Weekend Interview: DC Blox CEO Jeff Uphues On AI And the Next Data Center Boomtowns
This series goes deep with some of the most compelling figures in commercial real estate: the deal-makers, the game-changers, the city-shapers and the larger-than-life personalities who keep CRE interesting.
As a boy in the town of Jefferson, Iowa, the first job Jeff Uphues ever had was with a local electrician, digging trenches for power lines.
In a sense, digging trenches for cabling is what Uphues has been doing ever since across a 35-year career in communications infrastructure that, since 2017, has seen him at the helm of Atlanta-based data center developer DC Blox.
Since its founding eight years ago, DC Blox has focused primarily on building colocation data centers in small markets across the Southeast in places like Birmingham and Huntsville in Alabama and Chattanooga, Tennessee. The firm also has a significant footprint in the optical fiber business, building and operating networks to send data from smaller Southern markets into Atlanta, as well as developing its own subsea cable landing station in Myrtle Beach, South Carolina.
With record leasing demand spurring an acceleration of the data center development pipeline in the coming years, the industry is expanding beyond its traditional hubs and attracting new investors. Uphues and DC Blox stand to benefit from that expansion, and the firm is making its first push into a larger market by building a pair of campuses near Atlanta.
Uphues spoke to Bisnow about the flood of new data center development in small markets, how artificial intelligence is redrawing the data center map and what CRE professionals dipping their toes into data centers need to know before taking the plunge.
This interview has been edited for length and clarity.
Bisnow: DC Blox has a distinct model within the colocation data center space. You build and operate facilities almost exclusively in small markets in the Southeast, but a significant part of your business is focused on building out optical fiber network infrastructure. How do these two pieces fit together?
Uphues: We build the buildings that house the computers that power the internet and the cloud. But the second thing we do is we interconnect all of those facilities together with networks and fiber. Our mission was to build purpose-built facilities in markets that did not have the type of infrastructure that you’d find in large markets like Atlanta. Our value proposition is that we can make tenants feel like they're in Atlanta by the network connectivity that we put in place for them. If you build a data center that doesn't have access to the internet, it's an island. So, we connect the island with bridges and roads and ramps that allow fast traffic in and out of these places, connecting our customers to the main core markets where the internet is accessed.
Bisnow: Have you given any thought to expanding DC Blox’s footprint outside the Southeast?
Uphues: We’ve built data centers capable of handling large-scale enterprises, city, county, state and federal government entities as well as universities and the cloud companies that are now filled with tons of different businesses that need a facility in these markets. The Southeast is the fastest-growing region in the country. So why should we focus on any other place when you look at all the markets here where we can be the single source for where all content, all cloud, all carriers, all wireless, all enterprises and all governments connect through? By having that type of facility there, we’ve become the de facto choice for anybody putting their computing infrastructure into those markets.
Bisnow: Over the past two years the data center demand landscape has shifted in your direction. Tech giants like Microsoft, Amazon and Google, the industry’s biggest tenants, have started building and leasing significant blocks of capacity in the so-called edge markets that are DC Blox’s specialty. Yet as this is happening, DC Blox is going the other way, making its first move into a major market with hyperscale-anchored campuses near Atlanta. What’s driving this strategic shift?
Uphues: We built in these edge markets knowing that the cloud customers would ultimately come into those markets. And we also then said, instead of going to the edge, let's build to the core because we’d established a track record of performance. Who was going to trust us if we came out and we said we're going to just build up a 500K SF facility and put in 100 megawatts? No one. We wouldn’t have had an established track record of doing that. When we started in these edge markets, we had to prove ourselves. Once we’d built five data centers on time and within budget for the most discerning customers in the world like major enterprises and the cloud companies, we earned a reputation that preceded us. Can they trust us to build another, larger data center? The answer is yes.
Bisnow: There’s a bit of a frenzy around AI right now. Major tech companies are making big investments in the data center infrastructure to support these technologies, which Nvidia’s leadership believes will double the global data center supply within five years. Is AI moving the needle for you guys at all as a demand driver?
Uphues: We're involved in AI discussions with customers every single day. Two years ago, we set out to really embrace artificial intelligence, which people in the data center world refer to as ‘high-performance computing.’ We decided to build a high-performance computing center within one of our existing data centers in Birmingham, Alabama, that mainly served research customers like university systems. Well, now that there’s all this AI investment, we’re prepared to handle that next generation of high-performance computing and can now drop these deployments into any one of our data centers.
The conversations around this have been amplified at least tenfold with customers. We’ve invested a lot in our engineering and design capabilities to meet the demands of the largest cloud providers. Our designers and engineers talk to those companies literally every day about how they want things done, from the weight of the equipment to the amount of power to the chip.
Bisnow: For the data center sector more broadly, AI is helping fuel record demand and unprecedented development pipelines. But at the same time, power constraints in the industry’s traditional hubs like Northern Virginia, Silicon Valley and Chicago are pushing development into new markets. As the AI demand wave grows, how is it transforming the data center development map?
Uphues: I helped build the infrastructure for the first iteration of the internet thirty years ago. Now, the cloud companies are reimagining and rebuilding the internet in their image. The general notion until recently was that you needed to be in the big markets. But as it has become harder and harder to build in these markets, largely because of the availability of power, you have large-scale data centers that are as large as a gigawatt moving out to wherever there is available land and power.
The future interconnection hubs are going to be at these massive cloud clusters instead of legacy facilities in major markets. All the internet traffic is going to revolve around where these large-scale data centers go. It's now much cheaper to move the network to where the data centers are than to put the data centers where the network is. So, developers went to where they could get a lot of land and a lot of power and are bringing the network to it. That is what has really changed a lot.
Bisnow: So, in other words, it now makes economic sense for developers and major tenants to invest in building out the fiber infrastructure needed to site facilities far from major markets, as long as there’s power available. What are some emerging hubs that are going to see a significant share of this development?
Uphues: You're never going to replace the main core centers like Ashburn, Virginia, Dallas, Chicago or places like Columbus, Phoenix, Santa Clara or Hillsboro. But Reno is growing really fast now because of the availability of power there. Utah has been big. I wouldn't be surprised to see more data centers in the Ohio Valley. A lot of computing infrastructure is going in markets like these. The point is, it can live anywhere. If you have access to power you can bring the network to it. But if a market already has lots of network infrastructure and traffic, that does make it particularly amenable to being the next big place. I wouldn't be surprised to see more data centers in places like the Carolinas or Georgia where there’s access to cable landing stations bringing traffic in from around the globe.
Bisnow: There’s so much more interest in the data center space from across the CRE landscape than there was just a year or two ago, whether it’s new investors trying to fund new projects or developers from other sectors exploring moving into data centers. Are there any common misconceptions you regularly encounter or specific things that newcomers to the space need to understand?
Uphues: There’s a real art and a science to finding the next data center sites, and it really is finding a needle in a haystack. It’s not very easy. Just because you have large transmission lines near a piece of property and it looks like it’s flat and like you could build a data center there, that’s not always true. That’s something I hear all the time. People will come to me and say, “I’ve got this great site, it’s flat and it’s got power.” Well, how do you know it has power? Just because it has transmission lines next to it doesn’t mean it actually has available power. Is there a substation nearby or do you need to build a substation? And then there’s the question of whether you have access to water on the site, If there are environmental issues or things you have to worry about with zoning. What’s the local tax incentive structure and how close is the site to fiber networks? There are so many things that you have to investigate. It’s complicated.
Bisnow: We’ve talked about power constraints, but another emerging challenge for the industry is growing community opposition to data center development, particularly in certain major markets like Virginia. When it comes to executing community engagement strategies to head off this backlash, data center developers have been behind the eight-ball compared to other sectors of CRE. But I know you just came from the Georgia state house, where you were advocating for data center tax credits. Has community engagement been a big part of your development game plan?
Uphues: We get very involved politically and very involved on a local level in all the communities we’re in. Whether it's city councils or mayors or congressmen or state senators or state representatives or governors, we take a really proactive approach. Understanding a community’s issues is really important to us, and it helps us in many ways when it comes to things like permits, incentives or economic development. I don't want to say we have a perfect track record of unanimous voting when it comes to having DC Blox come into a market, but it's pretty darn close.
Part of it is that we have spent decades in the communities in the Southeast. We know how to do business here. Would I use the same tactic if I was in New York City? Probably not. But if you do a lot of engagement, you're likable, you understand local issues and spend time in these communities and talk to people, these are the things that work.
Bisnow: Give us a bold prediction for the rest of the year.
Uphues: There’s going to continue to be unprecedented demand over the course of 2024. No one is about to stop using any of the applications that have driven growth. We’re not about to do less video conferencing or use fewer streaming services. Then you add in AI and the growing consumption of data from things like wearables. We have 10 years of things ahead of us that are just unprecedented. It’s the opportunity of a lifetime to build in the data center business.
Bisnow: What is your weekend routine or favorite weekend activity?
Uphues: When I need a clear head and have the time I get out and road cycle. I ride with groups of 20 or 30 people, but sometimes we have groups of almost 100. It’s like the Tour de France. I'm kind of a big guy so I’m known as the wind wall where people will just get behind me and draft. Most of these cycling guys are not 6-foot-4 and 240 pounds like I am. They're more like 5-foot-7 and 160. But I've got pretty strong legs, and I can pull the peloton pretty well. I can send it for a while before getting tired and dropping to the back.