global rulebook. 

Different countries are basically experimenting with their own approaches to see what sticks. But amidst all the trial and error, one thing is clear that it has officially grown out of the IT department and stepped into the real world, becoming a major challenge for city planning, environmental sustainability, and everyday governance.

Looking Beyond 

The data centers being built today are not short-term projects. Many of these giant buildings will still be running years, or even decades, from now. That means everything that comes with them:

• the cooling setups, 

• the power lines, 

• and the heavy resource needs - is pretty much locked in for the long haul.

At the same time, a lot of the cities getting these tech investments are already dealing with problems like struggling with water shortages, overloaded power grids, and very fast city growth. 

Now, to be totally fair, this does not mean AI suddenly caused these problems. And it definitely doesn't mean countries should just completely stop building these infrastructure.

If anything, the exact opposite is happening. Governments and companies all over the world are spending more money than ever on these projects.  They see this tech as a must-have for economic growth, new ideas, and staying ahead of the competition. 

In the end, the questions about the basic resources needed to keep those machines running are only going to get louder.