The geography of data center real estate has been changing rapidly since ChatGPT arrived in 2022, kicking off huge demand for IT infrastructure for AI hardware. This AI boom has transformed how data centers are designed, and where they are located and powered.
To understand these changes, DCR host Rich Miller spoke with David Liggitt, the founder and CEO of datacenterHawk, which provides research and analysis on data center real estate. David is one of the sector’s sharpest analysts, and has been tracking data center markets for nearly 20 years. With the in-depth market data compiled by datacenterHawk, he has a unique vantage point on how AI is changing data centers, and what it may mean for our digital future.
In our discussion, David shares his perspective on utility power constraints in key real estate markets, the super-sizing of data center buildings and campuses, and how that has altered many of the traditional approaches to data center site selection. We also examine the influx of capital and new players in the data center market, whether there is a bubble developing, and why experience matters in delivering hyperscale campuses.
David also offers a glimpse into what datacenterHawk’s research suggests about which markets might emerge next as data center destinations. Some of them might surprise you. Here’s our conversation.
00:00 – Introduction from Rich and David’s background
02:03 – How the AI boom has changed the data center market
04:55 – Following the power, and the new geography
07:25 – The growing role of natural gas in site selection
09:45 – Is there a bubble? Why execution matters.
12:45 – The outlook for consolidation and M&A
20:25 – What are the interesting emerging data center markets?
24:40 – Community relations and resource impact of new projects
26:21 – The Road Ahead for Data Center Supply and Demand
At Data Center Richness we demystify technology and explain why data centers matter, and what comes next for the Internet, cloud computing and AI. Host Rich Miller is a journalist and analyst who has been covering the data center sector for 25 years.
Subscribe to our channel for more insights:
https://www.youtube.com/@DataCenterRichness?sub_confirmation=1
LINKS FROM THIS SHOW:
datacenterHawk
datacenterHawk on YouTube
TRANSCRIPT:
Rich Miller of Data Center Richness: Welcome to Data Center Richness. I’m Rich Miller, and I’ve spent 25 years telling the story of data centers, cloud computing, and AI infrastructure. In this podcast, I’m sharing conversations with the innovators building our digital future. Now, here’s our show. David Liot, welcome to Data Center Richness.
David Liggitt from datacenterHawk: Thanks, Rich, it’s great to be here, great to see you, and very excited that you’re pursuing this path of communication. It’s going to be awesome.
Rich Miller of Data Center Richness: Well, thanks so much. You and I have both been tracking the data center industry for a few years, but things have been different lately. We’ve kind of entered a new era of data center growth driven by AI workloads that really kind of kicked off with the arrival of ChatGPT and its growth in late 2022. Since then, some things have been different both in terms of volume and where data centers are located, so I thought a good starting point would be to get your take on what’s changed since AI has become the new hotness.
David Liggitt from datacenterHawk: Yeah, you know, we’ve been at data center conferences for almost like decades now where we’ve talked about the internet of things and 5G and all these different self-driving cars, all these different technologies and applications that could shift the data center industry dramatically. Really, it’s been AI, AI has been the one that has had the biggest impact on our business as a whole. You mentioned it, 2022 was really when we started to see the beginning portions of the growth that would take place.
A couple of things related to AI growth and how it’s impacted the data center market:
- It’s shifted the geography where data center locations or data center markets end users have traditionally built. For years, we tracked 10 major markets in the US, and that has shifted dramatically over the last two years because a lot of the growth has been centered around machine learning projects that can really go anywhere because they’re less latency-sensitive.
- The size of these projects has changed dramatically. Five to ten years ago, data center projects were 10, 20, 36, 72, 150ish megawatts. Now it’s shifted to seeing projects that were 300, 400, some up to even like a gigawatt or gigawatt and a half of power.
These have been the two largest changes in the way the US data center market has operated. It’s pushed demand into places historically that we haven’t seen it, but it’s also driven demand in the major markets as well, so the Dallas, Phoenix, Northern Virginia, Chicago, Atlanta have certainly grown.
Rich Miller of Data Center Richness: And with the sort of extraordinary volume, there’s been a lot of interest in following the power, as you guys have talked about very much. How has that changed how people think about location?
David Liggitt from datacenterHawk: Yeah, it’s changed it in a number of ways.
- First would be how quickly can the power be delivered, and those timelines have been extending now for the last three years. We’re not seeing areas in the US where the power story is getting a lot better; it typically is getting a lot harder. This creates an opportunity to differentiate yourself if you’re a data center developer or provider.
- It’s really shifted how people look at investing in certain areas and sites. It’s put more emphasis on speed to market related to power perspective, but also alternative energy. Those of us in the data center space have had to almost become energy experts as there’s now a path of interest in different energy sources, most of them off the grid, that could potentially power data center projects. This is a whole new part of our industry that five to seven years ago was a lot less interesting.
One story I tell related to the changes in the market is from my time at CBRE, doing site selection projects. We always had a long list of risk assessments for each site. One thing that would always kick a site out was its proximity to natural gas pipelines. Any proximity meant you would never want to utilize it. Now that’s a positive; people are looking for sites that are close to natural gas pipelines for alternative energy opportunities. The market has shifted quite a bit, and these are some of the things impacting how people are looking at one site versus another.
Rich Miller of Data Center Richness: I wanted to ask about the natural gas in particular because that seems to be a larger piece of the data center energy landscape now. And even I heard some discussions this week about thinking about natural gas basins as places that suddenly loom larger in the data center landscape. I mean, Pennsylvania is a state that I wouldn’t have thought just a little while ago is getting attention, but now it is. Does that track with what you’re seeing?
David Liggitt from datacenterHawk: Yeah, I think if there’s areas around the US that are rich in natural gas, today those would be areas considered very interesting for data center development.
There’s still a chasm to cross as it relates to end users really getting comfortable with this. The top 10 to 15 end users have significant sustainability goals that are almost clashing with the challenge of using some of these alternative energy sources.
Last year, I talked to someone focusing on financing alternative energy projects. I asked if you could solve this with one type of energy, what would it be? He quickly said natural gas. His comment was that although the barrier for getting those users to embrace it is still challenging, it is the one energy source that we have enough of here in the US and it is readily available. This was an interesting moment for me. While other alternative energy types are worthy of evaluation, some have different timelines, community pushback, or financial implications. You’re right, natural gas is certainly a far more interesting discussion today than it was three years ago, and if areas of the country like Pennsylvania have an abundance of this, chances are you’ll see data center developers kicking the tires on sites in those areas.
Rich Miller of Data Center Richness: One of the things that’s been interesting is that there have been so many announcements of gigawatt and multi-gigawatt projects. You guys had a great data point that there are now, I think, 34 markets that have over 500 megawatts coming. And just all of the announcements have given people some concerns, a little bit of anxiety, about, “Well, is this a bubble? What’s going on?” I noticed that particularly after the PTC event, one of the big conferences in our industry, it seemed like after that, some of the discussions coming out of that were much more about who can really execute on all of these. You’ve talked about the importance of teams that can deliver in this environment. What do you make about where we’re at and what the path forward is going to look like with all of these announcements, and at the same time, all of the traditional providers who are trying to get work done as well?
David Liggitt from datacenterHawk: Yeah, I mean, the data center industry is a great market to be in. We’ve known that for a long time, but with the demand growth that’s taken place over the last several years, and with other commercial real estate asset classes maybe not performing as expected, the data center space has kind of been the shining star of all these different places. A lot of people have noticed what has been taking place in the data center industry. So when that happens, you get a lot of interest from other investment groups and different developers that want to get into the space.
From our perspective, we’re really excited that there are new people entering the space. Number one, we should all be excited from a labor perspective because that’s one of the biggest challenges our industry is going to have moving forward.
Number two, it’s important to do due diligence on some of these projects to make sure that they actually have agreements with power providers and that they can perform. That’s the question that data center users are asking. You won’t be able to fake out a large data center user; they do this themselves, so they’re going to ask a series of questions pretty quickly to understand if these site developments are real or not. When there’s one person behind a 4 gigawatt development, people will scratch their heads and go, “How realistic is an opportunity like this?”.
This isn’t to discourage those who want to get into the space, but we’re talking about hundreds of billions of dollars of investment coming over the next year, let alone five to ten years. I would just encourage anyone getting into the space and putting developments together that’s never done it before: credibility, capital, creativity – those three things are what you have to have in order to get these deals done.
Rich Miller of Data Center Richness: It seems like you mentioned the things that are sort of requirements: capital, having a team, having power, and in some cases, having customers with whom you have that credibility. And obviously, some of the people who’ve been doing this for a lot of years have all of that. It seems to me that a lot of the new players are backed by folks who have lots of capital and plenty of experience, maybe they’ve got the right land. I’m wondering if there isn’t a little bit of a consolidation going forward as maybe folks who may not have all of those things lined up, but might have a site with potential, wind up becoming candidates to work with someone else to get it to the finish line?
David Liggitt from datacenterHawk: Yeah, absolutely, that’s a great point. There are going to be clear winners in this space, there are going to be some losers; that’s how these industry significant growth time periods go.
I think there’s opportunity to put together joint ventures (JVs) with different groups that might not have all the things they need but have enough to become a partner with some of these larger data center operators or developers or even end users themselves. You just have to be open and realize, like we talk a lot at our company about humility, and our view of that is just being open and honest about what your strengths are and what your weaknesses are.
With the development side of things, you really have to understand your strengths and weaknesses, and then how to position yourself with groups that might be able to fill those voids so that you can partner. Or, there is going to be some M&A that allows these groups to succeed, and I think that’s a really important part of our business today.
Rich Miller of Data Center Richness: It used to be back in the day where it felt like there was a pretty high barrier to entry to the data center industry, both because a lot of capital was needed to get a lot of these projects going, and because it wasn’t just real estate, there’s this technology piece, having the right team in place, it was like just not just anybody could come in and do it. And that’s still true to a certain extent, but it also seems like there have been some new players coming in that have been able to make a pretty big impact quickly, partly because of the shifting tides in the energy space. What’s been your experience on that, and do you think that’s something that will continue?
David Liggitt from datacenterHawk: I do think it’s something that will continue. I think the key there is picking your partners as well. The data center industry has certainly expanded over the last three to four years, but it’s a small space; most people know each other.
Typically, when we’re evaluating management groups of development projects, we want to understand who’s behind the project, what their experience is, and what they’ve done. That’s a big part of it. So, partnering with groups that have been there before, have done the work, have executed, that’s one of the things I think the end-user groups are looking at more than anything today: how are you going to deliver this project on time and on budget?. That question has never been harder to answer than it is today in our space. That’s not a shot, that’s just the market conditions have changed; it’s got so much harder to really understand what you have when you have a PPA signed from a power provider that says they’re going to give you 200 megawatts by this date, the question is, is that really going to happen?. They could have their best efforts to do that, but the reality is there might be changes that take place that don’t allow them to perform like they said they would.
A lot of it is making sure you are partnering with the right groups. If you look back over the last three years, the groups that have come in and made the biggest impact have understood both the macro and the micro trends happening in our space. They’ve picked partners that have allowed them to move fast and really develop with credibility, and I think those two things have made a big difference.
Rich Miller of Data Center Richness: With this sort of expansion of new players, new technologies, new energy sources, I don’t know about you, but this has always been an industry where you learn something new every day. And some days, simply to keep up, you’ve got to learn about a lot of new things. So I’m wondering, you mentioned Data Center Hawk, you guys are ten years in, you focus on helping data center professionals make better decisions. What does the decision-making landscape look like now? How has that itself changed? And how have you and the Data Center Hawk team tried to adjust what you do to meet the needs in the market?
David Liggitt from datacenterHawk: Yeah, I think any data center investor, provider, end user, consultant, any group that’s going to go out and spend a hundred million plus, a billion dollars, five billion dollars, ten billion dollars, they’re going to do that, or they should do that, with great data.
As the cost to make these decisions grows, the importance of having really great data to help support your decision-making is a really valuable tool. We’ve seen that the data points have changed over time. It’s certainly more about understanding how much power certain regions can accommodate over a period of time. It’s understanding all the components that it takes from a supply, demand, pricing standpoint to make a decision where you want to grow. Sometimes we talk about it’s not necessarily where everyone is, it’s where everyone isn’t, so it’s just understanding the landscape.
The market is changing faster than we’ve ever seen. With all that, we’ve found that the common thread is that people need data to make great decisions. We’ve really tried to design a platform that helps people do it quickly, build that framework quickly so that you can feel comfortable about going and spending a billion, three billion, five billion dollars. These projects are so capital intensive. When you hear a company raising a billion dollars, you might think, “Man, that’s so much money,” but in the data center world, that’s like one project, one development project essentially, before they’ve got to go back and say, “We need another billion or five billion more or ten billion more”.
So, what we’ve had to do over time is just change the approach, change the menu of data that’s out there for our customers, and then really recognize that this is a global business now. We talked about data center markets in the US 10 years ago—Dallas, Chicago, Phoenix. Now we’re still doing it here in the US, but it’s London, Frankfurt, Singapore, Sydney, Brazil, Spain. It’s a global business. So, that’s been a part of our push too: having team members in markets around the world, boots on the ground, to help our customers really understand what’s happening.
Rich Miller of Data Center Richness: Along those lines, you guys having the sort of data-focused approach, one of the things about looking at the data is that it can give you some early clues about where things are headed. Everyone’s always interested in what the hot geographic markets are. You mentioned you’ve got the global perspective now. What’s the data telling you in terms of what some of the interesting markets might be, first in North America and then maybe globally?
David Liggitt from datacenterHawk: Yeah, I’d say the interesting markets in the US are ones that have power and have it in 2025 and 2026.
More broadly speaking, the major markets still receive a significant amount of growth. One trend we have seen is new submarkets appearing in larger major markets. This is for a number of reasons, including a term that I think the industry is talking about or should be talking about more: site fungibility. Can a site be used for both AI or cloud deployments?. The end users who are buying land to either mature a cloud portfolio or put inference sites on the ground are looking at site fungibility in major markets across the US. So, we will see continued growth there.
Somebody told me the other day that Dallas is the most interesting market right now, and we’re seeing more people look at this market. If you look at Northern Virginia, I always say it’s almost like it’s becoming “middle Virginia” because it’s moving down the I-95 corridor. Seven or eight years ago, we were talking about Ashburn and Manassas and whether Manassas was too far. Now we’ve gone four or five times that distance south. Frederick, Maryland, for example, has a really interesting development, and that’s an example of a submarket in a larger market. I think you’ll see that globally, especially as data center operators, investors, and cloud service providers wrestle with how deep into the language learning model game they want to go, how far out into these markets they want to go, and how many chip refreshes they can get in a facility before the facility design is out of date.
Rich Miller of Data Center Richness: I’m having to learn all new county names in Virginia, just about, getting going all the way down to the border with North Carolina. It seems that everybody’s looking for opportunity there, but I think that’s been part of the state-level strategy, particularly with some of the deals they’ve signed, is to spread it out a little bit and see if other parts of the state can see some of the benefits that particularly Loudoun County has.
David Liggitt from datacenterHawk: Yeah, and I think one of the challenges as we move into these secondary markets and some of these big projects get done in cities that we never thought would be data center markets, is really like project two, three, and four. Because the initial projects coming in are so large, if a two-gigawatt project goes into a rural area in Mississippi or Louisiana or Arkansas or somewhere that traditionally has not had data center development, then the question is, “Okay, well, what’s left for project two or project three or project four that comes in?”.
I think that will be an interesting thing to watch: how mature can some of these secondary or tertiary markets get, only because a lot of times that first project can take the infrastructure that is there from a power perspective or even like a water perspective. That’s an interesting thing to watch too in some of these areas: the impact that the data center growth has not just on power, but on other community things, and how much more data center developments will these secondary or tertiary markets really want. That’ll be an interesting thing to watch.
Rich Miller of Data Center Richness: Because communities are noticing, and they know to ask about this upfront now.
David Liggitt from datacenterHawk: Yes, they’ve gotten smarter, and they’ve gotten more organized on how they might push back against data center projects. So, data center end users and developers are doing a much better job, in my opinion, on interacting with the community and trying to think through where they go and the impacts that might have and what they can do to help the community in that process. But it is a back and forth, and I think we’ll see that continually in the US, especially as the projects get larger.
Rich Miller of Data Center Richness: Yeah, I think so, because when you see some of the numbers, people are like, “Well, what does that mean? You know, what does a gigawatt look like in our community, and do I want to live next to it?” Those are real questions. I wanted to close with a question that I think is on a lot of folks’ minds, which is, right now, and for a while, it seems like there’s been sort of an imbalance between the supply available and the demand from folks who want AI. Is that imbalance going to change, or are we going to be almost fully pre-leased to the horizon here? What’s your take about what the next few years look like? You’ve got to be careful with crystal balls, but.
David Liggitt from datacenterHawk: Sure, you bet. I’m sure we’ll look back at this in two years and go, “David was completely wrong,” but I think for the next through 2027, we’re in this environment. It’s really hard to change anything within a two to three-year window, given the macro and micro characteristics in these different markets. So, I see the next few years beyond that, it really depends on how many groups are willing to first pay the new amount of money needed to do the power studies, and then pay the money that it’s going to take to actually fund the power to get the site up and going. It’s one thing to be buying a site and trying to flip it for data center development, it’s another thing to be doing a power study, and it’s another thing when they say, “Okay, we need $40 million for a new substation.” That list of people that are interested goes from here to here because they don’t want to put up that $40 million or might not have the capital to do that.
The good news about the data center space is that if this was the office market, it would be oversupplied in like a year and a half, and it would totally change. The capital is so intensive on one of these projects—a billion dollars for one project—that it’s really hard for groups to get ahead of themselves. Now, I say that, the industry has been ahead of themselves before, but not significantly. In my almost 20 years of doing this, there are not times that I can remember someone having a 36-megawatt site sitting there available to be leased; people don’t release the capital that way, they do it in tranches and different phases.
I think if everybody could keep their head screwed on straight, the market’s still going to be in a really good position over the next three or five years, even in the midst of some of the user readjustments with their demand. There are some changes happening, and I think those changes are good; it’s maturing the market, it’s creating opportunities for other users. So, long answer to a short question, but that’s really what I think. I think over the next two or three years, we’re in this market, and then after that, we might see a little a softer market, it might lean back towards the end user, but it’s going to be a while.
Rich Miller of Data Center Richness: Well, ask a lot of short questions, but the data center market these days is not easy to break down to a snippet. Always appreciate the data and market studies that you guys do. If folks are interested in learning more about Data Center Hawk, where can they go?
David Liggitt from datacenterHawk: Yeah, datacenterhawk.com is our site. We’re a subscription-based platform for data center investors, providers, end users, consultants, vendors, so you can jump on there. Our team is very responsive. We also do a lot of YouTube podcasts and things like that, so I think you can find us probably on any of those platforms. But Rich, as always, appreciate the opportunity, very excited about, I think it’s Data Center Richness, I think that’s what you’re calling it.
Rich Miller of Data Center Richness: Business that is. I love it.
David Liggitt from datacenterHawk: Well, anyway, we’re big supporters and excited about your days ahead.
Rich Miller of Data Center Richness: Well, David, thank you very much for your time, and thank you to all of our listeners. We appreciate you and your interest in data centers. Stay tuned, and we’ll be bringing you more good stuff on Data Center Richness.
