Today’s episode features an interview between Matt Trifiro and Colby Synesael, Senior Equity Research Analyst at Cowen and Company. In this interview, Colby offers his unique perspective on the investment decisions that are powering the growth of edge infrastructure, the elements that will enable True 5G, and how the pieces work together to usher in the fourth industrial revolution. The Edge Computing World Conference is October 12-15th, 2020. Fully virtual. Register at www.edgecomputingworld.com and use the promo code OVERTHEEDGE for 30% off the Edge Executive Conference
Today’s episode features an interview between Matt Trifiro and Colby Synesael, Senior Equity Research Analyst at Cowen and Company
Colby has been covering the communications infrastructure and telecom services industries as a senior research analyst since 2006, and recently co-authored Cowen’s extensive Ahead of the Curve report on edge computing and the future of the internet.
In this interview, Colby offers his unique perspective on the investment decisions that are powering the growth of edge infrastructure, the elements that will enable True 5G, and how the pieces work together to usher in the fourth industrial revolution.
Key Quotes
“While edge computing is not dependent on 5G, 5G is dependent on edge computing…The edge infrastructure that is obviously tied to edge computing is going to be required to enable what we refer to as ‘True 5G.’”
“When you take all three of those aspects--the bandwidth speed that’s north of a hundred megabits, the latency that’s sub-10 milliseconds, and the massive machine connectivity or IOT—that, combined, is the recipe to enable True 5G.”
“The beauty of True 5G is it's going to enable a lot of new businesses to leverage a wireless network beyond the phone that's never really happened before…And when we talk about True 5G, that's really what we're referring to, and to make that happen, absolutely a physical edge infrastructure is going to have to be out there.”
“With these three key attributes–the speed, the latency, and that massive connectivity–ultimately it enables what many people refer to as the fourth industrial revolution. And that's going to create a whole new set of market opportunities that are going to happen across just about every industry.”
“When you think about the internet in its current iteration, it's really been built for humans…This next iteration of the internet–which very much is tied to edge computing and 5G– is actually going to be built more for machines…It’s going to unleash this huge opportunity, and that altogether is what we refer to as the fourth industrial revolution.”
“The way I think about it [edge] is that it's not an explicit definition. It's not black and white. It's an evolution. It's a concept. And we're already starting to see that happening today. And at some point…we're going to see an inflection point where there finally is enough infrastructure that you start to see some of these newer use cases start to come to market.”
“I don't think of the edge as cannibalizing the cloud. I think of the edge as expanding the TAM–the total addressable market–for the cloud...You're not cannibalizing the demand, you're augmenting it.”
Sponsors
Over the Edge is brought to you by the generous sponsorship of Catchpoint, NetFoundry, Ori Industries, Packet, Seagate, Vapor IO, and Zenlayer.
The featured sponsor of this episode of Over the Edge is Ori Industries. Ori Industries is building the world’s largest edge cloud. Their products power the next generation of intelligent applications through unparalleled access to major communication networks worldwide. Ori is laying the foundations for application developers to seamlessly deploy to uncharted edge computing infrastructure across the globe. Learn more at ori.co
Links
CLICK HERE to Register for the Edge Computing World Conference, October 12-15th, 2020. Fully virtual. Use the promo code OVERTHEEDGE for 30% off the Edge Executive Conference
[00:00:00] Matt: Hi, this is Matt Trifiro, CMO of edge infrastructure company, vapor IO, and co chair of the Linux foundation, state of the edge project. . Today. I'm here with Colby Synesael senior equity research analyst at Cowen and company.
[00:00:11] We're going to talk about Colby, his career as a technology analyst, his current views on edge and all things 5g and where we're headed over the next decade. Colby. How are you doing today?
[00:00:22] Colby: I'm doing great. Thanks for having me. I'm appreciate it.
[00:00:25] Matt: Yeah. We really like getting a lot of interesting guests on the show. I'd like to start a little bit about just your background. as a technology analyst, how'd you even get into technology.
[00:00:35] Colby: Sure. I came out of undergrad in 2000 and, my first job was for a company called allegiance telecom, which some people might remember, and it was a C like a competitive local exchange carrier. And, at
[00:00:48] Matt: what was your degree?
[00:00:50] Colby: yeah, my degree was actually believe it or not in psychology. And I had a minor in English, so not necessarily something you would tie into a what I'm
[00:00:57] Matt: I was philosophy and cognitive science, so I totally [00:01:00] get it.
[00:01:00] Colby: Oh, fair enough then. But, I did a sales job. at that time I was pretty focused quite frankly, on I'm trying to make a few bucks. And, I thought that would be a good way to start off my career. And I was doing sales and I did it for probably about three or four months, not a very long time at all.
[00:01:15]and quickly realized I wasn't probably meant to be a sales person. realized at that point that I probably should be doing something different. And, I had a friend who was working for an investment bank at the time called Thomas Basel partners. And he mentioned that there was an opening and this was when I was living in Boston at the time.
[00:01:32] And, I applied for that job and got hired as an equity research associate. And, that was somebody who basically did a lot of the grunt work, a lot of the heavy lifting, if you will, for the senior analyst and that senior analyst was carving at that time enterprise software. So you could think of names like, Microsoft as an example, Oracle, et cetera.
[00:01:53] And I did that job for. Probably about a year. And then after that, there was a promotion that was made [00:02:00] available for somebody to work on the telecom services team. And believe it or not just given my three or four months of experience working for allegiance, I was asked if I'd want to apply for that job given by, experience in telecomm.
[00:02:12]and I did, and lo and behold, I got that job quickly moved to New York city and I've been doing, telecom services, equity research ever since.
[00:02:20]Matt: That's amazing. for those of my audience that don't know what. Cowen and company does. I became aware of you a few years ago, but only indirectly through the legendary Aspen conference you hold every year, which I've had peers call it the Dava data centers. but you and I came to know each other when you were doing some research on edge computing, and I've learned quite a bit about Cowan.
[00:02:45] It's a very impressive firm, but most of my audience probably doesn't know. What you are, who you serve and why you need research.
[00:02:54] Colby: Sure. so I got to correct you that the conference is actually in Boulder. but everything else that you said about the super-light is, are [00:03:00] accurate. it's a great event. And hopefully some of the listeners have some time we're able to get out there for the event, which always happens in August.
[00:03:07]so Cowan is an investment bank. And if you think about it, lots of peoples know the Morgan Stanley's and the Goldman Sachs and the cities of this world, but there's a second tier of investment banks. which Callen is one. And, that would be named like cow, but also names like Jeffery's or our RBC.
[00:03:25]maybe Oppenheimer is other examples. And we're a publicly traded company. so you could find this listed under the ticker C R O w. And I've been working at Cowen for, about 11 years now. And what Cowan does is firm as an investment bank, is we really focused on a few different things.
[00:03:44] Number one is we do have a banking practice. That's not the group I'm in. But we do provide advisory services when companies are coming public or being sold. we also have a, trading desks though, and that's really where my kind of broader focus happens, which is [00:04:00] we help our clients with their typically hedge funds and mutual funds to buy and sell stocks.
[00:04:06] And what I do on the research side is I provide advice. so I cover, as you mentioned, communications infrastructure. Which we consider to be the power of the data centers, the fiber name. And then I also cover the. services, space, predominantly names like 18 T and Verizon. But what I do, I put buy, sell and hold ratings, or at least at town, we consider them outperform record perform and underperform ratings.
[00:04:31] And I go, and I talk to our clients again, those hedge funds and those mutual funds. And I tell them the name that we think they should be buying the names effectively. We think that they should be selling and ultimately we, if they find our advice, a value, when they actually go to buy or sell those stocks that are trading best.
[00:04:47] And our trading desk will make a few pennies for every share that we trade on behalf of our customers. And you can imagine that when you're trading millions upon millions of shares each day, that starts to add up to a decent amount [00:05:00] of money and that's effectively how we make money for our firm.
[00:05:04] And I would say that in the equity research role that I have. I spend probably about 50% of my time talking to those clients. And another 50% of my time actually getting into the field to which I cover communications infrastructure and telecom services to better learn and understand obviously the companies to which I'm covering
[00:05:22]Matt: Yeah, that's really interesting. And so Cowan doesn't sell it's research. It provides it as an information resource and thought leadership in order to hopefully get that trading business.
[00:05:33] Colby: That's exactly right. it's funny. we're one of the few businesses where we actually don't put a specific price. on the services to which we sell, we effectively go into our clients and say, listen, if you value what, we're, what we're telling pay us what you feel is appropriate.
[00:05:47] Now, we obviously have some threshold that you have to hit, but there's really no explicit dollar amount tied to what it is that I do for a living. Quite frankly. And as a result of that, when we are writing our research, it goes to a pretty wide audience. some [00:06:00] people trading with our firm every day and therefore would be considered a great client.
[00:06:03] And others probably never traded it at all with us. I'd also point out that, it's a value to me to have people within the industry, to which I cover reading our research. So over the years I have probably built up a list of, I'll say about 500 different people that are within the industry. So they would be.
[00:06:20] Senior executives at power and data center and wireless companies. It would be private equity and venture capital type investors, which again, aren't really my core clientele, but my research is also going out to those people and they don't pay us a single dollar. but ultimately by seeing our research and helps to quite frankly, expand my brand with people that I want to know who I am, so that when I do engage with them, whether it's over email or phone call or see them at some type of event, they're willing to speak with me.
[00:06:47] And ultimately that helps me to expand my role of X so that I could learn more from people who are actually in the industry.
[00:06:55] Matt: Yeah. I'm one of the fortunate people that has seen some of your research because you've sent it to me. [00:07:00] And in particular, yeah. The five part series living on the edge. it is a incredibly impressive piece of work. it is comprehensive. there was a tremendous amount of thoroughness to it. And it's the kind of report that in private industry, you would not bat an eye to pay $10,000 for.
[00:07:19]and so it's, It's
[00:07:20] Colby: It's funny you say that because, people sometimes ask, we're say, I'm like, I'll put you on my research distribution list. And I'm like, it's how much is that? And I'm like, free. I'm not going to charge you anything to get our research. He's Oh, of course.
[00:07:30] And then sign me right up.
[00:07:31]Matt: Yeah. So let's go back to you. You mentioned the areas that you cover and, you rattled them off as if they all belong together, but I heard telecom infrastructure and data centers, and I think traditionally, I wouldn't have thought of those as being in the same, sector. Why are they together?
[00:07:51] Colby: sure. So if you think about, Telecom services. going back into the 1980s, you're talking about just traditional phone services, but if you think about the [00:08:00] tangential services that telecoms began to provide, in the eighties and then into the nineties and beyond. It included other services.
[00:08:08] So it included, for example, data center services. If you're an enterprise customer, Verizon, for many of those customers, they needed some type of physical place where to put some of their it infrastructure. And that was effectively data center services. at the same time, You have, fiber, because we all know today there's effectively an independent industry, a pure play companies that are providing those types of services, like dark fiber and wavelength services.
[00:08:34] In addition to the, big names that we all know, like the TMPs enterprises, the incumbents, if you will. And then also there's the tower space, which also, at first, actually those towers were owned by the telecoms themselves. And then over time they got spun out into these third party.
[00:08:49] Business
[00:08:50] Matt: And now they're buying fiber assets.
[00:08:51] Colby: and that's it. And now they're buying fiber assets, the crown castle. and the point though, is that, there's these other tangential spaces outside of just traditional telecom services, they've [00:09:00] expanded from the traditional telecom services industry. And those are the towers in the data center, in the fiber space.
[00:09:07] And I'd also take it a step further and say, though, that what's interesting about. those three verticals, tower fiber and data centers is that they now fall under note under their own umbrella, which is what we refer to as communications infrastructure. They're effectively the infrastructure assets, that enabled communications and the internet to work.
[00:09:25] And I'd also take it a step further and say that those business models, in many cases are actually much more aligned with the real estate model than they are with a traditional telco or technology. And what I mean by that is. When you think about, technology or oriented businesses, unit pricing, over time goes down, thought was Moore's law.
[00:09:45]and basically if you think about, dial up service back in 2000, we're all taking it. you are getting, I don't know, a 64 kilobit connection for a little, say 24 hours, $20 per month. You flash forward to where we are today and you're paying maybe $50, [00:10:00] but you're getting a 300 negative.
[00:10:02] Connection. If you look at it on a per bit basis, your pricing has gone down, but that's obviously enabled you and I as consumers to consume more. And it's enabled new business models like Netflix, opposed to just email back in 2000 communications. Infrastructure is different, though. It actually follows more of what I referred to as his real estate type pricing or inflationary type pricing, said differently.
[00:10:24] The cost to being on a tower actually goes up over time. There's actually escalators. Built into that pricing. The same could be said for dark fiber on the fiber side. And then on the data center side, also, you're seeing pricing going up. And when you start to think about the financial characteristics of those three businesses, they're very different quite frankly, than telecom service.
[00:10:44] And that's actually why you've seen a lot of those businesses actually become REITs, over the last several years.
[00:10:49]Matt: Yeah, that makes a lot of sense. my company is to a very real extent in the real estate business, and building edge infrastructure. And, the thing about it edge is there's a very, a lot [00:11:00] of it has to do with physical proximity and there's no way to get physical proximity without contemplating real estate.
[00:11:07] So I definitely see how those two. Worlds merged, back to telecommunications. So one of the things that I've heard you say, or maybe you've written, and I'm going to paraphrase it, et cetera, number the exact quote, but it was essentially that, five G is important, but true 5g won't be here until we have a computing.
[00:11:26] What do you mean by that?
[00:11:27]Colby: we've said is that, while edge computing is not dependent on five G. five G is dependent on edge computing. And what do we mean by that is the, that the edge infrastructure. And we'll talk more about that later on, but the idea infrastructure, that is obviously tied to edge computing is going to be required to truly enable what we refer to as true five G.
[00:11:46] And if you think about five G similar to edge, it's a term that kind of gets, abused and has many different meanings to many different people. But if you think about it today, Most people associated with their phone. they think about their phone [00:12:00] is getting faster in terms of bandwidth connectivity.
[00:12:02]and that's absolutely accurate with a 5G connection, opposed to a 4G connection. Your physical phone, will be faster than it would have been on a 4G network. We're talking about, on average, speeds North of a hundred megabits per second. Whereas today's Forgy networks probably have speeds closer to 50 megabits per second.
[00:12:20]But that's not the only aspect of five G and that's where the true 5g argument comes into play here. There's two other key aspects. One is latency. today's 4G networks have, probably somewhere around, I'd say maybe 50 milliseconds of latency. but it 5G network is intended to have sub 10 milliseconds of latency.
[00:12:42] And then the third aspect here is, IOT today's. 4 G networks and 3g networks, and even two D networks could only say so many physical devices connecting to that network at any one time. And what 5G promises is to basically support a heck [00:13:00] of a lot more devices in a given area that can connect to the network.
[00:13:03] At one time, we can get with four G and when you take all three of those aspects, the bandwidth speed that North of a hundred megabits, the latency that the sub 10 milliseconds. And the, massive machine connectivity, or IOT that's combined the recipe to enable true five G. And what's interesting is that while the faster banner speed is gonna be great from a phone experience perspective, having lower latency or being able to connect all these different types of devices really has nothing to do with the phone, but the beauty of truth by G and it's going to enable a lot of new businesses to leverage a wireless network beyond the phone.
[00:13:40] That's never really happened before. And that's when we start to think about things like sensors and connected cars and all these other things leveraging and utilizing a wireless network. And when we talk about true 5g, that's really what we're referring to and to make that happen, absolutely an edge infrastructure, physical edge infrastructure is asking to have [00:14:00] to be out there.
[00:14:00]Matt: Yeah, that makes a lot of sense. And so let's connect the dots on those. so When I have this physical edge infrastructure and I've deployed a five G network that gives me the bandwidth latency and the ability for massive machine connectivity. what happens? what devices are going to be connected, what sensors are gonna be connected and what are we going to do with them?
[00:14:24]Colby: that's the exciting part. and unfortunately that the punchline to some degree is as, I don't know, but if you think about 4g, Ford, you is, was really the first time a wireless network, enabled us to have broadband connectivity that was. similar to what you actually could get at home broadband connectivity.
[00:14:41] And once that happened, new services were created. So think of Uber before 4g, when you're on a three D network, you're able to get email, but you weren't able to get that broadband experience and therefore things like Uber didn't exist, but once you enabled Uber, actually, once you enabled that broadband speed.
[00:14:59] Lo and [00:15:00] behold, you had Uber. The same thing is going to be happening with five G, which is I'm going to give, not me personally, but we are going to give to them to the world, these three key attributes, the speed, the latency, and that massive connectivity. And it will be up to the creators, the innovators out there to come up with what is that next solution or service now, generically speaking, people refer to it as things like.
[00:15:24] Augmented reality. So solutions that might use augmented reality or virtual reality, people will talk about things that we'll use. autonomous vehicles, as another example that people talk about things like remote surgeries and stuff. So far, there's all these different ideas that are out there.
[00:15:38] But the admittedly right now, conceptual. We really haven't seen them come to fruition yet because the physical network hasn't been built. But when you think about companies like 18 T and Verizon, at least here in the United States, They have incubation labs there. They're working on trying to figure out what are some of these services or working with some of those creators to try and figure out how do you actually create these services?
[00:15:58] And then ultimately, how will you bring them to [00:16:00] market? And I think that's going to happen over the next several years. And when it does, what we think ultimately does is it enables what many people refer to as the fourth industrial revolution. And that's going to create a whole new set of market opportunities that are not just telecom specific, but really are going to happen across.
[00:16:17] Just about every industry. if you think about autonomous vehicles that obviously has huge ramifications on the auto industry, but if you think about things like remote surgery, that's going to have to draw implications on the healthcare industry. And if you think about things like autonomous, can you think about things like virtual reality or augmented reality?
[00:16:36]that's going to have potentially huge impact on, on sports and entertainment and it's really, again, it's taking these core components. and then leveraging them across a whole broad set of industries. And when that happens, we think it often creates a whole new set of demand that's just not in the market today.
[00:16:53] And one of the interesting things that we talk about is that when you think about the internet in its current, iteration [00:17:00] it's really been built for humans. So think about how we use it for email. We use it to go onto an application or a website. we use it to watch video this next iteration of the internet, which pretty much is tied to edge computing and five G is actually going to be built more for machines.
[00:17:19]and it's going to be, think of things like, robots, think about things like sensors. think about things like drones and those are all, examples that are going to need, not only that faster broadband connectivity, but they're also going to need that lower latency. they're also going to be many of those devices, IOT effectively.
[00:17:39]and that's when you get to that world where it's going to unleash this huge opportunity. and that altogether is how we think of the, what I referred to earlier as the fourth industrial revolution.
[00:17:50]Matt: Yeah. And that line of thinking, lines up nicely with the thesis that we had in the 20, 20 stave yets report, which, begins with this idea of the three acts of the internet. [00:18:00] And we're now in this third act and the way that I explain it to what just about everybody, but particularly lay people, as I say, the big change that's driving the need for five G and edge computing is exactly what you said.
[00:18:12] It's we're moving from a world of primarily people. Talking to machines or people talking to people like you and I are two machines talking to machines and the latencies and bandwidth for humans are orders of magnitude, less than the latency and bandwidth that machines, can consume and desire.
[00:18:32] And that's going to drive a need for a whole new internet, but what's really fascinating about. your business is looking at taking this pan optic view of all these industry and all these trends, and then backing it down into investment decisions. And, one of the things that in a new industry like this, where.
[00:18:52] In order for it to really realize the, financial growth, the platform has to be built. And that [00:19:00] means many companies spending billions of dollars and lots of time to put, A lot of infrastructure in place, hanging radios off of towers and light pole and street, side cabinets and micro modular data centers and regional data centers and all of this.
[00:19:14] And, there's a little bit of a chicken and egg problem, which is the people who want to deliver these applications. So why should I invest in the application until the infrastructure is in place? And the people building infrastructure is why should I build the infrastructure until I know which applications are going to make money and you ended up in this kind of, deadlock right.
[00:19:29] And investment. I see that breaking, but I'm super curious, like, How the company is you, how you see the company as you cover justifying these investments. So far ahead of the curve in many senses, and how you advise your clients who are making investments in these companies, how they should think about this chicken and egg problem.
[00:19:48] Colby: So you, yeah, I absolutely have clients that, want to know what stocks going to go up tomorrow. And if I'm lucky, maybe they want to know what's going to go up in the next quarter. But we also have other clients that are very much [00:20:00] longterm oriented and do have two, three maybes in five year time horizons in terms of their investment.
[00:20:06] And that's on the public market side. And admittedly, as I mentioned earlier, My core clientele is not private equity. but they're also very much involved in communications infrastructure. And I have lots of relationships with those types of people and they absolutely have a much more long dated, capital in and are making investments, for five years, perhaps even 10.
[00:20:26] And as a result of that, When I'm trying to do my research, I'm trying to really place all those constituencies. And when we write these ahead of the curve reports, which is we have calendar referred to as our, how we branded our white papers, it's really those longer time horizons in terms of how we try to think about this stuff.
[00:20:43] We typically, by the way, for what it's worth, each analyst account, we'll offer maybe one or two of those. Per year. But when I think about the opportunity set for companies like a crown castle or an aquaponics or Verizon, the way that we think about it is, as it relates [00:21:00] specifically to the edge is that this is going to just bleed out over time.
[00:21:03] And it's not as black and white as the chicken rag. Argument to what you made, meaning that, we think of edge as a concept meaning that it's not an explicit definition and the definition of edge is going to be very dependent on each company. And it's gonna be very dependent on the actual application or use case to which that edge solution is.
[00:21:25] And when you think about, someone like a crown castle, as an example, who's building small cells. Today, they're building small cells, predominantly for two use cases, it's actually to alleviate some of the congestion that's happening on the macro tower side, so that those macro towers could effectively propagate and work the way that they were supposed to.
[00:21:45] When you build that small cell below it, if you will, to take some of that traffic. Or congestion off of it. And that's where they're there. They're building up these small cells today without necessarily having an explicit 5g use case. if you think about that, if they continue to move in that path, [00:22:00] we're going to get to a point at some point, whether it's in a given market or perhaps across the country where they built enough small cells, that others that are you going to use them for five G an edge use cases can actually now start to use those at some level of scale.
[00:22:15] And then once you hit that level of scale, then it's. success begets success, meaning that you're going to start to see incremental demand from five G and from edge, which will then push more investments onto small cells. And the same thing can be said against across the fiber space. The same thing can be said against the data center space.
[00:22:31] In fact, in the data center space, while there's a lot of talk about what an edge micro or others like a vapor IO are doing in terms of building these modular, specific edge locations. whether it's at the base station of a tower or it's somewhere else, in the field, it will today.
[00:22:50] You're actually seeing, demand shift from what was these, core data center sites. for example, those owned by, an Amazon or Microsoft, we've seen that hierarchy [00:23:00] shift to where you're seeing incremental demand. Now in more regionalized locations, such as, for example, in suburban markets, a huge population centers.
[00:23:09] And they're starting to see, an Amazon or a Google or Microsoft deploying in their locations. And therefore that could be arguably considered edge. And then ultimately what we'll see is we'll see that bleed even further to where you actually are starting to see the Amazons and the Googles, the point in the vapor IO type locations.
[00:23:27] And again, so the way I think about it is. it's not an explicit definition. It's not black and white. It's an evolution. It's a concept. And we're already starting to see that happening today. And at some point, and it's gonna be different for each application and end user, we're going to see a, an inflection point where there finally is enough infrastructure that you start to see some of these newer use cases start to come to market.
[00:23:49]Matt: Yeah. Yeah. agree with that. so Colby, you mentioned small cells, and I'm sure that a lot of my audience may think they understand what a small cell is, but, actually is a lot of nuance to [00:24:00] it. can you help us understand what you mean by small cell and why those are so important?
[00:24:04] Colby: Sure. if you think about, the way that mobile phones work, what is the United States or just about anywhere? they work by leveraging off of the physical infrastructure, which is a macro tower, and we've all seen these when we drive down a highway and, depending on where you live, some cities and.
[00:24:20]Neighborhood associations don't want them to have their backyard, but the reality is we all need them if we want our phones to work well, as we've all shifted to a point where we no longer have just plain old flip phones, but we actually have smartphones, the, demand and requirements of each of those macro towers is going up.
[00:24:38]we're consuming more of the radio frequencies, if you will. that our phones are dependent on to actually, communicate with these macro towers. And that's leading to, as I mentioned earlier, some level of congestion and the more people you have using a tower in a given area, the less efficient that power actually is.
[00:24:57] And it also, it degrades the quality of service [00:25:00] that you and I actually have with the phones when we're using them. one of the ways in which you, alleviate that tension, do you do something called cell splitting, which is, if you actually had a macro tower location, you had a macro tower and another location.
[00:25:14] You're not putting a new macro tower in the middle, effectively alleviating the congestion that were previously being supported by just for example, those two towers. Now, for example, you'd have things supported by three.
[00:25:26]to take
[00:25:26] that
[00:25:26] Matt: Yeah. So a way to think about that was if you've got two towers and everybody in that area, or every device in that area is connected. These two towers, you add a third tower now you're load balancing across three towers instead of two towers. Is that correct?
[00:25:36] Colby: that's absolutely correct. And to take that step up that concept a step further, what a small cell does is it really targets in a very small area. typically a few hundred yards, To really take the congestion in that given area. So where you would typically see a small cell being put is in large population centers.
[00:25:54] So it could be on a school campus. it could be in a very popular tourist [00:26:00] destination, right? Such as, city in New York city. or it could be at a stadium. we're, out of the people gathered to watch an event and by putting small cells into those areas, it really does the same thing as a macro tower.
[00:26:13] It has the same, radio antennas. And in all that infrastructure on there that will actually communicate with your phone. But ultimately what it does is it alleviates some of the pressure that a macro tower net given area had previously been supporting. And, the concept of small cells are probably, I'd say at this point, Maybe five years old, maybe a little bit longer in terms of the popularity of using small cells, but it's still very much in its infancy.
[00:26:40] And to give you some concept of size, there's not perfect data on this, but there's probably somewhere between 150 and maybe 200,000 macro towers in the United States. and that number is probably growing a few thousand every year. from a small cell perspective. There's probably today, I would say about 200,000.
[00:26:58]small cells, be in the United States [00:27:00] today. But that number is expected to grow, by the hundreds of thousands, over the next several years. So there's a lot more small cells that are going to get put into the marketplace, but effectively what they're doing today is they're alleviating the congestion for that four G network.
[00:27:14] But in the future, that thought is that we're going to start to use these small cells more for these five G use cases. these true five. Do you use cases? He's not a phone use cases, that we've been talking about.
[00:27:25] Matt: Yeah. And as I understand it, another, important reason that small cells will be deployed at scale for five G, is that a lot of the higher frequencies using five G can't go through walls and sometimes not even trees. And so having more. Points to connect to, in and around buildings and walls becomes important.
[00:27:43]is that a correct statement?
[00:27:45] Colby: Yep. So when you think about frequency, there's effectively low band, mid band and high band spectrum. And we could think of low band as sub one gigahertz, and you could think of mid Dan is between one gigahertz and six gigahertz. And you can think of this high is really [00:28:00] millimeter wave, a type spectrum, which is typically above 20, gigahertz.
[00:28:04] And. The higher you go up in frequency, to your point at the lower, the signal strength actually can travel. And therefore, if you actually want to use millimeter wave spectrum, you're also going to need a lot more cell sites, for it to work, in a scalp fashion. And the reason that you would want to use by the way, higher frequency or millimeter wave spectrum is the other directional argument.
[00:28:28] Is that the higher, the frequency, the greater the band with capacity that spectrum has. And that's just a function of physics. And therefore, if we want to start to improve the bandwidth connectivity, I, we talked about getting to on average North of a hundred megabits per second. And in many cases, moving towards one gigabit type speeds, particularly when you're using millimeter wave spectrum to get that you, you have to start to use these, new, higher frequency spectrum bands, which we have historically not really used that much.
[00:28:54] When you think about wireless networks.
[00:28:55]Matt: Yeah, that's really interesting. This, that, and I think deployment of, you mentioned [00:29:00] 200,000 today, in five years, is that going to be 500,000 a million?
[00:29:06] Colby: Yeah, so when I look at other third party research, that takes more of a top down view of post bottoms up, which is what I would do. I would look at things from a company's perspective and then build up, there's others, like the Gartner's example of this world that can take top down to use on various industries.
[00:29:20] But when I look at those types of forecasts, you're absolutely right. And it depends on who your who's forecast you're looking at, but you'll see, somewhere around a half million and perhaps upwards of a million in the next 10 years. in terms of forecasted small cells being deployed in the United States.
[00:29:34] And therefore, while we're going to, as I mentioned earlier, see probably a few thousand macro towers every year, getting built out, there is expected to be, see this huge inflection, this growth curve, if you will, for small cells. But again, until, and it goes back to your chicken and egg until we hit that inflection point.
[00:29:50]you're seeing small cells being deployed today, but, maybe if, tens of thousands per year now, but at some point we're going to really hit a hockey stick permanent and we'll see that accelerate even more.
[00:30:00] [00:30:00] Matt: Yeah. And you can see how this drives the other components of, this sort of telecommunication infrastructure, because presumably each of those small cells is going to have a fiber connection, which means probably putting a lot more fiber in the ground or at least lighting existing dark fiber. Is that correct?
[00:30:17] Colby: absolutely. So when you think about, communications infrastructure, we mentioned data centers and we mentioned, powers, and you could even put within the power category, small cells, but the glue, if you will, for whether it's a data center or a tower or a small South. is fiber. And as a result of that, we're going to need to see a lot more fiber being built out, into the network than there actually is today.
[00:30:39] In fact, a lot of fiber. When you think about it, there's two key issues. One is that a lot of fiber was built, between core networks. so for example, between, different markets, between, Chicago, for example, in a, in the Milwaukee, but then when you get into a Milwaukee there wasn't a lot of.
[00:30:55] Fiber being built within that, given Metro, and we're going to see need to see a lot more. It [00:31:00] we're going into those metrics to support things like a data center or a macro tower or a small cell. The second thing too is then the actual density of fiber. And we think about density in terms of actual fiber strands.
[00:31:12] And when you think about legacy networks, there was never this recognition or realization that so many different occasions, whether it was a small cell or a macro tower or data center, even an enterprise type location. Would consumer require fiber connectivity. And as a result of that, when you look at, outdated fiber networks, they don't have physically enough fiber strands to connect all those different locations.
[00:31:35] So in many cases you're seeing new fiber networks going in that are overlaying on top of fiber that may have already actually been in there may have been in there for the last 20 years, but wasn't to the same level of density required to support these new use cases. And as a result of that, there's actually a tremendous amount of new fiber.
[00:31:52] Being built, particularly in the United States right now, to support many of these needs that we're talking about.
[00:31:57]Matt: Yeah. So let's shift to let's start at the [00:32:00] tower of the small cell and kind of work backward up cloud. and in fact this is how you and I met, looking for standardized terminology to refer to these. and as from looking at the industry, Everybody had a different word for it.
[00:32:13] One of the things that I've tried to do at the state of the edge in the Linux foundation is bring some sense and order and common agreement around what we call these things. Although interestingly enough, it went through another change while you were writing the report. but on this, in infrastructure edge or service provider edge as it's called, which.
[00:32:29]roughly stretches from the internet access point in a major metropolitan area to literally the tower, the small cell or the cable head end. There's a lot of stuff happening in there. Can you help us understand what happening in sometimes called middle mile?
[00:32:43]Colby: sure. first off, I just, take this opera and say that when we were doing our research on writing this report, the LS edge 2020 report was very helpful. And to your point, help provide some definitions around. How to think about some of these things. And we use a lot of those same [00:33:00] definitions in our portfolio today.
[00:33:01] Point I think Justin, in July, you guys put up an update and change some of those, which is understandable, just given how fluid the idea or concept of edge seems to be changing. with that said, when we think about the fiber connectivity, there's really three different.
[00:33:17]Components there's the core network or the backbone, which as I mentioned, connects various cities and typically the end points, of that backbone network is this is what we refer to as interconnect oriented, data centers, or some people might refer to it as carrier hotels. So for example, in New York city, 60 Hudson, or, one 15 Halsy.
[00:33:36]one 11, eight, those are all three addresses. Those are physical buildings in New York city or in the case of policy in New Jersey. and those are, key areas where a lot of different localized fees are in their service. Providers have their fiber connecting. Yeah. Now in a city like Chicago, you have three 50 Surmac, which is also a big carrier hotel where the localized peas and Matt.
[00:34:00] [00:33:59] Local region would connect into, but the backbone network would serve between for example, one 15 Halsy and, sir, max three 50 Surmac, that would be considered backbone and quite frankly, there's a lot of, connectivity already in the backbone and pricing is pretty efficient, but what you're finding is actually more telcos today are actually building.
[00:34:20]incremental backbone network to avoid, to reduce latency. So they're finding physical geographies to get there quicker. the other thing then is then the middle mile, as you pointed out in the middle mile is the part of the network that connects between, the last mile network, which is the local ISP, where the end user resides.
[00:34:38]and the backbone and the unfortunate part about the middle mile is that the middle mile, the routing is very efficient. it's uses something called, BGP, which is a type of routing to figure out, which is the best path to get traffic to where it ultimately needs to go. And because of things like competition and cost analysis, it doesn't always take the most efficient route [00:35:00] to get to where the end user device that's.
[00:35:03] Needs that data actually resides. And as a result of that, what we are seeing is we're seeing increased investment in middle mile. but then also then there's that last mile. And that last mile, as I mentioned, could either be local telco network. It could be the local fiber networks, could people cable network, it's typically, being supported by the local telecom or cable company.
[00:35:24] And that's the physical network that can take you to where that end device is. And even there we're actually, we're starting to see incremental investment being made, particularly as broadband connectivity, particularly a pan, has become so critical to so many people. And that too has now become a pretty, ripe area for
[00:35:41]Matt: Yeah. So let's talk about use cases. so you mentioned some, at the top of the interview, and it's ones that we hear all the time, and I've heard for years, Thomas cars and augmented reality and all this, and. at some point I can see all that happening. I was a fan of star Trek growing up, and I believe the holodeck is going to be something where good experience and, medicine's going to have a [00:36:00] tri-quarter and we're going to have all this, amazing computer at our fingertips in order to create these new use cases.
[00:36:05] But what's potentially more interesting is to look narrowly at what's emerging today. what are you seeing? Emerged today, but starting to look like production use cases, things that people are actually willing to pay for things that are justifying, ongoing investment from an end use case that uses this kind of infrastructure.
[00:36:25] Colby: sure. first off, I would say that I'll twist your question a little bit and I'll say that we're starting to finally see is actually the physical infrastructure being deployed. to enable, these cool solutions that you're talking about. So when you think about Amazon, announcing local zones or Amazon, announcing their wavelength solution, we're actually starting to get services there, that could enable some of these, ideas that we're talking about.
[00:36:51]second thing I would say too, though, is that when I think about an edge, location, And I think about who is the actual underlying customer [00:37:00] who is going to be able to provide the platforms to enable these solutions. It's very much in our opinion, going to be those cloud companies, the Amazons and the Googles and the Microsofts of the world.
[00:37:10] It's gonna be other cloud, like type businesses. so for example, a CDN, like an ACA Meyer or a Fastly, it's also going to be, other companies, such as, cloud security companies, like a Z scaler and those are the companies that we see actually going into these physical edge locations.
[00:37:29] To provide the platforms to create these or support these ed services. and that's the starting to happen now. a result of that, we're starting to see in terms of the initial use case, as you mentioned, as you're starting to see some of that start to populate, and Verizon actually has been working, as I mentioned earlier.
[00:37:47] With some of these companies. And one of them for example, is that's the gaming. So gaming is going to be probably one of the first I would argue edge services. You could appreciate things like augmented reality [00:38:00] or virtual reality solution. You can that the criticality of latency,
[00:38:05] in
[00:38:05] those types of
[00:38:06] Matt: my 10 year old knows more about latency than I do sometimes.
[00:38:09] Colby: Yeah, it's funny. It's like people get bandwidth and latency confused, in terms of, what do you actually need, to have a really great gaming experience, but you think about gaming that's absolutely. I think going to be one of the first, areas where we're actually going to start to see some of this come to fruition.
[00:38:23]Matt: Yeah, either improving the real time nature of. a multiplayer game and, or creating these new rich experiences that, might in fact exist in three dimensions. If I'm wearing a set of goggles or might exist in, in the real world, quote unquote, if I'm using an augmented reality application and that requires a lot of processing, so one of the things that's, a common trope in.
[00:38:47]in the field of computing in general, is this idea that there's this giant pendulum that swings back and forth between, mainframes to personal computers back to the cloud, back to this. and I've heard people place, edge computer, in [00:39:00] that, that metaphor. And I'm not sure that's actually true.
[00:39:04]and I'd like your opinion on it. This is why I don't think it's necessarily true because in the past, when the pendulum swung it really swamped. Meaning the bulk of the workloads actually stopped running in one place and started running in another place. And instead, what I'm starting to see an edge computing is.
[00:39:22] It's plus meaning the edge workloads do not seem to be cannibalizing the cloud workloads. It seems that the edge workloads are actually augmenting and in some cases actually increasing the need for additional cloud workloads. And I'm wondering you have an opinion on that,
[00:39:39] Colby: Yeah, absolutely. if you go in and look at the history of computing, back in the 1940, through the 1960s, we saw the, really the beginning of computing in the form of the mainframe and the mainframe was a effectively, a centralized. Computing solution. And the reason for that is a, it was really expensive.
[00:39:57]and bead mainframes, certainly [00:40:00] back then took up a lot of physical space. so you would put them in a centralized location and then, effectively people would be able to connect into that, to harness that computing power. And then I'll tell you what happened is that was replaced by, personal computers.
[00:40:14]and that was in the 1960s through we'll call the 1990s. And, that was, yeah, probably our first version of what we refer to as distributed computing and what that helped usher in it was an option. More people were actually able to use computing resources than what was possible when we were utilizing mainframes.
[00:40:34] And then the third phase of this was the cloud. As you just referenced, we went back effectively to a centralized computing platform, but the leveraging the internet and that allowed us. And that was probably in the two thousands and through the 2010, and that allowed us to gain a level of scale.
[00:40:51]you're able to effectively do more, by harnessing the cloud opposed to. What was possible just with your own personal computer. And now what we're [00:41:00] seeing is you're seeing fourth shift, which is effectively the edge, and we're going back then to a distributed, architecture.
[00:41:08]but what's interesting is it's still actually leveraging or using the underlying cloud architecture. that we saw come to fruition in that 2000 to 2010 time period. And therefore you're still maintaining that level of scale that you got by consuming off of the cloud, but you're gaining that efficiency by going back to that distributed model and here point that I don't think of the edge as cannibalizing the cloud.
[00:41:34] I think of the edge as expanding the Tam, the total addressable market, for the cloud. And. When I think about the internet, a hierarchy that the physical infrastructure that makes up the internet, I think about it in the form of data centers. And as I mentioned earlier, You have these large data center facilities in areas where taxes are lower and empowers cheap, that are typically [00:42:00] owned and controlled by these large hyperscale companies, like for example, on Amazon or Microsoft.
[00:42:06]and then you ultimately have, as I mentioned, those data centers that are in those regional areas where the population centers, those they're actually consuming these services reside. And now you're actually going to see a third layer, which is these localized data centers, which are effectively the edge, augmenting that previous hierarchy to which we had, which was predominantly those poor and regional locations.
[00:42:28] And by doing that, You're not necessarily cannibalizing the demand that's going to the, or it's going to those regional facilities. you're augmenting it. and I think that a lot of the data that's going to be created by these new services, because if you think about the autonomous vehicles, you think about that.
[00:42:42] Augmented and virtual reality solutions at the edge is going to enable. It's going to create a whole new set of data that doesn't exist today. And not all that data is going to be stored and computed at the edge. Some of it's going to move back to those regional data centers. Some of them are going to move back to those core data centers and dynamically.
[00:43:00] [00:43:00] It's going to change where that data is stored and computed, and it's going to be very different depending on the application. And the workload itself, but ultimately what the edge is going to do is it's going to create, I think, a tremendous amount of new demand, for all those constituencies and at the heart of it is going to be, the cloud.
[00:43:18] And as I mentioned to you earlier, a lot of the demand today for the cloud has come from human oriented use cases. and it's come from, shifting for example, the way that we work from an on premise type IP architecture to a cloud based. Architecture and it's come from us shifting how we consume, for example, video from a linear based solution to an OTT based solution.
[00:43:42] These new use cases that we talked about that are going to be machine oriented are gonna create a whole new set of demand. I think it's going to be really the next phase of growth for cloud computing.
[00:43:51]Matt: Yeah. Yeah. that's a, that's really interesting. so you mentioned earlier that a lot of the growth and edge is going to be driven by the cloud providers, but there's an [00:44:00] interesting dynamic between the last mile service providers, particularly the telcos and the cloud providers, because in the first wave of four GLT build out, as you mentioned, the.
[00:44:10]telephone network for finally capable of carrying rich data and supporting, the rich applications enabled by smartphones, but the. telcos feel that with the OTT, the over the top services, the Netflix is the applications that run on examples that they got cut out of lot of the revenue capture.
[00:44:27]I've heard them say, sometimes in really harsh terms, Over my dead body. Are we going to be a dumb bit pipe with five G? and yet you see some partnerships being at least publicly talked about, you mentioned Amazon with wavelength and Verizon, and I believe at, and T is also doing something with Microsoft.
[00:44:45]but I suspect there's. There's both, it's coopertition right. there's tension in that relationship. because ultimately the cloud providers have immense power and they have the application developers. and back to Steve bomber's famous words, [00:45:00] developers.
[00:45:00] It really does create. Tremendous gravity around application creation and deployment. what are you seeing happening in these backroom conversations? How are the telcos viewing their relationship to the cloud providers and vice versa?
[00:45:13]Colby: sure. You to your point on I'm on four G when you think about value creation, Yes. the wireless carriers benefited, from the emergency support GE and you saw tremendous growth is because people shifted over to smartphone devices and it started to consume more data. so by no means, zero sum game for them, but when you think about where the
[00:45:31] majority of
[00:45:32] Matt: Yeah, I have a $350.
[00:45:34]cell phone bill because I get data.
[00:45:36] Colby: that's right. So a week. we've all had those, bill shock moments, not just, what used to be on the voice ed, but now even on the broadband side, particularly don't have an unlimit plan still, but, but when you think about where the majority of the value creation, happened, it occurred at the application layer.
[00:45:53]you think about the Instagrams. The Facebook's, not to take tax in terms of, the report at $30 billion that they're [00:46:00] seeking in the United States. that's where the overwhelming majority of valuation came because of the creation of a Forgey. And it's your point that the carriers don't want to see that happen again, they want to make sure that they participate in the monetization of five G in a more meaningful way.
[00:46:18] Than they did with 4g. And the good news is I think five G actually is going to give them an opportunity to do that. it's through something called network slicing. And if you think about a five G networks and we think about true five G depending on the actual application service, they're all gonna have different requirements.
[00:46:40] Some for example, might need very low latency. There's might need very high bandwidth. but some might not. if you think about for you example, sensors that are in the marketplace some way it's actually really neat Lacey, so I might just be clicking the data and then once a week pushing that up into the cloud, you also would start to think about things like, service level [00:47:00] agreements in terms of, having to ensure that this definitely works.
[00:47:03] Because it's for example, a healthcare oriented and if it doesn't work, someone's gonna lose their life. whereas others are going to be for entertainment purposes and therefore, having a very high service level agreement may not be as important and slicing. What it's intended to do is basically allow each of these application, companies that are that are creating these files, these services, it go in there and say, I need these specific resources and I don't need these.
[00:47:29] Specific resources, but the fact that they actually need to go in there, into the network and basically say, I need this and I don't need that. It forces them to have a conversation with the carrier. And it also allows the carrier to monetize that in a way that they haven't been with foresee. Because if you think about boards, you is one plain vanilla network.
[00:47:49] Everybody had access to the same network and therefore, or once you created the actual application and your end user was paying for their. mobile services like you and I, there was really [00:48:00] no opportunity for Verizon or at and T or anybody for them. And to go insert themselves into that.
[00:48:05] Conversation five G is going to be different. And as a result of that, there is thought to be about monetization opportunity. The second thing I have to say, though, in terms of where are we today, it's really in exploration phase. And you think about all the different things that Verizon is doing in terms of, in, at and T for that matter.
[00:48:21] And T-Mobile, in terms of building out their, Various edge infrastructure. They're there, what they refer to as their Mac, mobile edge computing, type locations. And they're working with various partners, as I mentioned, but that's the gaming and Amazon, and Google, particularly in the case of, At and T they're trying to figure it out.
[00:48:40] And ultimately, I think we're going to get to a point where there is going to be more of a partnership between these clouds and these telcos, because these clouds are going to actually become increasingly dependent on the wireless networks. I think for where their next level of demand
[00:48:57]Matt: Seems to make sense. And I think five [00:49:00] years ago, There was a, a notion in the. telco industry, that was supported by, some of them buying up data centers, that they, ultimately, mostly divested of. but this idea that the telcos would offer their own cloud. And I think as appealing as that was from a potential revenue capture standpoint, what that idea missed is it's going to be really hard to attract.
[00:49:22]massive developers to a new cloud, there are, three to five major clouds, depending how you count. and, there's not an even distribution among those five. And the top three kind of make biggest impact and those developers, who developed for Amazon or Microsoft or Google cloud platforms, don't want to.
[00:49:40]Distribute to a new cloud. It may not be cost effective. the other thing that's interesting is that, you have multiple carriers and as a developer, I don't want to have to go and negotiate an arrangement with each carrier and. want to be able to develop an application and deploy it one time.
[00:49:57] And so I could see a [00:50:00] business arrangement being made where, maybe there's API integrations between the clouds and the service providers, where I as say an Amazon customer, today I can provision an ECE two instance in one of two locations in the U S well, in both locations, I want to us Westmead USC.
[00:50:14] So those are my choices. and there's. the time in the future where I imagine I'll be able to deploy instances, not just in U S West, in USC, but in Chicago, West and Chicago East. And if I'm deploying in Chicago, Western Chicago, East, I imagined that I would also want to be able to check a box that says, provision me and SLA guaranteed network slice, because that's what my application requires.
[00:50:37] And there'll be some business relationship where money would flow back to the carriers in that case. So I definitely see that happening. I wonder if you have an opinion on, the carriers collaborating to offer. A way for developers, federating the resources so that I don't have to go into every market, and talk to every carrier in order to have an application [00:51:00] that runs anywhere and not just us, but global.
[00:51:03]do you see any hope that. We're going to see some sort of open API or some other capability that would allow a developer to write an application that works across multiple,
[00:51:12]Colby: Yeah. so there lies the rub. so if we look about where we are today, and the various press releases that are coming out tied to five G and often tied to the edge, they are very, carrier specific. So AWS, partnering in a specific city like Chicago, with Verizon, No we've asked for
[00:51:32] In fact, we actually just had our communications infrastructures summit for tool this year post to Boulder. and we had Verizon there. And one of the questions that we asked them is, if for example, somebody has deployed a edge solution. On AWS and they're on an at and T network, but AWS is deployed in your facility in Chicago, w will that work.
[00:51:58]and the unfortunate part is, yeah, [00:52:00] they really aren't able to give an answer right now. And I think that the risk that we, yeah, Ben is, will AWS then have to go and deploy in Chicago, not just with Verizon, but also with, at and T and then also at T-Mobile. And, you would think that.
[00:52:12]AWS, isn't going to want to do that because there's, it's obviously not very efficient. so that's still something, that's a question Mark. and we'll have to get solved. to that point, one of the arguments we have broadly speaking French computing is I do think that given the sheer cost of making all this happen, in fact, that we're putting a lot more physical infrastructure and a lot more places, it actually will argue more for the neutral host bottle.
[00:52:34] Then we've seen historically, because give me a way to alleviate some of these costs. And then ultimately if that happens and therefore the future model isn't necessarily the UWC deploying an OB. Verizon facilities, perhaps in a neutral location and then Verizon isn't necessarily gonna be cut out.
[00:52:53] The business model is going to be, again, as we talked about more tied to things like that, like slicing, which they'd still be able to do [00:53:00] the ability to actually physically have that edge, a computing infrastructure in their network, or at least physically in there, edge data center type locations that may actually map, sustain
[00:53:14]Matt: I'm glad you brought up the neutral host model, because I agree that is one of the most important concepts that has been around for a while. but is now being applied in new ways. obviously, for a long time, carrier hotels have been neutral host. Anybody can bring their network and connect to anybody else's network.
[00:53:32] And to some extent, the tower industry moved from, vertically integrated with the telcos into, a neutral host model. And I'll say neutral, certain quotes. Cause it doesn't quite mean the same thing, but it's shared infrastructure, meaning. costs me a lot less to deploy my radio head on a crown castle or, an American tower being advertised with.
[00:53:52] Two or three other companies Radiohead's on that tower. and so I can yield a time to devalue, acceleration. I can potentially lower my [00:54:00] cost of deployment, my cost of maintenance. And you could see that being applied to other pieces of infrastructure, like fiber routes or, data centers with co-location, things that weren't typically, To the best of my knowledge in the U S carriers don't share street side cabinets.
[00:54:17]but you could see a world that says, okay, it's going to cost us, X to build a street side cabinet. and it's gonna, occupy this space and somebody has got to, as you pointed out, by the real estate and get the permits and put the physical thing there, why don't we build one that can hold.
[00:54:31]the equipment of three carriers and, reduce everybody's costs and, keep the city a little less cluttered with street side boxes and stuff like that. So what are you seeing in reality on the ground from a shared infrastructure perspective
[00:54:44]and where the trend is going?
[00:54:46]Colby: like everything else. it depends on the price. No. When we think about small cells as an example, in really the crown castle argument. they would make the same exact argument that you just made. But in some situations, when we've spoken with the carrier there, their argument has been that the [00:55:00] price point to which they're charging, no, it doesn't reflect shared economics.
[00:55:04]the way that one might think it does. And as a result of that, they're choosing to build that as opposed to using the neutral host model. No, our thought is twofold. One is that I think over time as they look to go into more and more places and also you'll have to spend more and more capital, I think that they'll increasingly, be open to using that third party model.
[00:55:25] So I think part of this is just a function of scout. And right now they're doing it at a small and a scout. They can't afford to do it, and are doing it to one city cities. And by the way, in some cases where they already have their own physical. A network to build off of. the second thing I would say there too, is it also, being first to market and if you think about Verizon.
[00:55:45] In particular, Verizon has always been known as being the highest quality network provider, when it comes to wireless it's, why a lot of people would argue that they're the best service because of the network quality. But what's interesting is that when you think about 4g and again, that [00:56:00] phone oriented experience, the quality of the network between our Verizon and it 18 P and T mobile is probably the.
[00:56:06]the smallest it's ever been in terms of differentiation between the three effectively T-Mobile and at, and T have for the most part caught up. What's interesting about five G in all the edge computing is that it does require this whole new network, typologies new with physical infrastructure and requires a lot of fiber and then ultimately small cells.
[00:56:25] And part of what we think Verizon is doing is they're recreating that, first mover advantage that they used to have. when it came to 3g and then 4g in the beginning, and that is that they're building their own physical fiber and they're building their own small cells. And by doing it themselves, they obviously aren't, they don't have a desire to go and, share that was an at and T mobile so that when 18 T-Mobile put more focus on the same aspects, 18 team, T-Mobile going to have a lot more work in and it's gonna be a lot harder for them to replicate what.
[00:56:56] Verizon has established that if you think about it fries, it had just been [00:57:00] purely dependent on partners like a crown castle to build office. you could imagine the crown castle be very happy to go and sell that exact same infrastructure to an 18 team at T-Mobile. So I think part of this is Verizon again, early on trying to establish that first mover advantage.
[00:57:16] But again, once we get to a point of scale and we get to a point of demand or we hit that inflection, there's just so much. Opportunity, if you will, for Verizon or for an 18 tier T-Mobile, our sense is that they're going to become much more comfortable utilizing a third party or neutral host model, very similar to what they become accustomed to on the macro type side.
[00:57:36] Matt: Colby. This has been such a terrifically, fascinating conversation. I realized that we could probably go another two hours. Maybe I'll have you back on season. And to, thank you so much for spending time with us, for people that want to find you and Callen online, where could they go?
[00:57:54] Colby: Yeah, the website is callen.com and my email addresses is my first name.my last [00:58:00] name dot com and, I'm always available and happy to have conversations with people in ministry, just so I can learn
[00:58:08]Matt: That's terrific. Thank you Colby so much for joining us today. I really appreciate it.