About this transcript: This is a full AI-generated transcript of Why AI Data Centers Are Suddenly Everywhere In India from Aevy TV, published July 17, 2026. The transcript contains 3,482 words with timestamps and was generated using Whisper AI.
"Achina is AI this is a comment I've had to see more often than I'd like on this channel but fortunately it is not true I am real but also just want to share this I'm also a big believer in the fact that AI is coming and just like the industrial revolution or the internet it is definitely going to..."
[00:00:00] Speaker 1: Achina is AI this is a comment I've had to see more often than I'd like on this channel but fortunately it is not true I am real but also just want to share this I'm also a big believer in the fact that AI is coming and just like the industrial revolution or the internet it is definitely going to change a lot of things because AI is a very powerful tool one that's going to reflect the best or the worst of us and right now I think it's reflecting the worst bits of
[00:00:34] Speaker 2: us big tech companies are spending billions of dollars to build data centers around the country sort of the largest infrastructure bill in human
[00:00:42] Speaker 3: history and it's a need that many communities across the nation are rejecting there's very little regulatory scrutiny of data center energy use if
[00:00:52] Speaker 4: policymakers fail to change anything by one estimate electricity rates for average households will spike by as much as 70% data centers are popping up literally
[00:01:01] Speaker 1: everywhere and we have no regulatory framework to safeguard against the consequences they bring about from draining our power grid skyrocketing our electricity bills displacing our farmers and even producing a mysterious and endless sound literally driving people insane yep these data centers have many side effects the worst of which is that they drain so much water that they are actively worsening India's water crisis and if you ask the internet it would tell you that your chanjipri prompts your AI queries those ghibli images you generate the AI slop you doom scroll over they all consume a lot of water and you know the liters of it evaporated in the data centers processing your prompts but you know what that's not completely true the real story is way way more complicated see the sudden boom in data centers is a byproduct of an AI race between the US and China a race that is prioritizing speed over safety and sustainability and well it has consequences but the big tech is okay with the consequences because they have found an easy shore to ship them to India now that does not sit well with me in the next 20 minutes I'm going to break it all down for you the true cost of a data center how much your water footprint actually is and how our country is it's becoming a scapegoat nation for this boom okay as always we have to start with the fundamentals what are data centers and why should you care see every time you swipe a debit card or backup a photo you cherish stream a video on our channel or just send an email to your boss all of it happens inside a building a big boring windowless warehouse full of servers these are the data centers that store and run our digital lives but around 2022 generative AI arrived and AI doesn't just store information it computes a lot at a scale that the internet has never demanded before so enter hyperscale data centers basically a massive campus of servers using up hundreds of megawatts and enabling the AI revolution that google meta microsoft amazon and many many others are trying to steer so mckenzie projects that by 2030 the world will spend 6.7 trillion dollars building data centers that is over 1.5 times the gdp of india global data center capacity is projected to jump from 82 to 219 gigawatts which basically means infrastructure will triple in just four years so naturally such speedy development comes with a few side effects first up is the energy crisis and over the course of the video you will see how it all progressively gets worse so a single large hyperscale campus can draw as much power as 80 000 households at any given time and such heavy consumption affects all the nearby areas which is why the us department of energy has warned that the blackouts in us could increase 104 by 2030 if data center center trends continue but that is just the beginning in georgia one of america's biggest data center hubs new hyperscales caused the state's energy demand to skyrocket which made georgia power the state's utility to file for grid expansions to meet that demand and the cost of this grid expansion wasn't paid for by the data centers it was paid for by the residents data centers dominated headlines across central georgia we feel
[00:04:36] Speaker 4: there's alternate uses of that space that can make it better for the community we want to see georgia develop we want to see georgia grow but we want to see georgia do it in the correct way through this
[00:04:46] Speaker 5: whole experience my eyes have kind of opened up to what i call like the falsehood of democracy
[00:04:52] Speaker 1: us states have only now started to create laws that prevent this scenario now let's look at what happens next the air starts to change see data centers keep rows of diesel generators for backup power and these emit nitrogen oxides and particulates that pollute the air reports also suggest that data centers emit so much heat the temperatures of downward neighborhoods can increase by up to four degrees and then we also have noise pollution data centers emit a peculiar sound that cannot be heard but
[00:05:24] Speaker 3: can be felt enormous reports coming from people about the noise they said it is so loud this is a 24 7 non-stop just in your ear humming hissing and buzzing and people just feel like they're going crazy
[00:05:43] Speaker 1: there the who in fact flags sustained nighttime noise above 40 decibels as a cardiovascular risk the sound from data centers sits at 55 to 65 decibels and finally the big one water see to prevent overheating of the servers in a data center a water-based open loop cooling system is used and the system works by evaporating water now the question here is how much exactly a single google search costs less than a milliliter of water a single air prompt and its response uses anywhere between five to 50 milliliters of water a hundred word complex gpt-4 response uses around half a liter if you video call somebody for an hour it would cost anywhere between one to five liters of water and an hour of your daily doom scrolling costs five to ten liters now these numbers are often contested but it is undisputed fact that a single hundred megawatt facility can easily consume about one to five million liters of water every single day for context an average restaurant is estimated to use around 20 to 25 000 liters per day now the un reports say that by 2030 the water footprint of ai data centers could hit 9.3 trillion liters a year equivalent to the annual domestic water needs of 1.3 billion people and the reason why all of this should concern you is because of this map that showcases the water stress regions of the country see we house 18 percent of the world's population but only four percent of its fresh water reserves so we recently made a video on how the 2026 el nino is draining our water reserves another on how the mismanaged ethanol push is doing the same thing and now finally we have the data center boom adding to this system these are the costs of hyperscale data centers with open loop cooling systems now luckily there is some respite newer data center designs incorporate a closed loop cooling system you fill the system once and the same liquid keeps recirculating which then cuts the water wastage by quite a dramatic degree however this is a recent development and still in an early rollout phase they're also a lot more expensive to build so majority of current data centers still operate under open loop systems and because they're cheaper faster and the easiest way for big tech companies of the west to try and overtake china if you jump over to china you'll see a very different landscape they have built way more data centers than their ai market even needs chinese outlets report that about 80 percent of the newly built data centers remain idle and unused the air is made them build capacity even before demand materialized now they're building a nationwide network that connects thousands of these data centers so instead of each data center finding its own customers compute capacities pooled and rented to whoever needs it and to avoid wasting water they're also starting to submerge their data centers into the ocean cutting cooling energy by up to 90 percent the first commercial underwater data center consisting of around 2000 servers was deployed off shanghai in 2025 which is why the big tech of the west is desperate to speed things up and since binding federal environmental standards for data centers have yet to properly materialize in the us they're able to
[00:09:04] Speaker 6: fast track building these hyper skills the ai industry today is the only industry in america that has
[00:09:11] Speaker 1: less regulations than sandwich shops now the air industry's defense is simple everybody uses ai so everybody needs data centers that sounds like we have no choice but to cause the water crisis but let's just put that into perspective while it's impossible to be precise with it we can approximate that an average family of four uses around 500 to 600 liters of water every day now this family of four processed 10 prompts each every day even then the max total comes up to around two liters which is less than 0.5 percent of their daily total use and on top of that according to the central water commission india gets about four quadrillion liters of water annually of which much evaporates or is unusable so even then it's estimated that at least one quadrillion is economically available for use and 80 to 90 percent of that goes into our agricultural efforts so our individual footprints while capable of adding up to a significant amount has never really been the issue and yet there's a lot of reports going around making us feel like it is why it's because of an age-old tactic you know british petroleum funded a campaign around the idea of your personal carbon footprint they launched a calculator in 2004 so that people could audit themselves instead of them the plastic industry sold us recycling for decades despite knowing most plastic products that would never be recycled the tactic was simple deflect downwards they want you to think you're responsible for the crisis that way data centers feel more justified and it works usually right now however not so much a tracker called data center watch counted more than 140 local groups that blocked or delayed around 64 billion dollars worth of u.s projects in about two years kevin o'leary from shark tank wanted to build the world's largest hyperscale in utah sized at 40 000 acres that's a single data center that is nearly the size of kolkata and it would have used 9 gigawatt hours of electricity every hour which is higher than the power use of entire countries now luckily the people of utah fought back hard and this forced o'leary to agree to scale back the project to 20 000 acres you know even erin brokovich the legendary reporter who took on pgne for poisoning a california town's groundwater and forced them into a 333 million dollar settlement the largest ever in american history joined the fight against data centers in fact she launched a whole website containing a crowdsource map tracking every ai data center in america allowing communities to report what's happening within weeks over 5000 reports emerged and they realized the number one issue was secrecy data center projects were only publicly announced after permits were already secured data center watch has reported that a significant amount of u.s data center projects in 2025 were cancelled or delayed due to public backlash because they realized that the real issue isn't the water use of ai data centers the real issue is data centers being built in water stressed areas and the ones building it choosing to not adhere to environmental safety standards so no more problematic data centers in the u.s so now big tech needs to build a data centers elsewhere somewhere with access to cheap land power water and a government that says yes quickly somewhere with no regulations or framework to safeguard against ai yeah i think you know where this is going now but before i get to that i want to show you something else tell me if you find this this is not just relatable it's 2 47 pm and you are struggling to finish your report you've already had two coffees and now a third one is calling your name but you say no because the jitters and dresslessness have already kicked in but you still can't finish the report because of a fuzzy slow can't think straight feeling that is rain fog and it's brutal on productivity gen d in the u.s have found a fix for this very feeling and it's called neuro gum built by two science grads who loved caffeine but wanted to make it smarter it's 40 milligram of natural caffeine paired with 60 milligrams of l-theanine basically cancels out the jitters and crash you get when you take caffeine only so it's just clean steady focused you know one study even found people perform cognitive tasks 21 faster after taking it after going viral in america it's now finally made its way into the indian market for av the most extensive and exhaustive part of the process is the research lots and lots of it and this little hack neuro gum is how i maintain the bandwidth to stop through so much information and misinformation and give you guys the stories you deserve so if you want to know more about this product check out the link in the description india today hosts around 250 to 300 data centers capacity has nearly tripled in five years from about 520 megawatts to 1.5 gigawatts and is projected to hit as much as 6.5 gigawatts by 2030 the adani group has pledged roughly 8 lakh crore rupees about a hundred billion dollars to india's ai infrastructure push adani google and airtel are building india's largest air campus in vishaka patnam 15 billion dollars one whole gigawatt and 480 acres of allotted land and the union budget this february the finance minister proposed a tax holiday for foreign companies running global cloud services out of indian data centers until 2047. a separate draft national policy offers up to 20 years of tax exemption plus gst credits this is honestly great for the big tech companies but what about for the city of vishaka patnam of the 480 acres allotted for the campus about 200 acres in tarluwada village belong to dalit families land the state granted them in the 1970s now on top of that videos about the displacement are being blocked on social media the human rights forum an activist organization put out a public plea condemning the hyperscale data center boom in vishaka patnam a city that's just not built to absorb it see the city's natural drainage network remains ruined the proposed sites sit at the boundaries of wildlife sanctuaries that are bound to be heavily disrupted and even the mudassar louval reservoir a vital drinking water source for the city will face disruptions now they call this move an impending ecological disaster that will worsen the city in the name of technological progress a hyperscale project of this magnitude has been greenlit in a city with acute water stress simply because there are no frameworks or regulations against this and according to a report by wri india over 50 percent of indian data centers are being made in water stressed areas and as as it stands only one of lnt vioma's 40 megawatt facility navi mumbai has been confirmed to have a closed loop cooling system the rest of the projects unspecified which seems to hint that those data centers might have the cheaper open loop cooling systems that u.s residents are no longer allowing in their lands so why is the boom fast-tracked here because india is yet to build a comprehensive regulatory framework for ai data centers see the efficiency of data centers are measured by a few metrics mainly the power usage effectiveness or pue a ratio of electricity used for operations versus the amount wasted in overheads a low pue of anywhere between 1.2 and 1.5 is fine manageable anything about 1.7 however is wasteful and detrimental to the environment which is why countries like germany singapore and china all have mandates to prevent it there's also water usage efficiency and carbon usage efficiency all of which our country's researchers are still looking into meanwhile the government continues to abide by the draft national data center policy 2025 which does encourage efficiency but also has no damn mandates for it see environmental clearances happen project by project with no cumulative assessment of what dozens of facilities do to a single water basin and yet we are letting foreign big tech bill whatever they want tax-free which means the water crisis is likely to worsen and that honestly is something we cannot afford neeti ayok reports that nearly 600 million indians face high to extreme water stress indian data centers consumed around 150 billion liters of water last year that number is expected to blow up to over 350 billion by 2030 and if any of it is happening in a water stressed area then the localized crisis will deepen severely the problem is not data centers they're useful and a catalyst for our economic development if built correctly the problem is that we are green lighting data centers that are not optimized for efficiency not sustainable for the environment and concentrating them in water bases that are already running dry much like the ethanol push progress is fast-track before we can build an infrastructure that can handle that progress but it is still not too late every fix that india needs is already being enacted somewhere else we just need to import those principles as always first up we have to go beyond the draft policy 2025 and start creating mandates that data centers have no choice but to adhere to and one of these mandates for newer data centers should be closed loop systems or treated wastewater second needs to be mandatory public disclosure of every liter and kilowatt hour you cannot govern something unless you measure it first and we don't just measure individual data centers we assess the cumulative impact of clusters of them and most of all we need to fund and fast track the research and regulatory frameworks not the business deals that make for good headlines now that's up to our government but even we can play smaller rules what we need is our version of erin brockowitz crowdsource website a platform where community members can report interact and monitor the ai data center push and just remember ordinary americans blocked or delayed at least 75 data center projects worth 130 billion through public record requests and city council meetings that playbook is free and we can use it too achina is ai that comment honestly always gets to me look at av and at eos the company behind av we use ai but we do it as a means to amplify the human touch not to replace it we try our best to ensure our use of ai is sustainable ethical and done in a way where it signifies a step ahead not behind which is why we had to dig into the cost of this revolution and the damage that corporate greed is costing us see right now these data centers and the potential benefits of artificial intelligence are coming at the cost of displaced farmers and the depletion of our water and energy reserves we are chasing intelligence and forgetting that that's just one of the many facets that makes us human creativity community purpose these are all starting to take a backseat as we optimize for just intelligence and if we keep chasing this progress through cheaper faster and dirtier methods that pushes the cost of this air evolution onto the people who cannot fight it then what we are really working towards is just an artificial humanity and that is something i am not aligned with still i remain hopeful and speak out against the way things are headed if everyone else has the same then maybe that becomes a step towards course correction