AI Has A Power Problem

Guest: Scott Willis, President & CEO of Dark Points Host: Jake Aaron Villarreal

Jake Aaron Villarreal: I'm Jake Aaron Villarreal, born and raised in Silicon Valley, here to take you behind the scenes to share what it's like to be a founder, CEO, president, and a leader. We have an incredible guest with us today that's going to talk to us about his journey and what they're building.

But our show really is about the problems that are being solved in the market, the products that are being built, and what makes it easier for our lives as we go through it. So I'm excited to have with us today Scott Willis, President and CEO of Dark Points. Scott, welcome to the show.

Scott Willis: Yeah, Jake, thank you. I appreciate the opportunity to talk to you. I'm always excited to talk about my industry, certainly data centers, and always excited to talk about Dark Points. So I appreciate the opportunity to come on and talk to you.

Jake Aaron Villarreal: Great. Well, a little bit more about Scott. He's essentially built Dark Points, which is a leading provider of regional edge collocation data centers, with over 30 years in communications and technology. He's built and transformed businesses across telecom, wireless, and data centers. So I guess before we jump in, Scott, where are you joining us from today?

Scott Willis: Yeah, so I'm coming in — I'm here in Dallas, Texas. So I'm in the middle of the country compared to where you are over there on the West Coast. But yeah, we are a company that's actually headquartered — and excited to be from Dallas. We've been here about 13 years and certainly enjoy this city.

Jake Aaron Villarreal: Yeah, well, it's a great city. I'm a huge Dallas Cowboys fan, so I like getting down there too. And, you know...

Scott Willis: I know there's mixed reviews on that, but it's always fun to go there. And it continues to grow and build — Texas seems to be leading when it comes to technology and innovation outside of Silicon Valley, where you're talking about big data center projects and infrastructure and technologies, and really large organizations just migrating there for various reasons.

Jake Aaron Villarreal: What was it that led you to Dallas? Was there a specific reason, or just a good spot to build?

Scott Willis: Yeah, no, I think it was opportunity, right? And a lot of the reasons that you described. I mean, Texas — it's a great business environment, very friendly. They make it easy for companies to want to make investments. Lots of talent in this city, particularly in North Texas, certainly in the Austin area, and even down further south into San Antonio and Houston. So geographically it's pretty easy to get east or west domestically with DFW. In America, there's just about nowhere you can't get to in one hop. So like you said, from a technology perspective, it seems to be one of the concentration points outside of Silicon Valley. When you think about other markets in the U.S., it's certainly not the only one, but it's one that — if you're looking at a business, starting a business, expanding a business — you want a business-friendly environment to invest in, and you want to draw upon very talented people in the technology sector. The North Texas region is a really good choice to make.

Jake Aaron Villarreal: Yeah, that's really cool. Well, today we're going to be talking about your line of business, which is the data center side of things. And a lot of us are not experts in data centers, but we know that it's critically important to where the world is going in terms of producing energy. And I like the analogy of a five-layer cake that Jensen from Nvidia was talking about — you know, really starts with energy, that's really line one in terms of looking at AI, and then from there it goes to chips, and then from there it goes to cloud, foundational-layer models, and then applications, which is what we are all using today. But you're really at the base of AI if you think about it. So talk to us a little bit about the data center space and kind of the area that you're focused on when it comes to producing energy.

Scott Willis: Yeah, I mean, listen — particularly, if you kind of think about the evolution of the data center, it kind of started out — it was really more of an economic driver, right? In terms of that — it was natural, where particularly around enterprise, you know, enterprises would find a closet or a small storage room or whatever in the early days, and they had three or four servers stacked in there, and that's what they ran their enterprise applications off of. And obviously, as complexity evolved, as demand evolved — heat, power, all those kinds of things — the data center industry has evolved with that, and that's a very simplified approach.

Whereas today — to your point — we're at the foundational layer, right? Core infrastructure, in terms of what really drives what will be the future AI economy. And, you know, it doesn't mean there aren't other fantastic areas in that layered cake that you described to be in — as an employee working within, or as an investor to invest in. But certainly, when you think about what are the critical components that really enable this AI economy that we're in the very early stages of, understanding data centers is really at the very foundational level. At the core infrastructure — we are the ecosystems that really hold and enable all of the value creation that AI promises, and that we talk about every day. So it's an exciting place to be as part of the industry. It's not the only place, but certainly it's a critical place, and I love being part of an industry that I think is playing such an important role in how this will evolve over the years to come.

Jake Aaron Villarreal: Yeah. Well, the word "data center" — I think of it as a lot of data in a center, and you have systems and technology and hardware and software and water and all sorts of parts of the data center that most of us don't really even know about. For the audience that doesn't know about data centers, can you give us just a synopsis or a summarized version of really what goes into building a data center?

Scott Willis: Yeah, that's changed a lot over the years, right? And it's getting even more and more complex. I mean, if you look at legacy data centers — how we built them — it was fairly consistent, around, if you're focusing on like a collocation, kind of simple space and power. It's having an environment, an ecosystem, that has the necessary power. It has the necessary cooling, right? It's got the necessary security. It's got the necessary redundancy.

When you think about our industry, the data center is the center of the ecosystem that houses all of that data processing and compute and the things that you're talking about. But an equally important component of that is the connectivity, right? I look at it that if the data center is kind of the heart of the human body, where everything's being processed, then fiber and connectivity is the central nervous system, right? That's moving all the signaling, all the things that the heart does in terms of the importance to the human body — but that central nervous system, which is the connectivity, is equally as critical. And so it's not just data centers, it's not just connectivity — it's all of this has to come together as a collective industry to ultimately be able to deliver and achieve what we're trying to accomplish as an industry, particularly on the tailwinds of AI and how our economy is moving forward.

So it's — what started out maybe in the early days as a little bit more of a simpler environment — today the data center is deeply complex. It's not just simply having power coming into the building, right? The cooling complexities, the density on a per-rack basis, how do you disperse heat in a very efficient way — because current configurations of AI infrastructure just drive demands into this industry and complexities into this industry that are very difficult to deal with.

Another challenge in how we think about that is how rapidly things are changing, right? To where what we're building and what we're deploying in our data center ecosystems today are going to be very different even three, five, seven, and ten years down the road. So that adds another layer of complexity in terms of not only how do you build and create a sustainable environment for today, and protect your capital dollars that are invested to create that — but you don't want to find yourself in three or five or seven or ten years in an environment where that data center is no longer relevant, because the technology has passed you up and advanced. And so all of the complexities associated with near-term, midterm, and long-term planning — thinking through all that — is what really has created an environment in and around data centers that, in many ways, is fun, right? If you're a type of person that likes and enjoys challenges. But it also adds risk and complexities into the formula that even as recently as three or four years ago, we weren't dealing with the way we deal with today.

Jake Aaron Villarreal: What's the biggest risk that you've now encountered as you continue to build out data centers?

Scott Willis: Boy, yeah, I mean, listen — the risks are all over the place, as you think about investing capital, building a data center. I think the biggest risk is what I described. We've got — and we'll talk about risk, we'll talk about headwinds, I'm sure, later on, right, that's a different element in terms of deploying — but I think the biggest risk is what I described: this industry is changing so fast and so rapidly. You mentioned a little bit — I've been in technology my whole career. I spent about just under two decades in wireless, and I really started at the early days of when it went analog to digital, which was GSM at the time, all the way through 5G. I thought that was a rapidly changing environment in the wireless sector as we evolved, and it was. But where we sit today, in the data center sector, on the tailwinds of AI, and as rapidly as this industry is evolving — it's on such an accelerated track, even if I reflect on my wireless experience — and that is a huge risk. You've got to understand that change. You've got to understand where the industry is growing. You've got to understand how your customer is changing, what their needs are, what's driving your customer. And then from an investor perspective, this is a very capital-intensive business. You've got to be able to understand that change, and how customer demands and market conditions are changing, so that as you're deploying capital into this sector, you're doing it in an efficient way, so that you're getting an acceptable return for your investors.

So it's hard to kind of pinpoint on one risk, but if I had to summarize it from where I sit and how I think about it, it is operating, building, and deploying in an environment that is just accelerating and changing and evolving so fast, and you've got to maintain and be able to keep up with that. That's probably — I would put it as the biggest challenge, and I would also put it as the biggest risk. So again, I love it, I enjoy it, but it certainly requires constant learning, constant change, to be able to keep up and maintain — directionally, where is this industry evolving to?

Jake Aaron Villarreal: Yeah. Talk a little bit about how big your company is. I mean, how many data centers have you been a part of, have you built out? What's that landscape look like?

Scott Willis: Yeah, so when we started out — really in about the time COVID came, you know, came out — about nine, ten months of investment thesis work with my, at that time, private equity partner, and then we put that to work, and that launched in about — like I said — March of 2020, with our first Arverse sites. We've had a really good run. Today we're twelve data centers in eleven markets. We just closed on another one late last week that we're excited about — you're going to see that out in the public domain probably midweek next week. So yeah, it's been a great run.

Where we sit is — we're an enterprise collocation, cloud data center operator, right? So that's where we sit versus targeting the hyperscalers, or in many cases wholesalers, type of customers. So we run space-and-power collocation. We've got an entire portfolio around multi-tenant private cloud managed services, and then we wrap that around a customer base really targeted at mid- to larger-enterprise. So that's who we are, and that's kind of the journey that we're on, and it's where we're strategically positioning the business.

We are directionally not necessarily focused on tier-one markets in terms of where we are today. We don't go to Dallas, or Chicago, or where you are in Northern California — kind of the primary concentration points of our industry. We're the non-tier-one markets — tier twos, tier threes. That's where we look to deploy. That's where we want to invest, as data distribution begins to evolve in our industry, particularly on the tailwinds of AI inference kind of workloads, that lays down nicely with our strategy, our investment thesis, and how we're looking to deploy capital into the market and build these data center ecosystems that support how we see this industry evolving going forward.

Jake Aaron Villarreal: Yeah, you had shared that everyone's talking about massive AI models for training. You've also shared that inference is becoming a bigger shift. Why is that?

Scott Willis: Yeah, I mean, listen — it's just a natural point of where we are, as we evolve and as an economy are kind of racing towards this AI race that we're all involved with, and what we're trying to build on, right? I mean, to enable it — the way you need to think about that — and these are largely the hyperscalers that are driving this, right, and again, those are not the types of workloads that we target — but those are big learning workloads, right? So if you want to think about it from a simplicity perspective, it's like — it's like you and I studying for an exam, right? Depending on whatever it is — whether it's medical, you're going to medical school — it's us studying for an exam. That's what the big learning workloads are all about, right? And we're maturing — by no means is that complete, but we're starting to mature. And so we've been talking about inference for some period of time, and there's no question in my mind — I see it in my pipeline, I see it in my conversations with my customers, I see it when I'm going to conferences and I'm talking to partners and peer companies. You know, '26 is the year where we're beginning to see inference take off. We're by no means on the scale part of the hockey stick — I think you'll see that in '27, '28, '29 and beyond — but it is happening. Inference is real.

And so when you think about inference, and you kind of think about that learning workload — if that's kind of studying for the exam — inference is about taking what you learned as you were studying for that exam, and applying it in the field of choice, right? If you're a doctor, and you're actually practicing and seeing patients, that's the way to think about inference. It's about predictability, it's about decision-making. And so that's where users, in our case enterprises, will take the investment in the big learning workloads and then apply inference workloads down into their business — whether it's operational efficiency, whether it's sales enablement, whether it's industrial manufacturing and you're trying to get efficiencies that way, whether it's customer care — how you interface with your customers coming into your business. All of those are examples of leveraging inference in a way that optimizes your business, enables your business to be more productive, more efficient. That's the way to think about it, when you think about learning workloads versus inference workloads.

So we're moving into that stage of AI where end users are starting to take advantage of those big learning workloads and applying them down into their environments, to ultimately achieve the efficiency that AI is promising as an overall economy. I'm simplifying it, but for people that really don't come from the industry or don't understand it, that's usually a pretty good way for me to apply what I'm talking about when we talk about learning versus inference kinds of workloads.

Jake Aaron Villarreal: Yeah, I mean, we see AI as productivity tools for how you just do your job better. And then we start looking at, okay, well, how can we automate our business? And then how can we use agents and create agents that can start to do the work for us, or with us — and then soon those agents are talking to each other, doing the work autonomously, sometimes with you and sometimes without you. And so I can imagine just the amount of information going back and forth, and the amount of energy required, when you have millions of agents and millions of data points going back and forth — it's got to come from somewhere. So I guess the question I have is — are we underestimating the amount of energy really needed as our industries continue to really take shape and use and maximize AI as we go forward? Where are we? I read in the news that other countries are ahead of us when it comes to AI, or other countries have different types of infrastructure for energy — and then you talk about, well, what does it take to stand up a data center, whether it's the hyperscalers or just in general — and where do you see, in terms of the demand, where are we in that race, if you want to just put a line in the sand? Curious to get your thoughts.

Scott Willis: Yeah, I mean, listen — we are, as an economy, the U.S. economy, we are certainly leading the race in a significant way, right? And I mean across the spectrum. Doesn't mean that there aren't other economies that we're competing against that aren't making milestone achievements as well. It's challenging, right? Because when you understand the ecosystem around data centers and how AI is evolving, the need for that ecosystem to keep pace, and be able to support what's required to be competitive and win that AI race — because if you look at the very front, which is mainly what you see in the press — this is always the front end, it's the hype, it's what people like to write about, it's what people like to talk about — software and silicon. Okay, rapid speed, right? Nvidia and others are at the forefront of that. That's what we read about, that's what we see about. There is no question — when I talk about the speed of how this industry is evolving, that is a driver. But none of that matters — none of that matters if there's not an ecosystem like a data center to put that in, for all of the processing, all of the learning workloads, all of the inference workloads, to be able to leverage that.

And so that segment, or that tip of the spear of our industry, moves rapidly, and it changes rapidly, advances rapidly — and that's what you see in the press all the time. Let's move to the middle, where infrastructure is — and let me — again, I'm not going to go through every building block of our ecosystem, I'm going to hit the major blocks. We, as an industry, given the challenges of labor, given the challenges of supply chain — whether it's impacted through raw material, pure demand — think generators, switch gear, cooling equipment — long lead-time items, in some cases north of two years.

Jake Aaron Villarreal: Wow.

Scott Willis: So my component — that's the infrastructure, building out data centers — that moves at a very different pace than the tip of the spear, which is silicon and software. So let's lock in on infrastructure, which is a data center — let's call that two and a half to three years of deployment, to be able to keep pace. We've got to build more and more data centers. Got to build, build, build more and more data centers. That's the window there.

Now, you've hit on it several times, right? You've hit on power. Nothing happens in my data center without an appropriate level of power to be able to support that. So now let's look at that — if you look at the power grid — stable, predictable power that needs to be able to grow, at the density level that the tip of the spear, the silicon, is driving. It used to be that half a megawatt, and five, six, seven kilowatts a rack, was pretty good, dense stuff. Today, we're talking regularly with our customers about eight kilowatts a rack up to two hundred kilowatts a rack — significantly more demand on power to support that. And if you want to move to where some of the industry is talking about — you're talking about a megawatt in a rack. So that's how quickly that industry is moving.

But the reality is, on the power infrastructure side, that is an industry that moves in decades. Okay? Because of what it takes to plan for power grids, in terms of building out generation — and then you need transmission, because you've got to transport it to wherever it's going. So you've got an industry there that thinks in ten-, twenty-, thirty-year cycles. So, but at the end of the day, going back to your question — what adds so much complexity to that is you've got the tip of the spear, that software-silicon piece, that moves rapidly, that moves quickly. We're in the middle — we're the ecosystem that you put all this in that enables all this data — let's just say between two and three years for ours — and then you've got to have the power to be able to support it, and they work in decades. So how do you put all of that together, as a single ecosystem, that really contributes to the success of the U.S. as an economy, to maintain our leadership in AI? So it's not an easy problem that you can solve, and depending on how you peel apart certain parts of that ecosystem, each of those has its own different challenges. And that is only going to exponentially grow.

And so that's where innovation comes in, right? Lots of innovation on the front end — tip of the spear, in terms of where I'm talking about — you're seeing more and more innovation there, and in where I sit, in the data center. And then on the power side, that innovation has to come in alternative sources of energy, right? And that's where there's lots and lots of talk in our industry around alternative — some of the basics around solar and wind, and some of that is going to play a role. You've got a lot of momentum building around turbines, in terms of that playing a role. You've got lots and lots of conversations — in some countries it's more advanced than in others — around nuclear, in terms of, can nuclear be a — but at the end of the day, all of those alternatives are important. Innovation is important. From my perspective, it's the traditional grid — of what I described today — that is going to play a hugely critical role in delivering sustainable, dependable power to this industry over time.

So how do we continue to accelerate that time? Those are going to be challenges that, as part of our industry, and those that are specifically within the power sector of our industry ecosystem, are going to be challenged with — and it's going to be an important part of how we solve the greater equation of how we keep up pace, how we maintain our leadership position in AI as an economy. And the way you do that is you've got to solve all these building-block challenges that enable us to do that. So it's all deeply interconnected, right? And very much like fiber and data centers — power plays that same critical role. All of those intertwined interconnection points that really enable the ecosystems we're talking about — that's really going to allow AI to achieve the levels of leadership and ambition that we, as an economy, want to achieve. They're deeply connected, and it's challenging, and those are the things that all of us that are in the industry are dealing with on a day-to-day basis.

I could go on and on in terms of other headwinds and other challenges, and I'm sure we're probably going to get to it a little bit later on — but I simplified it, because there are a lot of other critical components. But if you really want to break it down into those simple ones, that gives you a sense for what are the challenges around speed to market, speed to delivery, and how do we maintain that leadership position within AI. Each segment of that ecosystem plays a different role and has different challenges.

Jake Aaron Villarreal: Yeah, it sounds like there's some constraints across different levels in all of that, but one that you mentioned was talent. Talk about the people that are required to build a data center — or just in general, where are you seeing the lack of talent, or the demand for talent, as you build this infrastructure?

Scott Willis: Yeah, I mean, listen, I think talent is always a challenge for any industry that achieves any level of reasonable — you know — high growth, much less hyper-growth, which is what I would consider the data center space to be in today.

Jake Aaron Villarreal: Yeah.

Scott Willis: In terms of that, it's all the way across the — if I look at the challenges that I have today in our capital program, in terms of expansions or new builds that we've currently got going on — you know, just the volume of activity in data centers — if you just think simply on the trades, right — these are welders, these are electricians — these are just skill sets that we need to be able to stand up a data center environment. So, given the volume that we're facing, it's just a shortage of that kind of skill set.

As you move more into the environment of the data center, as innovation is playing a much more critical role — whether it's cooling innovation and technology, whether it's switching and electronics, whether it's how do we deal with an environment as density gets more complex, and heat distribution, of all that heat that's generated by these servers that are processing all of that data — the complexity of what that has created inside the data center, versus even three, four, five years ago, is completely different. And the skill set, the knowledge, and the workforce that you need to be able to — not only in your organization, you know, me internally within Dark Points, but within my partners — nothing happens in this industry as a standalone. I have a deep and wide partnership portfolio that we rely on every single day to accomplish what we do. They have to have the appropriate level of talent to be able to guide my team as we're evolving through this process.

So, and then when you overlay all of that with the speed of change that you and I talked about earlier on — how do you take that talent and continuously educate them, continuously evolve them, so that they're trained and they're knowledgeable on how this ecosystem is evolving and changing at such a rapid pace? It's a real challenging part of the industry, getting the right talent, the right mix. And when you throw the overall demand in — it's even harder, because there's just not enough people out there with the skill sets — again, from just traditional what it takes to build a data center, all the way through the technical skill sets — and we need to solve for that.

And you're seeing some real leadership in our industry evolve towards that. I mean, you've got universities — you've got SMU, right? With the strong participation of a few individuals here in Dallas at a leadership level who recognized this and approached SMU in their engineering school, where they actually have curriculums now focused on the data center sector. So you now have institutions that are recognizing the need for this skill set, and we're teaching it at the university. All of that innovative way of thinking is the way that we're going to solve for the challenge of how this industry is evolving, how AI is changing the complexity and the demand of the industry. And that's the way that, from a talent perspective, we have to solve for it — because it's not a world where, a few years ago, there was plenty of skill sets and the environments were much simpler.

Today, it's about making sure that when you come into the workforce — again, regardless of where you enter the ecosystem — that you're prepared for it, you're trained for it, and you're productive — you can step into a role and start contributing at a relatively high level from day one. And that's going to continue to be an important need of our industry, as complexity only enhances, as the speed of this industry only enhances. That's going to be an important equation that we need to solve for us to be successful.

Jake Aaron Villarreal: Yeah, I mean, we're in the people business, on the talent side of things really, where it comes to engineering and product and leadership and operations and different areas. But when I think of a data center, like you shared, it's really hands-on, and you're building physical products and infrastructure. And so if you're out there as a young kid coming into the market — maybe you went to school and you're an electrician, or you're a plumber, whatever it happens to be — but you're hands-on, but you see this as an opportunity, you have to adapt and accelerate and learn really what you have to do when you join a sector like that. What would you recommend that they do? I mean, are there — you mentioned SMU as one institution — but are there schools, are there curriculums, that you could get hands-on experience so that you could join an organization that's building in that space, where you would hit the ground running and be able to get up to speed on the complexities of where tech is headed? What are you seeing there?

Scott Willis: Yeah, 100 percent. A lot of it's what I'm talking about, and you see a lot of it in the press — there's a lot of advancements in the trades, where they're trying to get more and more people to consider, maybe instead of a four-year college degree, they go to trade schools and learn a trade. And we're certainly an industry where there's such demand, and that demand is sustainable. Historically, technology comes in cycles, right? It would lift up and then it would fade. You'd realize a return on your investment, and then the next technology — this one is sustainable. I sit and listen to thought leaders and people that are really perceived to be experts in the industry, and, you know, no one really knows how sustainable this capital cycle around AI, and particularly around data centers, is going to last. I get that question constantly from my investors and prospective investors, in terms of how long this capital cycle will last — and from my perspective, it's hard to predict that.

So I think it's a wonderful opportunity for any young person coming out. I mean, I know that's the approach I took — I worked with my staffing resources on campus, and met with them, and laid out what was important to me, what were things I was looking for. So if you're that type of individual, and you're looking for long-term opportunities in a growing industry, if you want to work with highly talented people, and that motivates you in terms of being in that technology field, if you want to look at opportunities where — better-than-average earnings, in terms of what you can make in your career — the data center sector — another one that was important to me is working in an environment that had a positive impact on society, and I would argue that this is one where, if that's important to you, those are the drivers, and I think the data center sector offers that.

And so again, whether it's the trades, whether it's more of the technical, more of the engineering — wherever you want to enter into this industry, there are opportunities for all individuals as you're a young person coming out and deciding what you want to spend your career doing. I'm — obviously I am in the industry, but I think this is a pretty good choice in terms of what you can accomplish. And I would just encourage you to be curious, be a constant learner — that's what's going to make you successful, I think, in this industry. I mean, there's a lot of press out there in terms of people that are afraid of AI and those kinds of things. I don't happen to buy into that. I think if you're an individual that embraces it, and embraces change, and you're curious and you're constantly learning, you don't need to be afraid of it. You come in and understand it, and figure out how you leverage it, and how you evolve your career. This is a great opportunity to be able to do that. But certainly it's one where change is rapid, and it requires that individual that really enjoys and is attracted to that constant-learning and constant-change environment, because this — Dark Points, or any other data center environment — will test you on that, because it changes so rapidly.

Jake Aaron Villarreal: Yeah. Well, you're in a space where there's a lot of demand for it. There's some constraints around talent and infrastructure needs and a lot of different challenges across the board — but one that most companies don't also have a challenge around is the community where you're putting data centers in. What do you think people are getting wrong about that idea of a data center coming into their city or their town?

Scott Willis: Yeah, that's very real, and I'm glad you asked that. This is — I'm going to stick with my thesis of change and rapid change. If you would have asked me in late summer of last year, or even early fall — would Dark Points be dealing with the political environment that is very real, at the state level and local level today, for our industry? I would have said, "No, it's not — it's not for Dark Points." I mean, yeah, a hyperscaler that's deploying half a gigawatt or a gigawatt kind of facility — yes, that's a natural cost of doing business, engaging. But even today, with Dark Points, I've got lobbyist firms that I've had to engage in the states that we are in. The headwinds politically at a state and local level are very real.

And from my side, it's a little bit frustrating, because the headlines that you get at a local level are very political, and it's coming from a position that's not really knowledge-based — particularly when you talk about Dark Points, where we kind of play in the ten- to, let's call it, fifty-, sixty-, seventy-megawatt range. Those are the ecosystems that we play in. Anywhere we go, anywhere we invest, generally the power is available on the utility that's there. I mean, that's what we want to do — take advantage of excess utility power that might have been used by a different industry ten years ago, but today that industry no longer needs that location. You've got stagnant power off the utility. That's perfect for us — we want to move in, we want to redo that brownfield, build out a data center, and we've got power right there ready to take advantage of it.

It's something that, last summer, I didn't spend a lot of time thinking about, and I didn't spend a lot of time in my day dealing with it. Today, that is a significant part of my day — as I'm engaging law firms, as we're shaping legislation, as we're trying to work with those legislators to educate them, we bring them into our data centers, we give tours, we want to show them we're not — we're not what the headline says. We're not the dark enemy here. We are good corporate citizens, right? We want to coexist in this environment with the communities, in the states that we do business in. And it's a real education of that community, in terms of the role we play versus the headlines that might be political, that some politicians are trying to use politically for some purpose.

So it's real, and it's an investment — personally, my time, my team's time, and then financially — we're having to pay for this, but it's important. We've got to work with our local communities. We've got to work with our local utilities. It takes that teaming effort. If we're going to be successful as an industry going forward, we've got to do it in a collective, collaborative way, and demonstrate we are good corporate citizens. And particularly in Dark Points's environment, being an enterprise environment, I strongly benefit the enterprises that are in the communities where I am — they can leverage my ecosystem, they can put their data there, they can have connections out into broader communities, from an interconnection and communication standpoint, with other enterprises. I, in the role that we play, we unleash the economic value in the communities that we're in, by what we provide enterprises within our ecosystems.

And it's just a constant education. It's talking to the community, it's talking to the leaders, and talking about, hey, we're not here driving up your power rates. We're not here using all your water. We're not some of the negative headlines that you do — that it really is a tremendous amount of energy, economic and personal resources, are going into that, where again, a year ago that really wasn't even on my radar. That's how quickly the political environment has changed around data centers, where everybody kind of wanted data centers to come in — they're great businesses, great economic engines — and now all of a sudden it's a negative.

So, when you talk about some of the other ecosystem challenges that we were talking about earlier, in terms of maintaining that leadership position on the AI race, this is another headwind politically, as we're looking to deploy new data centers or new ecosystems into these communities, to make sure that we keep up with that infrastructure demand that you and I talked about. So again, it's just another part of the industry that you've got to focus on. You've got to have a strategy on, and you've got to be a real collaborative partner at a local level, in your community, as you're trying to educate the leadership there of the value that you really bring, and that we're not really taking away from your community the way that sometimes it's portrayed in the media. So yeah, don't want to underestimate — that's a big part of the reality of what we're dealing with as a data center sector today.

Jake Aaron Villarreal: Yeah, well, we've heard about the negative aspects of what we read about when it comes to having data centers go into a town or a city. What are the positives? You talked about the enterprises, you talked about bringing value to the local economy there. AI is being used by just about everybody, it seems like — so if you're going through a ChatGPT or on Claude or something, and you're in any part of the world, that energy has to come from somewhere, data has to come from somewhere, that system has to come from somewhere. So there's benefits to it — what are some of the other benefits that you're having to convince, you know, politicians or people in the community that you're bringing to them, that's really not hurting them? Like, you know, if my bill went up a little bit on the electric, but I was getting value out of AI, I probably wouldn't think about it much. If my water went up a little bit, I probably wouldn't think about it too much. But, like, how much water is it taking, how much electricity is it really taking from me as an individual? Just kind of paint that picture a little bit if you can.

Scott Willis: Yeah, I mean, that's interesting, right? I mean, here's — I'm going to speak from Dark Points's perspective on this one. It could be different, depending on where you come from in the industry, but from an economic development perspective, it's about jobs. It's about education system, it's about housing, property taxes, and it's about those employees coming in and spending money in their community. Okay, that's kind of the historical, legacy view of what a local community's economic development team wants to talk to an enterprise like me about, to be able to draw them into their community and attract them in.

We're — the reality is, we do create a lot of jobs, particularly if you think about the broader ecosystem and all of the building that's going on, and the partners, the general contractors, their subs — we do generate a lot of jobs. But as a standalone data center, we're not like a big manufacturing plant that comes in and is going to bring two thousand jobs into the community. It just doesn't take that many employees once you've got that data center. But here's the message that I try to work with, as I'm trying to educate them, that I think is a little bit underestimated in our industry. Yes, the value that you're talking about, from an economic development perspective, to the enterprises, the businesses in that community, to individuals — whether that's on your iPhone or whatever your device is, on your laptop, on your computer — that aside, we all, I think, generally understand what AI is bringing to that. But the shift that I try to get the local economic development people and politicians towards is — don't underestimate the value of sales tax that the state gets as a result.

Okay, so think about it this way. If I've got a rack that's doing about thirty kilowatts a rack, in my data center today, that's plus or minus approaching a million dollars of servers and equipment that goes into that rack. Okay, that equipment has to be purchased and shipped to my site. It hits my dock, we either rack and stack it, or our customers come up and rack and stack it, or a third party comes up and racks and stacks it. That's all sales tax for the state. So you get up to fifty kilowatts a rack, that's about one and a half million dollars in equipment. So if you think about it, I've got hundreds of racks in my data center, and you think about all the sales tax it takes to buy all those servers and all that equipment to stand up all of that into that data center — that is a significant sales-tax uplift for the state that just seems to fly under the radar in terms of local communities understanding the benefits.

So even if I don't talk about the output, and I talk about all of the efficiencies and gains that enterprises in that community are going to get — let's just set that aside, if we believe that AI is going to deliver that — that's the piece that I try to get people to understand: it's a sales-tax benefit to the state, whether it's that local — and we've got a data center in Greenville, in Greenville County, and we're doing an expansion project there, and the sales tax is billions and billions and billions of dollars in infrastructure equipment that goes inside my data center. And this isn't — I'm not even talking about generators, and PDUs, and CRAC units, and switch gear, and all the heavy infrastructure gear that enables my data center to work, all the cooling, all the things that's big money — I'm talking about the customer's equipment that is brought in, the racks and racks and racks of equipment. That is a tremendous benefit to that community from a sales-tax perspective.

So those are some of the ways that I try to educate, as we're talking to the communities — is I may not bring you lots of employees, that's going to create a certain level of income that you're expecting across your community, because that's the way that traditional economic development at the city level tends to think about things. So it's just a different way to think about it, a different way to approach it. And when you're talking decades of ingrained thinking, of local governments, in terms of how they attract local businesses to come to their communities, and I'm trying to pivot that a little bit and ask them to think about it differently. That's certainly one of the big areas that I spend some time on, trying to educate them to say, you know, we are different — we don't bring lots and lots and lots of employees that are going to buy lots and lots of houses and those kinds of things. But all that equipment sitting in my data center, oh, by the way, all of that sits on about a four- to about a six-year refresh cycle. So once they deploy it, and then it lifecycles out, and that enterprise has to reinvest in all that new equipment, because it's kind of run its life — and in some cases it's even more rapid than that — all of that repeats itself. So it's a gift that keeps giving, in terms of — it's an education, taking them through that.

Those are a couple of different ways to kind of think about it, as we're working with our local communities and we're talking about future investments, and we're trying to go to them and say, hey, we'd like to get some property-tax incentives from the local county or the local city there, because I'm going to build a new data center, or I'm going to expand a data center, and I'm going to invest this many millions of dollars in your community. Those are some of the trade-offs, and the ways to kind of have them think about why it's beneficial for a partner like Dark Points to be a really solid citizen in your community, adding value to the broader community. We're not just a negative, or a takeaway — we're actually additive to what we're trying to do. And so it goes back to that — it's just education. You've got to invest the time. It's about partnerships with that entire local and state-level constituency, and just making sure they understand the value proposition that you bring into that community.

Jake Aaron Villarreal: Yeah, well, that's well said. I want to take a different step here and talk about other options. You talked about alternatives when it comes to power and data centers. What are your thoughts about what we're starting to read more about — data centers in space? Is that like a dream? Is that hype? Is that reality? I was listening to — it was on X a couple days ago — about a leader talking about millions of satellites hitting the sun 24 hours a day, and there's an opportunity to kind of put a strategy together that is sending that energy down to Earth, and you've got Bezos, you've got Elon Musk, you've got others talking about this as a direction that we could be looking at. What are your thoughts there — is that reality, or is that hype?

Scott Willis: No, I mean, listen — you're talking to someone who believes that — listen, I mean, a satellite today that is launched, right — and we've been doing that for years — I would argue that is a mini data center, in terms of that that already exists. But so, listen, you're talking to someone — I'm a big believer in innovation. I'm a big believer that nothing remains constant. And I think that people that are investing, thinking, and working in disruptive ways of thinking like that are critical to the sustained success of us as an industry.

Do I believe it'll get there? I think probably so. I mean, you said you listened to one — I think it's been out there. Bezos has been a little bit public recently, in terms of some interviews that he's done, and he was asked that question. You know, we may be a few years away, but listen, I think that is very real. If I look at our sector, there's example after example after example where innovation, and that kind of creative, disruptive way of thinking, is very real.

So do I ever think that it's going to completely replace what our primary core of how we deploy data centers today is, and the value proposition that we're providing for the industry as a whole? No. But do I believe that could be an alternative solution? Right — very much like power innovation, and the way that — I do think the grid is going to, for the foreseeable future, be the primary source of sustainable, dependable power for our industry. But that doesn't mean there's not going to be alternatives, and disruptive and innovative ways of doing that. I look at that space the same way — I think there's a similar approach, although space seems to be getting a little bit more momentum than, say, data centers down on the floors of the ocean, in terms of the potential benefits there. So I love that kind of thinking. I appreciate the people that are thinking about it that way and working in that part, because eventually I do think there will be enough innovation that is realized, and it'll be a complement — it'll be accretive to what we're trying to do — because the amount of demand around just infrastructure, the foundational infrastructure, which is where you and I started this conversation, is only growing and growing and growing. So the more innovative types of technologies that we can come up with, as a collective industry, that can support that, I think is the better.

So I'll — maybe to close that down as a predictor — do I believe it'll get there? I do. I think, at some point in time — probably a little more innovation, a little more time needs to take place before we get there — but I think it's a viable solution, that could be an alternative way to think about infrastructure.

Jake Aaron Villarreal: Yeah, well, I want to end on a couple of notes. One — we get calls and emails all the time from kids in college, and they're all asking, "Where do I go? What should I do? What should I be studying? Should I drop out of school? Should I start vibe coding? Should I understand business and workflows and building agents and getting into AI? What should I be doing?" Because what we're hearing is, within institutions today — and it's getting better — but it's more about just going to school and getting your degree, and trying to prepare students for the real world. But we're seeing this massive change across all positions, and AI has to be not just something you think about — it's what do you use it for today, and how can you bring that value to our company if we hire you? And that's just — think about knowledge workers. Let's talk about the people that are building data centers too, because I know that, like, you had shared earlier with me on a call, that if you're an electrician coming out doing electrical work, it might not necessarily be the same type of work that you would have to do from a technology perspective in a data center. There's other forms of education there too — but really, across the board, if they wanted to get into your space in data centers, what path should they take now?

Scott Willis: Yeah, I mean, I think it's fairly consistent with what I'm saying earlier. Listen, I'm in this phase with my daughter, where I have this conversation with her as well. I come from a belief that, you know, life is a journey, it's not a destination, and it evolves, it changes — there's failures, there's successes, there's resets — you make choices. So that is inevitable for anybody. And so I've always kind of — the advice, as people sit down and talk, is: if you can find your purpose, if you can find what drives you, what motivates you, what you're deeply interested in — versus just, you know, not just a job for the sake of a job — but if you can kind of find that inner self that drives you, that, to me, is really — if you can tap into that, and if it's into data centers, if it's into technology, if it's into AI, once you've solved for that, the rest of it — you just naturally evolve with it, because you're naturally driven by that passion.

And so to me, that's the way — that's the first thing — and I realize, in saying that, that's not easy. You've got to really self-reflect. You've got to understand yourself — what drives you, what are you interested in, what is it that you want to do? And if you can get grounded in that, that really will serve you well as you step in and say, okay, now that I understand that, what are the opportunities that are out there for me, and based upon that, where can I achieve that? Where — if these are the things that are important to me, and this industry offers those things — then it's an easy step to do that.

But that's the first way that I would advise. And as you think — if I'm a young person today — and again, there's so much fear in the media around AI, and extremism, in terms of, it's going to replace everybody's job, unemployment is going to be at 99 percent — those kinds of — I'm just not a believer in that. We have had industrial shifts since our beginning, and we're very adaptive as a species — humans are. And if you're curious, and if you're a constant learner, and that's important to you, you've got an advantage. And I look at AI as no different than that — embrace it, understand it, figure out how you can leverage it and take advantage of it.

And I think if you come, as a young person, and you're reaching that point in your life where you're trying to assess what is it, what do I want to do, what's important to me, what's going to drive me — because again, it's a long career, and if you can recognize that and approach it from that perspective, my personal view is it's going to be a lot more satisfying, it's going to be a lot more rewarding, and you're going to enjoy that journey along the way. But you've got to tap into that, and understand kind of the basis of where your interests are. And then from there, you can make choices into — what's your on-ramp into those industries that will support what's important to you? And that's the way that I advise and talk to people that are younger and thinking about, as they — what did I do, what choices did I make? Don't be afraid, embrace change, take risks — take the jobs that nobody else wants — that's where you're going to differentiate yourself. People recognize that. If it's a job that's easy, and it's in a wonderful location, and it's not disruptive — well, you've got ten people that want that job. Take the one that nobody wants, take the one that's riskier. Yeah, you may have a few failures along the way, but I promise you, if you adopt that mindset over time, you'll be successful, you'll be recognized for it, you'll be rewarded for it, in terms of that type of an approach to — over the long journey — what is it you want to do with your career?

Jake Aaron Villarreal: Yeah, that's great advice, and I like how you state that — take the risk, and also do the hard work. Do the stuff that no one wants to do.

Scott Willis: A hundred percent. You will learn, you will get better, and pretty soon you'll have people working for you doing that hard stuff. So we've seen that evolve over the years.

Jake Aaron Villarreal: I like to always leave a little space at the end of the show for you to talk about — first, where you see the growth in your company over the next 12 to 24 months, and then also maybe internally, some of the roles you're looking to fill for your own organization as you grow. Feel free to talk about anything there that makes sense for you.

Scott Willis: Yeah, listen, we're just — we're in a stage now where we're trying to — we think we've got a really good strategy, we think we've got a really good vision, we think it's very clear, it's very defined. Our customers understand it, our employees understand it, our investors understand it, right? And we think it's based on really strong tailwinds, particularly on AI, as we look to grow and scale and invest into Dark Points. So now it's really about execution — if we've got that conviction and we've got that passion, and those pieces are in place, how do you execute? How do you operationalize that vision? How do you create that ecosystem of what it takes to realize that vision? How do you do that in the most efficient way, the most cost-effective way, the most repeatable way? That's what we're really focused on. We're at that stage in our journey where the vision, the strategy, the customers, our employees, and our investors are bought into that. And now we're really working on that operational piece — how do we industrialize that in a way that we're really efficient, and we're really good at deploying these data centers to support the demand that we have? And it's hard stuff — it's really hard stuff — but that's kind of where we are, and that's an important piece.

Listen, where you started with this — it's really talent across the board — but if I really look at where we're really, really looking to solidify, it is that technical engineering expertise, in terms of dealing with this world of constant change, this world of constant speed, constant evolution — how do I maintain Dark Points and our relevancy across that dynamic that's going on? And so those are skill sets and talents — again, doesn't mean that finance and marketing and sales and operations, all of those are important too — but certainly, having really good, solid core engineering expertise, as you're dealing with the challenges, the innovation, and the speed that you're trying to evolve as an organization — again, to meet your constituents, your customers first and foremost, your employees certainly right there, and secondarily, and then certainly lastly your investors, that are behind you, that have confidence, and that are bought into your vision, and are providing the capital that you need to continue to grow and sustain yourself going forward. It's a real challenge. It's a lot of fun. I enjoy it every day. But it comes with a mindset that you've got to be an individual that thrives in that constant-changing and challenging environment.

Jake Aaron Villarreal: Yeah, well, Scott, well said. I like what you're building, I like the space you're in. I know there's lots of opportunity and a lot of evolving areas within that industry, so it'll be great to continue to hear how things go for you. If anyone wants to find you or find Dark Points, where do they go?

Scott Willis: Yeah, listen, just dartpoints.com. That's everything — everything you need to know is on our website — who we are, what we're about, anything that you might need to know. Certainly me on LinkedIn, or however you need to hit me up — yeah, happy to connect that way. So, you know, I think society in general is — we are just more and more about interconnection, right? We're all interconnected. That is very true within the data center sector. It's very true within our environment, within Dark Points. So yeah, I certainly embrace that, as we're trying to grow and scale and evolve what we're trying to do at Dark Points, and what we're all about.

Jake Aaron Villarreal: Yeah, well, really cool. Well, thanks for coming on the show and sharing your story, and for the listeners, for listening today — it means a lot to me if you spent your time with us. I'm your host, Jake Aaron Villarreal, signing off for now. Can't wait to catch up with you all on the next episode. Until then, Scott, take care.

If you like what we're doing, don't forget to subscribe, leave a review on Apple Podcasts or wherever you listen, and follow us on YouTube, where we go behind the scenes to learn what it takes to be a startup founder.

‍