The Case Against Off-Grid Data Centers with Tim Hade

What's actually happening is like you have developers throughout the country, right now, that are buying used aircraft engines from recycled military planes. And they're putting those on trailers and they're bringing them to locations where they're going to build data centers. And they're trying to build power plants out of these recycled aircraft engines. And people think that capacity is going to come online. [Music] Hello, my name is Jigger Shaw and I'm a clean energy entrepreneur. And I'm Jamie Nolan, a clean energy communications consultant. And this is energy empire. Very exciting, Jamie. We've got our good friend Tim Hade joining us. And in perfect timing, right, we just got a bunch of crazy dumb bills that got, you know, released by Congress. And Bernie Sanders had a video saying that we should ban all new data centers in the country until we figure this stuff out. I think Josh Holly and Richard Blumenthal basically said that all data centers should be expected to run off-grid, I think, until, you know, we could justify that they weren't going to add to people's electricity bills. Like, I don't know who's advising these people, but just crazy town. I mean, obviously both of those proposals are super unrealistic. I think very narrow chances of passage. But that being said, I do think it is reflective of the fact that there is rising and very loud opposition to these data centers and communities across the country. And while energy isn't all of it, it's certainly a big part of it. Just like at a dinner party at my house a couple of weeks ago, someone was asking me about rising electricity bills here in Virginia and saying, isn't it the data centers? And of course, I was like, well, it's complicated. But, you know, when these become kitchen table issues, it's amazing for those of us who work in the industry to see people that invested in what we're working on. And yeah, I mean, there's just a lot of nuance to it. And Tim, our favorite friend of the pod, so happy to have him on. I think if our listeners are not familiar with him, they're going to love him. He's the best. But he has a really interesting view as someone who has done a lot of off-grid, like microgrid projects, development throughout his career, through his company scale microgrids. That being said, I think he mentions a couple of terms that some people may be familiar with, but others may not. So I think we should just clarify. So can you just really quickly explain what's the difference between a battery, for example, that's in front of the meter versus behind the meter? Yeah, it goes, it's so hard to have these conversations without using some, you know, like of these terms. So when you do behind the meter, that means that that is hooked up after the meter that, you know, records how you use power from the grid. So that means that if you lose power, you can back up your house with that, that, that battery, or your business. If you install the battery in front of the meter, then it might be using part of your land, which you get rent for, but it's not available to back up your house. And so it's a grid asset, and it's part of the grid. So like, like the battery's right there, but like if you run out of power, you can't use it to, you know, turn your power back on. And so it's just controlled by the utility, and you're getting paid rent to have some of your land being used for it. Yep, absolutely. So like, for example, most people's rooftop solar systems and battery backup systems would typically be behind the meter on the customer side of the meter. So they essentially own or lease and control those systems, but in front of the meter is typically owned and controlled by the utility. So that's a really important distinction. The other thing that we didn't talk to Tim about, but I just want to mention here is that we have so many veterans that serve in the Clean Energy sector, right? Because many people served in Iraq and Afghanistan and realized that those wars were largely about oil, and you know, energy resources, and we're like, screw this. Like, I want to work in clean energy, so that we don't have to have people die overseas for something so dumb. And so like, I think there's just a lot of people who serve in the Clean Energy industry that serve in the military. Actually, no, for a fact, because I worked on this when I worked in the solar office at DOE, that there is a higher percentage of veterans in the solar energy industry than there is working in the overall economy. The solar industry has done a great job at recruiting and retaining veteran, a veteran workforce. There are a lot of workforce, like transition programs out of the military into the solar industry. So we're really proud of that and proud to have Tim on the pod to talk about his journey from the Air Force to our sector. So I hope everyone enjoys his conversation with Tim Hade. This episode is supported by S2G investments. If you've been listening to me, you know that I'm done talking about R&D. We have the technology. The real question is, can we deploy it at scale? When I joined Sunjeev Krishnan on the S2G podcast, we skipped the hype and talked about what actually matters, affordability, fit for purpose capital, and getting real infrastructure built. S2G investments is one of the only firms who brings that same deployment first mindset to energy, agriculture, and the oceans, connecting the dots across the real economy. Listen to the S2G podcast on your favorite podcast app now and start with my episode on energy affordability. We are joined by a friend of the pod, probably the first friend of the pod. You know, Tim Hade, super excited to have you here and we are just gonna dive right in because you are already one of us. Well, first of all, I'm putting that on my resume that I'm the first friend of the pod and secondly, I'm just here for Jamie. So let's let's do this. We're all just here for Jamie. Let's be honest. Let's tell you Simon yesterday, Tim, when I was describing you, that I was like he's like, he reminds me of Jigger. He's like, "A mini Jigger Shaw." Very different personas, but it's your authenticity and you're so uniquely you and your enthusiasm is infectious and you and Jigger and your optimism is just lovely to be around. So I've been really looking forward to you joining us. Wow, very kind. But the difference is that you can bench a lot more than I can. Those broad shoulders clearly like Cove and Handy. But I can buy off the rack where like you probably have to get all your stuff tailored. Well, look, I mean, I've done, I've said many nice things about Jigger and how much he's meant to me over the years and I have certainly learned a ton from him. But look, I think, you know, there's a lot of reasons to be optimistic right now, right? In terms of where we're at as an industry and sort of what's happening despite, you know, the chaos of the world, I think we're in a really good spot from a distributed energy standpoint. And, you know, look, more people need to be talking about that, right? I think we should, you know, this is kind of our moment and we got to capture it. And so, you know, hopefully there are more people than Jigger and I running around telling people this is a good idea. And, you know, let's make that proliferate. It only takes two of us, Tim. We don't need everybody else. Yeah, kidding. All right, let's go back to the beginning. So like, where are you born? Patterson, New Jersey. Patterson with one team. I mean, I feel like you've got that New Jersey boy look to you. Yeah, yeah, it's, you know, I am, I am a California transplant, but New Jersey born in Brad and definitely according to my wife that is part of my personality. But yeah, like I grew up in a pretty, I don't know, like lower middle class, you know, sort of upbringing in New Jersey. I loved it. I had a great education. I lived in a great community. I was surrounded by great people. My community was really, you know, a huge part of me growing up and sort of having some of the opportunities I had. My father passed away when I was young. And so, you know, it wasn't really, like the best nuclear family situation, but people in my community really helped me out, you know, coaches and teachers and neighbors and friends, parents and things like that. I'll kind of look out for me. And, you know, that, that sort of set me up to be able to go to, you know, the best undergrad college in the United States, the United States Air Force Academy. I mean, is it the best? I feel like you just went because it's a Colorado. I mean, like being from Maryland now, like, I feel like everyone thinks Annapolis is the best. I don't even think we need a Navy. But the, oh, my goodness, you heard it here. Okay. Yeah. I mean, look, you know, it's, as we say at the Air Force Academy, it's 7,258 feet above sea level, far, far above West Point and Annapolis. But yeah, you know, so I was in the military for 10 years. Wait, did you have to go through all of that stuff? Did you have to have like a congressman write a letter for you? Yeah. So who wrote your letter? that Bob Menendez, who's now I believe even jail was the senator who recommended me to the Air Force Academy. So if you're listening from prison Bob, thank you. And this was pre, I think, controversy. And then yeah, I was in the military for 10 years and did, I was an intel officer, but did sort of a variety of different things in, in, while I was in the military, got to see the world, but a lot of places in the world that not a lot of people get to see. And I think that's had a pretty, a pretty impact, like big impact on sort of how I think about how I want to spend my professional time and the type of stuff I want to work on in my, you know, post military career. And then when I got out, right, it was really at a time when the Department of Defense was, you know, talking about climate change as, you know, the number one national security threat, which is something that I believed. And I don't know. I think it's like a code champion with AI, but like, well, I mean, getting like fuel to forward bases was how we were losing the most, you know, soldiers in Afghanistan. And so like, so when did you discover electricity? Yeah. So I think, I think that was kind of it, right? So I mean, I've told you this story before, but like my initial, my initial, you know, sort of appetizer in the distributed energy was, you know, one of the biggest problems in Afghanistan was fuel logistics, right? So we had all these forward operating bases kind of on the border of Pakistan. And the main US base was Bogram Air Force Base, which is kind of on the, you know, western side of Afghanistan. And so we had to get diesel fuel from the outroom to the forward operating bases, which either meant you were flying diesel fuel and then dropping it out of planes with parachutes and trying to land it on the forward operating bases, or you were driving fuel convoys across Afghanistan. And it was the worst job in the military, right? And the casualty rates were preposterously high. And so there was sort of an experimental program at the time where they were looking at, hey, can we power forward operating bases with solar and batteries? And so we don't have to do as many fuel convoy runs to the forward operating bases. And we couldn't really figure it out. I mean, this was like 2007, 2008. So technology wasn't where it is today by any means. But even then, right, you could start to see the writing on the wall, which is like, oh, like battery technology is actually evolving really, really quickly. And solar technology is evolving really, really quickly. And so you could kind of see where this was going. And then, you know, that's sort of what sparked my interest in the commercialization of these technologies. And, you know, I've been working on it ever since. I love that. So you left the Air Force. Did you immediately start your company scale microgrids? Like, what was that? What was your evolution out of the military and into the private sector like? Yeah. So I went to grad school for two years, right after the military, predominantly because I had no idea what I wanted to do. I didn't even know anything about the private sector. And so, you know, I think that's just because you wanted to be in California, right? Yeah. I mean, it wasn't, you know, it wasn't a bad experience in grad school. I had fun. But the, but yeah, it was, it was mostly like kind of trying to figure out like, okay, I have this skill set that I've developed. How does that translate into the private sector? And I didn't really figure that out, right? So I was still kind of working on that. And then fortuitously, my best friend from growing up, who's currently the CEO of scale microgrids and is still my best friend in the world. He was a Boston consulting group at the time. And he decided he was going to leave Boston consulting group and go back and take over his family's diesel generator business. They had this like little diesel generator business in northern New Jersey. You know, that provided service to events and really their calling card was like, they did resilience projects for like big complicated projects in New York and the surrounding area. So they did like Goldman Sachs's data center and Verizon's 911 call center, like they did all the backup power for that. So my buddy decided he was going to go back and take over that business and try to turn it into a sustainable energy company. And we started with co-generation. And so he kind of pitched me on that and I was like, yeah, I don't have anything better to do. So let's go. And so we started a co-generation company and went through an acquisition there. So that business was bought by Central Cut in the UK. And so like sort of five years after that, I was really convinced that the future of distributed energy was solar and storage. And Central Cut just purchased a co-generation business and was very interested in me doing co-generation projects. And so I kind of decided, look, like I think I'm going to go focus on solar and storage and create this new thing. And so that was sort of the precipice of scale. And then Ryan came with me. And so Ryan and I started scale with Duncan Campbell and Pete Crosco and Ryan's Dad Howard in 2015. And you know, scale micro grids became my life for the next 10 years. And then I sold that in February as you guys know and I've been kind of a free agent ever since. So when I met you at scale, we had already seen a bunch of these co-gen projects. Like for those people, don't know what co-gen is, it's largely natural gas powered generators that then, you know, you take the heat off of the natural gas generator and you also use the heat. So that's the co-gen part and use it for cooling, for heating, for lots of things, right? And so one of the things that we realize very quickly at Generate Capital after meeting you guys is that like you could look at any system design and basically in like five seconds go, yeah, second work or yeah, it probably could work, right? Why is it that you and Ryan and the team had that level of like insight on like these types of projects where everyone else basically couldn't figure it out like without doing $50,000 worth of engineering work. Well, because he's brilliant, but I'll let him answer. No, like, look, I think I like maybe the answer is because I'm not brilliant, right? Look, I think, ultimately, right? Like what we figured out about co-generation really, really quickly was you basically needed the perfect customer in order to pull a project off, right? And what we saw from a lot of our peers in the space was they would spend so much time trying to jam round pegs in the square holes, right? And so they would go and they would look at these projects and, you know, maybe the, you know, HVAC system wasn't perfect or maybe the incoming service entrance wasn't perfect, but like we can work around that and we'll figure it out and we can engineer our way around this stuff. And, you know, look, when you, when you talk about, you know, bigger projects and, you know, we'll talk about data centers in a little bit, right? To some extent, you can afford customization and you can afford for those projects to be kind of snowflakes. But when you're talking about the types of projects that we are working on, which is sub five megawatt scale, like it either works or it doesn't work, right? And it basically requires sort of this perfect set of circumstances in order to work. And so what we started to do is we started to say, look, instead of trying to, you know, model these projects based on 10,000 different parameters up front, let's just come up with a heuristic that we can use to kind of, you know, take all the bad stuff, all the projects that aren't going to work and just dismiss them very, very quickly so we can spend our time working on the stuff that, you know, has a higher probability of working. And, and candidly, I don't even remember exactly how we got, you know, we developed that strategy. I think it was mostly trial and error, right? So it was just kind of, yeah, you know, I spent a lot of time in the early days working on projects that I would spend six months on and then we would find out like, oh, there's a $200,000 engineering upgrade that we don't have a budget for. And so like now that project's dead. So I think a lot of it was iterative and trial and error, but ultimately we came up with this ability to kind of, you know, source these projects really, really quickly. And that allowed us to spend the bulk of our engineering time and our money and our, you know, limited sales dollars and marketing dollars and things like that on projects that had a high probability of working. And that allowed us to have a higher conversion rate than most other people in the space and ultimately get more projects done. And again, I think, you know, that, that's kind of the same approach that I've taken throughout my, you know, career as a developer, right? Which is ultimately what matters is delivering a really compelling value proposition. Well, the fail fast and throw away the projects that are going to work. I mean, like what's interesting is that Congressman Caston, his father and his company, recycled energy development sort of did something similar, right? And they were very successful and sold. And I think their expertise was, yeah, exactly the same, which was they said, no, to like 99% of the opportunities that came in because they were going to make money on it. In your case, though, I want to dig deeper in this a little bit because like the biggest problem with co-gen is you know, is electric utilities like had your number and really wanted you to fail and they had these standby charges that were priced exactly where they needed to be to make sure your economics didn't work. But like, but then the big breakthrough that came through, I think, was with the cannabis industry, right? Because they were sort of like buying up 500 kilowatt interconnection warehouses and then trying to stuff two megawatts worth of lighting in there. And they couldn't get interconnected fast enough at two megawatts. So they were like, we need a co-gen system. So like talk about like, talk about the like role of a lock versus preparation, right? Because you clearly like knew what you were doing. But that was all like grinding it out business, right? Like the first big, like, you know, group that was just willing to pay you cash was really the cannabis industry, right? Yeah, I mean, look, I think the, you know, my entire career and distributed energy, right? I've been in search of product market fit, right? And I think ultimately, when we looked at the cannabis industry in 2015, we saw a few things, right? The first thing was we had a theory that Hillary Clinton was going to be the president and Hillary Clinton was calling. And then she was, she was going to like smoke cannabis all the way to town. Yeah, she was going to, she was going to kind of, you know, facilitate the legalization of cannabis on a federal level. And obviously that was a great hypothesis that worked out perfectly. Perfectly. But, you know, there, there, there, sort of that movement was happening, right? This is post, you know, legalization in Colorado and Massachusetts and California and some of these bigger markets were, you know, coming up and at the time it was like a $6 billion a year business, but it was also uninsurable and unbankable, right? And so if you think about, you know, how cannabis is still to this day, mostly cultivated, it has to be within state lines. And so if you're going to grow cannabis in Massachusetts, you're not going to do that outdoors, right? You have to do that in sort of a controlled environment, which is both legislatively and from a weather standpoint dictated that that's how you have to do it. So now you have this warehouse full of cannabis that's worth millions of dollars, but you can't ensure it. And so power resilience becomes really, really big for you. You're also, you know, at the time, and I haven't looked into this in a while, but at the time energy costs were like 40% of total operating costs for cannabis cultivators. So energy economics mattered a lot to them. And so the idea was basically like we have this product in a microgrid that adds resilience to your facility and can be done cheaper than, you know, the utility rate can reduce your cost and you care a lot about both of those things. And then oh, by the way, right, like there tends to be a big overlap between people who are growing cannabis and people who care about the environment. It wasn't like a one for one thing, but sustainability definitely mattered. And so, you know, again, we thought that that was really good product market fit. And so we started to work in that industry. And look, like there's, I think, a somewhat infinite, you know, infamous version of the energy gang where I went on the energy gang in 2015. And I like told this whole story. And then Jigger was like, you're fucking crazy. This isn't bankable. And he was right, right? And so like we made some progress in the cannabis industry, but all the time. And then I think we funded you a generate. Yeah. Yeah. Yeah. It wasn't a big market just because, you know, a lot of these deals we just couldn't get, you know, the credit, worthiness wasn't there. The collateral wasn't there, things like that. But like in building those projects, we made a bunch of huge technical breakthroughs, right? And so right now scale is doing some of the most pioneering work in the co-ocated data center space, right? And they're doing that by combining solar and storage and natural gas and power electronics to be able to support the needs of data centers and AI compute clients. And a lot of the reason they're able to do that is because of stuff that we figured out when we were building these cannabis projects 10 years ago. And so, you know, I think that's, you know, part of, you know, my big lesson learned from, you know, this whole experience I've had in the distributed energy industry, right? It's successes in linear, right? So one of the things that we talked about in the first episode of Energy Empire was around my journey at Sun Edison. And that when we sold the company in the end, right? Because of the 1X preference, you know, that we had ahead of us, like we didn't, you know, all make a lot of money off of there. I think that was a similar result for you guys. I'm curious, like if you had to do it all over again, would you have taken the money from Orberg Pinkis or would you have just sold the company then? I mean, look, it's a good question, right? And so, you know, I think what Jager's touching on, right, is something that a lot of clean tech founders deal with, right? Which is, you know, historically, the sticker price on these transactions is really, really high. How that flows through to founders and employees and the people that build the business is in, as big or proportion as one might think, right? And so, I don't know, right? It's created some social problems for me, right? Where if you like go on the internet and you look up, like scale microgrips is sold for whatever number, you know, they wrote in the press release, you know, people assume that like, you know, I can buy Monocito or, you know, whatever. And that's definitely not good. You're Bitcoin while it is getting hacked every day. Yeah, exactly. You know, and so look, like with that said, right? Like, I've made more money than I ever thought I was going to make and I can't really complain on that front. But, look, yeah, I think, I think, you know, thinking about the business model and the funding mechanisms for clean energy is, yes, something I wish I had known more about, right? At the time that we did all this stuff, I mean, look, I think part of this thinking, right, is it's all contextual. And so, yeah, I mean, look, the short answer is like, no, if I had to do it again, I would have taken the Warbird Pinkist deal again. You would have. Yeah, like, look, the reason for that is that, you know, if you go back to sort of 2019 when we closed our big funding round with Warbird Pinkist, it was essentially a $500 million line of credit. There weren't a lot of people that were willing to make that kind of bat on scale micro grids, right? Like, Warbird Pinkist was. And, you know, we didn't have, like, you know, 50 or 100 people, you know, running an auction process to try to fund our company. We had a few people who were interested in doing it. And then ultimately, you know, we thought that, you know, Warbird Pinkist and their skills and their capabilities were sort of the best strategic fit for us to take the business forward. And, you know, look, like overall, that experience was really, really good. And so, you know, again, I think, you know, ultimately, there are a lot of things I've learned since then. And I think, you know, the work that Jigger and Jonathan Silver are doing at multiplier, like, this is why it's so important because if you're the founder of a company or you're an early employee at a company, like, you should be thinking about how the funding for this business is structured and what that's ultimately going to mean to you when you have an accident opportunity, whether that's two years or five years or 10 years down the road. And I didn't know any of that. I didn't really think about that a lot when I was doing it. And, you know, hindsight being 50/50, do I wish I knew more about it? For sure. Do I wish multiplier existed when we were doing this? Like, yes. But, you know, again, I think, you know, part of this is, you know, now I know, right? And so, as we start looking at like the next things I'm trying to take on or the next challenges in my career, or, you know, how I'm going to sort of, you know, build the cap table for companies that I work in, or advise moving forward. You know, now I have that, you know, sort of experience to lean on and I, you know, know some things now that I didn't know then. Well, I'm not going to cry for you. It's all during Barbara. So, yeah. And I live in Santa Barbara. So it's not the end of the world. All right. Let's, let's transition to your hot take. All right. So like, let's go to the data center stuff. Yeah. So you're now in a place where our friend Elon Musk, not really our friend, but, you know, has, you know, turned on a data center in record time. It was like 22 days or whatever it was in Memphis, right? Using basically, you know, natural gas generators and then connected to the grid and then put in a bunch of Tesla megapacts, but it is somehow convinced the entire state of California and all of the Silicon Valley bros that you can run a data center off grid. Like what the hell? Like, like, set us straight. They use all of this technical knowledge that you've amassed to like, help us understand this behind the meter thing, off grid, data centers, whatever. All right. So, um, all right. So the basic idea, right, is we, there are a lot of problems with the US Electric Grid supporting the data center build out at the scale that Silicon Valley wants to build out. And so the idea that's become very, very prominent is we don't need to connect data centers to the grid. We're just going to build data centers with co-located power generation and then they won't be on the grid and then we won't need to use the grid for data centers, right? And I think right now, you know, according to the report, Jagers semi last week, there's 56 gigawatts of these co-located data centers in the queue. People who have lost their goddamn mind. Yeah. So like, my, let's credit the author of that report. That's Michael Thomas at CleanView. We love his stuff. I love his data visualizations. Of note, in that report, he put out, there's a chart that shows that data center power plants would use five times more electricity than New York City. 46 data centers with a combined capacity of 56 gigawatts, gigawatts, plans to build their own power plants. They would use five times more electricity than New York City. Yeah. So, so look, I guess like my hot take, right, is that my best gas is 80% of those projects won't get built, right? And so like, is 40% of them are so dumb that they're just going to finish the project? No, no, look, I like, I think they're, all right, so the way to think about this, right, is the thing people miss is is putting a power plant next to a data center is like 20% of the work, right? The hard part is the power electronics. So like the way to generally think about this is it's the difference between building a power plant and building a utility, right? That's the difference between building like a natural gas power plant and a co-located data center. Is a co-located data center is its own utility, right? And the hard part is in the middle, right? And so when you think about like, I guess the best way to like summarize this is like data centers are essentially like power electronics, divas, right? They have all these like weird quirks and weird characteristics and they have in-rush currents and harmonics and these fast ramp rates and like non-linear current. All these like crazy electrical characteristics they make it really, really, really difficult to load balance locally, right? I'm gonna quote you in your paper that you sent me which I know has not been published but we might have to force you to publish it. You say when people say we need more electrons what they're actually saying is we need more short circuit strength, more reactive support and more tolerance for fast ramps. What does that mean? - Yeah, so essentially what you're talking about here, right? Is you put like a data center is essentially a shell that hosts these GPUs, right? And how GPUs behave in a training capacity is something that candidly everyone is still learning about, right? But what the initial data shows is that some of the power characteristics of these things are crazy, right? So they'll go from like using five watts to one watt in a millisecond and the system has to be able to respond to that. So if you connect data centers to the electric grid which is how we've connected data centers to the electric grid forever, that's the only way we build data centers until recently, right? You benefit from the inertia of the grid which means you can have these like momentary lapses in frequency or harmonics or in rush current or wherever the case might be and the griddle backstop you. - And that's what you mean by short circuit strength? - So the short circuit strength is basically that if the data center has a generator on site, right? And that gets out of phase with the data center characteristics it'll force that generator to shut down, right? It'll have a short circuit trip and it'll force that data center to shut down. If you're connected to the grid and you have that same system act, that same system characteristic happen, then the grid has the inertia to be able to backstop that problem so you can keep the data center online. So ultimately what you're talking about is system uptime which matters a lot to the data center providers, right? And so like stepping back from this, the high level way to think about it is building a co-located data center is a massive power electronics challenge. Now the reason I think 20% of those 56 gigawatts of plan projects are gonna get built is because there are companies that are good at doing this, right? So like my former company, scale micro grids can build co-located data centers and Sheldon, Kimber, Intersect Power can build co-located data centers and like the Crusoe guys can build co-located data centers, right? There's like, - Well, maybe, like I think that like, I think scale micro grid could get to three nines reliability, which is not five nines reliability, right? - Sure. - And then like Sheldon is not really building off grid data centers, he filed for interconnection permits for his solar projects five years ago in Texas. So those are gonna be grid connected systems with a hybrid like power plant behind it, which like I think is the right way to go, right? Is to have behind the meter resources, whether it's natural gas or solar, but to have a grid connection, right? Like he's not foregoing the grid connection. And I think Crusoe like God love Cully and company, but like I think they've like, you know, created like two big like power quality issues on the Uricock grid or any. - Yeah, I mean, again, right? I think so the question of like, is it possible to build co-located data centers is yes. Is it a good idea? Is it different question? - All right, so the next thing I'm gonna quote you on is you say, but co-location is not a hack around the grid. It is the decision to become your own utility, not just in ownership but in technical obligations. Their grid interconnection point is not merely a wire. It is access to an infinite bus, a massive system that provides voltage stiffness, frequency stability, high fault current and an ocean of inertia. When you step behind the meter, you don't just change who owns the electrons, you change the physics of the power system. Correct. All right, and then the next thing you say is, none of this is to say co-location is foolish. It will work in some places for some projects, especially when executed by firms at know how to build and operate power plants. But the idea that co-location can meet, can scale to meet the entire wave of AI load growth, assumes a labor market is supply chain a permitting environment that does not exist. Yeah, I mean look, right? That's, so again, right? Like there are firms that can build co-located plants, right? But it's not a lot of firms, right? There aren't a lot of people that know how to do this. And then on top of this, like the people that you talk to at those firms, like the guys who run the engineering teams at those firms, like they are so focused on every single piece of equipment that's part of that ecosystem and making sure that that piece of equipment is like a tier one product that's been tested and works well within the context of the rest of the ecosystem and all this kind of stuff. So there is a way to build them. Now let's look at what's actually happening, right? What's actually happening is like you have developers throughout the country right now that are buying used aircraft engines from like recycled military planes. And they're putting those on trailers and they're bringing them to locations where they're gonna build data centers. And people think that capacity is gonna come online. - So how much do you have to hate your fellow neighbor to do that, right? So the next like, the next like quote here, the core mistake in the build a plant next to every data center worldview is treating the grid as a static obstacle. The grid is not just where power comes from. It is the reliability machine that makes modern electricity usable, even co-located projects lean on it for backup, for trading, for reserve sharing. And often for the very fuel and maintenance ecosystem that keeps generation online. - Yeah, right, it's just like the grid is the answer for data centers, right? And so like we kind of have this problem backwards. So the idea that basically like data centers are the things that should come off the grid is the exact opposite of what we wanna have happen, right? And so first of all, from a power quality standpoint, the best way to build the data centers to hook it up to the grid. And the only reason you should be messing around with co-location or anything else is if you can't hook the data center up to the grid, right? So then the question becomes like, well, can we hook up more data centers to the grid? And the answer is yes. - Yes, yes, right. - I mean, I feel like this is like in part, like people who move to Florida and then decide to live in a private gated community with private security, sending their kids to private school and then voting for like, you know, cutting taxes for everybody who lives in Florida. It's like, well, just 'cause you have the money to build a private neighborhood with private security, with a private school, doesn't mean that you don't like want to be a part of an overall system that actually has a robust like public safety network, public school, like public ecosystem. Like I just feel like, like in some ways, like it's not only will it not work. Like I don't think even scale micrograd or alpha structure or China electric or whatever is gonna do better than 99.9% reliability. But like on top of that, it's so frickin selfish. - Yeah, I mean, I think it's even dumber than like the private school, private community example, right, because presumably those people in Florida are paying for a premium service, right? But like what the hyperscalers are paying for with these co-located data centers is something that's not as good as the grid, right? And so like the fundamental way that people should be thinking about this is there is a symbiotic relationship between a good grid and building a lot of data centers, right? And there is a way to connect a lot of compute to the grid in a way that not only allows hyperscalers to build all the data centers they want, but also lowers retail rates for consumers by increasing grid utilization, right? And that's what we should be focused on, right? And the way you do that is you make grid utilization a metric, so we can talk about the fact that the state of Virginia just did this, right? And so you're forthcoming with everyone voting for it. - 99 and nothing in the house, right? It's like something that there's bipartisan agreement for, which is like you know what we should measure is grid utilization. And when you measure grid utilization, what you're likely going to find is a lot of utilities are operating their distribution systems at 40 to 50% capacity factors or 40 to 50% efficiency. And then the question is like, well, is there a way to increase that efficiency? Because if we can increase that efficiency, then we can move the same number of kilowatt hours through the grid for the same fix, or more kilowatt hours through the grid, for the same fixed cost, which lowers rates. And the answer is like, yes, we can do that, right? How do we do that? Load flexibility generally, right? And that's generally batteries. It's taking the 2% of commercial sites, which is like 150,000 locations across the country, adding 250 kilowatts to five megawatts of batteries at each site, right? And so the customer gets resilience. It's like a community lighthouse if you put it in a school, right, or a church or whatever. They get revenue for hosting it there and like people pay rent or whatever or they have a free battery. There, and the data center companies pay for the capacity from the battery. So in the end, the data center companies are actually paying for the batteries, right? And so it's a win-win, right? And like, one of the quotes you have in here is like, if you're a community starting at a data center, you're basically saying, trust us. We'll bring you zero permanent jobs. That's not really always that compelling. But if you say we're paying to heart and local businesses and public buildings with batteries and backup capability, that's tangible. And it's a way to convert local opposition into local benefit and to convert a permitting fight into a deal. Yeah. And look, I think this is important, right? Because part of this argument is about energy economics, right? And just on energy economics, the answer is like, don't make the data center flexible, make the grid around the data center flexible, right? And the plan to do that is like measure grid utilization and then improve load flexibility via the deployment of batteries, right? So that, like, if you just focused on the energy economics issue, like, that's the answer. But like, what's actually happening in the data center world right now, right? Is there's two problems. There's speed to power, which is how long it takes you to get interconnected to the grid. And there's speed to permit, which is, can you get a permit from your local AHA to build the data center? And increasingly, the bottleneck is becoming speed to permit, right? And so like, I didn't know this until, you know, our knob and the fine folks that deploy action gave me a PowerPoint presentation. But they went out and they like collected all this poll data. And the polling on data centers is awful, right? Like I think they showed us the 80% of the people they talked to were opposed to data centers being built in their local community. Yeah, Microsoft is still waiting for permission to start construction of their data center in Wisconsin, which is like a state that like features Ron Johnson. And like, you know, generally are like, yeah, do whatever the hell you want to do. We don't care. And like, and like the people of Wisconsin are just pissed. They're like, you guys just try to railroad us. It's kind of make people really like data centers is putting a bunch of old airplane engines right in their back yard and pumping out a bunch of local air pollution. That's going to be great for everybody. Yeah. And like, you know, I think one of the things that's like kind of bothered me in our conversations with the hyperscalers, right, is like, you know, their view on this is basically like, well, the local people just don't understand, right? Like, they don't understand that like actually data centers put like downward crashes on rates and they're misinterpreting the data and all this kind of stuff. It's like way more complicated than that, right? Like when you talk to these people and like, you know, I had this great conversation with Arnav about this, but like the feedback they were getting from local community members wasn't like ignorant feedback, right? It was like, we, first of all, like this isn't creating jobs for us, right? It's creating a bunch of construction jobs for the 18 months it takes to build the data center and then like basically no jobs, right? Um, also like, we don't really know how we feel about AI and how fast this shit's going, right? And we have some concerns about that. We know that's that goes to cost over build, right? Like, I mean, if you, if you say yes to like Duke Energy's $103 billion dollar proposal to like build out their grid, right? And then the data center load doesn't show up then someone still has to pay them back, right? And so like part of the reason why great utilization and batteries are so smart is that it's like one 10th the cost and like if, if the data center load doesn't show up, which that's the other point that I think we need to make, which is that like the data center companies are combining training with inference, right? Like, and like, it's very clear that you need a thousand megawatts of chips in one place to do the training. Like doing training in a distributed fashion is like super complicated and hard, right? I get that. So the new people tell me I, I don't know that I get it, but I, I believe them when they say that, right? Yeah. But like that number, according to ontropic is only 25,000 megawatts that they need for training, right? But they're asking for like a hundred thousand megawatts of new data centers in these 1000 megawatts blocks. Even in video came out with a format like a month ago saying we could build about your five megawatt data centers and fit them nicely into the existing infrastructure for inference. And inference is like when you pull up Jatchy BT and you just like use it, right? That's just inference, right? That's not the training. That's just using it, right? And so like, like part of what paces me off is that all of the data center companies are deliberately not answering the question of like how much, how many megawatts you need for training. And I think the answer is 25,000 megawatts, but like they won't answer it. And so now we have this feeding frenzy of developers who are going all over the country, lying to people on a regular basis, trying to get them to give them permission to build the data center for a data center. They probably aren't even going to have permission to build because we don't need it for training. Yeah. I mean, look, I think that like everything you said is right. Although I would I wouldn't say like I don't think anyone at the tech companies is lying per se, I just think they don't know, right? And so like one of the things that was fascinating when I was like, right, writing this little paper for you guys, right, is so like LBNL did this, their most recent forecast, right? I wrote this down. So that's our Lawrence Berkeley National Laboratory. Yeah. So there are projection for electricity consumption in 2028, right? Which is two years away is between 325 and 580 terawatt hours. So like most of the time when you looked at this, like when I was growing up and you would look at these spreads, that spread would be like, it'll be between like 30310 terawatt hours. And you'd be like, ah, all right, like narrow band. Like now it's 325 to 580, which basically means like no one knows anything, right? But people don't know how much compute and how much electricity AI is going to use, right? It's anyone's guess. And so when you think about this, right, it's like part of the planning process needs to be planning for uncertainty, right? And so if you build a one gigawatt power plant in the middle of Louisiana and it turns out you were wrong about your compute forecast, you have a one gigawatt facility that's fucking useless, right? But if you take that same one gigawatt and you deploy that in the form of, you know, 10,000 distributed batteries around a distribution network and it turns out you don't need that gigawatt, now you have a better network, right? And so ultimately, right, like part of this is like understanding that there's so much uncertainty around the load forecast and building a flexible infrastructure so that we can adapt as we need to, right? And that's one of the things about distributed energy, right? We build it from the bottom up, which means it's inherently more flexible than building it from the top down, which means if we're off by 10 or 15 or 50%, right? We can adapt in real time and either ramp up or ramp down appropriately, which you just can't do with with bigger stuff. And so again, I think like people are just thinking about this the wrong way, right? The idea is not like take data centers off the grid. It's build a grid that's more accepting of data centers. And if we do both of those things, right, we're going to have a better grid and lower rates and we're going to have all the compute power we need. And again, right, I think sort of spending time on alternative strategies at this point is just kind of a waste of time. Well, the batteries end up becoming like an emergency reserve, right? Like, I mean, like, who doesn't want an emergency reserve? Like it's great. Yeah, right. Like if you do it like that, right? And like the load doesn't manifest, right? Now all of the sudden you have this load flexibility that you can use to be really attractive to like a next generation robotics company, right? Or a manufacturing facility that wants to stand up, right? You can now offer people cheap, flexible power in spots where they want to build infrastructure. And so again, right, like it should be a national mission, right, to upgrade the grid to the 21st century. We've been talking about that forever, right? That was a big part of your industrial policy in the Biden administration. Like we should do that. And the load growth that data centers are creating is providing the vehicle for us to do that. And the fact that we're not doing it is, I don't know, annoying. But we're going to be good. We are. We are. We are. We are. There are definitely bright spots, right? But it still seems like, I mean, who does this message need to reach? Like is it the hyperscalers? Is it the utilities? Is it like, you know, who's creating the permission structure here that we need to like make this a reality? I think it's largely the governors. There's over 30 governors races this fall. Every one of them is going to lose their governor's race unless they have an answer to how they're going to manage all this data center stuff, right? Like people do not want to be trampled by hyperscalers. And so the governors are going to have to ensure that they're going to protect their voters from being trampled by the hyperscalers, right? And they're going to like have to ensure that all of these best practices are followed, right? And so like I hold our elected officials. accountable to making sure that they're not just chasing the $25 million a year of property taxes, that the hyperscalers are going to pay, but they actually are making sure that it is a net positive for the community in the way that Tim outlined, which we passed as law in Virginia. But I think, you know, lots of places around the country are looking to do the same thing. Well, then you can call up Tim Hayden. He's their guy. He's got the playbook. No, look, like, like, Jamie, I think I think jiggers, right? Like the governors have a big responsibility. But like, look, I think this is candidly why like leadership matters, right? And I think ultimately, like, I don't know who's going to lead the charge on this. It doesn't seem like it's going to be secretary right, right? And, and, you know, ultimately, I think, you know, what's happening in DC right now is really, really fascinating because, look, I think candidly, right? Like, and Jigger and I have talked about this. I mean, you know, part of the issue here is that, you know, for the hyperscalers and really we're talking about seven companies, right? You know, those folks have always been big procurers of energy, but energy has never been a core competency of those businesses, right? So if you go talk to like the C-suite people in these in these organizations, like they're new to energy, right? They're thinking about this stuff for the first time. And I think that's where you're seeing the most progress, right? Like, they're pretty quickly coming to the realization that like, load flexibility matters, right? And we need to figure out how to increase load flexibility and work with the community. Our Tyler Doris wrote that paper they got hired by Google. For sure, right? Like every Google tiring everyone that we know, right? And we got to get invited to the next Google party. Yeah. I know they're like packing the bench over there and I'm like, I'm excited to see what they're going to do because it looks exciting from the outside. Well, look Tim, it was such a pleasure to have you on. You are most certainly going to be on, again, probably as like a co-host when like I'm on vacation. But like, you just talk shit about Jigger the whole time. I mean, like, that's the best. But like, I am just so happy that you agreed to come on. Like I said, you know, front of the pod number one. But thank you so much for being here. No, I'm so proud of you guys. This is awesome. And I can't wait to subscribe and listen to every episode. This is the best. This is Tim. Wow. I love talking to Tim. He is truly a tour de force. It is so hard to like come out of a conversation with him without learning so much. But we covered a lot, right? We covered his background, right? Why he's such an expert. But also the fact that he doesn't think most of these off-grid data centers are going to get built. Yeah. I mean, Tim is the guy on this issue. And I love that in preparation for our recording, he just like shot over a very heavily researched eight page position paper for us on this topic so that we could study up and prepare. Clearly he has a perspective. And I just wonder when the hyper scalers are going to be knocking on his door because I feel like he's got it all figured out. Well, I feel like they're already knocking on his door. I mean, when you think about his trajectory, and it's one that I think we heard a little bit from Dean Solan about, but it's certainly my own trajectory at Sun Edison and then at Generate Capital, which I started with Scott Jacobs and Maton Friedman in 2014. A lot of this is preparation, right? And then you're waiting for luck, right? So like in his case, you know, he was trying to do the right thing for like farmers and agribusiness and other types of companies that wanted to build these sort of co-gen microgrids. It was slow going. I mean, I was there with him. Every step of the way we were funding a couple of those deals. So it was slow going. But then the cannabis industry took off and there was a lot of good business there. And then, you know, Warburpinkus came in with a lot of money. And then, you know, EQT finally came in and said, we're going to pay a lot of money because we have several big investments in data centers. And we want you guys to build off-grid data centers just for our data centers. And so they paid a lot of money for, for scale. Like, I mean, it's unfortunate. It sounds like he didn't become like filthy rich because he, if anyone deserves to be filthy rich, it's him. But it's actually a clear story of like what happens in this industry is like timing is a lot of it. But also, you know, what your financial structure and capitalization looks like is part of it too. I mean, Warburpinkus clearly negotiated a deal where they got a lot of the spoils from the EQT, you know, pay out and the employees got less. He didn't seem bitter about it. He seemed very at peace. I mean, if I lived in Santa Barbara, I would probably be at peace too. What was it when you first met him more than 10 years ago that led you to like kind of take him under your wing as, you know, your mentee? And like, what did you see in him? And kind of how has that relationship evolved? Because I know that that's, you play that role for a lot of folks in the space. Well, I think you know that I really just can't tolerate assholes and there's just so many extraordinary people in our industry that are assholes. And so I just don't like, I just don't follow up and hang out with those folks more than I have to. But I think the other thing that we just talked about with his exit right was people who are grateful, right? And some of that's like, you know, growing up in a small town like you and I did in, you know, Illinois for me and Maryland for you. But like, you know, for him and Patterson, New Jersey, I think there's, there are folks who are actually grateful for the position that they're in. And I think that that comes through in ways that like are, you know, appreciative of the people around us who make our lives easier and, you know, like in all the, you know, ecosystem that supports us and like all that stuff. And like I just, I have very little tolerance for people who just think that they're the center of the universe and that everything revolves around them. And Tim is definitely not one of those people, which is fantastic. And it's very obvious that the guy is brilliant, like just like just like gab's backing, like how brilliant he is. And enthusiastic. And he makes you feel like you, you leave, I think after talking to Tim feeling empowered, like you're like, let's go build some shit. Like, yeah, we can get this done. I think that's, that's his best quality. I think Tim is evident of the fact that this is, this is the day for distributed power. And, and now we, as he mentioned, need to be thinking in a billion dollars scale and said a million dollars scale. And that's, that's the theme of the year. That's, that's what we're here to explore on energy empire. So I'm excited to talk to more change makers like him that are actually really building big shit. Yeah, what do they say? If not now, when, if not us, who? Absolutely. As always, thanks for listening to energy empire. We are produced by Simon Bergbaum review us on Apple podcast and Spotify. It really helps if you leave a comment. That's how people find us. And I don't actually know why, but that's how the algorithm works. And then I think my son told me that you have to smash the subscribe button on our YouTube channel. So smash it. You can also reach us at our website, which is energy empire.fm.