Aired Aug 25, 2026 · 36 min

Listen on: YouTube · Spotify · Apple Podcasts

AI data centers are driving unprecedented electricity demand—but who should pay for it all?

Energy Central’s Matt Chester sits down with Ari Peskoe, Director of the Electricity Law Initiative at Harvard Law School, to unpack how the data center boom is reshaping electricity costs, utility regulation, and the power grid. Peskoe explains why adding a hyperscale data center can be like “plopping down a large city basically out of nowhere” from the power sector’s perspective.

The conversation also examines utilities’ incentives to attract data centers, the push for greater transparency, and whether tech companies should simply build their own power generation.

Plus, Ari answers questions submitted directly by the Energy Central community about behind-the-meter power, climate impacts, data center noise, and the future of electricity regulation as AI-driven demand continues to grow.

Matt Chester: Why don't we require data centers to provide their own power? Is there a real chance that data centers lower bills for customers? What will be the economic impact on increasing climate disasters on data use and data center development? These are just some of the questions that came in from you for our guest today. Ari Peskoe currently serves as Director of Electricity Law Initiative at Harvard Law School and his regular writing covers energy policy issues at the state and federal level.

He has a background in engineering and a law degree from Harvard Law School, so he's uniquely qualified to sit at the intersection and not only teach the next generation of electricity lawyers, but also educate today's policymakers on energy regulation issues, directing them towards what's urgent and feasible and pressing. For those of you unfamiliar, electricity law was once only a niche wonky topic that would have insiders buzzing, but an article about power grid policy decisions probably wouldn't make the first several pages of a newspaper. But all that has changed recently as the intricacies and impacts of electricity law are now more in the public landscape than maybe ever before. From data centers to rising utility bills to the ever-present clean energy and climate regulations, these topics have moved outside of niche insider conferences and happy hours and really into the mainstream headlines and town halls. And during this moment in time when we're recording amid summer heat waves that seem to be exacerbating these issues and a new hyperscale data center is announced seemingly every day, we wanted to get the inside scoop from somebody who speaks the language and we found perhaps the best possible guest to do just that. Stay tuned for this whole episode because we go deep into everything that's happening in data center development from cost structures to policy, the things that you probably haven't even thought about yourself.

Oh, and Ari answers questions directly from members of the Energy Central community at the end of this episode. And you don't want to miss those. And if you're watching on YouTube, subscribe and like this video. And now let's get into it.

So I just want to dive kind of straight in Ari. And for those of you who don't follow the nuances of rate cases, let's kind of start at that high level. Who generally approves price increases? Is it regulators at the state level, federal level, or the private utility companies or tech companies building the data centers?

Ari Peskoe: So the prices that you and I pay for electricity service and what an ordinary business pays, those are approved by state regulators. And typically what happens is that a utility will go to the state utility commission and apply to change its prices. And that initiates a rate case. And there's basically two steps to a rate case. One is, figure out how much money the utility is going to collect in total. In general, that amount covers the utility's expenses of providing service and also provides a return on its investments in physical infrastructure like power lines and power plants where that's allowed.

And then the second piece is figuring out how to share those costs among everyone who takes service from that utility. And these rate cases can attract a lot of public attention. Big energy users like industrial customers and Walmart often participates in these cases. There are residential ratepayer advocates in most states charged with looking out for ordinary consumer interests. And a number of other parties with their own business interests like solar developers, EV charging companies may also participate in a rate case, and ultimately, the regulator makes a final decision based on all the evidence before it.

Matt Chester: Excellent. That's really helpful context. And you mentioned the major power consumers kind of being involved. And that's what's brought a lot of attention, even more so attention, to these rate cases recently. And as an example, Meta's Hyperion Data Center, it's a $50 billion project in Richland Parish in Louisiana. And it's expanding to 5 gigawatts of IT capacity, which you're quoted as saying is five to six times the total power demand of New Orleans.

So when people hear a data center is using more electricity than a major city, they understandably worry about, well, what's that going to do for my electricity costs? Another example is Los Angeles, where they don't really have excess power just sitting around and regular extreme heats pushing their power grid to the extreme rates rising 30 percent from 2019 to 2024. Yeah. All that's to say customers are understandably frustrated. So all that is prelude to, you know, from a practical implementation standpoint and maybe a messaging perspective, how do the utility leaders, the utility CEOs, how can they confidently talk about, hey, we're going to add this hyperscale data center amid that backdrop? You know, what's the messaging that they should be bringing to frustrated and anxious ratepayers?

Ari Peskoe: Yeah, I mean, look, so the basic issue here is the utility business model, which is, as I've said, you socialize the costs of providing service. And when you build a new power plant, you build a new transmission line because demand is increasing. There's typically a formula, a cost allocation formula for sharing those costs among all ratepayers. And what's different here with data centers is that there is a single customer that's one of these giant gigawatt scale facilities that can drive billions of dollars in infrastructure costs. And so I would argue the goal then is not to figure out how to share those costs, but rather the goal should be, can we isolate those costs and assign them all directly to the data center? I think we've seen over the, you know, I've been following this issue for a few years now, And I think we went from a place where there was a lot of denial, both from the utilities and from the data centers, that this was even these costs are being shared at all, to now more of an effort to isolate perhaps some of those costs.

But I think there's still a challenge to really look at the full spectrum because this sort of socialization is really just so baked in. Yeah. In August 2026, that this is a potential issue, something the public is concerned about, than there was in 2023 when this data center build-out started to ramp up in response to the introduction of ChatGPT.

Matt Chester: Yeah, absolutely. And I think that's that's the landscape, like I said, in 2026, where we're seeing a lot of headlines. And, you know, one recently came in The New York Times. They published an article about that you're quoted in that where President Trump announced, hey, we're going to have more than 200 companies and politicians, some of the largest utilities, NextEra Energy, Duke Energy. They're all signing on to this voluntary pledge to prevent AI data centers from driving up electricity bills for millions of people. It seems like they're trying to address the anger and the anxiety. But in this article, you said that the pledges are clearly symbolic and a lot of theater and really what happens is what happens at the utility level. I guess my first question is, why does it seem that these voluntary pledges are the best that we can get, at least for now?

Ari Peskoe: Yeah, they're not the best we can get. I think the fact that the pledge even exists reflects a failure of the utility industry to not get out in front of this problem two or three years ago when they saw all these requests coming in from these data center companies developers. I mean, the company that kind of was a first mover here was AEP in Ohio, and they put forward a proposal in 2024 to actually create a new class of ratepayers, a data center rate class. And what that could do is provide a mechanism in a rate case to, when we're figuring out how much money each type of consumer is going to pay, how we're going to share the cost of the system, there's now this new ratepayer class. We can say they are responsible just the data centers for a certain percentage of the cost. And that was a real step forward in terms of isolating the ratepayers.

But what, and so that's something that the industry has been building on and something they can continue to build on. And these sort of White House pledges are performative. They don't mean anything. The legal teeth to hold everyone accountable happens at the regulatory level level by the tariffs and the contracts that are reviewed by state regulators and, again, sometimes federal regulators as well. We can get into that issue too.

But just creating that class is an important step one. The next step then is figuring out what costs ought to be assigned to those data centers. And that's where I think there's still a lot of progress that needs to happen.

Matt Chester: Yeah, and I think for that progress and for getting the teeth onto these regulations, like you mentioned, inherently, it seems like the government needs to push back almost on the tech companies, but there might be some challenges with that. What influence do you think those tech companies have on the policymakers? Are they kind of exerting a lot of influence and making sure that those teeth aren't as strong at this early level when they think their growth is so critical?

Ari Peskoe: Yeah. Yeah, look, I think the buck stops with utilities and the utility regulators more so than the tech companies, because for the most part, and we can get to off-grid data centers later if you want to, but for the most part, data centers are connecting to a utility-owned system. And so it's the utility that has sort of the pen on these contracts and on these tariffs, and it's the utility that goes to the regulators and files whatever rate structures it wants to propose. And the data centers... are certainly active participants in a lot of these proceedings, but they're not the entities that are writing and filing these contracts. I think also what you're alluding to, though, is where the political power is these days as well.

Certainly, there's a lot of lobbying at the state and at the federal level from both the utility industry and from the data center industry. To a large degree, their interests are aligned in that the utilities want data center customers to, the data centers want to get power as fast as they can and get the utilities to build stuff. And so there's certainly a large overlap in what each of those two parties want, but they may diverge as well. Data centers, in many cases, want to go outside of the utility system. If they think they can get power faster by going to a private power plant developer, then that's what they'll want to do. And the utilities may be standing in the way in some cases.

We have some pretty strong monopoly protections in many states that can prevent a data center from going outside the utility system. And for the most part, we've seen that monopoly system stay intact. There are a few states that are allowing data centers to go off grid and have passed laws recently. Yeah. But for the most part, I think there's a lot of overlapping interests here between the two kind of big, big commercial players here.

Matt Chester: It sounds like in a lot of ways it is up to the utilities to play that pushback role if, you know, the tech companies and the politicians are happy to lead with that so-called theater. And does that create any tension where the utilities have to be the one to do that? And what are the best tools that the utilities have to kind of hold their ground there?

Ari Peskoe: I think the theater aspect I would probably blame the White House for if I were going to assign any blame. And I think at the first press conference on this pledge was back in March. And I think the president gave away the game by saying this was PR help for the AI industry, because obviously this White House is a major supporter of AI development. And let's set all that aside. But I think that was really the goal was to remove what the White House perceived was a sort of political obstacle to, you know, I guess what they would say, sort of winning the AI industry. This is really not conceptually that hard of a problem.

It's a matter of figuring out how much the utility is spending to bring these data centers online and making sure that the data centers are paying the full costs of service and being transparent about it. There are a lot of cases across the country, and we found this when we were researching these issues, where deals are filed under seal between a utility and a data center. Even right now, actually, as we're speaking right now, the Louisiana Public Service Commission is holding a hearing about that giant Meta facility and whether or not the Public Service Commission will compel Meta to provide more information about how much electricity it expects its data center to consume because it's not being forthcoming with that information. So there's just a lot of secrecy. And that's not happening just at the utility regulatory world.

It's happening governments with their sort of economic development incentives, land deals that are happening. I think one of the big sources of pushback to these data centers is just sort of the lack of transparency about so many aspects of this development. So a lot of so a lot of parties think have responsibility for that and can be more forthcoming with the public about how much this is costing and who's paying for those costs.

Matt Chester: When you say the lack of transparency, it seems like the, you know, anybody who's involved might be benefiting from not having to disclose that. And who it hurts in the end is the public who is already upset about this. Are there any tools that, you know, the ratepayer advocates and the general public would have to force some more of that transparency?

Ari Peskoe: Yeah. There's been a change in that, and I think when one of these data center deals would come into a regulatory commission because they often require regulatory approval. When these were happening in 2023, 2024, what we often saw is that no one was really paying attention, right? And so these things would be filed maybe the state ratepayer advocate would would participate. But at the time, nobody really understood how big this was all going to become. And so there wasn't really robust participation in these proceedings.

That's clearly changed now. We understand that the scale of this issue and the potential impacts for everyone. So you can't kind of sneak one of these deals under. And so at least the ratepayer advocate itself under the you know, a protective order can get access to, you know, the terms of the deal and can respond under seal. So that's certainly better than nothing.

We're still not the public isn't getting a full picture and utilities and the data centers will argue that this is commercially sensitive information. And so it ought not to be disclosed. I don't know how I feel. I've mixed feelings about that sort of argument. But I think certainly in rate cases where the utility does have to disclose what it's spending, what it's spending for, I think certainly that's where we need to make sure that we're being clear on why the utility is choosing particular projects, why it's building certain transmission lines, why it's making decisions about certain types of capacity. And if it's because of a data center, then I think we need to make sure that we're appropriately charging the data center for those costs.

Matt Chester: That only seems fair enough. I want to go back to something you touched on earlier, which was the idea that for about 100 years, utilities have socialized infrastructure costs onto customers. And I want to dig into, is what's happening with the current situation, socializing the AI infrastructure, is that fundamentally different than what's always occurred with major infrastructure?

Ari Peskoe: Yeah. Well, again, I think the scale here really matters. I mean, you mentioned the Meta facility, which is something like five gigawatts. I'm not 100% sure, but maybe the largest energy-consuming facility in the country. If it's not, it's certainly way up there. And there are a number of these sort of gigawatt-scale facilities.

So for the power sector's perspective, it's like plopping down a large city basically out of nowhere. The only thing I think you can really compare that to is really a major industrial facility facilities, an aluminum smelter, a giant refinery that may even approach that sort of scale. And a lot of the, and so there's been a history of utilities cutting deals with those types of facilities. And a couple of things I think are different with these AI data centers. One is that a lot of the, these sort of large industrial facilities, many of them were built generations and the economics of the power industry were just different.

And so it used to be the case that if you go back to the 50s, 60s, when you added new demand to the system and you had to add new supply to match it, the new supply was actually cheaper. Yeah. than the old supply on the system. And so by a utility building a new power plant, you could actually lower the costs for everyone because new power plants just had captured efficiencies of scale and just efficiencies of operation compared to inefficient plants built in the 30s or 40s or something like that. And so you would lower everyone's average cost. That's just not the case right now.

For a number of reasons, it's just expensive to build new infrastructure. And so when we're adding capacity, we're raising capacity the average cost of the system. And so if we just socialize, we're probably raising everyone's costs unless you really get the rate structure just right, which again, we can get into how you might go about thinking about that. So that's one issue. The second issue is, you know, one of the justifications for... allowing an aluminum smelter, let's say, to have sort of a slightly discounted rate that gets subsidized by everyone else, is that it's an economic boon for everyone.

And so we're going to use our utility rates to sort of have an economic development effect. And I think there's a real debate around there about whether these data centers are similar. Certainly, they're huge construction projects. And so we're seeing sort of an immediate, you know, uptick in, you know, construction jobs and things like that. But long term, it's not clear that these are, you know, real, real job creators, but I don't, you know, sort of beyond the scope of the sort of utility rate issues I normally get into.

Matt Chester: Yeah. And it's interesting when you look at the major industrial facilities of the past, you mentioned it's an economic development. It used to be that, and still is, that utilities would almost compete and say, we want you to build this in our jurisdiction, and that's going to be such a great thing. And that's where the discounted power rates would come from. And We're looking at kind of a different environment when it comes to the data centers for those reasons you just mentioned. But at the same time, do you see something to be learned from that past of the major manufacturing facilities and using utility rates to lure them? Is there some applicability for what utilities can and should apply to the data center side of the equation?

Ari Peskoe: I mean, I really don't think so. Again, because the economics of the industry are so different. I think that's the big distinguishing factor and the scale of this. I mean, there are some utilities now that are seeing their loads with these new data centers increase by 30, 40, 50, 100% just because of data centers. So it's not one facility that is maybe increasing somewhat the load of the service territory. It's just, it's really... overwhelming, you know, existing, existing load. And so I'm not sure that the traditional way of just allowing a subsidy to flow through based on some economic development potential is really the right approach here.

Matt Chester: Yeah, that's fair. And then, you know, looking at kind of merging or compromising between the different models, is there an option that looks at like a compromise between totally socializing the costs and totally putting all the costs on the data centers? You know, we hear strong opinions on one side or the other, but is there a middle ground that we can or should be working towards?

Ari Peskoe: What we're seeing now is... isolating some costs and socializing others. And so the big one that we're still socializing for the most part is the high voltage transmission system. Utilities right now are building tens of billions of dollars of projects that are really needed because of data centers. And those costs are all, for the most part, not, I can't say all, but for the most part, are being socialized to everyone based on the theory that we all sort of draw power from the high voltage transmission system. The basic principle that federal regulators have applied to figure out how the costs of the system get shared, at least at the regional level, is that if you're benefiting from the new projects, then you pay for those projects in proportion of the benefits you receive.

And right now, the primary beneficiaries, if not the only beneficiaries, are the data centers of these projects, but the costs are still being paid. And then a lot of the projects have sort of a different theory of cost allocation, which is that while you're part of the utility's local network, and so everybody that's sort of captive to that network, all the local ratepayers have to share the costs in a formula set by state regulators that's usually based on how much energy each type of consumer, residential, commercial, etc., draws from the system each year. And so again, even if the only reason utility is expanding the system is because of a data center, we're still beholden to that old way of thinking about socializing the cost. So we still see that happening, even if there is a mechanism to put in place for a data center to pay for the new power plant capacity that might be needed for it.

Matt Chester: And then there's almost a door number three option that you alluded to earlier, which is these tech companies building their own generation are taken to an extreme. You know, people talk about could they become their own utilities? What are the chances that we see that come to fruition and what would it even look like?

Ari Peskoe: As I said before, in some states that's not allowed. Only the utility is allowed to sell electricity to somebody else. And so, you know, you could only maybe do it if all the infrastructure is owned by one company, right? One LLC owns everything. Maybe you could pull it off. But in a lot of states, it's still not going to be workable.

And so... We are seeing a couple of new laws. Utah, West Virginia have kind of allowed this sort of off grid, you know, outside the utility system developments. And then in a number of states like Texas, it's already allowed. And that's where we're really seeing a proliferation of data center and power plant combinations that at least initially expect to be off-grid.

And I think if the only goal of the data center is we need power now and we just want to get it as fast as we can, then that may be the way to go. You probably have to overbuild it to make sure that it stays reliable at the level that these data centers but that's an economic choice that they can make. But, you know, once you start to connect to the utility system, then you have to figure out sort of what's going to be the arrangement between this off grid, the formerly off grid system and the utility owned system and how are they going to exchange energy. And then you have to worry about what the price is is going to be.

Matt Chester: A lot to figure out there if they want to go that route. I want to pivot now. Recently, DTE Energy and financial analysts devoted much of their second quarter earnings call to data center negotiations. And we saw that you said that a consequence of this is that utility management is, as they're devoting more time to data centers, that means there's less time and less space to go to everyone else. I'm curious what you've seen have been the consequences of this, if any. Who's not getting the airtime they need? And do the data centers really need that much coverage? Or is that all just part of the theater we've been talking about?

Ari Peskoe: I mean, they get this attention because they require so much new infrastructure. And so whenever a utility builds a new power plant, new transmission line, there's a permitting process, there's a planning process. And so that would happen anyway, but usually it would happen for infrastructure that was needed for a broad swath of the utility ratepayers. And now we're having all these proceedings that are primarily because there is one customer that needs all of it. So yeah, there does have to be regulatory attention paid to these issues unless we just say what really ought to happen here is the data center should just build all of this on its own. And we don't have to worry about the cost impacts on everyone else if just the utility is contracting with some private party to build all of this infrastructure.

But as long as it's going through a utility... We have all these existing processes to get this infrastructure built and paid for. And look, if you go back in time of just a few years before this data center boom started, you would see, obviously, we still have rate cases. We still have long-term planning processes. But both of these are now... to a certain degree about data centers that weren't part of the conversation before.

And so that means necessarily everyone is spending more time thinking about all of the other issues. There was certainly a lot of momentum prior to this current administration in D.C. about clean energy issues about, you know, transmission issues related to clean energy and reliability and other purposes. And so, again, there's just, you know, regulators only have so much time, so much resources. And necessarily, a lot of those resources right now are going to issues related to data centers. So you think about all the technologies, whether it's alternative transmission technologies, grid enhancing technologies, things at the local level, there's a lot of potential progress that we could be making to figure out how to take advantage of these things, but there's only so many hours in the day. And budgets are, unfortunately, I would love for every public utility commission to have double the budget, double the resources that it has, but I don't know that that's a politically viable dream.

Matt Chester: I kind of want to distill a lot of our conversation down to, you know, the main crux of it, which is what are some of the alternative policy models that are being explored or should be explored for who should be footing these costs? And kind of a part two of that is for up to you, if we have a perfect world, what solution would be the best one that isn't coming to fruition for whatever reason?

Ari Peskoe: We talked before about a lot of utilities creating this data center rate class. And I think that's important as a step for assigning costs to that rate class. What we've seen so far, though, is that rate class being used to... relieve the public of some of the risk of this infrastructure build-out. This is an issue we haven't really touched on yet, which is one concern here with this build-out is that a data center tells a utility, you know, we're coming, and so we need you to build billions of dollars of infrastructure for us. And then for whatever reason, the data center never materializes, right?

Right. And so the concern would be that the utility, having spent the billions of dollars for the data center, goes to regulators and say, we want to put some or all of these costs on everyone now. And so the way to protect the public from that possibility is to require data centers to to sign a long-term contract with the utility before the utility starts building this infrastructure. We've seen a lot of progress on that front in a number of states, but we're still not seeing the sort of more granular cost allocation tied to those contracts. So the amount that these data centers are paying is often kind of picked out of the air.

It's, let's say, 80% of some charge that we set based on an earlier cost allocation scheme without updating it based on how the system is going to change due to these data centers. So I think the key thing is continuing on that line of progress of creating these data center classes. The next step is figuring out how to allocate costs to that ratepayer class or alternatively, I think that I'll just leave it at that. I think that's the important progress that needs to be made. And we also shall be doing that at the federal level as well, where a lot of these transmission costs are set.

And a lot of the, we have these interstate markets in many parts of the country run by the regional transmission organizations. We're seeing some increases in some of those markets due to data centers as well. And there's some progress being made in figuring out how to isolate or allocate costs to data centers. We got to keep going down that path. I think the one area where we could use more progress, is in competition.

The challenge for the public is when the utility builds the infrastructure, because the utility has the unique ability to put the costs on everyone's bill. But if you have a private deal between a data center and a power plant developer or a transmission line developer, those costs shouldn't be able to trickle down into everybody else's bill because it's a private contract between those two parties. And so if we can do more of that across the country, we'll reduce the risk that any of these costs go to the public. But a lot of the utility monopolies are protected very strongly, and the utilities defend those protections. And so that's one thing that's blocking the use of more competition that I think could benefit the data centers, because they would have more choice and be able to go out and procure the system that they want. And they're increasingly very sophisticated at knowing how to do that. And it would prevent these costs from trickling down to everyone else.

Matt Chester: Excellent. I know our listeners will be happy to kind of hear and parse that, you know, a game plan, you know, what we should be doing and, you know, mentioning our listeners. I know I teased at the beginning that we had asked our community ahead of you being on this episode, you know, what would you want to ask Ari? What would you want to ask Ari? you know, the electricity law professor from Harvard Law School. So I want to go through a couple of those questions we got from our community members and, you know, kind of get some quick hit answers for you. And the first one is, why not require data centers to provide their own power behind the meter?

Ari Peskoe: You could do that. I'm not opposed to that. I think, you know, this is also called co-location potentially, depending on the exact definition and configuration, but you could you could call that co-location. I don't see any harm in telling the data center to go out. And again, that's just what I mentioned. Have your own deal with a power plant developer that's not the utility.

And then there ought to be no way for those costs to trickle down. But that said, if you are going to connect it to the utility system, you're going to have to have some kind of contract in place for it exchanging energy between the two systems. And the data center is going to probably still lean on the system for emergencies and things. And so you need to work some things out. And that's not trivial. But I think that's certainly better than having the utility build the whole system.

Matt Chester: The next question we got is, does Ari see a real chance of data centers lowering bills of the other customers, as stated by some data centers?

Ari Peskoe: It's certainly possible you have to get the rate structures right. And the way I think about this is in terms of the transmission system. So let's get into it just a little bit. Typically, there's a rate for transmission service that we all pay that's based on basically the value of the assets in the transmission system that the utility owns, and we all pay for a tiny slice of that. Now, when a data center connects to the system, it uses so much energy that typically the utility or multiple utilities need to add to the system. They need to build more to the system.

What that can do is raise the value of the system and then raise everyone's rates. So the way you mitigate that is you tell the data center, okay, you're going to pay for the full cost of all of those upgrades. You're causing them, you pay for them. And you're also going to pay your share of the value of the system that everybody else pays. So what that does is it holds everyone harmless from those upgrade costs because the data center is paying for them. And it reduces everyone's rates a little bit because now the data center is going to pay for its share of the value of the entire remaining part of the system.

And then everybody's rates can go down slightly. We're not seeing that actually happen, but that could happen. And I think that would be a way for the data centers to help everyone.

Matt Chester: I think everybody would like to see that outcome. So we'll stay tuned on that one.

Ari Peskoe: Yes, I wouldn't say everyone would like to see it.

Matt Chester: Fair enough. Next question is, I realize this may be outside your field of expertise, but I wonder what the economic impact of increasing climate disasters will have on data use and data center development. And is anything being done to protect the states hosting their construction from potential economic collapse of the, you know, from said disasters?

Ari Peskoe: I think the short answer to that last part is no at least as far as I'm aware. I mean, look, we're seeing a surge of natural gas-fired generation due to data centers. You know, obviously has greenhouse gas emission, emissions associated with it. We're seeing this administration keep old coal plants online in the name of, data center growth. We're seeing some utilities do the same as well as some merchant private generators trying to revive old coal and oil-fired plants due to data center growth.

So there are emissions consequences. Translating that into specific disasters and disaster costs is beyond what I can do. But certainly there are environmental costs starting with greenhouse gas emissions, but you can go beyond that associated with this data center build out.

Matt Chester: Is noise raised as an issue in many data center controversies during the environmental impact assessment? Have you noticed that being something, you know, a linchpin or something people are talking about?

Ari Peskoe: I certainly see people talking about it online and you can find some videos of data center noise online and people are writing about it. I haven't followed in detail sort of specific permitting proceedings where these noise issues come up to know how, you know, local authorities might be handling them in one way or another.

Matt Chester: And then our last one from the community is, as AI data centers become one of the largest drivers of electricity demand, should energy law reconsider the traditional principle of cost causation to ensure that hyperscale customers bear an appropriate share of the cost created by their electricity demand growth? Or would that conflict with the principles of economic development and innovation?

Ari Peskoe: Cost causation is kind of the flip side of what I said before, which is the beneficiary pays principle. And I think cost causation is exactly the right principle here because it is the data center that's causing this new infrastructure to be built. And so I think the data center should be paying for all those costs.

Matt Chester: Great. Well, thank you for indulging our community in their questions. And thank you for doing the same with me, joining us on the podcast and having a really great discussion. I've appreciated listening to everything you have to share here.

Ari Peskoe: Thanks for having me.

Matt Chester: Of course. And thanks again to our listeners for joining us today. And once again, this has been Power Perspectives, and we'll catch you next time.