Not Your Father’s Data Center - What’s Happening with Cloud Computing in 2020?

Episode Date: February 18, 2020

Cloud computing appeared to experience a bit of a lull in the market in 2019. Raymond Hawkins, the host of Not Your Father’s Data Center and Chief Revenue Officer for Compass Datacenters, noted that the apparent 2019 lull came after a robust year in 2018. There are two primary questions that come from this observation. Was there actually a lull? And what does 2020 look like? Hawkins posed those questions to Ty Miller and Joe Goldsmith on a recent episode of Not Your Father’s Data Center Podcast. Goldsmith said that he thinks the market volatility is regional. First, he noted that he separates the corporate enterprise side of the business, because it is a bit more predictable. However, that is not the case for the cloud and service delivery portion of the market. “[For] the folks that are making money with (cloud computing), it has been incredibly volatile,” he said. “I think that some providers are building out infrastructure hoping for more business, and there are some regions wherein they are so far behind the capacity curve, they’re scrambling to catch up. “ Miller said that he doesn’t think there was a lull at all. He cited an example in Northern Virginia that may have looked like a lull. However, Ty noted that absorption was up from 2018, but was outpaced by supply. He also echoed comments by Goldsmith regarding the hybrid enterprise portion of the business. “Five years ago, you’d see a large enterprise deployment in the megawatt measurement range between 1 and 2, or even 5 megawatts,” Miller said. “Now, about 75% of that workload seems to be going to the cloud. Those are from entities who formerly procured as many as 5-times that amount.” He said this speaks to massive growth. For the latest news, videos, and podcasts in the Building Management Industry, be sure to subscribe to our industry publication.

Welcome to Not Your Father's Data Center podcast, brought to you by Compass Data Centers. We build for what's next. Now, here's your host, Raymond Hawkins. Thank you for joining us for this first episode of Not Your Father's Data Center podcast. Today, we'll be talking with my good friends, Joe Goldsmith, Chief Revenue Officer at NTT Data Centers, and Ty Miller, Chief Revenue Officer at Stack Infrastructure, and learning what they think about 2018's big year, 2019's reality, and what 2020 will look like. 2018 was a great absorption year for our industry. The data center business continues to grow like crazy. 2019, I think, left a lot of people feeling hollow, feeling a little

bit empty, feeling like we didn't have the absorption that we had. I think that that's not necessarily the case. It's really more to do with all the new players and new entrants. And love to talk about how we think 2020 is going to look and where we see trends in the industry going from a why is the business continuing to grow and where are the workloads coming from and going to. And we'll look forward to sharing that with our friends Joe and Ty. I think our industry looks around and goes, wow, 18 was an incredible year for absorption, an incredible year for the biggest cloud guys. And 19 was a bit of a lull and just wanted to say, A, I'm assuming you saw that. B, what'd you think? And what do you think 2020 looks like?

I think I get asked that question about every third day from my capital partners. I'm guessing both of you do. I'd just love to hear how you guys both think about that concept. Joe, if you want to hit lead off, that'd be great. Yeah, you know, we're seeing the same trends, I think, where we're seeing that corporate enterprise business sort saw that people weren't putting in huge colo installs, and they're not putting in their own tape arrays and big systems and stuff like that. They're really leveraging the cloud and the cloud services

and the service providers that are delivering their software products via the cloud. So the enterprise piece of it, I sort of put off to one side and say no surprises there. On the cloud and the service delivery side, the folks that are really making money with their compute, it has been so just incredibly volatile and dynamic. And I think what's happening is some of the providers are building out infrastructure, and they're not quite sure what kind of utilization and load they're going to get on it.

And so I think they're building out availability zones and regions where they're hoping for more business and they're finding it to be less absorbed than they expected. And then there's other regions where they're so far behind the capacity curve that they're just scrambling like crazy to catch up. So I'd like to say that, oh, maybe there was, maybe there was a lull, there was a dip, or there was a high point, but I think it's really regionalized and it's sort of by, call it AZ, call it region, but we're seeing incredible dynamic absorption in some regions. And then others are a little bit more predictive. So Joe, you talk about, you know, almost, for me, it feels almost like an arms race in the, I got to have AZ stood up or I've got to have AD stood up.

It's interesting. Do you see that or does it feel like an arms race? Because to your point, I think I see people going places without even deciding exactly what needs to be there. Just I know I have to be there. And I think maybe the most recent one would be Microsoft's Noel's announcement that, hey, we're going to be in Israel and watching everyone respond to that. Yeah, I think that people are recognizing that the only way you're going to capture market share, it's kind of like our business. If you and I, Ray or Ty, you and I, we wanted to go capture market share in Israel for data center, the only way you're doing it is to go there. Now, is there demand for that? I don't know all those pieces about Israel and some of these other markets, but the reality

is that those that are getting there first, they're winning. Those that are getting there first, just like in the data center business, they're winning business. They're following where the customers are wanting to buy services and then they're doing well now it might take longer than they thought in market X and it might go faster than they thought in market Y but the reality is it it is a bit of an arms race and that's not unlike our business too yeah I definitely feel like we're frankly arms dealers really providing the infrastructure for where this heads because I think if you look down the GDP rankings by country, that's where they're trying to go.

They're trying to support the markets in order. And, Joe, I think you make a great point. Who gets there first, that's who's winning. Because at the end of the day, these guys care about can you meet my timeline and can you meet my growth horizon long term. I'm shocked at how often those questions are asked. Are you in my market? Can you meet my timeline? And can you handle my growth horizon long-term? I'm shocked at how often those questions are asked. Are you in my market? Can you meet my timeline?

And can you handle my growth? And Ray, I'll just pile on and say that the notion of a well-built building, of quality parts, of mature operations, of solid security and SOPs and MOPs and all those things, that's sort of table stakes these days. And so the providers, the sophisticated folks, they expect that of us. And I think they expect

it of you and of Ty and everybody else. That's not the differentiator. I think the differentiator is speed to market, capacity management, capacity delivery. Joe, I think you're 100% right. I think our industry has matured to the point that, of course, you have security and you have solid operations and we know your building is supposed to work. That's a great way to say it. It's table stakes. Ty, I know I'm going to assume you might agree a good bit with what Joe had to say, but love to hear your thoughts on 18, 19 feeling like a l low, and what do you see in 2020?

Well, I think Joe made a good point that it's market specific. You know, there's a perception that you mentioned at the top that 2019 was a little bit of a low. I don't think that that's actually true. I think if you look at Northern Virginia and the Ashburn market as an example, it turns out at the end of the year that absorption was up over 2018. The challenge was that supply outpaced absorption. So we've had so much supply come in Northern Virginia, including sub markets like Manassas and Weisberg, that we've had quite a lot of choices for buyers. And the multi-tenant market has had some price softening. That doesn't mean that there weren't single tenant deals. You look at the amount of data center growth

from AWS in that region. Now they're self performer largely or almost exclusively, but it doesn't necessarily mean that the region was negative or that we had a low trend. There were markets inside of North America that had tremendous performance. Certainly look at a market like Hillsborough, Oregon, where Stack has a presence

and Joe and NTT data centers are coming into that market now. That's a market that historically had been, you know, an 8 to 10 megawatt type market. And we saw results last year that were significantly higher, more than double that in terms of absorption. So I think there are some workloads that are shifting, but really the macro trends are, simply put, just bigger and more and geographically distributed. So Ty, I like the point you make about Ashburn. Total absorption, I don't think to your point was net down year over year, but it was how people felt about it.

I think 18, end of 18, there was so much activity and so much absorption and so much new capital coming to our business that all that capital started asking hard questions of its management teams in 19. And that caused a lot of tough questions in 19 about activity. And to your point, Amazon doing a good bit of it on their own didn't make all those new capital players necessarily feel that great about it. I think your point's well made that Ashburn's not going anywhere. I do get asked that question occasionally. Hey, do you think Ashburn's going to not be a top market long term?

I don't think the trend line there can't project anything for forever, but I don't see anything slowing down in Northern Virginia. And to your point, we may get down to Prince William County and some other places, but that area is going to continue to explode, don't you think, Ty? I absolutely do. And you look at Joe's comment, which I completely agree with. The enterprise workloads are complete hybrid deployments. And whereas five years ago, you'd see a large enterprise deployment in the megawatt measurement range, one megawatt, two megawatts, even five megawatts, really three quarters of that workload

seems to be going to the cloud now for a hybrid architecture. And so you're seeing enterprise RFPs come through in the industry. And they're 250, they're 500 kilowatts, 750 kilowatts. And those are from entities that formerly had procured 3, 4, 5x that amount of capacity direct for their own hardware. And what does that speak to? It speaks to massive growth in the public cloud's providers, right? You're talking about Google's platform

was 53% up year over year, 19 versus 18. Cloud growth at Amazon, fourth quarter over third quarter was 34% year over year. And so, you know, Azure's 62% growth over the previous quarter. I mean, just incredible growth rates at the public clouds. Joe, you see that UCLA education paying off all the numbers with Ty? All that counting and stuff is crazy. Yeah, it's unbelievable. And he wasn't even on the crew team. I know. They must have taught arithmetic over there at UCLA. Yeah, they had all the art. In any event, my point is that there's this massive cloud growth,

right? It's the shift of workloads to the cloud. Does that mean death of the data center? Hardly, hardly. I mean, folks, unless you believe that the digitization of humankind is going to somehow reverse itself, it's going to be more compute, more storage, more network, and all that lives in a buildable. Yeah. Here, here, Ty. When I get asked, and I keep thinking of questions, so I apologize, guys, I'm going to bounce around. That's another one I get asked a lot, Ty. I think the misunderstanding of, hey, isn't the cloud going to damage the data center industry? It's just, it's just, we're the home for the data center industry. The data center industry is the home for the cloud. This is where it lives.

And although I think that we'll see Google Cloud and Azure and AWS continue to grow, I also think there's another round, as Joe referenced, the enterprise business waning. Not all your workload is going to end up at Azure GCP or AWS. There's going to be meaningful workload in other places, whether that workload is application specific like, you know, Salesforce or whether it's just an enterprise application on the back end with Oracle. I think you're going to continue to see workload shift. And I don't think we'll ever see the end of the enterprise data center, at least not in our three lifetimes, but that workload is going to continue to shift. And it's not all going to just three places, but it continues to move. Joe, can I ask just one question? I know I saw y'all's announcement at PTC. Do you mind giving

us a one minute commercial on Raging Wire to NTT data center? What's the right way to think about that? So the right way to think about it is we've taken this sort of disparate group of data center companies, eShelter in Europe, NetMagic in India, the legacy NTT comm facilities throughout Asia, including Japan, Raging Wire here in the US. And as they were bought and put under the NTT umbrella, there was no real integration with regard to commercial structure, operating structure. We ran the businesses independent of one another. Now, what we found, and it's our belief, that a customer is looking for ubiquity across a global platform. If you look at, say, the business Equinix has built over many, many, many years, they've done a great job in being able to deliver service to the same customer in Frankfurt and the same application in Singapore and in

Northern Virginia and in South America. And so customers really, they're looking for that, I won't call it ubiquitous because I'm not quite sure that there is such a thing in our data center business, but they're looking for more choices on a globalized basis. And if you look at some of the global and hyperscale customers we're talking to, they're not just interested in Raging Liar United States of America. They're interested in the global portfolio. And they also want to have consistency in contract terms. They want to have a solid view of what the inventory looks like and our speed to market. And so it's really taking regional companies, putting them under this umbrella and globalizing it.

And then on top of that, putting some of the legacy NTT products and services on it, like the NTT Global Network. So Joe, and because I'm a simple, stupid chief revenue officer, all I think of is sales and marketing. Does that mean that your team is going to end up having the ability to sell in all of those markets? Or is that all going to consolidate and all end up working for you? So it does mean that my team will be able to sell in all those markets. Certainly, I'm not responsible for what happens in Europe and capital decisions in India and

in Asia, et cetera. But when I have a customer here in the United States and they say, look, I don't want to figure out what's going on in Europe. I want you to tell me. I don't want to have to go deal with five different account teams. I want to deal with one account team. And if I say I need capacity in Singapore, I want to know how much you've got, when I can get it, how I can deploy it, how I can tip it up.

And so we're really simplifying the buying process for the customer by giving them points of contact so that they can have a global purview. Love that. That's super cool for you guys, Joe. I think that's exciting. Thank you. And this is not a dig at before, but I think having the NTT name is frankly exciting for you guys to be able to have that across the platform

instead of taking a jab at Raging Wire. But having the NTT name, I think, is a big positive. I hope that's good for you guys. And I appreciate that. We've gotten great, overwhelming, positive feedback from the customer base. And frankly, hearing it from my contemporaries, hearing it from you or Ty, and frankly, the fact that we compete head to head, I appreciate that. And that means a lot to me.

So thank you. I think it's good for you guys and exciting. Joe, I'd say for years, I've, I've viewed NTT globally as a sleeping giant. And well, I just hope you wake up slowly. Ty, just so you know, I've already got my resume in to be Joe's driver. So you're going to have to come up with another role, Ty.

Thank you. Thank you're going to have to come up with another roll of tie. Thank you. Thank you, Ray. I think the reality is, folks, if we pick our metros and our capital deployments carefully and in conjunction with our customers and we pay attention and we listen, I think there's enough business for all of us to do really, really well. Yeah, I think you hear, Joe, I think that theme enough business for all of us to do really, really well. Yeah, I think you hear here, Joe. I think that theme can continue throughout all of our conversations. Joe, you touched on something I'd love to ask you both to comment on it.

I think that there's two different ways that people think about our market or our business in general, because I'm going to ask a question about market. Where you need to be, we're in the real estate business. The market matters. But at the end of the day, I tend to believe that the relationship with the customers matter. And inside Compass, we talk about cloud 60. We think of the 60 biggest guys solving problems with technology. And so that's, you know, someone like AWS, but also someone like Uber, even though Uber is not really a cloud provider. I think relationships and starting at the customer matters first, and they'll tell you what marketplace to go into. How do you guys think about that?

And having access to multiple markets is how you take care of those customers. Do you think about it market first or account first, if you both would comment on that? Ty, why don't you take a hack at it? I have some ideas, and I'd certainly like to pile on, but I want to hear from Ty. Yeah, past. I'm just kidding. You know, our point of view as a smaller company in just our second year at Stack is that we want to be in the markets where we know clients have to be, right?

Major data center markets. We're in seven of the top ten. And that's purposeful. As I think about markets and I'll come back to relationship to the moment, Raymond, but I think about markets. I think a little bit about edge. And, you know, Joe, a decade ago when you and I would talk about edge, it was that was tier two and tier three cities, right? The conversation about edge has evolved to be how many points of compute or network can you have in a dense urban area, almost. One of the things I'm seeing, I'm sure you guys are too, is large wholesale and cloud workloads and compute centers moving into the urban core, not out to cornfields. Started at a macro level out in the hinterlands, but as 5G comes into play, eyeballs, it's all about low latency and eyeballs.

You're seeing all the major cloud workloads, whether it's media, whether it's compute, et cetera, it's coming into the dense urban core. And so location matters, I think, to a large degree, but you haven't seen wholesale workloads moving into the urban core. On relationship selling, Raymond, I think our clients take us to new markets.

Trusted relationships allow you to expand. We've seen that within each of our respective platforms and in some of our competitors. None of us, well, maybe Joe, but certainly you and I Raymond are not selling a thousand points of light around the globe like some others, right? But when you deliver on your promises and you treat your clients tremendously well,

they'll come back to you for solutions time and time again. And that helps our platforms grow in the places that perhaps aren't right on our radar today. But when you can go with an anchor, you can go with confidence to a new location. Yep. Love that answer, Ty. Joe, thoughts on markets versus, and not so much versus as in

they compete with each other, but which is a greater focus? Do I focus on developing in a market or do I focus developing in an account relationship? Well, I think the answer is probably both. And Ty really touched on it. If you're talking about the huge, huge, multi-hundred megawatt deployments that you see some of the cloud providers do, up in the DALs or in San Antonio, and I don't even need to name names.

We all know who they are and where they are. There's these huge deployments that they've created on their own, independent of any of the data center providers, et cetera. And those sort of, in my opinion, huge deployments that they've created on their own independent of any of the data center providers etc and those sort of in my opinion those call it top five amazon google microsoft maybe facebook which is a little more distributed but they're sort of uh in a separate bucket they're going to go do their own thing and they're going to do it at a scale that's just unprecedented but then they're going to fill in the gaps with providers like ours,

like Stack, like Compass, where they've got go-to-market strategy and they can't go build a huge campus, and it doesn't make sense to build a huge campus. So I put those, call it, top half dozen cloud providers over on one side, but then you start getting into, and you mentioned, Raymond, the top 60 or top 50, however many you kind of target in terms of that global service provider or whatever you might categorize them,

but let's use NVIDIA as an example. NVIDIA has a GPU as a cloud service. Well, if you don't understand it and you don't know what it's for and how it works, it's a very low likelihood that you're gonna be able to walk into NVIDIA and say, hey, we've got a great solution for you and a great data center in market X,

because the NVIDIA cloud product is gonna support all the autonomous vehicle traffic that's out on the roads, and those vehicles are gonna be hitting this cloud platform in a very distributed world. And so what makes sense for NVIDIA may not make sense for AWS. So I think the answer there is sort of the second half of your question, which is that's very much account and relationship centric and understanding what's that product do?

How does it work? How does a consumer buy it? How do they consume it? Yeah. So we've tap danced around some of the terms. So, so just, and I don't know that these become industry terms, but certainly stuff for us. So inside compass, we call the big five. I loved Fang, but I couldn't get there. So it's Facebook, Apple, Microsoft, Google, Amazon. So we have it. FAMGA are the big five. And then the

cloud 60, the next 60 biggest guys, um, where, where so much of the workload is getting driven. And I liked your NVIDIA reference, Joe. I mean, there's what they're doing and what their video drivers can do. Their GPU is just fascinating, and the amount of power it chews up is unbelievable. So I can think of all kinds of subjects, guys. I know, Ty, I want to be sensitive to your time. Things I'd love to talk with you about just like this, just have a chat. Ty, you talked about the edge.

You brought up latency. I also think that backhaul avoidance is a great subject. It might bore some people to death, but I think it'd be something the three of us could talk about and what the edge really means. Love to do that with you guys in the future. Guys, I just appreciate getting to chat with both of you and look forward to getting to do it again in the future. Guys, I just appreciate getting to chat with both of you and look forward to getting to do it again in the future. I will tell you as we're just getting this podcast off the ground, but as a thank you to our guests, we are offering a tour of some of our data centers free with Arlington's finest, Cameron Cooper, who's our rep here at Market

Scale. And we took some time to be careful that we wanted to find places you'd be interested in. So you can tour our facilities in Kandahar, Madagascar, or Wuhan anytime you guys ask. So just let me know when you guys are ready, and we'll get that set up. The staff will get your information. So, guys, thank you so much for joining us, both the great Ty Miller and the Ryan Gossing of the data center industry, Joe Goldsmith. Thank you, Ray. Thank you, Ty. Appreciate it, everybody. Bye, guys. Great catching up.

Thanks for the opportunity. Adios, amigos. We'll see you next time.