#329 - Marc Ganzi - CEO @ DigitalBridge ($DBRG) - $70B Company Investing Globally in Digital Infrastructure

Chris Powers: Let's start with how someone got into this business. You got into this as soon as it was not a thing. How did you find your way into the digital infrastructure world? 

Marc Ganzi: Well, like all great stories, it's accidental, right? And so my journey began in 1994. I came out of business school and focused on buying distressed real estate.

Then, the Resolution Trust Corp was unwinding, and a series of distressed real estate portfolios were being sold off. The real estate investment firm I was working for bought tranches of these portfolios of assets, and one of the tranches of assets came with a series of what I would term Class B office buildings and the Mid Atlantics.

One of the office buildings was a tall building with a massive radio tower with a series of antennas. And we looked at the rent rolls, and someone said, well, the rooftops are doing, you know, a million dollars a year in revenue, and the rest of the buildings are doing like 2 million a year in revenue.

We said, well, that doesn't make any sense. So we started looking at it, and it turned out that there was this unknown white space of, you know, antenna real estate or rooftop real estate. And it turned out that those antennas belonged to TV stations, radio stations, and these new cellular telephones.

And we needed to understand how to negotiate. So we took this building out of bankruptcy, and these customers came at us and said, we want you to confirm our leases because, you know, in a bankruptcy process, you can wipe out the leases and start over again. After doing a little research, I found out two things.

1, once somebody puts an antenna location and a spot, it's tough to move it. So you have great tension as a landlord with the tenant, which is that you can control the outcome because of the amount of electronics and equipment they put into their space. It's hard to pick up all that stuff and move it elsewhere.

And if you have a perfect location, you can't move it anywhere else, so we found that out. The second thing I learned was that I needed to learn about the technology. So, I found some guys who went to Wharton and knew something about it. And they were telling me about this thing called digital telephones.

So, cellular telephones were analogue, like these vast bricks you'd pick up, carry, and walk around with. They said you don't get it. There's this whole new wave of digital technology coming. And these, you know, cellular phone companies are going to need hundreds of thousands of locations called towers.

And the light went off in the head, and I had that kind of Eureka moment, which was like, wait a second, you mean they'll need 200,000 of these locations in the next five years. And these two guys they met were like, yeah, and we're, you know, we did one of the start-ups, we built one of the first cellular telephone companies.

They said it is better to be on the real estate side of the trade here instead of owning the technology and building the cellular phone networks. And so we all agreed, we got together the three of us who wrote a business plan to create the first tower business in 1994.

And just because I was a real estate guy, I accidentally bumped into a bunch of antenna leases. It changed my life forever, and so began this 30-year journey with, you know, digital real estate and digital infrastructure in 1994. We built our 1st tower company in 1997, we made our 1st fibre optic cabling business where we put high-speed Internet and office buildings, and, you know, 30 years later, we've owned and operated 50 different companies around the world. You know, all myopically focused on digital infrastructure and digital real estate. 

Chris Powers: Regarding those towers, you said there would need to be 200,000. How many exist today in America? Was 200,000 really what we needed? Or was that 1994 thinking?

Marc Ganzi: That was 94 thinking. Today, there are over 350,000 tower locations in the US. When we got started, there were probably somewhere in the neighbourhood, about 20,000 towers in the United States. So it's been, through three decades, there's been a lot of digital transformation. And to do that requires, of course, more and more and more antenna locations. So right place, right time, good industry.

And we're still growing because, as you know, it's about more than just 5G technology. It's about artificial intelligence, cloud computing, Internet of things. There's so much going on in our sector today. It's more, how should I say this, more cross Polonized than ever because cellular networks talk to the cloud and the cloud talks to the fibre optic cabling and the fibre optic cabling comes back to a private 5G network for an enterprise and more and more a lot of this stuff.

It's software-defined, so the cloud is the bond that ties it all together. We are working in these different verticals, building all this infrastructure to support the digital economy. It's exciting. It's a big 13 to 14-trillion-dollar marketplace now, and 30 years ago, it didn't exist.

Chris Powers: Has defining where a cell tower should go back in 94 changed today when you're thinking about location based on all this cross-pollinating or do the same location metrics matter when you're looking to put a cell tower somewhere? 

Marc Ganzi: Three things have changed in the last 20 years.

One, it's challenging to zone these things. Most community boards or city councils want to avoid seeing a cell tower. So visually, they could be more attractive things. So, as you can imagine from a zoning and planning perspective, nobody wants to sell a tower in their backyard.

So that makes it harder to build these things than it was, you know, 20 years ago, even if you explain the public benefit and public safety and everyone, you know, everyone needs connectivity at the end of the day, you know, the sort of moral minority shows up at a zoning board meeting, 100 to 200 citizens out of, you know, maybe 20, 000 show up and say, we don't want to sell tower in our backyard and.

So, you can imagine city councils, listen to that stuff. Zoning has gotten tougher. The 2nd significant change is that the location of the technology and the radius are getting smaller and smaller. So, cell towers were spaced 5 miles apart. Then we got to 2G, and it was 3 miles apart.

Then we got to 3G, a mile and a half apart. 4G a mile apart. In 5G, you know, they're less than half a mile apart—the cell towers. So, as the technology has improved, the radio propagation has gotten smaller and smaller. And so you need more and more towers, which is why you've seen the proliferation of these towers.

A radio spectrum wave travels in the air, and based on the speed of that radio frequency, that signal will drop off. And then, as it drops off, you need to build another tower. So it hands off. And you keep building. Hence why, it's called a cellular network. So, as the density has gotten more intense, there's been more demand for more towers, which makes our business harder because if somebody says I need a cell tower on the corner of the state, Maine. There's no real estate available in Maine, and you have a problematic zoning board, and it gets harder and harder.

So, density plays a big part in it. And then a 3rd thing that's happened, which is very interesting, is the height of the antennas; Chris, over time, has come down. So, 29 years ago, you'd see a cell tower that was 300 feet tall. You know, 20 years ago, you'd see a cell tower that is 150 feet tall.

Five years ago, you saw a tower, and it's like 80 to 90 feet tall. And now, in 5G, you're seeing antennas 25 to 30 feet off the ground again because of that density factor. So now we're seeing more of this technology increase, but we're seeing it at a lower height. And the antennas are a little smaller, and you can conceal it better.

And so the game is changing, and now it's more about the ability to show up and deliver at scale. And that's what we're doing with our business days. We're doing that here in the US, but we're doing it in Asia and Europe, and we're just meeting the demands of our customers. 

Chris Powers: I love it; I assume the same folks who show up to zoning meetings not wanting a cell tower are also disappointed when their internet speed or digital speed is not in tip-top shape.

It probably cuts both ways. Is it more accessible than putting cell towers, call it in New York City, where everything is a dense building, and you can hide the tower on top of a building versus suburban locations where there's not a lot of height? And so they're going to be more visible. 

Marc Ganzi: So, in the suburbs, we're building more pine trees, flagpoles, and other types of concealed infrastructure.

It could look better. You can tell a cell could tree better if you drive by one, but this hoarding board thinks that's a better look for some odd reason. And yeah, it does and takes away the visual pollution. And so we built cellular cactuses. We built cellular adobe huts. We made cellular palm trees, pine trees, flagpoles, church steeples; you name it.

In the 29 years, I've concealed every kind of antenna you can think of. Now, New York City, it's interesting because in New York City, we're putting the antennas down at the street level, so we're not putting them up at 40 stories or 50 stories. What we do is we use this new technology called small cells.

And small cells are small antennae we put on a street corner. We run the fibre underground, which carries the radio waves through optical light cabling. So, we run that RF signal through optical light. We bring it up the utility pole or the traffic signal, and then you'll see a small antenna on top of it.

And inside, we have a small radio, which transmits that signal. The only problem with a small cell, Chris, is it only goes one block in each direction. So, in New York City, we're putting these small cells literally every other street corner. So, we have a company that does that. One of the digital bridge businesses we own is called Extinet.

Extinet has over 4,000 small cell locations in New York City. We're the largest owner and operator of outdoor small-cell infrastructure in New York. And that's how cellular networks and mobile networks work in New York City. You have to have access to that small cell technology. So, there are different ways to build these networks now.

Chris Powers: Okay, so is the business, I guess what I'd consider building spec where you can look at the demographics, the population, the density and go, you clearly need cell service here and then you build it or are the primary carriers coming to you saying, Hey, we want locations in these spaces or is it a combination of both?

Marc Ganzi: It's a combination of both. When we're in New York City, we know every street corner where the carriers will need us. Being 1 of the two companies with the New York City franchise agreement is valuable. And, you know, sometimes we are responding to a request.

From Verizon or T Mobile or AT& T, or sometimes we're going in on a spec basis. So, for example, we have the small cell network at Madison Square Garden. So if you go to a New York Knicks basketball game or a New York Rangers baseball game, you're running on an Extinet small cell network because there are no cell towers; obviously, the garden today has over 400 antennas.

So, the guest experience. What the Dolans want to promote is that they want to make sure your phone works everywhere. They want to make sure that, because to get in there, your ticket is now on your phone, you order your food on your phone, you call merch on your phone, and if you're logged into the New York Knicks network, at specific points in the game, they may say, look, Julius Randle's shirt is on sale, go here to concourse 3. You get 20 per cent off the shirt.

The LA Dodgers are doing that by example. We have another company called Boingo. That works with the Dodgers ownership group with security benefits. And we put in a completely neutral host network that controls, you know, the Dodgers VPN, their virtual private network. And so, as you go to a Dodgers game and enter that VPN, your phone can click on the Dodgers network.

And once you're on that Dodger network, it's tickets, it's statistics. It's like interactive fan engagement with the players. It's cool; it's a great guest experience. And that's a private 5G network, but also on that network again, you've got T Mobile, AT& T and Verizon. So, all the mobile carriers operate on that one shared network. Still, the Dodgers, you know, own it in combination with one of our companies, and we're helping the Dodgers monetize, you know, that guest experience through their phones.

And it's pretty cool. We're just scratching the surface of how small cell infrastructure can impact. You know, the guest experience when you go to a sporting event, there's more coming there.

Chris Powers: And whether it's a small cell or a typical tower, does AT&T sign a lease or a contract, and I'm assuming T-Mobile?

Marc Ganzi: Absolutely. 

Chris Powers: What if it's like Boost Mobile or one of the smaller ones? Does everybody who wants to benefit from that small cell or that tower have to sign? Is it a lease, or is it a contract? How's it structured? 

Marc Ganzi: It's a lease. It's a lease agreement. So, on a small cell tower, the customer enters into a.

You know, five, 10, 15-year lease agreement with us. That lease agreement is like any other lease; 30 years ago, we structured it just like an office lease. We said, here's your demise premises. Here's the tenant build-out, and then you sign a long-term lease with us, and we're your landlord. So, we applied to digital for many of the skills that I learned at Wharton and, you know, attending the Zell Lurie real estate centre.

We said, look, it's a space at the end of the day. It's a space lease. And there's a certain amount of equipment they will put in place. And that's how the digital leasing business began, just applying traditional real estate principles to technology. And we do that today in our fibre business.

We do that today in our data centre and cell tower businesses. 

Chris Powers: And if you took it to real estate, you would assume AT&T, one of the largest carriers, is probably at Madison Square Garden; most of the people are using AT&T as opposed to Boost Mobile. What AT&T are they signing? It is an enormous lease since they will have more customers, then call it a minor boost.

Like, how do you gauge? Especially 15 years out, Boost Mobile doesn't become some enormous company where, you know, who knows what the future holds, but they become more fans are using Boost than AT&T, but when it started, it was reversed.

Marc Ganzi: So Boost runs on another person's network. It doesn't even run on; they don't have the spectrum that is their radios.

They buy capacity on other people's networks. And that's the first sort of salvo, which is the same thing with cricket or Metro PCS. Those brands below T Mobile and AT&T are more value-oriented. But at the end of the day, those networks now run on other people's networks.

So they have a wholesale agreement with a, you know, a T Mobile, like Metro, Metro PCS is a subsidiary of T Mobile, where Cricut is a subsidiary of AT& T. So you've got these situations where you have the high-end product, Verizon's or AT& T or T Mobile's product. Then you have these prepaid products, which are also very convenient: light contract and light touch on the contract.

You don't get the fancy iPhone 15. You should pay a little more. There are different Value points in the ecosystem, but ultimately, in the US mobile market, our big three customers are mobile: Verizon, AT&T, and Dish networks dishes. Now, they are starting to build their own private, you know, their own, public 5G network.

Chris Powers: Okay. And in a previous episode you were on, you said that the chessboard changes every seven years. And so, what have been the significant inflexion points over the last 29 years? As you look back, I can think of mobile 5G, social media, and AI. What have been the massive inflexion points along the way?

Marc Ganzi: The most significant catalyst for me was that jump from analogue to digital in 96. We had that first jump at a digital PCS. Another catalyst was the Telecom Reform Act of 97. When the U. S. government opened up, you know, the wireline infrastructure, anybody could compete and become a competitive local exchange carrier, which was the breakup of the Bell monopoly.

That was huge, the ability to create competition and to allow the cable companies to go into voice communications that changed the entire landscape of, you know, fibre Internet to the home and cable TV And vice versa, the guys that had the copper and fibre lines could go into cable TV and could provide, you know, TV service.

So that was another game changer. Many people talk about something other than the advent of the cloud, but in 2013, when a public cloud was launched, that was a big deal. Cloud computing has changed a lot of things, but it's changed how you and I communicate. It's created this portal where you and I are having this conversation.

It completely changed retail. You could argue that it decimated retail or disintermediated retail, but one thing that both you and I could agree on is that it changed retail. The Internet has completely changed how we shop and conduct commerce. And now, as we move into AI, it's probably right up there, if not the most significant step change in how we work business.

And how fast we're going to go. Ultimately, initial language-based models only do a little for you and me today, but as those models start learning, They become intuitive, and we move into inference; generative AI is a big deal because, in generative AI, you start to get enterprise based applications that start thinking and that move from language-based models into inference is enormous.

And so, there will be a significant change and, ultimately, software systems. Software systems are going to be a lot faster and a lot more intuitive, and you're going to see a massive shift in IOT networks. So, the ability to take the Internet of Things and then weaponize it for autonomous vehicles and public safety and all these applications will come through generative AI.

You and I can't even think of all these upcoming applications, but guess what? Chris, they're coming. They're coming, and they're going to change our lives. And so it's going to change the way you drive home. It's going to change how you go grocery shopping. Your house will tell you everything that's wrong with it.

In 2 to 3 years, you'll have a full diagnostics kit for your house where you won't have to guess what's wrong with your thermostat. You won't have to think about what's wrong with your Internet. AI will help you predict when those so-called pattern recognition moments happen.

So, it's scary, it's exciting. It is the most transformative thing I've seen in my 30 years doing this. The public cloud was the significant catalyst, but as we move from the public cloud into this AI realm, we're swamped because we're building a lot of the AI infrastructure.

We're building the data centres that power AI. We're building the fibre cabling that delivers that AI infrastructure. So it's been a fantastic, you know, 18 months for our business because we're tasked with being the axes and the shovels for this next generation of infrastructure. So it's exciting.

Chris Powers: Let's talk about data centres then, and we can start with just continuing on the AI. Will they be built differently? We need more capacity. But are there other ways this infrastructure plays out differently in AI than before besides capacity needs? 

Marc Ganzi: So, yeah, the short answer is yes. There have been changes. From the way we design data centres, the way we build them, and the way we operate them, the most significant change so far has been power. So, the power of the energy that comes into an AI data centre against a cloud-based hyper-scale data centre is much more significant. So, you know, AI servers, mainly these NVidia, you know, new servers that are out there, they're consuming, you know, three to four X more power than a traditional cloud server.

So, AI is very power-hungry. That's one theme. You got to know. So we're consuming more power. So, we're designing the data centre differently because we're putting more power density into a smaller location. So you're getting slightly smaller data centres with more power with different servers and then ultimately.

You know, they've got to be super secure. So, security is now at a premium. Backup power is at a premium. The providers of these, you know, artificial intelligence networks, they can't afford to have the power go down, right? As you can imagine, it just it just can't happen. So there's a lot of pressure to create the data centre of the future.

That is, Not going to be susceptible to losing power, which can continue to keep going. You need more connectivity if the power goes down but more energy into a denser area and more fibre. It would help if you had bigger pipes to deliver these applications. So everything's a little bigger, right?

The power requirements are more significant. The fibre connectivity requirements are more extensive. Initially, where they're built is a minor deal. You don't have to have an AI data centre in downtown New York. You can put the initial big language-based model AI data centres in places like Ohio. Atlanta, Georgia.

Dallas, Texas. But it would help if you saw these guys building these big AI data centres in Connecticut, Los Angeles, or Menlo Park. Reno, Nevada, is a hot AI data centre location. Why? Land is cheap. Power is cheap. And there's a bunch of it. And you're very close to Silicon Valley from Reno. So that's become one of the hot, you know, desirable data centre markets.

Chris Powers: So is that, when you're talking about a data centre location, obviously you need access to affordable power. It would help if you had cheaper land. Is there anything else that goes into a significant data centre location? I know Virginia is hot. Texas is starting to see a lot of them. Is there anything else that makes a great data centre location?

Marc Ganzi: It depends on what you're trying to do, right? So if you're trying to build a public cloud, private cloud, AI, or you're building on edge computing. Those are all very different requirements. One of the things that people need to appreciate more about Chris is that the data centre sector is six industries wrapped up into one industry.

So, we call it the data centre industry. But believe it or not, six different business models are inside the sector. And so you've got, you know, at the low end of the stack, you've got managed it, which is like traditional, like a rack space who will you're a small company needs to want to host your information for you.

You go to a managed it provider, and they'll outsource everything for you. And so that's one business model. You then move into what's called a hybrid cloud. We've got a bit of managed services. But that same managed services provider creates your private cloud where you can keep your data super secure.

And so we call that business model hybrid cloud. Then, as you move further up the value chain, you get an enterprise data centre. And so those are co-location facilities where different corporations put their servers into a location highly connected with fibre with interconnection that goes back out and allows an enterprise to increase a spread network.

So that's enterprise. Then you've got the fourth vertical, edge data centres. So these are data centres in tier two and tier three markets where the cloud guys build out the next phase of their network deployments. So this is staying away from Virginia, staying away from Texas.

They are staying away from Atlanta and Reno but going into smaller cities like Cleveland and Pittsburgh or Kansas City or Salt Lake City, where many of these cloud guys are starting to increase their workloads because they need to be closer to the customers, industrial users. So that's an edge.

The 5th vertical is hyper-scale, which comprises big public cloud campuses and big data centres. So think 200-megawatt campuses, 400 to 800,000 square feet of space. These are big data centres, and these are the data centres that fuel the cloud. And then the toughest vertical is the last one, which is the private cloud.

So, the level of security there is called Tier 5. There's no downtime. The data centre never goes down. These are Fortune 100 customers and government agencies that demand the highest security and reliability. So the rents are a lot higher because you're getting super value-added services.

But you're also really secure, knowing your data is safe. And so for government agencies that are seeking, you know, data sovereignty. It is a prevalent business model. We own the most significant private cloud data centre business in the United States. We bought that business a little over a year ago because we thought the private cloud would blow up, and we were right.

We've grown EBITDA by about 40 per cent in one year. It's one of the hottest Verticals in the data centre space today, which is the private cloud, and we own the largest personal cloud computing business called Switch, which is based in Las Vegas. 

Chris Powers: And all of those business models rely on the size of the business using the product.

Like, the more you move up the chain, the more significant your business gets, or is there any other reason why you would bounce between any?

Marc Ganzi: You scale into that I.T. Stack. And, eventually, when you hit the fifth leg, many big companies say, Hey, I want to go public cloud. I want to move all this stuff out to AWS, or I want to move it out to Microsoft because I don't want to be in the business of owning data centres, and I don't want to be in the business of having to manage multiple locations. Whereas if I'm in a public cloud, all I need is a connection to the Internet and access to my data. 

Chris Powers: Do Amazon and Microsoft own the cloud? Is there room for others, or do they have it wrapped down? 

Marc Ganzi: They've done a pretty good job. I think Google would also have a say about that, but as it relates to the public cloud, you've got to say that, you know, AWS and Google dominated Microsoft. It's also done an excellent job from a cloud perspective. Still, they focus on the enterprise and conquering corporate America, from Outlook to teams to relationship software.

Those are all the things that Microsoft is doing, whereas Amazon is potentially more commercially minded and focused on getting you as a customer and winning your house. 

Chris Powers: I read somewhere that my question is how far out these companies can see their demand for data storage. I've read something that said even the most prominent companies can't even predict what their data will be a year or two years out.

Is that still true today? Or are they good? Do they have a ten-year outlook on how and how much data they'll need? 

Marc Ganzi: Every time I talk about my company, we try to predict where we will be in 10 years. We look up in two years, and we've doubled the amount of data we thought we'd have in 10.

So, one could say. Eventually, Moore's Law takes effect, in effect at some point on this, but we have yet to see that. And the reason for that is less than that, you know, we're creating many documents or writing many emails. The big consumer of data right now is applications.

As an organization, if we're running Oracle, we're running Salesforce, we're running Microsoft Teams, Those applications are living and breathing, and as our employees spend more time on those applications, as you can guess, we're consuming more data, a lot more data, and it seems to have an exponential growth factor.

It's not linear; it's exponentially linear as you add more users and, more importantly, use the applications. So, the more my team uses Salesforce, the more the accounting team uses Oracle, and the more we as a firm use Teams to collaborate. All that happens is exponentially; we keep taking up more and more data.

Chris Powers: This is a dumb question, but let's say you're, you have a 20-year company and the data from 20 years ago, while it's still stored, is probably not as relevant today as the data being created today. For the sake of a company saving data, is there a value-weighted on, like, the more recent data that takes up larger capacity, and it's cheaper to store the data from 20 years ago, or is all data equal?

Marc Ganzi: Well, it depends on your document retention policy at the firm. You know, you have to decide what you will do there. We only keep things for about a year. That keeps our data storage small. And so when I say we keep it for a year, that's instant messaging emails.

Whatever the tells us to keep, we keep now on the documentation and email side on the actual, you know, due diligence and models. I see memos and all the things we use to run our business that go on to a private cloud, and we would never delete that because those are corporate records.

And so it comes down to what your head of IT will do. What is your CIO going to do? What's the policy? What's the document retention policy? And that will ultimately define how much data you will end up storing. If it's a live application, Chris, I want it as close to my office as possible.

We call that, you know, on-prem. We want it to be as close to the premises as possible. If it's like a series of documents from a loan closing or something, and I don't need it like this. I can put it in Iceland if I need to. And if it takes me five seconds or two seconds to download it, I'm okay with that latency of two seconds.

And it's an application. I need the application to decide on one 10th of a second. There's just a different level of intimacy in terms of the data. 

Chris Powers: So they say like you once, if I were to get on like Twitter or something. I tweeted 12 years ago that Twitter still holds the data.

It's my tweet or any content I put out, but are they storing it the same way? Are they holding what they would consider older stuff in cheaper locations? Or are all data points created equal, no matter when it was posted to Twitter? 

Marc Ganzi: No, definitely; the older the data is, the more they can shift it.

They can move it. They can move it to a data storage location in North Dakota. It just depends on what they want to do at the end of the day. 

Chris Powers: Okay. You all do; it looks like you do value-added digital and core plus. Is that within the data centre realm or across all verticals?

Marc Ganzi: Well, look for us, value investing is our bread and butter, so that's, you know, traditional infrastructure, high teens, low 20s returns, own a business for 5 to 10 years, buy something and then build it. So, make a great management team, finance it correctly and grow it for ultimate harvesting. And that's value-added, core stuff for us, or core plus.

If I've got 2,000 cell towers in Germany and a 30-year non-breakable lease with Deutsche Telecom, that's about as core as the core gets, right? Then, it will have a 4 to 6 per cent accretion, based on the escalator and incremental antennas. But also know that things are cash on cash in winter, and it's a dividend yield play, so, Some investors Chris wants, right?

They want safety and yield, and they want to know that they can come back. They know that in 20 years, a lease with Deutsche Telecom will be in place. Some people like value investing and say; please buy a residential fibre business in the United States with a two per cent churn. It's got 79 RPU, and the weighted average contracted cash flows are year to year.

That's a different risk profile. If we run it and exit correctly, we can get the mid- to high-twenties IRR. But it's not a core plus, right? I'm not taking the dividends and running it off over time. So, there are the same industries with digital infrastructure, just different return profiles, risk tolerance and durations of cash flows. 

Chris Powers: What does RPU mean? Mean? 

Marc Ganzi: Revenue per user. 

Chris Powers: Got it, so is it similar to cell towers? For if Amazon needs a data centre, are they usually coming to you all saying, Hey, we need a data centre, and we need your expertise to show us how big it needs to be, where it needs to go, or is a lot of it spec building and build it and they will come?

Marc Ganzi: Most of it is, 90 per cent of our development today in data centres is not only with a customer telling us where they're going to go but with a customer lease already executed. So we'll do a certain amount of work for site selection, but once the customer commits to the location, they sign a lease. Then we go, and we go with everything. 

Chris Powers: And because things are evolving rapidly, these projects take multiple years, from when you hear from them to when it's open. I'm not saying it's obsolete in some situations, but the finished product is the world's moved so far in three years that it would have been different to keep up with the current time. Does that make sense?

Marc Ganzi: Well, there's this whole notion of functional obsolescence versus whether you can continue to upgrade something and have it work, right? Every asset in our business has a useful life and an accounting life. And sometimes, the accounting and practical life need to match up.

So, for example, a cell tower has an accounting life of 30 years. Well, cell towers have been built to last longer. They're steel. You know, a data centre has an accounting life of ten years. But if a contract, every six or seven years, the technology is changing, and the leases are shorter than that, you could argue that the financial life of that asset might be faster than the physical obsolescence of that asset.

So you're matching, you know, the cash flows. With the technology curve and ultimately the physical deterioration, that asset like a data centre in 25 years might deteriorate its natural state of the physical building where a cell tower, if it's a steel pole, you know. It's sitting on I-95, somewhere outside of Georgia, where the weather's pretty relaxed. That tower should stay longer than 25 to 30 years.

Chris Powers: So we have all these data centres. How do you think about it from the standpoint that we need more data centres versus if servers get smaller and they can hold more capacity that will be able to host more capacity within existing data centres rather than having to build new ones?

Marc Ganzi: So what's unique about our pricing model now is we don't price on the physical rack.

So somebody drops a server in, we used to price that and say, okay, that's 900 a month. That's 1, 200 a month. It's not how we do it anymore. The way we do it now is based on power consumption, and the reason I did that is because, look, if a server gets more efficient, Chris, and it gets smaller, but you're still conducting the same amount of commerce from that server.

I should get paid for that. Right? So, what happens is that as more data moves through a server, something happens. It consumes more power. And so what we do is lease the demised premises for a cloud-based player, which isn't the number of servers they drop on the floor but rather the total amount of power they consume.

Chris Powers: Got it, so it's a variable, like a lease. It might look like a variable rate lease over time as they consume more power and lease payments go up. 

Marc Ganzi: Yep, exactly. 

Chris Powers: How long are those leases, like 10 or 15 years?

Marc Ganzi: Exactly. You nailed it, 10 to 15 years. 

Chris Powers: And who's responsible for the server and maybe the cyber security of it versus the physical security of it? I'm just making sure nobody breaks in and steals it.

Marc Ganzi: The server is the property of the customer. They own it, they install it. We put a little cage around their equipment and locked the door. They have a retinal scanner. They have a key code or a thumb code. And so everything outside that server belongs to us: the security, the cooling, and the power.

And we must provide Amazon with the best experience they can get, right? And if we don't do a good job, they don't come back to us. So we're laser-focused on ensuring they have an outstanding customer experience, but ultimately, they own the server. If the server breaks, that's on them.

If the data centre breaks, that's on me: the cooling, the connectivity, the security. So I provide all that. I provide everything, you know, but the server. And there may come a time when we decide we want to own the server for the customer and go further up the value chain, right? When we go, ask for more rent. 

Chris Powers: When you think of fibre, let's say it's laid in the ground, and the fibre technology continues improving. Do you have to replace existing fibre every time there's a technology enhancement, or are there ways to enhance the current fibre in the ground?

Marc Ganzi: Remember, fibre optic cabling is just; it's a glass. And so all we're doing is we provide the conduit. We provide the number of pairs; if it's a dark fibre route, the customer lights it. We don't burn it. We're just the train tracks. And so, on that basis, it helps the business become more future-proof.

And the upgrade cycle on those particular routes. They have a useful life, just like a cell tower. It's exciting. It has about a 25-year helpful life for fibre optic cabling. 

Chris Powers: Knowing that a lot of it comes down to power. We will have all these; I'm sure you get this question.

We're going to have a lot of empty office buildings sitting around the country. Are a lot of them going to be candidates for data centres or not really? Cause the power's not there. 

Marc Ganzi: That's interesting. I hadn't thought about that, but Chris's power is the biggest issue in data centre land today.

And as you know, our transmission infrastructure grids in the US are ageing. Most of the transmission was built in the Eisenhower era. So then, here we are asking about these transmission grids to deliver hundreds of megawatts of power to a 10-acre parcel. That's different than what was envisioned two years ago. I don't think the United States has a generation issue.

We're generating plenty of power, whether it's hydro, solar, wind, or traditional, you know, coal-fired plants or nuclear plants. We know how to generate power. We must learn how to move, distribute, and transmit it correctly. So that's why you see what ERCOTs are doing in Texas.

Redoing the entire, you know, transmission infrastructure grid. It's what you see in Dominion Resources. What's happening in Virginia? What's happening in California was with PG & E and so-called Edison having to dig their lines and build new lines, so we're going through this renaissance per se of our transmission infrastructure, and it will take time. It will take time, so in the meantime, as an industry that, as someone who's the most significant donor data centres and, I think, the world now, we're very focused on energy independence. So everything we're doing right now is towards that, whether we're building solar power, whether we're building wind, whether we're buying hydro, whatever we're doing, we're trying to figure out how to get behind the meter and not have to be subject to the existing infrastructure of the US or Europe or Latin America. 

Chris Powers: Do you even think about the rise of electric vehicles and the energy allocation they will take up regarding what might be available for data centres? Or is that a blip on the radar?

Marc Ganzi: If I'm thinking about the future and how much power data centres and electric cars are taking off the grid, our transmission infrastructure will run out of power in three to four years. So the same problems that you're seeing in Northern California, the same issues you saw in the Northeast with Mohawk power and Pacific gas and electric, where you'll start having brownouts, you'll have to turn off the grid. That's a real problem. That's a short time away.

You talk to some people, and that's three to four years away, just given the demands of the grid. The electrification of the automobile industry, the movement towards AI and how much power is being spent on data centres. And maybe I'm being a bit dramatic about that, but even if it's four or five years away, we still have a problem. Right? 

Chris Powers: Well, could you see politicians or government being like, Hey, tech companies, sorry, you can't store this data anymore? It's more important that power goes to people's houses, and they can, you know, use basic power needs instead of saving the following YouTube video on YouTube. 

Marc Ganzi: Well, politicians are not thinking that way right now; I would tell you that Glenn Yunkin and Mark Warner are thinking about it. You know, after Virginia ran out of power, the data centre industry became one of the great producers of jobs and tax income for Virginia. Young can prioritize that. So did Mark Warner, and they got in and, you know, instead of waiting five to six years to upgrade their transmission infrastructure, they're doing it in two years.

So they've prioritized that because they know if they don't fix the grid, Virginia stops hiring jobs, and tax rolls go down and, you know. The government ends up having a real problem on its hands. Virginia is one of the great states that's benefited from cloud computing because of Amazon's location there and the number of, you know, data centres that exist in Data Centre Alley, but this is going to be a continuing dialogue. You and I will discuss this for the next three to four years.

Chris Powers: And you've mentioned AI, but is Crypto putting any dent in the space? Is it, or is it not near the impact of AI?

Marc Ganzi: No, it's nothing near AI. We own one of the largest crypto mines in Europe, and we sold it, you know, right at the beginning of COVID.

Crypto could have been more sustainable. The data centres were consuming a lot of power. Crypto consumes a lot of energy. And ultimately, these were data centres located in Iceland. So, we didn't have a huge desire to keep them, and we've grown EBITDA by about 200 per cent. So somebody came to us and said, you know, we want to own crypto infrastructure.

We said, great here, take it. And you know, we sold the business for like 28 times even, and we made five times our money. What we do is a trade. It was a simple two-year trade for us, And it was all based on crypto hype and momentum. And I don't believe in investing in the hype, and by the business that's being built off of future expectations, and it's trading in the twenties, and we don't have long-term contracts with customers, then I'm an easy seller every day of the week.

Chris Powers: Getting back quickly to cyber security is what we are told year after. It's a huge threat, and it's more of a threat again; you all secure the physical aspects of the location. It's on the company to manage its cybersecurity, correct? Or is there anything that falls in your?

And since you're just in this world, do you have any comments on how we're doing as a country around cybersecurity? It's a huge threat, but I assume we have talented people working on it.

Marc Ganzi: You have to assume we have highly talented people working on it.

And the cyber-attacks you do here are basically, you hear about 1 per cent of the cyber attacks daily. You see, we're getting targeted every day just in our network and our own small digital rich network. We have dozens of cyber attacks on our network every day.

Welcome to the world of being a CIO today or CTO; it's protecting the Ford at all costs. The last thing you want to do is go to your CEO's office and say, we've got a ransom letter for 2 million, and we got to, we got to pay tomorrow or else we lose all our data. 

Chris Powers: Is there any? There's no magic bullet like Crypto's; they say it's impenetrable because of the algorithm.

Marc Ganzi: There's Blockchain; blockchain keeps Crypto safe. No, there's no magic bullet to save people from hacking. 

Chris Powers: Does everything move on to the blockchain, or is that a sound good line, but it won't happen?

Marc Ganzi: That's the feel-good version of Sports Centre. I don't think that happens. 

Chris Powers: When you think about the Middle East and Asia, which you all own a lot, is the business run the same way over there in America, or is it a lot different, more of the Wild West?

Marc Ganzi: No, Asia's been a great surprise for us in the last three years. We see things happening in Asia that are, candidly, more progressive from a digital perspective than the US. So, we see a regulatory regime open in many of these countries.

We see a lot of under-investment. You know, the U. S. has been investing in digital infrastructure for decades, and in Asia, you know, they're probably one cycle behind. Which is, we see parts of Asia about three to five years behind in terms of infrastructure building. Europe has always been a laggard economy compared to the US, except in the Nordics, where infrastructure is progressive and modern. One of the more intriguing regions in the world today is the G. C. C. When you look across the Gulf, you see economies levelling up based on digital, the fastest growing social media and YouTube markets in the world or Saudi Arabia, in the UAE.

These are companies that have embraced Digital transformation. They're moving fast, and the younger demographics in those countries are using digital technology to start pulling the economy and culture into more of a Western, what I would call a democratic economy or capitalist economy.

Those countries get it; they're progressive and investing heavily in digital. So when we show up in the GCC or Asia, those markets are working well because they're embracing digital transformation, and to get there, you need the infrastructure.

And so it makes it an excellent investable area for us. And also from a fundraising perspective, you know, the Asian pension funds and sovereign wealth funds, along with the Gulf, you know, sovereign wealth funds, they have a lot of money. They're under-allocated, you know, from an asset allocation decision, looking at their pie, they've traditionally been in fixed income, public equities, and a little bit of private equity.

And now, these sovereign wealth funds are getting sophisticated. And now they're doing much bigger digital allocations into infrastructure and private markets. And so I'm finding that when I travel, and I'm fundraising for what we're doing, our most significant opportunity to form new capital today is in Asia and the Middle East.

Chris Powers: And some people lump it in with real estate. Real estate had a challenging fundraising environment. The last call was 18 months ago. Does digital real estate follow that trend line, or is it different? People think about it differently. 

Marc Ganzi: Differently, why? Because the returns are there, right?

And the growth is there, so, unlike traditional commercial real estate, which has been declining, you know, rent rolls are coming down, rents are coming down, and occupancy is coming down. Let's go to digital and flip the script. Occupancy is up; growth is up. Rental rates are up, so just in one year alone, data centre rents have moved up 21%.

Our occupancy is up 80% year over year, 80%. Chris, I've been doing this for 30 years. That's never happened in my life where any of our businesses have been up from an occupancy perspective by 80%. So we deployed seven and a half billion of CapEx last year into our data centre businesses. We own six different companies around the world that do this.

We finished our budget planning season two weeks ago and will deploy 12. 9 billion of CapEx next year. 7. 5 to 12. 9 billion of new CapEx. It's stunning, it's incredible. 

Chris Powers: And is the answer to that increase in demand, obviously AI is coming, and just people are using digital products more and more and the globe, America, we've been on it for a while, but the rest of the world is just continuing to get more and more online.

Marc Ganzi: The rest of the globe is catching up. That's, that's the key. I mean, that's the critical headline. As we said, they're just now building cloud infrastructure in Asia and the Middle East. We were doing that ten years ago. So, cloud adaptation is moving fast in those regions. And AI will move just as fast. So we're swamped and extremely busy.

Chris Powers: Did you see anything coming out of COVID? Was there a permanent shift also out of COVID that drove this or accelerated it?

Marc Ganzi: COVID taught us that the digital infrastructure in the suburbs and the secondary and tertiary markets needed to be more robust.

So, we spent a lot of time in COVID building out edge computing, and we spent more time fortifying, you know, broadband networks to the home because there wasn't enough fibre sitting out in the suburbs. And so we had to increase connectivity issues. We had to grow the cloud. We had to increase the cloud.

And so, COVID was all about spreading networks further and broader and ultimately performing at a high level, which networks in rural areas and secondary and tertiary markets still need to achieve well. 

Chris Powers: It all sounds like really bullish. My last question is that some people, some of the prolific short sellers, have been bearish. Like what is their case against digital that could be rationally taken seriously?

Marc Ganzi: When I merged my company with another public company, we reverse-merged into an old real estate investment trust called Colony Capital.

We had an activist shareholder who was trying to short us. And I said, look, you don't get it. Digital real estate is what is making real estate obsolete. You know, the big disinter mediator to malls and office buildings is what we do. And so, anyone going to short digital infrastructure or fast digital rich must know what you're doing.

We ran into this guy that was shorting the data centre space. And I won't name his name specifically, except I never named his name specifically. I just ran my race and said, look, if that guy wants to spend 3rd party investor capital shorting me, it's a bad run. He will fail because we've been able to grow our assets under management from 14 billion to 75 billion in 3.5 years. We sold 50 billion of commercial real estate. We rotated into 75 billion digital. At the same time, we do leverage the company from, you know, 14 billion to 330 million in debt. So, the short thesis on us was that we would choke on our capital stack. We didn't plan to de-lever and would get stuck with all this old real estate while pivoting to digital.

Well, we just went fast. We went faster than everyone expected. Then, the other short thesis on us in the sector was just that. We're a competitor to Amazon and Microsoft. And again, it's just naive thinking, right? I'm the landlord to them. I'm their partner. I have no interest in being in their business. Still, I am interested in facilitating their business's growth, ensuring their networks stay up and reliable, and providing them with fibre and new ideas.

And we've had customer relationships, Chris, for 30 years. I've had the privilege of maintaining networks for, you know, Verizon and AT&T for three decades now, T Mobile for three decades. These are long-lived customer relationships built on trust and the fact that we run their networks for them.

We help build their networks for them. And that is something you can only accomplish after a while. It takes decades of experience to get there. You can ruin it one night by having the network go down, but we don't do that. And ultimately, the short thesis on data centres failed. It was the perfect storm because no one had done the homework around cloud adaptation, the growth of public cloud, private cloud and AI simultaneously.

It was a perfect tsunami for growth, and at the same time, everything you and I just talked about, the ageing transmission grid; there's not enough power. So, the existing data centres that are built are in great locations. They have power. They can serve the customer, so rents increased 21 per cent in one year.

So, you have this fantastic demand and supply trade imbalance. The supply is the data centres, and the market is the cloud players and the AI providers. And these guys have an enormous appetite, more than the industry can deliver. So that's why rents have moved, you know, moved up consecutively for seven straight quarters.

A lot is going on here. Exciting. A lot of fun. Someone may think of a way to short us again, but it's a bad bet. It's a terrible bet. 

Chris Powers: Mark, this was awesome. I appreciate your time today. It was a treat.

Marc Ganzi: Yeah. For me as well, Chris. Thank you.