S1 EP2: Power Hungry - The Hidden Toll of Datacenters

Datacenters are here to stay as they are critical to winning the battle of AI supremacy! It is also a reality that the datacenters place a burden on the society - in terms of impact on air, energy, water, land and other short-term phenomena!  Beyond the impact, it is important to be cognizant of the additional stress these datacenters will place on the health systems in the community.  Drawing a parallel from climate change, one can say that both mitigation and adaptation will be critical to ensuring a sustainable future for humanity.

Transcript

SPEAKER_03 0:37

All right, um, welcome to episode two of What's the Point? And uh today we're going to be discussing the hidden toll of data centers. You know, uh, so first thing we want to say real quick is just a hello to our audience and give everybody you know a little bit of a shout out that's been following along, giving us some feedback. We're trying a different format and style today as we're experimenting through just a few things up front as you know, we're learning as we do this. Um, that being said, you know, we want to introduce you to Dr. Shri, which is our other co-host for the podcast that you didn't have the opportunity to meet last time. So, Shree, before we dive into the topic, uh let's give you a minute and do an intro and talk about you know your background a little bit before we dive in.

SPEAKER_01 1:22

Sure, thank you, Sunny. Hello everybody, this is um Sri Kant, Shri Fashad. Uh my background is uh I'm an environmental economist by training, did my PhD in environmental economics, worked in India for a few years on international developmental projects, focused on climate change, sustainability, before transitioning over to the energy sector. And over the last 20, 25 years, I have worked in different sectors of the energy sector, both fossil energy and renewable energy. And I have been a consultant to both the private sector and the federal sector. And here I am.

SPEAKER_03 2:02

Fantastic. So uh to our audience, you know, we just now wanted to introduce you to again the entire Watts the Point co-host group here, which is the three of us, including Varesh, Dr. Shri, and myself. And you know, we're gonna more or less just dive a little deeper from where we were last time, right? So we started talking about data centers as a as a whole. Um, today we really want to dig a little deeper, just based on some of the conversations we've had, some of the feedback we've received, and and factor that in, dig deeper and go into what are some of the hidden tolls of data centers, the stuff that you know may just get glossed over or isn't making the news. And you know, just as a caveat before we get into this, we're not really picking a are we for it, are we against it? We're trying to be as neutral as we can, share our collective knowledge, and um, you know, just also use this as an opportunity to put some of that out there. So, you know, as we as we dive deeper, one of the big things that's coming up a lot is uh what is the impact to ratepayers?

SPEAKER_01 3:04

Uh great question, Sonny. Uh look, uh I would like to answer this question by explicitly stating that data centers do spike up, the residential electricity prices. But that's a caveat. I think we need to be more nuanced and we need to be more discerning before we start talking about the quantum of the impact. Because what I'm seeing is I'm seeing a lot of analysis and information uh uh on social media, on LinkedIn, et cetera, where people I think overblow or over exaggerate the impact of data centers on residential electricity prices. And the reason why I'm saying this is uh sometime back I saw a news item which attributed Warren Buffet uh to saying that uh the data centers in Virginia have resulted in a 267% increase in residential electricity prices over the last five years. And uh more recently, there was another quote attributed to him where he kind of said he was misquoted. So uh clearly the 267% uh increase refers to the wholesale price. And uh, as utility customers, as we know or should know, uh we do not pay the wholesale price. The wholesale price is the supply part of the bill and it's accounts only for 30 to 50 percent of the bill. Right? So, and there are other reasons too for you know the prices going up.

SPEAKER_03 4:35

Yeah.

SPEAKER_01 4:35

Uh the infrastructure, the grid infrastructure is aging. I think that's a known fact. Despite the efforts in the last 10-15 years, we still have a largely uh aging infrastructure, and then you have the associated uh grid congestion issues. So these two uh these two play a role as well. So it would be incorrect to make any assumptions uh to the effect that data centers alone are responsible for um the spike in residential electricity prices. Having said that, I have also seen a contrarian point of view in the recent days. Um report that was uh you know from the Data Center Association where they stated the opposite. Essentially, they rebutted the claim that they spike up the residential electricity prices. Uh I do not have a point of view on their claim because I have not delved deep into it. Uh so, but based on the evidence that I have seen thus far and what I have read so far so far, data center definitely plays a role. The emphasis is on the word plays a role in resident in this uh residential price increases.

SPEAKER_02 5:50

Yeah, no, good points.

SPEAKER_01 5:53

And uh it's important that uh, you know, by default, the focus seems to be on electricity. We also need to look at the impact on the other two utilities, gas and water.

SPEAKER_00 6:06

And I think Vresh, you'll be able to sure absolutely so equally important is the water as a commodity, as a utility item that is needed by this new entrant, per se. Uh, and it's not a small-scale entrant either, it's an industrial scale usage of water. Um, data centers use water for sure for cooling and for um all the heat dissipation that happens over there in the actual crunching of the work. Uh the AI clusters, obviously, all of that takes a lot of power to cool. Uh, there's some other uses in the data centers also for water, but uh water needs to be delivered to data centers 24-7 at any price available. Obviously, that will have an effect on the finite amount of water resources that the utility companies manage. Uh they produce water, they treat water, they deliver water, they need water, and all the maintenance. So, this additional entrant, which is a big consumer of the water as a commodity, is going to be sharing with other people that live in that neighborhood. Uh, this has a direct impact, especially if it's an urban area, semi-urban area, not necessarily off-the-grid area as much, but wherever there's a community water, city water, county water, or private water companies. So obviously that'll have a trickle effect on um the actual bill that surrounding people pay, unless there's some special arrangement, which is very unlikely because they want clean treated water for cooling. Uh, so that's uh just a because resources are not infinite, that has a competing effect, and that has an effect on the bill itself. Same thing goes with the natural gas as a fuel, especially areas where utility connection or utility feed is not available. Many hyperscalers or co-location companies are going for on-site generation, which needs fuel. And the most available fuel is natural gas, molecules of natural gas. It will come over the same pipelines or maybe newly newly laid pipelines, but it creates another big node of demand, just like an industrial customer coming to your backyard. Again, that's 24-7, not just seasonal, uh, especially because it's prime power, continuous prime power for data centers. So that's that quantum of jump and the demand of gas, natural gas, which runs either into supercutting engines or it could be gas turbines, it's incremental load and uh incremental demand on the same supply line. So that also has uh an effect on the rates that the utility people pay for natural gas. So both, I think I would just concur with Sikant there, that these three major essential utilities for gas companies uh for data center companies would have some kind of diluting effect or some kind of spiking effect on the surrounding other consumers.

SPEAKER_03 8:58

Yeah, you know, and I think from the average consumer perspective, right? Yeah, appreciate you guys unpacking that. You know, it's I look at my utility bills, and all I see is it's going up month over month, year over year, right? And it's the I think the question that you know the average citizen would have in that sense is like, is this ever gonna stop? What is it, what does it look like? Because, you know, all they see is the impact to them personally, their family, their household. They don't necessarily see the the benefit or the other side of it from, you know, I can't see it, I can't touch it, I can't feel it. I just know it's costing me more to do the same thing I was last time. Right.

SPEAKER_01 9:33

Yeah, interesting that you you know mentioned that because just this morning I was reading the article which said that data centers, you know, increasingly they are trying to go behind the meter.

SPEAKER_02 9:44

Right.

SPEAKER_01 9:44

And this analysis said that uh if if the data centers go behind the meter, it's going to actually spike up their bills even more.

SPEAKER_02 9:52

Yeah.

SPEAKER_01 9:53

So yeah, so the impact on the prices, the utility prices, electricity, water, gas, I think it's definitely going to be there.

SPEAKER_03 10:02

So I think that's tricky, right? Because the idea of you want to connect or you want more of the resource, we will self-fund or find our own internal way to do it. But going behind the meter and still has a transfer cost back to the average consumer, you know, it leads to the natural like, well, then why are we doing this question in that's unanswered for a lot of people? On you know, what what's the reasoning? What's the rationale? What's the point? And yeah, unintended.

SPEAKER_00 10:32

And one more aspect I think you can consider is uh because data centers are typically define markets where there is fiber communication, yeah, there's plenty of electricity, gas, and and manpower.

SPEAKER_02 10:48

Yeah.

SPEAKER_00 10:48

Uh but these just happen to be semi-urban areas where there is utility serving customers. So you're you're having this other entrant or sector of the entrance that um drives uh competition for the same same commodities.

SPEAKER_03 11:03

So well, so let's let's kind of shift gears, right? I mean, it's uh what are your thoughts on you know what's the impact from an air emission? You know, is this polluting the air? We're hearing a lot of this is going to affect uh heat and how heat's transferred, or naturally just heating up this surrounding area of the data centers. Um what are your guys' thoughts on on those aspects? Or again, the drain on natural resources being clean water that's usable?

SPEAKER_01 11:37

So uh in terms of the impact on air pollution, look, it's not a myth, right? These concerns are not out of place, then or uh they it's it's not that they are misplaced. These are genuine concerns. Um and there are serious apprehensions on whether you know we have you have a lot of clean energy, renewable energy today, but whether that is going to be sufficient of that to meet the energy requirements of a data center is a big question mark. Uh, and essentially what that means is uh we are already seeing uh increased reliance on natural gas to meet these requirements, but we are also seeing uh the retirement of coal plants being delayed, right? And this is happening in the democratic states, uh, in the middle, Pennsylvania, Colorado, where uh the retirement of the coal plants have been delayed, but primarily to meet the needs of the data centers, right? So which means that essentially that the concerns are legitimate about um the air quality emissions of data centers. Now, if we look at the air quality emissions per se, um there has been some analysis done in this regard. Um and I think they have looked at uh data centers in the United States. And if you look at the top 10 states, you know, which has the highest concentration of data centers, it is obviously Virginia, Texas, California, Illinois, and Georgia. These are you know, these are some of the states that uh feature in top 10. And the global external damages through a data center nationally uh uh is about estimated to be about 10 to 30 billion dollars. The median is 24 billion dollars. So, what do we mean by global external damages, right? So, external damages, I think it's nothing but negative externalities. I think everybody understands that. But here the global refers to the fact it's not that it's gonna affect the rest of the countries, but the global here refers to the fact that the pollution, the emissions is gonna move in the air. You know, it's not going to be restricted to where the data centers are, right? It's gonna affect the neighboring states and beyond, right? So the global external damages is estimated to estimated to be $24 billion. And of that, uh Virginia and Texas is estimated to account for 30% of that.

SPEAKER_03 14:19

Right.

SPEAKER_01 14:20

Right? Just two states, 30%, which is seven billion dollars.

SPEAKER_03 14:24

Now, and and just to stay with that thought for a minute, you know, that's today's numbers based on what's active today.

SPEAKER_01 14:30

Absolutely, that's a great question. Yes, that is based on the current capacity of LSE, not taking into account the plant capacity. Right. Uh and keep in mind, so also uh if you look at the global external damage damages per megawatt hour, it is more than the commercial electricity prices in these states.

SPEAKER_02 14:57

Interesting.

SPEAKER_01 14:58

So, what does that mean? That the data centers are not paying for the negative externalities that they cost, right? Uh and this is a topic which is a bit close to my heart because some years back I did an analysis on gas peaker plans, and you know, they also cause a lot of emissions. And my theory was uh it's not just enough to look at the levelized cost of energy generated, you have to also include the social cost of the social damage that they cause into the into the levelized cost. And if I are to extend that argument, I would say the data centers should be made to pay for the damages that they cause.

unknown 15:41

Right?

SPEAKER_01 15:42

So if the commercial electricity price is six cents per kilowatt hour and the damage that they cause is 10 cents per kilowatt hour, they should be made to pay 16 cents per kilowatt hour, not six cents. This is easier said than done. You know, there are a lot of legal and regulatory and policy hurdles to cross before we anywhere come anywhere close to doing this. Uh but yeah, but you know that that is an important issue to look at when we uh talk about the air emission impact of data centers. Yeah. On the special, the the related aspect is the data heat islands, right? Uh there is also evidence to suggest, and this is again based on not just the data centers in the US. It's based on uh data centers across the world. And about, I think there are about 11,000 or 12,000 data centers across the world. And that is evidence to suggest that that is a two to three degrees Celsius increase in the land surface temperature neural data centers in the immediate vicinity. And this effect extends to up to uh you know six to seven miles beyond where the data centers are located, right? So it is, and if we want to you know go beyond just these numbers, I think the bigger question or the bigger issue is uh, and you know this is my perception, and I could be wrong. I think a lot of the discussion on data centers today is happening uh around the US and a few Western countries. Uh, I don't think a lot of thought uh has gone into the discussion on the uh feasibility and the impact of data centers in the rest of the world, right? The question and and what I'm gonna say has relevance even for the rural America. Because when you talk about global external damages, like I made uh the point I made earlier, what does it actually mean? It means you're gonna have more respiratory diseases, you're gonna have more uh hospital admissions, so there are gonna be more morbidity outcomes, right? And if you're gonna have data centers in the rural areas in this country, or if you're gonna have data centers in the rest of the world, uh, you know, which are not so well developed, are the health systems prepared in those areas prepared to handle the additional stress? I think that is an issue to think about. And you know, and that is the kind of uh you know, the hidden cost uh, you know, we are talking about here during the data centers. It's just not about the numbers, you know. We need to look beyond the numbers and see what the implications of these are.

SPEAKER_03 18:30

Yeah, yeah, those are all good points.

SPEAKER_00 18:33

So I would I would add the previous um answer on that uh utility rates for gas and water. Specifically, water, let's just consider water as an important commodity, how it will have a societal impact, the community impact, the indirect impact. It could be detrimental or hardship. Um water as a as a commodity is essential for data centers for cooling, indirect cooling as well as direct cooling. And that there's always losses in evaporation and leaks and so on. But that becomes like a constant node of very high demand. Obviously, that is non-negotiable even during times of droughts or other hardships of water scarcity. These things must operate. So that actually takes an essential uh quantum of essential commodity away from the rest of the community. Obviously, that might have more water restrictions on the surrounding communities. Um, if it's also other agricultural areas or areas that require water to operate, industries and so on, some somebody will have to give it, somebody will have to give up or trade off, right? So those things are considered indirect effects on the surrounding community, not necessarily from health perspective, but availability perspective. Somebody has to compromise as such. And uh typically the data centers are generally well funded, so there's also they pay premium for it, uh, but it's a commodity that they won't live without. Um, the other thing I would say from a community perspective, impact perspective, is when you do, especially when you do on-site generation, is and also a lot of cooling, there's going to be a lot of acoustic noise that is almost like noise pollution that you don't want all the time around you where people live, or people operate, the schools operate. Uh, so that adds like a almost like a background noise that is not of the desirable type. Uh, it's also not a white noise that you can just ignore. Um, so it adds this additional disturbance to people's living, uh, their surroundings, their environment. Um, I think that's something that some communities have started to kind of voice saying that's additional noise for us. We don't want that. Um, so noise introduction as well as water tensions that it can create during even times of hardship for the rest of the community. Yeah, I guess those are considered societal costs.

SPEAKER_01 20:56

Absolutely.

SPEAKER_03 20:57

Yeah, and a little bit of a tangent, you know. Um, but I I you guys know this personally, but for our audience, I do real estate on the side, and one of the things that's got me thinking about is you know, there are clauses or documents written into real estate contracts when you're near an airport, right? Like that's a reportable thing you have to talk about. I wonder maybe someday through regulation and stuff we talked about, if there's gonna be a I'm so close to a data center, and now there's a noise pollution aspect that we have to, you know, declare in a real estate contract, right? And um this actually kind of dovetails nicely because one of the next things we're gonna talk about is land availability and prices and things to that effect. But you know, kind of a hook in where I was going with that was the the the real estate costs, uh not necessarily in a commercial sense, but also in a residential sense. So what you know, let's talk about some of the hidden costs or the hidden toll around land and pricing and how that's starting to show a trickle effect.

SPEAKER_01 21:54

Sure. And I think I think uh you know you you raised a great uh point about the impact on the real estate value. Right. Yeah. So say land is a finite resource.

SPEAKER_03 22:04

Yeah, we're not making more.

SPEAKER_01 22:06

And you know, uh, I I I hear a lot of people say uh land is not an issue in the US. We have plenty of land, for sure, today.

SPEAKER_02 22:14

Yeah.

SPEAKER_01 22:15

Right. But it's a finite resource. At some point, we're going to start running out. And there are going to be competing uses for land, right? So we cannot take the land for granted. So also because of you know the perceptions about the impact of the data centers, uh, there has been a pushback, you know, from the residential community on the takeover of the land by the data centers, right? Yeah, and what that has done, and that has been actually one of the factors leading to the data centers moving to the rural America. Right. So, and even there, they are getting a pushback from the farming community. Because I mean, I would say uh not all farming community, you know, some of them have you know given sold their land to the data center companies, but some farming community are pushing back. I think the concern there is uh whether there is going to be enough to uh you know feed the nation. And this brings me back to the days of 2006-2007, during the days of the ethanol craze, right? And there was this classic food versus fuel debate, right? And we may see uh we may end up seeing a similar debate if the use of land for uh data centers is not regulated.

SPEAKER_02 23:34

Yeah.

SPEAKER_01 23:35

And um yeah, and also, you know, again, similar to the point that I made earlier, is great, we have a lot of land in the US, but you know, what about countries where land is at premium, right? Yeah, uh, and countries which are primarily uh agrarian. What about the feasibility of data centers there? You know, and if data centers are going to start taking a plan on the scale that we are seeing in this country, that is definitely going to be an impact on food prices. But before that, that's definitely going to be a lot of pushback from both the farming and the residential community.

SPEAKER_03 24:12

Well, it and real quick on that, you know, as you talk about the agricultural aspects, um, back to something we were just talking about a minute ago. You know, if there's uh air emission pollution, if there's noise pollution, if there's it's just a warmer climate, is that also going to have an effect on the crops or you know, the animals from a farming standpoint, right? Like again, I'm way out of my wheelhouse at this point, but it's as we talk about these things and go deeper, it just leaves me wondering, and I have more questions, you know, than than answers at this point.

SPEAKER_00 24:44

But so I'll I'll I'll I'll put land as a another essential site selection criteria in the whole discussion here.

SPEAKER_02 24:53

Yeah.

SPEAKER_00 24:53

Um, you want to put data centers where you can connect to power, water, and fiber, for example. Now there are lots of parcels of land around the world where you can still give it out because it's non-usable, non-arable, too arid. It's also called surplus land, uh, things that nobody wants. But but that's it doesn't have the trifecta of other commodities that are needed for data center setup. If you do this in the middle of a desert, yeah, you got plenty of free land or free available land that could be cultivated, uh, that could be converted. However, uh, it'll be a hard place to do any kind of fiber communication or even for the workforce part of it. Uh, you can do that for like remote large solar photovoltaic farms. Yeah, that's perfectly fine. Uh, that can be, it doesn't need to be attended so much. Um, also, it's it's where land is available and the natural conditions are available for power production. But for data centers, you've got to have a feed for communication. There are there are um Starlink-based or satellite-based, communication-based data center models coming up, but that's not the mainstream. The mainstream is near semi-urban areas, near urban clusters. And um that itself means that demand is also going to compete for parcels of land. Uh, the the other flip side, the good flip side would be um that the today's data center designs are a lot more compact, so they don't take massive acres and acres of land. Uh, so they do take less land per per megawatt of compute. However, it is still too close to some communities where communities are trying to do a sprawl and expand and build out, they will be competing against the same parcels, right? So, all in all, the land um aspect itself is an integral part of the the play here, yeah, how it competes with the other needs of the society. Sure.

SPEAKER_01 26:55

So actually, uh you spoke about the impact on the real estate prices. So, from what I'm seeing, is uh on the rural side, uh, because of the you know attempt by the data center companies to take over the agricultural lands, that is pushing up the prices of the farming lands. And what that is doing is that is restricting the ability of the small and marginal farmers to buy the lands, right? So uh, and uh I just wanted to turn this over to you because you know you're on the real estate side. Yeah. What what do you see an impact on the real estate prices in the residential communities in the data centers?

SPEAKER_03 27:35

You know, I'm I'm not well versed specifically in anything in the data center area because I'm I'm sticking more to the Maryland market, but I would imagine you know there's gonna be there's gonna be an effect at some point, right? And as I was saying, with you know, whether it's noise pollution or other things, I would imagine the immediate surrounding towns beyond the data centers themselves, I would expect there's probably gonna be some type of cost increase. Um you you brought up workforce, places like data centers certainly are gonna drive you know jobs as a positive, which has a boost to the economy because then you know people don't want long commutes if they can help it, right? So the immediate towns become the places these folks all want to live. So, you know, we we drop this big thing that brings everyone there from a business standpoint. The surrounding areas tend to have a boost in home value. There's you know, um market conditions that suggest this is favorable to do, it puts them close to work, it allows surrounding towns to have a benefit one way, but you know, the the other side of it still turns into, again, it's the the affordability, right? It's are we gonna price them out and this becomes more challenging? Or is this in some shape or form gonna be the new Silicon Valley situation where everything just exponentially grows and you start to price out everyone around it? Um who knows, right? And it's it's hard to it's hard to quantify some of that just based on where I think we are right now.

SPEAKER_01 29:08

Right. Uh you know, you you you you made a reference to uh the benefits of data centers, right?

unknown 29:14

Yeah.

SPEAKER_01 29:14

And you made a reference to citing. And I think this um you know struck a kind of um tube light in my head. I think when it comes to uh uh boils down to the authorities, the regulatory authorities, uh when it comes to deciding whether to give approval for uh data centers, yeah. I think they need to look at both aspects, not just, I mean, they just can't say, hey, you know, you're just gonna create pollution and you know, we we don't want to have you, right? So you they also need to see uh the economic impact these data centers create. So they need to have a metric like uh you know gross damages to GDP ratio. And in cases where in areas where the damage to GDP ratio is greater than one, do not locate the data centers there, right? And in areas where damages to GDP ratio, yeah, or that content when I say GDP, that contribution to GDP ratio is less than one, surely you know it makes sense to locate the data center there. And but today what is happening is if you look at California, right? So 82% of the data centers are located where the air quality is already low. Similar to the case in Virginia. Right. So it's interesting. In fact, I was mentioning this to Varesh yesterday that you know, given the huge water requirements uh of data centers, if they are located in cool, you know, cooler areas, the water requirement will significantly reduce. But most of the data centers are being located in the dry arid areas, right? So that is so the whole thing is kind of the the decision to locate the data centers by the operators, by the builders, and by the regulatory authorities who approve them, in a way, kind of to me seems in a bit self-defeating.

SPEAKER_00 31:18

Yeah, so the geographic location of your data center itself is just one aspect. Right. The other aspect communication and the connectivity to uh to power, other systems like that. So, I know, for example, up upstate North Dakota, upstate on New York's New York uh state could be good location from a cooling perspective for land availability perspective. Sure. May not be the greatest choices for workforce, connection, for electric power, uh as well as uh communication. Um so these things are not a perfect, it's not a perfect solution here, but uh there's always case by case.

SPEAKER_03 31:54

Yeah, yeah. But uh no, those are those are all good thoughts, right? And it's it's back to what you were mentioning. It's that three-prong criteria, that triangle of of all of it. You know, I mean it'd be real easy to say, well, let's just go put them all in Colorado or Canada or something. It's cool, it's you know, high mountaintop, the conditions are good, but it's it's not quite so simple. Um you know, but we've been talking about some of this, and uh I'm I'm also curious around like how how's this going to affect clean energy? You know, is this is there a trade-off there? Um, what's the impact on you know, when you when you plan in decades or half centuries and centuries, and we have these long-term goals we want to try to get to, based on all the demand and everything we're seeing, right? It it it all seems to be a pretty aggressive push to get data centers and everything going. But I wonder where that tipping point is on the clean energy side of it.

SPEAKER_01 32:50

Yeah, I think the whole clean energy transition, uh I'm not sure. I think uh that's that's taking a hit.

SPEAKER_03 32:57

Um doesn't seem to be the focus of the day.

SPEAKER_01 32:59

It doesn't seem to be the focus of the day for a number of reasons. For two reasons. One is obviously it's not the priority of the current administration.

unknown 33:06

Sure.

SPEAKER_01 33:07

Right? Uh, but also um uh uh it has been compounded by the whole data center thing. You know, um as I said, their energy requirements requirements are huge, and there's not enough of either the conventional or the fossil energy or the renewable energy to meet those needs. So you need you need a combination of both. Yeah right. And today you have more of fossil than renewable. Yeah so what, as I said, is I mean, like I said uh earlier, that is increased reliance on gas, right? Also, when we talk about gas, there are uh two kinds of technologies, and Varish should speak to that. That is um combined cycle um gas turbine, and that is this reciprocating engines, right? And the combined cycle gas turbines are not going to be available, uh that's going to be a scarcity of that until 20, at least until 2030, if not 2035. And which means if you're looking at if a data center is looking at gas, um they need to look at they need to be looking at reciprocating engines, which are much more polluting than the combined cycle.

SPEAKER_00 34:15

And noisier.

SPEAKER_01 34:16

And noisier and noisier noisier, yeah. Yeah. So and then, you know, there are states, like I said, which are bringing back the coal plants and which are delaying the retirement of the coal plants. They may say that they are deploying clean coal technologies, but we know that clean coal technology is probably not as clean as renewable energy, right? So um, yeah, so the clean energy transition, I think, is definitely taking a hit uh you know, due to the priorities of the current administration and due to the whole data data center uh game.

SPEAKER_00 34:54

But part of that is like Shikard was saying, it's the cause is primarily the magnitude of hunger and the haste, like we covered in the first episode.

SPEAKER_02 35:04

Yeah.

SPEAKER_00 35:05

If if there was a little bit of relief, there was a little bit of a roadmap or planning horizon, you can do this slowly but deliberately and add this additional stuff, additional items, additional data center capacities, and so on. And it can still fit into the clean transition picture. You have time to plan it out and take all these things into account growth and where, what, where, and all that. But the average data center timeline for coming operation is 12 to 18 months. That's not compatible with pipeline growth, gas pipeline growth, power delivery growth, power delivery of the right type of power, too. Um, if you can add other dimensions like small modular reactor-based power generation, it's a great idea on paper. Maybe it's also pre-commercial right now, but the ability to do that quickly and today and tomorrow is just not there yet. So we I don't think the society has had the uh comfort of roadmap planning, timeline planning, because of the haste and the hunger. Hunger itself uh being a big quantum of need, and here and now. The other aspect is um the the thing that Shrikan mentioned about gas turbines, which is probably a cleaner burning system for power generation for on-site, like independent power production, is severely supply chain restricted. The the big drivers are already there in the industry for power generation. That's for the utility companies, as well as mechanical drives that is already filling their order books for 12 to 18 months to two years out. And you have this additional demand from just data centered on-site power, it's obviously a not as easy match, even with even with uh exponentially premiums they are ready to pay. It's still the factories don't turn that fast, right? So, supply chain for gas turbines is a determining factor that you would love to do that, but it's not compatible, it's not available that that fast. Um, and that's what I think he also made a reference that there's that other option is to go back to the old-fashioned clean coal or coal fired power plants, just to up you know, up the capacity of the power plants itself, the utility power plants. That obviously is general in generally speaking, it's against the principle of going away from fossil, the clean energy transition promise, right? So it's affecting the speed, the velocity, the implementation of clean energy transition. But it's something that's we just have to acknowledge, analyze, and cope with and find strategies to mitigate.

SPEAKER_03 37:39

Well, and it's it's back to you know, not only the conversation we're having today, but also the one we had on episode one with Elliot around not only the hunger and haste, but the aspects of you know regulatory, the process pieces, the scheduling, the timelines, all of it. These seem to be lagging things that are still trying to catch up. And by the time they catch up, you know, the demand's continuing to increase, right? Like it it's if you could future forecast and say, all right, if there's 3,000 active data centers hypothetically in the world, and in four years we were gonna double that number, I want you to plan for that problem now, right? I don't know how the equation would change from a planning standpoint or what we would need to do, or it's here's all the infrastructure things we need to have, here's all the clean energy things that are gonna have to be delayed, deferred, because we have to keep these other systems live active now, or we're never gonna get there. Um all of these things start to build on themselves, right? Also as a challenge point.

SPEAKER_01 38:38

It is, it is. But uh, you know, Sunny, I think um I'll go back to the point that I made at the beginning of uh you know this session today.

SPEAKER_02 38:46

Yeah.

SPEAKER_01 38:49

Yeah, I mean, we are talking about the cost of data centers, the hidden cost and all that, that's fine. But I I don't think the idea is to portray a kind of uh impending doom because you know yeah, you know, because what is happening is thanks to NIMBY, not in my backyard, and thanks to pushback from uh the farming community. Uh legitimately, uh legitimately so because of concerns over uh the capacity, right? The many of the data center projects are getting paused or getting cancelled. Uh it's happened in several states.

SPEAKER_02 39:25

Sure.

SPEAKER_01 39:26

So uh it's gonna take time for uh these data center projects that are in the planning stage to actually come online.

SPEAKER_02 39:34

Yeah.

SPEAKER_01 39:35

So we may see uh so what I'm saying is we may not see a drastic consequence of the growth of data center in the next few years. You know, this pause or this you know temporary cancellation of uh the projects uh may actually result in a staggered impact, which can only be good for us. It will give us time to actually be better prepared to absorb the growth of the data centers.

SPEAKER_00 40:06

Yeah, exactly. So we I think the society as a whole hasn't had the time or the intent to put long-term growth plan in place for this technically uh fourth revolution, fourth industrial revolution.

SPEAKER_03 40:18

I mean, that's what it feels like for sure. And there's so much focus on it. You know, um one of the things we touched on earlier, let's go back to this, right? We we briefly touched on the whole bring your own power type of initiative, right? The going behind the meter, I think is the words you used. What are your guys' thoughts on that? On the what's the impact going to be from a bring your own power initiative standpoint?

SPEAKER_00 40:40

So obviously, um the investors, the hyperscalers, as well as the co-locators, the co-located data center companies, uh, depending on their own barometer of haste and hunger, yeah, they're ready to put more capital down and say, I want it done today. I can't wait two to three years for utility interconnect or extra feed. So they are choosing the do-your-own self option, the make your own power option, right? Um that is one clear way to do this. Uh the limiting factors, again, the supply chain is the main issue. It's not about technology, it's about the supply chain. Um, gas turbines being one of the best options for on-site combined cycle generation. But you can also get beneficial heat and you can use that for cooling through absorption chilling. There are various ways to wrap it if you had the ability to tap into a flowing supply chain. Um, I think the world in general was not prepared for that because gas turbines are also in demand for like aviation markets and mobility markets. Um, so some of that is putting an extra pressure on the gas turbine supplies supply chain. I think, as far as I know, the the synchronous generators that I need, the alternators on the driven end, uh, that is not a big supply chain constraint at the moment. However, the order books are filling fast for a lot of aging equipment, right? Things that need normal change out anyway, overalls and stuff. Uh, and then there is all these other um other power switch gears that you need, the substations you need, the relays you need, the giant power transformers that you need, all of that is also getting added to the order books of these apparatus companies, right? Electrical apparatus companies, I'm saying, switch gear companies. Uh, I think the know-how is all there. Is the ability to do it is also there from an execution perspective. It's just the ability to catch up with the uh supply chain backlog. That is what sometimes gives sleepless nights to the independent power production plants. Yeah. Uh, but if you gotta go live, you gotta go live. And unfortunately, there's not an easy option to say I'll just use that for now to get going and delegate connection. For example, small modular reactor-based solution. Not there yet. If it was, of course, that this debate will be so much lighter, right?

SPEAKER_01 42:58

Yeah, sure. And similarly, long duration energy storage. You have some data centers that are looking at uh you know using renewable energy combined with storage. But uh you really need long duration energy storage for that to be effective. And it they have potential, but uh, I think they're at least a couple of years, two, three years away from being a reality, right? So um, so uh in terms of uh my my perspective on uh bring your own is uh as Varesh said, uh it is early. I mean, my assessment is it is early days yet. Uh there are still a lot of unknowns in terms of how things will play out. I think uh I would like to wait and watch uh that space a little bit more.

SPEAKER_02 43:44

Yeah.

SPEAKER_01 43:44

Uh because as I said, just this morning I read the news that behind the meter is going to spike up the builds even more. I don't know how much truth or how much uh rigor is there to that analysis, but uh yeah, very early days yet. Uh we need to watch the space.

SPEAKER_03 44:02

For sure. And I mean you touched on something earlier about it being the fourth industrial level revolution. You know, with that part of it, I feel like this is also driving a lot of innovation and a lot of different conversation that's now happening. You know, to your point of small modular reactors, other technologies that are going to surface, or how do we capture heat or transfer in different ways, or how do we not lose a lot of that so we can recycle it back in, you know, doing those heat transfer types of things. It's gonna force innovation, I think, which is gonna be interesting.

SPEAKER_01 44:34

Absolutely, absolutely, absolutely.

SPEAKER_03 44:36

It's not all doom and gloom.

SPEAKER_01 44:37

No, absolutely not, no, not at all. I don't think it's doom and gloom. We just need to uh wear a bit of caution. That's all that's all I'm gonna say. Yeah, and there are solutions. Yeah, there are solutions today, and there'll be solutions a few years down the road as well.

SPEAKER_03 44:53

Yeah. So, you know, um, we're we're getting close to bringing this one to the end here for this episode, but a couple more things we want to touch on. We talked a little about the the G Von's paradox. Let's talk about that a little bit, and you know, what's the impact of that on AI?

SPEAKER_01 45:08

Yeah, no, the G1's paradox is essentially essentially, I think we need to be cautious. And uh, there's a lot of people are saying uh you know that'll be efficiency gains, and once the computing need the per density, the density of that starts to go down, the energy requirements will not be so high. But essentially what G1's paradox says is if something is going to be more efficient, then you're going to use that more. And essentially then that's going to actually negate all the efficiency gains. Right? So we yeah, so it could be possible that it it may play out this way, that you know, five years or ten years down the road, and despite all the efficiency gains uh in terms of the technology, we find that uh the emissions are still the same. And you know, that could be energy once paradox playing out. Uh, you know, something to think about. You know, we don't know whether that's going to happen for not. It's happened in other sectors. Um we have to wait and see whether that's going to happen in the data center project. Um, yeah, yeah, that's my thought. But I think the other thought is um we we spoke a lot about the about the impact on uh the utilities, the land, uh, water, etc., etc. Uh, I'm gonna basically uh uh focus on it on the aspect that is not talked about enough. And uh so and this is an issue, clearly an issue when it comes to critical minerals in some parts of the world, um, like Africa, and where uh the labor has been that work in the sector, you know, the the mineral supply sector, they have been abused and they have been exploited, and there are a lot of labor issues around there. There are similarly a lot of labor issues around um energy storage manufacturing in China, and there are talks that there could be similar labor exploitation when it comes to the data center and AI market. And I think uh as part of the corporate social responsibility, it is incumbent on all the you know you know, data center designers, builders, operators to be more conscious of the fact and make sure that uh we do not repeat, we do not see a repeat of what we have seen in Africa with critical minerals and energy storage in China.

SPEAKER_00 47:42

So for so, from my perspective, the organizations could be hyperscalers or could be co-locating companies that build the shell and let people reuse there as tenants. For them, if the corporations themselves have made a pledge of being responsible socially, corporate social responsibility, it'll be for them to kind of demonstrate the least amount of impact they've had on the surrounding communities, surrounding areas, and the environment as such. Uh, I think it's also seen as if the capital markets, which is looking at these firms, which are building these, it's seen as a gauge of success in some way. If they're doing the CSR right, the corporate social responsibility right, uh, it'll be considered a win for them showing responsibility to the society itself, besides delivering this essential AI compute capacity. AI is like we were talking about the G1 paradox, right? Earlier, it's something that hosts storage, compute, or cloud, as well as all the AI applications, and it also leads to it's a host for uh the the um cryptocurrency market as such, yeah, the blockchain market. So all that is only going up in terms of intensity, sure, the need. So the the the diminishing returns that you were referring to is possibly very, very true because it's people are going to get more and more used to this as a methodology. So um, overall, I think the CSR, back to the CSR point, the corporate social responsibility, it will be a challenge for firms who who build these, operate these, maintain these, to show responsibility to the environment, responsibility to the community, as well as if they've built to a lead standard, they have to show demonstration of all the principles being met consistently. That's where some of those uh the metrics will come out.

SPEAKER_01 49:34

You know, uh, you know, uh you triggered up my memory. So, you know, there are also short-term impacts uh, you know, because of data center, right? Uh apparently in some rural uh town of uh north uh of this country, uh during the period the data center was being constructed, that town saw a 600% increase in traffic-related deaths. Oh yeah, uh in in 2025 as compared to 2024.

SPEAKER_02 50:09

Okay.

SPEAKER_01 50:10

So which means that you know the construction activity of data centers also has its own risks, which is going to impact the community. So then these are going to be these are not going to be long lasting, these are more going to be you know short term, but uh I think that also fits into the corporate social responsibility for safety practices. Yeah.

SPEAKER_03 50:35

No, those are those are key points too.

SPEAKER_00 50:37

Um yeah, and so from a CSR umbrella perspective, they can also do programs for the community once the data center is live for education, internships, training, cybersecurity, yeah, a variety of trade jobs, and so on. So those things are some things I think the corporations that build it can demonstrate you know their redemption of their CSR commitment through these activities. It doesn't have to be donations, it could be inclusive, could be uh a little bit more participatory for this community around them. So all of those could be other factors where they can gain cookie points towards CSR goals.

SPEAKER_03 51:11

Makes sense, right? And and the point you just brought up, like the safety side of it, active campaigns to you know minimize safety incidents, right? I mean, if you if you worked in different safety environments, right? My background being in the defense sector, uh, I walked a lot of production floors in those areas, you know, and it was they're always holding their safety standards as as the the book, the guide, the touch point they're going to. And the goal, of course, is minimize safety standards, certainly minimize you know, broken bones, deaths, things of that nature. And um, it it's as about getting as close to perfect as you can for you know zero defects in that sense or zero incidents. Um, but those are those are key things. But you know, there's one there's an aspect of like, are we just paying it lip service or are we actively doing things and we can prove we're doing the right thing from a corporate social responsibility standpoint?

SPEAKER_01 52:01

That's a great point because a lot of what a lot of companies do is they do greenwashing.

SPEAKER_03 52:06

Yeah, it's a policy.

SPEAKER_01 52:08

It's sitting on a shelf, we've not touched it in 10 years.

SPEAKER_03 52:10

Yep. Yeah.

SPEAKER_01 52:12

So uh uh so also you know another important point to keep in mind is you know the AI supremacy battle is very important for this country. Yeah, we touched it. It is, it is it is very important for any country, right?

unknown 52:25

Yeah.

SPEAKER_01 52:26

And data centers are here to stay. So I think, yeah, sure, there are these impacts of data centers, but we need to not just look in terms of mitigating these impacts. I think the equal focus has also has to be on helping the communities adapt to the impacts.

SPEAKER_02 52:47

Yeah.

SPEAKER_01 52:48

You know, similar to what uh has been happening in climate change. You just don't have mitigation, you have adaptation. I think a similar approach needs to be there for data centers, is focus has to be on mitigation and adaptation, not just mitigation alone.

SPEAKER_03 53:06

That might be a good setup for a future topic for us, right? I think collectively we've talked about uh haven't quite settled on the right title or what we're gonna cover next, but that might be a nice thing on well, what can we do? This is here to stay for sure. So let's get beyond just all right, is this around for a long term or not? But more of the what's a proactive approach or measure we need to start taking. How do we start holding ourselves to that standard? Um, more grab the bull by the horns, if you will, right? We're gonna be dealing with it.

SPEAKER_00 53:38

Yep. So I think we we've been going through these episodes like first one was awe, now it's analyze, then we'll get into activate, evaluate, activate all these other steps, steps of uh of our journey. And it's a matter of adopting to it, essentially, not resisting it. Yeah, it's not gonna go away by wishing it away, right? So we had to find a way to cope with it, live with it, and expand it. And and use look, keep the humankind benefit in mind all the time. Yeah, productivity, safety, communication, quality of life, all of the above, right?

SPEAKER_03 54:10

Sure, sure. Well, guys, we're we've been going for a while. You know, let's let's bring it to a close. Any, you know, one last thought, one thing you want to echo or share that you may have already stated, you want to recap, or one more closing thing you want to just share before we wrap this one up?

SPEAKER_01 54:27

Uh yeah, the impact of data centers are real, but uh there is no need to panic. I think uh we need to take a nuanced and uh approach towards mitigating and uh mitigating the impact of the data centers. Yeah. No need to press the panic button.

SPEAKER_00 54:46

Agree, and I see more of an opportunity in this kind of uh challenge uh for innovation as well as better ways of going getting this done and meeting this uh this um unprecedented demand surge as such, find ways to do this. Um so I would I would challenge that we go into the innovative mode so and stay that way so we can find a consistent solution for this.

SPEAKER_03 55:11

Sure. No, I I think those are great thoughts. And you know, for me, it it's it's been a great conversation so far as everything we've covered, and uh really excited to talk more about the well, how do we get the bull by the horns and how are we gonna innovate? How do we take that more deliberate, thoughtful, proactive approach to whatever's coming next? Um, so with that, I think we were gonna wrap this one up. Uh so for our audience, we'll say thanks for tuning in today. You know, please like, subscribe, share. Those those are the best compliments you can certainly give us. We are we're always open to your feedback. You know, please reach out, let us know how you're feeling about it. Um, give us some ideas for future topics. Our next episode, we think, is gonna drop around mid-July, around the 19th or so. And until then, stay charged up. We'll see you next time.