DigitalBridge Group, Inc. (DBRG) Earnings Call Transcript & Summary

Well, thanks, and good morning, and welcome to the second day of Citi's Global TMT West Conference. For those of you I haven't met yet, I'm Mike Rollins, and I cover the communications services and infrastructure categories for Citi Research. Just for reference, we do have disclosures available on the conference registration site. I'd like to welcome back, Marc Ganzi, President and CEO of Colony Capital. Marc, thank you for joining us today.

Thanks, Michael. Good morning. It's great to be here.

Great to see you.

So we've started in some of our past conversations with just a high level understanding of what your operating and strategic priorities are. So if you can run us through how you're looking at the year to come, that would be great.

Yes. Thanks, Michael. Look, last year, it was really, from my perspective, about putting our house in order and giving us the opportunity to make the pivot to digital. And we achieved all of our goals that we set out last year. And so as we thought about last year as being more about how to fix and stabilize Colony and bring investor sentiment rotating back into our new story. This year, for us, it's about growth. And we see a tremendous amount of opportunity ahead of us. And so I think from our perspective, we want to continue to keep doing what worked last year. And so the key thematics for us will be to continue to form capital. We had a tremendous year last year. We almost raised about $10 billion in third-party capital, which is, from our perspective, a great outcome. We acquired over $22 billion of digital assets last year. So we want to continue to invest in high-quality assets. So as we think about that and our ability to raise capital, inform new ideas, the fourth quarter was fantastic. We made 3 new investments in the fourth quarter. And we'll make significant new investments in the first quarter. So to do that, obviously, we have to continue to rotate. And our rotation at Colony has been about selling legacy real estate assets and rotating into high-quality digital assets. And so we got to finish what we started, which is we've got to end the journey in selling legacy assets. So in my hierarchy of objectives, one will be in the first half of the year to find a good home for our health care assets. We have our 400 health care assets that we're looking to find a home for. Ultimately, continue to find a good place for CLNC. A lot of great progress there with Marc Ganzi and the team, and we want to make sure we find the right home for that. And then we want to continue to methodically sell our OE&D portfolio. We had almost $300 million of monetizations in the fourth quarter. It was a great outcome for us. And we believe there's about a little less than $1 billion left in our OE&D portfolio that we can take that cash and rotate it back onto the balance sheet. And that will give us the firepower to add great digital assets and continue to invest in great digital investment management products. So I'm really pretty optimistic about 2021. A lot of the heavy lift behind us around fixing the capital structure, fixing our cost basis and our G&A load. And proving that we could rotate those 3 key themes we proved out in 2020, and now we're moving forward.

Great. So we're going to bring our audience into the conversation early with our live surveys. And I'll throw out the questions and then we'll talk about something else while people get a chance to poll in and tell us how they're feeling about things. So the first question that we'd like to throw out, and we've done this before, is which infrastructure assets represent the best business model to own? And the choices that we're giving today are domestic tower assets, international tower assets, fiber infrastructure, including small cells, retail data centers, hyperscale data centers and multiproduct data center campuses.

Yes.

So we'll see what our audience comes back. But before we get there, Marc, you started on the strategy to pivot to digital before anyone would have imagined there would be a pandemic in 2020. So can you talk about the impacts that Colony has seen from the pandemic? And has the second wave that's been hitting some of the cities, is that affecting the assets differently than you would have expected?

Well, look, I think it's been kind of the tale of 2 cities, right? On one side of the coin, in our digital business, it's been fantastic. We've had incredible organic growth. All of our '19 portfolio companies that we own globally posted positive organic growth in 2020. So we continue to see incredible demand from our customers. And that's irrespective of what asset class that we've invested in. And of course, we've invested across the entire ecosystem. I would say during the pandemic, our relationships with customers deepened, because I think our customers really made a conscious decision to prune their vendor list on who they work with in terms of new builds. And so from a greenfield perspective, we've never had a better year in greenfield than we did in 2020 in Europe, in North America, in Latin America, now, our operations in Asia. We continue to win our outsized share of jump balls at our various '19 portfolio companies. And so the reason I have a lot of optimism around 2021 is because we're doing good things for customers. We've never been closer to our customers. And most importantly, we've delivered in terms of new construction. So this is a business hasn't changed. Since I got in the business in 1994, it's the same. You got to wake up every day, and you got to execute. And we had a great execution year. We always can do better. And we definitely are pretty self-aware about what we didn't do right in 2020. But I do think the momentum of all of our leasing pipelines across the globe have continued to accelerate and grow. And that -- and you and I have talked about this before. Your pipeline really is an indicator of what's coming. And so we saw our pipelines sort of come down a little bit in Q2 and Q3 as that was really the impact from the first wave of the pandemic and then we saw in the fall of this year, our pipeline is accelerated because, as you know, CapEx is always most prevalent in fourth quarter. The question is what happens in Q1 as we proliferate the vaccine and as we start pivoting through the C-band auctions and what happens next. Right now as I look at our pipelines, they all feel pretty good. And there's pockets and there's opportunities, certainly, that are differentiated in each of the verticals we're in and each of the geographies we're in. And I can go as deep as you wish in any geography or any vertical, but I think the punchline is, if you took a snapshot of our pipeline a year ago when you looked at our pipelines today across all of our businesses, they're generally higher than they were 1 year ago. So that gives me some optimism for '21.

And just following on that. So what we've been hearing around the industry is that as we've exited 2020, the pipeline is bigger. But some of the decision-making has been slower. And so from a smaller percentage against a larger potential pie, it's kind of been a similar type of bookings environment. I'm generalizing. Of course, you have, as you said, all these holding companies and different products in different regions. But are there certain places where you would say, no, there's been no change in decision-making. It's been very prompt and that there have been places, whether it's product or by region, where you'd say, yes, we are seeing some extension of that decision-making cycle and how you might expect that to change in 2021?

Well, look, let's just quickly go geographically, and I'll tell you my thoughts. Let's start in Europe. I would say, in Europe, data center leasing massively outperformed. We performed 200% of our budget in 2020. So in hyperscale leasing, we won a lot of jump balls against some of our public brethren. We do compete head-to-head with CyrusOne and Digital Realty. Those are really the 2 prominent logos we face in Europe, and we won our fair share and then some. Big victories in Berlin and Zurich and Cardiff and Offenbach. I couldn't be happier with what's happening in hyperscale leasing in Europe, and I think we've done a good job there because we -- once again, we've delivered. We were shovel-ready. We were prepared. And that's what the customers care about. Can you deliver? Can you be on time? And so I'm really happy with what we've done in Europe. I would say tower leasing there was okay. It wasn't a great year for tower leasing. The carriers were pretty muted in terms of CapEx spending as I think they're all looking at this dinner bill for 5G. And candidly, there's a lot of captive tower cos that are being developed by the carriers. And we all know how the captive tower cos don't perform the same as an independent tower co. And the way they get growth is by building their own towers, and there's no real independent lease-up. So Europe is a troubled tower market because we see all of these MNOs wanting to do their own tower cos. If everybody does their own tower cos, there's going to be no colo. So as we look at the projections for some of these soon to be announced public vehicles from the various carriers, we have a high bit of skepticism. And we look, of course, at what recently what Cellnex did and what they paid for some of the assets they acquired, and we just can't make the math work. So we're pretty muted on European towers right now. I would say in European fiber, we have 3 fiber assets over there: Digital, Freshwave group and Zayo. All of those companies performed exceptionally well in the fourth quarter, particularly Freshwave, which does fiber and small cells. We're finally beginning to see adaptation of small cells in the U.K. And they had their best quarter in company history. So that was a bright spot for us. Lastly, I think that's really a small cells, towers, fiber, data centers, that's kind of my view on Europe. So in Europe, we're kind of excited about what's happening in data center land. We're pretty muted around towers. And I think fibers performed well and small cells are beginning to come as we start pointing at 5G. In North America, I would say, on the edge compute side at DataBank, another really solid year there, positive organic growth, a lot of new conversations and new logos going into our 60-plus data centers in the U.S., a lot like probably what you heard from Equinix. We're pretty excited about the ecosystem that we've developed at Databank. And having that highly interconnected environment where we can provide interconnection, we can provide space and power. But most importantly, we can get that connected out to our towers and small cells. That's a real differentiator. So as we look at a world where there's decentralized RAN and we see more edge computing, data bank is really well positioned to take advantage of that. And I think you spent some time yesterday talking about the edge and what's the importance of open RAN and what's the importance of having the base stations close to the compute and the applications. I watched my good friend, Tom Bartlett's, interview with you yesterday, and that's where the future is going. And if you don't understand C-RAN or open RAN architecture and why having edge compute facilities is important to towers, you're going to be vacant to a critical piece of the narrative, I think, going forward. And I watch some of your polls that you asked some of the investors yesterday about edge computing and whether it's important to American tower or not. The question isn't whether it's important to American tower not. The question is, is it important as a digital infrastructure owner to have an edge solution. And I think the answer is unequivocally yes. That's the bet that we've made. I think in hyperscale leasing, we had a very good year. We actually were ahead of budget. We had a strong fourth quarter finish with good victories in a couple of our key markets. We really are sort of confidential about where those victories are. But I would say exceeding budget and leasing was important because a lot of the narrative around hyperscale leasing in the U.S. was that it's slow. It's competitive. People didn't hit their bookings. Some did. Some didn't. We did. So we're very happy with hyperscale in the U.S. We're very happy with edge compute. I would say U.S. towers. We had a good year. We didn't have a great year. We were very ambitious about what we wanted to do. We were hoping to get double-digit organic growth this year. We're going to end up in the sort of high single digits between 8% to 9% organic growth, which will obviously be, on a pure basis, will be probably industry-leading, but there was some disappointment. I think some of that deferral you talked about pushing out into 2021. When do we really accelerate with DISH? When does T-Mobile turn back on? And what will the impact be of a very expensive C-band auction on AT&T and Verizon. So we've sort of taken our forecast for '21. We brought it down a little bit vis-à-vis 2020 because we had a slight miss on leasing. But still, I'll take 8% to 9% organic growth in the U.S. all day long. And we had very little churn. I think that was important. We weren't exposed to Sprint. We have the youngest U.S. cell tower REIT in the U.S. Vertical Bridge is the youngest. Our average tower is about 7 years in age. So we didn't have a lot of exposure to Sprint. So we perhaps didn't feel the same level of churn that American Crown and SBA felt. So that was good. I would say in fiber, Zayo continues to perform. We posted positive organic growth in the fourth quarter. First quarter was spectacular. It was incredible. We were sort of 8.9% organic growth, net of churn. And then Q2 was a slower growth rate. And then Q3 was a little bit of a disappoint. We had a little more churn than bookings. But I would say net-net on the year, it was a great year. Backlog continues to build. We've never had a bigger backlog at Zayo as we head into Q1. So we're pretty optimistic about that. Our Canadian Fiberco greenfield field posted double-digit organic growth up in Canada where we serve Montreal and Toronto. It's a metro fiber business. It's done incredibly well. So we're happy with where we are in terms of our fiber assets. Small cells was pretty -- got off to a good start in Q1. It was very decimated Q2 and Q3. As you know, the small cell sector, Michael, relies on permits. We had a lot of difficulties with municipalities, but we had a tremendous fourth quarter. We've never built more nodes in a quarter than we built in fourth quarter. We had our best construction quarter in company history. And we did over $6 million in new bookings in the quarter. So we had our best quarterly bookings. So a lot of our key customers turned back on. We're busy working for Verizon. We're busy working for AT&T. We're optimistic about T-Mobile in the second half of next year. We're optimistic about DISH and small cells in '21 and '22, late '21, probably more of a 2022 event, and we continue to build a lot of C-RAN hubs. And so ExteNet's never been busier, and I'm really excited. Our pipeline is over close to $63 million heading into next year in terms of new bookings. So we've never had a bigger pipeline. We never had a big fourth quarter. So we're really excited about what we're doing at ExteNet. And maybe that experience is a little different at Crown. But once again, it's about winning the jump balls. It's about getting out there and competing. And I love the way ExteNet's competing right now. Latin America was quite good in hyperscale leasing. We had a strong fourth quarter at Scala, big, huge booking with a webscaler in São Paulo. So very excited about that. We exceeded our leasing forecast there. I would say, in small cells, which touches ATP and highline and MTP, we own 3 tower companies in the region. Really, the small cell space is starting to heat up in Brazil. We had a very strong year in Colombia and Chile in small cells and really Peru kind of shut down a little bit on us, and Mexico is pretty muted in small cells. But that being said, we've continued to deploy open RAN architecture in Mexico, which is pretty exciting. We've built a series of C-RAN hubs for Alton, where we've got the likes of Telefonica, Total Play and AT&T co-locating in those open RAN hubs. So we're very excited about that business. The C-RAN business in Mexico has never been stronger. We'll probably finish this year somewhere between 60 to 80 C-RAN hubs in Mexico. It's interesting we got to go south of the border to learn about open RAN and sharing of networks. But we've been on that journey with Alton for about 3 years, and it's great that Total Play and AT&T and Telefónica have co-located inside that open RAN architecture with us. I'd say the tower business was good in Latin America this year. We saw a real acceleration in Brazil in the second half. Peru has been a tough political environment. So we didn't hit our expectations in Peru. We massively over-exceeded in Colombia. We hit our budgets in Santiago, Chile. Chile was a good year for us. And then I would say in Mexico, a slight miss on leasing, but we had pretty muted expectations in Mexico. But I do like, once again, what we've been doing in the open RAN and C-RAN space there. That's worked out exceedingly well. We've exceeded our plan there. And then in fiber, we own fiber assets in Chile, in Colombia and in Mexico. The best year for us was really Colombia. We saw a lot of dark fiber leasing. We own a significant amount of dark capacity there, and we're starting to bring dark services to some of our mobile towers. And we're also own a couple of metro rings that we built for some customers, and we're also doing fiber to the home on a wholesale basis where we'll build out the architecture and the network for some of the major carriers in Colombia and Chile. So net-net, we're very positive on Latin America. And Asia is a new market for us. We just acquired -- we have 2 businesses in Asia now, Agile and Edgepoint. Really no operating metrics because we closed on both those investments in the fourth quarter, but we continue to be pretty excited about some of the things we're doing in Asia, particularly in towers and in hyperscale. Those are 2 areas where we're focused on. Sorry, I took about 7 minutes there, but that's a quick trip around the world in digital infrastructure.

This give us a lot to talk about. So thank you. So let's also just see -- before we get there, let's just see the answer to the survey question from our audience.

Sure.

And we'll talk about that for a moment, and then we'll drill down into some of what you're seeing in your operations. And this is super helpful.

This is interesting.

So for those on the audio, I'm just going to read out the responses. So what's the best business model to own? Domestic tower assets, 60%; hyperscale data centers, 13%; international towers, 7%; fiber, 7%; retail data centers, 7%; multiproduct data center, 7%. And just for our audience's sake, all your responses are confidential. We're just aggregating the percentages. So Marc, we talked about this concept of what's the best business model to own for some time. And over the years, and the businesses have expanded in infrastructure, your views have evolved. What's your current view on what the best business model is to own?

Well, I'll start out the same place we've always started the dialogue, which is at what price do you own it at. The entry price is probably the most important thing. And I think look, towers are really expensive today, Michael. They're -- as I like to say, they're priced to perfection. And obviously, we'll see what happens here in the first quarter with these results coming out of Georgia and what the market does. But some would offer that perhaps maybe tower stocks are perhaps priced a little bit too much to perfection. So we'll find out what happens. I think, look, at the end of the day, our bet at Colony is you don't have to choose, right? You can own it all. And so I think the digital REITs of the future aren't making investors choose which vertical to own. I think having exposure to all of the verticals and being able to deliver integrated solutions is the future. And even -- Tom Bartling even admitted that yesterday, that you have to have multiple solutions for customers to stay relevant. And that's what we're doing. That's what we've been doing for the last 3 years at Colony, specifically since I got here 18 months ago, is we are building the most diversified and fully integrated portfolio of digital infrastructure globally, and we think that's the right way to play it. And in doing so, Michael, it gives us a chance to be responsive as we just -- I was able to just move around the globe in 7 minutes with you and talk about each vertical, talk about our customers. And look, by the way, just European towers, that could all change in 18 months, right? We could end up having a really unique solution around C-RAN or open RAN. And we do a ton of organic leasing with all the carriers in Europe because we have a unique product but, most importantly, we have a unique customer-focused solution. So it really comes back to how do you respond for customers. And so I look at what we're doing in Japan for some of the web scalers in the U.S., and I get excited about that because that's a customer-driven opportunity. And everything we're doing today at Colony is customer-driven. We're listening to customers carefully what they want. Nobody would have said, if we -- if you and I would have sat down 2 years ago and said, Marc, why are you investing in C-RAN in Mexico? That seems totally crazy. You should be investing in C-RAN in the U.S. And my answer is, well, we are. We've built close to 800 C-RAN hubs in the last 5 years at ExteNet, but we don't run around bragging about it or talk about it. We just go out and we execute, because that's what customers want us to do. And so I think I find that survey to be interesting. If they were all priced the same, I would agree. I would want to own domestic U.S. tower assets. If I could own domestic U.S. tower assets at 16, 18x, I'd be really happy. But I don't have that luxury today. Most of what we're seeing today is priced north of 20x. We saw American Tower's most recent domestic acquisition at 30x. And so these things are priced on the screws. And so our pivot to that is we're a builder. Certainly, we're a very selective buyer. We've been very active in towers, as you know, extremely active in Brazil and extremely active in Southeast Asia. Those were 2 theaters where we felt like we could find value in the last 6 to 9 months. And when we found that value, we executed flawlessly. And so we went from 300 towers to 4,000 towers in Brazil. We went from 0 towers to over 4,000 towers in Southeast Asia. So we're finding opportunities, pockets where we see value and where we have good customers that want to grow with us. Both Southeast Asia and Brazil are 2 real hot markets for us because, once again, we're winning our outside share of new builds.

We're going to bring up our next survey question, and we'll take a couple of minutes to get to it. I think it'll be a little preview of where we go with the conversation. So we're on survey number 2. So there's a question about demand for data center assets and how that looks in 2021 over 2020. So we're going to give a few choices and just say, yes, bookings will broadly exceed last year's level for data centers; yes, but for hyperscale deployments, principally; yes, for retail-centric deployments; or no, bookings will just be flat or down in '21 over the 2020 performance. So we'll see what our audience thinks about that. Before we get there, though, you were talking about the strength of hyperscale bookings. And one of the questions has been the development yields that are coming out of hyperscale. Are you seeing those stabilize? Or are you still seeing competitive pressures tightening those initial yields that you're getting from the bookings and in the marketplace?

Well, look, I would say it's once again very geographically-driven. And in instances where we're shovel-ready and we have the power and we have the we'll serve letter and nobody else does, we're getting great returns. And so in Europe, by example, we made a conscious decision to stay out of the flat markets. Other than our presence in Offenbach, which has worked out really well, we have a great joint venture with the power company there. We've got a fairly unlimited supply power because they're our partner in that location. But I look at the other places where we have a presence, where we have either data centers active or where we're going shovel ready. We like sort of staying out of the fight in places like Amsterdam and London. Frankfurt is kind of a unique market. We're an Offenbach, which is a really good place for us to be. But we've really had our success in places like Cardiff, Dublin, Warsaw, Berlin, Zurich, Milan. And so looking at those, what I would call those Tier 2 markets has really worked out well for us because we've had the land. We've had the capacity. We've been able to deliver for customers. And so I think it's just sort of being intelligent about where you go. Same thing in the U.S. I mean we are in Ashburn, but Ashburn is not our biggest market. We're very long on Santa Clara, where we have a unique footprint. We have a unique set of will serve letters. We've got a lot of power density in Santa Clara, and that's why we're the market leader in Santa Clara. We're in places like Quince, Québec City, Montreal. We're shovel-ready in Goodyear. And so it's about site selection. It's about understanding where your customers are going. I don't like the returns in Ashburn. Never have. But it's a market we had to be in. And we've done okay. We actually secured 2 very competitive wins in Ashburn in Q4. But it's a fight to the death in Ashburn. There's a lot of private players there that have come into that market with private capital, a little new to the space, a little green to the space, and they've really underpriced the megawatts there, and it's caused some market friction. And you'd hear the same from QTS. You'd hear the same from DLR, and you'd hear the same from Cyrus. I think the good news is each of us have pet markets we really like. You've heard people talk about Atlanta being a good market. You've heard about Chicago. You've heard about Dallas. You've heard about Austin. Some of these perhaps Tier 2 hyperscale markets have worked out really well for our peers. And like for us, I mean, Quincy, Washington has worked out great. Québec City has worked out great for us. Sometimes you've got to go find those, what I call those seams in the marketplace where you can get the right returns. So I would say São Paulo, Brazil has been great. We've got a very rational competitor in Digital Realty. In Osaka in Tokyo, we've got a very rational set of competitors. So we continue to find the right returns in hyperscale. And it's an asset class we've been able to finance. I think we've raised -- someone told me the other day, we've raised close to $6 billion of capital in the last 2.5 years to support our hyperscale efforts. We'll raise another $6 billion in the next 18 months to continue to grow globally in hyperscale. Our ambition is to be the largest hyperscale provider in the world with Vantage. And we're on our way to doing that.

So let's see the survey results from the data center side just to see how people are feeling about the bookings environment. So 36% yes for hyperscale; 29% broadly better; 21% more retail-centric bookings; and 14%, no, flat or down. And you talked a little bit about this earlier, Marc, but just to put a finer point on it, is there a certain category for data centers that you're the most bullish on for 2021?

Well, look, it's just looking at the pipeline, right? So I think our hyperscale pipeline is just a lot bigger and a lot deeper than what we're seeing at DataBank on the edge side. Now look, DataBank operates in the U.S., 60-plus data centers in the U.S. So we're a pretty domestically-focused edge compute business in the U.S. Our pipeline is building in the fourth quarter. We're hoping to put up double-digit organic growth at DataBank next year. Vantage had double-digit organic growth this year at Vantage Europe and Vantage U.S.A. and Scala had double-digit organic growth as well in Latin America. So all of our hyperscale businesses are growing at double-digit organic growth. A lot of that is greenfield construction. Some of it is leasing, but all of it is new leasing. Everything we're doing in our hyperscale business today is around new builds and lease-up in existing parks. So I think that's what you're seeing in that survey. That survey is actually a pretty accurate indication of where the market is. Remember, there's 5 ways to make revenue in data centers: hyperscale, edge, retail, managed services and interconnection. Those are kind of the different business models. I would say, as I look across those 5 business models, hyperscale had the highest growth rate of the 5 products we offer. The lowest growth rate this year was managed services. We posted positive organic growth, but it was tough. Our business app up in Canada, which is our managed services, our hybrid cloud business, had a good year. It could have had a better year. There was churn. There was new bookings. And that's probably a sector, as you look at some of the public peers and you look at people that have gone away from managed services, I think Rackspace is probably the best proxy for what's happening in the hybrid cloud world. And Rackspace had a pretty good year. And they've got a great management team and a great CEO. I would say retail was probably, net-net, a little disappointing. Started off the year great. Midst of the pandemic, probably didn't perform as well as we would have liked in Q2 and Q3, began to see a rebound in bookings in Q4. But I continue to think that IT managers will sit on their hands in the first half of 2021. And then the edge business has been great. I mean DataBank's organic bookings on edge have been good. We have an investment in edge presence, which has grown significantly this year. And I think for me -- maybe -- I'll just put out there a hypothetical, but I think hyperscale on our edge data center businesses will have similar organic growth rates in '21. I think the edge is very interesting. And we define edge as what DataBank is doing but also what -- certainly what Freshwave and MTP and ExteNet are doing around Open RAN hubs, C-RAN hubs and building those smaller little edge facilities, where we're bringing together the mobile base station, decentralized rand side-by-side with compute and applications.

I want to come back to that in a few moments. Let's go with the third survey question, just to get that out there, and we'll see what the responses are. And this is sort of framing an earlier question differently, what's the most exciting product category for growth on a 5-year view? So if you're just looking at pure growth, what's going to have the greatest opportunity. And we broke this down fiber in small cells, the hyperscale data centers, enterprise data center deployments, edge and far edge concepts for data centers, and then new business models such as C-RAN, hosting and other things that people might imagine that just aren't articulated in the first 4 choices. So we'll see what people come back with on that. But one thing you said earlier I want to come back to is you talked about how you're very value-focused when you're thinking about assets and your views of assets. And over the last few years, we've seen broader participation from infrastructure funds, sovereign wealth funds, joint ventures in the category. And as you look at whether it's buying assets and the potential competition, or having more possible parties to sell into, how has it affected the buy-and-sell dynamics that you look at for the portfolio of assets that you manage?

Yes. Look, I would say, broadly speaking, it's kind of an interesting conundrum, right? I mean, in one hand, sometimes we're a seller and other times, we're buyers. And that's part of our investment management platform. We own 19 companies globally, and sometimes we are a seller and sometimes we are a buyer. I would say, if there was one impact -- we talked about it in hyperscales, by example. The folks that have entered the space on the private side are mostly backed by infrastructure investors. And there's a flood of liquidity with infrastructure investors today. And so they're coming into the space. And they're resetting what's an acceptable return. And so what previously was an acceptable return 20 years ago was mid-20s to low 30s IRRs. Then 10 years ago, as we started to see the public story anchor in, in towers and data centers, prices got more expensive, returns got compressed, and we saw an acceptable return would be a 20 -- 18 to 20 IRR return. And now with infrastructure capital in place, people are underwriting these deals to a 10 to 12 IRR. So we've literally seen massive compression in returns from the 30s to the 20s to the high teens. And then you look at what some of the public tower companies have done in M&A in the fourth quarter, and we're seeing yields kind of in that 2% to 3% range for European tower assets. For example, running the Hutch deal through our model, we barely scratched out to a 3% IRR. That's not acceptable for our capital. So the environment has gotten very competitive and certain CEOs are going to underwrite to certain returns. But there is definitely a flood of capital that's entered the space. And so as we think about public auctions and processes of that nature, we're typically, Mike, not very competitive in public auctions. We could have chased the Hutchinson deal, for example, against Cellnex, but why am I going to chase a deal to 30 times? That doesn't make any sense. So we've got to keep doing what we're doing and what we're doing works. And as we think about, for example, the 3 new deals we did in Q4, they were all proprietary, not involved. What we did in Edge Point in Southeast Asia was proprietary. What we did in agile in Japan was proprietary. What we've done in working with our new wireless Eastman portfolio in the U.S., that was proprietary. So all of the new ideas that we're cooking are proprietary ones, and that's where we find the best value. And what I've learned in doing this for 26 years is you've got to have patience. There are cycles in this. Things go up. Things go down. And we had our best years in 2008 and 2009. I think there's definitely some correction coming at some shape or form over the next couple of years, just given the dinner bill that's going to have to be settled up for the pandemic. And we'll have a lot of access to capital, and we'll have a lot of really good ideas. And I'm optimistic. I'm always an optimist. So -- but the key is we've got to run our race. We don't run the race the other guys are running. And so that's what's different at Colony. And that's worked really well for multiple decades. We've been able to put up great returns for investors because we've had the discipline to understand when you go chase M&A and when you go chase greenfield. Everything you've heard about my narrative today is we're very busy building. We think that's the right value proposition. I can build -- I was just thinking about fiber the other day. I can build fiber for $60,000 per route mile. Good dark fiber, high strand count. And I look at some of the M&A deals that have been done at 350,000 to the Astound deal that was done at $525,000 per route mile, and that just doesn't make any sense. We're talking about deals that are getting done at 3 to 5x replacement cost. And so I look at that and say, if stuff is trading in the M&A market at 3 to 5x replacement cost, I better hurry up and get a shovel in the ground and start building. And so that's what we're doing in all of our businesses today, whether it's Zayo building new dark routes, whether it's Freshwave in the U.K., building new small cells. We told you about what we're doing in Vantage Europe. We're builders everywhere across the globe today. And I think that's the big theme for 2021 at Colony, is having a deep pipeline, having great customer relationships, having formed a ton of capital at the end of the year. We'll continue to form capital in '21. And we're going to be active builders. Yes, we'll be opportunistic in terms of the M&A deals we do. We haven't left the M&A market. We'll continue to be pragmatic about it and chase M&A where it makes sense. We were very active in the M&A market in the fourth quarter, but we were very selective. We picked the right pockets where we felt like we could, once again, go chase value.

Everybody says now for which asset is going to have the best growth over the next 5 years?

Yes.

Okay. So we have...

Agree.

54% edge and far edge, 46% fiber and small cells. And then that was it. Those were the respondents, coupled into those 2 buckets.

It's interesting everybody based C-RAN hosting. That, to me, is fascinating. But I think -- look…

And that's what I want to come back to.

I think Edge and C-RAN hosting are kind of the same. I mean, I know what we're doing in our edge data center business, and we're hosting the RAN in those data centers. So RAN hosting is probably the same as edge data centers, really, if you think about it. In the far edge, which is really these micro data centers, that is an exciting product. I think the investment we made in edge presence. There's 2 other players in the market, Vapor, Edge Micro, all the tower companies have made their various investments in these companies. And I think if I was sitting in -- if we were sitting at the table with Jay and Tom and Jeff, we'd all be talking about those strategies and what do we think it looks like and what's the deployment schedule. It's nascent at this point. I mean, we've deployed 9 edge facilities across our portfolio. We'll do more this year. We've got probably anywhere from a dozen to 20 on the drawing board. And so far, the returns are good. We're getting good rents for people that are co-locating in their shelters. We've got 2 sites that now have 3 customers -- more than 3 customers in them. So it's coming. But it is -- I agree, it's a 5-year -- you have to take a 5-year view at this stuff. You can't take a quarter-to-quarter view on the -- on what I would call the far edge data center business.

I want to come back to a question on the C-RAN, O-RAN and some of these deployments that you're doing. But before that, real quick, is why isn't spectrum an infrastructure asset today? And could it be in the future, just given the evolution of technology. You have companies that facilitate sharing. Any perspectives on whether spectrum enters into the equation in the future as a digital infrastructure asset?

We've debated this ad nauseam for the last 8 years in our investment committee. This and satellites both. And ultimately, we've gotten a no on both those asset classes. They're not infrastructure. We arrived at no for various -- for different reasons. I think let's start out with spectrum. Spectrum, for us, is a permission from a government to operate a spectrum band. It can be revoked, and it has technology risk because the spectrum itself does not produce long-term sustainable cash flows, which is one of the core fundamentals of infrastructure, whether it's a toll road, renewable, midstream or digital, you have to have a long-term contracted cash flow associated with that to be infrastructure. And so to buy spectrum on a speculative basis, unless you're going to lease it to Verizon for 25 years, it's really hard to call it infrastructure per se. In addition to that, infrastructure usually has a physical aspect to it, where you're taking fee or you have a long-term lease or you have a long-term easement. And ultimately, where you can finance it and you can record some sort of instrument against that to finance it on an infrastructure basis. So there's 4 corners to underwriting digital infrastructure. There's the physical location. There's the -- ultimately, the collateral, which is underlying the real estate. There's the cash flows, and then there's their ability to put incremental growth on it, can you provide colocation? And so as you know, spectrum doesn't have colocation characteristics. So I think on the sort of 4 core pillars of underwriting digital infrastructure, spectrum just sort of falls out a bit. It just doesn't fit the box. I would say satellites, you could make the argument but really, from a real estate collateral perspective, a real asset that's affixed to something that you can record and ultimately finance the way this asset class should be financed from a securitized perspective, satellites have been hard. And I think one of the big things that really where satellite struggle is, it has a useful life. And that useful life is not very long. Typically 7 to 10 years. Maintenance issues are really hard. What are you going to do? You're going to fire up SpaceX and go up and send somebody up to fix the satellite. There's just a lot of different moving pieces in satellites that make it -- that don't make it digital infrastructure. Now it's a good investment class. There's people that are going to make a lot of money in satellites over the next decade, but we just don't think it's digital infrastructure at the end of the day.

That's really helpful. And our last question is to C-RAN, O-RAN, the hubs, mobile edge compute nodes. You've been talking about the C-RAN and O-RAN now with us for a couple of years, and it sounds like you're really getting traction with that. I think what our clients wonder about is what's the natural home for those opportunities? Is it in the fiber and small cell business? Is it in a data center business? Is it all of the above? Is it just going to be a mix of everything? And so as you have portfolio companies across the full spectrum, sorry, to bring that back into the conversation, of assets, where does this go? And how do we think about where to look for these opportunities for C-RAN, O-RAN hubs, et cetera?

Well, look, the natural home for these hubs is in a REIT vehicle. That's the natural home. And so whether the 3 public tower companies here in the U.S. get more aggressive in owning that type of infrastructure or developing it, it remains to be seen. I would say the last 3 to 4 years, most of the RAN hubs we've built have been at ExteNet. And some of those RAN hubs we turn over to the customer, and some of them we own. In San Francisco, we own 2 really significant hubs that have hundreds of radios in them. And now in that ecosystem, we're starting to see public safety. We're starting to see IoT. We're beginning to see cloud providers co-locate there. So we're very encouraged by that. And I also think that the work that DataBank has been doing on the edge, they're starting now to see some of the OEMs move into their data centers. And look, when we see Nokia and we see Ericsson move into our edge data centers, we know what the next step is. The next step is they're moving sort of some of the virtualized core there. Then you get base stations there, and then you get ecosystem. And so we think ultimately that for DataBank and Equinix, that will be a big opportunity. And then the question is whether or not companies like DataBank and Equinix end up marrying up with a tower company. That could make a lot of industrial -- could have a lot of industrial logic in the future. I just think the customer relationships and who we're doing business with at DataBank ultimately probably marries up pretty well with tower companies. That's sort of a prediction for the future over the next 2 to 5 years. And I think, look, listening to some of the -- some of my public tower companies CEOs talk in the last 90 days, they know they have to have an edge solution. It's inevitable. And so for these guys, they -- knowing that they have to have an edge solution, they've got to figure out how are they going to do it. Are they going to buy? Are they going to build? Probably pretty hard for them to build because they don't have the web scale relationships. So it's probably a little combination of initially buying. You saw SBA and American Tower put their toe into water buying colo facilities in Atlanta and Jacksonville. You've seen Crown's involvement in Vapor IO and their involvement in fiber. So I do think, ultimately, the edge opportunity probably ends up resident in data center companies, or it ends up resident in tower companies, one of the 2.

That's really helpful. Marc, great to see you. Thanks for sharing all your thoughts with us today, and best wishes for the year to come.

Yes. Absolutely. Thanks, Michael. And hopefully, we're all together in Vegas next year.

Yes. I look forward to doing this in person again. Thanks for our audience for joining today.

Thanks, Mike. Take care, everybody. Happy New year.