Ari Peskoe:

Welcome to Clean Law. I’m Ari Peskoe, director of the Electricity Law Initiative and I’m here with Eliza Martin, Legal Fellow, and we are going to be discussing our new paper, Extracting Profits from the Public: How Utility Ratepayers are Paying for Big Tech’s Power. Thanks for being here, Eliza.

Eliza Martin:

Anytime.

Ari Peskoe:

So I got interested in this topic because data center energy demand has been the big issue for the past year or so in the power industry and the forecast for growth from data centers, these giant warehouses filled with power-hungry computer chips, has been astounding. So Eliza, what are some of the projections you’ve seen out there for how this industry might grow over the next few years?

Eliza Martin:

I mean it varies by utility, but we’re seeing in Texas, Encore has said that they have about 82 gigawatts of potential data center load. INM in Indiana has said they expect its system to rise from 2.8 gigawatts to seven. Georgia power anticipates its sales doubling. So it’s just like all of the existing investment that we’ve made into the utility system, just doubling or tripling that is what we’re seeing by 2030 or the middle of 2030.

Ari Peskoe:

Yeah. So we’re seeing some of these utilities who have built their demand growth over a century now saying that over the next five to 10 years, they think it might double, or even, in that Indiana example you gave, triple. But we should say that the utilities do have an incentive to inflate these projections. So why is it that a utility might want to give a sort of very rosy picture for growth in its service territory?

Eliza Martin:

Well, utilities make their profits by investing in capital infrastructure. So that’s like the transmission lines, distribution lines, like everything that we see, that’s how they make their profit. So they have an incentive to project that there will be a lot of growth and then try to build that growth because shareholders get a return on their investment or a chance to earn a return on their investment for capital infrastructure costs. So they have an incentive to say that there’s going to be a bunch of data center growth and then go build the infrastructure that they think that they need to meet that demand.

Ari Peskoe:

I’ll just add that these are publicly traded companies. Their stocks are worth more than a trillion dollars these days, though I haven’t checked quite recently enough to know that that’s still true. But they like to go and tell their investors about all the stuff they’re going to build because that’s how they profit. I get these emails on a daily basis from stock analysts about how data center growth is going to be fueling profits for the next 10 years. So these CEOs of these companies get paid based on the financial performance of the companies that they run, and so there’s a lot of incentives to push up the stock value, to promote earnings growth and to try to capture these data centers in order to fuel their profitability for the long run. So let’s get into some of the mechanics from the public’s perspective about what happens when the utility does build something, it builds a new power plant, for example, for utilities that do that sort of thing, or it builds another new piece of infrastructure to meet growth. How do we end up paying for the utility’s growth?

Eliza Martin:

The public ends up paying through the public utility commission or public service commission rate setting process. So generally the way that it works is a utility says we spent X amount of money this year on capital investments, on operation and maintenance costs, we have this amount of money that we need to recover from our captive ratepayers, so me and you, and then everyone fights about that amount of money. So the utility comes forward with this amount of money, parties are in the docket responding and saying, no, you should have lower utility revenue requirement. And then after the public utility commission sets the revenue requirement, then as part of that process, they allocate those costs across consumer groups.

Ari Peskoe:

So yeah, let’s stop there for a second. So we have this two-step process. One is setting the revenue requirement, which is a really utility friendly term. How much money does the utility want to collect from the public? And a key component of that is called the rate base. I’m bringing up that term because I think we’ll probably use it later. And that’s just the value of the utility’s capital assets, which is all the things that it has already built that are not fully depreciated plus whatever new stuff it has just built. And so the public is interested in these proceedings because we pay for the utility’s infrastructure. And so you have big energy users like industrial concerns, factory owners, they participate. You have some big retailers like Walmart in many states is a huge energy user, and they participate in these proceedings. And in pretty much every state there’s a rate payer advocate that’s a state government office that is there on behalf of residential ratepayers.

And so they’re all there arguing for a lower revenue requirement. Ultimately though the regulator then decides this is the amount that it’s going to be, and then there’s sort of like the food fight after that. This sort of cost allocation process can be a bit messy.

Eliza Martin:

Yes. And the cost allocation process is really about determining what amount of that big utility revenue requirement, how that pie should be split across all the different people, all the different groups of consumers. So Walmart or industrial consumers will pay a different amount of that pie, I guess, as opposed to like residential homeowners who will pay a different amount. And generally these amounts are theoretically supposed to align with the cost that the utility incurs to serve those people. So like to serve a bunch of homeowners, you need a lot of distribution infrastructure. So you need the lines and wires to deliver power to people’s homes, which can be spread across a longer distance versus if you’re serving a factory, you need different sorts of infrastructure, so your costs theoretically look different. But this is a really subjective process just because every party’s like looking out for themselves and trying to reduce their share of costs.

So well lawyered parties are usually going to do better in these cases because they’re really technical, they’re really long proceedings, and so if you have a good advocate who is diligent about looking through the utility filings, is diligent about trying to interrogate the other party’s claims, then you can end up reducing the amount of money that this group pays and then increasing another groups, but it’s kind of a fight.

Ari Peskoe:

Yeah. I’m going to go back to this high level principle you mentioned, cost causation, that the rates that a particular type of consumer class pays should align with the costs that the utility incurs to serve them. And so that principle ought to apply to growth. So if the utility is building out its system because society is demanding more energy than under this cost causation principle, then if you’re the sort of class of consumers that’s driving that growth, you should be responsible for paying those costs. So if we go back in time, the last time that the utility industry saw the kind of growth that some utilities are projecting now is back in the post-World War II era. In the 1950s and 1960s, electricity consumption in this country was growing at something like 10% per year. But there it was being driven by sort of all sectors of society. We had massive industrial growth, we had the build out of the suburbs, growth was everywhere, and so the costs of that growth were sort of shared among all rate payer classes.

Everybody was paying for it based on the idea that this growth was going to benefit everyone, population growth, economic development, the sort of things that are just sort of like hallmarks of Western civilization. But now we have this growth being driven by data centers and that seems to change things a little bit, or, at least in our view, ought to change how these costs are allocated. But our concern in the paper though is that utilities are interested in attracting these data centers to come to them, and one way they can do that is by offering discounted rates to the data center and surreptitiously sort of requiring everyone else to cover any shortfall of those discounts. And we can talk a little bit about sort of why we’re skeptical of the utilities, what about the utilities and their history makes us skeptical that they’re not doing this?

Eliza Martin:

Well, utilities have a long history of doing this. In the paper we get into like a couple of reasons why we’re skeptical that utilities are doing a good job on this because the rates process is really driven by utility companies. So they will exploit their advantage. Like, they’re the only ones with access to their cost information. They don’t open their accounting books for people in these rate cases. So there’s a history of utilities lying in their rate cases. There’s been some big scandals recently that utilities are only caught when there’s a federal investigation, First Energy in Ohio, this has been a big issue. But every year FERC, the Federal Energy Regulatory Commission, puts out its audit report and every year there’s instances of utilities lying in their rate cases about attribution of costs to certain accounts and there’s informational issues within that process. So that’s, I guess, one of the first reasons why we’re skeptical of utilities being, I guess, good Samaritans of this process. So there’s that.

And then I think the other issue is state PUCs are political economy bodies, but inherently state PUCs in every single state, they’re either elected positions or they are appointed by the governor. And state public utility commissions are generally underfunded, I think is fair to say, so they struggle with sort of responding to these claims that utility companies make if there’s not other good parties in the case that are challenging what utility companies say. And because state PUCs make decisions based on what’s in the record, what’s in the evidence, it can be difficult for a state PUC if there’s not other good parties in the proceeding to sort of challenge what a utility company is saying. So you have that aspect of it, but then you also have this issue, which is like a state PUC commissioner who maybe wants to be reappointed by the governor, they can be responsive to maybe the governor’s needs over a rate payer’s needs or vice versa.

If you’re elected and you want a donation from an industrial group or something, then maybe you’ll favor them over something else, which I think is just a classical kind of political economy story. So those are two of the reasons.

Ari Peskoe:

I’ll add just one more, which is many states have policies favoring investment of data centers. They want these investments in their states. And so if you’re a utility regulator, you’re a potential veto point through these rate cases and other proceedings we’ll talk about. But that’s a lot to ask of a utility commissioner to veto a billion dollar investment, one maybe that the governor has already been promoting. And so that might give utilities a little bit of extra leeway because there’s this political momentum for these deals to happen. And I think you mentioned the history of utilities exploiting their monopolies. One recent case we bring up in the paper is this Duke Energy case, which was a Fourth Circuit decision from 2024 where there was a natural gas developer that was trying to compete with Duke in North Carolina. The natural gas developer was going to sell its energy to municipal utilities. These are, for the most part, small towns in North Carolina that own their local distribution infrastructure, but relied on Duke to supply power.

And this natural gas developer was able to supply power at a much lower rate than Duke, and so it was picking off Duke’s customers. And when one of Duke’s larger municipal customers was getting ready to renew its contract or potentially sign with this new developer, Duke offered a major $300 million discount and some internal Duke documents disclosed through litigation revealed that Duke had come up with a plan to shift the cost of that discount to other Duke ratepayers. In other words, Duke was going to force its captive ratepayers to subsidize its competition with this new natural gas fired power plant. And so we see utilities do things like this, try to basically abuse their monopolies in this way, and we’re concerned that they may be doing the same sort of thing here with regard to data centers.

We talked about this rate setting process, but one of the, I think, main findings of our paper is that data centers often sign up with utilities for service through a process that completely bypasses the normal rate setting process. So these are called special contracts or secret contracts. Why don’t you talk a little bit about what you found about these deals between utilities and data centers.

Eliza Martin:

So these are contracts that essentially, like you said, allow any customer, I guess, that has a special contract, but what we saw was data centers, that will take service from a utility company with terms and conditions of service that aren’t otherwise applicable to any other party. So I guess theoretically like the origins of this are, you would have a large customer who maybe you can bring onto your utility system without having to add a bunch of new infrastructure to serve them so you have extra capacity, I guess, on the existing system. And so if you can bring on a customer without having to build out new power plants or having to build out new infrastructure, then theoretically that is good for everyone because it lowers your cost. But what we’re seeing with data centers are these terms and conditions that are otherwise not applicable to anyone, but the load is so big, like data centers are so enormous that there’s no way for a utility to serve these customers without having to build out their infrastructure.

So the result of that is that the utility can end up shifting costs because if it loses money on that contract deal, like what you were talking about with Duke, then it can use its monopoly to shift those costs to other ratepayers like you and me who don’t have a choice in where we go to get our electricity. So these data centers can choose where to locate. A utility can offer them a sweetheart deal, but we as ratepayers can’t, and then we get hit with the cost of it.

Ari Peskoe:

And I think one of your key findings here is that these are secret contracts. So why don’t you talk a little bit about the sort of confidentiality of these deals and what we do and don’t know about what’s in these contracts?

Eliza Martin:

Well, we don’t know a lot about what’s in the contracts. Basically a utility will file a contract with the public utility commission and claim that the information in the contract is proprietary. So they’ll make a claim for privileged treatment of the contract because on the theory that their terms of service to this data center need to remain confidential because otherwise another utility could undercut them and then you would lose that customer. The results of those confidential claims for treatment vary. In some states, you can’t even see the contract or it’s all blacked out basically. In other states, you can read the words, but there’s no pricing terms, there’s no information on the total load. So that makes it really difficult to understand the scale of the subsidy that’s potentially there. So basically we know that these exist, but it’s really difficult to measure the wealth subsidy that’s sort of being offered. So it’s a real problem because a public utility is a public service, I guess, and we have no idea what terms it’s offering to its customers that we’re potentially paying for.

Ari Peskoe:

Yeah, these are monopolies and we have a public body charged with regulating them to protect consumers. And that’s hard here. One reason you mentioned earlier is because there are not a lot of other parties participating in these proceedings, and I think that’s largely because even though these sort of special contract or secret contracts have existed for decades in many states, consumers often thought that they didn’t really have a direct interest in these proceedings because there would only be an indirect effect on prices. If this wealth transfer that we’ve been talking about was actually happening, it may have some small increase on future bills for everyone else. But what’s different here, I think, is the scale of these data center contracts. We know that these facilities are hundreds of megawatts that particular utilities have already announced, they have gigawatts of contracts already signed. And so the effect on other consumers could be quite significant.

And I’ll just bring up my favorite example here of a special contract, which is one where it’s far worse than the example of hypotheticals we’ve just been talking about because there is no regulatory review at all. It’s not even blacked out. So this is an example in Louisiana. The utility there is called Entergy. They’ve signed up a data center client with Meta. They signed a contract with Meta, the company that owns Facebook. And this is a facility that Entergy says it’s going to be building more than two gigawatts of gas-fired power plants for. And just to give a sense of what that means, on a hot summer day when it’s using the most energy, the city of New Orleans consumes about one gigawatt. So this is enough natural gas-fired power plants to power two city of New Orleans on a hot summer day. And Entergy has taken the position that it does not have to file this contract with the public service commission, with the regulators there.

Entergy has also said in the course of the proceedings where they’re trying to get approval to build these giant power plants, that it does not know how much electricity Meta is going to be using at this data center. So we don’t know how much of this two plus gigawatts of gas-fired capacity will actually be used by the data center and how much Meta is just trying to take advantage of sort of the political winds here to get additional gas-fired capacity through this sort of what it’s hoping is an abbreviated regulatory process. And this raises a huge concern that we talk about in the paper. It’s not just that utilities can transfer costs to ratepayers, they also transfer risk. And so there could be any number of reasons why Meta decides to try to back out from this deal, or maybe it only wants to build a facility that’s going to use half as much energy as it had originally anticipated. But if Entergy goes through with building all this infrastructure, it’s going to be at least $3 billion, and we don’t know the extent to which Meta is going to be on the hook for any of these costs.

And so you could have a situation where it’s just the people of Louisiana are left holding the bag for all of these infrastructure costs that are these stranded costs if Meta either disappears or just doesn’t use as much energy as it anticipates. This is another one of these hidden terms from these special contracts that could have huge consequences for ratepayers, and we just don’t know who’s holding the risk in these deals. So secret special contracts are one huge red flag that we raise in this paper. Another area we focus on is the fact that these cost allocation processes we’ve been talking about can happen at the state level and also happen at the federal level through the Federal Energy Regulatory Commission. So Eliza, maybe you could walk us through the sort of PJM regional transmission development process and how it’s potentially causing the public to pay for the data center build-out in Virginia and other parts of that region.

Eliza Martin:

So this is a normal process. We just highlight it in the paper because it has the potential to result in residential ratepayers paying a lot of transmission costs for data centers. So basically when PJM, which is this massive utility alliance in, I guess, the Mid-Atlantic, that stretches all the way to Chicago, so it’s kind of a behemoth. So PJM decides that it will be building a lot of transmission lines that stretch in the region. So it’s in the PJM region. They need to build transmission lines mostly to meet data center load growth in Virginia. So Virginia has had this explosion of data centers locate there, and as a result of all these data centers locating there, we need new transmission. PJM goes to FERC and FERC approves those transmission costs. So at that point, there’s a very complicated process at FERC where these transmission costs get allocated to the member utility companies within PJM. Then those member utility companies at the state level have agreements on how to allocate costs to its various rate payer groups.

So the results of this very complicated process just means that residential ratepayers in Virginia and Maryland are going to be paying a bunch of money for transmission that’s really just built for identifiable data center customers in Virginia. So this PJM example is $5.1 billion, which is an insane amount of money. And the fact that residential ratepayers are getting stuck with that bill because of Virginia’s policies about this kind of violates this idea that the group that’s causing this should be paying for this. Residential ratepayers in Virginia and Maryland aren’t the ones who need this transmission. It’s built for data centers. It’s not built for the broader public.

Ari Peskoe:

Yeah, I mean the premise behind the PJM cost allocation process is that new high voltage transmission is broadly beneficial, and so costs are spread out based on that premise. But then when it gets to the state level, as you were describing, some states are operating under very old models for how to share those costs at the state level. So in Maryland, for instance, the Maryland utilities will pay about 10% of that 5.1 billion, which is about 500 million, and under Maryland’s approach, residential ratepayers then pay for about two-thirds of that 500 million. So something like $300 million for regional transmission expansion, as you described, is largely being driven by data center growth. So that’s a mechanism that we talk about in the paper. There’s a couple of other subtleties about it that we don’t have to get into here about some stranded cost concerns related to that. Basically if some of the data center load fails to materialize, even if a utility has strong contractual commitments with its data center customer that might protect the utility’s own ratepayers through some of the mechanics of the regional process that neighboring utilities and their ratepayers could end up being stuck with some of these costs.

So there’s a lot of sort of mismatch between things that FERC is doing and things that are happening on the state level. There’s another issue involving FERC that we talk about in the paper as well, and that’s called co-location. I guess one of the main examples of this is where a data center locates next to a nuclear plant or sort of physically next to it and is literally connecting to the nuclear plant. Of course it takes more than just an extension cord. There’s some complicated, expensive infrastructure that goes into this, but there’s a contract between the nuclear plant owner and the data center and then the complexities involved here with how does that contract interact or not interact with the regional market. So what’s the state of debate on co-location now?

Eliza Martin:

Well, there’s a lot of different issues. It’s currently a live proceeding in front of FERC, so they may hopefully make decisions on it somewhat soon. But there’s a lot of different issues. So first, basically the phenomenon that you described were it to happen, were a data center customer to co-locate with a nuclear plant, you would essentially remove that entire generation asset. So all the power that this nuclear plant would normally be bidding into a competitive auction to supply a utility with its power, that entire generation asset is gone from the market. So because we run, in some regions, regionally, we run competitive auctions to supply power. If you take out a whole big chunk of that generation asset that normally clears the market, so that normally is sold to a utility company which then sells it to the customers, to ratepayers, or passes those costs on, if you remove that whole chunk of generation assets, then you have higher priced generation that will end up clearing the market.

So functionally that means dirtier generation of coal plants will probably clear the market or you just won’t have enough. So that’s another concern is you just will not have enough generation. So you’ll not only have wholesale power prices increase, but you could potentially have a shortage of generation.

Ari Peskoe:

Yeah, the bottom line there is that for consumers they’ll see higher power prices. And then there’s this other issue about who pays for the transmission system. So what’s the controversy there?

Eliza Martin:

So the controversy around who’s paying for transmission is up for debate in front of FERC because when co-location happens, a data center connects directly to the power plant behind the plant’s point of interconnection to the utility transmission network. So theoretically it’s a way to cut out the utility. No longer is the utility having to use its transmission or distribution infrastructure to move power from a nuclear plant to its customers. So there’s debate around, if that infrastructure isn’t being used anymore, should the data center customer be paying for it? And there’s a lot of technical issues like is it possible for a nuclear power plant to really ever be fully isolated from a utility transmission network? That’s up for debate, I guess, in front of FERC, but utilities are obviously upset about this arrangement because it cuts out potential profits for them. So they profit from having and operating their distribution and transmission network and making any of these upgrades that would be required.

Ari Peskoe:

So we have this competition between the utilities who want to sign up data centers as their customers and these nuclear generators who also want to have data center customers. And so this co-location fight right now before FERC is really pitting these two power sector interests against each other. We’ll see how FERC decides. FERC is probably going to issue some decision about the rules of the road in PJM perhaps over the summer that might inform what happens in other parts of the country. This transmission rate issue that you mentioned, Eliza, I think is really important for consumers. It’s an example perhaps of where new data centers could actually help reduce costs for consumers. Because if data centers are paying for part of the regional transmission system, then perhaps that means the rest of us can pay somewhat less. Although at the same time, data centers are also driving the need for a new transmission so it’s not clear how this all exactly will shake out for consumers.

We have some recommendations in the paper about what we think state regulators and state lawmakers ought to be doing. I would say that there’s sort of classic no silver bullet here sort of issue. There’s not one thing that regulators can do to protect consumers here. But if there’s one recommendation you wanted to highlight, what would you point to for something that either regulators or state legislators ought to be considering here?

Eliza Martin:

Get rid of special contracts for data centers. Shifting to a tariff proceeding obviously isn’t a foolproof issue because we still end up with these cost allocation fights, but the procedural processes around these secret contracts is just so egregious that it’s shocking that public utility commissioners and legislatures are okay with this happening still. So I guess that would be my one recommendation if I were in charge of anything.

Ari Peskoe:

Yeah, and unfortunately we’ve seen lawmakers go the other way. We have an example in the paper of a Mississippi law passed last year that’s designed to attract a $15 billion Amazon data center. And there the legislature says explicitly in the law that regulators have no authority to review the contract between the utility, Entergy, again, incidentally, and Amazon. The one recommendation I would highlight is about competition. In about 35 states, utilities are still the dominant owners of power plants, and these data centers, as we’ve been talking about, use a lot of energy, and to meet this growth, utilities are potentially making a lot of money by building new power plants specifically for these data centers. That’s an industry that is now highly competitive. And we can protect consumers by requiring the data centers to sign contracts with a generation company that completely takes consumers out of the contract. So those two parties, the data center, the independent generation developer can allocate the risks and costs between them and ratepayers would be totally isolated from the deal.

That would be the recommendation that I would highlight. But of course, utilities have been fighting against that sort of competition for a generation. But I think we have this moment, perhaps even this crisis, due to rising costs due to data centers that ought to cause legislators to take another look at this issue and perhaps see competition as one potential solution here. One last thing we talk about in the paper is about, you know, if we look at all these subsidies for data centers and we kind of say, well, are they worth it? And we have a long history of cost shifts in utility rates, some of them are unintentional because cost allocation is complex, some of them are intentional. The example we give in the paper is about state energy efficiency programs where a utility or a contractor will come to your house, for example, and give you a free home energy audit and will give you discounts on insulation and other things that can save an individual money on their energy bills.

And we justify those sorts of subsidies because the system as a whole benefits because if you lower energy consumption in one particular home, it can help reduce system-wide costs. And so is there any evidence, Eliza, that we’ve seen about whether these data center deals are good for ratepayers?

Eliza Martin:

Well, we have no public accounting of if they’re good or not. Virginia has started the process of trying to, I think, account for the impact of data centers on its infrastructure, but I think we still don’t have any answers about whether this benefits ratepayers. And for all the reasons that you mentioned about why this hurts ratepayers, maybe we should take a step back.

Ari Peskoe:

Yeah. The Virginia study came out I think the end of last year. It’s a great report from the Virginia legislature. The scope of the report, there’s a whole host of issues about data centers that we don’t touch in this paper. There’s clean energy issues about how we’re powering these data centers and whether it’s causing an increase in greenhouse gas emissions, the water consumption from these data centers is tremendous because it gets quite hot with all of that computing power in these warehouses and so you need a lot of water to cool the place down. There are backup diesel generators at these facilities which have their own environmental issues. There are issues about land use, about noise. So there’s a whole host of issues that we’re not touching in the paper that are covered in this report. And on the one issue that we are covering about utility rates, the legislature had commissioned a study that found that roughly speaking, so far, the data centers were paying for their energy costs, but they projected that that was no longer going to hold true in the immediate future.

As this growth really expands, it sort of found that residential ratepayers are going to be paying quite a substantial subsidy for data centers moving forward. The last issue that we’re concerned about in the paper is just with power sector reforms. The utility industry is beholden to a very old business model that, as we’ve described, rewards utilities for investing in large-scale infrastructure projects. And we have a lot of new technologies today that can try to sort of squeeze more efficiencies out of our existing system. Utilities generally disfavor them because they’re not capital investments that they profit from. There’s a whole host of new operational planning practices. It’s been a long effort to try to get utilities to build, sort of connect to each other in a way that’s going to benefit consumers by improving reliability, allowing more clean energy onto the system. These are all very complicated issues that people have been talking about for 10, 20 years, and now utilities just see these shiny data centers as an opportunity for massive, easy profits. And it’s just another reason to kick the can down the road on some of these hard-to-achieve reforms.

And so I don’t think we are particularly optimistic that data centers are going to be benefiting consumers. I think one conclusion that I take away from the paper is that the power system is really changing with these data centers. We could be looking at some utility systems where these are huge long-term customers and that might really change some of the dynamics in utility regulation as these data center companies may be looking to protect their long-term deals and kind of potentially change some of the dynamics at public utility commissions. This is going to be a big story to watch for a long time, I think. So Eliza, thanks so much for coming to talk about the paper today, and we hope folks will take a look at it. The link is in the show notes. We will see where this issue goes from here.

Eliza Martin:

Thank you, Ari.