Revealing How We Evaluate New Data Center Providers for Our Portfolio
I'm Max Clark. Let's talk about evaluating new providers and how we add them to our portfolio and whether we choose to offer them to to our customers or not. This stems recently at a data center company or bare metal provider reach out and ask, you know, hey. What do we need to do to get your portfolio? We'd love to work with you, basically.
Speaker 1:Now there's only so many data centers on the planet. I mean, new ones are being built, but this isn't like a huge, you know, size. Right? Like, when you come into major markets let's just use Los Angeles for a little bit. Los Angeles as a data center market and I'm, by the way, I'm gonna include Orange County.
Speaker 1:And I mean, technically, it's not. I mean, Orange County and LA are completely different animals. When you look at the LA market for data centers, the center of the LA universe is downtown Los Angeles. And for colocation, really, that is 1 Wilshire 624 South Grand. By the way, this is a byproduct of deregulation and how, you know, select start and the birth of the Internet and all these different things.
Speaker 1:So some of this stuff is just geography and availability and, you know, hey. You got a building that's a couple blocks away from AT and T's wire center, a survey wire center, and it has an empty floor and it's cheap and nobody wants to be in it. And if you can get fiber to it really easily and boom, you've now you have a dominant carrier hotel for, you know, one of the densest interconnection points on the planet, and that is 1 Wilshire. So you look at these buildings. And so if one Wilshire is the center of the universe in downtown Los Angeles, everything else exists within a certain radius and proximity to this building.
Speaker 1:600 West 7th is a converted department store that was converted into data center space. And it was, you know, I think originally done by an entity that, of course, you know and basically this other consistency in telecom and data centers is when most of these were converted or built the original operator went bankrupt what's a good example that exodus right access built the most of the original data centers are still used today exodus goes under is acquired try to think if I can remember the exact chain of operations here so I think it's Exodus to cable and wireless to Savvis to CenturyLink which becomes Lumen which then gets spun off to 6Terra which then is bankrupt and then merged by private equity with Evoque. And by the way, technically, there's another transaction in there with another company called software. Now you're talking about a data center. And by the way, this is true for all these original data, you know, excess assets.
Speaker 1:So I don't care what market you're talking about. These facilities have phenomenally long useful lives as long as you're maintaining the plant. They might not be a 1000 watt per square foot density, but who cares in terms of what the owner and the operator of that data center is getting out of them? They're making money on these facilities. They're making a ton of money on these facilities.
Speaker 1:Let's be real. Back to Los Angeles. So Los Angeles is quirky because of 2 things. So first, you have the center of the data center universe is is one will ship because that's where the and network interconnection is. Then you get pressure based on cost for floor space, cost for building, cost for power, power availability, and then taxes and taxation overhead.
Speaker 1:Right? So most operators then have a decision to make where they're either converting and doing stuff in downtown LA. So in downtown LA, we have the 1 Wilshire building 624 South Grand. We have 600 West 7th Street converted department store. Then a lot of XO Communications operations were in this facility.
Speaker 1:Then it becomes Equinix is on the top floor. So 600 West 7th is really x you know, Equinix. We have level 3 build out their gateway at 818 West 7th. We have another major data center, Mike, you know, process happen, 1200 West 7th, which was a cash vault in a Pitney Bowes mail processing facility. It's already raised floor generators in the building.
Speaker 1:1200 West 7th is where, you know, Myspace, for instance, had the bulk of their operations. And then while this is going on and people are trying to build data center in Los Angeles, you have Irvine facilities coming online. So access builds their facilities. There's the original OC one facility, which is no longer around. You had OC 2 come on in Gillette.
Speaker 1:You know, XO has a facility Santa Ana that gets built El Segundo why El Segundo different power grid so Edison versus DWP cheaper building it's easier to acquire the actual building itself and probably one of the best tax bases for a business to operate in the LA metro versus Los Angeles City. Going to El Segundo is a great thing for you if you're actually looking at and thinking about Texas. So access builds a facility at 200 North Nash. We have Equinix comes down, 1920 Maple. You know, like, there's luster of data centers that all exist right around the same area.
Speaker 1:Also in Los Angeles, we can go a little bit deeper. We can talk about Burbank and facilities being built built on Burbank. Same thing. Right? Land, power, cost.
Speaker 1:Now the buildings rattling off are all what I would consider the pre minute locations. Metromedia fiber starts building a facility, which then becomes above net, which then goes through bankruptcy then is acquired by 3 65 main. And, this is, 225 El Segundo, 2265 Rosecrans. It's a little hazy. So again, you know, built bankrupt, bought rehab, built, maintain, like, these things just kind of trade.
Speaker 1:Anyways, okay. So these are all what I would say, like, the good buildings. Right? Good operators, good infrastructure. People have spent a lot of money.
Speaker 1:And then you get into buildings, which are the adjacencies. And I had a friend years ago referred to it as ghetto colo. So I will refer to these things forever for the rest of eternity as the ghetto colo locations. I won't name the buildings by numbers, but if you're around the markets, you'll understand which what I'm talking about. And these are buildings where it was, hey.
Speaker 1:We just happen to be right next door to one of these buildings, and we can run, you know, conduit at one window down across the mechanical floor up up through another window, get into the building. Now we have conduit access or we're you know, there's plenty of conduit in the street. We can just tap into it, and we can provide service, and we can offer a low cost data center option next door or a block away. And usually what you see with these buildings is small floor plates. They're not large buildings.
Speaker 1:They weren't designed to be data halls. They don't have generators or they have one generator for the entire building. Not a lot of power provided to the building, relatively low power provided to each suite. And operators that exist in this space are really trying to come to market with a cost per kilowatt that is so far under market for that metro that when they provide, you know, put their margin on it that when a customer looks at this location and says, hey, this building isn't ideal, but you know, it's so much cheaper. We're gonna go here instead.
Speaker 1:So dubbed 20 years ago to me as the ghetto call and this exists. I'm I'm picking on LA just because it was fresh in mind and we're just doing stuff out there. But this exists in most major metro markets this is true in New York it's true in New Jersey San Francisco Seattle Chicago Atlanta you know Miami go to a major market where you had heavy interconnection and heavy get us our deployment and you see this stuff now 10 minutes segue aside right we're really jump the shark on that one pretty good. You know, when we get an inbound inquiry of a, hey, we're, you know, we're offering you know, we're a bare metal company. We're a data center company.
Speaker 1:We're colocation. We have a private cloud. Whatever these whatever the thing is, we're gonna look at it and we're gonna say, okay. Do we know you? No.
Speaker 1:We don't know you. Why do we not know you? Okay. Are you operating your own facilities? Are you using somebody else's facilities?
Speaker 1:Okay. What facilities are you in? Are you in a, you know, tier 4 major data center or not? Right? And it's the or nots that are really interesting because usually the conversation is, you know, you get pitched this like or this phenomenal, like, top tier unbelievable best operator in the market.
Speaker 1:And then you look up the facility like, yeah. But you're in a total POS building that has a single generator that I know is blown up 3 times. And why would somebody do this? Right? And by the way, let me let me actually put this out here.
Speaker 1:Right? I'm not knocking on this in terms of a selection process. I've joked about this for a lot. I mean, jokes not right to work. There is a a massive price differential between going you know, when you talk about, like, uptime institute tiering.
Speaker 1:You know, if you wanna be in a tier 4 data center and the requirement to be in a tier 4 data center, I mean, it is expensive. It is expensive to build that facility. It is expensive to maintain that facility. It's expensive to have generators enough. I mean, everything about that facility is expensive.
Speaker 1:There is a reason why it's a tier 4 data center. Right? Like you're buying into that approach. If you have infrastructure that doesn't require to be in a tier 4 data center, like, absolutely don't put it in a tier 4 data center. If you're looking for a cheap place to go put your, you know, fill in the blank, whatever it is, you can do this thing a lot cheaper.
Speaker 1:I would get into a thought exercise with you of, like, how much redundancy do you actually even need at that point? You know, like, if you're not gonna be in a tier 4 data center, do you need diverse fiber entrances? Do you need diverse power entrances? Do you need diverse grids? Do you need a redundant generators?
Speaker 1:Do you need redundant UPS systems do you need redundant air conditioning like what do you need and what do you not need and what this all impacts into is a statistical probability of outage per year so on a given time frame how many seconds minutes hours of downtime can you accept with this facility and what are you trading in terms of cost that's all we're really getting into here is we're just talking about trading cost with a statistical probability of an outage over what duration of time and just because there's a statistical probability of an outage doesn't mean that the outage actually occurs. It could not at all. You could run forever and, like, everything's fine, or you could have more outage than what the statistical probability actually is. And when I say that, I mean it like, you know, and by the way, this weird black swan events happen, you know, tier 4 data centers. What's a crazy one?
Speaker 1:Drunk driver crashes into a utility pole feeding into a data center, and the fire department walks into the data center and says you have to power off your data center so we can cut the occupant of the car out of the vehicle. Right? And the data center is like I mean, it's already cut over to generators running their own thing, and they're like, what do we do? You know, they weren't a fake. They couldn't they you know, the authorities are now telling you you have to do this thing, period.
Speaker 1:Like, was that a statistical probability that you have to worry about, you know, that kind of event happening? Of course not. Like, it's relatively unusual. Right? Like, I can think of 1 or 2 stories like this.
Speaker 1:But, you know, stuff happens. I've seen utility substations blow up. You know, not not like a terrorism bomb blow up but like a something that mean literally explodes the transformer explodes and to the utilities credit they repaired that extremely fast but then you get into what is your actual outage duration that you can take is the 8 hour repair cycle for the utility to fix that substation acceptable to you or not and if it's not make plans for it on the other side lower Manhattan had significant weather event that went through and while a lot of buildings had generators on the upper floors they had diesel storage tanks and pumps underground And when it was flooded and their pumps couldn't work and their diesel storage tanks were underwater, you know, people had to run 5 gallon buckets. I mean, when I say run, like, we do not running, but, like, people physically carrying 5 gallon buckets of diesel fuel of flights stairs in order to keep the generators running to keep the facilities on. Does that fit your risk profile or not?
Speaker 1:Like, what would happen, you know, if that facility just turned off? Like, does it fit your risk profile? What rubs you the wrong way is this presentation of something that's not like if you're a service provider and you are a value low cost, we're coming to the market. We're gonna give, you know, our positioning is a product that is going to be 50% of the competitor because we're making these decisions or in order to capture and to attract a different, you know, segment of the market. Just lay that out upfront.
Speaker 1:I 100% respect and appreciate that. Just lay it out upfront. This is what we're doing and why we're doing it and what we're focused on. And it's fantastic. If you come to us and you say, hey, we're the best thing since sliced bread.
Speaker 1:And we do a little research on you and we find out that not, we're not gonna continue the conversation anymore because you started with a falsehood. You've started with a lie. And if you're lying about things that we can discover very easily, what are you not telling us that we have to dig into farther? And is it worth it? And again, for us, when we start talking about helping our customers through this process, the risk reward, there's no value at that point.
Speaker 1:There's no point where it's gonna be interesting or valuable to us to take that risk where we can't have an open conversation. If I can have an open conversation with a customer and say, hey. There's this product in this market that might be great for you. But by the way, there's these trade offs. But, you know, for these trade offs, you're gonna have a 50% reduction in your cost.
Speaker 1:Is that something that's interesting to you? And then they can make an informed decision. Like, that is a great conversation. Like, fantastic. Maybe they come back and they say, yes, we're good with this.
Speaker 1:Or maybe they come back and they say, no, we're not good with that because we've decided these other things. And it's, like, absolutely. Totally understand. Let's go a different direction. You know?
Speaker 1:And sometimes people maybe don't really know the answer to that until they're actually confronted with it. Like, you know, until you actually have to make the decision and get ready to sign, you know, like, actually contemplating signing a contract. You might think that you're in one bucket, and then you find out that you're just in another bucket. And that's okay too. So, you know, summarizing, wrapping all this back up and circling around, whatever you wanna call it, too long didn't read.
Speaker 1:The point of all this stuff is, especially the data center world, just be straight up. You know? And and when you're looking for data centers, you know, it's really easy to evaluate, look at these things, look up their history, you know, and and know what their operations are and figure out what's going on. You know, if you're talking to a data center and you find out that their entire facilities team is x Submariner nuclear, you know, engineers, like, they probably have the thickest book of SOPs that you've ever encountered in your life. And then the only question you have is how long has that facility been online, and have they encountered every edge case that you can possibly ever think of?
Speaker 1:And And then there's gonna be SOP for it, and they've got preventive maintenance. And you're great, and things are wonderful. I'm Max Clark. If you have any questions, comments, let me know below, and, we'll get back to you.
