How Surging Demand Is Shaping the US Power Sector, with Brad Harris

Notable quotes

• Growing demand for electricity due to three main factors: “Load growth is really being driven by three things: Electrification (think about electric vehicles using the grid, not from oil and gas). It’s being driven by an increased investment in manufacturing following the CHIPS Act and the Inflation Reduction Act; there’s huge investments in manufacturing in the United States. It’s also being driven by data centers … [which are] a very big source of load growth in a lot of jurisdictions.” (4:11)
• Balancing climate targets with increased energy demand: “A lot of these companies care about what they’re doing, and they have climate commitments. When you have big companies that are sophisticated and care about this, you have a lot of potential revenue … We have a lot of new technologies, a lot of solar and wind. If we can build it, that can help meet this demand.” (13:53)
• Increasing the efficiency of data centers insufficiently addresses challenges of load growth: “These data centers have become a lot more [energy] efficient in recent years, and that has been a big part of the reason why they are only 2.5 percent of electricity usage [in the United States], even though the amount of data we’re storing in all these systems has grown exponentially … It’s important to mention, though, that the more efficient they get, the cheaper they are to run, and the more we might use these types of technologies. If you increase the usage, obviously that still maintains that load growth. Efficiency is something that can help us offset this load growth.” (19:22)

Top of the Stack

• Harlan County, USA documentary
• Enron: The Smartest Guys in the Room documentary

The Full Transcript

Daniel Raimi: Hello, and welcome to Resources Radio, a weekly podcast from Resources for the Future. I'm your host, Daniel Raimi. Today, we talk with Brad Harris, the new director of government affairs here at Resources for the Future. Brad has years of experience working in the electricity sector and brings a wealth of knowledge, particularly about utilities, to our conversation.

In today's talk, we'll identify why demand for electricity is increasing dramatically in the United States, how utilities are planning to address that growth and demand, what it means for ratepayers and carbon dioxide emissions, and how policymakers can address the challenges as they grow in the coming years. Stay with us.

Brad Harris, our new director of government affairs here at Resources for the Future (RFF), welcome to RFF, and welcome to Resources Radio.

Brad Harris: Thank you. Excited to be here.

Daniel Raimi: We are excited to have you. Your background is a little bit different than a lot of the folks we interview on the show. You don't have as much of a research background, but I'm curious how you got interested in working on things related to energy and the environment?

Brad Harris: I think I've always been really interested in public policy and business. My dad did international business all over the world. He was always traveling, talking about what he was doing. My family's always very focused on public policy and government. I don't think it's a coincidence that my sister majored in international relations, my brother majored in government, and I majored in political science. I think part of that is my parents grew up in apartheid South Africa and were very involved in the anti-apartheid movement, and that permeated through the family. I was always interested in where I could find a way to mix these two interests.

I did a bunch of different things, and then, in grad school, I thought I had found it. I was going to do public-private partnerships to redesign downtown areas in North Carolina. I was all excited, and I was going to do my master of business administration at the University of North Carolina, and I went to a real estate club meeting at the university, and they were talking about building the same condo building in every major city in the United States. I was like, "Maybe this isn't for me." But I was always interested in reading the news about energy stuff. I went to the energy club meeting and they were talking about utilities and climate change and how utilities serve the public interest and I thought, "This is it. I'm home. This is what I want to do." That's how I got involved in the sector.

Daniel Raimi: That's great. You ended up working at Duke Energy for a while, right?

Brad Harris: I wrote my master's thesis on rate design in North Carolina. I did that, because I took a course taught by an executive at Duke Energy, and I was listening to what he was interested in in class, and he was really excited about rate design, which is basically how utilities price electricity. I said, "I'm going to write my thesis on that."

I did that. I got a job on the rate design team at Duke Energy and helped lead a bunch of initiatives there, including how we analyze. I led the analytics for what was called the “Comprehensive Rate Design Study,” where Duke looked at our pricing for all of our customers in North Carolina and South Carolina. I was in charge of the analytics of that—determining how much revenue and how much cost different customers would bring and our pricing policies around them. That forms the basis of what we're going to talk about today.

Daniel Raimi: Absolutely. Today, we're going to talk about this issue that a lot of people have been talking about lately, which is load growth; in other words, increasing electricity demand. This load growth is happening at the same time that we need to decarbonize the grid to get to net-zero emissions and address the climate challenge. We're going to talk about all that, but first, can you just give us the fundamentals of what's driving the increase in electricity demand currently and what we anticipate driving that demand in the future?

Brad Harris: This is a really interesting phenomenon. Load growth has been pretty subdued for at least the last decade. There's been a lot of advances in energy efficiency, particularly light-emitting diodes, or LEDs. The introduction of them has been great for efficiency. But now, load growth is really being driven by three things: Electrification (think about electric vehicles using the grid, not from oil and gas). It's being driven by an increased investment in manufacturing following the CHIPS Act and the Inflation Reduction Act; there's huge investments in manufacturing in the United States. It's also being driven by data centers, and it's this last one that I think is particularly interesting, because the other two are really a shift in how we use energy. We're using electricity instead of other sources. Data centers are really a new source of energy and have some very unique characteristics, and this is a very big source of load growth in a lot of jurisdictions. I’m looking forward to talking more about it.

Daniel Raimi: Can you put a little more meat on those bones? Let's imagine a hypothetical data center out there. Maybe it's powering some cloud services. Maybe it's being used for artificial intelligence (AI). Can you give us a sense of how energy-intensive some of those facilities can be?

Brad Harris: I should probably give a little bit more background, as well. Data centers can support things like the internet and all this information technology infrastructure. Right now, according to the Boston Consulting Group, it's about 2.5 percent of US electricity demand. We're seeing big growth right now, primarily driven by generative AI and cryptocurrency mining. The example I'll use to answer your question—I actually pulled the capacity of a Bitcoin-mining operation from a public website out of Georgia. It's 195 megawatts of demand. That's a lot of demand. Let me explain and put that in context a little bit.

It's important to note that these data centers have a very high load factor. Basically, they run very consistently. Let's say it has a 90 percent load factor. A 100 percent load factor would mean that you use the same amount of power all the time. The lower the load factor, the more peaky, as we say, the usage is. 195 megawatts at 90 percent load factor uses about as much electricity in the year as 140,000 households. That's about the number of households in the city of Pittsburgh.

Daniel Raimi: So, a Pittsburgh-sized Bitcoin-mining operation.

Brad Harris: Pittsburgh-sized, for all the houses. Another way to conceptualize that is Commonwealth Edison, according to US Energy Information Administration data, has the largest energy-efficiency program in the country, and in 2022, this Bitcoin-mining operation would've used 78 percent of that entire program. That's the largest in the country.

One final example: If you wanted to power that one operation with solar energy, it would require about 700 megawatts of solar, which would mean covering New York City's Central Park at least four times over—probably five or six times over. That's a lot of energy. By the way, if you want it to run 24/7, you'd need more than 700 megawatts of battery storage, and if you want it to be carbon neutral, you have to account for all the carbon to build all that solar and batteries. It's a lot of energy is the answer to your question.

Daniel Raimi: It is. My understanding is that some of these facilities, especially the generative AI facilities that are at least being proposed and being talked about and also some of the chips manufacturing facilities, like the chip fabrication facilities, use even quite a bit more than that hypothetical Bitcoin-mining operation you were talking about.

Brad Harris: I wanted to pick an example that I could pull off a website, but I've heard anecdotally that it could be much more. When you look at some of the utility integrated resource plans, they don't always pull out all these different things. But suffice to say, we're seeing big growths in load demand that are changing integrated resource plans. We're seeing some revisions having to be made, because the difference in forecast from year to year is big enough to change the overall plans. This is a big deal.

Daniel Raimi: For people who don't know: What is an integrated resource plan?

Brad Harris: Oh yes, thank you. It's basically the plan of what generation sources the utility is going to need to build or have to maintain firm, reliable power.

Daniel Raimi: Great. All right. There's a lot of new demand coming onto the system, and utilities are going to need to meet that demand or at least do their very best to meet that demand in ways that ensure that the grid is reliable and as affordable as possible, as well as try to clean up the grid at the same time. Trying to achieve multiple objectives at once—how are utilities starting to respond to this?

Brad Harris: They're thinking very hard about it is the short answer. Because these customers don't just want a lot of power, a lot of these companies want clean power. Even if they're in states that maybe don't have clean energy commitments, the companies have clean energy commitments. I think a lot of utilities are trying to acquire as much clean energy as they can, as much solar and wind as they can.

Now, as I believe you've discussed before in this podcast, there are constraints to that, and I want to highlight some of the work RFF is doing to look at those constraints. But in short, there may be some constraints. You can't just add more and more solar and wind to compensate. You're going to have to look a little bit creatively. I think some utilities are going to look at adding more gas plants. They're going to try to add peaking gas plants—gas plants that only use power a little bit during the year, rather than plants they have to run all the time. But this is a challenge.

I think also they're going to look at expanding energy-efficiency programs and what are called “demand-side management” programs. Now, I can talk a little bit more about what those mean, but this is a big challenge, and I think utilities are trying to use all the resources that they can to meet this demand.

Daniel Raimi: Why don't you say a little bit more about demand-side management strategies and to what extent some of these new centers of load can participate in those programs?

Brad Harris: It's probably helpful to think a little bit about the history of high–load factor customers like this. Again, these are customers that use power very consistently over time, and they’re large customers. Traditionally, utilities and regulators love these types of customers, because the theory goes that it basically costs a lot to build a power plant, but running it is a lot cheaper. A customer that uses load consistently over time is going to be overall cheaper to serve, and it's going to put downward pressure on prices for all customers, because overall, you’re running your whole system at a much higher utilization rate, if that makes sense.

I'm a big fan of history. You can actually go back and look at utility leaders back in 1910 discussing how they needed to give better pricing for high load factor customers like this, because it'll help everyone. There are actually economic development tariffs that encourage exactly these types of customers. There's a lot of good to be said under normal circumstances.

The question is, Are we in normal circumstances right now? That all assumes that you can build a power plant tomorrow that is about the same cost as a power plant you built 10 years ago. You can build that power plant, and that's before accounting for environmental factors or climate policy. I think that's really important to understand the basis. These should be customers that utilities traditionally want and are set up to attract. There may be some utilities out there that are surprised at how much they've gotten and maybe some regulators that might think, "Is this too much of a good thing?" That'll be very interesting to see how this plays out. But traditionally, it's a very good thing for these utilities. What I'm driving here to answer your question is demand-side management programs basically say, “If you reduce your usage during certain periods of time, then we'll give you a credit for that.”

Think, like, the utility is going to have a peak on a really cold winter morning. The data centers could ramp down their usage. That's basically creating capacity in the system. You don't have to build a power plant. That is wonderful economically. That helps this theory that I was just talking about really work great, because you're not having to build another power plant to serve these customers, and you're using your system more efficiently. The challenge, though, is that it helps with capacity and cost but doesn't necessarily help with climate policy, because the carbon emissions are in the energy, the kilowatt-hours, rather than the capacity of the kilowatts.

Daniel Raimi: Interesting. Can you say a little bit more about the climate implications of all this? I know one of the big concerns from environmental advocates and some analysts that I've seen is that this dramatic expansion of load and the natural gas plants that some utilities are proposing to meet that expansion might just blow climate targets out of the water. What do you make of that concern?

Brad Harris: I think it's a legitimate concern. To strike a more optimistic tone, a lot of these companies care about what they're doing, and they have climate commitments. When you have big companies that are sophisticated and care about this, you have a lot of potential revenue. Because there's a lot of excitement and a lot of potential revenue here, there's a lot you can do with that. We have a lot of new technologies, a lot of solar and wind. If we can build it, that can help meet this demand.

But I think there's a real need for regulators, people at utilities, and policymakers in general to think about understanding this challenge and coming up with creative solutions to solve it, because this is a lot of load, and if we don't think creatively about offsetting it in a real way, it's a big problem. Just as one example, Duke Energy, where I used to work, had to revise its integrated resource plan for how it's going to build its capacity moving forward because of an increase in load growth.

There's a carbon plan in North Carolina, and I don't want to talk about too much specifics, but I don't think anyone thought that the carbon plan was easier with all this load growth. Even though, as I've mentioned, there are a lot of benefits to load growth, particularly of these customers, but it's challenging. It's also my understanding you can build a data center a lot quicker than an electric vehicle manufacturer, because when you build electric vehicles, you have to think about the whole supply chain and everything. When you're building a Bitcoin-mining or a generative AI system, you can ramp that up much quicker, because it's more modular. This is a really important part of the national story.

I think this highlights the importance of policy, because it's actually quite hard to verify that any load you've added is carbon neutral. I think we've seen this a lot with 45V, the hydrogen tax credits, where basically the US Treasury Department wants to make sure that any tax credits go toward "clean hydrogen." Our colleague Aaron Bergman has done quite a lot of work on this. I'll let him speak to that, but just note that that's quite a hard thing to actually prove out.

Daniel Raimi: Definitely a topic for another podcast or three in the future. We'll see.

Let's turn to policy now explicitly. What are the policy measures that you're seeing decisionmakers consider either at the national level or in certain states?

Brad Harris: One thing that I'll mention—I talked about this theory about load growth driving down rates for everyone, and a really important assumption is that you get the load growth right, because basically, the assumption there is you're going to have revenue to offset those costs and actually more revenue than the costs to help drive down the price.

Daniel Raimi: When you say "get the load growth right," do you mean the utilities are correctly anticipating the amount of load growth?

Brad Harris: Exactly. If you expect a lot of load growth, and then it doesn't happen, then you've built a lot of power plants you're not going to use, and the costs to build and maintain those power plants still need to be maintained. There's actually a cost to planning for this load growth.

One thing that I would think about if I were a policymaker is, Is this load growth going to actually come online? Can we be careful, when you're a utility, in how you sign contracts and just be mindful that, if this is coming online, this is firm demand, and we need to meet that. That's part of what utilities are responsible for. But if it's not actually going to come online, how do we protect ratepayers? That's one thing I think about very carefully.

Additionally, we discussed the need for these companies wanting to be carbon neutral, and it's hard to do. Helping make sure that they have pathways to do that and giving that seal of approval that this is a good pathway, or this pathway may not be as good—I think that's a really important thing for utilities and policymakers to think about, because again, I think there's a great desire here.

Maybe, I should have mentioned at the beginning of this podcast, I'm only talking about the load growth side of this. What this energy is being used for I'm not commenting on here on purpose, but I'll note that you could have many guests on this podcast that can talk about the great things generative AI could do for the grid and the energy transition. I want to mention that, as well.

Daniel Raimi: Absolutely. A lot of promise in these technologies, especially generative AI. People will probably have more divergent views about Bitcoin mining and things like that, but again, that’s a topic for another conversation.

One thing that we haven't talked about a little bit, although you have brought it up a couple of times, is the issue of energy efficiency and demand-side management. A lot of times, when people talk about the energy system, there's a tendency to focus mostly on the supply of energy and the emissions associated with that supply. We sometimes forget about the other half of the equation on the chalkboard, which is demand.

Can you say a little bit more about the potential for some of these facilities to be more energy efficient, whether it takes the form of the demand-side management programs that you mentioned earlier, or if there are technological innovations that might allow these facilities to just simply use less energy while they're doing their work?

Brad Harris: I think it's important to mention that it's my understanding that these data centers have become a lot more efficient in recent years, and that has been a big part of the reason why they are only 2.5 percent of electricity usage, even though the amount of data we're storing in all these systems has grown exponentially. There's obviously potential for that to continue, and the more data centers we have, the more financial incentive and smart people looking into how to make them more efficient. It's important to mention, though, that the more efficient they get, the cheaper they are to run, the more we might use these types of technologies. If you increase the usage, obviously that still maintains that load growth. Efficiency is something that can help us offset this load growth. I don't think it's sufficient by itself.

As I mentioned before, the demand-side management, often called “demand response,” where they ramp down the amount of usage when the utility needs it, is something that data centers are uniquely good at, particularly cryptocurrency mining. A lot of advocates of cryptocurrency mining advertise this. They can ramp down really quickly in a way that a big factory or something can't. That's great from a ratepayer point of view, in terms of driving down the price of electricity, because you don't have to build more power plants. From an environmental or the cost-of-complying-with-carbon-reduction-policies point of view, it doesn't help you very much, because you only use that in a limited number of hours, and they're running full bore at all the other hours.

Now, expanding your question a little bit more globally, energy-efficiency programs, in general, are wonderful. I worked on the energy-efficiency team at Duke Energy. We could do a whole other podcast on all the things we can do to expand energy efficiency. But I'll remind you that Commonwealth Edison has the largest utility energy-efficiency program in the country, and that one data center, just using some basic math, is 78 percent of that whole program. This is a big challenge that I think energy efficiency by itself can help but not fully solve.

Daniel Raimi: Absolutely. A helpful tool, not a silver bullet.

One other question, just jumping back a moment. You mentioned earlier the planning that utilities need to go through and that system operators need to go through to ensure reliable and affordable supplies. I was just gaming out, if I am OpenAI, for example, or Microsoft, or some other large company that is building lots of data centers and is going to need a lot of computing power—do those companies work directly with utilities to plan their siting of facilities two, three, four, or five years into the future? Is that something that utilities actually can know with a large degree of precision?

Brad Harris: First, let me just say, because I was in some confidential conversations, that I'm not going to speak about any one company in particular. Let me just say that up front.

But yes, there are large account-management teams and economic-development teams that talk directly between utilities and companies, and there's a lot of communication there. That doesn't mean that there can't be miscommunication, as I mentioned. The planning process here is a really complex process, and sometimes everyone thinks they're going to have a big load center, and then it doesn't materialize. That can certainly happen, but there's a lot of coordination that goes on.

Daniel Raimi: That's interesting. I never knew that.

Well, Brad, this has been a really interesting conversation. I've learned a lot. I'm sure our listeners have, too. I'd love to ask you, now, the same question that we ask all of our guests, which is to recommend something that you think is really great. It can be related to the topic that we've been talking about today, or it could be something else entirely. So, what's at the top of your literal or your metaphorical reading stack?

Brad Harris: I've been looking forward to this question, because I knew it was coming. By any chance, have you seen Harlan County, USA?

Daniel Raimi: Gosh, have I seen it? I've definitely heard of it. It's a movie, right?

Brad Harris: It is a documentary.

Daniel Raimi: Is it about the coal-mining strikes?

Brad Harris: Yes.

Daniel Raimi: And the police and the union-busting response? I have seen it, but it was a long time ago.

Brad Harris: It's a great portrayal of a community with energy at the heart of not just its economy, but really at the heart of the community, and the strike that happens there. It won the best picture for documentaries. It's a really good documentary. It's been a few years since I've seen it, but it's one thing that comes top of mind.

Daniel Raimi: Yeah, I'll second that. I remember it being really powerful and disturbing in some ways, too.

Brad Harris: Because I really like movies, the other one that came to mind was Enron: The Smartest Guys in the Room. I haven't spoken to anyone who's seen that and hasn't said, "It's a wild ride and a great movie." It’s another documentary this time about Enron and some of the scandals there. I say some of the scandals—a lot of the scandals. It's just another great piece of media.

Daniel Raimi: That sounds like fun. Great, we've got our movie-watching experience all lined up for the next couple nights. One more time, Brad Harris here at RFF. Thanks so much for coming on the show. Welcome again to RFF. It's been a really great conversation.

Brad Harris: Thank you. I've enjoyed it.

Daniel Raimi: You've been listening to Resources Radio, a podcast from Resources for the Future, or RFF. If you have a minute, we'd really appreciate you leaving us a rating or a comment on your podcast platform of choice. Also, feel free to send us your suggestions for future episodes.

This podcast is made possible with the generous financial support of our listeners. You can help us continue producing these kinds of discussions on the topics that you care about by making a donation to Resources for the Future online at rff.org/donate.

RFF is an independent, nonprofit research institution in Washington, DC. Our mission is to improve environmental, energy, and natural resource decisions through impartial economic research and policy engagement. The views expressed on this podcast are solely those of the podcast guests and may differ from those of RFF experts, its officers, or its directors. RFF does not take positions on specific legislative proposals.

Resources Radio is produced by Elizabeth Wason with music by me, Daniel Raimi. Join us next week for another episode.