Not Your Father’s Data Center - David Porter: From Grid Load to Grid Asset: How Data Centers Can Support the Grid

Episode Date: July 14, 2026

In this episode, Raymond Hawkins, Chief Revenue Officer at Compass Datacenters, is joined by David Porter, Vice President of Electrification and Sustainable Energy Strategy at EPRI—formerly known as the Electric Power Research Institute. With a career spanning mechanical engineering at Duke Energy, stints across the US energy landscape, and over 17 years at EPRI, David Porter brings deep industry insight into the intersection of energy and data infrastructure.Their discussion explores the rapid evolution of data center power demand, shifting from kilowatt to gigawatt requirements within a decade. The conversation unpacks the unique challenges this poses for power utilities, such as mismatched development timelines, grid transmission constraints, and the need for adaptable grid-responsive infrastructure. Key topics include EPRI’s nonprofit, science-based research mission; the rise of grid-interactive data centers; the DC Flex initiative for flexible load management; and the critical impact of load growth on grid affordability and efficiency. The episode offers a detailed look at how collaboration between data centers and utilities can enhance grid resilience and lower costs for everyone.Timestamped Overview00:00 Intro03:25 EPRI's public benefit mission09:50 Data centers and electric transport13:02 Energy demand forecasting for developers15:00 Challenges with data center power constraints17:49 Transmission vs. distribution systems21:45 Improving energy transmission efficiency23:35 Bridging data centers and utilities28:06 Evolution of the Data Center Industry31:42 Explaining the Flex Mosaic framework35:30 Planning for grid capacity challenges38:31 Importance of Frequency in Power Grids41:05 Energy demand and frequency issues45:56 Golfing chat

are statistics looking just at the U.S. that where states have load growth, average electricity prices are going down. States that don't have load growth, the average electricity prices are going up. We have historical data that shows that. So, David, mine and your podcast may get the most serious editing ever. This podcast may get just reduced to those 20 seconds. All right. Welcome to another edition of Not Your Father's Data Center.

I am Raymond Hawkins, and today we are super excited to have David Porter, who is the vice president of electrification and sustainable energy strategy with EPRI. David, thank you for hanging out with us for a little while. Thanks for having me, Raymond. David, we want to talk about level set everybody on what EPRI is. I know most of our audience is data center related people, and so they've heard the acronym, they've heard the name EPRI, but they don't know much about it. But before we get into Epry, do you mind taking a few minutes and telling us about you?

How'd you end up in the power business and where'd you come from? Let's get oriented on you a little bit. I wound up in the power business largely because I have an engineering degree, mechanical engineering degree, from North Carolina State University. And my first job out of college was with Duke Energy. Then it was Duke Power, but I started with Duke. It was a great beginning to my career. And I have essentially stayed in the energy business.

for the rest of my career. I've been with Eprey now for, it'll be 17 years next month. All right. And so it's been a great ride. Is North Carolina still home? Are you? It is home.

Yeah. It is home. When I came to Eprey 17 years ago, I was on the West Coast. So we came back here, but I've lived on the West Coast twice, fortunate enough to get moved to Florida at some point, which is where I met my wife.

And I did spend time in Dallas. for a little over a year. All right. In one job, too. So I moved around a lot. We're settled now. We're not moving again. I've said that before.

Hopefully that will stick this time. Knock on wood. I'm not going on for you. You made the transition from Duke to Epry 17 years ago? No, I went from Duke to my first stint with Epry. This is my second time with Epry. Okay. The first time I was with Epry.

Epre was just starting to get into working with international utilities. I came on board to work with those international utilities. And when I left there, it was because Florida Power Corporation recruited me to come to Florida and be part of their team as deregulation was starting to hit the industry. And then when I left there, I've been in a few other related energy businesses, both in energy procurement, energy management, and a lot of other energy-related services and then came back to EPRY 17 years ago. So most of our folks that are going to listen to us, they're data center people,

so they're people that are either building or developing or procuring or supporting the building of data center. So it's pretty much a data center crowd. They all understand that power is a massive challenge for us these days. I know most of them have heard of EPRI, but I think very few of us understand EPRI. Can you give us a download on what EPRI is and what EPRE isn't? Yes, EPRI is now in our 54th year of existence. Founded in the early 70s. We're a not-for-profit research and development and deployment company. All our work is science-based. One of the unique things about EPRI, though, is that our charter calls for a public benefit mission, which is a key part of what we do. So work that we do is not intended just to benefit.

energy companies, it's intended to benefit society as a whole. And whether that's an individual customer connected to the grid or something that has to do with sustainability or clean energy that everyone benefits from, all of that plays into the work that we choose to do and the strategic direction that we choose to follow. So help me with the acronym because I know I will mess it up. Originally, and we only go by EPRI now, Raymond, but it was Electric Power Research Institute is what it stood for. And it originally was dedicated just to electric utilities, but that's expanded over the years now. And we operate in 40 countries around the world today, everything from nuclear energy to the end use of energy and all that's in between in the value chain of electricity. but we also do work in natural gas and alternative fuels, advanced nuclear, as well as all the other

associated pieces around the electric utility industry. So you guys don't, you're not really a lobbying arm, but you are a nonprofit. So you are, is an advocacy, not advocacy? What's the right way to think about the work? We do no advocacy at all. And we are not a lobbying organization at all. Everything we do is science-based, factual results that we produce. You know, it's one of the things we do.

We're not biased towards any technology or approach. We look at everything agnosticly, and we report out what we find is the best answer to a problem or solution. Gotcha. Well, problems or solutions are what we're here to talk about today. I'm going to guess you've heard about the data center business. and you guys might have thought about us a little recently. Just a tan. Just a little, yeah, yeah.

I will say, so I'm in year 13 in the data center business, which is not a terribly long career, but compared to the age of the data center industry, about half of the industry's life, it really became a commercial enterprise in the early 2000s. And I've watched the industry go from, it was very normal to do leases with customers that measured in the kilowatts to go from kilowatt leases in my early days just 13 years ago

to now we almost never talk in anything sub hundred megawatts and it is routine to talking gigawatts that I think I believe has caused an incredible challenge to the industry that you guys support with research and and information so from your from from from EPRI's view can we just start with, wow, it's a lot and start there. How do we get all of this power? How do we get, and not just the generation, but how do we get the distribution? How do we get the transmission?

Where in EPRI's view of the world is this challenge of, oh, my gosh, they called and asked for how many gigawatts? That's all literally the last four or five years. I mean, nobody said the word gigawatt four years ago. That's a fair point. When people were talking gigawatts four years ago, they were talking about a new nuclear generating unit. That's right.

That's right. Not a single load source that could consume that. Right. Not a single customer. That's right. Not a single customer. And we're really Raymond in an interesting time for the industry because the industry over the last two to three decades, load growth for electricity has been very flat.

In some cases, declining depending on where you were located. And that was a combination of a combination of a. number of things. Some of it was the offshoring of industry. Some of it is a continuation of energy efficiency programs that, you know, every new HVAC system that comes off the line is more efficient than the last. So you take an old one off and you put a new one on. It uses less energy. All those things together kind of flattened out load growth for the utilities. So hold on, David. I want to jump on that one. When you say the load growth had gotten flat, What time frame are we talking about?

You talking about the 2010s? You talking about when did we start to see that trend of flattening? Because as a consumer, I think, oh, there's more people, there's more electrification. It's only going up. I obviously have a very limited view. So when's that flattening started? That flattening started pretty steadily back in the 90s. Okay.

All right. So it's been a while. Now, if you look at different components of load growth, And we've charted out over the years data centers. And you can clearly see changes in data centers from, you know, where we started with, you know, enterprise and some of the things are going on. And yeah, you can see in the growth when the social media trend started. But then you also see a flattening out there. And now you're seeing a steady upturn with data centers because of the growth of AI.

But that's not the only growth, as you well know. for data centers. Yeah. The other types of data centers that have been around longer than the AI data centers or AI factories are still growing too. All those businesses are growing. Here, here, yeah. All right, so I'm going to orient everybody.

So even as late as the 90s, so literally someone's entire working career, 95 to 2026, three decades, they've seen a relatively flat, if I'm, if I'm Duke Energy, at that point in time, Duke power, my global view, my third. footprint view has been largely flat for decades. And then what? Then we reach a combination of factors, right? So a couple things are going on. And I'll compare two things that are very different. One is the large point loads that are represented by data centers where you get this very large customer that's locating in one spot and they need a lot of power and they need it in a hurry. And that's one form of growth. Another that is a challenge for the industry that

looked to be a bigger challenge before the data center boom came along was electric transportation. The growth of electric transportation has been enormous around the world and that's going to continue. Electric transportation, largely electric cars, right, but not exclusively. So all of these batteries needing charged to move people around buses, scooters, but mostly electric cars. Is that when you say transportation? You're not talking about moving electrons from a, from a distribution or transmission perspective. You're talking about moving humans with electricity batteries, right? Yes, absolutely. So I'm going to see. Humans or freight or one.

whatever. Or stuff. Yeah, right. Yeah, whether it's an individual personal vehicle all the way up to a class eight over the road semi. Freight liner. All of that will be electrified. And it's very different.

You hit the point right on the head. It's a moving source that needs to be charged. But that also makes it potentially a valuable resource to the grid between charging and discharging and how you match up with what the grid has going. on charging vehicles during the day when there's an abundance of solar energy on the grid in a lot of places around the world, that is just a perfect match for the cheap solar energy coming on to the grid to charge those batteries for vehicles. All right.

So first, I sidetracked you, but you said electric transportation. I just want to clarify electric vehicles, truck, scooters, all of that stuff. That's the first blip. Sorry, I took you off your train and thought. Go ahead. No, no, that was the first blip. But the two things that utilities are looking at much more closely now are data center growth and reshoring of industry.

You know, giving an example where there's a tie there, there are a lot of new battery factories that have been built in the U.S. to be able to support not only the electric transportation business, but also the energy storage business, using batteries as a grid resource. So that's growth. But the data centers are much more unique. because they are a very large point load, and they can build and be operational in a very short period of time. And when you're thinking about the electric utility industry, most of the investments the utility industry makes are 30-year assets.

So they're planning for the long haul. It's a slower cycle than what the tech companies operate on. And so, you know, that makes it challenging for both parties to get to the same place quickly. Can we talk about that phenomenon for one second, David? So I've tried to relate because people say to me, hey, Raymond, just tell the energy company to produce more stuff. It'll be fine. They do this. And I'm like, guys, understand the history.

And this is a question because it's the way I try to think about it. I want to know if this is right. If a developer comes along who develops houses and he says, I'm going to buy all that land, I'm going to zone all that land, I'm going to clear all that land. and I'm going to start building houses. And when I'm done, that neighborhood's going to need two megawatts worth of power. He's got a 10-year project. From the time he puts the first shovel in the ground, the time he builds the last house, he's going to take 10 years to get to two megawatts.

And by the way, two megawatts, that's a big neighborhood. I mean, that's a giant. It is. But he's got a decade of development and a decade of coordination with an industry that's got a three-decade-packed to power generation. Now the data center guys show up and go, hey, I don't need two megawatts. I need 500 megawatts. And oh, by the way, I need them in 24 months.

Because they're not building, you know, 4,000 houses. They're building a dozen buildings or 10, you know, and they're building them in one. You use the point, I'm the phrase you said, a point customer. Hey, I got one place. And I'm going to be there in two and a half for three years. I want it all done. And you try to mirror that two and a half for three years up to an industry that thinks in decades. It's a challenge.

It is. And that's what I think we are, we, the data center business and you, the power industry, are trying to figure out how to weave us back together. Absolutely. And that's what really led us to the initiative that we started 18 months ago called DC Flex, which compass is part of, one of our first members in that program. And we've made significant progress in the last 18 months. So can we talk a little bit about the whole idea of grid responsive infrastructure,

or how that works. You and I are both familiar with DCFlex, but there's a lot there. Can you talk through the components of DCFlex? Traditionally, as you know, Raymond, data centers have been a pure off-taker of energy from the grid, which for many years was just fine until you start to see these low growth projections that are out there today to meet what the needs are from the data center industry.

And we get into that speed to power discussion, which is so important to the developers. and the hypers. The challenge is not just about putting electrons onto the system. In many places, there is more than adequate generation to handle the data center loads that are looking at specific areas to build. The bigger constraint in the U.S., particularly today, and this is really a global issue too. It's transmission capacity that is the constraint. And just to give you an idea, part of that challenge is because in this country, from design to permitting to building a transmission line, it's typically a seven to 10 year process. That does not match up well with

24 months to build a data center. Well, you walk people that will listen to us through those three Because you and I will talk about transmission. I know there are people don't fully grasp what we're saying there. Will you walk through that real quick? So the first piece is the energy supply, and that's the generation. Where are the electrons coming from? Is it nuclear, power plant, natural gas, wind, solar, hydro, any of those sources produce electrons that go on to the grid? Those electrons first go to the transmission system, which is the bulk carrier of electric energy all around the world.

It's very similar to a gas transmission pipeline. It's like the big pipe that you're moving a huge volume through at a very high voltage. Big wire, high voltage, really reliable, but lots of energy on that wire. Very much so. That's very accurate, yes. Okay. Just because I'm not a mechanical engineer or electrical engineer. Do we do it at high voltage to make it go further?

We do it at high voltage because it's more efficient. Because the higher the voltage, the lower the current for the total energy. that you're transmitting. Okay. And the biggest strain on energy losses is the higher the current is being pushed through that wire, the more losses there are. So it's much more efficient. Yeah, yeah.

Okay. So I get to run more efficiently. Okay. So I'm dumping a lot of energy. I'm running at low current, but high voltage. And I'm moving at long distances. Long distances. Very long distances.

Also, the transmission system is used to directly serve large loads. such as data centers. Right. So that is another aspect of the transmission system, and that's a big part of where data centers come into play. When you start talking about 100 megawatts and larger, that is not a load to get served at the distribution level. So that third piece of the puzzle that you're asking about,

when you get to more local, like you mentioned, the neighborhood, that two-megawatt neighborhood is served off the distribution system. Your local mall is served off the distribution system. A lot of smaller industrial customers are served off the distribution system. So those are lower voltage and shorter distance deliveries. The other thing that's very different between the two, the transmission system is a looped system. So that if there is a fault somewhere, you have the ability to quickly continue to move bulk power. so there are not expansive outages.

The distribution system has always been designed as a radial system, from the substation out to the loads. It operates very differently than the transmission system. Can you give me numbers, and I know they'll matter to technical people, but I would like to hear them. Can you give me numbers? Transmission is how big and distribution is how big. In terms of voltage, transmission typically runs between

100,000 volts all the way up in some cases to 765,000 volts. Okay. A typical distribution voltage is typically 35 KV, 35,000 volts, and down to, in some places, older systems still have 4,000 volts, but most of it's 12, 24, or 35 KV at the distribution. So as I think about this, if I'm looking at a picture, I got a, I got generation, that generation is pumping that gazillion volts up onto a wire. That wire is running a long distance. Then it's getting carved off to a substation. And that substation steps it down and puts it into the distribution system. Are there substations on distribution beyond that one off the big wire?

Oh, yes. Okay. The distribution substations get more centrally located around the load centers that they're going to serve. Okay. So it's not just, it's interesting because you started with, Hey, Raymond, the real problem is. building distribution. Transmission. Excuse me. Transmission in a timeline because it's a decade-long project. Now that we've laid out the three pieces, let's go back to there. Why does it take 10 years to build transmission lines? So transmission lines fall under the jurisdiction of the Federal Energy Regulatory Commission, FERC. Okay. So there's federal permitting that has to take place to

approve the need for the transmission line. And once that is received, then the utilities can go about securing the right-of-way to be able to build the lines where they want to go. So if you can imagine doing these things cross-country type of building and having to buy right away from lots of different landowners, that can be a very slow and tedious process, particularly when these days the old NIMBY principle, you know, not in my backyard applies to about everything. So it's very challenging. I can tell you that some of the work that we're doing under one of my colleagues in our transmission and distribution infrastructure area is called grid enhancing technologies. And this is work that we're doing to be able to push more energy through existing transmission corridors and using things like advanced conductors that can carry more energy in the same space that an existing conductor.

So we just changed the material out? It's a different material. It's a better material that's more capable of carrying greater amounts of energy. Along with that, we're doing a lot of work around dynamic line ratings where you truly do, in this case, use artificial intelligence to understand the weather parameters, wind, rain, temperature, all of these things. But the rating and the loading on those cables can change depending on. the ambient air conditions so that at times you may be able to move more or less energy down a particular corridor. And then there's also some work around power line flow controllers, which help move power around different avenues through the transmission system where one is a

better opportunity or more efficient opportunity than others. So there's a lot of work there that can help get more out of what we have to do today as opposed to just the need for new transmission? So the biggest answer is, and I'm going to paraphrase, I just want to make sure I don't paraphrase incorrectly. There's a big federal agency that has to approve, and then we've got lots of homeowners, because if we're doing transmission, we're doing long run, so I'm going across lots of people's tracks of land, and I've got to talk to all those people and tell them what's coming, and either get them to agree or eminent domain them or whatever I have to. I got to work it out, And that's not a simple process.

It's not a construction problem. It is a getting all the people and the interests aligned to get the line for like of a better word, sorted out, right? Okay. That makes a lot of sense. All right. I know I got sidetracked thinking through the three layers. Okay.

Can I go back to DC Flex? And I know you started talking and I sidetracked you. Can we go back to DC Flex and talk about what's going on there? Yeah. So about 18 months ago, we saw an opportunity. in meeting with developers, hyperscalers, and utilities to help close some gaps.

One of the things we talked about earlier was this difference in speed of doing business between the utilities and the data centers. And we knew there was a serious gap in communications and speaking the same language to try to reach an endpoint that was the same for both,

which is providing service to a new customer. So that was part of our objective with DC Flex. But the other was, as we worked with the hyperscalers in particular, and understood what they were doing with the AI factories, what the different workloads on the compute side are in an AI factory between building the model, training the model, and executing the model, there's a lot of inherent flexibility in those different workloads on the compute side of the data.

data center. So we saw an opportunity there for the data centers to not only get connected more quickly, but also be a resource back to the grid in times of emergency through a variety of different ways, not just the potential to flex compute workload, but as you know, all the data centers have an enormous amount of backup generation. That could also be a great grid resource. We have some other issues to get over there with a diesel fuel backup. But there's some work that we're doing as part of DCFlex on that with renewable fuels. And we've just completed some testing the last couple of months at a compass facility in Texas. We run HBO, not to make a Compass commercial, but for just to where you're going,

we run HBO to try to minimize the environmental impact of when we go on Jen. And to your point, we've paid for a massive amount of Gen. capacity on all of our campuses, there's unlocked potential there. That's part of what I hear in DC Flex. Absolutely. And we believe that with the testing that we've done with HVO, it is a much superior fuel to diesel in terms of overall carbon footprint, less NOx output, not by a huge amount, maybe in the 8 to 10 percent range. But the big difference is in particulate emissions, which are the big no-noes from runtime on those gen sets. Gens, yeah.

So with the testing that we've done, we're preparing reports that can then be taken back to state and federal environmental regulators and make the case to allow longer runtimes for those facilities that utilize HVO. And that gives them not only greater operational flexibility, but what a grid resource. We're not talking about asking to operate those generators a thousand hours a year. But if you think about Winterstorm Uri and Texas a few years ago, what a godsend that would have been if those generators were able to run during that time. David, I think about both ends of the spectrum, right?

So you talk about the winter storm where we'd call Snowmageddon, Texas, or on the other end of the spectrum, let's think about Phoenix, in July, Dallas in July, right? There's four, five or six days a year where the grid's tapped. Yes. Right? It's 107 degrees outside. I don't care how efficient everybody's air conditioners are. We've used it all.

And if I've got hundreds of megawatts and data centers and industry have gigawatts of generation capacity on the grid in Arizona, wouldn't that be amazing? It's such a huge potential asset to the grid. That's why we're trying to move this just to get 100 or 200 hours a year, allowance would handle most of the types of problems that you just described, Raymond. It's typically, you know, the big challenge for the grid is usually less than 100 hours a year where there's a super peak condition where temperatures are extreme, maybe some generating units are down, whatever it is, but it's not long periods of time. But if you call on those

today, you're going to get the owner of those gen sets out of compliance with their air permits. So this is really an important aspect that we're pursuing. And it can be a big, big plus for the grid. David, you're welcome to punt this next section I ask you about. But if you don't want to touch it, is the third rail here for a second. I've been amazed that in the, you know, five years ago, nobody had any idea what I did for a living. You tell them you're in the data center business and they, on to the next subject, you're not very interesting. I don't know what you do, Raymond.

And then kind of halfway through that, two and a half, three years ago, you know, data centers started being on every newscasts, and then everybody knew what we did, and then everybody's really interested in what we did, that every investor on the planet was pursuing what we did. And literally in the last four or five months, now it's gone to knowledge,

we know what you do, but you're worse than big oil. You're evil. You're ruining all of the water, and you're making everybody's electricity rates go up. And so you don't have to touch it as an eprio resentative, but what I hope gets heard from people who listen to this is, we, Compass, the data industry,

we are coming and trying to help enhance the grid because it is our lifeline. We are not coming and trying to take anyone's power or raise anyone's rates. We're, I think, actively trying to be the opposite of that. And I can only speak for compass's approach. But I think my industry has started to recognize we have to be a serve with the power industry. We can't be a to just take care of me. We've got to help the grid grow, help the grid perform, help the grid upgrade its assets. those are all things we should be doing.

And just this current narrative that's grown in last three or four months that we're ruining people's power infrastructure, I think is irresponsible at a maximum a lie and in a minimum irresponsible. And you don't have to comment if you don't want to. But I'd love to hear from a third party if I'm too far off base there. You're not. And I'm happy to comment on it because it's not something that's out of the realm of what we look at and work on.

And we actually have a couple of reports that we've put out about affordability in general, as well as the affordability impact from data centers on the grid and on other customers connected to the grid. There is not a correlation between a new data center and electric prices going up. In a right scenario, that could be the case. But the more normal scenario is that load growth, on the grid benefits all of the people connected to the grid. It makes the operation of the grid more efficient.

It spreads out the cost over a greater expanse of kilowatt hours. So everyone's cost comes down. And you can go to our website, Raymond. It's epri.com. And there's a DCFlex microsite underneath that that you can go to. And if you type in affordability, it'll pop up some of the work that we did. And what it will show you are statistics looking just at the U.S. That where states have load growth, average electricity prices are going down.

States that don't have load growth, the average electricity prices are going up. We have historical data that shows that. So, David, mine and your podcast may get the serious, most serious editing ever. This podcast may get just reduced to those 20 seconds. I wouldn't be surprised if companies just played those 20 seconds over and over, where the grid grows, people's prices go down. Can you talk to me a little bit about how Mosaic, does it help you do this? Does Mosaic help you look at this stuff?

Or am I misunderstanding what FlexMosate does? Well, Flex Mosaic is a framework that we built Raymond in conjunction with developers, hyperscalers, most importantly, the grid operators, the people who run the structured markets in the U.S. like Erkot and Texas, as well as the individual utilities and energy companies that serve the data centers. But it's a framework that defined five different levels of flexibility. It gives the data centers a starting point to say, okay, on the low end, we'll be flexible for some amount of our load for this period of time during the year. That's the smallest increment. It's at less than 100 hours that I was mentioning earlier.

All the way up on the far end, the fifth level, where the data center is fully grid interactive. They're providing ancillary services back to the grid. They're willing to island and be able to come off the grid for up to 24 hours at a time. They'll meet short calls to reduce load or day ahead calls, but they are fully interactive with the grid. And there's a few things in between. What the framework does is give the utility and the utility. the developer, the opportunity in the first time they sit down, the data center company can say, we are able to commit this amount at this stage, how fast can you serve me for this new data

center? That for the utility makes their interconnection study much, much more straightforward. Because now they know they're asking for 500 megawatts, but for 200 hours a year, they can be flexible for half of that. So that means my need to build additional facilities to meet that one to two weeks a year when I'm stressed to meet that load, they're capable of making it work without me having to shed load elsewhere.

That makes all the difference in the world to the utility planners and the operators. Because if I'm a utility planner, I don't get to plan for average load, right? I have to plan for, in the data center world, we call design day conditions. So, right, I've had an engineer come up with a perfect scenario on design day condition. This is, you know, day one, this is what it looks like. But I have to plan for what's the absolute worst condition it's ever going to be when it's, you know,

four standard deviations beyond the hottest day of the year or whatever. And I think about that from how I designed cooling in the building. Same thing for a power generation company. They have to plan for in July 17th, in the worst heat stroke of the year, they got to be able to provide the whole grid, right? And so if I'm a 500 guy and I'm never less than a 500 guy, they got to start planning, okay, how do I get, not only do I get 500 generated, but how do I get 500 moved around on the worst day? That's what you're saying, right? That's exactly right. And if you look, and this is a broad statistic, Raymond, but if you look at the grid in the United States, there's a

about 15% of the grid capacity that's built for that peak, which is maybe 10% of the time in a year. So if we can more efficiently do things on the grid in total, we can start to capture some of that excess capacity and put it to work more effectively. So it's not sitting around idle 90% of the time. Yeah, that's that flexibility of being able to, hey, if I can plan for some flex, you can take some load for me. worst days. Now I'm not having, I can, I can run at a higher utilization all the time because I don't need as much headroom. Exactly. And it's no different in the grid than the way you play in your scenarios in the data center too. The more of that headroom that you have to plan for,

those last increments of capability and capacity that you have to bring on are far more expensive. That's right. And all of those that preceded it. And for the utilities, when they make that investment, and then, you know, they're in rate base in some regulated, vertically integrated utilities. But if you're in a competitive market and you've built that increment to meet those peak times, and it's just sitting there idle, as an investor, you're not making money off of that either. That's not earning any return. No. So the better opportunity, is to better utilize everything connected to the grid, load and supply side, use it more efficiently.

And again, back to our earlier points, that drives down the overall price for everyone when you can run the assets longer, more efficiently, and more effectively. And a better utilization. All right, David, can I ask you, it's not really an EPRI question. I'm going to get, if you're willing to answer,

as I, I'm a data center guy, degree in finance, you know, I'm not smart enough to be an electrical engineer, but as I've spent more time in the data center business and learn more and more about, you know, electricity and the grid, one of the conversations I had on the podcast one time talked about, and I'm not going to use the right terms, but how the grid needs to stay in balance, right? I've got these turbines spinning and the turbines, you know, keep, and I'm not even sure what the right term is, keep the grid running at a certain, I don't know what's called. Harmonics is not the right word, but it's something like that. It's the frequency.

Right. And as generation comes on or comes off, it changes that frequency on the grid, right? And I don't understand how that works. And I'm not even sure if I've done a good job asking the question. But can you talk about that? Because as a as a consumer and you hear, hey, they took the grid down to avoid damage. They're managing that frequency getting out of whack. That's what I think is happening. But as a consumer, I don't think, we just think, hey, wait, my power's on or my power's off. There's way more science behind it than that, right? There is. And we could get into some really long-winded engineering discussions here,

but I don't want to do that and you don't want to hear it. The one thing that's really unique about the electric grid is it's really a living, breathing system that changes moment by moment because the generation follows the load. And by that I mean, and this is what the system operators do. they are watching 24 by 7 365 days a year what the load is at all the different points on their system and when they need more load they dial up more generation or you know reach out to somebody else in the market and say hey we need some more energy can you bring on this unit or that unit or whatever it is but they follow and so then when the load goes back down some of the generation will come off the

grid. So there's this whole following path between load and supply. In between that is in North America, the frequency of electricity is 60 hertz. Most of the rest of the world, it's 50 hertz. But that frequency and very tight tolerance on that frequency is the backbone of what keeps the grid running. Because once the frequency starts to get a little bit out of whack, then you can get runaway voltages on the high side or in some cases the low side, but the high side is the bigger challenge because if that happens and the voltage being supplied to in-use customers is higher than allowed by code, then you can start to ruin a lot of equipment inside homes, businesses, factories, data centers, those kind of things.

So that's why those tight controls are built into the. the grid. And if the frequency gets, I mean, and we're talking like two-tenths of a hertz out of 60, if it gets out of whack, then there has to be some regulation to bring it back into control so that you don't have that runaway situation and create a cascading blackout. Yeah, David, so you, I love the description. And as a lay person, this is where I think we get confused. I'm going to blow up transformer, substations, you know, equipment at your home. I'm going to blow stuff up because I've gotten outside of the 60 hertz range and now I'm pushing more loads coming down at you faster than your equipment can handle. And now I start breaking things.

That means all the connection is lost and I've got to go fix that thing before you can get generation again. And that's what the, hey, I'm trying to avoid a system blackout when I'm trying to turn the dials. So I think back to Snowmageddon. Why did they shut the grid down? And Urquot made the statement, hey, we were about to lose the whole grid if we didn't take everybody down. And that's because demand was so out supplying, was so out running supply that the frequency got out of whack. And it's, hey, we better. The only way we're going to get it back in a whack is just shut people down and back off the load requirements. Is that a fair way to say it? Some of that played into it. I think what you call Snowmageddon, I think there was just a complete shortage of anywhere close to the

ability to meet the demand, but when you're still trying to meet some of it, that does create those operational problems that we talked about with voltage fluctuations and the frequency fluctuations. You know, one of the best examples that happened in the last couple of years, you may remember the Iberian Peninsula blackout. That was caused by frequency regulation. issues. What happens when this occurs, sensors are set on generation equipment for overvoltage conditions. And when the frequency starts to oscillate, and particularly when it gets low, then there's an attempt to bring up the voltage to correct that. When that happens, sometimes the voltage gets too high, and then this is what happened in Spain and Portugal.

the voltage got so high trying to correct the frequency challenges that the sensors on the energy supply equipment tripped out for over voltage conditions on the grid. So that created the cascade to where the entire system was lost. And then you have to go, but then you're in a black start situation where you have to start all over from scratch, which is what they did in Texas and Snowmageddon. That is a very slow, painful process. So if there's a way to avoid it, we want to avoid it at all costs. And it's slow and painful because we now have to build the demand and the supply side by side to keep it right. You've lost this steady state you were running at and you've gone back to zero and now you have to slowly build back up.

Is that a layman way of saying it? It is. And what happens, you know, what you see a lot of times in, like Snowmageddon where the temperature is very cold. When the system does start to come back up, the cold-blowed pickup from all these people that have been without energy for so long is very large. So it spikes and you can run into other issues there.

Same thing when it's really hot in the summer. So it's an art as well as a science for the grid operators to be able to do that. Yeah, back to your living and breathing. It's a constant dance. Yeah. Well, David, this has been awesome. Eprey's vice president of electrification and sustainable energy strategy.

We got to, before we started recording, we got to talk a little bit of golf. I'm going to try to get you on the record to agree to come back into a whole episode just on golf. I would love to do that. We make sure we'll talk about some data center energy stuff to keep our bosses from yelling at us, but we'll largely focus on golf. I got to hear what golf balls you have in the cabinet over there and why you save them. There's about 600 in my study right now. And I have them in order of my affection for the golf courses.

Oh, I like that. I like that. So I thought, how will I organize these? And so what I did is I started with, this is my favorite golf course and this is my least favorite golf course. And then I just went. That's how I organized them. Which one got the top billing?

So you only get on the wall if I've played you. Yep, me too. And so Royal County Down is the best golf course I've personally played. I would not argue that point with you. I love County Down. I love Port Rush. Port Rush is amazing. Have you been to Dornick? Bornick is on the list. Yep. It's on. It's, Dornick's probably 12th on my board. Port Rush is I think seventh or eighth. Yeah. Yeah. County Down is just a special place. Number two is Pine Valley. And so film people yelling me about that. I've never been able to get on Pine Valley. Never. So some people yelling me about that when they're like, how do you not? at Pine Valley number one. I'm like, well, I really like Troyo County Down. Pine Valley's number two.

So I'm a fan of Lynx Golf, so I'm going to lean that way anyway. But, you know, the big three for me, I hadn't been on Pine Valley, Augusta, and Cyprus. We actually had a foursome on Cyprus in 2020. And we had to bag because of COVID. Oh, 2020. Yes. Yeah. So mine is, Not all of the courses I've played, but I keep track of Golf Digest, top 100 courses. Yes. And so one of my goals is to play all of them. Yeah. And I'm, you know, somewhere in the about 60 at this point.

Yeah, you're doing great. Have you played Bandon? I'm not played Bandon. No, every time I've looked at going to Bandon, I can do a trip to Ireland for less money than going to Bandon. Okay. So you're not wrong, but I want to say this to you. So I go on a trip every July for my birthday.

birthday with three buddies and we went to band in 2020 and we've gone every july since and we've been to ireland and we've been to scotland and we've been to england it's a challenge to get there but they are the most magnet for me you mentioned lynx golf so right his whole philosophy was to build golf link style golf course was on the water the way golf was originally designed there are no golf carts and for me that's golf so i agree the caddy every day you walk the golf course you're on the ocean You know, I hear you, there's great places in Ireland, but it is a magical trip to go to, to, to abandon dudes. There are now seven golf courses there if you count the par three. And two of my top five are abandoned.

Really? They're that special golf courses. It's a magical golfers experience. That's a ringing endorsement. So I'll have to figure out how to put that one on the list. That's not one. You know, my wife will go to Ireland with me, and she plays golf, but she's, She doesn't want to play those hard links courses, but I don't think I can get her to go to Bandon.

I will tell you, it's a dude's trip. First of all, it's not easy to get to, right, to fly into Portland and you still have to drive over the mountain and out. It is a getting there, but once you're there, it is golf heaven. I'll remember that. We'll stop talking about golf and give back to energy and data centers. Francis, you want to jump on and wrap us up.