Lawfare Daily: Energy Security, Data Centers, and AI

[Intro]

Mandy DeRoche: The price of Bitcoin is correlated with how much mining happens because the reward is a Bitcoin or a percentage of a Bitcoin for mining more. And so as the price of Bitcoin goes up, the rewards for mining goes up. It means the energy intensity and the energy consumption goes up, meaning the pollution goes up.

Kevin Frazier: It's the Lawfare Podcast. I'm Kevin Frazier, a senior research fellow in the Constitutional Studies Program at the University of Texas at Austin, and a Tarbell Fellow at Lawfare, joined by two leading environmental attorneys from Earthjustice: Christie Hicks leads Earthjustice's nationwide efforts to phase out fossil fuel power generation, and advance equitable energy access across the regulatory landscape. Mandy DeRoche has led significant legal challenges over data center energy use from contesting cryptocurrency mining operations in New York to addressing grid stability concerns in Texas.

Christie Hicks: So it's really harmful when communities then are paying both for upgrades to the grid that they're not benefiting from, and that they are having to pay more for the privilege of having these data centers or crypto-mining operations located in their communities.

Kevin Frazier: Today we're examining how their work in environmental protection and energy regulation intersects with the Biden administration's recent National Security Memo on AI.

[Main Podcast]

So approximately 5,000 data centers ensure that our digital world keeps going. They keep the cloud operating, they allow for advances in AI, and they ensure that e-commerce can continue. And that number is only going up. According to National Security Advisor Jake Sullivan, quote, if we don't rapidly build out this infrastructure in the next few years, adding tens or even hundreds of gigawatts of clean power to the grid, we will risk falling behind, end quote.

Well, those facts leave us with some difficult questions. Realizing national security priorities and national goals while also making good on climate pledges. Answering those questions is, though, way beyond my pay grade, and so I'm very lucky to be joined by Mandy and Christie today. And Mandy, let's start with you.

You've been at the forefront of challenges involving cryptocurrency mining facilities and data centers, particularly cases like Greenwich in New York. Can you walk us through what initially drew Earthjustice’s attention to these facilities and how these cases have kind of proceeded?

Mandy DeRoche: Let me start by, I don't know if our listeners need to know that cryptocurrency mining is a particular type of validation method. Most cryptocurrencies do not use it. Proof of work means that there is energy use that is needed to validate the currency. I think there are something like 3,000 coins out there that don't use proof of work validation method, and don't use any, or very small amounts of energy. For example, Ethereum switched away from proof of work mining to proof of stake and their energy use decreased by 99.5 percent or something really large like that.

And so what we're talking about here is proof of work crypto mining. It is highly energy-intensive. It's racks and racks of specialized machines. They're not regular computers anymore. They are very specialized and they need lots of energy and lots of actually fans. They're very hot to cool them. And those cooling fans also use a tremendous amount of energy.

And the case that you mentioned Greenwich. It is in upstate New York. It's in the Finger Lakes. We have here at Earthjustice have worked with community groups and environmental groups in the Finger Lakes for many years. And in 2020, our longstanding clients up there called us and said, this power plant on Seneca Lake, it was not really running for years. And it was running a little bit, but now it's running all the time. What is going on over there?

Well, what was going on over there was that a cryptocurrency mining company called Greenwich Generation had purchased the power plant. At first with the intent to use the electricity markets to bid in, you know, into the market, sell electricity into the grid when it was needed. So that plant did not operate from 2011 to 2016, zero emissions, zero operations.

Then Greenwich Generation bought it in 2016 and it operated during sort of what we call peaking times, you know, hot days in July, cold days in January, but not really the rest of the year. That company that bought the power plant wasn't making sufficient amount of revenue and profits. And we have some, you know, documents saying we need to find another way to make money here.

The way to make money there was to start mining Bitcoin, what we call behind the meter, meaning using the energy at the power plant and not sending the electricity into the grid, but keeping it on site to supply the energy to these hungry, hungry machines.

So luckily in New York, we have a really strong climate law that passed in 2019 and went into effect in 2020, right when this plant started mining cryptocurrency. And so the big litigation that we've had there, is over its air permit and its substantial increase in both greenhouse gas emissions and local air pollution.

The plant also is generating a tremendous amount of water pollution and noise pollution. And these machines, these specialized machines, they either burn out quickly or are replaced quickly with faster machines. So there's also a tremendous amount of electronic waste from these facilities. And unfortunately, after we learned about Greenwich, we started learning from our clients in other parts of the country that this was happening elsewhere.

And it really ramped up in 2021 when China banned crypto mining, for many reasons, but including environmental impacts and climate impacts. And that's when we really saw an exodus. Most of the mining in the world was in China and it, a lot of it moved here. Estimates are now that the United States is home to about 50 percent, somewhere between 40 and 50 percent of the world's crypto mining.

And so we started working with clients in a lot of different places. There's another behind the meter gas plant up in the Niagara region in North Tonawanda, same thing. Bought the plant, wasn't operating very much, is now operating 24/7. Generating huge amounts of greenhouse gases, local air pollution, water pollution, etc.

There's two waste coal plants that we're actively involved in. Waste coal is the dirtiest type of coal out there. And these two waste coal plants also sort of same fact pattern, were uneconomic, were on their way to retirement or not being used very much, and are now running 24 seven to mine cryptocurrency. We saw a coal plant in Montana, similar. And so that was repeated over and over again throughout the United States.

Kevin Frazier: Wow. And so there are so many threads to pull on there. Number one, I think it's just important to call attention to the fact that we are talking on November 6th, the day after the presidential election.

And it's worth noting that in addition to waking up to the news of an incoming Trump administration, I also saw that cryptocurrency stocks and prices have surged in response to the administration. And so, one question I have would be, are we anticipating that proof of work crypto mining will continue, or proof of work as a mechanism for crypto, is that going to increase, or are we seeing more coins change to more sustainable measures?

Mandy DeRoche: Any new coin that becomes a thing, is not proof of work. All of the new ones use proof of stake or federated consensus or some other validation method. Bitcoin and a handful of others, maybe three or four others use proof of work. They're the older coins. You can't do as much with that validation method, from a technology standpoint.

So these are the older coins that aren't really going away. So, when the price of crypto generally goes up there, there still needs to be a pause and to consider whether it's a proof of work like Bitcoin or if it's something else that doesn't really have an environmental impact.

But the price of Bitcoin is correlated with how much mining happens because the reward is a Bitcoin, or a percentage of a Bitcoin for mining more. And so as the price of Bitcoin goes up, the rewards for mining goes up, means the energy intensity and the energy consumption goes up, meaning the pollution goes up.

Kevin Frazier: This is such a fascinating dynamic, thinking about how this is all playing out at a hyperlocal level, and/or a state level while often being framed in terms of national economic priorities or national security interests.

So Christie, I was hoping that you could explore for us a little bit about this state dynamic where on the one hand, as Mandy pointed out, Earthjustice's claims against some of these companies are reliant upon positive state laws protecting the environment. While, on the other hand, we see that states are kind of leaning into this data center craze as a economic tool and as a source of greater community prosperity. So can you kind of walk through those dynamics and how you've seen states think about the cumulative impact of data centers on energy goals, while also the sort of economic concerns that are also at the fore for a lot of these states?

Christie Hicks: Yeah, well, as you heard just a little bit about, a lot of the legal issues here are going to be at the state and the local level. And that's because most of our energy regulation really happens at the state level. There's incredibly important federal regulations that come into play, especially with respect to the pollution from power plants, but a lot of this is regulated state by state, even utility by utility. And so that's sort of where the rubber meets the road on how the federal regulations actually get implemented in the real world.

And it is the energy intensity and the fact that these data centers and proof of work crypto mining operations. Not only are they using up energy or bringing energy back on the grid from fossil fuel plants that might have otherwise retired or even had already retired, they're also pulling a lot of the energy that would otherwise go to other customers. So that has ripple effects economy wide.

If we want to talk about the economic impacts of data centers and proof of work crypto mining that everyone's paying for grid infrastructure that's benefiting these really large load customers, and everyone is suffering the public health impacts of the fossil fuel plants that are staying or coming back online in order to serve them.

I think it's a really interesting question about the supposed economic development potential, especially of data centers, because we know that has been touted by both the large energy users themselves as well as local elected officials.

We know a few things to be true about this. We know that they often do not actually have a net tax benefit once you compare the incentives that they're receiving, the tax incentives that they're receiving and then that they are not also contributing to the tax base. We know that these are not a significant contributor to jobs in the areas, and we also know that jurisdictions that have previously welcomed these facilities with open arms, thinking that it was going to be an economic boon for the area are backtracking on that. And Mandy, if I could kick it back to you, there's been some really fascinating statements including from elected officials on this.

Mandy DeRoche: In states where there's been an influx of cryptocurrency mining operations, less so from data centers. Mostly, I'd like to give an example from Texas and from Arkansas. Both red states, both which their state governments welcomed crypto miners with open arms, I think on inflated promises of jobs and high tech industry coming to places.

Well, Arkansas in 2023 passed a bill called the Right to Mine bill. A right to cryptocurrency mine bill. It basically preempted state and local governments from regulating the extra negative externalities of crypto mining. That bill passed in the Arkansas legislature, I believe within a week's time, I don't, not sure that folks knew what they were voting on.

And in fact once the crypto mining operation started flooding into Arkansas, there was an outcry from local communities of all political and economic stripes. Arkansas called a special session in the next legislative session to repeal almost all the provisions of that Right to Mine bill. There was really a case of buyer's remorse and legislators went on the record saying…

Kevin Frazier: Mandy, I believe, sorry, you meant miner's remorse?

Mandy DeRoche: Hahaha! I meant buyer's remorse! But I like that.

And then similarly, we've seen, you know, Texas is also and has been a hotspot for Bitcoin mining. The politicians there for years have welcomed Bitcoin mining with open arms. And said that they're open for business and please come and they did, they flocked there. And recently both the grid operator, the public utility commission, state and federal politicians there, have questioned the value of these operations on their grid.

The lieutenant governor, I'll pull up the tweet in a moment, but, you know, after a legislative hearing on the impacts of crypto mining and data centers on their grid, the lieutenant governor was shocked to know how, how little economic development and how few jobs were produced for the amount of energy that was coming onto the system and how much Texas ratepayers and taxpayers will need to pay out of their own pockets for the new generation to serve this large digital load.

Kevin Frazier: Wow. So we see this hope that data centers and crypto mining activities will produce jobs and there's sort of that economic rationale. So that obviously usually has political persuasive effect. If you compound that now with this sort of national security issues and motives that are being touted by, for example, National Security Advisor Sullivan, that makes it even harder to push back against efforts to invest in data centers and crypto mining.

So I want to point out, though, Mandy, you've highlighted that there's another sort of disruption to our national wellbeing as a result of this use of our energy infrastructure. So you wrote, for example, quote, that between the price of Bitcoin and extreme weather, the combination is a danger to our grid and a danger to externalizing costs on other rate payers and on the environment, end quote.

So can you kind of talk about how we do have national security incentives, perhaps, in making sure, for example, the U.S. leads on AI, but what are the sort of national security threats to using our energy infrastructure in this way?

Mandy DeRoche: I first like to start with making a clarification between Bitcoin mining operations and data centers for cloud computing, for AI, for other things. Those are very different uses for these centers. And yes, you know, there are several administration officials and politicians talking about the national security needs of making sure that data centers and AI are housed in the United States. Obviously I'm not an expert in that.

I think that's a very different proposition with Bitcoin mining. These machines only mine Bitcoin. They don't do anything else. There really isn't that sort of national security benefit that data centers provide, and that people need to pay close attention to there has been reporting that Bitcoin mining operations near military bases may be a national security threat. There has been writings about that again, not the strength of the area that we focus on.

And I also would say, after making that distinction between data centers and crypto mining. Crypto mining's incentives to turn off in high peak demands for electricity are different than other users because it is correlated with the price of Bitcoin. They have different economic incentives to turn on or turn off their machines.

And so in Texas, most specifically, because that's where the most mining is and because it's super hot down there in these summer peaking months, ERCOT, the grid operator down there, has noted that Bitcoin operations operate inconsistently or unexpectedly when called upon for their demand-responsiveness or their interruptibility, that they can't always rely on them acting like other users because their economic incentives are different.

Kevin Frazier: Thank you for clarifying that distinction between the crypto mining and the data centers. I think it'd be helpful also to get a sense of what are some of those similarities with respect to the impacts on the local communities? Can you kind of share more about why a local community might not exactly be thrilled to learn that a data center is being built next door or down the street?

Mandy DeRoche: Yeah, I think Virginia is a good case study where there has been a huge buildout of what I would call more traditional data centers, less so for Bitcoin mining because it's near Washington, D.C. and for other reasons. Now, what we see from folks that live in Virginia is that there is local water problems. You know, these huge data centers, most of them need water for cooling, or for other operations. They're noisy, they're humongous. If they come into your neighborhood, they're clear cutting, you know, land, trees, et cetera.

But mostly we've heard, I think that from an energy perspective, which is what we do here in the Clean Energy Program at Earthjustice, is that there are impacts on rates, on their local rates, on what they're paying for electricity, because there's a, you know, at least right now in Virginia, not enough energy to supply all of the demand and who is bearing the cost of these new loads coming on? Are the new loads paying their fair share? Are residential rate payers being protected? And so we're seeing both the environmental impacts, but we're also seeing the impacts on people's pocketbooks as reasons why people are having sort of buyer's remorse like we see in Virginia now.

Kevin Frazier: Christie, I'm curious to hear more about how state officials are responding to local concerns about data centers, increased prices, for example. No one likes to see their energy bill go up, while also trying to adhere to those state renewable energy goals, for example.

Are we continuing to see state officials kind of walk back on those goals? Or are they coming up with new solutions to realize those goals? What's, kind of, the net outcome as a result of this buildout of more data centers?

Christie Hicks: Yeah, it's a great question because it's something that we are learning more and understanding it better every single day. And part of that is because this is unprecedented. We have not been in a scenario of significant energy load growth for many, many years. The current predictions are that the amount of energy that data centers use could be as much as 9 percent, by the end of the decade, of our total U.S. energy consumption. It's truly staggering, double their current share and just a significant portion of energy load.

And also, I think it's important to understand how different this energy load is than other energy users. So a few different things there. One is that as Mandy was pointing out, particularly with respect to energy used by proof of work crypto mining operations, it's considered really volatile.

So for many years, energy has been used on a pretty consistent curve. You can predict with some degree of certainty, what the highest use is going to be in a particular area. That's what the grid operators actually are paid to do, is to understand how much energy is going to be needed, make sure that it's going to be available on the grid at the times when it's going to be used and because these crypto mining operations can ramp up and down so quickly poses new challenges for grid operators that they just have not previously experienced.

And then you add on top of that the, just, intensity of the energy use from both data centers and crypto mining operations. As we've talked about there, there is not currently enough energy on the grid to serve all of that load. And then, as you pointed out, it is in conflict with state climate goals to reduce emissions.

So, I think the real challenge for states right now, and particularly in this moment, is going to be to lean into clean energy development to match the new load with clean energy, as opposed to relying on either continuing to use fossil fuels or bringing more fossil fuels online, because not only is it going to be cleaner and better for the environment, it's going to be cheaper in the long run.

We're learning so much more about this load. And, for example, you think about artificial intelligence, presumably it is going to get significantly more efficient. It is supposed to make itself get more efficient. So accurately projecting what the energy load is going to look like even five years from now is something of a challenge for utilities, for grid operators, for states.

We can expect it's going to get significantly more efficient and investing in, for example, a new gas-fired power plant that has an assumed useful life of at least 30 years today. It's like investing in a fleet of horse drawn buggies in 1930. The energy use is going to get more efficient. We know that. It's incredibly important that we get energy projections as right as we can right now and to understand that efficiencies are coming and to accelerate the development of clean energy to match that because that is the way of the future and those are the investments that are going to pay off in the long run.

Kevin Frazier: Well, Christie, I may have to bring you on a future podcast about investing tactics because I could use some more gems like avoiding the horse and buggy investment in the 1930s.

I do think though, this prediction challenge is so fascinating. My wife who, full disclosure, works for the development team at Earthjustice would tell you that I am horrible at predicting anything and let alone the sheer complexity of where AI is going to head. You know, are we going to see some sort of killer app where OpenAI actually starts to make a revenue? That would be a novel, crazy outcome that would change the energy demands. But to your point, Christy, we also don't know if AI is just going to solve all of this for us, right? Maybe it's going to find some way where it's able to train in some incredibly efficient manner. Maybe it comes up with some new way of generating energy we haven't even thought of!

But I think what we can expect based off of current trends is that a lot of these labs are going to take their power generation into their own hands. For those listeners who didn't see, you know, we have labs, for example, kickstarting Three Mile Island again to generate energy for their behalf. We see more companies, for example, leaning into small nuclear reactors for their own energy generation.

And so, Mandy, I think it would be great to get a better insight into what sort of clean energy would actually be feasible. What sorts of investments do we need to see to meet some sort of expected huge increase in demand? What sources should we be looking to? Is now the time for wave energy? I'm from Oregon originally. We're all obsessed with wave energy. And then we're like, oh, well, geothermal is the new thing. And we're like, ah, I don't know about that. What should we be looking to? Where is the promise areas?

Mandy DeRoche: Okay, let me start with saying that I think we have a time, sort of, calendar management issue, which is data centers are coming on now and in the near term. And as Christie said before, there are long queues of clean energy waiting to come onto the grid. And we see utilities now proposing to build brand new gas plants to meet this demand.

Now, we don't know what the future holds. We just talked about how we can't predict what's going to happen, in terms of demand over time and what's going to be more efficient over time. But what we could be doing, and we don't want to be doing, is building new gas plants that we don't need in a decade's time. And that's stranded asset risk. That's capital that other ratepayers on the system are going to have to pay for when it's not needed, in time.

Building a gas plant now is not going to match the demand now, right? So there's this time component of data centers coming online, clean energy coming online, and then these big, humongous, dirty gas plants that may come on later and may not be needed years down the road that we're all going to pay for.

What needs to happen, and especially, you know, a lot of these data center companies have very deep pockets and are well resourced, okay? So they've got different, they have the clean energy goals that they themselves have made to their investors and as corporate entities that they need to adhere to, and they have the resources to do that. So that's great.

What we would want to see is clean energy that meets what we call the three pillars. It's new, meaning new clean energy. You're not using clean energy that's already on the grid. You're creating your own new clean energy. Two, that it's now, meaning it's matching what you need at any one time and what the grid in, for example, in those peak demands needs or doesn't need at that time. And that it's near, that you're using energy, new clean energy that is nearby. It isn't in a different utility service area. It's not in another regional transmission area, that it's nearby.

The other ways to talk about this are, one, additionality, so that's the new clean energy. You're creating new, you're not using old. Time matching, meaning everything is matched to when the grid needs it. And then regionality, meaning you're using it nearby. Because what we have seen is carbon offsets or renewable energy credits, that's sort of a lack of accountability. And that there are some companies that use that but really aren't decreasing their operations or their emissions. They're just purchasing offsets and credits that actually don't decrease local air pollution or local greenhouse gas emissions.

Kevin Frazier: Yeah, and I think that's really interesting to get to as well, because we can have these national strategies about creating clean sources of energy for data centers, for cloud computing, for AI, for whatever emerging technology comes next, but as you two have pointed out, these are just such local state issues.

And so, Chrissy, can you walk through how states have kind of taken a different approach to regulations with respect to data centers? Have we seen different models start to emerge? We saw earlier, Mandy walk through Arkansas, perhaps wishing that it hadn't had doubled down on miner’s rights. Are we seeing and miners there— I'll emphasize is M I N E R. In this instance though, are we seeing two different models start to emerge with respect to regulating data centers or how is that playing out at the state level?

Christie Hicks: A few interesting components to your question here. One is that in terms of the energy rates and the way that states are handling this a lot of this is in the jurisdiction of state public utility commissions, or in some places they're called public service commissions.

Some states have some different names. But they are the state agencies that are charged with, in many states, deciding what investments utilities can make, including the power plants, deciding what rate structures, different customer classes are going to be on, and for example, allowing or not allowing different rate structures, for large loads, such as data centers, crypto miners, et cetera.

One of the most troubling parts of this from an economic perspective is places that have offered what we call preferential rates to these exceptionally large loads. In utility regulation, we talk about cost causers and that they should be responsible for, you know, yhe investments that a utility or that the grid has to make in order to serve them. And in many areas that have wanted to initially incentivize data centers and crypto mines to come to their territory, they've offered them millions of dollars worth of incentives to upgrade the grid in order to serve them as well as lower rate per kilowatt hour of energy used than, for example, a residential customer is paying.

So it's really harmful when communities then are paying both for upgrades to the grid that they're not benefiting from and that they are having to pay more for the privilege of having these data centers or crypto mining operations located in their communities. The good news is that a lot of jurisdictions, as we've been talking about, are understanding now that these are not necessarily operations that they want or even can afford to support in their territory.

And when I say afford, I mean that there's not enough energy to go around in order to serve them. So rather than offering significant incentives just to try to bring them in, they're designing in, public utility regulation parlance, we call these tariffs. But they're policies that require these large energy users to have some different rate structures and some different requirements in the way that they're going to be served.

So a few examples of these are having to pay essentially a deposit for some of the infrastructure that's going to get built to serve them. Having to pay up front for a certain amount of energy use because maybe they're going to relocate, some of these are highly mobile loads and if they did that after just three years when they have just upgraded the grid to serve them, again it's going to be families who are left footing the bill for that.

Similarly, especially for data centers, it can take them several years before they actually hit the amount of energy that they have said that they're going to use. So having those customers paying, for example, 90 percent of the energy use from the get-go, that can be one way to, to ensure that again, residential customers, families, aren't going to be basically subsidizing these large customers.

So there's actually a number of states and utilities out there, sort of, trying on different approaches here. And it's sort of the hot topic within utility regulation right now. So I do think that there's an increasing understanding of what some of the best practices are in order to protect customers and what some of the programs are and the policies that can make sure that these customers are paying their fair share and are actually contributing to the grid.

For example, what we call demand response programs, which is where a customer reduces their energy use when the grid is otherwise at a high peak time, so that they can help flatten the load, as we say, requiring these large loads to participate in demand response programs is another example of a best practice that we're seeing out there.

Kevin Frazier: Wow. Well, so now that we've hit on the hottest topics in utility regulation, I don't know if we can improve upon that. So, unfortunately I'm going to have to let you all head back to being the Earth's lawyers. I'm sure we'll have a lot to talk about down the road with respect to these data centers and maybe even crypto mining.

We'll see. Crazy crypto markets ahead, crazy markets ahead. So the future will be interesting and I look forward to touching base down the road. But for now, I'll see you all later.

Christie Hicks: Thanks so much.

Kevin Frazier: The Lawfare Podcast is produced in cooperation with the Brookings Institution. You can get ad-free versions of this and other Lawfare podcasts by becoming a Lawfare material supporter through our website, lawfaremedia.org/support. You'll also get access to special events and other content available only to our supporters.

Please rate and review us wherever you get your podcasts. Look out for our other podcasts including Rational Security, Chatter, Allies, and the Aftermath, our latest Lawfare Presents podcast series on the government's response to January 6th. Check out our written work at lawfaremedia.org. The podcast is edited by Jen Patja, and your audio engineer this episode was Max Johnston of Goat Rodeo. Our theme song is from Alibi Music. As always, thank you for listening.