Powering AI _ Global Leaders Session _ AI Impact Summit India Part 2

Good evening, distinguished guests. Welcome to the session on powering AI. As AI scales at speed, so does its infrastructure demands. Data centers are facing unprecedented power and cooling requirements. A single large AI training run can consume as much electricity as thousands of homes use in a year. This raises critical questions like how do we plan for rapidly rising and uncertainty energy demand? Can edge computing reduce the load, or is centralization inevitable? To address these critical issues, we are joined by our exceptional panelists. Mr. Vineet Mittal, Chairman of Avada Group. Sir, I request you to please come on stage. Mr. Natham Blom, Vice President, Cooling Chambers. Professor Raghav Chandra, Professor at IIM Calcutta, Founder and CEO of Consult and former Chairman of NHAI and Secretary to Government of India.

Moderating this important conversation is Mr. Ashish Khanna, Director General of the International Solar Alliance. Mr. Ashish Khanna, Director General of the International Solar Alliance. Thank you panelists for being here with that I now request Mr. Khanna to please take the discussion forward

Good evening everyone not easy being the last panel especially when we are probably starting at the time that we are supposed to end but we hope and we will try and make it more interesting for all of you we are here to talk about Powering AI the format will be that I will begin in terms of framing some of the issues at heart and also tell you a little bit about what International Solar Alliance is going to do then I will hand over to each of the esteemed panelists to make an opening kind of a statement of what’s their vision on this question of Powering AI for about 5 minutes each and then I will ask them one question each and then I will ask the question and then I will ask the question on some of the specific issues for which they are probably an expert on.

And finally, if there is any time left, we will see if any audience member wants to ask a question. Let me start off by saying, why is International Solar Alliance in this session and in this AI Impact Summit? We are here primarily for two reasons. The first reason is, the world has done 1000 gigawatt of solar doubled in just last two years, what was done in last 25 years. Almost 40 % of that is decentralized, which means it’s either solar rooftop or pump or others. That figure is only 15 or 20 % in India and obviously very low in a lot of developing countries. And a distribution company often does not like decentralized solar because it impacts the distribution. It impacts the distribution system and finances.

But the right amount of digitization and AI can actually help them absorb it and reduce the cost of the system as a whole. And therefore, India’s ability to more than double decentralize renewable energy, but in general, world over, will require AI. That’s issue number one, for which actually I will say that we launched a global AI mission for energy in the AI Impact Summit. We call it AI for Energy. The session is going to talk about energy for meeting AI demands. But let’s first talk about AI for Energy. Why? Because there are some elements that the world has not seen, which is, if some of you were part of some sessions earlier, can consumers trade power based on what rooftop and batteries do you have, P2P trading, that requires certain digital enablement of the trade of millions of consumers, producers and consumers, that right now needs a lot of regulatory evolvement.

An IT architecture, so that each distribution company in India, but for that, that matter anywhere in the world will know what will it make ready to actually trade that power. all. It’s about jobs. Today, a lot of AI engineers do not understand energy. Energy engineers do not understand AI. We at International Solar Alliance, which is now 125 country member body, headquartered in India, is creating an ISI Academy to train people to bring together AI and energy skills engineers together. This intersection of energy and AI will be the fundamental shift over next five years, the way Amazon changed retail. This is what is going to happen, we believe, in renewable energy. Third, is about innovation ecosystem. We are in the AI Summit.

A lot of startups are having fundamentally disruptive ideas on both decentralized renewable energy, as well as the way you manage generation, transmission, and others. The fourth is about financing. How will all this financing and de -risking be done? Because not all places have a lot of venture capital or commercial loans and equity possible. We are in the process of creating a new industry. We are in the process of creating a new industry. We are in the process of creating a new industry. We are in the process of creating a new industry. We are in the process of creating a new industry. We are in the process of creating a new industry. And finally, there’s a global dimension where International Solar Alliance is involved.

What are going to be the interoperable standards? Because the world is not united on how all of this will be done. So that’s a lot about AI for energy. But there’s also an equally important element of energy for AI. The world’s largest sources of increase in electricity consumption right now are only two. Data centers and cooling. Some of it is going to happen through electrification of cars, EVs as well. Now, 70 % of all data center demand today is US and China. But in times to come, it’s increasing by almost more than 50%. A lot of it is going to happen in developing countries. And we’ll hear some of that in addition to global elements. A lot of that is also having a lot of innovation that will renewable energy provide that energy.

Can 24 by 7 solar and storage provide cost -competitive energy to some of these data centers, whether they’re small or hyperscale, hyperscale being above 100 megawatt? What’s happening on innovation on cooling? We will hear some of the experts on the private sector who are trying to come out with a lot of innovation on that and what happens on the ecosystem. Obviously, today’s data centers are consuming a grid equal to Spain right now, and it’s going to double every three years. So this is a very important segment. Without further ado, I’m actually going to go probably to the esteemed panel. I’m going to request Mr. Raghav Chandra. Sir, you have been part of the government and now teaching. When you look at this big…

element of powering AI, how do you see it?

Thank you, Ashish. You’ve done a fantastic job in a short time period covering the larger macro issues connected with this sector. Friends, as we gather here in a nation racing towards digital sovereignty and sustainable growth, I want to emphasize and putting on my academic professorial hat, the single greatest constraint on AI’s future, which is not algorithms, not chips, but it is energy for AI -based data centers. And, you know, I’m going to mention a few such instances. In late 2024, Mark Zuckerberg made a confession that stunned his employees. A nesting colony of bees had torpedoed. Meta’s plans to open the world’s first nuclear -powered AI data center. That single environmental snag exposed their deeper vulnerability that Meta’s AI strategy depended on a single resource that it did not control and command, which is electricity.

Power outages and energy shortages have increasingly disrupted major tech companies’ operations, particularly as AI -driven data center demands strains global grids. There has been another very famous incident of March 29, 2025. A sudden loss of utility power in Google Cloud’s Columbus, Ohio unit triggered a critical failure in the uninterruptible power supply UPS batteries that created a major havoc for society. Several hours. This caused a cascading outage of six hours, in fact. Over 20 services were hit. Various customers experienced degraded performance or total unavailability, affecting cloud -dependent apps and websites globally. No direct apology was, of course, issued, but the event underscored energy reliability in a big way in an era of AI growth. In September 2019, utility power failed at one data center in Amazon Web Services, AWS’s North Virginia zone.

Backup generators activated but ran out of fuel after about an hour due to faulty automated refueling systems exacerbating the blackout. And it affected about 7 .5 % of the volume of… Apps and databases and some customers lost data permanently. Backups weren’t in place. Services like Slack and Netflix saw major ripples. And this has happened not only with Google Cloud or with Amazon AWS, but it has happened with companies such as Microsoft’s Azure, which suffered a major setback in 2018. It has affected TikTok, that’s ByteDance’s new USDS joint venture, causing widespread system failures. And what it underscores is the need for ensuring that there is suitable energy availability for data centers and that there is suitable backup for data centers.

Otherwise, you will not be able to have high powered. So energy guzzling, AI based data centers, which are the basic, basic. unit for AI to be implemented across the board for simplifying and for making and achieving our goals of ensuring that we have AI which is responsible, ethical, efficient, and which can do our job effectively. There is one county in the U .S. which my friends here on the dais would be aware of, of Ludon County, Virginia, just outside Washington, D .C., where data centers now outnumber people in density. And this 40 -square -kilometer area of computer server farms is Christend, the data center capital of the world. It hosts about 200 operational facilities, and another 100 or so are coming up.

Their peak draw is nearly 3 gigawatts. That’s enough to power a small country. Over 70 percent of the global Internet traffic passes. This is the clear area. What brought Ludon and its implications to the world’s notice was the massive outage at Amazon, causing tripping of crucial banking services and various social media companies. In Ireland, their data centers consume already one -fifth of the nation’s electricity, more than all the urban homes combined. Data centers traditionally began as largely in -house centers for proprietary computing data storage. They have since evolved, and today they are largely remote facilities or networks of facilities owned by cloud service providers, housing virtualized infrastructure for the shared use of multiple companies and customers. They need tons of electricity.

With all the power -hungry hardware and cooling systems, a data center today uses, higher -density racks, and whereas earlier… the data center typically used something like 150 to 300 watts of electricity per square foot. Today these higher density racks can consume as much as 100 kilowatts per cabinet which equates to 10 ,000 watts per square foot. And therefore a data center power problem can have global ramifications for the company. AI is supercharging data center boom that will recharge global energy systems. Global data center electricity consumption today is 415 terawatt hours. That’s about 1 .5 % of the world’s total consumption of electricity. And by 2030 it’s predicted to be nearly 945 terawatts or 3 % of the total consumption. So AI is not a side story. It’s the main driver with accelerated servers growing 30 % annually.

in the United States, which is the current epicenter. Data centers use 176 terawatts in 2023, or 4 .4 % of the national electricity. Projections are staggering. That’s like adding, in fact, the entire power demand of countries like Australia or Spain. So, you know, when we look at powering AI, we have to look not just at the upstream issues of creating the requisite demand, of creating the requisite power supply, but the other factors which come into play are the downstream effects and the hidden costs of progress. Environmental, if we rely only on fossil fuels to bridge the gap, emissions soar, so you have the debate between thermal and between renewable. which my colleague here will talk about. Big tech’s scope to emissions are already up 30 to 50 % since 2020.

Globally, data centers could claim 40 % of new fossil generation if clean supply lags. And so, while on the one hand, AI can help to accelerate decarbonization through optimal strategies and with intelligent working, but at the same time, the very fact that they are power guzzlers, they have an environmental issues which is inherent, and therefore there is a need for choosing a virtuous path. They have economic and social costs, while power prices are spiking. For instance, in the US, in and around areas which have data centers, the power cost has gone up significantly. In fact, wholesale electricity has jumped 200 to 250%. . in five years in certain areas, and the households are feeling that pinch. There is an issue of reliability.

Grids weren’t built for this. Voltage swings in Virginia have already tripped dozens of centers. In a warming world with rising AC loads, blackouts aren’t theoretical. They’re a governance failure waiting to happen. You have the equity issue. Who bears the burden? Communities near data centers face noise, heat, and land -use conflicts. In developing nations such as India, the digital divide widens if energy access for AI crowds out basic needs. So there’s a need for ingenuity when we’re dealing with this issue, and efficiency has to be the best weapon for dealing with the larger social, environmental, and other issues connected with this. And, of course, you know, in India, a lot is happening about which we’ll talk about.

But there is, indeed, a moment of great… happiness that AI is powering us, but there is also need to be concerned about whether we will be able to power AI effectively and whether we will be able to effectively and efficiently manage the downstream effects of powering that AI effectively. Thank you.

Thank you so much, Raghavji, for the different elements of sustainability risks for the society. Nathan, your opening statement especially from the cooling perspective.

that keeps these northern Virginia, as an example, data centers from adapting to more efficient and effective technologies. But when you’re starting with new builds, with white space technologies, you have the opportunity to actually build for the future instead of build for the past. And so that, to me, is the most important element as to how we’re going to solve powering AI in the future.

Thank you, Nathan. I’m sure you educated a lot of us in terms of what’s really happening on the cooling side on the innovation. Vineet, over to you, that from one of the leading renewable energy developers, how do you see?

Good evening, everyone. So I see AI as one of the biggest opportunity. For the renewable sector, historically, people believe that renewable is intermittent, which it is. It is difficult to predict when the sun shines and wind blows. So we needed the technology which can help us intermittent power dispatchable at 15 minutes interval so that the grid can operate in a stable environment. So what AI has helped that with the help of a lot of climatic data, which your weather department collects, company like renewable companies are collecting, defense department collects. And then you can get real time data from low earth orbit satellites. If you use all of them in the right way, you are able to predict using AI that what would be my generation like.

And then you go a step further and you can schedule and dispatch that power like a conventional thermal power would do. So that makes AI for energy and energy for AI. And that empowers the grid to have always on. Clean power. which is the uniqueness India offers. So let me tell you, friends, when India started adding solar and wind some 15 -16 years ago, we didn’t even have 5 megawatt of operational asset. And this year alone, India is going to add 50 ,000 megawatt of solar and wind capacity, making us the second largest green energy player besides China. And what it gives power to India is that, like the previous panelists were saying, in the U .S., in Malaysia, even in Ireland, which used to be the data center capital, every country started charging some surcharge on powering the data center.

But the reality of life is that there are not going to be 50 megawatt or 100 megawatt data center. Now we are talking about 500. 500 megawatt, gigawatt data center because the… compute requires so much of eating as Nathan has explained. Without impacting the society and affordably if you have to do, India is the place. And the reason I say that we are blessed with abundance of sun, wind and water. So using the pumped storage because of our geography, we are actually getting a natural ability to do storage. And largely in most of the states, sun and wind are complementary in nature. So what happens is using sun and wind alone, you can generate 14 to 18 hours of power and then you complement it with pumped storage and battery.

And if you combine with the AI and you build your AI stack properly, you are looking for round the clock green power. So India is the perfect location India is adding 50 gigawatt It’s not competing with the normal consumer. India has a lot of very good policy where using green power, they are able to move even farming activity from the night shift to the day shift. So and our per capita power consumption is one of the lowest in the world. We are less than 1500 kilowatt hour per year per capita. So if India has to become a Vixit Bharat, you can’t become Vixit without data. And data is the new oil. And unfortunately, what is happening today is that we are we have 1 .4 billion people and out of which a billion people are connected.

And we are one of the cheapest data connectivity package in the world. So we are. The largest user of YouTube in the world, almost 700 million. user on the YouTube is from India and we are the largest content creating economy whether you take Insta, whether you take YouTube, you take any social media. It’s a repeat story even on WhatsApp we are more than half billion user. And all of this data as previous speaker was saying resides in some other countries. So why should we generate so much of data and the data should reside in any other country because probably earlier we didn’t focus to use all this abundance of energy and power that data center and now today the scenario is it makes economic sense in US.

Now you cannot get any power before 2030. All even the gas machines are sold out. So if you look at the grid waiting time in the US is typically 7 to 8 years. Permitting you can get during that time but if the world has to adopt TI at a massive scale India offers that opportunity where we can set up multiple gigawatt data center. We can provide them green power using solar, wind and storage. And we actually have a very unique situation. Unlike US or Europe, India has a single grid. You can insert the power in Rajasthan and can pick up in Mumbai in real time basis. And India has invested heavily into the grid. And we continue to grow that national grid where the whole country is connected.

So the best location for solar, best location for wind, best location for pumped storage and battery can bring power to the data centers in Mumbai, Chennai, which are already connected. So the latency does not become the bottom line and it becomes the ideal choice. What is needed is probably more data sovereignty type of act. Indian user content has to be located in India by certain time frame and so that developers can plan for the grid they can plan for the large data center capacity and can bring that to light so it’s one of the greatest opportunity for India Indian ecosystem is purely geared up for that and on top of that we have if you look at even in the AI probably more than 25 % of talent resides in India and that talent currently is working for other countries so they will be based in India, work for India and provide services and intelligence to the rest of the world and that’s the way moving forward

Great so let’s have a little bit of a discussion and I do hope we get time for one or two questions so there was a little bit of questions that, Raghavji, you talked about, but a lot of optimism on both sides. I will ask a question combining two elements, which are important. One of it relates to the whole policy and regulatory landscape. Is India, or for that matter, developing world’s policy and regulatory landscape conducive for promoting data centers? I think, Vineet, you talked about the importance of data, the policy and regulatory landscape related to data sovereignty. Even Africa, I remember, was thinking of having a legislation like Europe, where the data for that particular continent or that particular country should be within that region.

But then there is also a policy and regulatory landscape for discovering price of power for data centers. India believes it’s very competitive. U .S. is struggling with the cost of providing power. Power probably is a limiting factor rather than the Nvidia chips. So that’s where the U .S. is. The second element is on innovation. I think you spoke about it, Nathan, but we’d like to hear what would an innovation landscape for cooling look like? Is it a lot of startups? Is it a lot of some of the larger companies doing some process efficiency? What would this innovation landscape look like? I want to request each of you to think about and say what would a policy and regulatory landscape change and an innovation landscape change accelerate both the speed and the cost of what meeting the demands of data centers look like?

Raghav.

So in the Indian context, you know, the stakes for us, as Vineet mentioned, with all the opportunity and the resources that are available to us in terms of land, in terms of water, in terms of the skilled manpower, the opportunities are enormous. And the data center capacity is all set to explode. Today, it consumes about one gigawatt of power today. We are expecting it to reach about eight or nine gigawatts by 2030. And it’s continuously growing. We have ambitious states like Andhra Pradesh, which can effectively be called the data cities or data states for the country. We have a coal -dominated grid, which India has fortunately allowed to continue in a very pragmatic way. We have rising cooling needs from extreme heat.

And as Nathan mentioned, that some of our states can have a power usage efficiency or effectiveness, which can be extraordinary. Because of all the heat, whereas ideally it should be. one which is the perfect index and we also have a net zero ambition of ensuring that we have complete renewable focus non -fossil fuel based energy dependency to reach by the year 2070 which is our global commitment which is I think again a very bold and generous commitment of India but the biggest issue that I find in this entire landscape if you ask me is about the ease of doing business in India and I am not being skeptical but having been an administrator who has been a managing director of the state industrial development corporation the state investment corporation the managing director of the road corporation the urban development principal secretary chairman of the national highway authority and various other such positions Now when I sit back and I’m on six company boards, I realize that the biggest bottleneck in India today is the lack of synergy between the states and the center, between the departments of the government and essentially between the states and the center.

And if India has to move forward to achieve this huge target that it has set for itself for ensuring that we become the data center country for the world, that we exploit our entire human resources, that we exploit our land resources, the solar energy that we have, we must have, you know, apart from the regulatory schemes, et cetera, and for the regulatory. On the regulatory side, much is being done. For instance, in the latest budget, we are all aware that how the finance. Minister announced the scheme for ensuring tax exemption for data centers that are set up in India with foreign collaboration for the foreign component part of the investment and for their revenues. Lots happening on the renewable energy front.

Lots happening on the various data centers that are being set up. However, lots needs to be done in terms of getting synchronized coordination, ensuring that the best technologies are brought in. One of the points which Nathan made was about leapfrogging and ensuring that India should capture technologies which other nations have faultily or by mistake adopted. We can certainly skip that and go on to the best technology. Water in the days to come is going to be a very, very big and critical issue. Thank you for India. And therefore, using liquid coolants and solutions such as that for cooling are going to be extremely important. And this has to be realized not only by the central government, by the states, and by everybody who is working in the field that they must facilitate ensuring that these things are adopted in a positive manner.

I had an example of a foreign company which the other day was talking to me. And they had signed an MOU with a particular state government for a huge amount of data centers to be established there. And they said that, you know, we are struggling. We’ve made eight presentations and we haven’t been able to move forward on that. That’s the kind of thing which with the best intentions and with our prime minister being so proactive that we should really have proactive chief ministers, everyone getting down to business, and using the large number of experts who are available all around to explain to them the best technology and moving beyond perhaps even L1 to be able to get the best configurations on the ground to ensure that we are not only efficient but we are effective.

Great. So a lot of potential but work to be done on ease of doing business, center state coordination and also from innovation a big potential for Indian companies to innovate on cooling, liquid cooling especially given water constraints. Nathan, over to you.

Yeah, I’ll comment on that innovation because that is innovation is the foundation that the IT industry is built upon and it’s built upon the idea that any one individual or small group of individuals can create an idea that changes the entire multi billion dollar industry itself and those who don’t innovate end up falling off the map. You know, you don’t talk anymore about AOL or Ask Jeeves or companies like that, and maybe we’ll say the same thing about Meta or Microsoft or, you know, Google or Amazon someday. Who knows? Because that’s the nature of the industry. And so as we look into the future, I think the innovation is going to require these smaller companies who are able to take risks and think bigger, especially around these cooling technologies.

And that’s what’s already happening today is we’re seeing people who are thinking outside the box of what we’ve normally considered to be advanced cooling technologies. Today, when we talk about advanced cooling that’s being deployed, what we’re really talking about is moving from that air -cooled ecosystem to just a simple liquid cooling ecosystem, which was developed in the 1960s for the Apollo space mission in the United States by IBM. And it’s been used for all of those years, including in the 1960s. If you’re a gamer at home, it’s been used in those large desktop gaming systems. And so this is an old and proven technology. You basically use… ethylene or propylene glycol mixed with water and you pump it through a pipe and it touches a cold plate on top of the hot chip and it captures it in liquid and moves away.

And that is a very simple and easy way of capturing heat, but it has limits. And what we’re facing is the limit that that liquid, as it leaves the chip, is getting so hot that you then have to have some coolant, some way to cool it back down. And that uses an incredible amount of electricity to cool that water back down, to use chillers on the roof of your data center to chill that water back down. And so the delta between the heated water, glycol, and the chilled water has to continue to get bigger and bigger and bigger, which means you have to cool that water lower and lower and lower using more electricity, so you eliminate the efficiency.

There’s now technologies emerging, and this is what my company is focused on, that is very similar to the way we cool. You can cool air in an air conditioner or in your car or in your refrigerator, and it’s called a two -phase technology, and basically what that means, instead of pumping liquid around… and it’s staying liquid, it actually, the liquid boils and vaporizes and that change of phase that is from a liquid to a gas is 10 to 20 times more effective and efficient at capturing heat. And that technology, though, is being spearheaded by small companies and those small companies will get bought up by large companies and they’ll be adopted into the ecosystem. And so expect to see that.

Expect to see the same basic use of refrigeration or refrigerants that we have today and we’ve been using for a long time, but using them specifically within the IT load of a data center ecosystem. That allows us to get those PUEs, that utilization efficiency ratio, not 1 .5 but 1 .05. You see that, I mean, that’s a massive step function increase in efficiency, which means the power generation doesn’t, doesn’t have to be strained nearly as much. And so I think that’s where the innovation is really going to come in the next three to five years.

great I think on the lighter note I’m always it’s baffled but amusing that the gaming industry was the start of GPU and now the cooling as well it’s fascinating how gaming industry is responsible for the AI revolution but lot of space for small companies if you have on the innovation side Vineet what do you think?

Gaming and best actually because the large batteries requires the same amount of cooling so the way I see innovation happening across the board is when the knowledge and cross industry expertise starts fertilizing and for that to happen you have to start creating local ecosystem see we can’t be sitting on the fence and be solving and innovating consistently that you are doing in theory but when you are building large data center of gigawatt scale you can find solution and use those skill because the similar challenge comes when you design the clean room. So how do you combine the expertise of building a clean room of millions of square foot with the expertise which is required for cooling the batteries and the expertise which is required for power usage efficiency in the data center.

How do you combine those skills and build the solution which is good for India where the humidity in some of the cities where optical fibers are terminating through the sea is large and how do you balance it out. So you have to use the external environmental data also to customize your PUE efficiency. So we see that efficiency is possible at all levels whether the ceiling height should be 6 meters or 8 meters. How close to India. So India is in that sense is fortunate that we are building those expertise locally without being building those expertise locally. building those 100 gigawatt off data center. In Morgan Stanley did a study. There is a $4 million opportunity cost for the power.

So they are saying the battle for the AI is no more compute. And it’s no more intelligence. It’s the power. Power is the biggest challenge. And there is a lot of innovation which is happening on power sector in India. You gave a good example of P2P trading using AI. And the policy in India is quite open on open access. So when I give power to the grid and I’m taking it out, I’m getting the power in the real -time basis, which is very few countries are able to do globally. And we have to account for on the monthly basis. So that gives a flexibility to the data center, which you always want. clean power and they want 24 by 7 365 days reliable power that is what is available in India and I agree with Raghavji ease of doing business is not similar across the states in the country but that’s why government of India is doing stack ranking of the state so today you can’t be just dependent on one state like look at Maharashtra the kind of support they are providing today if you want to build data center is amazing like permitting land everything is fairly streamlined and on top of that they incentivize so I think government has got it not every state is on the same page that if you have to become a developed nation your data is the biggest enabler if you have to win any kind of manufacturing battle data is the biggest enabler like if you look at today even our financial data most of the software companies whether it is Oracle or SAP or Microsoft they want the data to be on the cloud and those data even your financial data now even SAP you can’t do ECC everything goes on HANA on rise which is on the cloud so you buy the space either from AWS or Microsoft because they have only partnership with those two so even the 40 ,000 odd companies which are on these ERP softwares in India where are their data going and so the opportunity wise I think India because of its own need will innovate consistently ease of doing business is a challenge and that’s where there is an opportunity to continuously work with the government on transparently on your challenges and suggesting a solution which is not benefiting one voice is a And then third is understanding the nuances of how the application layer is working across the industry and educating government also that why should they have the data localization initiative.

And I see all of this getting combined and India becoming probably the third largest country where the AI adoption and data center would be one of the enabling block for the future growth. Thank you.

So a lot of optimism. I did promise one question. I have one question space only. So please go ahead. If you can identify yourself and have a brief question.

My name is Umesh Prasad Singh and I’m an associate member of Indian Institute of Public Administration. Sir, my question is directly to you. In your paper you have mentioned about global ramification. that particular aspect of global ramifications are of both types that is positive and negative. With respect to that, will you just have the clarification on that note? I wanted to know on that.

When I said global ramifications, I’m talking of both essentially the downstream effects of focusing on data centers and the implications it has on the fact that it will have an implication for the environment because they are power guzzling, as Nathan mentioned, that today earlier we had data centers which were full of CPUs, today they are full of GPUs and you’re going into all kinds of even more complex computing units so because of the storage, the networking, they are becoming far more complex. So it’s going to have an impact on the environment because of the heat. That is generated intrinsically because of the data center, because of the environment that will be affected, because when you’re consuming coal to produce that power, you’re using water, that same water which could feed millions of people and pay the, you know, today we are not able to feed enough people for, provide adequate drinking water 24 by 7 to all our cities, yet you would have water effectively being used for the cooling of the data centers.

You will have social issues, because people today already for thermal power plants, they are creating issues where they find that their land, especially in the scheduled areas, et cetera, is being consumed for coal mining, so there are issues connected with that. Likewise, there are all kinds of social and environmental issues that are likely to happen. There are issues on the side of… You know, whether we, you know, what other implications it can have… So all these are things which are not essentially just localized, though they are local problems, but they will affect global companies which can have the benefit of India is that we can leapfrog in terms of technology. And hopefully, as one of the speakers earlier in the previous session mentioned, chips are also becoming more and more efficient.

So, you know, as they become, computing becomes more efficient, the chips become more efficient. So you will require a lesser amount of energy. If we can leapfrog, adopt the best technologies in terms of design and infrastructure, that again will be a great saving. So today, no nation is an island. Everyone is connected. And anything which impacts one nation affects the entire thing because data centers, if they are located here, as I mentioned, the case of the Ludon County outage, it affected billions of people all across. So it has a global ramification. while you have to think of your own benefit you have to keep an eye also on the impact whatever you are doing has on all across the nations and which is why when the Prime Minister talks of Manav, it is the human being who is at the center of it and the human being is not just you it is the larger mankind and the larger human community

Thank you, unfortunately we do not have time for any more questions but it’s pretty late I’m ending without summarizing but it’s pretty apparent huge optimism on the power of India and developing countries to meet the demand for AI both through solar storage, innovation on liquid cooling and of course the ecosystem with ease of doing business please join me in giving a big round of applause to all of them and thank you for staying very late thank you everyone for joining Thank you. Thank you.