← Inevitable
Inevitable
Accelerating Data Center and Renewable Energy Siting with Paces
00:49:09
James McWalter is the CEO and co-founder of Paces, an AI-powered software platform that helps clean energy project developers identify and evaluate the best sites for new projects. Paces streamlines the process by integrating interconnection, environmental, zoning, and legal data on land parcels—key factors developers must navigate when selecting viable locations. MCJ is proud to be an investor in Paces through our venture funds. In renewable project development, one of the biggest challenges is quickly pinpointing sites with the highest likelihood of reaching operation efficiently. As James explains, this process depends on multiple factors: the capacity of the local grid, whether upgrades are needed to connect a project, the feasibility of permitting, and more. And as we discuss, the recent AI-driven data center boom is intensifying the urgency and complexity of each of these considerations. Paces’ platform is being used by both clean energy developers and data center developer...
Loading summary
Transcript109 lines
- [00:00]
A
Today on Inevitable, our Guest is James McWalter, CEO and co founder of PACES. PACES is an AI powered software platform that helps clean energy project developers to find and evaluate the best sites, helping them navigate interconnection, environmental zoning and data in land parcels they are considering for new projects. MCJ is proud to be an investor in paces, but via our venture funds in the world of renewable project development, one of the largest challenges is in quickly identifying sites that have the highest likelihood of getting to power quickly. As James shares, this process is dependent on multiple factors from how much capacity the local grid has, to whether upgrades are needed in order to connect a project to the grid, to the ability for a site to get permitted, and much more. And as we discuss the recent AI driven data center boom is increasing the urgency and challenges associated with each of these factors. Pace's platform is being used both by clean energy developers and by data center developers who are trying to get projects built as quickly, cheaply and cleanly as possible. A few months ago Pace's co authored a white paper with Scale Microgrids and Stripe Climate titled Fast, Scalable, Clean and Cheap How Off Grid Solar Microgrids Can Power the AI Race and it generated quite a stir. In the paper they contemplate whether forecasted load increases driven by AI can be met with off grid behind the meter projects. Oh, and James also hosts his own podcast called Build Repeat, featuring conversations with project developers and others who are actively building and deploying clean energy projects. And with all that said, James and I discuss this, his background, the origins of the company and much more in this episode. But before we start from McJ, I'm Cody Sims and this is inevitable. Climate change is inevitable. It's already here, but so are the solutions. Ships Shaping Our Future Join us every week to learn from experts and entrepreneurs about the transition of energy and industry. James, welcome to the show.
- [02:30]
B
Good to be here Cody.
- [02:31]
A
I'm looking forward to learning from you about the work you're doing really at the front end of the development process for clean energy projects, data centers, everything that a lot of people are talking about these days and trying to understand more about. So maybe take it from the top. What is Paces?
- [02:48]
B
First of all, really appreciate you taking the time with us today. Paces we are a software platform to enable the siting and due diligence of any sort of large scale power project, whether that's on the generation side like large scale solar battery storage and wind projects, or increasingly on the load side big data centers and similar projects that have many tens or hundreds of Megawatts of power requirements.
- [03:10]
A
And how did you land on this project being the thing you wanted to focus on?
- [03:15]
B
We met when I was working on a previous project. The first time we met a very long time ago. So I'm originally from Ireland. I grew up on a farm, which is actually relevant for the pathway. And then out of university, working across multiple industries, from finance and then later into industries like market research, customer success, generally applying machine learning and data techniques to different business problems. Had a great old time. I was working at big companies, small companies. Saw the startup life as an early employee or founder at a couple of failed companies. Started getting stressed about climate change in 2019, I was like, look, nobody I know who's working on climate. How dare they? Then I have the cognitive dissonance of how dare I? I'm not doing anything either. So started thinking through what would be a potential problem around climate that I could spend time solving.
- [04:00]
A
Your background was classic. It looks like Enterprise SaaS and a lot of basic tech. Though it looks like you worked in a couple companies based in Mexico. What a fascinating experience.
- [04:08]
B
Yeah, yeah. So I lived in Mexico for nearly five years, waiting on my green card. These are all US based companies, but I was down in Puerto Vallarta, Mexico. It was not supposed to be nearly five years, it was supposed to be one, but it was a great time. But honestly, like, a lot of the reason why I even thought that I could start something in climate was Covid because all of a sudden you had this situation where I didn't feel as isolated from the different startup communities in Silicon Valley, et cetera. Everyone was on the other side of a zoom call or a slack community. And so all of a sudden I'm at the same amount of access to folks and obviously Covid very devastating and so on. But in terms of access to smart folks who are working on climate, it was transformative.
- [04:45]
A
McJ was very much born of the COVID sort of zoom pandemic world. It resonates with me for sure.
- [04:52]
B
So yeah, so I was like, okay, I want to work on something in climate. And what did I know? I grew up on a farm. I saw firsthand what it was like to do different farming. But interestingly, my mother was actually head of organic standards in Ireland. And so we converted to Organic in 1997, saw that transition and I was like, oh, agriculture has a massive carbon footprint. How could that be decarbonized? And this is also like the period which you know even better than I do of everything was carbon markets, everything was carbon credits this is like the very, very hot topic. And so I was like, oh, let's get in on that.
- [05:22]
A
I feel like food and ag is the gateway drug to working in climate stuff. It's the thing people kind of gravitate to first I think because it feels less wonky and scary than energy when in reality it actually is pretty wonky and scary too. It's hard.
- [05:36]
B
I jokingly said it's somewhat similar, but I would say everyone goes through a kelp phase where they're like oh, the scale of the ocean or the scale of soil or whatever it might be. And at the time I thought energy was so boring. I was like, I don't think we just have to build a lot. Why don't we just do that? We don't have to invent anything funky and cool. But yeah, so I was like, all right, let's work on something on ag. And was interested in basically how do you change farming practices? Started iterating on an idea there just on my own. I didn't have anybody with me at the time and then very rapidly realized or even re remembered that selling the farmers is closer to B2C than B2B. And so all that enterprise B2B sales skills were just very different when you're trying to sell generally on fairly emotion based type selling for folks who are very obviously protective of the family farm and all those kind of things.
- [06:21]
A
And so then you did what, what was the sort of moment into moving into Paces?
- [06:26]
B
I think the other piece of this was I was also actively co founder dating some folks and was really interested because I had always spent time selling very strong data machine learning solutions to folks. I was like, oh, I wanted to work with folks who had a lot of that skill set. I met a few people, one of which became co founder of Paces Charles who ended up with somebody more impressive, more wonderful to work with. And it was actually this funny group where I found out later that there's like seven or eight of us all co founder dating each other. It's like this like poly situation of folks potentially trying to collaborate on climate related projects. And as part of that I was working with this particular gentleman. He was really strong, he's now a founder of a company called Windscape. Jason Yasinski. I was working with him on a couple of ideas and he was like stop looking at all this carbon stuff, we should look at the energy system. And he was at the time thinking more from a trading perspective which I was like less interested in. But it did drive us to start having Some conversations with some project developers. And that reminded me as we were talking to product developers about some of the challenges they were having. That back when I was trying to sell to these farmers, some of the farmers were complaining about solar developers pinging them about land and this kind of thing. So I started to form this idea around, okay, it's very hard to build these projects. This industry is not using anything close to the cutting edge of different types of software and data solutions. There might be something there. And so I was like, okay. And this is exactly around the time that I met Charles. He was very, I think, really interesting for the time that we're in right now. He was literally working on optimizing models within Facebook and Meta's data centers for lower power use. It's called sustainable AI. Very hot topic. Of course today with deep SEQ and everything else that's happening. But he was trying to start that as a startup and he was like, oh, but nobody was interested in optimizing power use in data centers at the time, right? This is not that long ago. It's 2021. Whereas I had a couple of renewable developers interested in struggling with particular set of problems. And so we were like, look, if we can sign a couple of letters of intent, that's a good proof. We had already started working on a few projects together just to make sure that we could work together and work well together and yeah, sign a couple of letters of intent. And that was our signal to start fundraising and start building up paces.
- [08:24]
A
You guys did yc? I think shortly thereafter we did as well.
- [08:28]
B
Yeah. So, and this is one of the funny situations. So we applied to yc, didn't get an interview for that batch, then raised a pre seed and then got into yc, usually the kind of other way around. So yeah, so we did YC Summer 2022, entered the batch, zero customers, a very basic MVP coded up, ended the batch with about 8 to 10k monthly recurring revenue and a product that was just about working. Even though if somebody clicked a button on the product, our co founder Charles, he would literally get a text message. Then he had to go run the thing manually and then he would send it to me in an Excel spreadsheet and I would clean the data and then I'd send it back to the developer through a like an email address that looks like it was like a notificationsaces.com, but it was actually just like an email address that I would log into and send it. And it was literally the first two months of the product out Was like something that was stringed together.
- [09:15]
A
I love it. The AI was actually Charles in the background.
- [09:19]
B
The AI comes way later. Like the AI was smart people trying to do manual things in the back end. As it should be.
- [09:26]
A
As it should be. You got to learn the problems before you just trust the computers to solve them. What did you learn early on that these renewable project developers were struggling with?
- [09:36]
B
Just for the folks who are maybe less familiar, if you're trying to build a solar farm, the basic workflow is I try to have a specific strategy. So I want to build a 10 megawatt solar farm in upstate New York. I then try to find sites that meet certain criteria. So they need to be able to be large enough, just enough land to put the solar panels down. Second, they need to be near the grid infrastructure that you can get your power on. And then most importantly, you need to be able to get enough capacity on the grid infrastructure to accept your power. You as a private developer, you are responsible for any upgrades in the grid to get your power onto the system. And the last thing is environmental review, permitting restrictions, all those kinds of things. And so what was happening is developers were finding sites, they would meet some set of criteria that looked reasonably good, but additional de risking was just super expensive. So they would start buying or leasing land and then going around this development pathway with projects or potential projects. And at the end of a multi year process, only one in five of these projects would actually get built, the other 80% failing primarily because of power interconnection and permitting. We say power and permitting is the big issues.
- [10:35]
A
This is the famous interconnection queue and these like long seven year cycles where a project may be ready to go and it just quite literally can't get plugged into the grid.
- [10:43]
B
Absolutely. So the way United States works is you literally a queue. The utility cannot privilege one project above the other. Even if one project technically would be better, they just are not allowed. It's a regulated monopoly. So what happens is they're like, I want to build a solar farm on the substation or nearby. And. But if another developer got there ahead of you and tried to build a solar farm or battery storage project on that substation, they are privileged for that initial capacity because they're ahead of you from a time perspective. And so a lot of times folks are taking queue positions, right? Developers and trying to hopefully some of the projects might fall out. And you all have this kind of situation where because utility has to take every single project as it comes. A ton of months of analysis are occurring on projects that are actually low likelihood of ever getting built. And this drives a lot of conflict between the different stakeholders because utility is like oh, these project developers are bringing nonsense projects to bear. The developers are not right. Developers want projects to be built because if they don't get the projects built they make zero money. But they also just don't have the same level of insight on a per project basis even though it's their own project for what the likelihood their project getting to the queue would be.
- [11:45]
A
What is causing these projects to not get through the queue? Does this go to the other problem.
- [11:48]
B
You mentioned which is permitting it's probably about 60% interconnection. So the grid itself just the upgrade cost is going to be too high to make the project pencil.
- [11:56]
A
They penciled out the project, they think it looks good, but they've underestimated how much it's going to cost for them to run new interconnect to the substation that's 15 miles away or whatever.
- [12:06]
B
Exactly. So transmission line that's very expensive. Just to be specific, if you are having to build a new substation that's $25 million. So if the whole CAPEX of the project is 40 million, $25 million substation won't make sense CAPEX the project is 300 million, then maybe $25 million substation.
- [12:20]
A
Will pencil and it's because when they specked out this piece of land, they didn't realize that there's no right of way between here and some existing to hook up to. So they're going to have to tunnel around it or something like that. Are those the type of scenarios we're dealing with?
- [12:33]
B
Those. But a lot of the time the utility themselves at the beginning of the process do not make available and often don't know it themselves what the constraints in the system are. Imagine you have a power line, you have a substation, you have a transformer, your solar farm power is going to have to run through all three of those systems. Let's say the transformer is really old, but it doesn't appear on any system diagram that it's really old and it needs to be replaced. If your system, when they do the full impact study triggers a replacement of the transformer, it could break the whole thing.
- [12:59]
A
It's like when you're installing an EV charger in your house and the electrician says oh, your panel doesn't have enough slots. You got to spend another whatever five grand and upgrade your panel.
- [13:08]
B
Exactly. It might feel like an upsell even for the big renewable Developers similar to the EV panel type scenario and all of the money spent on development costs trying to de risk that project, that all just gets burned. There is no further away to monetize that project at that point.
- [13:22]
A
Thank you, super helpful context.
- [13:24]
B
And then just on the permitting side, just real quick, so that's the other 20, 30% of the projects failing because of permitting. And one of the really, and this is moratoriums. You can't build solar because there's some sort of restriction. One of the things that folks have not realized that we've discovered independently at paces, is that most of the queue, where there's a high likelihood of the project actually moving forward because interconnection is going to be cheap enough, is actually in jurisdictions that have negative permitting outcomes for renewables. So there's basically wherever there's most capacity. So the places where you can most cheaply add electrons to the grid, build a solar farm, et cetera, from a grid perspective has the worst permitting outcomes. And these are directly related. Basically what happens is a developer or set of developers will be like, hey, this substitute has capacity. They'll do some analysis using paces or otherwise they'll flood in to that community, try to capture that capacity, then they'll start the permitting process. And a small agricultural community that had maybe never even dealt with a solar farm, all of a sudden has five developers trying to build projects. The community often freaks out and it's like, okay, there's a moratorium here or some sort of restriction that makes it very hard to build. So there's this hidden time bomb in queues like PJM and a lot of the other queues. We've done this analysis on where there's a lot of projects near to the end of connection. But we have done the analysis. This project literally is in a moratorium and will not be allowed to be built. Because it took so long to go through interconnection. The permitting regime has changed in a negative way for that development type.
- [14:41]
A
I saw on a recent renewable energy trend report that you all put out that it said that you're forecasting that there's a 21.9% decrease in suitable sites from January to October 2024. I guess it's not forecasting that's actually counting. Is that some of the reason why.
- [14:57]
B
So that's actually interesting. That is just capacity on the grid. That number. So we ran the analysis. How many sites have capacity from the grid, from our own analysis and have flat land nearby? That makes sense. That's just decreasing grid capacity. If you add in the permitting, it's way worse than that.
- [15:13]
A
And why is grid capacity decreasing? That's a rapid rate.
- [15:17]
B
Grid capacity is a number that is an estimate until the utility basically tells you the exact number after they do a lot of complicated studies. So we can generally say, okay, based on the information utility, this much grid capacity, end, developers are trying to grab that capacity. As the queue increases, that capacity is being grabbed by other developers. So that's what we mean. It's like the queue is increased and it's consumed a lot of that capacity.
- [15:38]
A
And that number of 22% decreased capacity is really a reflection of just the number of new projects that are trying to get built. And so because this land grab is happening, there's just fewer sites that are likely to be available going forward.
- [15:52]
B
That and also you have, as the utility is getting more of these projects, they start doing more studies and they realize, oh, actually they had less capacity in the first place than they realized as well.
- [16:02]
A
Really helpful context. So all of these challenges are out there and are real. Where does paces come in? What does your software do to help with these problems?
- [16:11]
B
The big thing and the big insight that we tried to bring on the market was you want to pull risk and understanding of risk as early as possible. So typically folks are not even doing permitting until after the interconnection is fairly well processed because it's so expensive to do permitting work. So it's very sequential, this development process. So we were like, look, if we could dramatically decrease, increase the cost, like discovering other risks using software and insights and AI and all the good stuff, we could actually pull forward the risk discovery and the de risking way early in the process. So developers first, only start on projects that have the highest likelihood of completion, second, can fail faster. Right. And understand exactly what's most likely to derail a project. Pulling all that information forward became a key piece. So what we did was we built a search tool for sites where a developer can run a search, end up with a list of sites. But most importantly, we're scoring the sites based on the likelihood that the project will make it all the way to the end of this process. And that's through building out some pretty unique data sets around power permitting and lan. And then once they have the site under site control, right? So basically they've been able to eliminate a lot of sites that they shouldn't waste time on, then we have a series of tools to help them de risk the project again from interconnection permitting perspective in a much more automated way.
- [17:22]
A
So you're somehow absorbing the world of utility data and you're understanding where there is grid capacity or trying to understand where the utilities, even if they haven't actually accurately forecasted it, where you think there might be grid capacity, I'm guessing where there are smaller interconnect queues so there's a higher likelihood to get through the project. And also where permitting requirements may be easier to navigate, I guess. Would those be the big factors you're having to think through?
- [17:49]
B
Those are the big ones. There's a few smaller ones around, let's say where the load and generation, how those things are behind the meter growth. Especially data centers that we can talk to in a moment.
- [17:58]
A
We're going to spend a bunch of time on the data center thing because I know you guys have done some specific work there.
- [18:02]
B
Absolutely. But those are the main ones and the main reason why it was very difficult to do a lot of this previously is permitting has 28,000 jurisdictions in the United States. There's 2,000 utilities. Every US state has a different approach to doing certain things. And so in Ireland we have 26 counties and we're a small little country and one utility. Things are much, much more straightforward in the United States. You can be a world class expert in the development of solar farms in PJM that are 100 megawatts plus and you have nearly no ability to understand what a distrib community solar farm development type is in another part of the country. Just because like things differ so much. There are certain fundamentals, but things differ so much in terms of data available.
- [18:42]
A
Is most of the permitting challenge that these developers are running into on site permitting or does it have to do with whatever the transmission or interconnect is required to get the project live?
- [18:53]
B
Most of it is on site. If you do have to build a transmission line, usually you'll have that. Most of these folks are trying to generally be pretty close to the infrastructure. Unless it's a very large project, people are typically within a couple miles. And so you very in people use paces for this as well. You very early will eliminate sites that don't have a very nice clear, clean right away basically.
- [19:11]
A
Just help me understand. I know we hear about the permitting challenges all the time. If you're building a solar farm which is a pretty known environmental footprint and you're building it on private land somewhere, why is it so hard to get permitted?
- [19:24]
B
It's behind a hedge and it's silent. It generates property tax income for many years. It is a remarkable thing how much folks hate Solar. So there's a few things that come up. The first thing is that there is NIMBYism, right. Not in my backyard. There are some historical concerns around property values is like the biggest one. The second concern is a lot of these are in agricultural communities and there's a concern about transitioning the agricultural nature of the community into something that is more industrial, commercial, et cetera.
- [19:54]
A
So this is farmland that's been farmed for 80 years, 100 years, and the plot is up for sale. It's a, call it a 200 acre plot. It's up for sale. And the neighboring farmers are like, like no, we don't want a power plant here, we want this to be farmland.
- [20:08]
B
Exactly. Now it's again a shocking thing for me not being American, that America, the land of property rights, is like, hey, why do people get to tell you what to do with your property when again it literally doesn't affect you in any way? It's behind a hedge, obviously there's a bit of construction for the year or so it takes to build out the project. It's on a side road, it's not that big a deal. And so these jurisdictions, so that's one big piece of it where you have community opposition.
- [20:28]
A
In my mind, having grown up in the Midwest, one thing I could see is if that entire farming community is dependent on a co op and dependent on pooling its overall produce in order to secure essentially offtake for the produce to find buyers for it. And now you've got a multi hundred acre plot of land that's being removed from that overall pool. You have less supply and therefore you maybe have less sort of market leverage. Like I could see that use case maybe I'm making this up, but potentially that's a thing.
- [20:58]
B
It's interesting. Yeah. When we've talked to these communities, right. Because we're also trying to understand, try to build ever better solutions to respond to these set of problems. And one of the things I've said to multiple folks is like, hey, are you completely against renewables from a theoretical perspective? And most of them are like no. And I'm like, look, how much max do you want to build renewables on this particular county, for example? And they'll be like quarter percent or half percent or 1% of the land. And I'm like great. The problem is where you have zoned for that quarter percent is not next to the power lines that actually can support that. So you have this mismatch. So even when the community is, oh, we're pretty open to it, it's going to be on a brownfield, it's going to be on a former landfill. That's great. There's very few other alternative uses for that land. But if that landfill is 20 miles from the nearest transmission system, it's not going to pencil. So the product doesn't make sense. So there's often like this misunderstanding in terms of what the requirements are to build these projects are and what the community is willing to accept. A lot of these communities are just overwhelmed. You'll have a single person, usually on the elderly side, who is navigating all these inbound permits and so on. And they went from permitting a couple single family homes a year to, to a 10 to $100 million CapEx project and they're like, oh, we don't know what to do with this and we don't have the support needed to actually execute on that project.
- [22:06]
A
Helpful detour. Thanks for helping me understand the local permitting challenges, particularly on site. I actually didn't think that's what you were going to say. I thought you were going to say it was all the transmission interconnect that was causing challenges.
- [22:16]
B
That's really hard as well. But for the individual project developer, that honestly affects more the utility when they're trying to build transmission. It's a lot easier to build gas in this country than transmission, literally legally. So it's, it's just very hard to build transmission lines. But that's more borne by the utility.
- [22:28]
A
So paces is coming in. You've got a software platform, it allows a project developer to see available sites, to score them based on their likelihood to get through permitting, their likelihood to get through interconnect and just their likelihood to pencil. They use you to do what they.
- [22:43]
B
Have those sites and then they can then use us to identify who the landowner is, the contact details for those folks, they will then engage those landowners and buy or lease the land. So that piece where they are engaging landowner negotiating the contract, that is what we call humans doing human things. So it's like we always try to do machines do machine things, what the machine is best for. Then humans do the human thing. The humans are best for the person who is going to call the farmer and talk farmer talk. Agricultural is a very key skill. I don't sound like an American farmer. So if I start calling farmers, they are not going to return my phone call. We'll say that even when I used to call farmers for some of these other ideas that I mentioned. Whereas one of the best LAN teams of any customer that we have they only hire folks from that community and those folks are just way stronger about getting those sites under site control. So that piece is off Paces, but everything else happens on Paces. So you find the sites, we have the contact details, they'll make the phone call, get the folks under lease option or under contract. And then once you have the site on Paces, you have a centralized place to track all the different things that might affect the project over time. And increasingly the ability for us to just complete specific tasks that are de risking the project. Hey everyone, I'm Yin, a partner at mcj, here to take a quick minute to tell you about the MCJ Collective membership. Globally, startups are rewriting industries to be cleaner, more profitable and more secure. And at MCJ we recognize that a rapidly changing business landscape requires a workforce that can adapt. MCJ Collective is a vetted member network for tech and industry leaders who are building, working for or advising on solutions that can address the transition of energy and industry. MCJ Collective connects members with one another with MCJ's portfolio and our broader network. We do this through a powerful member hub, timely introductions, curated events, and a unique talent matchmaking system, and opportunities to learn from peers and podcast guests. We started in 2019 and have grown to thousands of members globally. If you want to learn more, head over to MCJ VC and click the membership tab at the top. Thanks and enjoy the rest of the show.
- [24:51]
A
What have you discovered in terms of the right customer type for you on the developer side? How many total sort of development shops are out there and is there a certain size shop that is the right kind of model for Paces that is helping you streamline your own sales and go to market?
- [25:07]
B
Market we started really tight segmentation, so we started Community Solar Developers in Upstate New York with fewer than 50 employees. That's what we started with and our first three customers were exactly that. And because so much of the data is geographically constrained, focus on a specific area. We just had a product for New York State for our first three or four months in business and then one of our first customers were like, hey, I also need Illinois. And so we were like, all right, if you sign an agreement with us, we'll let you out of the agreement. We will ship Illinois in two weeks. And then we felt our heads down, do all the startup thing and got the Illinois.
- [25:39]
A
This is all the data aggregation and the scoring and all the actually backend.
- [25:43]
B
Work you have to do exactly to make a state work. So we did that over time built up to about 30 US states that people were asking for. And then we just finished off the last 20. The people don't ask for the Dakotas as much as you think, that kind of thing. So once we had all of those in place, so that's how we did it. And we started just with distribution grid, DG solar community solar developers. Then we added in DG battery storage developers. Then we added you don't really have DG wind. We're open into that if I'm businessing. And then we had our first utility scale developer working on the big transmission level projects. 100 megawatts plus. I think they're our ninth or tenth customer. They were also interestingly looking just at New York at that time. So we're able to again build on the data set. We didn't have a lot on the transmission data side at the time, but we were able to build on that. So we've always been really purposeful that we don't want to just, hey. There was a lot of opportunities thrown at us where it's like, hey, do this other thing right. Green hydrogen, it's going to be super hot. You should do this thing over here. That's like a different data set and a different market. But siding is really important to them. We're like, no, we need to make sure that the geographic and data bets we're making compound in ways that actually increase our potential set of markets that we can go after.
- [26:47]
A
Grid intelligence, grid capacity, intelligence permitting knowledge related to connecting to the grid. Land is the constant. But land as it merges with these other two things.
- [26:58]
B
Exactly. And then like the environmental factors related to that, the ability to generate risk from wetlands and things like that, which.
- [27:04]
A
Everything you say would be like, oh, couldn't paces just be the same thing for oil and gas projects? And I guess the answer would be not really. Because if it's not leveraging this underlying grid intelligence sort of data infrastructure you built out, it's not quite in scope for you exactly.
- [27:19]
B
First of all, we would not support oil and gas siting. We've had people ask for that. But the mission of the company is to enable a clean transition. So we would not support that use case we say it has to be over a megawatt. If it's smaller than a megawatt, we can probably still support it. We have a few CNI customers, but generally over a megawatt is when we're a really good fit on the generation or load side side across the United States. Sorry, over megawatts of electrical power. So even when we support the bind meter solution where grid connection is less important, it's still an electrical system versus like an oil and gas well or something like that.
- [27:47]
A
So would you in theory support nuclear projects as they were trying to come online?
- [27:51]
B
We would love to support nuclear enhanced geothermal. Amazing. Some great stuff happening with different project types. We will support any. We're even supporting some gas as part of a larger system for data centers, et cetera, where gas is a requirement. But still, even in that case, we're working hard to make sure that it's sequestered and things like that.
- [28:09]
A
Let's go there on the data center side. So I think you're referring specifically to probably gas backups on a mostly solar powered data center project.
- [28:17]
B
Yes. We very rapidly, in the history of paces, developed this theory about as more parts of the economy electrifies, you'll have different asset classes start to compete. Like what does a shopping mall and some other type, you know, data center, what do they have in common? Very little from a real estate siting perspective. But as more and more parts of the economy is competing for the same constrained set of electrons, you're going to see different types of asset classes start to compete for the best sites. The massive ramp up in AI data centers that started 18 months ago, that pulled a lot of things we were expecting to happen maybe three years from now to now. And so we actually closed our first data center developer customer over a year ago. We weren't even expecting for that. They reached out to us, we started going through the process. Can you use the current product? We're like, we're not shipping anything new. This is exactly what it is. We closed them. They're still a customer today. We later closed folks like Rivian on the EV charging station side, et cetera as well. So we started being like, okay, we can actually start supporting load here and continue to build that out. Then last summer, a lot of folks were using PACES to find where's their load on the bridge to build these projects. And they were finding fewer and fewer opportunities for that or they would find that the load is not going to be available until 2032. So we were like, look, there's fewer and fewer places to build these data centers. And so then they started coming to us and like, I just want to build a 300 megawatt gas plant to support this data center behind the meter. That's not the business we want to be in. They're like, hey, can you help us build a 300 megawatt gas plant? And I was like, are you going to sequester the gas, Are you going to try to figure out where the salt cavern is to get the thing? They're like, well, we'll do something on the back end with some carbon credits, but we're not doing any of that. So we were honestly kind of like, what should we do with these? Should we cut some of these customers loose? Should we think through exactly what to do? And that's when Nan from Stripe and Duncan from Scale Microgrids, they had been been talking about this exact same problem and they reached out to us being like, hey, what are you seeing in the market? We're trying to figure out a way to actually support large scale behind the meter projects that are not primarily fossil fueled. And that was like what became this white paper that has driven a ton of activity. And basically 80% of what I'm doing right now is either productionizing the white paper or porting that use case.
- [30:16]
A
The white paper made quite a stir on social media for folks who haven't seen it, you can go to the Paces website and it's linked to from there, I think. Yeah. Or there's a specific URL for it too.
- [30:25]
B
Yeah, yeah. Off grid AI us. Yeah.
- [30:29]
A
So when I think about the data center world and when I reflect on all the challenges you mentioned at the start of this conversation related to just building new solar on the grid, from interconnect to permitting to finding the right sites and all of that, and just the grid capacity problem that we've talked about, it feels obvious to me that these big data centers shouldn't necessarily be trying to connect to the grid to pull power, but should be funding their own behind the meter solar or solar and storage capacity next door to the data center. Why has that not been the default case? What is causing people to think, oh, now, oh, this is what we should maybe be doing?
- [31:10]
B
Basically we just were in an era of like not much load growth. So if you need a data center in two to three years, you can get the power for that you build out in pjm, you're in the relationship with the grid operator, the utility ISO though you get that project built, it might take a year longer than you expected. It's fine. Right? Then all of a sudden because of this AI, it becomes existential. Right. It's who's going to build the digital God first? Right. You got to deploy so much capex so quickly and the behind the meter piece, there just wasn't a need to do it because the biggest thing is redundancy. Like the data center itself needs what's called 5, 9 redundancy. And so grid connection has the best redundancy because the grid itself is already inherently redundant. It's like a literal network of power lines and everything.
- [31:53]
A
Distributed redundancy. You don't have a single point of failure.
- [31:56]
B
Exactly. And even if you're grid connecting, ideally you're doing system analysis of that grid connection to make sure there are transmission lines that can get around if a line goes down, all that kind of thing. Interestingly enough, the first ever large scale off grid system was built in Ireland. Because Ireland, which is like 20% plus power consumption by data centers today, I think it's the highest in the world, started having a massive cultural backlash to so much power going to data centers and the cost that the ordinary person ratepayer is paying in power. So there was a large data center that was mostly gas that was built just outside Dublin a couple of years ago and that was the first. So even the model of building 300 megawatt gas powered by the meters, that was like new. So like even when people were breaking ground in that nine, ten months ago, even that was like a new concept. I think that was also when we started thinking through hey, could we have a more renewables forward approach. It was like the market has appetite for newness. Right. Because it's moving into this behind the meter model for the first time anyways. Can we prove out a way that's actually cleaner than the fully gas option.
- [32:52]
A
From a cost perspective, is building an entirely new essentially power plant just for your data center. Does it pencil relative to entering into a power purchase agreement and buying power from the grid, or does it not matter because there is no power to buy from the grid is I guess the point you've been making.
- [33:08]
B
Yeah. So these three categories of important things, right? There's speed to deploy the data center and the power system associated with that. The second is cost of that power power. And the third is the cleanliness of that power. So until 18 months ago it went cost, cleanliness and speed were about the same. Now no one really cares about cleanliness, unfortunately. And so we had to argue in the white paper that speed is at least as good, if not better. Do you use a combination of renewables and gas than just going gas alone? Because you can go a little bit higher in cost. Like cost is literally number two on that stack rank. But you can't go slower. Just to give you a sense of it, if you build a 100 megawatt data center, you will completely return the entire return on investment on that in months, not years. So if you can get the product built six months earlier or 12 months or 18 months earlier, there is literally tens of millions of hundreds of millions of dollars in additional money that you can generate.
- [34:01]
A
Time to power becomes the primary metric that you're building against.
- [34:05]
B
Exactly. So we also did this analysis of saying, okay, what are folks on the data center side paying from a levelized cost of energy perspective? And so I think people will be familiar with Three Mile island being refired. That is a pretty expensive but cost of energy. So we had that as our top line and then what a fully gas system would have as our bottom line. And then we did a lot of modeling on what a binyomin meter power plant would look like and what percentage of renewables needs to be there or doesn't breach those two lines. So you can stay within the levelized cost of energy and you can get pretty high. You can get to 75 to 80% renewable system while still being within that range and and being price competitive with fully gas system.
- [34:44]
A
Are you seeing a tipping point in new construction moving this direction yet or was this more a white paper that you're hoping generates more interest in moving this direction?
- [34:54]
B
I'm spending 70 hours a week talking about how to actually deploy this. We have over a dozen projects that we're actively involved in that folks are using this as a structure. So we did not expect this level of activity. We basically had to ship multiple things in paces very rapidly in the month of January to support this use case and basically production reductionize the white paper. But yes, people still need the gas, right? They'll need some gas backup. But you also need gas backup for the Three Mile island refiring. Right? They literally need gas backup for that as well. Of all the things that we've put our name to at paces, this white paper has had by far the most impact. Not just from us as a company perspective, but from a climate perspective as well. Because this has genuinely caught the eye of all the right players and people are deploying large dollars of siting and de risking of projects that follow this model.
- [35:39]
A
Obviously your co author in the report Scale Microgrids has since you authored the report has been acquired is there by eqt. Is there specific technology that needs to exist to enable these type of off grid microgrids to work effectively? What's the sort of the management technology here that needs to exist as opposed to just plugging something into an interconnect queue and shipping it off to the utility?
- [36:02]
B
You definitely need to have folks who are smart at building out these hybrid systems. What is the right component of, of the solar, the best, the battery, the gas, et cetera. That is an important component. And also even fully behind a meter and operating them.
- [36:15]
A
If you are a data center developer, you aren't used to being a power plant operator.
- [36:19]
B
Absolutely. There's definitely that very specific skill set for how a developer would be able to get that project up. And then the asset owner. I will say, though probably the biggest constraint that still exists is actual access to the equipment needed to build this stuff. The procurement for gas turbines. GE has a two to three year wait list for anything at all. One of the first questions we ask when product developers who come to us and say, hey, can you help us do these projects? And the first thing I always say to them is okay, do you have any sort of line of sight to the equipment needed to build these projects? Because there's nobody just giving away the solar panels, the switchgear needed to build these projects. You actually need to have relationships and most developers do not have those relationships. So either you have a need to have a really big balance sheet so that you can have a specific amount of investment to get those relationships, or you already have to prepare these relationships pre existing to get the equipment like the equipment is. I would say after the generic set of problems that we've already talked about or general set of problems we've already talked about. The equipment is the number one thing that folks are not spending enough time and money on. And part of that is also just the capital. If you're going to put a big deposit on some of this equipment, it's tens of millions of dollars. And where the capital sits in the capital stack is still sorting out. You have a lot of risk associated with some of that.
- [37:27]
A
And it's a relatively short term sort of capital gap need. Right. It's call it 18 to 24 months to buy the equipment to getting it in the ground and deploying. If you're off grid, if you're doing interconnect and all that, it may be multiple years, but for an off grid project I think you're talking a couple years.
- [37:43]
B
Exactly. I was talking to a developer the other day who actually doesn't have the money to do these types of projects. But they're like, oh, I've got $40 million of captured equipment. Should I sell that on to somebody? Basically because if we can get $10 cents on the dollar, then we just like nearly flip it. So you're even starting to have kind of near your Brokerage on the equipment side, which is fascinating.
- [38:01]
A
Who are the experts in the procurement side here today? Where are these large project developers going for that particular need? Is that an area you see paces.
- [38:10]
B
Growing into that is in house generally it is probably the main area we are unlikely to grow into because we do not think this is a very strong technology solution there. It's very relationship driven. We've looked at it. You can come up with a list of all the equipment where it's available. But it's similar to like can you build a technology solution to get Nvidia chip? You could, but in the end Jensen Huang has to tell you yes or no. And so it's down to relationship is the primary mover. And so in that case, maybe a startup could figure out something better than what we can on the procurement side, but probably not a fit for us.
- [38:38]
A
Now the report that you published took some heat online in the climate Twitter world for basically advocating that these systems are not 100% renewable. Unpack that for us a little bit.
- [38:50]
B
Yeah, absolutely. So some folks, I think it's completely understandable. Initial reaction were like, look, you guys are still trying to build gas. I think there was even Matt Iglesias had this oh, take off the mask like meme. This is really just like it's not a solar and battery microgrid, it's like a gas microgrid. Basically. Basically we go in depth in the paper on this, the reason why we did this and even the reason why Stripe's climate team did this. And those folks in NAN would be the first to say this, they were basically are funding a lot of frontier carbon capture technologies, et cetera and they were looking at the numbers and just the amount of gas that's being scheduled to build for data centers are going to wipe out all these gains. First carbon emissions. For all the other work that they're planning to do in the next couple of years. And so if there is not a way of having every single or as many of the kind of electrons that will be generated by gas moving to a cleaner source, we're just going to overbuild a lot of gas. And once you build gas, like anything that's built it's 30 year lifespan gas.
- [39:44]
A
Is being provisioned and on site. Carbon capture and sequestration of natural gas isn't really ready for prime time yet in full scope. So I guess a lot of these projects were assuming they would do some kind of director capture, carbon, carbon removal, carbon credit thing that was unrelated to the site they were building themselves A.
- [40:02]
B
Lot of hand wavy stuff was happening basically. And so we were like, look, if we can basically decarbonize 1%, obviously we want to decarbonize way more of these systems on a persistent basis, but that is 1% that's dramatically saved. And just also the speed and the scale, this is not stuff that, oh, we have three to four years to figure out a good technology mix. It's, there's going to be what, $200 billion CAPEX spent on the low end this year by hyperscalers in the United States alone. So that is probably going to be right now 40% plus G unless you start to kind of bend the curve and add more renewables.
- [40:33]
A
You're saying a lot of these projects come online might be 300, 500, 800 megawatt pure play gas plants. What mix are you generally recommending or advocating or starting to see in the market of renewables to gas in a hybrid off grid project?
- [40:48]
B
So the initial few projects that we've been involved in tend to be something like 60% gas, 40% renewables. We're starting to be able to get more projects in the reverse of that. The other piece that we are trying to actively show in the product because again, you can't just say hey, do the cleaner one, right? You have to show with hard numbers is if you also have a line of sight to some great interconnection, even if it's three or four years away, you can not overbuild the gas piece and basically have a mix where maybe the total emissions over time are lower than you expect. But basically, in general, a lot of these systems, we're going from 100% gas to 50, 50 if you gas renewables on average.
- [41:26]
A
Putting in the mix of renewables I think would also primarily mean that you're talking about the US Southwest as your primary land footprint. And coming back to our original conversation, one of the things Paces is particularly strong at is identifying local sites. How much of everything we're talking about here is like a Texas, Arizona, Utah thing?
- [41:48]
B
A lot. I would say 80% of the folks who are actively working on large scale data center campuses are in those states. Plus Oklahoma, Oklahoma, Louisiana and Missouri. Those are the states. And partly that's just because a lot of folks would love to build in PJM and to build in Pennsylvania, Virginia, but you're very land constrained. And for all of my utility friends, I'm sure you won't feel too bad, but like utilities in those jurisdictions are hard to work with in some cases. So generally people Are like, look, you go to Texas, air cost in Texas is the easiest place to build renewables. As it is, you have a lot less restrictions around certain types of permitting. You can build a lot faster. Land is relatively cheap. So air it's Texas is like 60% of it and then another 20, 30% is spread between some neighboring states.
- [42:28]
A
So does all this mean that 100% renewables, such as 100% solar and storage is not a feasible scenario for behind the meter microgrid power for data centers?
- [42:40]
B
So yeah, I guess. What would the 100% renewables case look like? So we did do that analysis and it actually does produce a levelized cost of energy over a 10 year period that is actually manageable in a lot of cases basis. You do see a spike once you get over 95% renewables. Basically you are trying to overbuild batteries to a very, very high degree to handle the one day every two years in Texas on a February where the sun wasn't really shining too much. And you basically need to overbuild batteries for that case. And that drives up some capex and that also drives up the lcoe. There's actually not a lot of difference in the levelized cost of energy between something around the 70% renewables all the way up to 80, 90%. So for developers who are interested in this model, you can actually get to a pretty high renewables percentage without breaking the bank. And then especially because you can deploy renewables faster than gas, that deployment premium is really valuable. Another way to think about it is you could ramp up the power mix over time. So a lot of time with data centers, you're trying to get the first 50 megawatts built and then that's deployed in the data center center, and then you're adding 50 megawatt increments over the course of every three to six months. So if you were able to deploy a renewable system fast and get those 50 megawatts built a year before the gas piece, you can actually do 100% renewable system early and then start layering in some of the gas and some of the other aspects over time. So this hybrid timeline approach where you start with renewables 100%, then you add in maybe a bit more gas behind the meter and then eventually, hopefully you might even have grid connection to build out the final piece of that power mix a few years down the line. I really do encourage people to go to the Off GridAI US website. We actually added a handy calculator at the very top where people can play around with inputs to the model and it'll spit out exactly what the LCOE, the CapEx, et cetera will be around different versions of what percentage renewables folks are using.
- [44:36]
A
As we think about this data center expansion, what are the current options? You've done all this math and basically you said hey, you can either expand the grid, which we know is hard and we've talked about a lot of the challenges there. You can restart big nuclear facilities like Three Mile island and all of that. There are only so many of those opportunities out there. You can build and co locate data centers next to clean energy generation. You can build these off grid sites and basically bring data centers to where you can build new power. Feels like like those latter two solutions are most likely to be the things that grow here. But you know, curious, we are always.
- [45:15]
B
An all of the above type company. From our perspective we're not going to do all the above, but we encourage all of those things to happen. But just from a time perspective, every single one of those CAPEX dollars, it's most likely to be a combination of buying a meter or just one thing. We didn't mention as much. But it's not just nuclear. If you have projects like a wind farm where the PPA is rolling off, there's also the possibility of siting there. So people are going very deep into current power mix. Right. If you've got some west Texas renewables that are, can you actually just bring your data center there and the power is available day one and then you figure out build out of the rest of the behind the meter system and eventually grid connection. So a lot of what we're doing is building a fairly sophisticated way on a per site basis showing the ideal power mix. Grid connected already, co located renewables, new co located renewables, the gas piece all together produce the pro forma, produce all the calculations and then make a strong suggestion and then rank the likelihood that this project will get to the later stages needed to actually build it Us.
- [46:10]
A
James, this has been an awesome conversation. We could keep going for another hour I'm sure but you guys raised a series A middle of last year or so. Maybe talk a little bit about how you finance the business and what you see as the path forward.
- [46:22]
B
Yeah, absolutely. So we closed our series A last summer led by Navitas, MCJ Collective, Amazing Partners as part of that as well and some other great folks. And so yeah, so we have basically been really doubling down and automating as much of development workflow as possible. We're just taking every single Aspect that a developer is spending a ton of time in front of their computer automating that so they can spend their time at that local community building relationship with utility, doing deal making with procurement processes, et cetera, and really focusing on this combined pathway of renewables plus data centers of renewable projects in of themselves. Data centers in of themselves. But then also as we talked about this combined behind the meter opportunity. So our aim is to be able to point to many hundred megawatts of deployed projects in the next 18 to 24 months. Dash Paces was directly responsible for enabling and we're not, we're, we're here obviously to grow fast and be a startup. But if we can't point to those projects that were built because of our work work, then we haven't done our jobs.
- [47:13]
A
Who do you want to hear from today?
- [47:15]
B
Yeah, so any folks who are interested, who are developers who might potentially be interested in the solutions we provide would love to talk to those folks, but also folks on the capital side for project development capital specifically. I love talking to VCs but we're good for the time being. But folks who are actually looking to potentially provide development capital, we've actually started working with a lot of those folks and talking through how to underwrite these projects and what the risks to look out for. And so I think this whole approach, especially the large scale behind the meter, a lot of it's new and there's a lot of like innovation that's needed on the capital side and so folks want to reach out. I'm jamesa.com you can always email me on that and always very happy to just talk the industry as well.
- [47:52]
A
James, anything else we should have touched on today?
- [47:54]
B
No. This is amazing. I can't say enough about how supportive folks at MCJ has been. I was listening to Jason Jacobs version of this podcast a very long time ago and been a massive inspiration and I can't say enough for the support that you guys have had for the community.
- [48:07]
A
Thanks for all the work you're doing. You're working at the intersection of some incredibly important developments right now, both as AI becomes our future and as need and access to clean power hopefully sets the future that we all want to have and see in the world. Thanks for your work and really honored to be a backer of Paces and glad you were able to join us today for this conversation.
- [48:30]
B
Conversation amazing. Thank you so much. Cody.
- [48:33]
A
Inevitable is an MCJ podcast. At MCJ we back founders driving the transition of energy and industry and solving the inevitable impacts of climate change. If you'd like to learn more about MCJ, visit us at MCJ VC and subscribe to our weekly newsletter at Newsletter McGlorin. Thanks and see you next episode.