‘It’s A Fine Line’: PowerHouse Exec On The Risks Of Building In Small Data Center Markets

PowerHouse Data Centers has announced hundreds of megawatts of new projects over the past two weeks, much of it outside the industry’s largest markets. 

PowerHouse Senior Vice President for Asset Management and Development Matt Monaco, who joined in July from data center giant Equinix, says building massive campuses in smaller markets means walking a “fine line” between more speculative, higher-risk land deals and tenants’ growing need for power.

PowerHouse, a subsidiary of American Real Estate Partners, last week unveiled plans for a 300-megawatt data center campus in Charlotte. Through a joint venture with real estate investment firm Town Lane, PowerHouse plans to build 2.5M SF of data centers across five buildings by 2027, with the potential to increase capacity on the site to as much as 500 MW.

The 122-acre site sits 10 miles from downtown Charlotte, a city that, while not a digital infrastructure hinterland, barely cracks the list of second-tier data center markets. 

This came just days after PowerHouse announced a separate joint venture with fellow data center firm Chirisa and asset manager Blue Owl to build a series of data centers anchored by artificial intelligence cloud provider CoreWeave. In a deal that could eventually see the deployment of more than $5B, the JV’s list of development sites includes locations like Pennsylvania, Kentucky and Nevada, with the first project a 120-megawatt facility on 350 acres near Richmond, Virginia. 

PowerHouse’s suddenly swelling development pipeline builds on a portfolio that began in the heart of Northern Virginia’s Data Center Alley but has steadily expanded into secondary and tertiary markets. The firm has 30 data centers in development or underway, representing 2.3 gigawatts of capacity. 

In an interview with Bisnow last month, Monaco discussed the company’s growing footprint that is increasingly pushing outside of primary data center hubs. He said delivering data centers in these emerging markets requires a balancing act that limits risk while anticipating the needs of the largest Big Tech tenants. 

This interview has been edited for length and clarity.

Bisnow: You just started at PowerHouse over the summer after a long tenure at Equinix. These are two companies with really different business models, development philosophies and roles within the data center ecosystem. Has that been a challenging transition? 

Monaco: I’d spent 10 years at Equinix, running the asset management group there. Equinix plays mainly in the retail data center space, small cages focused on specific types of workloads, almost all in major metro areas, selling small chunks on a retail basis. That is a great business, and that will continue to be a great business.

But what's exciting about PowerHouse is that it’s focused on this transition from campuses of 100 MW with a few buildings of 20 or 30 MW apiece to now the buildings being 100 MW, and the campus is maybe a gigawatt. That’s one of the reasons I’m here. 

These gigawatt campuses are needed not only to support AI and this massive AI demand surge but also just to sustain growth of the standard cloud. AI is just another layer on top of an already very fast-growing business. That’s really exciting to me. If I can spend the next 10 years helping to enable that, I'll be very happy. 

Bisnow: I’m curious about PowerHouse’s development and site selection strategy. Your portfolio and development pipeline are spread across a diverse range of markets, from the heart of Data Center Alley in Northern Virginia and other primary markets to the Carolinas, Reno and places like rural Spotsylvania, Virginia, that are way outside the box. What’s the driving principle or North Star when it comes to where you’re looking to buy land and build?

Monaco: Speed to market is very significant, but it's also about other factors. We have relationships with a lot of the major players, and it's about understanding them and their needs. That means thinking about different kinds of workloads as it relates to geography.

For example, Ashburn in Northern Virginia has a vacancy rate so small you can't even see it on the charts comparing it with other markets. And so, if you were to offer a big player 50 MW or 100 MW or really any quantity in Ashburn, they’re going to jump to take that. 

However, as tenants move towards more AI workloads, some of that is less sensitive to latency. So we try to think about the path of development: Where are other locations where they're close enough that they’re still going to be suitable for most of these other workloads? In the case of Northern Virginia, that's very clearly coming down Interstate 95 where there’s a lot of fiber, places like Spotsylvania and Richmond. 

The Northern Virginia market used to have a certain radius. Now maybe the radius of the circle gets a little bigger. Crystal balls aren't available, but in a lot of global markets, it's easy to see where the opportunities are to expand the circle. Our goal is to be proactive, figure out where those opportunities are and lead the way there.

Bisnow: That balancing act when it comes to secondary and tertiary markets is interesting. On one hand, you need to bank land in locations that can get power quickly so that when major tenants say they need hundreds of megawatts in that region, you can meet their timelines. On the other hand, you don’t want to be too speculative and get too far ahead of demand, no matter how robust it is. How do you make sure you’re not overextending when it comes to land acquisitions? 

Monaco: It’s a fine line for sure. We talk to the tenants a lot. The relationships and the amount of sharing that happens is pretty significant. It’s really more of a partnership in most cases, where you’re looking at road maps together and figuring out where things are going. 

There’s always going to be some risk, though, of course. But in general, having so much demand out there is a luxury. There’s also a real premium placed on having a lot of capacity in one place — 500 MW or a gigawatt in one place is very attractive. For the largest tenants, being able to build in that scalability is really important. 

It’s also worth mentioning that this kind of development — and the different things that need to happen on these projects — are inherently somewhat modular. Take a switching station that needs to be built by the utility: Most switching stations are 300 MW, so even if you’re building a gigawatt campus, it’s not like everything goes in on Day 1. 

Bisnow: So you’re saying that you mitigate some of the risk by building these massive projects in tiers over time? 

Monaco: Exactly. It’s about figuring out places where the type of scale that tenants are looking for is possible, then, over time, figuring out the right way to meet those needs with one tenant or two tenants and being ready for different scenarios. 

There are some players out there that have gone after massive sites with thousands of acres and things like that. That’s not for us. I don’t want to say that’s a bad strategy — it might end up being an amazing strategy. But it’s not for us. 

Bisnow: In terms of site selection and willingness to go outside of major markets, to what degree is it a game of follow the hyperscaler? It seems like when Amazon, Microsoft or Google builds a big campus in a new region, third-party developers start snapping up land around them. 

Monaco: They do. A lot of times when you see a lot of operators move into a region, it’s just that the area has good infrastructure or there are good incentives like tax abatements being offered or things like that. But there’s the natural tendency for operators to follow the large players. 

Sometimes the thinking is just that if the big hyperscale tenants are in an area, they’re going to need more capacity eventually, so you want to be there and ready to serve them whenever that is. But the second part, which is more interesting, is how the network effect takes hold to turn a few hyperscale flags planted into a self-reinforcing loop.

There are real benefits to clustering, especially when it comes to fiber. You get a couple hyperscale flags planted in a new area due to incentives or water availability or something like that, and all of a sudden, you’ll have a lot of fiber in an area that didn’t have a lot of fiber in the past. All of a sudden, you’ll have a community of tradespeople and experts, and you get an economy of scale and significant value in clustering in the area.

Bisnow: At the moment, data center development is all about trying to get power quickly as utilities struggle to keep up with demand. I see firms taking a couple of different approaches to expediting power at their sites. Companies are working more proactively with utilities to jointly solve some of these issues. At the same time, some are taking more creative approaches, like on-site power generation, co-development with energy firms and “behind-the-meter” deals with power providers. What’s PowerHouse’s approach to energy acquisition? 

Monaco: With utilities, it's always about partnership. We have our business case, the utility has theirs, and it’s about finding the best risk-adjusted way to get there. This is a big part of our approach, to be very collaborative and make sure all parties are on the same team.

There's a lot of different ways to partner. In some cases, it's the small stuff. For example, if there are utility lines that have to be run to bring power from a high-voltage line to a data center building, we can divide and conquer on that. We’ll work in parallel where we focus on negotiating easements and things like that to help accelerate things while the utility focuses on the elements that only they can do. 

In terms of on-site generation, co-generation and that kind of thing, it’s very much on a case-by-case basis. It can be a bridge in the first phase of a project until utility power is available. There are others in the industry who feel that co-generation should be the next big thing and that every data center should have a giant power plant next to it.

Bisnow: That kind of project has been in the news over the past two months, with Amazon acquiring a data center campus abutting a nuclear power plant and two other nuclear-adjacent data center campuses being proposed. 

Monaco: The AWS deal with Talen Energy was a huge move for them. It's really the epitome of “behind the meter.”

With the speed to market such an important element in the landscape right now, in order to get the amount of power the industry needs in the amount of time we need it, the land has to be in close proximity to existing power infrastructure. 

We need to power fast, so we need to go to locations where that's possible. It takes a good 10 years to get new generation and new transmission capacity built, so bringing the data center to the power rather than bringing the power to the data center is how we’re thinking about it. 

We’ll continue to see creative dealmaking and solutions happening in the market to be able to get a lot of scale on accelerated timelines. The specifics will dictate what utilities are willing to do, but we’re going to keep seeing these types of unique arrangements. 

Bisnow: Utilities were really caught off guard by the surge in demand from data centers, with data center requests for grid connections doubling or tripling in many markets. But as utilities invest billions of dollars to upgrade grid infrastructure, there’s also concern that some of this data center demand is a mirage — that they are speculative requests that may never materialize. Is it possible to separate the wheat from the chaff here? 

Monaco: The first hurdle is just looking at who it is requesting power. If the requests are coming from some of the very largest users for a self-build, that’s not a guarantee, but it's a pretty good indicator that request is probably real. That’s especially true if there’s existing relationships and history with the firm making the request, as is often the case between the actual humans who are having these discussions.

That said, we’re talking about very large-scale campuses here, and there are companies out there asking for very large numbers of megawatts just to see what’s possible. I think that’s a tough strategy to take and may not always work out so well. 

For the utility, it’s about them understanding developers’ strategies and business plans. What are they really about? How informed are they? 

There are a lot of real estate developers out there who haven't done data centers before who are looking at this industry and thinking, how hard could it be? They’re jumping in and asking the power company for 500 MW. There's a fair amount of that kind of activity driving those numbers up.