Todd Murren - Bluebird Networks: An Undergound Data Centre の全文書き起こし

if you're into roller skate or if you're into uh if you're into uh uh uh different kinds of adventures, it's a really cool thing to be underground when they're blasting. 00:05:40.079 --> 00:05:43.120 Um and then if you're not, it's kind of a scary place to be. 00:05:43.680 --> 00:05:46.240 Yeah, and if it's not something you're into, it's probably not not on. 00:05:46.959 --> 00:05:49.600 If it's if you're not into that, you probably ought not to be. 00:05:49.839 --> 00:06:02.720 So um anyway, um started the design of a data center to connect to our network that that transporting bits back and forth, uh, underground facility. 00:06:03.199 --> 00:06:09.759 Underground facilities, quite honestly, where you know, ask yourself some really simple questions. 00:06:10.079 --> 00:06:14.879 Where do governments go when they want a highly secured facility? 00:06:15.199 --> 00:06:16.959 They go underground. 00:06:17.199 --> 00:06:33.600 Um, you know, I don't want to harp on the gloom and doom of things, but but EMF and radiation and plane crashes and all of these things are, for all intents and purposes, not even on the scope when you're underground. 00:06:33.759 --> 00:06:43.519 Now, it doesn't mean that it's a panacea of of easy stuff, because when you're underground, you've got other, I like good design challenges. 00:06:43.600 --> 00:06:58.399 So you've got other challenges as far as that intake for generators and mechanical systems and intake and exhaust and all that, but there's they are things that can be uh addressed, they can be engineered around. 00:06:58.560 --> 00:07:06.160 Now, like most things that you engineer, you can take the easier path, the wide road, or you can take the narrow path. 00:07:06.319 --> 00:07:30.560 Um, we chose uh to take the little more narrow path, the challenging one, and and that was all for very good purposes, and I and I I translate back to my experiences with an above ground facility because our generation and our fuel storage and our mechanical cooling systems are also underground with us. 00:07:31.439 --> 00:07:35.279 And you go, well, where else are you gonna put them when you're underground? 00:07:35.439 --> 00:07:44.319 Well, you can put them above ground, and it certainly is a lot easier, but that's a critical function to a data center. 00:07:44.480 --> 00:07:53.519 Uh, the generation, uh, power, cooling, if any any one of those components fail, game over. 00:07:54.079 --> 00:07:57.199 Um, so they can't or they shouldn't. 00:07:57.360 --> 00:08:05.439 Um, so the design, all of everything that it takes for us to live, so to speak, underground is with us underground. 00:08:05.680 --> 00:08:15.439 Um, and uh I I uh now it may or may not be a good picture to paint, but it it makes the point. 00:08:15.680 --> 00:08:22.399 It's like, okay, where are most of my critical organs for my body? 00:08:22.639 --> 00:08:26.000 They're under my rib cage, folks. 00:08:26.480 --> 00:08:27.600 That's for a reason. 00:08:27.680 --> 00:08:29.040 It's called protection. 00:08:29.680 --> 00:08:35.120 And I need every one of those organs to work properly for me to continue on with my life. 00:08:35.360 --> 00:08:37.519 Same thing for the data center. 00:08:38.159 --> 00:08:48.240 So now you get and you go, well, yeah, but you can have diversity and with with 10 gig and 100 gig connectivity, data centers can be anywhere. 00:08:48.480 --> 00:08:51.200 And that to a degree is true. 00:08:51.519 --> 00:09:02.960 I get back to the virtualization where most of our customers are operating in two locations, um, and and and that's a very appropriate way to operate, quite frankly. 00:09:03.200 --> 00:09:09.600 But then you get into the edge data centers and and where is your data center located? 00:09:09.759 --> 00:09:17.759 And and and and then this whole vulnerability thing kind of folds back into it a little bit, and I'll explain what I'm talking about. 00:09:18.399 --> 00:09:30.320 So Bluebird Network is primarily a Midwestern, North American, 48 contiguous state operations, just right through the heartland of the United States. 00:09:30.559 --> 00:09:34.399 Uh now, we have these things called tornadoes. 00:09:34.480 --> 00:09:46.960 And I was gonna I was gonna talk to you about weather because I thought that's obviously you know, we always talk about weather when we're talking about data centers, but it's not quite as extreme as the as the weather situations or challenges you might have. 00:09:47.279 --> 00:09:56.240 Yeah, well, you know, um the strange thing about weather is two strange things about weather. 00:09:56.480 --> 00:10:01.759 Weather is everywhere, and weather is not controllable. 00:10:02.000 --> 00:10:13.039 And now I, you know, so my background, the engineering background goes, well, is there any way you can engineer those things that you have no control over? 00:10:13.279 --> 00:10:14.960 Can you let risk mitigation? 00:10:15.120 --> 00:10:16.080 Can you lessen them? 00:10:16.159 --> 00:10:17.360 Can you get away from them? 00:10:17.440 --> 00:10:21.440 And well, getting away from them isn't gonna work because the weather is everywhere. 00:10:21.759 --> 00:10:35.759 So, you know, it's it, you know, if it's hurricanes, if it's tornadoes, if it's earthquakes, if it's flooding, if it's wildfires, if it's temperature extremes, those are all weather events. 00:10:36.159 --> 00:10:41.840 And and now I'll be the first to say I probably have a bias here. 00:10:42.799 --> 00:10:52.240 But some of those things I just mentioned aren't things that the Midwest has to be concerned with. 00:10:52.399 --> 00:10:59.440 That would be hurricanes, uh, Midwest, not really wildfires, those kinds of things. 00:10:59.600 --> 00:11:02.080 Midwest, you know, we have temperature extremes. 00:11:02.159 --> 00:11:06.159 We saw 10 below zero here a few months ago, which was kind of crazy. 00:11:06.399 --> 00:11:13.759 Um, but they are at the very best minimized tremendously. 00:11:14.080 -->

00:11:21.360 That day that it was 10 below zero here, not where I work, it was 62 degrees. 00:11:21.759 --> 00:11:25.759 Now, and and afterwards I'll I'll send you a video I've got. 00:11:25.840 --> 00:11:43.279 It was it was absolutely bizarre driving out of the underground mine because it's it's pulling in, we have to pull in so many cubic feet per minute of outside air into the mine, and the mine is naturally, depending on well, the mine is naturally about 60 degrees. 00:11:43.600 --> 00:11:58.879 So the cool thing is when it's 10 below zero outside, and you're pulling in this extremely cold air into this warm environment, poof, we have this cloud bank that is just incredible. 00:11:59.120 --> 00:12:00.879 And remind me, and I'll send you the video. 00:12:00.960 --> 00:12:05.440 It's just but anyway, I digress here a little bit. 00:12:06.080 --> 00:12:10.399 But our tornadoes, um, they're non-existent. 00:12:10.639 --> 00:12:19.679 Now, if we would have gone the simple path, the wide path, and put stuff above ground, I wouldn't be able to make that statement. 00:12:20.000 --> 00:12:22.159 Um, but they're underground with us. 00:12:22.240 --> 00:12:26.000 So tornadoes is a non-event, temperature extremes are a non-event. 00:12:26.240 --> 00:12:34.639 Um, we did look because flooding, I mean, when you talk to somebody about being underground, they, you know, two things come to mind. 00:12:34.879 --> 00:12:36.879 Whoa, what about flooding? 00:12:37.120 --> 00:12:38.960 And and what about earthquakes? 00:12:39.120 --> 00:12:41.039 Now I'll get to the earthquakes. 00:12:41.360 --> 00:12:45.120 But the flooding thing, it's called geography. 00:12:45.279 --> 00:12:46.639 You look at where you're at. 00:12:46.799 --> 00:12:52.960 We we happen to be in an area, a region that's called the Ozarks Plateau. 00:12:53.279 --> 00:12:55.840 There's also nothing I can do about the geography. 00:12:55.919 --> 00:12:59.200 So if there's nothing you can do about it, pick your geography. 00:12:59.440 --> 00:13:01.840 Plateau, basically tabletop. 00:13:02.080 --> 00:13:08.000 I'm taking a lot of liberties with the English language, but water does not stay on the tabletop. 00:13:08.240 --> 00:13:13.279 As we all know, when we've spilled that glass of water at dinner, it runs off the table and onto the floor. 00:13:13.440 --> 00:13:14.960 Well, that's what it does here. 00:13:15.120 --> 00:13:21.039 You know, it it it we get biblical amounts of rain above ground, but it runs away. 00:13:21.360 --> 00:13:29.120 Um, most of the natural flooding that occurs comes from bodies of water. 00:13:29.279 --> 00:13:33.200 So if you live on the coast and a hurricane, that's called the oceans. 00:13:33.360 --> 00:13:36.799 Here in the Midwest, it's typically rivers. 00:13:36.960 --> 00:13:43.440 Navigable rivers uh are your largest potentials for for flooding, obviously. 00:13:43.679 --> 00:13:48.159 Well, this uh getting back to that whole Ozarks plateau thing, guess what? 00:13:48.240 --> 00:13:52.399 We don't typically rivers, you don't find them on plateaus. 00:13:52.799 --> 00:14:01.759 So now some of this, you know, I would I don't want the reader or I don't want the listeners to go, wow, this dude is really smart. 00:14:01.919 --> 00:14:11.039 Some of this is just blind luck that we we after the fact have gone, wow, this is a really good place we've selected here. 00:14:11.360 --> 00:14:13.759 So now let's talk about the earthquakes. 00:14:13.840 --> 00:14:22.960 Uh, because that clearly, um, being 85 feet underground, um, that has some great potential to be disastrous. 00:14:23.120 --> 00:14:28.480 Um, there's a couple of things about earthquakes, actually, about three things about earthquakes. 00:14:28.720 --> 00:14:32.240 Number one, now let me qualify this. 00:14:32.559 --> 00:14:38.559 Natural earthquakes, have we proven that we can produce our own earthquakes by fracking? 00:14:38.799 --> 00:14:39.200 Yes. 00:14:39.360 --> 00:14:43.919 Now they're not, you know, they're not eight-point on the Richter scale. 00:14:44.080 --> 00:14:48.080 Richter's a little dated, but um, you know, they're they're generally smaller. 00:14:48.320 --> 00:14:57.039 But earthquakes occur in specific areas, either tectonic plates or faults. 00:14:57.360 --> 00:15:01.919 Well, in this particular region, we don't have either of those. 00:15:02.080 --> 00:15:04.240 Now, a couple of hundred, yeah. 00:15:04.320 --> 00:15:06.159 Let's see, I want to be as accurate as I can. 00:15:06.240 --> 00:15:20.559 I'm I'm going to say a couple of hundred miles away, we do have uh the Numadrid fault, and we do have some tectonic plates over in eastern Kansas that not a lot of people know about. 00:15:20.879 --> 00:15:33.360 They're not particularly active, but nevertheless, so the earthquake thing is minimized, and you may go, well, minimize is great, but what about just getting rid of that? 00:15:33.600 --> 00:15:36.879 Then then you take a look at the rock structure. 00:15:37.120 --> 00:15:41.840 Um, all rocks are not created equal. 00:15:42.080 --> 00:15:45.039 And you may go, I don't really follow this guy. 00:15:45.120 --> 00:15:46.559 What's he talking about? 00:15:46.879 --> 00:15:52.720 There are, we happen to be in a limestone rock, um, and there are different. 00:15:52.799 --> 00:15:58.080 There's dolomite and there's granite, and and there's a whole, you know, there's more rocks than I can even imagine. 00:15:58.320 --> 00:16:06.320 But but rocks have uh dissimilar characteristics as far as their abilities. 00:16:06.480 --> 00:16:11.440 Um, some rocks, when you hit them, they just crack and they just break wide open. 00:16:11.759 --> 00:16:18.879 Limestone, well, cracked rocks, that has a lot to do with how they're formed. 00:16:19.039 --> 00:16:22.320 And no, I'm not talking about how they were formed 50 years ago. 00:16:22.480 --> 00:16:24.799 We're talking millions of years ago. 00:16:25.120 --> 00:16:29.440 We live within a

what's called Burlington limestone. 00:16:29.679 --> 00:16:33.279 It's approximately 250 to 300 million years ago. 00:16:33.440 --> 00:16:39.200 Now, at this point, the listeners are going, I thought this thing was about data centers, not geography and rocks. 00:16:39.360 --> 00:16:39.919 I'm getting there. 00:16:40.000 --> 00:16:40.480 I'm getting there. 00:16:40.720 --> 00:16:42.559 Every day, every day's a school day, as I say. 00:16:42.879 --> 00:16:44.080 Yeah, there you there you go. 00:16:44.240 --> 00:16:45.840 Even even Fridays, right? 00:16:46.080 --> 00:16:46.320 Exactly. 00:16:48.240 --> 00:16:54.559 Um, so so anyway, limestone is is compressed ocean bed. 00:16:55.360 --> 00:16:58.960 So it takes many, many millions of years. 00:16:59.120 --> 00:17:05.519 And the iron the ironics of any of your any of your listeners go, you know what, I'm not believing this guy. 00:17:05.839 --> 00:17:06.799 Come and see me. 00:17:06.960 --> 00:17:08.720 I'm 85 feet underground. 00:17:08.960 --> 00:17:19.119 I can show you a fossil of a shell just outside of our conference room in the solid rock to prove that that this, what I say, what I speak is true. 00:17:19.279 --> 00:17:21.039 But anyway, getting back to the point. 00:17:21.119 --> 00:17:23.440 So limestone grows in layers. 00:17:23.759 --> 00:17:29.920 And you can actually, when you see the the profile cut where they've done the mining, you can actually see the layers. 00:17:30.240 --> 00:17:32.960 Okay, so here's our next lesson. 00:17:33.119 --> 00:17:34.400 It's about lumber. 00:17:34.559 --> 00:17:38.079 It's about two by fours versus a sheet of plywood. 00:17:38.240 --> 00:17:44.880 Plywood made up of very different layers laid in different directions of wood grain. 00:17:45.599 --> 00:17:48.319 I'm fairly small statured. 00:17:48.480 --> 00:17:51.759 I can, with the right kind of leverage, I can break a two by four. 00:17:51.839 --> 00:17:52.720 It's fairly easy. 00:17:52.880 --> 00:17:55.039 It's one wood, it's one grain, boo. 00:17:55.359 --> 00:18:00.160 But I have a real problem with breaking a sheet of plywood. 00:18:00.799 --> 00:18:04.079 Limestone rock is the same thing. 00:18:04.240 --> 00:18:07.599 It's this layered from a from a mine standpoint. 00:18:07.759 --> 00:18:19.920 A limestone mine has some resiliencies that a dolomite or a granite mine, you know, typically granites for minerals and things like that, doesn't possess. 00:18:20.079 --> 00:18:37.039 So it's got this rigidity during, remember, we were talking about earthquakes, during an earthquake, you've got this rigidity so that if this limestone, Burlington limestone shelf were to move, it would move in its entirety. 00:18:37.279 --> 00:18:45.359 So the picture I can paint is when you're inside of your car going down the road 70 miles an hour or kilometers, whichever the case may be. 00:18:45.519 --> 00:18:49.839 Um, you know, everything around you is is fine, but you're you are moving. 00:18:49.920 --> 00:19:05.119 Now, the other fact about earthquakes, the damage that earthquakes do typically they manifest themselves at the surface of the earth and they will liquefy, if you will, the soil right at the earth's crust. 00:19:05.200 --> 00:19:26.960 And that's where we get into challenges with buildings, because we tie them into the bedrock or the rock structure, but the surface of the earth liquefies, and you get this damaging effect as earth shifts right and building shifts left, and that's how you get cracks and collapse, heaven forbid, collapses and all of that. 00:19:27.279 --> 00:19:29.200 I say all of that for this. 00:19:30.160 --> 00:19:32.559 I don't have to worry about that down here. 00:19:33.039 --> 00:19:36.640 It's just that risk mitigation mentality. 00:19:36.799 --> 00:19:43.200 Now, most data center customers go, wow, I haven't really thought about that to that degree. 00:19:44.880 --> 00:19:47.839 And I I'll admit that is an extreme. 00:19:48.160 --> 00:19:53.519 But our customers that are down here, we're a relatively small data center underneath. 00:19:53.599 --> 00:19:56.000 We're about 80,000 square feet. 00:19:56.240 --> 00:19:59.839 Now, the mine is huge, it's 5 million square feet. 00:19:59.920 --> 00:20:02.160 So we've got I was going to say that as well. 00:20:02.240 --> 00:20:06.160 And obviously, because you've just had that quite a big expansion project, haven't you? 00:20:06.240 --> 00:20:09.359 Where you've added, was it 11 just over 11,000 square feet? 00:20:09.599 --> 00:20:09.759 Yes. 00:20:10.000 --> 00:20:17.440 I was I was quite interested in the the complexities of of that project of expanding an underground data center. 00:20:17.519 --> 00:20:22.000 I I genuinely don't know, is it is it easy or is it very, very difficult? 00:20:22.160 --> 00:20:23.440 Or a bit in the middle. 00:20:23.759 --> 00:20:26.559 Okay, so so the obvious, uh, it's both. 00:20:26.640 --> 00:20:27.519 How does that sound? 00:20:27.599 --> 00:20:28.240 It depends. 00:20:28.480 --> 00:20:30.079 That's exactly what I expected. 00:20:30.400 --> 00:20:31.039 But it depends. 00:20:31.279 --> 00:20:37.119 No, uh, from an easy standpoint, my floors, walls, and ceilings are already constructed. 00:20:37.599 --> 00:20:39.359 So that's the easy part. 00:20:39.680 --> 00:20:47.440 And, you know, they're they're anywhere from oh, my floors, wall, you know, they're anywhere from 85 to 50 feet thick. 00:20:47.680 --> 00:20:53.200 So it's like from a structural standpoint, you go, cool, this is that's the easy part. 00:20:53.440 --> 00:21:08.000 Now the challenging part is all of the irregular surfaces that you have to accommodate, and and then you get into the the mechanicals where's air intake, where's exhaust, and all of those things. 00:21:08.240 --> 00:21:12.480 Um, we have uh, which is a little unusual. 00:21:12.559 --> 00:21:18.799 Um, being that the underground is always a consistent 62 to 64 degrees. 00:21:19.359 --> 00:21:24.079 I have more cool air than

I can possibly use. 00:21:24.799 --> 00:21:26.160 So we leverage that. 00:21:26.319 --> 00:21:29.039 We use that cool air for our intake. 00:21:29.279 --> 00:21:34.720 So we're drawing in from that 5 million square feet all of the time, 24-7. 00:21:35.359 --> 00:21:37.599 Now, our exhaust. 00:21:37.680 --> 00:21:40.000 Let's talk about that hot, humid air. 00:21:40.079 --> 00:21:43.359 When when our generators are not running, it's hot and humid. 00:21:44.079 --> 00:21:48.880 When our generators are running, you've got a little diesel smoke mixed in with that. 00:21:49.200 --> 00:21:53.359 But we exhaust that, if that's the right term, to the surface. 00:21:53.519 --> 00:21:59.759 It it's we have a 13 foot, if you will, vent pipe to the surface so that that. 00:22:00.079 --> 00:22:08.720 We don't spoil the environment that is everything we need to expand into that next 125,000 square feet. 00:22:08.880 --> 00:22:16.319 So we've got this constant flume of warm, humid air off of our mechanical systems blowing up. 00:22:16.559 --> 00:22:24.559 The other thing, which is beginning to be a big deal in the industry, you talk about PUE, power usage efficiency. 00:22:24.799 --> 00:22:30.319 And it clearly is an indicator that everybody looking at co-a-space should ask that question. 00:22:30.480 --> 00:22:31.599 What is your PUE? 00:22:31.920 --> 00:22:40.160 You know, PUEs, of course, there's the extremes, but quite frankly, you should be a PUE of you know one point. 00:22:41.039 --> 00:22:56.319 Uh personally, these PUEs that hover around 1.0, I'm not sure I'm buying that, quite honestly, but I think a really, really, really great data center is probably a PUE of 1.3, 2 to 2. 00:22:56.480 --> 00:22:58.000 I think that's in the margin. 00:22:58.160 --> 00:23:00.240 That's a question you need to ask. 00:23:00.640 --> 00:23:11.920 But the water usage, I don't know if you if you've done any of this, if you've looked at any, but there's another term coming that's that's catching, beginning to catch a little more attention. 00:23:12.079 --> 00:23:16.960 It's called W UE, water usage efficiency. 00:23:17.119 --> 00:23:18.880 And no, I didn't just make that up. 00:23:19.039 --> 00:23:19.920 That is really true. 00:23:20.000 --> 00:23:21.359 You can Google that. 00:23:21.759 --> 00:23:31.039 Um you can have an incredible PUE if you use water to chill with. 00:23:31.200 --> 00:23:37.359 So you're you're you're taking water in, you're using this water that most water is potable water. 00:23:37.599 --> 00:23:44.160 It's already been processed to where you and I can drink it, and it's not going to cause us any any internal problems. 00:23:44.480 --> 00:23:49.599 But you use it for its temperature and then you dump it out into the storm sewer system. 00:23:50.000 --> 00:23:52.640 You can't put it back into the water table. 00:23:52.960 --> 00:24:01.839 So data centers with low PUEs will likely have, if they're even calculating a WUE, will be huge. 00:24:02.000 --> 00:24:07.759 They will use potentially millions of gallons of water a day. 00:24:08.079 --> 00:24:12.799 And you've seen it a lot in the news lately as well with some of the hyperscalers, haven't you, about water usage. 00:24:12.880 --> 00:24:15.039 It's suddenly big story news. 00:24:15.119 --> 00:24:16.559 Yeah, everyone's interested in it. 00:24:16.799 --> 00:24:23.200 Well, in in our country, you know, in our nation, you go to the southwest, you go to the Arizona and the California. 00:24:23.440 --> 00:24:24.559 That's a big deal. 00:24:24.799 --> 00:24:30.799 And and you look at it, it's like I can go for three or four days without food. 00:24:30.960 --> 00:24:35.920 Now, I may not be particularly happy about it, but I could probably use the weight loss, but that's okay. 00:24:36.000 --> 00:24:37.279 That's neither here nor there. 00:24:37.519 --> 00:24:39.519 Water, different story. 00:24:39.680 --> 00:24:42.079 You know, in a day, I'm beginning. 00:24:42.240 --> 00:24:45.359 So water is a big deal, and it's getting to be a bigger deal. 00:24:45.599 --> 00:24:49.839 This whole closed loop system here, it's it's a closed loop system. 00:24:50.240 --> 00:24:52.640 I don't use millions of gallons. 00:24:52.799 --> 00:25:01.279 I calculated our WE last year, and I can't quote the number, but it was it was to the right of the decimal point. 00:25:01.599 --> 00:25:06.480 Last year we used um, wait a minute, I almost misspoke. 00:25:06.559 --> 00:25:08.160 My last year is in 2021. 00:25:08.319 --> 00:25:09.920 I looked at it in 2019. 00:25:10.160 --> 00:25:13.279 We used 3,000 gallons, and that was it. 00:25:13.920 --> 00:25:30.720 Um, so that whole using a using a uh water, using water, a large amount of water to get a tremendous PUE is just I think that's going to change, quite honestly. 00:25:30.880 --> 00:25:35.599 So that's where another advantage to underground comes in, comes into play. 00:25:35.920 --> 00:25:39.200 Yeah, what about one thing I also wanted to ask about your connectivity? 00:25:39.359 --> 00:25:49.839 So obviously you've you've explained where you are, but connected, where are you connected to and you know how where do your customers, where are they placed, and and what are you offering from that perspective? 00:25:50.160 --> 00:25:52.640 Um that, sir, is a great question. 00:25:52.799 --> 00:25:54.000 Thank you for asking that. 00:25:54.160 --> 00:26:04.160 Connectivity, I mean, just to be quite honest with you, customers that move into a data co-location center, they need to have choice. 00:26:04.319 --> 00:26:06.480 It's called being carrier neutral. 00:26:06.799 --> 00:26:13.039 Would I love for everybody down here to simply buy Bluebird Colo and Bluebird Network? 00:26:13.200 --> 00:26:14.640 I absolutely would. 00:26:14.799 --> 00:26:16.799 That's I have a bias there. 00:26:16.960 --> 00:26:20.240 But the fact of the matter is, folks need choice. 00:26:20.400 --> 00:26:24.480 And and for diversity purposes, they need

to be dealing with carriers. 00:26:24.640 --> 00:26:35.920 So we have down it at this data center, and I'm gonna qualify that because we've just recently, over the last six months, acquired another data center. 00:26:36.000 --> 00:26:40.880 So I gotta give fair, I gotta give to my other data center, I gotta give fair play here. 00:26:41.039 --> 00:26:43.519 We have multiple carriers in both. 00:26:43.599 --> 00:26:49.440 So you can make that decision for connectivity in and out, diversity, different paths. 00:26:49.680 --> 00:26:52.559 Um, and that's a really important thing. 00:26:52.799 --> 00:26:54.240 It's a very important thing. 00:26:54.559 --> 00:26:54.880 Definitely. 00:26:54.960 --> 00:27:00.480 And what with regards to being an underground facility, obviously that's you know, it's unique. 00:27:00.960 --> 00:27:10.000 What are the main sort of selling points to your customers of why they should go with an underground facility as opposed to an above-ground facility? 00:27:10.319 --> 00:27:17.200 Well, see, now I've got to be very careful with how I answer that because our other data centers and above-ground facility. 00:27:17.440 --> 00:27:19.839 So they both have their positives. 00:27:20.160 --> 00:27:27.519 So, well, and there's reasons for, you know, as most things in life, there's reasons for why things happen. 00:27:28.000 --> 00:27:32.000 We specifically looked for an above-ground facility. 00:27:32.160 --> 00:27:33.519 And I've had that question before. 00:27:33.599 --> 00:27:39.519 It's like, well, wait a minute, I'm I'm just buying this whole underground thing, and you just why did you do that? 00:27:39.759 --> 00:27:41.039 Because you need both. 00:27:41.200 --> 00:27:42.240 It's like a car. 00:27:42.400 --> 00:27:44.960 For it to be effective, I've got to have four tires. 00:27:45.200 --> 00:27:47.279 Two is not gonna do me much good. 00:27:47.680 --> 00:28:01.839 Um and I get back to that, to that virtualization and living in two different places, and and and our customers, you know, if they're using both data centers, because there's advantages and disadvantages to both. 00:28:02.000 --> 00:28:11.519 You've got some you've got some operational benefits to being underground, but you've got some above-ground benefits, plus they're geographically separated. 00:28:12.079 --> 00:28:31.119 And I don't want to make this too much of a bluebird commercial, but the fact that that these data centers are owned by a network company, I see that as a real cost advantage to our customers because transport between the two data centers is included. 00:28:31.440 --> 00:28:38.079 We want you to be in both places, um, because it's best for you and it's it's best for us. 00:28:38.400 --> 00:28:42.000 Um but there are there are subtle differences. 00:28:42.160 --> 00:28:48.319 You know, I can the the most obvious um cost power costs. 00:28:48.640 --> 00:28:57.680 You know, I'm in two separate states and two different electric grids, uh, and the cost structures are different. 00:28:57.759 --> 00:29:01.519 You know, our charges per kilowatt hour or KW are different. 00:29:01.680 --> 00:29:05.599 So there's some advantages there because companies can operate. 00:29:06.160 --> 00:29:10.559 Um we have some time of use rates difference at our above-ground data centers. 00:29:10.720 --> 00:29:19.279 And if you're operating in a virtual, you shift your load and operate where you cost efficiencies, but you've got your disaster backup underground. 00:29:19.519 --> 00:29:24.559 Um, it saves people money, it saves companies money. 00:29:24.799 --> 00:29:25.359 Yeah, exactly. 00:29:25.519 --> 00:29:27.839 And that's that's the main driver, I think, isn't it? 00:29:27.920 --> 00:29:29.359 For most people, it absolutely is. 00:29:29.519 --> 00:29:39.279 I mean, all of these other things that I've spent what, 34 minutes talking about are nice to have, but at the end of the day, we're accountable to somebody else. 00:29:39.359 --> 00:29:45.680 And it usually boils down to what's this going to man, I love everything I've heard. 00:29:45.839 --> 00:29:47.279 What's it going to cost us? 00:29:47.440 --> 00:29:53.759 And that's in a trail, and that's I I think we have some very strong uh positions there as well. 00:29:54.079 --> 00:29:54.720 Yeah, definitely. 00:29:54.880 --> 00:30:00.559 And the obviously the expansion that you've recently conducted was, I guess, as a result of increased demand. 00:30:01.519 --> 00:30:03.519 So is that something over? 00:30:03.839 --> 00:30:11.759 I mean, obviously, all of us in the industry have seen a significant increase in demand in the last 12 months, but is that the the main driver at the moment for your growth? 00:30:12.079 --> 00:30:17.359 Um the yes, the the main driver for our growth. 00:30:17.680 --> 00:30:23.279 The simple answer is yes, but I'd like to, I'd like to, if you will, chase a rabbit for a moment. 00:30:23.519 --> 00:30:45.920 You know, there was a period of time where as, and and I'm gonna date myself a little bit here, but but when IT, when the when a business's IT organization was essentially breaking apart, um, and and businesses were looking at this and going, the cost associated with my IT staff and my business is just it's a cost of business. 00:30:46.079 --> 00:30:50.079 There's no other way I can frame it, and it's rapidly increasing. 00:30:50.240 --> 00:30:51.440 How can we change this? 00:30:51.599 --> 00:31:04.400 So you get into the hosting and service provider and cloud, and and so for a period of time, you saw the traditional co-location market shrink. 00:31:04.640 --> 00:31:17.359 When I say traditional, the you know, the the business with their own enterprise IT coming to you and wanting space to, and if more times than not, it was disaster recovery, business continuation. 00:31:17.519 --> 00:31:19.440 Sometimes it would be operating. 00:31:19.759 --> 00:31:21.359 But you saw that shrink. 00:31:21.599 --> 00:31:33.599 And as as the

market has kind of levelized, and you've got we've got managed service providers that will do the sheet metal, will host the application for you. 00:31:33.920 --> 00:31:43.359 You've got, and we're beginning to see this, where you've got artificial intelligence, you've got these customers that are running GPUs, which is a very interesting thing. 00:31:43.599 --> 00:31:59.519 GPUs, graphics processing units, which are typically found in your gaming machines, but and and you know, but possibly for some of our readers, these things consume massive amounts of power, and they also produce tremendous amounts of heat. 00:31:59.599 --> 00:32:05.359 It's not uncommon to see on the board of a GPU temperatures in excess of 100 degrees. 00:32:05.680 --> 00:32:11.119 Um, and that can that can have a substantial impact on data centers. 00:32:11.359 --> 00:32:14.400 You've got to have some flexibility to be able to accommodate. 00:32:14.480 --> 00:32:25.680 But now you're beginning to see with cloud, because quite on, well, I don't want to be a commercial for Amazon or Azure, but storage in the cloud can be very, very competitive. 00:32:25.839 --> 00:32:31.279 We can't compete with that, but we need to have that connectivity, that private line connectivity to them. 00:32:31.759 --> 00:32:41.519 So this market, if you will, is maturing, and and I'm and I'm we are seeing co-location really coming back strong. 00:32:41.599 --> 00:33:01.119 And for this reason, I don't care who you are, you can be on the IT now, with the exception of Google as your da-da-da, you know, you if you can be an IT enterprise or you can be a managed service provider, or you can be a cloud operator, you don't want to, you don't want to get into this data center thing. 00:33:01.279 --> 00:33:14.480 You know, that whole rock and humidity and and tornadoes, and let the experts, and I yeah, I've got a bias there, let us manage all of those Maslow level one kinds of things for you. 00:33:14.720 --> 00:33:16.160 You know your business. 00:33:16.400 --> 00:33:27.200 Your business, it needs to be in places like data, co-location data centers, whether they be above ground or below ground, you know, that's a business decision. 00:33:27.440 --> 00:33:30.880 But so we're seeing this this influx. 00:33:31.119 --> 00:33:39.359 It's not what it was seven years ago, and it's not what it was three years ago, but it's it's really coming back quite strong. 00:33:39.519 --> 00:33:48.319 And that's yeah, we're we're seeing a huge demand for new build co-location facilities at the moment across the world, all regions. 00:33:48.720 --> 00:34:07.599 And then and then you mix in this whole edge data center because as as we cash more, and you know, it depends on who your customers, you know, if you're if your ultimate, if a business ultimate customer is is uh, you know, is Netflix watchers, then you got to get that cash close to where those are at. 00:34:07.680 --> 00:34:13.199 So there's got this next level of consideration as far as placement of data centers. 00:34:13.360 --> 00:34:17.679 And and I think you'll, you know, you'll now some would get this. 00:34:17.840 --> 00:34:22.639 If you get into, you know, how I got started, I love taking things apart and I love electronics. 00:34:22.719 --> 00:34:36.000 So if you'll visualize taking your computer apart and putting different pieces of your computer in different places, put that hard drive storage closer to where your users are so they can access it. 00:34:36.159 --> 00:34:46.079 Now, your processor has got to be fairly close to that, but your I/O, your screens, and your keyboards can be back at your office, or maybe everybody's working from home. 00:34:46.400 --> 00:35:01.679 Some may not get that visualization, but I really think you're seeing the whole data center, because we're having customers come to us and go, I need a place for my storage, or I need a place in the case of artificial intelligence, I just need a place to put my processors. 00:35:02.159 --> 00:35:05.199 Yeah, it's it's um it's a wild ride. 00:35:05.840 --> 00:35:06.559 It certainly is. 00:35:06.639 --> 00:35:07.920 It's a it's a crazy industry. 00:35:08.079 --> 00:35:15.920 A lot of we were saying before we went online, it's a very, very busy industry at the moment, and we're all equally as as busy as each other, I think. 00:35:16.239 --> 00:35:34.400 Yeah, you know, and I and I would be remiss if I didn't mention the people side to this, because I can I can have a tool, I can have a really cool bag of tools, but if I don't know how to use those tools or I don't have the people that know how to use those tools, then it's just a bag of tools. 00:35:34.559 --> 00:35:35.199 That's all it is. 00:35:35.280 --> 00:35:36.400 You can't do anything. 00:35:36.719 --> 00:35:55.760 So the the team that you have supporting, like our network technicians in and out of here, and our data center technicians, and just managing that whole installation changes, uh, because businesses are always out of time. 00:35:56.239 --> 00:36:00.960 When the decision made, it's made, hey, we got to get this done by next Tuesday. 00:36:01.199 --> 00:36:06.800 And if you don't have a good team, um which next Tuesday? 00:36:07.039 --> 00:36:18.880 And so team is a big deal, and and uh and now we're getting to your podcast where it's like we need people, we need more exactly we need more people. 00:36:19.360 --> 00:36:26.000 Again, like we were saying earlier, you know, the whole idea of this is to increase the exposure to attract more people, and that's why we're talking. 00:36:26.320 --> 00:36:26.639 Exactly. 00:36:26.719 --> 00:36:36.320 And if data centers are doing their business, we want to be, we don't want to be in the headlines because being in the headlines about something bad happening, no, no, no. 00:36:36.480 --> 00:36:41.280 So it's so easy, especially an underground data center that's not visible anyway. 00:36:41.440 --> 00:36:51.360 Yeah, I mean,

you visit visibility of data centers is is I'm not gonna say non-existent, but you really have to be looking. 00:36:51.599 --> 00:36:54.880 And there it's it's a good career. 00:36:55.039 --> 00:36:57.119 I've been in this thing for a long time. 00:36:57.280 --> 00:36:58.960 I go back to that point. 00:36:59.440 --> 00:37:07.599 That look at all the networks that exist and all these bits, these terabytes and whatever, whatever above terabytes, ziggabits, or whatever. 00:37:07.679 --> 00:37:08.559 We'll make something up, dude. 00:37:08.880 --> 00:37:09.440 Yeah, it doesn't matter. 00:37:09.519 --> 00:37:10.000 Say what you want. 00:37:10.239 --> 00:37:10.800 It doesn't matter. 00:37:10.880 --> 00:37:12.159 Yeah, it doesn't matter. 00:37:12.400 --> 00:37:15.599 You know, they're going from point A to point B. 00:37:16.000 --> 00:37:17.440 Where are those points? 00:37:17.679 --> 00:37:26.079 More times than not, they're they're at data centers, and they're probably at our cell phone or our mobile device. 00:37:26.239 --> 00:37:28.960 I mean, honestly, that's where it's that's where it's ending up. 00:37:29.280 --> 00:37:29.599 Exactly. 00:37:29.679 --> 00:37:32.320 And that's the easiest way to explain it to people, isn't it, these days? 00:37:32.480 --> 00:37:35.519 It's you know, if you use a mobile phone, you you should know what a data center is. 00:37:35.599 --> 00:37:36.719 And we all use a mobile phone. 00:37:36.880 --> 00:37:38.480 Well, not all, but pretty much all. 00:37:38.800 --> 00:37:46.400 Yeah, well, it's in and but anyway, so so the the getting, I say, you know, people. 00:37:46.559 --> 00:37:49.679 This can this this is not guys' work. 00:37:49.840 --> 00:37:52.400 This is this, it's anybody. 00:37:52.639 --> 00:38:02.880 We need knowledgeable people that understand and can think through the process, are challenged um with are challenged by things that are difficult to understand. 00:38:03.039 --> 00:38:03.920 That's a great thing. 00:38:04.159 --> 00:38:06.159 Just means you're working those brain cells. 00:38:06.800 --> 00:38:07.119 Exactly. 00:38:07.280 --> 00:38:12.159 And it it brings us on quite nicely, actually, to my last question that I ask everybody at the end of my podcast. 00:38:12.320 --> 00:38:15.599 You kind of already answered it, but we're gonna we're gonna ask it anyway. 00:38:15.760 --> 00:38:21.679 But if you could give one piece of advice to anybody that's looking to work in the data center industry, what would it be? 00:38:23.119 --> 00:38:24.719 You pretty much said it a minute ago. 00:38:24.880 --> 00:38:26.880 Definitely, definitely do it. 00:38:27.119 --> 00:38:35.679 You know, I um so so if I can offer offer the perspective from being in this industry for a really, really long time. 00:38:36.239 --> 00:38:50.239 Um we need we need people that can think for themselves, common sense, you know, you can you can always become more educated, you can always get the certifications that are necessary. 00:38:50.400 --> 00:38:55.920 When we when I go out looking for people, I'm looking for people that can think on demand. 00:38:56.079 --> 00:38:57.679 You don't have to know everything. 00:38:57.840 --> 00:38:58.880 I don't know everything. 00:38:58.960 --> 00:39:00.559 That's what Google's all about. 00:39:00.800 --> 00:39:05.280 You know, it but you've got to have and you've got to have the commitment. 00:39:05.360 --> 00:39:12.159 And it doesn't mean you know, showing up to work on time or those things, because we do a lot, we I do a lot of my work remotely. 00:39:12.320 --> 00:39:15.760 That that's the pandemic has really changed everything. 00:39:15.920 --> 00:39:21.199 Yeah, thinking on your feet, showing initiative, wanting to learn new things. 00:39:21.280 --> 00:39:30.320 Uh, and you're gonna have to search out data centers like through your podcast here to find out about them because they're typically, you know, we're kind of like bats. 00:39:30.480 --> 00:39:31.840 We only come out at night. 00:39:32.400 --> 00:39:41.679 No, but I think I also think, and why I was keen to talk to you, is the uniqueness of an underground data center is attractive to people, it's different, it's unique. 00:39:42.480 --> 00:39:47.119 People that want to get into the industry, you know, if they know there's facilities like this. 00:39:47.280 --> 00:39:52.880 I always say, you know, like the switch pyramid, the you know, the underground data centers like yourselves, like the one in Oslo. 00:39:53.280 --> 00:40:02.239 There's so many amazing facilities and buildings that people don't know about, and they're the type of thing that would trigger the attention of somebody looking to join the industry. 00:40:02.480 --> 00:40:06.800 We would we would love to entertain all of your listeners that want to, you know, want to consider it. 00:40:06.880 --> 00:40:07.679 Just look at it. 00:40:07.760 --> 00:40:09.280 You know, it's it's technology. 00:40:09.360 --> 00:40:12.000 There's so many different flavors of technology. 00:40:12.239 --> 00:40:15.599 Data center is one of them, but but it's a concentration. 00:40:15.760 --> 00:40:20.880 So you deal with networks and you deal with you know infrastructure and high availability. 00:40:20.960 --> 00:40:27.119 And, you know, if you happen to be that, well, I'm kind of a hands-on guy and I really like the mechanical or the electrical. 00:40:27.360 --> 00:40:30.239 We have got so much of that around here, it's not even funny. 00:40:30.320 --> 00:40:35.199 And that in and of itself is a science, it really, especially underground. 00:40:35.440 --> 00:40:48.079 Everything is controlled, variable frequency drives, because we don't want to run things just 100% because we can only run them to the degree that you want, because that lowers everybody's utility bill, ours and our customers. 00:40:48.159 --> 00:40:50.159 So it's just it's smart business. 00:40:50.320 --> 00:40:57.440 But uh, yeah, it's a cool place, literally and figuratively, and yes, pun intended, it's a cool place to work. 00:40:57.840 --> 00:40:59.679 What a perfect way to end the podcast. 00:40:59.760 --> 00:41:00.800 It's a cool place to work. 00:41:00.960 --> 00:41:02.639 I think really enjoyed that conversation. 00:41:02.719 --> 00:41:15.440 I think you're giving a great insight into what it's like in an underground data center, why we have an underground data center, and also the geology around it and the reasons for it, which is equally as beneficial, hopefully, to everyone that's listening to this podcast. 00:41:15.840 --> 00:41:16.480 I hope so. 00:41:16.639 --> 00:41:17.119 I hope so. 00:41:17.360 --> 00:41:18.079 So this is great. 00:41:18.320 --> 00:41:19.840 Thank you, Tom today, Todd. 00:41:20.000 --> 00:41:23.440 Obviously, I'm as I said earlier, I'll go off to the pub now at the end of my Friday. 00:41:23.599 --> 00:41:28.320 I hope your Friday is is a good one, and um, I'm sure we'll catch up again in the future. 00:41:28.559 --> 00:41:31.440 Well, I'm I'm not quite ready to go to pub, but it is on the list. 00:41:31.840 --> 00:41:32.480 Not along the list.