Inside the New Energy Arms Race to Power Artificial Intelligence

As enterprises rush to deploy copilots, automation and secure private AI, energy has become the defining constraint. In this episode of the AI Proving Ground Podcast, visionary leaders including NightDragon's Dave DeWalt, Morgan Stanley's Jeff Holzschuh and Beacon AI Centers' Josh Schertzer unpack the new reality unfolding across the grid. They reveal why power — not land or fiber — is now the first question in every AI strategy, and what executives must do now to de-risk growth, secure their operations and stay ahead of the coming infrastructure supercycle.

Editor's Note: This special episode of the AI Proving Ground Podcast was recorded during WWT's Business Innovation Summit, which took place at the PGA TOUR's World Wide Technology Championship in November 2025. 

The AI Proving Ground Podcast leverages the deep AI technical and business expertise from within World Wide Technology's one-of-a-kind AI Proving Ground, which provides unrivaled access to the world's leading AI technologies. This unique lab environment accelerates your ability to learn about, test, train and implement AI solutions.

Learn more about WWT's AI Proving Ground.

The AI Proving Ground is a composable lab environment that features the latest high-performance infrastructure and reference architectures from the world's leading AI companies, such as NVIDIA, Cisco, Dell, F5, AMD, Intel and others.

Developed within our Advanced Technology Center (ATC), this one-of-a-kind lab environment empowers IT teams to evaluate and test AI infrastructure, software and solutions for efficacy, scalability and flexibility — all under one roof. The AI Proving Ground provides visibility into data flows across the entire development pipeline, enabling more informed decision-making while safeguarding production environments.

Chapters

• 0:00 The AI Power Crunch Emerges
• 0:43 Guests And Stakes Introduced
• 4:09 Energy Investment Supercycle Explained
• 7:35 Power-First Data Center Strategy
• 9:06 Grid Resilience Meets Sustainability
• 12:32 Energy As National Security Risk
• 16:40 The Grid Becomes A Digital Attack Surface
• 19:21 AI For Cyber Defense And Reliability
• 22:33 Where Capital Flows And Who Bears Risk
• 27:22 Microgrids, Gas Bridge, And SMRs
• 31:23 Workforce, Supply Chain, And Timelines
• 34:15 Executives’ Playbook And Partnerships
• 36:14 Closing: Power, Security, Compute United

Transcript

SPEAKER_00: 0:00

From Worldwide Technology, this is the AI Proving Ground Podcast. Today, as trillions of dollars are being poured into new energy capacity to support AI, a new scenario is unfolding. The race to build smarter models is now colliding with the hard reality of how we power them. Data centers are going up faster than the grid can evolve. The digital economy has leapt ahead of the physical one. And suddenly, the real limit to AI isn't silicone. It's megawatts. This is all more than an engineering challenge. It's a geopolitical issue, a national security concern, and a leadership test for every enterprise betting big on AI. If we can't modernize the grid, secure it, and scale it, the entire AI revolution is going to slow down. So today, you'll hear from leaders across finance, infrastructure, and cybersecurity who are confronting this energy paradigm shift in real time and what it means for the future of AI. We'll be joined by Jeff Holshoe, chairman of Morgan Stanley's Institutional Securities Group, who has spent nearly four decades advising on some of the largest energy and utility transactions in modern history, and now has a front-row seat to how AI is reshaping global capital flows and infrastructure strategy. We'll also hear from Dave DeWalt, founder and CEO of Night Dragon, a legendary cybersecurity leader who has guided companies like McAfee, FireEye, and Mandian, and now counsels governments and enterprises on the rising national security risks tied to our rapidly expanding digital and energy systems. And finally, Josh Scherzer, CEO of Beacon AI Centers, which is building some of North America's most advanced AI-ready data centers, redefining how power, sustainability, and infrastructure must evolve to keep pace with AI-scale computing. This conversation took place just a few weeks ago at WWT's Business Innovation Summit, which curated some of the most engaging and influential minds in business and technology today, and equip leaders with the clarity, conviction, and playbook to compete and win in an AI-defined future. So let's jump in.

SPEAKER_02: 2:17

So Jeff Holshu's the chairman of uh institutional securities group at Morgan Stanley. Over his nearly 40-year career, he's advised on some of the largest MA deals in the energy and utility business. Josh is CEO of Beacon AI Centers and senior operating partner at Nadia Partners, previously the CTO at Blackstone, and now leads one of North America's most ambitious AI infrastructure efforts, building high-capacity data centers designed for the AI era. And Dave, if you don't know Dave, he is like royalty at RSA and Black Hat and other industry conferences around cybersecurity. But Dave is the founder and CEO of Night Dragon Security. He's a veteran CEO and investor. He's led companies like McAfee and FireEye and created over$20 billion in shareholder holder value. And also prestigiously served as an advisor to U.S. uh presidential administrations on cybersecurity. Uh and for those that weren't following along yesterday, happen to win the pro-am.

SPEAKER_03: 3:13

Uh and by the way, it was a it was a team effort for sure. I mean I know Jay's here and Steve's here and uh Chris Pronger, so we had a wonderful time, and big thank you to WWT for sponsoring it. But it was really nice to get a golf trophy, I have to admit. I don't get very many loads.

SPEAKER_02: 3:31

So thank you for the I'm still calling shenanigans. Uh integrity is a core value at worldwide. So I think we need to do a little inspection on that 21 underscore that they they shot out at Cardell.

SPEAKER_01: 3:42

So by the way, uh Brian, it's not it's not often that you get an energy guy, a finance guy, and a security guy on the same stage. So hopefully, if we're somewhat productive, we'll solve some of the world problems by lunch.

SPEAKER_02: 3:52

I I will say in our our little pre-meeting out in the hallway this morning, uh after we first introduced each other, I don't think I got a word in edgewise. So I might be pretty quiet over here while the three uh informed individuals up here do all the work. So uh Jeff, we're gonna go ahead and get started just to set the stage. Uh AI adoption is accelerating at an unprecedented pace. How is this surge reshaping the energy demands and data center design? And there's this concept floating around of an infrastructure supercycle. Like, what does that mean for corporate leaders and investors around energy?

SPEAKER_04: 4:24

Well, thank you, Brian, and thank the WWT uh team for everything. It's been uh a great couple days, and I continue to be inspired by your own innovation. So secondly, I want to thank you for that high school yearbook picture you had in it.

SPEAKER_02: 4:40

Um that was that was Can we put that back up? That was pretty pricey.

SPEAKER_04: 4:45

Uh and uh third, I've got a bone to pick with G2 uh because he just ruined my next 72 hours because I type like this. And so I'm I'm going back to uh I'm going back to cancel golf for the weekend and learn how to type. Uh and also being part of the energy power sector for many, many years, uh, no one has ever used the word cool in association with us. So I'm announcing I'm now going to be known as J3. And we'll be cool. All right, on to some some substance. Um look, the the energy spaces couldn't be more in transition right now. And and frankly, I don't think it's hyperbole to think that it's reshaping the global economy. The explosive growth and capbacks, uh, the grid modernization, uh, the safety issues associated with it. Uh it it's you heard G2 use the number, and I think McKinsey's and others are using it. It the expect expectation is five trillion dollars to be spent between now and twenty thirty in this space, and about a trillion eight of it will be in the United States. So it's an incredible you you look at the you know GDP in the United States for the first six months, it was about 1.2 percent growth, and 1% of it was the spend on AI and data centers. So could not be more critical. Uh it's just a massive transformation. And the infrastructure question, Brian, is uh there's no question that the digital economy is greatly advanced beyond the physical uh ability right now to service it. The grid wasn't built for any kind of this magnitude, uh, and that's where a lot of focus. And AI, frankly, is a friend and a foe. It's the foe in that it's growing so fast and has so much demand on power, but it's also a friend in that it will innovate the grid, it will innovate the way we transmit power, it'll innovate the way we use power, and that's all going to be beneficial at the end of the day. So um it's an incredible uh you just look at the four or five announcements in the last 30 days. We've never seen anything like this in magnitude. I mean, U.S. government partnering with Brookfield for$80 billion of new nuclear build, a$45 billion transaction for Alliance data centers, uh, you know,$40 billion announcement for this Hyperion data center in Louisiana. Those are staggering numbers, and it's it's a relatively tight community that's doing all that. So uh I am, as you know and heard last year, I'm a huge believer that this is a mega trend. I'm probably a little bit more cynical that it's gonna go as fast as people think it is, um, but it's for real, and uh we'll take being a cool kid for five minutes.

SPEAKER_02: 7:33

So any anything to comment on?

SPEAKER_01: 7:36

I mean, I would just add that, you know, obviously the demand is is unprecedented and something we've never seen. And sort of, you know, because of that, you're sort of seeing this inversion, you know, of priorities when it comes to the build-out of data centers. Whereas before it was about where can I find land in the NASCA utility, how to get power, where can I find land around um ideal fiber optic networks. Now it's like, where is the power and we'll bring the land and data center to it? So that will be the trend that continues for the foreseeable future, and whoever secures the power will ultimately be the uh the ones that win.

SPEAKER_03: 8:06

Yeah, absolutely. And I was working on my uh new nickname too. If you're going with J3, I'm gonna go with D2. Rhymes better with G2. So I'm I'm working on it. So just for the All for it. Yeah. Yeah. Now do you type like this or do you Oh yeah, definitely. I have two thumbs though. That's all I got. It's two thumbs to work on. But listen, you know, I think you said it really well. I mean, maybe we live in interesting times, right? It's like a curse and a blessing. And, you know, a little bit of this panel to talk about that. But what a what a world we're living in right now. If you can't get excited about technology and inertia and things that are happening and AI and quantum and blockchain and security, it's uh, you know, you're in you're in the wrong time. And what a what an opportunity we have on this planet right now, I think, to really make a difference. But you heard G2 say things. Power and security equals trust, if we can solve those things. So it's uh it's a big deal as part of all this that's going on. I know we're talking about it's a great point.

SPEAKER_02: 9:01

Uh Josh, you're building some of the most advanced AI data centers in North America. It is a different mindset when you think about an AI-ready data center as opposed to a traditional data center. What is what does sustainability and grid resilience look like under this new model?

SPEAKER_01: 9:16

Yeah, I mean, it's interesting. I think I might have a controversial opinion there. So, like, you know, so grid resilience is obviously of the most utmost importance, but as you guys know, um the the sort of like traditional electronic uh electrical grids uh are gonna struggle for the perceivable future to keep up with the demand. And so that's obviously uh forcing more of a hybrid of approach, whether that be microgrids or on-site and behind-the-fence generation. And that's and that sort of hybrid approach uh will create some challenges around how you see how you how you look at and how you tackle uh sustainability. And so uh what we're seeing now, and what we're also hearing from some of our hyperscaler friends, is that because the demand for compute is so strong from them, because they're essentially bleeding from the scenes for this compute, uh, and because only going with electrical or only going with renewables is not going to be enough uh to power that, to, to enable that compute. Uh, essentially what you're seeing is them sort of, I don't say lower, but maybe temporarily lower their target such that they can get to that power and to that compute ultimately faster. It doesn't mean that sustainability is not front and center in the conversation. It means that the order in which it comes uh tends to be different. So maybe where you would you would build a data center out where it will have better sustainability in it, but you might do things like layer and carbon capture and things like that, uh, you know, in a phase two and a phase three part of it for the sake of getting to market and getting that energized as quickly as possible.

SPEAKER_04: 10:32

Yeah, the only thing I'd say, and I probably should have said going in is that if we're talking about AI and data centers primarily, but the really the reindustrialization of manufacturing, electric vehicles, there are so many other things that are happening fast calling on power that you're gonna compete with that, those entities just as much as competing for other data centers and other AI providers. Interesting.

SPEAKER_02: 10:57

Revitalization of manufacturing in the U.S. is a huge component of this, just the power demands that that's gonna put on on the grid.

SPEAKER_04: 11:02

The only time I've ever been quoted, I was down at a conference in DC and I said that you know electrons are gonna be the new oil of the digital economy. And I've seen it everywhere, but no one's ever given me credit for it.

SPEAKER_01: 11:15

We like to say that the uh that uh the future of AI is gonna be gated on megawatts, not chips. Yeah.

SPEAKER_02: 11:20

Yeah. And by the way, I'm just glad Sarah quoted Jim instead of Jim quoting Jim, because that that tends to get awkward when you quote yourself. Uh Dave, does does grid resilience mean something different when you look at that kind of a statement?

SPEAKER_03: 11:33

Oh yeah, absolutely. Um funny, Jeff. Last year uh you were on stage too, and uh it just created such an epiphany for me. You know, I've spent 25 years in security and national security, and I'm like, I gotta learn energy. I just have to learn this. And you inspired me. I think Jensen inspired me too, especially his shoes he was wearing that day, which might better have thought later. So um, but I I decided I had to really learn and be curious. And uh I ultimately over the last year joined the Exilon board, right? I don't know if you know the company called Exilon, you know, powers a lot of the smart grid in the Northeast and Midwest. And a lot of people were asking me, like, why are you joining an energy board, especially a monopoly of uh, you know, of a grid, and you know, with all the political things that are happening there, I just had to learn it. And talk about like a major epiphany. I mean, you want to talk about the biggest national security risk we probably have in the United States and most sovereign nations around the world, it's energy. Without a doubt, it is energy. And I always thought, well, misinformation, you know, cyber attacks, you know, some other type of problem that we might have in the invisible domains. It is literally power. And so much so that when you start to look at this perfect storm of issues facing the grid resiliency, you start to realize like we have to catch up. And I always use a joke what did the mama tomato say to the baby tomato when they crossed the road? Catch up. Right? So we had that terrible dad joke. Oh, Jim Kavanaugh. I made you laugh with that one. That was a quick one, wasn't it? Please laugh at that. Just make it feel better. See, that just rolled right off my tongue. But, anyways, um, you know, what's the grid resiliency issues, right? We have such little capacity for the demand that we have right now. And I was asking the team, like, okay, what's the biggest uh portion of our grid as measured by how much capacity? And it was like 35, you know, megawatts. And like, how much demand do we have? Uh we'll add a zero to that. It's measured in gigawatts, right? So, well, how do we get from megawatts to gigawatts? Well, there's generation issues. Oh, wait, well, not just generation issues. We've got all these resiliency issues because not only do we have to have data centers stay up all the time, we got to figure out how to create surge protections. We got to start to create resilience to that grid in a different way. Oh, what do you mean the power goes one way and now we have to go both ways and we have to manage that? That's right. So I started to learn, okay, what's a PMU, like these phaser management units? Oh, what's like programmable logic arrays? How do we figure this out? What's a DER? And all these acronyms of energy meets security in a way that we just have to catch up. And it was a real epiphany over the last year to learn this is probably our biggest weakness we're facing right now. And because, from a national security perspective, we don't want these data centers to leave the United States or whatever country you're from and go to the Middle East. And when you compare that to what's happening in China right now with their power, with their nuclear, with their fossil fuels, we are so far behind in this thing. It's scary at the moment. So we as a village building data centers and working with security and trust and coming together, if we don't solve that underlying foundational problem of power and energy and supply of this, we are gonna have a bigger problem. And there's a lot of politics to this from hyper-local to federal. There's so many layers to this challenge that it just is mind-boggling. But I'm not sure if I'm more scared after the one year of really learning or more optimistic, but we have technology that can solve it. That's what's great about this. And we have the effort. And when America and these countries really get going, we can solve a problem fast. So I'm excited. I'm hoping from one year from now, we're much further along with this problem than we are right now. But we got a lot to go.

SPEAKER_01: 15:14

It's it's interesting when you think about it. When you think about the amount of investment and innovation that's gone into securing the enterprise, and then you focus on the physical infrastructure and the challenges around it, it's quite, it's quite different. Uh, I look at some of this, and and you'd probably be more familiar uh familiar with this than me, but you know, when I was a CTO uh CTO of enterprise at Blackstone, uh, you know, securing the internal infrastructure was challenging but doable. When it got down to like building management and physical security aspects, those these legacy systems with you know outdated firmware, with not the right process behind them, not the same process behind them that you would have at the enterprise layer that was driven by a CISO, uh, it's quite frightening when you think about the impact and the impact being not just digital, but being physical.

SPEAKER_03: 15:53

Yeah, and one more point to that, uh, and you know this well, like the the realization of the climatology issues on the grid resiliency really struck me too. Because in the first safety and security report across the eight operating companies at Exelon, I won't give you all the numbers, but let's just say, you know, the tornadoes, the hurricanes, the winds, the fires, the heat. And it hasn't been since the 1950s that we face an existential crisis in energy. And what happened in the 1950s? Air conditioning, heating units came about, and suddenly the energy industry, as you know Jeff really well, like face like, wow, we got to solve this problem. Suddenly AI and data centers created this times a million. And suddenly we're sitting there with physical issues, cyber issues, resiliency issues, agility issues, and it's really Yeah.

SPEAKER_04: 16:40

I mean, the thing that it just always amazed to me is uh you go back a decade, 10 years, and we had an analog power grid in this country. So we've switched it now to a digital power grid. But if you asked anybody the question of what's the largest digital network in the world, no one would answer power. But it is. So if you go to the 350 million people in the US and the rest of the world from their home back to the generation source, every piece of that has a data entry mechanism, meters, substations, everything else, all connected digitally. So the surface to create the problems that David tries to solve on cyber and just just harm is in it's the largest surface in the world. And it's it's and expanding quickly. Yeah.

SPEAKER_03: 17:28

And that's what's incredible now. I don't know if you all heard this. Uh there was a big attack over the last two years called Bolt Typhoon. You guys hear this one. Michael knows this one. Well, we now go by typhoons now in the cyber industry, but it was such epiphany again around the access and ease to that attack surface, Jeff, that we learned. And this was China really creating a uh a signal to the United States, I'll just say, over the South China Sea, where they were able to really penetrate our energy and utility industries through some new uh attack surfaces that we really hadn't seen before. And suddenly we're realizing our grid is vulnerable from this, both physically, as Josh says, but from a cyber point of view. But again, I feel like the inertia is coming. And, you know, you can hear all the leaders like Jay and G2, we all talking about we gotta catch up here in this area, and we will, but we're vulnerable at the moment. Yeah.

SPEAKER_02: 18:21

It's it's uh a very interesting time. You know, I think about the the attack surface and the level of connectivity across the grid itself is is actually, you know, we talk about this in terms of how fast things are moving. That is actually an incredible use case for AI, not only for resiliency and uptime against natural disaster and other situations that might occur on the grid itself, but also to defend against cyber attacks and nation-state actors that exist out there. That wasn't a possibility a year or two years ago. And so the paradigm is changing in terms of the way that we would defend and create stability. And we've got some outstanding clients in the room that are living this every day in the energy and utility sector. And the level of innovation and forward-looking opportunity that we are working on collaboratively with them and with our vendors in the room is incredible. It's changing the way that we deliver value and stability and reliability and protect the grid, uh, which is a requirement because the grid only has to grow in order to serve AI use cases for the rest of the enterprises that we have. Um, Dave, I'm gonna come back to you on the topic of cyber, though. When you think obviously the grid is a key tenant in this when we talk about energy in the new paradigm of AI, what what does that mean in terms of maybe executives in this room for innovation speed and how you balance that while the growing imperative to secure both physical and digital assets? It's a it's a different time for executives on the enterprise side as they think about this.

SPEAKER_03: 19:45

Yeah, I mean, for many years, uh, you know, as a CEO of Fire Eye mandate, uh, I have like a tattoo of my arm literally, or figuratively, but um not literally, figuratively, but we watched all these attacks, and I always said the number uh 5572, which is 5,572 confirmed IP thefts on American soil from China during my time just as CEO of Fire Eye. I mean, just take that in. 5,572 confirmed IP thefts, right? So this is the world we were living in. This is in a five-year period of CEO of Fire. So like you're just you're just living with this kind of problem that's been going on. And, you know, we now we now see like an era where we we can catch up with AI again. And it's it's really optimistic for me to kind of feel this way because suddenly autonomous pen testing as an example. We can create autonomy and every layer of the stack for the good and create self-inperity and self-learning networks. I had to learn new words like autonomic and autodidactic. There's some new AI ones for you as well. But like self-learning, self-healing systems brought on by AI. And it's the reality is here. So to improve the grid is really a matter of us applying a lot of that technology. And one of the biggest areas we've learned is these autonomy platforms are really here to say. I'll give you a stat really quick. As of uh the end of October, we had 19,972. I'm good with numbers, 19,972 CVEs. So y'all know the CVE is a critical uh vulnerability exposure. So we had almost 20,000 of these in the first 10 months or so of the year. And um now for CISOs and CIOs, how do you keep up with 20,000 vulnerabilities disclosed by the United States government alone? Well, you got to prioritize them, you got to understand whether you have them in your attack surface. And there's a thing called IT now and OT. So you gotta understand, you know, how do you patch a system on OT that might have been air gap? Do I touch it? Do I not touch it? What's the vulnerability? What's the exploit? What do we do? But now with the capabilities of AI and autonomy, we have the ability to start to do this self-healing, self-learning system on both IT and OT, which is really, really encouraging for me to watch. So now while the bad guys are also doing the same thing, the good guys, all of us, have to use these tools in a way to make all this safer and quicker. And time's not on our side because every time a vulnerability does come out, the bad guys get it too, and they try to attack you.

SPEAKER_01: 22:13

That's right.

SPEAKER_03: 22:14

So the good guys have to find that vulnerability and patch it quicker, and it's a race. And maybe we live in interesting times, right?

SPEAKER_02: 22:21

Yeah. Uh let's not forget physical security while we're at it. If uh the Louvre can be breached with a lift and a broken window, that still matters in the midst of all this cybercrime. Um, Jeff, I'm gonna come back to you. Uh, as we think about capital markets and policy signals shaping the future of energy, where is capital flowing? We've got a lot of different places that we can build the grid out to serve this growing demand. It could be renewables, nuclear, microgrids. We talked storage last year a little bit. Uh give us a little bit of insight in terms of what you see shaping up out there. And and love your perspective on this too, Josh, more from a technological standpoint and trying to meet the timelines that uh our clients are demanding of us. Sure.

SPEAKER_04: 23:01

Yeah, it's probably helpful to take a bit of a half step back and say, how did we get here where we we're in this power conundrum of not having enough? It's been almost 20 years since the demand on the power system grew more than one or two percent. So now the demands are 50 and 100 percent all at once. And the the the universe is call it 80 percent in a regulatory environment and 20% in a non-regulatory environment. And so the speed at which we can fix this system and generate all this power is contained by a couple of things. One, the supply chain had narrowed down to a point to just be maintenance and incremental supply. We got to rebuild the entire supply chain: turbines, transformers, basic boring stuff, but you can't have power without it. And so that's kind of been a huge focus for the industry. And the second part of it is in a regulated world, who pays for this stuff? So if the if your capital expenditures, which they have in the industry, have now doubled, expectation is they will triple. So$250 billion this year just for utilities alone, um, that pass through in rates to customers like all of us is 15%, 20%. You do that two or three years, of course, you've got a huge political backlash. So the struggle for these, you know, entities to continue to build as fast as you need it, is in fact that. So the next stage is do you build these facilities outside the regulated world? And then who's the credit supplier for the capital markets to raise the kind of money we're raising? The obvious answer so far has been the hyperscalers doing contracts on the other side of the infrastructure, and then the capital markets will allow debt to be raised and equity to be raised to support it. Capital is not the issue right now. That's right. We have access to enormous amounts of capital globally. Uh as I said, in the last month, we've done two$40 billion debt deals for single facilities. You'd never would have thought that was an option. Um the, you know, the equity performance for the people in this industry has been, you know, 5x, 6x, 7x. So I just don't in in other cycles, capital was clearly constrained. I don't see that being the case here. It's the pace of the things I talked about, supply chains and getting the actual physical things and and labor. The one thing that you should hear from us too is they're not enough electricians, pipe fitters, carpenters, things that actually need to build this massive amount of infrastructure. And so that's gonna take some time to gear up as well. All of that, you know, is a is a challenge pile. Um, but I have no doubt that innovation, others are you know gonna solve that problem.

SPEAKER_01: 25:51

Yeah. Yeah, I mean, I think I think Jeff nailed it. I don't have too much to add there. What I'd say there is the the the conversation now, it's not a shortage of capital, but it's really like who's gonna own the risk. Uh so like while you don't see speculative building, which is indicative of us, by the way, not being in a bubble, um, what you do see is the need to proactively procure things like transformers, turbines, breakers, things like that. Those are very capex intensive. And you know, for any data center developer, even a well-established one, doing that off the balance sheet is just too much. And so you're seeing that sort of solid for in two ways. You know, you're seeing that like almost like pre-proactive project financing where you're getting some capital to put to work to pro to engage and get those things. Some of these things have two, three, four, five-year lead times. Uh, and then you're also seeing something that's really interesting and different over the last couple of years, which is you're seeing some of the customers, the hyperscalers in particular, backstop it. So, like before, as part of the lease, as part of the letter of intent and as part of the leases, you're actually seeing that these these uh uh the hyperscalers are putting money up off of their balance sheet to get ahead of these things if they're if they don't already have it within their procurement system already. So there's kind of like this game of ping pong a little bit between like, is it gonna be the data center company, is it gonna be the private equity investor, is it gonna be the customer? And really what you're seeing is on almost every project, it could differentiate. It's very customized. It's uh based on that solution.

SPEAKER_03: 27:09

And Josh, you and I were talking about this earlier. You know, if I go to the partner side, like an Exelon side, you know, we're sitting there, and another one of my learnings this past year was just how many data center applications and proposals Exelon was receiving. That's right. And one of them I could say pretty publicly was that we received in the last two quarters over 90 new data center applications for one of our operating companies, ComEd and Chicago. So 90 new data center requests that come into one grid. And so you multiply that across just what Exelon has, you start to see the demand coming in. But to your point, we were talking about like, how do we get them to give us upfront dollars? How do we make sure we're secure? We can't go build all that generation and all that delivery without, you know, kind of a really good partnership with the customer and making sure, because many times they're portending a lot of demand, but you know, if we're gonna go build it, is it gonna really be there too? So on the partner side, we've got to be careful with that. And we're um, you know, sort of a regulated industry, not sort of we are in a regulated industry. So we have to be really careful how we use the capex. So we have very sparing kind of capex to use to deploy. So it's a careful. Absolutely.

SPEAKER_01: 28:22

And then and then you you also see now, you see uh as a result of uh, you know, this the grid congestion, whether it be in the US, North America, or you know, more broadly, uh, you're obviously seeing folks, you know, move to a more hybrid strategy again, where you have microgrids and self-generation. And at those times, those present opportunities where those generators essentially will put up the money because they're interested in selling gas. And so like they'll put up that CapEx all day, essentially. Not all day, but for a good portion of it. And Jeff, keep me honest here, but they will do that because ultimately their goal is to sell gas over a 10, 15, 20-year period, and they can still make they can still make a nice return on that.

SPEAKER_04: 28:54

Yeah, I I I think again, capital is ubiquitous. We've all said that. I think the one risk that is yet to be the the capital markets are ignoring at their own peril here is that the average life of 90% of the cost of a data center is chips. And their average life is three to five years. The infrastructure is 25 or 30 years to be recovered. So this thing's gonna turn over. And what we've seen historically is when you finance long-term real estate with short-term mortgages, it doesn't end well. So there's gonna be some true up here in this process as we go. Hopefully, it's that the the entire economy grows sufficiently that the that solves all the problems and the credits become okay. But I think long term, uh I we're gonna be very sensitive about credit cycles.

SPEAKER_01: 29:46

I think that's right. And I thought the last thing I'd say is you know, this is also forcing, you know, mainly developers and data center developers and operators to uh to essentially explore markets that probably traditionally wouldn't be, you know, they wouldn't go for. So for us it'd be For example, we actually started our business out of Alberta, Canada, very similar to Texas, although the grain the grid was less constrained, but also it was a deregulated energy market and superabundant in natural resources. And so you'll start to see whereas before, again, you'd go in top-tier markets, very close to fiber, close proximity to the things that matter, you'll start to see, and you're seeing this in the US, you'll definitely can see this continue throughout least Alberta, Canada, where you have to go to, you're gonna have to shift to markets where you can actually do that.

SPEAKER_03: 30:28

Can I do a little quote follow-up for Jeff and Josh for a second? I'm really curious from the partner side. Yeah are we seeing SMRs as a future here? Because these companies are exploding in value in ways that I was I've been shocked at. And I don't know if you know much about small modular reactors and things. I'm curious on your viewpoint of that, because boy, that could solve a problem if it's really capable.

SPEAKER_04: 30:49

Yeah, I'm I'm sure I'm gonna offend somebody here, but um I've seen every SMR lab that I think exists. Uh there are two or three that are clearly have a lead at this point. I don't see a commercial SMR solution, David, till 35, 36, 37. That's right. So it's 10 years out. It it will be a solution at some point, but the cost to get that down the curve, it it's$200 a megawatt hour now versus 90. So that's going to be the constraint. We did just see, as I mentioned, this big announcement from the U.S. government with Brookfield to build large-scale nuclear again. We'll see how fast that goes. I mean, on the positive side, it's clear the administration is is rescinding climate control restraints, rescinding uh the time frames for permit approvals. You know, that'll get better. It's not going to take us 10 years to get through a nuclear process, but it's still going to take five or six or seven.

SPEAKER_01: 31:46

That's exactly right. And I would say, I mean, the the world, the world, I don't, you know, in reality, the world doesn't have any choice but to go nuclear, whether it's SMRs or full-on plants, just given like given the demand that we've been talking about today on stage. Uh what I'd say is, you know, in the short term, natural gas is that bridge.

SPEAKER_04: 32:00

Yeah, and I didn't really answer your second part, Brian, which is is capital flowing to renewables, storage, you know, the grid. It's flowing to everything. I mean, everybody is in a mode of all the above. We're extending the lives of coal plants, we're turning crypto facilities uh into you know data centers. Any access to reliable power is being turned into this growth in AI and data centers.

SPEAKER_02: 32:25

Yeah, I think one thing's for certain, too. This is another use case for AI. I think we're gonna find new ways to solve the power challenge with AI and and turn the AI against the problem set of how do we power itself. And that just creates new opportunity and a level of innovation that didn't exist again two or three years ago. So we're already seeing some of that in the world of storage. I am looking forward to seeing a year from now, when we've got you back up here, what the conversation looks like and how things have changed and and the market has evolved. We've got just a couple minutes left. Would love to hear from you guys just as you look at the audience here and you think about the the group of executives and leaders that we have across a whole slew of industries. Is there any guidance or a single point or bull to action that you would suggest that they think about applying into their world uh as enterprise leaders?

SPEAKER_04: 33:14

Yeah, man, the first thing I'd do is hire WWT. I spend uh I spend a lot of time uh with senior management teams at boardrooms, and and I would say my my outsider advice would be a hundred percent focus on integrating not only the technology innovation, but the energy security part of it and the uh just the profile of your energy strategy. You're gonna be a partner with energy companies now. You're not gonna buy power from them. This is not gonna work unless they you know they coalesce. And so um I think that's the skill set within your organizations to have regulatory relations, to have a technology person, and to have an energy person, not many of these organizations have all three. I think you gotta put all three together and turn them loose.

SPEAKER_01: 34:14

Yeah, I'm gonna uh I'm gonna piggyback on something that you know Jeff, I think, alluded to previously, uh, which is um whether you're you know, whether you're investing in physical infrastructure like like my, you know, like like us are like us a beacon, uh, or at the enterprise level, if you're investing in AI software and you're you know you're looking to get funding from a budget perspective. Either way, you know, our view is that like the returns in AI are are fairly like they're long, they're long term. So find a patient partner that has patient capital, uh, whether that's again coming off the balance sheet or from an investment firm, set the expectations that you intend to move fast, uh, but the return will come over time.

SPEAKER_03: 34:50

Yeah, maybe I'll just add on. Not a lot more to say. I thought Jeff and Josh said it well, but you know, education, education, education. Like we we've got to learn this problem. We've got to learn this opportunity, we've got to work together as a village, so to speak. And it's gonna take more than one voice here, one company to do it. And we need public-private partnership to make this happen. And what I've learned about energy in this my short time was boy, I mean, it's a federal level to a state level, or regional levels to a state level to a hyper-local level. It is a very complex regulatory environment. We got to clean that regulatory environment up. We've got to create more capitalism to that environment, more innovation to that environment. And that's gonna take a village to get done. But we've done that in other industries as well, so it can happen. So excited for that.

SPEAKER_02: 35:35

Yeah, I think just Jeff, on your point, um, one of the things that we've observed is in that MIT study that recently came out, many of you may have seen that 95% of use cases for AI have failed in the enterprise, 5% have succeeded. We believe there's a lot of gray area in that. But the the common theme that we've seen in this in the use cases that have succeeded and had significant positive impact on the business are where the right partner was selected and embedded in that process. And we would be humbled and honored to be a part of that journey with you as well. Gentlemen, thank you for the time today. Really appreciate this. Uh, they made it easy on me.

unknown: 36:10

Thank you.

SPEAKER_00: 36:11

Okay, in the end, what we heard today is unmistakable. Energy strategy is AI strategy. And the organizations that win the next decade will be the ones that treat power, security, and compute as a single integrated challenge, not separate problems to solve later. This episode of the AI Proving Ground Podcast was co produced by Nas Baker, Kara Kuhn, and Diane Swank. Our audio and video engineers, John Knoblock. My name is Brian Felt. Thanks for listening, and we'll see you next time.