The Hidden Cost of Your AI Chat: How a Single Prompt Is Quietly Draining the World’s Drinking Water
You typed a question. Somewhere, a server farm got hotter, a cooling tower hissed, and a few drops of drinking water disappeared forever. This is the story nobody tells you when you open ChatGPT, Gemini, or Claude — and why it matters more for India than almost anywhere else on Earth.
Do you also believe that AI has made your life easier? If your answer is yes, pause for a second — because behind every convenience, we are quietly paying a price that almost nobody talks about. Every time you ask ChatGPT or Google Gemini a question, you are spending a sliver of the world’s drinking water. When OpenAI founder Sam Altman was once asked how much water a single ChatGPT prompt consumes, his answer was striking: roughly one-fifteenth of a teaspoon. In other words, every question you ask quietly costs a drop of water that belonged to someone, somewhere.
But here’s the catch — experts say even that number understates reality. Use the advanced, more powerful versions of these AI tools, and the water cost climbs sharply. Add it all up, and somewhere between 10 and 15 prompts on any AI tool can burn through nearly a litre of water on your behalf. The culprit isn’t the question itself — it’s the army of data centers racing around the planet to process it.
And we’re not just talking about your one prompt. We’re talking about the unimaginable volume of data processed every single day by thousands of data centers scattered across the globe — banking systems, government records, your social media feeds, and yes, every AI prompt ever typed. This data is measured in zettabytes, and one zettabyte equals one trillion gigabytes. Processing that much information, every single day, quietly drains away crores of litres of water that should have gone to people.
The Numbers That Should Worry Every One of Us
According to a United Nations report, by 2030, data centers worldwide will consume 9.3 trillion litres of water annually — enough to keep 1.3 billion people hydrated for an entire year. That is not a typo. That is the scale of thirst hidden inside the machines that answer your questions in milliseconds.
Why are we telling you this story now? Because the next big destination for these data centers is India. Google, Meta, Microsoft, and other tech giants are setting up shop here, and the government is rolling out the red carpet — free electricity, free land, and a tax holiday stretching all the way to 2047. But here’s the unsettling part: the MOUs being signed with these companies say almost nothing about how much water these data centers will consume annually, or where that water will come from.
This isn’t only a story about water. It is the story of data centers multiplying like mushrooms after the AI boom — and their next destination is our own backyard, India.
What Actually Happens Inside a Data Center?
Before going further, it helps to understand what a data center actually is. At its core, it has four components:
- The server room — rows of powerful computers that collect, store, and process data around the clock.
- The power grid connection — supplying uninterrupted electricity 24 hours a day, 7 days a week.
- Diesel generator sets — the backup that kicks in the moment grid power fails, keeping servers alive.
- The cooling station — arguably the most critical and most water-hungry part of the entire facility.
When data centers process crores of gigabytes of data, their servers heat up dramatically. To keep them from melting down, cooling stations pump in vast quantities of water. And here’s the brutal math: roughly 80% of that water simply evaporates and is lost forever, while barely 20% can be recycled. Cooling alone eats up 30 to 40% of a data center’s total electricity consumption — and AI has multiplied this demand many times over, because AI-focused data centers consume far more power than traditional ones ever did.
So the next time you stream a video, save a file to the cloud, or ask an AI chatbot a question, remember: behind that simple action sit thousands of humming computers and the cooling machines working overtime to keep them from overheating — burning through enormous amounts of electricity and water in the process.
How Did We Get Here? A Short History of the Data Boom
Back in the 1990s, the world was just discovering the internet, and words like “email” were entering everyday vocabulary. That era came to be known as the dot-com boom. As internet usage exploded globally, a new question emerged: where would all of this data actually live? That question gave birth to the modern data center — giant buildings housing the servers of every internet company, holding your data and mine, running 24 hours a day without rest.
The United States understood this game early. It adopted a clever policy called the “free flow of data” — essentially, let the world’s data flow freely and let it land on American servers. That single policy is why the US today controls more than 74% of the world’s high-end computing power, and why American tech companies dominate the globe.
Why only a handful of cities?
Data centers can’t simply be built anywhere. They need land where earthquakes, floods, or cyclones are unlikely; dense electrical grids and fiber-optic internet cable networks; and above all, abundant water to keep servers cool. That’s why some companies now build along coastlines — or even attempt to sink data centers under the sea, using the ocean’s natural cold water as a coolant.
But when these same companies turn toward India’s landlocked cities, the equation changes entirely. India today already hosts more than 268 large data centers, of which around 150 are fully commercial and operational — concentrated mainly in six cities: Mumbai, Chennai, Noida, Bengaluru, Hyderabad, and Pune. Mumbai and Chennai sit on the coast, where international undersea internet cables make landfall at what are called “landing stations,” giving these cities the fastest international connectivity. Bengaluru and Hyderabad already have IT infrastructure and talent. Noida, close to the national capital, has become a natural hub for government and corporate data.
The Cambridge Analytica Wake-Up Call
Before India built its own data infrastructure, most of the country’s sensitive banking, government, and corporate data sat on foreign cloud servers in the US and Europe. A handful of deeply classified government files stayed locked away in National Informatics Centre server rooms or government office cabinets. But as Digital India brought hundreds of millions of citizens online almost overnight, the volume of data exploded — and the government realized a hard truth: data stored abroad could one day be used against the very country it belonged to.
The real alarm bell rang on March 17, 2018, when a joint investigation by The Observer and The New York Times exposed how a firm called Cambridge Analytica had harvested the private data of roughly 87 million Facebook users. This wasn’t ordinary data theft — it was used to build psychological profiles of people’s preferences, fears, and political leanings, which were then weaponized to micro-target voters and influence the outcome of the 2016 US presidential election. Chillingly, the same firm had run similar operations in India.
“If our citizens’ data sits on foreign servers, foreign powers can one day reshape our own government.” — the realization that triggered India’s data localization push
That single revelation gave birth to India’s data localization policy: a clear principle that Indian citizens’ data should not leave Indian soil. In 2018, the Reserve Bank of India mandated that all payment-system data stay within the country. After years of back and forth, the Digital Personal Data Protection (DPDP) Act arrived in 2023, allowing data transfers abroad only under conditions the government could revoke at will, including the power to blacklist any country.
The US pushed back hard. Investigative reporting revealed that US Secretary of State Marco Rubio sent firm instructions to American ambassadors and diplomats stationed around the world, urging them to pressure foreign governments and lobby aggressively against data sovereignty laws. The secret cable branded Europe’s GDPR privacy framework a direct threat to American AI and cloud computing dominance. Within three days, the leaked cable made its way to Reuters and splashed across front pages worldwide, with the US openly insisting that the world’s data belonged on American servers — because if it stayed local, billions of dollars in American corporate profit would be at stake. But India held its ground — leveraging the strength of over a billion internet users that no global tech company could afford to walk away from. Eventually, Google, Microsoft, and Amazon relented, agreeing to build data centers on Indian soil, sweetened further by uninterrupted power supply and a 20-year tax holiday.
Three Superpowers, One Data War
Today, three major blocs are fighting for control of the world’s data. The United States, with 74% of global high-end computing power, wants the world’s data parked on its own servers. China has built its Great Firewall and demands that any foreign company operating there store data locally with a Chinese partner — government access included — a strategy that birthed homegrown giants like Alibaba, Tencent, and Baidu, sweeping American competitors out of the Chinese market. And Europe enforces the GDPR, the world’s strictest privacy framework, which once slapped Google with a fine of 4 billion euros for violations.
These are the grand geopolitical chess moves made in air-conditioned rooms in Delhi and Washington, where billion-dollar investments sound like pure progress. But on the ground, the reality looks entirely different — and it’s tied directly to ordinary people’s land, water, and daily lives.
What’s Already Happening Around the World
Ireland’s data centers now consume so much electricity that it rivals the combined usage of every urban household in the entire country. In the Netherlands and Germany, farmers have been driving tractors onto highways in protest, accusing Microsoft and Meta’s data centers of swallowing fertile farmland and hogging green energy — driving up electricity and food prices for ordinary citizens.
The sharpest blow landed in Indiana, USA. In September 2025, Google was forced to completely shut down a billion-dollar data center project there — not because of political conspiracy, but because local residents launched a massive movement against the project’s relentless water and electricity consumption. Their message was simple and final: the water from our taps will not be used to cool your machines. And the company backed down.
Think about that for a moment. These are developed nations with no acute water or power shortages, where democratic voices are actually heard — and even there, people are fed up with data centers. So imagine the scale of land and environmental disruption when these same tech giants set their sights on Indian towns and villages.
Vishakhapatnam: When Land Was Quietly Taken From Those Who Had Least
Between Vishakhapatnam and Anakapalli district sits a small village called Tarluwada. In April 2026, this is where Google laid the foundation for a $15 billion data center — its ribbon cut personally by Andhra Pradesh Chief Minister Chandrababu Naidu. The state government allotted 480 acres of land for the project — but where did that land actually come from?
As part of the deal, the project will receive a guaranteed 1 gigawatt of uninterrupted power — meaning the supply will never be allowed to falter. To put that in perspective: just 100 megawatts is enough to power an entire small city. This single data center has been promised ten times that.
Reports indicate this land was originally distributed in the 1970s to poor Dalit families who had no livelihood beyond farming it. In the name of this data center, that land was abruptly taken away from these same families. Human rights groups have stated that the entire acquisition process was carried out with little transparency or public consultation.
And the land was only the beginning. The project reportedly received a tax holiday of 20 years, land at throwaway prices, heavy subsidies on power and water, and infrastructure benefits worth roughly ₹22,200 crore — about $2.4 billion. On one side, farmers lost their land. On the other, the public treasury was opened wide for a foreign tech giant.
The usual justification is jobs. But how many, really? A large data center typically employs barely 150 people; a smaller one runs with just 20 to 25 staff. And yet, the tax breaks keep flowing regardless of how few jobs are actually created.
The water question that simply isn’t asked
Environmental activists who dug through this project’s clearance documents found something glaring missing: not a single line about how much water the data center would actually consume. Everything around water usage was left deliberately vague.
In India, when a steel, chemical, or textile factory is set up, it must account for every litre of water to the Pollution Control Board. But data centers are officially classified under “IT and IT-enabled services” — since they don’t dump chemical waste into rivers or burn coal in a furnace, they’re slotted into the “green” or “white” category by the Central Pollution Control Board’s industrial classification system. In the government’s eyes: no smoke, no pollution. And that loophole lets these companies sidestep strict water regulations entirely.
Indore and Pithampur: When the Sea Excuse Runs Dry
A data center needs water around the clock to tame the brutal heat generated by its servers. Around 80% of that water evaporates into thin air and is lost completely — meaning groundwater that could have hydrated millions of people simply vanishes. Tech companies often claim publicly that they use seawater to cool their servers. But Indore and Pithampur sit in the heart of Madhya Pradesh — nowhere near a coastline. The only water available to these foreign data centers is the same sweet groundwater that local people depend on for drinking and farming.
This is exactly why Indore witnessed a protest last month unlike anything seen in the past 30 years. According to NITI Aayog, 163 million Indians still lack access to clean drinking water — yet this precious resource is being served to tech giants without question, while agreements conveniently drop any mention of water consumption, exactly as happened in Vishakhapatnam.
That jump — from 150 billion litres in 2025 to a projected 358 billion litres by 2030 — represents enough water to quench the yearly thirst of 3.5 crore (35 million) Indians. Put another way: it’s roughly the full annual water consumption of entire megacities like Delhi, Mumbai, and Bengaluru combined. Or imagine the parched villages of Bundelkhand and Marathwada, where women still walk miles for a single pot of water, having their thirst quenched for years. That water isn’t reaching those villages. It’s flowing into data centers instead.
Cities like Noida and Greater Noida see their large data centers draw thousands of litres of groundwater every single month. And here’s a fact that should alarm every Indian: according to the Council on Energy, Environment and Water (CEEW), 54% of India’s data centers are built in cities that are already starved for water.
Fifteen hyperscale data centers alone have consumed more than 216,000 kilolitres of water. To put that in perspective: imagine the 10,000-litre water tanker that rolls through your neighborhood every summer. These 15 data centers together drank the equivalent of more than 21,000 such tankers — and roughly 1,000 of those tankers’ worth came straight out of the ground as pure groundwater.
“Day Zero” Is Closer Than We Think
This relentless extraction is pushing several Indian metros toward what experts call “Day Zero” — the day when the last drop of water leaves the taps and municipal supply simply stops. The World Economic Forum has flagged water scarcity as one of India’s single biggest environmental risks for the 2025–2027 window. Bengaluru, Chennai, Hyderabad, and Delhi NCR are now standing right at that edge.
Researchers have recommended that cities like Delhi NCR and Chennai should not approve new data centers until operators can prove concrete water conservation measures. So far, the Indian government has imposed no such condition.
How This Actually Touches Your Daily Life
It’s easy to treat all of this as a distant, abstract policy story — until you realize how directly it could disrupt your own day.
Your UPI payments and banking could stop working
Picture peak summer. A severe water shortage forces the cooling system at your city’s massive data center to fail. Servers overheat and crash. Reports warn that a scenario like this could simultaneously knock out UPI payments, net banking, hospital systems, and even metro rail operations — leaving you standing in a shop, unable to pay even one rupee from your phone.
Your electricity bill could quietly rise
A single hyperscale data center consumes electricity equivalent to 100,000 homes. To attract these foreign companies, the government offers them cheap power and 20-year tax exemptions. That cost doesn’t simply disappear — it tends to circle back, loaded onto ordinary household electricity bills.
In Oregon, USA, residents revolted against soaring electricity bills caused by exactly this dynamic. The backlash forced a 2025 law requiring data centers to cover the full cost of their own power consumption. India has no such protective law for ordinary consumers yet. And on the privacy front, until the Data Protection Board becomes fully operational, the honest answer to “who really has access to your data” remains uncertain — even if that data technically sits on Indian servers.
So What Does India Actually Gain From This $180 Billion Digital Empire?
There’s no denying that data centers will shape India’s economic direction for decades to come. For a nation of 1.4 billion people, digital sovereignty matters — sensitive data shouldn’t sit at the mercy of a foreign government’s discretion or a foreign diplomat’s secret cable.
But what should never happen alongside this digital growth is just as clear: poor Dalit farmers in Vishakhapatnam should not quietly lose their land. The drinking water of thirsty cities like Indore, Delhi, or Bengaluru should not be drained to cool foreign machines while ordinary citizens go without.
“Our government is looking at these data centers purely through the lens of investment and profit — not as a political and economic concern that touches ordinary people’s lives.” — Indu Malhi C, lawyer, Internet Freedom Foundation
Until that mindset shifts among policymakers, this wheel of “development” will keep enriching a handful of giant corporations. These data centers will undoubtedly become symbols of India’s digital power — but for the families who lost their land, their water, and their share of electricity, they risk becoming just another chapter in a new kind of exploitation, dressed up as progress.
Frequently Asked Questions
How much water does a single ChatGPT prompt use?
OpenAI’s Sam Altman estimated about one-fifteenth of a teaspoon per prompt. Critics argue this is understated, especially for advanced AI models, and that 10–15 prompts can collectively cost close to a litre of water once data center cooling losses are factored in.
Why do AI data centers need so much water?
Cooling stations use water to keep overheating servers running. About 80% of that water evaporates and is permanently lost, while only around 20% is recycled.
How much electricity does a hyperscale data center use?
A single hyperscale facility can consume as much electricity as 100,000 homes. Globally, data centers used about 448 terawatt-hours in a year — more than Saudi Arabia’s national consumption.
How much water will Indian data centers consume by 2030?
Projections show a rise from roughly 150 billion litres in 2025 to about 358 billion litres by 2030 — enough to supply 35 million people for a full year.
What is “Day Zero” and which Indian cities are at risk?
Day Zero is the point at which a city’s municipal water supply runs completely dry. Bengaluru, Chennai, Hyderabad, and Delhi NCR are considered closest to this threshold.
So the next time you open an AI chatbot on your phone to ask a quick question, pause for half a second. Somewhere, a server hums a little louder, a cooling tower works a little harder — and a drop of water that belonged to someone else quietly disappears.
Pankaj Dubey is an entrepreneur, business analyst, digital marketer, financial researcher, and brand strategist. He specializes in developing marketing strategies, building and positioning brands, and conducting in-depth business and financial research. He is also known for creating detailed case studies on reputed brands, analyzing market trends, and sharing insights through his writing and blogging. His work combines business intelligence, strategic thinking, and digital innovation to help businesses grow and strengthen their market presence.
