The World’s Tech Giants Are Running Out of Power. This CEO Plans to Deliver.

0:00Before we get into it, this is a message for all those bold name fanatics who aren't just my mother. We need your help because we're cooking up something great. Yeah, this is your chance to ask us questions directly. So all the juicy material that ends up on the cutting room floor, you can ask us. You know, what did people tell us when their mics were off, for example? Send us those burning questions via selfie video or voice memo to boldnames at wsj.com. Or you could just email us those questions. We might use it on a future episode. Just in advance, I want to thank everybody. and I cannot wait to see what we get.

0:40The world's largest tech giants are clamoring for more power. Electricity, that is. We've heard Elon Musk talk about how there could be shortages in the near future. My guess is people are going to start hitting challenges with power generation maybe in the middle of next year, end of next year. Sam Altman and Jeff Bezos have been pointing to outer space. We're going to start building these giant gigawatt data centers in space. I would love to go build the Dyson sphere on the solar system and like, you know, make the world's gigantic data center with the entire energy output of the sun. How does it feel to be one of the few people on Earth who might be able to help these guys?

1:17It's not just for the hyperscalers and the tech companies. It's really for the world at large. Now, this is going to be a much more growth-oriented next decade than the prior few decades, but we know how to build things at scale, and a lot of the technology exists. So I like our chances. Sounds electric.

1:43I'm Christopher Mims. And I'm Tim Higgins. This is Bold Names, where you'll hear from the leaders of the bold name companies featured in the Wall Street Journal. Today we ask, how prepared are we to meet our exploding demand for energy?

2:04Scott, we're so glad you could join us today. I think most people are familiar with what General Electric is, but maybe not as familiar with what GE Vrnova is. Could you break down your business for us? What do you do? You bet. I mean, GE Vernova was born on April 2nd of 2024 as an offshoot or spinoff of General Electric, a company that has over 130 years of history. And that represents all of what was General Electric's energy businesses. We produce more than half of the electrons in the U.S. every day with our equipment through our customers. And that's across technologies, gas power, nuclear power, wind.

2:43We play a critical role with the grid. So it's really about applying the right technologies where the right resources exist. And that's really what GE Vernova is focused on doing every day. I mean, you're really manufacturing at scale the energy industry products, right? Gas turbines in particular are important. Why are they so important right now? Well, power density matters. How much energy or electricity can you create with the least amount of space? Gas is a very power-dense solution in that regard, which is very important. You also just are getting into a world where the load dynamics with things like hyperscalers are very extreme.

3:25And what I mean by that is the megawatts going up and down can happen quickly. The reality is gas turbines are very capable of ramping up and down very fast to follow the load. And that's why it's so well suited for this moment. So I like that you jump straight to AI. Certainly one of my favorite topics. Hyperscalers. That sounds like something out of space. I mean, I fall asleep every night with that on my lips, Tim. So these hyperscalers, is this the reason that your stock has been on a tear since 2024 and the split from your parent company? Or is it bigger than that? I mean, you're up like 300 percent, right?

4:07I mean, geez. It's a contributor, but it's by no means the only factor. I mean, if you take a step back, the world is only getting today about 20 percent of its energy from electric power. The other 80 % today is still coming from molecules, fuels, coal on industrial applications. And it's very hard to find someone that doesn't believe the electric power as a proportion of the total energy mix isn't going to double over the next, let's say, 25 years. So you mean especially as EVs replace gasoline for transportation and industry electrifies that kind of thing? As industries electrify, EVs is a very common one.

4:47I mean, heat pumps are becoming more and more prevalent in new homes that are getting built. So all of these dynamics are going to lead to a world where both industries and products electrify. And as they do, that's going to drive a greater need for our products. So hyperscalers are part of it, but it's by no means the only factor. So we've got EVs or electric cars out there. AI models need a lot of compute, which requires a lot of electricity. We've got heat pumps. Just demand for electricity is huge. But I think I once heard you talk about how this moment almost rivals the post-World War II build out.

5:26Yes. What exactly did you mean by that? Well, if you look in the U.S., from 1945 to 1975, every 10 years, the capacity for the electric power system double every 10 years. And then we got to 1975 and it continued to grow, but at a much slower pace. So it took almost 35 years from 1975 to double again. And then in reality, what happened from the early 2000s until directionally now is demand was flat. And demand was flat in the U.S. for a lot of reasons. It was flat because jobs were leaving the country, manufacturing demand. It was flat because of energy efficiency with things like our refrigerators and microwaves.

6:12It was also flat because technology with things like software and the internet drove productivity that hampered the need for more electrons. And the inflection point we're at now is those same drivers that were leading to a very flattish demand environment now need more electrons because what's happening? more industries coming back to the U.S., whether it be chip factories, whether it be data centers, that requires electrons. We also have things when it comes to the internet and software, well, all of a sudden AI, the next productivity driver, needs even more electrons, and it's a net add. So we're really getting into a phase that I'm not going to say the total U.S.

6:59power system is going to double in a decade like it did from 45 to 75, but the clear drivers of growth are substantial and are going to be substantial for a long period of time. And with it, it takes us back towards a period like 1945 where we've got to build things at scale. So what steps are you taking to prepare for this, you know, potential explosion in demand for electricity? investing every day. I mean, we've talked this year about investing seven to eight hundred million dollars in our existing factories in the U.S. We're adding over eighteen hundred workers, primarily production workers, into our factories to balance today's products, making more of them.

7:46But we're also investing in new products for tomorrow. So our research and development budget this year is going to grow almost 25%. Our real objective every day is to balance expanding our existing capacity more with the products we know how to make today, and we're doing that, while also keeping our eye on the next decade when we're going to need some new technologies to continue to meet this moment. And sorry, what's your R &D budget now? You said it's going to go up 25%, but what's the baseline? 1.2 to 1.3 billion dollars annually, growing at 20 to 25 percent. I mean, we're in a bit of an unusual moment in history because typically when people are trying to project demand for something like electricity, it's very hard to make that prediction.

8:32But we know that, you know, for instance, Sam Altman at OpenAI has said that they're going to spend more than a trillion dollars on new infrastructure over the next few years. So it doesn't take, you know, much time with a napkin to calculate what that's going to translate in terms of gigawatts of fresh electricity supply. Realistically, is your company or you in combination with your competitors equipped to build enough capacity to address this explosion in demand over the next five years? Yes, I think we can meet this moment. And I think the way to caveat it, though, is it's not going to be solved in five years.

9:13It's not just about the equipment. It's about having the EPC or the engineering procurement and construction companies to build the plants. It's about getting the gas pipelines available. It's about working through the permitting and land constraints and some of the dynamics in the country with different communities either wanting or not these investments in their community. What's clear is the direction of travel is going to include a substantial amount of investment into this space, but I wouldn't narrow it down to five years. I think we're really looking out the next 10 to 15 years. And can we meet that moment, Vernova and the industry at large?

9:54I'm highly confident we can. After the break, President Donald Trump says there are signs the U.S. could face an energy crunch in the not-so-distant future. That is why today I will also declare a national energy emergency. We will drill, baby, drill. How is Scott thinking about this looming crisis? That's next.

10:34your gas turbines are sold out through i think 2028 so how is this not a manufacturing bottleneck well what we say every day is we'll make the gas turbines as fast as the pedestals are ready for the gas turbines to arrive. And that goes back to the dynamic before. It's not just about manufacturing the gas turbines. Now, if incremental gas turbines are needed by 2028, because you're right, we're largely sold out between now and 2028. We'll work to solve that dynamic. What we see today is an environment where through 28, we likely are meeting the moment. But we also acknowledge that as the orders continue to come in as strong as they are, and as the demand continues to rise, we may need to add incremental capacity to meet it.

11:26But based on everything we see today, that incremental capacity would come later than 2028. Yeah, more broadly, just beyond your business, there are real signs here in the U.S. that we face an energy crisis. The administration says a new executive order is aimed at improving manufacturing, transportation, agriculture, and defense industries. That is why today I will also declare a national energy emergency. We will drill, baby, drill. I know our colleagues here at the Journal have reported on just the sheer backlog of energy projects and the fact that our national grid is basically running out of room.

12:04In some locations, for example, data centers won't be able to plug into the power grid until the 2030s. Get a big picture thinking, how does the U.S. overcome this challenge? Well, I think in the next five to seven years, a lot of those data centers may get built with the power attached to the data center. That's good for you, right? Because you're selling to a different customer. Clearly an opportunity for us, yes. Now, I think at the end of the day, those power foundries and data center perks will get connected to the grid. But if the hyperscalers can't wait for the grid connection. We'll build the plant.

12:42We'll connect it to the data center. As the grid ultimately catches up, that can ultimately be a very positive thing for the grid system in total. Because in that case, once the grid connection comes, at moments where the peak power is very high, or said another way, the reserve margins in a system get very low, those power foundries with the data centers can feed the grid and provide that peak power. So this is a dynamic that I don't think is necessarily well understood today, which is this is a real opportunity to support the utilities and support the local communities by giving them the excess power they need during very fractional periods of time during the year.

13:29And it saves those communities or those utilities from having to build incremental what we call peak power that's really only needed when it's really hot in the summer or really cold in the winter when electricity demand rises. Right. And then those facilities aren't being utilized the rest of the year. It's kind of wasteful. Exactly. And that's where this is a real opportunity. You know, the idea of OpenAI or Google or whoever having their own kind of energy plant, that vertical integration almost reminds me going back many, many years to Henry Ford and the rush to build out the automotive industry and having to vertically integrate the supply chain, going down to the jungles to get out the rubber to make the tires for the cars.

14:10I mean, it's kind of a signal of how fast this kind of growth is occurring. Yeah, I think good businesses are very effective at identifying what their next bottleneck is and eliminating it. And if the electron, going back to data, electron, and chip are the three key components, businesses are going to find ways to de-bottleneck what could slow them down. And today, as the electron is becoming more challenged, you're right. Great companies are then re-evaluating what model they're going to use to demodel that. And that's exactly what's happening today. We've talked a lot about gas. You're also in the nuclear business, which is having its own renaissance, at least from a regulatory perspective.

14:56What role is that going to have in your company's future? It's a material role. I mean, we're in construction on our first 300 megawatts small modular reactor today. I'm sorry, what's a small modular reactor? That sounds like a suitcase nuclear thing. It sounds pretty cool. Less suitcase, more size of a football field. But think to yourself the size of a U.S. football field, 300 megawatts. It's generally 300 megawatts to contextualize that is enough to power 300 ,000 homes in the U.S. So that's not small by any means. It's a decent sized power plant on its own. For sure. We have our first plant or application into the Nuclear Regulatory Commission for U.S.

15:41build for a plant in Tennessee called Clinch River with Tennessee Valley Authority. We expect the Nuclear Regulatory Commission will approve that build or the beginning of construction by the summer of next year, summer of 26. And if that happens, we expect to be in construction on our first plant in the U.S. sometime in 2027. Now, once construction starts, it's about a four-year build right now to go from beginning construction to having incremental electrons on the grid. So for the U.S., we see this starting to become a material part of the equation in 2031, 2032, something like that, and then growing.

16:25So will nuclear power solve the surge of demand this decade? It won't. Can nuclear power start to be a material contributor 2030 to 2040? You bet it can. We've got a lot of conviction that it can and that it will. After the break, how might AI affect GE Renova's future? There's going to be very material winners and losers. But in my view, based on everything that I see, for whoever wins in this journey, the infrastructure buildout is substantial. That's next.

17:11we've talked a lot about this surge in demand for electricity. Um, people who, who follow me on social media or can parse the subtext of my columns know that I think that within the cone of possibility in the future is a significant crash, uh, in terms of, uh, financing, uh, and then ultimately construction of these new AI supercomputers. Um, If something like that happens, I'm not saying it will, what happens then? Are you caught flat-footed if suddenly people are canceling orders? Or does that affect your hiring, your build-out? Well, I would take a step back and then say when people talk about bubbles or talk about crashes, so to speak, it's often in the vein of the equity markets at large and kind of how different companies are valued today.

18:07I do expect with the AI surge, there's going to be very material winners and losers. But in my view, based on everything that I see for whoever wins in this journey, the infrastructure build out is substantial. So you win no matter who else wins because you're the pick and shovel provider. They can't do it without electrons. There's going to be a need for a substantially larger amount of electrons in this country, regardless of who wins and loses. And I think it goes back a little bit to the dimensions of really our business today. I mean, we've talked a lot about gas turbines. Our explicit backlog today on gas turbines only has about 10 % of the demand attached to AI or hyperscalers.

18:54Now, admittedly, we also talk about another metric that we call slot reservation agreements. These are future slots for equipment that aren't yet orders, but the customers have put a substantial amount of cash down to secure the slot while they figure out who's going to build it, while they figure out what sites they're going to apply it to. In that case, it's about a third of our slot reservation agreements are tied to hyperscalers or AI. So I give those numbers to contextualize. This is only a small proportion of the electrification of industry. It's an important one. And it's one we're very motivated to serve.

19:36But just it's very hard to not believe that the world doesn't need more energy and that for the world to meet its goals, a large proportion of that energy needs to come from electricity. So this infrastructure build out is going to happen regardless. I'm just curious if AI reflects only a small portion of that projected demand. What's the pie chart look like? Like, what's the rest of the demand going toward? Well, we have regulated utilities across the U.S. that are simply seeing their reserve margins come down. Sorry, what's a reserve margin? The amount of excess energy that exists relative to moments of peak power.

20:19Ah, so these utilities are saying if something happens, if there's a lot of demand in the winter or the summer, we don't have enough of a buffer to make sure that the grid stays intact. That's exactly right. And what also is happening is as we add more and more wind and solar as the capacity, what's needed simultaneously is more gas to follow it. Because in reality, as you have more intermittent or power that comes from resources you don't control, like the wind or the solar, you need something else to be the force multiplier that protects for those moments those resources aren't there. That fact alone is driving regulated utilities to need more gas while they build out those other technologies.

21:07But this also isn't just about the U.S. I mean, we've commissioned in the last few years over 10 gigawatts of new power in Taiwan. A lot of that, admittedly, is attached to the TSMC build-out for their manufacturing. We've done 6 gigawatts in the last couple of years in Saudi Arabia. You look at India today, our most populated country in the world, 80 % of the electricity is coming from coal. Over time, that's going to evolve. So hyperscalers and AI is important, and we're very motivated to serve that opportunity. But there are a lot of variables at play that we're investing into to serve all of these markets.

21:50You know, we've talked a lot about AI here, about the need for electricity to power those data centers. But I'm curious also how AI might affect your business in other ways. It is a technology that has the promise of at least of being able to upend all these traditional industries. And when I hear Microsoft AI CEO Mustafa Sullyman, who has been a guest on this program before, he paints a future where energy costs can come down dramatically because of advances in AI, AI being very good at finding efficiencies and maybe taking cost out or just making things, the power go further, perhaps. How do you think about that potential?

22:31What does it mean? Because he's talking about a world where power cost comes very cheaply. Well, I think if you take a step back, most manufacturing processes over the medium to long term should have cost competitive benefits. We won't meet this moment without a combination of using AI, physical AI, and a growth of our labor pool, because ultimately we will run into challenges with labor availability. As we get those technologies to a place that they're effective, that will drive a lot of productivity. And that productivity will bend the cost curve, especially with the amount of volume we expect to see going forward.

23:18I also agree with him in the vein that there are a lot of inefficiencies in the system that with AI, we'll be able to better point the electrons where they're most needed. Software and AI is going to help us with that because as our planning really leverages weather patterns, really projects with another level of sophistication how much electricity is going to be available in one location, we will drive efficiency in getting that excess electron to the place that's most important for it. So there's real opportunity here. I want to get your perspective as a GE lifer, which you are, of course. so you know some of us on this call remember during the internet boom right uh just before 2000 there was this projection that uh we were going to need a lot of energy for this new internet that didn't work out i mean kind of went the opposite way right it led to a lot of efficiencies what lessons do you take from history from that history in particular do they temper your current expectations for what's coming next or do you feel like oh no this time is different because you don't seem tempered right now.

24:31You seem very enthusiastic about it. Yeah. Grounded enthusiasm with what we work on every day. Grounded enthusiasm. Is that an electric joke? If you just take a step back and go back to the late 90s and the early 2000s and think about Enron, think about the deregulation of the US that was happening at that time. There was a level of fragility where there was concern, especially in markets like California, that was leading to a need to create more power for what was really, again, peak margins or reserve margins where during small periods of time, there was a risk that the existing system would not meet the peak demand.

25:20And during that time in the late 90s and early 2000s, there was a very substantial buildup in the U.S. for the risk of brownouts for peak demand. Sitting here today, there is that dynamic in certain parts of the U.S. where simply because of renewables growth, simply because of the electrification of other things, they need to invest. But there's too many other variables at place. When you look at the number of chip foundries that are being built today in the U.S. that need electrons, it's a very different demand cycle. When you look at the data centers being built out, it's a very different demand cycle.

26:02And I think the other difference between 99 and today is what we talked about earlier with other global markets, Taiwan, Saudi, that are also demonstrating real demand. Now, that doesn't mean every day we're not scrutinizing our orders book, iterating with our customers on timing. Because I think the most important question is not where this is going, but exactly the time everything is going to be needed. And good businesses build things exactly when they're needed and not too early to the point that they stood on working capital as an example. So where we do have a lot of conviction is the direction of travel over the next 10 to 15 years.

26:50exactly the timing of what's needed in 28 versus 31, we have a lot of humility that the market will tell us those answers as this thing moves forward. But what we're investing into is to be ready for this moment. And once we're ready for that moment, we will only build on purchase order on demand. So we don't build on spec, so to speak, hoping to sell something. What we will do is make the investments in our factories, make the investments in R &D to ensure we're ready for the moment, and then wait for the customers to tell us they're ready. Well, Scott, this has been fascinating. Thank you so much for taking the time to be with us.

27:38Christopher, Tim, thank you very much. I've really enjoyed it. And thanks for having us and look forward to future conversations.

27:48And that's bold names for this week. Our producer is Alexis Green. Our video producer is Kasha Broussalian. And our fact checker is Aparna Nathan. Jessica Fenton is our technical manager, Jessica. And Michael LaValle are also our sound designers, Jessica, wrote our theme music. Our supervising producers are Catherine Millsop and Katie Ferguson. Our development producer is Aisha Al-Muslim. Chris Zinsley is the deputy editor, and Philana Patterson is the Wall Street Journal's head of news audio. For even more, check out our columns on wsj.com. We've linked them in the show notes. I'm Tim Higgins.