Welcome And Guest Backstory

SPEAKER_01 0:02

Welcome to Uplink, where we explore the world of digital infrastructure, uncovering the technology fueling AI and cloud innovation with the leaders making it happen. I've been waiting a while to do this one. Finally got him. This is uh this is exciting. So uh yeah, I want you to introduce yourself and tell me what you think you do. And I've got a whole list of all the stuff that I know that you do as well that you haven't got written down. Uh what we've got here, ladies and gentlemen, is uh something I had to get printed, actually. Here it is. We're the king of t-shirts, and today we get to interview the godfather. I don't know if you can get a zoom in on that bad boy, but there it is. The godfather is in town. Tolney. Hey, Dolmy, tell me the godfather. Uh hey man, look, thanks for catching up. We're in your city. Uh, it's not often we get to get you down and actually chat like this. So tell me, what what would you describe you as right now in your role or your many roles?

SPEAKER_00 0:57

So we uh I've uh we launched a company called Beyond, which has been fantastic. It's a data center construction operations company.

SPEAKER_01 1:04

And that's what I want to drill into a bit after this as well. That sounds that's incredible. Yeah.

SPEAKER_00 1:08

So we started that. Uh oh, we've been working on it for a couple years. We launched uh earlier this year, but sort of got our we partner with WAFRA, which is the uh effectively the um U.S. arm of the Kuwaiti uh Social Security arm. Awesome. Um that's that's it for funding and stuff. That's for funding. Yeah, yeah. Yeah. Um these things are not cheap. These things are not cheap to build. Yeah. Uh this is my fifth deal with them. They've been a great partner. We did uh a car wash business together. We did it. We're gonna get into that too. Which is was incredibly successful. Uh we made you know six plus times our money. It was a great launch with uh Otto Lagerboss and and um and uh Ryan Brenner. Um, you know, with these guys literally driving car washes and what happened. Uh yeah, it's express wash. It's called and the best name, it had the best name was called L Car Wash. Um, so uh uh in Miami and two incredible entrepreneurs, uh Justin Landau and Jeff Karas, they ran it. They were fantastic guys, came up with the idea. So that they've been with you for a while. So you've car washes. So we did car washes. Uh the original investment was they invested in Tier Point. So they invested in Tier Point was on was on the executive team, and then we met each other. Uh I subsequently left Tier Point to run a company called Recovery Point. But uh I was involved with the car wash was on the board, and we've done we did an incredible investment in Connect Base. Um so we did that with uh Ben Edman. Yeah, that just had a nice exit. Yeah, yeah, yeah. Yeah, great exit.

SPEAKER_01 2:43

You know what he's been on the show? Has he been? He has. Oh yeah, yeah. Ben's Ben's awesome. Early days. We got there. I was there before he became famous.

SPEAKER_00 2:50

Yeah, he has been. Exactly. Um, so I've known him for 25 years. We did that investment in 2022. No wow. Um great exit. Ben took the business from zero to 200 million. Yes, huge. Uh it was a great, great win for him. And then we did uh I'm on the board of a company called CLC, which is another year, it's an 18-wheeler trailer leasing business. Of course it was. Of course it is. And uh and then we've done beyond. And so it's been a great success and uh a really good run with Waffer. They've been fantastic to me. I'm I have a great relationship with them. Um Otto and Ryan have been the leaders with Michael Coleman as well. Um and we've uh worked incredibly well together.

SPEAKER_01 3:32

So we'd say serial entrepreneur in some way, shape, or form. I'm sure you don't want to be called that, but we just called you the godfather. So say anyone, so you've these are all the companies that you've built. You're also on the board of the colour.

SPEAKER_00 3:42

I haven't built all those, but I'm on the board of those. I'm on the board of a few, and then we built beyond, uh, and then we built Ocolo. Yes, Ocolo is, yeah, really. Ocolo we built in 2000, uh, really started in sort of 2021. Yeah. Um, kicked it off. Ocolo is an online marketplace for data centers. It's uh my partner there, Sean McCarthy. We've done, you know, kind of 20 plus years, and the dinosaurs in the space have decided that that they uh they wanted to, you know, ideally

Ocolo And The Data Center Marketplace

SPEAKER_00 4:10

work with the younger folks. And uh but we we created this awesome platform where we have, I don't know, 250 plus data center providers, about 8,000 providers. Um and it's uh it's a marketplace. Yeah. You know, and we've we've enabled it with AI. We have an online contract management tool, which is all AI enabled. It's been, I will tell you, is starting uh getting into a startup at 50 was uh was not ideal, but uh, you know, it's we've made it work and uh company's very profitable. Um we're small or six people, so you know it does does the job.

SPEAKER_01 4:44

Yeah, and and and that's I think the hotter we s we've talked about this a lot on this particular podcast, but it's getting getting harder and harder to find space and power. Uh and it's it particularly with this sort of rise of inference, and we can chat about that as well. There's there's multiple phases. There's people that are looking to try and build a gigawatt, which is then folks that have to build, find land, find power to build something. Then there's a lot of folks who are utilizing existing data centers who need to land like now or tomorrow. And I think the Ocolo piece is really powerful to say I need space and power now, but I don't know where to look. And you can scour pretty much the globe, but um, I think a lot of it's probably US at the moment, or is it it's worldwide?

SPEAKER_00 5:26

Surprisingly, the first three years were basically all non-US.

SPEAKER_01 5:31

Yes.

SPEAKER_00 5:32

Oh, non-US.

SPEAKER_01 5:32

Yeah.

SPEAKER_00 5:33

Oh wow. And now we've come back, we've sort of repatriated back to the US. Um, but we cover every continent except Antarctica.

SPEAKER_01 5:39

Yeah, well, we're working on that.

SPEAKER_00 5:42

We're Sean and I are taking a journey there to see the. Exactly. Take one of our switches, pop that place for me.

SPEAKER_01 5:48

Yeah, exactly. It's the only country we're missing.

SPEAKER_00 5:50

Yeah. Um, but uh other than you know, it's it's been a it's been an interesting ride seeing. I mean, I think there's been two constants in this data center space, right? You're looking for reasonable space and reasonable power. Yeah. Just the enormity has been insane. Yeah. I mean, we have seen the last 18, 24 months, the growth, you know, far exceeds 20 plus years before. Yeah. It's just it's just it's exceptional. And I think the next five years are gonna be even more exceptional.

SPEAKER_01 6:15

Yeah. You know, I mean we should dig into some of it around like what you're seeing entering this space. Stephanie, we were just at ITW and it was the conversations we're having. In fact, I talked to one of our customers who uses, they provide a lot of compute and so forth. Uh, and they were telling us that they've found data center companies, or there's a lot of VC money or something, we'll call it new money in data center space. And they're just building things that are substandard. They're rushing or they're not getting the parts or they're not ordering in advance. And so there's these folks who've been doing this for a long time, um, who provide great product. There are folks, all these sort of cowboys, I want to say, like that have entered into this space that have just been producing. And it's actually I it's the first time I've started to hear that the product is actually not a great uh experience for them. They gave examples of liquid-cooled floors or liquid-cooled facilities that don't have floors that were raised. Like I don't even understand some of these specs, but like we're talking just design specs that haven't been designed properly, which I find almost that's really hard to understand as well, because you could outsource most of these pieces. But it's like pieces of the puzzle are missing.

SPEAKER_00 7:20

No, I I I think you're spot on. So what has happened is there's so many not I wouldn't need to say entrants. There's there's new entrants, there's current folks that are just trying to get the challenge right now is just time to market. Yeah. And so I think that folks are trying to produce something rush without the quality. Um the other piece is what you brought up a liquid cooling. I would say that less than 10% of the industry fully understands liquid cooling.

SPEAKER_01 7:45

Yeah, that's probably one of the pivots. Even if you've been building data centers for the last 20 years, this is different. Yeah, it's really hard.

SPEAKER_00 7:51

Yeah. Um I think that obviously NVIDIA and AMD and these folks have captured it. They understand it. Yes. They're engineers. I think it's challenging for, you know, folks that have been doing air cooling for 20 years to pivot immediately and say, hey, we're going to liquid.

SPEAKER_01 8:07

Yeah.

SPEAKER_00 8:08

Yeah. It's hard. And I think so.

SPEAKER_01 8:10

What's an engineering, completely re-re-engineered these buildings. But also, frankly, it the it's changed. What's the slow part of the equation is the build to build a data center. It's even slower to get the power, but let's just say the build component for this, what's what's challenging is NVIDIA keeps changing their chips. That's right. And the requirement of Vera is so different to what B's were, which was H's and so forth. And so now all of a sudden you're trying to build something for three years, and every year Nvidia's bringing out something different. And so that is also challenging. And the amount of power they're punching into a single racks, the density of it, changes how much liquid you need to get to it, how that works, how cold the liquid is or warm, and they've now sort of got it to a point where actually warm is better. People looked at submersion for a period, and that's probably just not easy to run. This sort of direct chips become sort of the answer, which I think probably makes a lot more sense. It's it's sort of makes sense that it's been hard to keep up with.

SPEAKER_00 9:06

I think it's been hard. The other thing I want to say is I I'm excited to hire you as my new engineer because you're you're you're talking, your your your uh technical speak is pretty incredible.

SPEAKER_01 9:15

Well, I just hear everyone tells me these problems. I'm like, that is a fascinating problem.

SPEAKER_00 9:19

Uh no, but we do have so I think a couple of things. One, the technology and the innovation has been so rapid in the space that for uh we're gonna go through a transition period within this, within this space. I mean, you think about a lot of data center has been around for 25, 30 years. And that tech that is in there today is quite outdated. Yes. But we've just been running on it. And now as we move to say, hey, how do we become more energy efficient? Yeah, how do we become more you know green and carbon emissions and things like that? Most of those are not equipped to do that. So I think there's gonna be a fairly decent business if folks want to engage in this is the adjacencies is hey, how do we retrofit data centers without interrupting service? Yes. And I think that's gonna be so I talked to uh a gentleman at um Geico a number of years ago, and they had just completely refitted all of their data centers. I went to a two and a half year um Geico insurance and they did a fantastic job. There are gonna be more and more folks that

The Scramble For Space And Power

SPEAKER_00 10:15

are gonna do that. Yeah, retrofit.

SPEAKER_01 10:16

But even if you retrofitted for something for two years ago, oh it's gonna be outdated.

SPEAKER_00 10:20

Yeah, it's uh it's gonna be outdated. Yeah. I mean, I think we're moving so quickly. And so as we talked about just a minute ago, is time to market is so important. I mean, and you know, look, we're in this crazy stage right now where I mean, I I think you and I talked about this a couple days ago. It was like when we started telex, a big deal for us in 2012 was 1.25 megawatts. It's a like exactly that's uh dropping the bucket now. Yeah, it's crazy.

SPEAKER_01 10:44

We took that down in a week.

SPEAKER_00 10:46

Yeah, a couple, I mean that's one cabinet.

SPEAKER_01 10:49

That's true. It's insane. It's crazy. Yeah, giant facilities are getting pulled down into its tiny. It's a sp space is the deal, it's it's not really space, it's power.

SPEAKER_00 10:58

It's power, it's power and how you cool that power. 100%. And I I think the challenge right now, you know, it's interesting. 2012 average was like two to four kva per capit. The average is like 40 to 50.

SPEAKER_01 11:09

Yeah, it's insane.

SPEAKER_00 11:10

It's it's crazy. It's just it's just crazy.

SPEAKER_01 11:12

And yeah, so we've got all of these these things changing so quickly. So if then Ocolo solves a big problem, which is I need a megawatt of power, I need two megawatts of power, I need it now and I need it somewhere. And it's like, well, there's hundreds of providers, actually, data center counts. There's actually thousands of data centers, tens of thousands. Each one's at a different stage. Everyone will tell you, yes, I can get it for you in a year, two years, three years, or now, but it's not this, or it doesn't have a this, and I haven't got that. And it's like really tricky.

SPEAKER_00 11:39

No, I mean, it's like, hey, I mean, I don't know if everybody knows about Raxio, who covers tier two markets in in Africa, and you know, and it's basically evil link in Bulgaria. You know, I mean, these are these are companies that we work with, right? And you know, we find uh I guess Ocolo finds this also sources for, you know, we do a job that a lot of people don't want to do. It's super and you know, we're happy to do it.

SPEAKER_01 12:04

We're we're but by doing it for one, you it's the same thing. You do it if it's you you're doing the same discussion for it's the same discussion. Exactly right. It's just maybe different scale, or it might be like a different power client, or that's really cool.

SPEAKER_00 12:16

But we're looking for unique markets. We've done pretty much all the major carriers and they are major providers, and then now we're saying, okay, how do we get to you know the secondary and tertiary? And uh yeah, it's fun.

SPEAKER_01 12:27

So that's a colour, that's an awesome company on its own, right? But what you're doing with Beyond, which is not even on your LinkedIn little stealth mode, it's a lot of uh that's that's really cool. So I uh my assumption is that you you've been looking for power and space, and you got to a point where for some of the hyperscale clients that probably needed to use you, you were like, Well, actually, there isn't any, yeah, and I'm gonna build it for you and find it. So tell me that story, because that that's pretty amazing.

SPEAKER_00 12:51

Yeah, so so we were looking at And did you bet you didn't think you're gonna be ending up doing that? We didn't. We didn't know. This is it was uh, look, we found a great customer um who's a hyperscaler, who needed purpose-built data centers in certain markets. Um Waffra has uh made some great investments. They've invested in Vantage and DC blocks, they invested in Tier Point. Um, they're astute in the data center space. I had a relationship with them. We brought our client, the client really liked them. It all everything came together.

SPEAKER_01 13:26

Yeah.

SPEAKER_00 13:27

Um, you never know if it is. You know, it could be a bump, and you know, the you know, customer might not like the financial sponsor, yeah, and it all worked.

SPEAKER_01 13:37

Um you've got to bring that's the piece to this, is and we'll get into it, but you've got to bring so many pieces of the puzzle together. That's right. You're gonna have to source land that has access to power, then you get the regulations and the approvals from the the local councils and all these, and this is not just one state or one county. No.

SPEAKER_00 13:56

This is multiple. It's all it's it's you know, we're starting with ten sites across the US, mostly tier two cities. Size of ten megawatts. Which is it's not it's not tiny, you know.

SPEAKER_01 14:06

But imagine ten megawatts three years ago, you'd be like, yeah. It's like three cabinets. Exactly. But I think, you know, I think one of the things that you broke. That's big. It's like 10 megawatts, but times ten across the U.S. and then we're doing Europe as well.

SPEAKER_00 14:24

So easy. Yeah. It's to drive the bug. Yeah. So look, none of how many you're putting across Europe? Uh we're putting nine in Europe. So different countries or across uh at least six countries. Okay. Yeah. So it's so I think the policy one of the things you brought up which has really changed over the last little bit is the policy. Yeah.

SPEAKER_01 14:46

I mean This must be the one of the hardest pieces.

SPEAKER_00 14:49

They permitting the policy now because I think there's a lot of, I would say, um there's challenges in terms of dealing with specific ordinances, where

Liquid Cooling And Rushed Builds

SPEAKER_00 15:00

you're getting pushback from the communities. Yes. Um, which has become, you know, it's become an issue.

SPEAKER_01 15:06

It's emotional, it's it's also totally misinformed.

SPEAKER_00 15:10

It's very misinformed, and I think I think our part within this space is we needed to do a better job of educating the community. Yeah. Right. And I don't know. I mean, I think that's gonna be all of us. It there's some work to be done. Look, I don't think anybody expected the pushback.

SPEAKER_01 15:26

You know, and I think we're Oh, the pushback is in many cases illogical and incorrect. That's right. But that doesn't matter, unfortunately. Perception is reality. So, yeah, this this particularly the the I mean the the liquid thing, the the the water usage one is just the most um bizarre one that is so obviously not true, but uh I think the water usage and the potential for radiation, I think, is is a challenge.

SPEAKER_00 15:46

I think the other piece is that we are bringing, you know, economic growth to these areas. I mean, we're we're putting people in there to do construction, we're leaving people there permanently to maintain the data centers, you know, we're bringing capital in um to markets that really in before, you know, you know, across the US that weren't that didn't have the data centers. And I think I look I I don't I'm not saying that you know it's panacea, but I am saying that we are doing, you know, a lot of great work. I think the other pieces we have to be careful, we want to make sure that they're green and they're carbon neutral. I think that's important. But I will tell you, I mean, overall, I think it's uh it's definitely gotten a lot harder. Yes. You know, it's it's I mean, I would say the last 24 months have been tricky.

SPEAKER_01 16:33

Yes. So in your case, how on earth are you finding land? That's so you're getting too it's like nearly 20, 10 megawatt data centers.

SPEAKER_00 16:41

Yeah, so we partnered with a great developer here in the US called Oppiton, based in Minnesota. They've been fantastic.

SPEAKER_01 16:47

Um, you know, uh the guys, you know, well So you find you hunt down somewhere with power, somewhere that can get the space, then you've got to get and I I chatted to the gentleman that we we caught up with the other day who was saying there's all these mineral rights that are decoupled from land rights as well. That's right. Which I found like a whole new industry, like in that space, which is fascinating, particularly in Texas.

SPEAKER_00 17:09

Well, in Texas, you can actually own mineral rights, which is different than without the land. Yeah, that's that's exactly right. It's quite interesting. Someone someone can have the right to drill a hole through your data center. Somebody else's somebody else's land, exactly. And find oil. The oil's theirs, but the land is somebody else's.

SPEAKER_01 17:26

I've chosen to drill a hole through your data center, through all your chips in there to get to whatever's beneath it. And I just might use that's right.

SPEAKER_00 17:33

Exactly it. It is no, it is. And uh, you know, so I think, but uh, you know, I think you met Drew the other day from Oppon, who's been, you know, Drew and Brian have been, you know, the driving force for us, helping us. They're out there sourcing the land. You know, we we uh we look at the design to make sure it's efficient for and and obviously the hyper the hyperscaler that we partner with approves that it's their design. They want to make sure it's the most efficient use of of the space. Um, you know, and and look, it's uh we got to look out for the obstacles. Like some some areas where we've had to pivot.

SPEAKER_01 18:06

So so can you get like let's say you get how do you get the power bits to it? Or do you come with that ability to access, or do you then have to build into a grid like what's the thing?

SPEAKER_00 18:16

So we've been primarily looking for where we can in these markets specific powered land, right?

SPEAKER_01 18:23

And they've got the power running by to that what do they have to put in? I don't even know what happens next.

SPEAKER_00 18:26

Yeah, what does that look like? It it depends. You know, we want to make sure that we can get, you know, the first the first priority is to see if we can get grid.

SPEAKER_01 18:35

You know, if not and if grid, there's there's already a power line running by.

SPEAKER_00 18:39

It might not be. It might not be. We just need to make sure that we can build a power line or a substation. And then they can supply that much. That's right. And then ensuring that they can give us 10 megawatts.

SPEAKER_01 18:49

And 10 megawat what does even 10 megawatts look like? Is that like a just single power? A lot of them, or how do I do that?

SPEAKER_00 18:53

Well, you know, the the way that the way that we've we're building obviously dual feeds to ensure that there's completely diverse. Um and you know, but you know, look, I mean if you think about plus diesel gen backups and all those that's all of that. Yeah. No, we need to make sure that we're all set with that. And then obviously, you know, crit we're on the critical list for uh for everything for specifically for diesel uh supply, supply and things like that.

SPEAKER_01 19:20

By the way, it's worth touching on that. People probably don't realize that um when there's a a problem with power, it's diesel starts getting diverted to certain locations. And so you need to be on the priority list. You know, they want to send diesel to a hospital, they want to send it to the government, whatever something, something.

SPEAKER_00 19:36

Hospital government, and then data centers need to be on there too, because we are spinning up critical data with these facilities. You know, we gotta make sure everybody gets their Netflix, you know.

SPEAKER_01 19:44

It's so important. You gotta monium. You can't watch your show.

SPEAKER_00 19:49

That's important. You know that diesel reroute ready button. Exactly. No, it's important. So uh, but yeah, we have to be on the critical list. I mean, I think uh, you know, during COVID, most data center workers were on the essential employee list as well. So uh because you had to go in. Yeah. I mean, um I ran a business which was called Recovery Point, which 30 plus percent of the business was uh uh federal government. Yeah, yeah, yeah. It was backup, it was recovery, yeah. So um, and uh it was an you know, we were on the essential work employee list. So and worker list. So that was really important. And you know, I think we're gonna continue to see that. I mean, the more the more we digitize everything, the more data centers are relevant. And and and and and Michael, you run an incr incredible company, you pass it in enormous. I mean, I can't even imagine how many bits you pass. It's just millions, you know. And you know, similar to you, what you're doing is the critical data that's being passed over your I mean, hospital records, you know, whatever it is.

SPEAKER_01 20:50

That's right.

SPEAKER_00 20:51

Lots of clicks, lots of shows. Um, but I think it's important to r that people realize that, you know, at we all joking aside, yes, there are other things besides content.

SPEAKER_01 21:02

Entire markets run across trading platforms, all healthcare records and all come across it. Military, yeah, government, critical infrastructure, power generation, like movements in terms of what pricing it needs to be. Like it doesn't matter, it just everything runs through these pieces.

SPEAKER_00 21:16

That's that's exactly right. And I think as the we are and more and

Why Beyond Builds 10MW Sites

SPEAKER_00 21:20

more is going to run through that as we take sort of legacy applications or legacy work projects and they're gonna be digitized. It the continued use of SD-WAN, fiber, towers, data centers is gonna be more more relevant. Never stops, yeah. No, it's never gonna stop.

SPEAKER_01 21:37

So I I'm gonna keep going in the trend of you building because th like this is a really interesting space. The data centers had become cool, uh, but then everyone's jumped into the space. People have realized that it's not that easy to build.

SPEAKER_00 21:48

Yeah.

SPEAKER_01 21:48

Um tell me something like everyone's been searching for 30 megawatt plus up to gigawatts of power. Do you fit into a spot where 10 megawatts is a little bit easier than trying to fight with the the the folks that are gonna gigawatt campus, 500 megawatt campus. Is that I'm assuming fighting for that is going to be almost impossible because there's just so many big companies doing it. Is that why you can fit into that space or it's just I don't know.

SPEAKER_00 22:11

Like how how how do you think of that we're we're a bit under the radar at this space.

SPEAKER_01 22:17

Which is perfect. And it's actually rare that a a hyperscaler needs that but it's more the specific um use case you're solved for is is unique. And that's as opposed to saying I need yeah gigawatt of all these chips that all sit in one location which is gonna be super hard to find, build and massive but this is like we've sort of Tetris your way into it which is we have and and you know looking at a data center it's kind of like Tetris, one cabin open yes it's exactly the same.

SPEAKER_00 22:42

But yeah we I think we were fortunate that had we gone probably 20 to 30 megawatts it would have been a lot more difficult. We would have had um I think policy issues would have been a lot more critical and challenging. Right now where we sit and 10 it's not I mean how big are they like in terms of size football field and all that so look so we take down roughly 15 acres. So if you think about that but yeah I mean it it's it's in the range we take additional land in case we want to do expansion um want to make sure but uh you know it's a couple hundred thousand we can build up to a couple hundred thousand we'll probably start with a hundred thousand square feet it all depends on you know we want to obviously make sure the cooling's right yeah you know and I think but we do have a a standard design and we we overbuy in terms of land just in case we want to expand right I mean the interesting piece is that you start with 10 and next thing you know you're halfway through your build and you say hey I need 30. Yeah. And can you is it easy enough to get access to it's I think the additional tower is going to be challenging for the next two to three years. But uh we'll start with our 10 and then we'll see see where we go from there.

SPEAKER_01 23:50

And is it more that the infrastructure can deliver up to 30 but the supply side needs to just give you an allocation? Is that what like like what's the difference between going from 10 to 30 for example? Is it that you need to upgrade the physical cables come in or is it that no actually it's the grid whoever's supply or the whichever grid you're sucking it down from needs to say no we have more capacity or I think it's a combination of both.

SPEAKER_00 24:12

We're gonna have to change you know the rack density will change you know depend because we're probably going to want more you know more higher density. We'll have to go back to the to the uh the power proof the utility and we'll have to change that. Do we need to build a substation?

SPEAKER_01 24:29

You know um there's a lot of different requirements once you go from that and a substation you need to build to take like um like high voltage transmission lines and send it down.

SPEAKER_00 24:39

Is that what that's for or yeah and I mean it's distribution, right? So I think I think what we're we're gonna want to do is depending on you know you it's all about proximity where where you want to build you know some sometimes you don't need a substation neck right next to you but sometimes you're gonna want to have it. I mean I think it's I think it's important that as we build um there's a plan for potential future and what yeah and and look if we look at it by building a substation that it gives us more optionality. Right. And we say okay then we you know we can do this these types of such and I think the key is really you never want to have a single point of failure. So um you want to make sure that we have at least five to six fiber providers coming in for for uh all coming out different paths yeah all coming out different paths in different in different ducks um ducks and different trenching um we want to make sure we're completely diverse from a power feed perspective um and then diesel gym what is your diesel you how long do you need to run something like that? Yeah I mean look we we want to have access for a number of days yeah um and you know uh so and we want to make sure we're on critical list. Um you want to make sure you also have pathway access to get into the facility. Yes um that we're not building you know like in the middle of woods that we can't have access um and that the streets etc are are accessible. Um because it's like this is the part that's it gets it gets it gets yeah more complicated once you start to think through this. Yeah.

SPEAKER_01 26:08

I mean I think when you put all these you need to have all these things it gets really tricky.

SPEAKER_00 26:13

I think you and I have talked about this is there was just so much movement of capital into this space. Right. And I think everybody wants to continue to move into the space. But I'm not sure uh you know look and this is folks who are sort of new investors in the space that really need to understand the mechanics are right into it.

SPEAKER_01 26:33

Because you you just hey I'll just dump money in here and and but I don't really you know following yeah it's easy just yeah start building well I found what I what I found really odd is I think

Permitting Pushback And Public Myths

SPEAKER_01 26:45

you know there's plenty of companies that can get into construction and build as in you you outsource to a construction company. We do we outsource to a construction but so it's surprising to see them get it wrong. That's what I was really surprised at at W to hear multiple folks come up with these like data centers that are sp that are randomly built huge amounts of capital gone into them. You would think that the easy part is to do a high consultant and build me a data center spec and hit all these certain requirements but obviously it's for whatever reason not occurring. I I that's really bizarre to me because I'm sure that's an easy problem. I thought that'd be an easier problem the design should be an easier problem to solve.

SPEAKER_00 27:21

I think the challenge you have is there's been this movement to get into the data center space of builders construction folks that you know these are first time yeah and that's a challenge. We partnered with Oppiton instead of Minnesota fantastic group they've been doing this for quite a number of years they're they've been building data centers they've been building data centers for quite a number of years um and I think that they not only they're familiar they're doing it for multiple hyperscalers yeah so I mean they're they know their way around I think and look we you know we're fortunate that uh the hyperscaler we partnered with knew them at familiarity with them and it actually ended up and they'll have a certain spec as well that's that's right delivered that's right and they can do soup to nuts design build construction and operate for us. Yeah yeah yeah right so wow so we've got you know incredible partnership with them.

SPEAKER_01 28:12

Yeah so you build um and then I assume these are long term leases that happen to be a lot of these are long term leases yeah yeah yeah these are long term leases it's like gold at the moment so let's long term leases and you know investment grade tenant you know um we're in the midst of uh doing a debt raise on this um which is exciting and you know you're going through I mean you do you do this you you're I'm a novice at this you're not not in this space we're publicly traded yeah us we can raise capital on the market we did so for the acquisition exactly which is incredible can you at someone else does all that for me you know I just instruct people and then bankers appear you know it's like yeah yeah no we're I mean look we're uh it's an interesting company because we're a very small we're small group we're uh it's we're eight people it's uh we outsource you're an eight you're eight people you're building almost 19 10 megawatt data centers all across the United States and across Europe um couple of years what Alan's putting it we have 24 months to do this so we're we're shoveling that's why you're not sleeping that's why you and I were up at working out at 4 a.m yesterday for the listeners who are listening to this uh watching it this banana I get a text message at 4 a.m for 10 a.m it just says uh you know a word that was uh suggesting that I'm weak and then saying get down to the gym man and you made it well the weird part is I am not a morning person but I was so jet lagged I was awake and silly looking at my phone and I was like I'm not going back to sleep now this guy's challenging me as I go down there thinking that I'll beat him uh I'm younger and he's actually punching out way more so he's like I got schooled well our good friend Mr.

SPEAKER_00 29:52

Shuck will give him a shout out right now um you know we're we're waiting for a push-up competition with them as well so uh yeah Henry where are you beating come on yeah we're waiting we're calling you out right now that's but uh look I I think this I I mean as do you I think we get excited by this industry right because this is only it's so exciting now because it's so incredible I mean the the pace that we're moving at now it's nuts is just nuts. It's just nuts. But and I don't you tell me but my view is we have exceptional growth of five to six more years at least.

SPEAKER_01 30:26

I can't see the the it's slowing. I can only see it accelerating at the moment. Yeah. This is this whole discussion about the bubble. Yeah is a bubble is a bubble and then it's like everyone in the moment in time would say oh it's not a bubble everything's amazing that no matter what bubble anyone's been in but the thing I liken it back to is I remember cloud came into existence sort of 14 15 years ago everyone started getting into that so it's what it's what megaport was founded on. So that's all cloud pieces is and the clouds exploded they just kept scaling and it's like the cloud bubble when's the bubble going to stop it all this but eventually what happened is people stopped saying when's the bubble going to pop because it they just got so big and then they were just punching out 20 to 30% growth per annum. So what the bubble popping so to speak was just that it wasn't now at a growth rate of that it was now just more consistent growth. But but by that stage 30% on the top of something huge is massive so I suspect AI is going to do that and but it's by the time it's right sizing itself is so big. And the the challenges is so different in that it's so much power requirement there's so much build it's like there's a natural dampner on how fast we can grow. When SaaS grew there was no restrictor um if you think of every SaaS company they could literally grow grow at infinity almost because you could sit all your software on AWS you could scale it anywhere in the world you could sell it on USD and just infinite growth would be okay. It could actually withstand that that's fine when Amazon's doing it and it's a small amount of compute required for software, any software business. This is rate limited by how much physical hardware can be actually powered on yeah and NVIDIA chips are not cheap. So there's a huge capital constraint that well the capital's actually available but I think it's how you monitor and manage the capital is key. Startups need access to them but they can't get the capital so then there's companies appearing to sort of provide in all these different ways there's weird funding mechanisms going on. There's a C of it but at the end of it there's this throttling in terms of how many how many GPS they can turn on and how many Nvidia can pump out and then how much um even this the the memory and just lip drives they can't get enough drives to turn these things on.

Land, Grid Access, And Redundancy

SPEAKER_01 32:40

So they can actually get the NVIDIA chip sitting there but they haven't got a drive to make it work and without that one piece of the puzzle you take one piece of all of this giant jigsaw puzzle out and it stops working. So then there's this natural dampener occurring we saw it with Anthropic like we just saw the other day Anthropic literally had to start dampening everything down. They just didn't have enough GPUs to hit and then they went and did that deal with Elon taking out all the X that'll get them for like 10 minutes before the next piece of growth so now everything's sort of this wave. So then people say well the the bubbles coming still there's still just not enough build coming for that Jensen called out like a trillion in chip spend this year just from the hyperscalers last year was half a billion this year it's a trillion the hyperscalers wrapped up and gave their forward looking guide or their guidance in effect not forward looking but they just gave guidance around capex spend which aligns and that's going to come through then there's all the neo clouds that are going to have to build for all those other folks. So I don't know it's like naturally getting slowed almost in a weird way even though it's growing so fast.

SPEAKER_00 33:40

Well I I think what's you know you you talk about companies value company valuations today. I mean you know Apple's at four billion four trillion trillions yeah these are these are exceptional. Five trillion plus you know I I think though the one thing that we've we're spending a lot of time thinking and talking about AI which is incredible. It's a great you know it is going to be it's going to continue to be the future but not to be sort of forgotten is that cloud's still growing at 30 to 40% a year.

SPEAKER_01 34:08

Huge amounts of compute getting deployed like unbelievable quantities compute.

SPEAKER_00 34:12

The the other thing that's really interesting is Hunt the IT adoption of fully outsourced services is growing at like a half a percent it's growing at a very small pace. So the global idea and adoption of IT outsource is still huge. I mean there's so much we're in the early innings. Yeah yeah you know and then lastly so I think there's sort of four areas it's AI it's the IT adoption enterprise adoption of IT outsource service it's you know cloud compute and we're still growing content is going to continue to grow. I mean we're being broadcast it's going through somebody's fiber data center tower et cetera so they have sort of these four buckets we talk a lot about AI but there's three other buckets that are still growing 30% a year.

SPEAKER_01 34:57

Correct and they all live in the data center. That's right yeah so they they all need fiber they all need the data center they all need a CPU a GPU some storage or a network that's right yeah I mean it's a great business to be in you know and I think you know the more we did that the DC side's far more like a real estate place. It's 100% for tech. That's right the other elements that they put into it is tech. But it's it's super complicated in terms of the designs now what those looking for so it's just I don't know it's like I I think the complexity has changed, right?

SPEAKER_00 35:28

We used to say you know it's space power and cross connects it's not just that I mean you're talking about if you think about where companies were you had maybe somebody who focused on in 10 years ago you had somebody who focused on power uh and that person had three other jobs at the company. Yeah. Now you have you have a data center companies have 40 people in power 30 people in power. Really? Oh yeah it's such it's so important to all as close to the amount of power they need to get to a single cabinet is that what's it's working with the utilities working with the community it's working with you know whether you're working with alternative use power you know whether you're you're trying to investigate into can renewables you know are there other sources and other means I mean the other area is policy. You didn't really talk about policy before now you have people have groups that are set up for policy. I mean I I think this is look it's just a maturation of the business right and uh we've moved through this which is great. Look I think we're creating more jobs you know which is great. So we're set creating We chat on that panel and and your mate was talking about the the electricians, the Sparkies that are we so we so there's some fairly significant builds going on in Abilene, Texas, right? I believe that electricians are being flown in there from 48 different states. Yeah because we so if you want you know moves for data center space, you know, we have experts in you know the plumbing and you know electricity distributed how they distribute power how they are yeah exactly and I mean I think there's gonna be there's going to be a shortage of that type of skill set.

SPEAKER_01 37:01

So actually he was explaining something that was quite interesting and that a lot of industries can get support like hands and feet support that aren't experts at a certain thing, but you can teach someone to do something pretty quickly. Like splice fiber. There's no degree for it that takes four years. There's no whatever you can teach someone you can start to get them good. But electricians there's what is that four, six years or whatever on the training just to get to that point.

SPEAKER_00 37:23

Yeah and so I mean I think to have that expertise yeah and especially from a data center perspective. Yeah that does take time. But there are some there are some areas that you can work in this space that require less training and less schooling that I think also then you can obviously mature through the space and and learn more.

SPEAKER_01 37:40

Yeah which is creating the you know new waves of jobs. And so then everyone's carrying on that AI is going to retake all these jobs and this is the the piece that I find funny.

SPEAKER_00 37:48

I'm far more of an optim optimist around that's like I mean look I I think it's gonna be AI in conjunction with human development and human interaction right I don't think you're gonna you know I think that there will be some jobs that AI will make more efficient right um I think there but there's going to be a lot of you need to have human development. I mean I don't think you you know unless we go to robotic sports which uh which I don't think people will find as fun yeah you know and exciting um you know that's that's you know there's sports there's there's so many things that but look I think the influential the only thing that's gonna happen I believe with AI it probably needs to is there has to be some regulatory around it because you're talking about financial training and things like that and using you know the efforts of some type of AI you know algorithm that can obviously influence. And so you know look but look at look at the other thing healthcare it can influence and create like you know can we find better and quicker ways to prevent or or treat truth specific diseases and things like that.

SPEAKER_01 38:50

That was always the case. It's like you um if you think something's going to replace something what usually happens it just speeds it up which means humans do it more. Yeah and that's and then and then we just do more and we get better at it. It's like someone gave me the example an MRI it's so hard to do it's so expensive takes so long to look at it. If I makes that really really fast then we just do lots more of them. That's right. Like I'll probably add one in my life you know we might do it every month or whatever. But who knows? Yeah.

SPEAKER_00 39:15

I I think you know 20 years ago computers were going to replace replace everybody. Yeah. We use computers but I think not replacing everybody.

SPEAKER_01 39:22

Yeah exactly I don't think that's a big worry. So what else is interesting in your work? What else you got going on?

SPEAKER_00 39:28

Yeah what else am I got going on? Well you know I think uh a lot a lot of engagement with this project yeah that'd be a huge management it's a big one um you know on on the personal level you got to keep the fitness up yeah you know you and I are big big balls on the oh I'm only catching you I've I had I was uh yeah this is my new yeah look I've got a little wood I can get fit this year that's I I love the wood that's great I gotta get one of those um look I I think this is you know one of the areas that I've been passionate about uh within this space is um my involvement with PTC yeah that's super cool I've been involved with PTC this I think I think this is my 30th PTC so you're now what is it president chair chairman PTC chairman yeah I'm the chairman I tried to get them to change that to Godfather no one they weren't listening to me yeah I said this how good would it be if they just we're officially uh that's something I said to you assuming that you're from my as my first act as chairman I'm changing the title to Godfather from henceforth it shall be known as the godfather of PTC we'll we we'll rewrite the bylaws exactly because it rewrite them you can I'm serious the buffalo that'd be great now I'm passionate about PTC it's a nonprofit um and PTC's evolved like PTC's different I mean I think we've I've only been in megaport three years and in the last three years just that the the change of who's turning up is like completely yeah it's the clientele the attendees etc have been incredible we've expanded a bond you know it's adjacency is the data center to the fiber space to the tower space cloud providers cloud providers neo neo clouds we had 1000 plus people there this year yeah you know and we want to continue bankers funders private equity you name it yeah yeah yeah that's more more suits were all in around you know we you can't wear a suit in a wire it is weird suit and tie come on who know you got yeah yeah yeah I mean I think the one thing that's really important about the mission of PTC is the the

AI Demand, Cloud Growth, And Capital

SPEAKER_00 41:30

it's a nonprofit we give back to the community we invest a lot in in the Pacific we invest we provide uh education to up and coming managers I you know we've have this uh class called PTC Academy we've partnered with uh Columbia University where they bring the f uh professors from the business school so look I'm super passionate about this um I've been uh part of the planning committee for the last 20 years yeah yeah I know you know when you you know you know I told you I'm you know I'm pushing I'm pushing Walker status you know so uh um yeah but uh you're just getting started about a hook setup exactly I'm pretty passionate about that you know and I think that's a it's a great organization Brian has uh you know um done a great job you know Sharon ran PTC for quite a long time she was amazing Brian she retired and Brian took over Brian's done immense work's the CEO I'm sorry Bri Brian Moon he's done a great job of kind of working through you know what do we want to see who do we want to you know uh how to curate something that people want to turn up to yeah and I think we we've had different you know I'm I you know we're gonna personally invite you to uh the Ala Kaye which I want to interview you there this time in uh that's our new stage a little less stories yeah you've got amazing stories we got amazing stories but I think you know um that's where I'm super passionate. We have PTC Japan coming up next week. Oh you are you going to that too? I will be going to Japan on Tuesday. Stop traveling next week you got you got 19 data centers you've got to be building and by the way and this is the other thing you're you're actually on the board of a lot of different companies I'm on the board of a few companies uh I'm on a void of vanish europe I'm on the board of uh Spanish Europe's data center companies Spanish Europe is a data center company I'm on a board of amazing little uh it's not little I shouldn't say that it's 1623 Farnham which is an Interconnect data center in uh in Nebraska they're amazing you're at building number two um shout out to Bill and uh Bill Severne and Matt Reid Brian Bradley um Brian you uh know I know you're leaving for Bahamas because you got a tough job no um but uh this weekend uh no I'm on the board of a company called CLC which is an 18 wheeler trailer leasing business uh she's yeah Mike Gore does a great job and that's a waffra uh okay is that three guys yeah and then I'm on the board of a company called Mock uh wireless mock networks what is that um mock networks is a 4G 5G uh Sense. And then uh what did I miss? I'm on Cade Lever's board, which is ColdCom, and they've done an incredible job. Of company started six years ago. Cade kicked it off. Actually, it was started in Australia and then they kicked off here. He's done an amazing job. I mean, we use them.

SPEAKER_01 44:19

They make actually all of our spaces look cool because they put all the little things.

SPEAKER_00 44:23

It's really nice. Yeah, they do it. He does a great job.

SPEAKER_01 44:25

And all the fiber connections in that, that must be just he introduced me to RM Williams too.

SPEAKER_00 44:30

So you got some boots. I got some new boots. Australia's greatest export. We graduate. So I left that board, but it's a great outcome for So that you've moved off that, I assume, with the new ownership or something.

SPEAKER_01 44:47

Yeah, so so new ownership, but uh I saw Ben, he's probably got I think yes, yeah, he's still gonna stay around for a few years or whatever.

SPEAKER_00 44:54

Yeah, no, Ben's Ben's gonna see it through. Yeah, he's done an Ben Ben's done. I mean, Ben built that on his credit card to a $200 million company. Yeah, it's insane. It's pretty uncre pretty unclean.

SPEAKER_01 45:04

And the amount of connections come I mean, yeah, what was it?

SPEAKER_00 45:07

Hundreds of thousands of cool. It is like and the fact, you know, kudos to him. Um, you know, you start and you rough it. You know, I went through that with Ocolo, except he started 10 years earlier than me at 10 years younger. You know, I was as I said, I'm in, you know, approaching uh Walker realtor status. Um yeah, no, I I mean I think and you know, the other thing I think you get from working with all these folks is you get to work with really smart entrepreneurs. Yeah, that's super fascinating. I mean, these guys are these guys are really sharp, really creative. Um, and you know, it's great to talk about, you know, for you and I, it's pretty lonely at the top. Yeah, you know? Yeah, yeah. And you're like, holy cow, who can I talk to?

SPEAKER_01 45:49

Yeah, that's hard. Um you also got something with Aussie Super as well, don't you?

SPEAKER_00 45:54

Yeah, so so Aussie Super is I'm an independent director for them for Vantage. Oh, that's okay. That's that's the linkage. Yeah, yeah. So but spend a lot of time with them. Um, you know, going out to Australia for for an event that they host in the next couple of weeks. Um they've been they want to be our bigger investors, and so yeah, they've been awesome for us. Yeah, they've they're great, you know. And I can't say enough about Aussie Super, you know, work with uh, you know, a myriad of different folks there and uh they've but they've been great.

SPEAKER_01 46:21

They're and they've also branched out um because they typically, I mean, they they're a superannuation fund in effect. So that's what's really interesting about Australia is uh 10% of everyone's wealth or income goes to actually these funds which help you retire so the company country doesn't have to pay for pensions and so forth. But with that comes unbelievable amounts of capital inflow constantly, and then they're constantly having to deploy that capital. So the bizarre thing is like the deployment of capital is like sort of perpetual. They then have to figure out where they put

Electricians, Workforce, And Real Jobs

SPEAKER_01 46:50

that and they have percentages of you know, I think it's probably a certain amount as you go to Australia, and then of that they break down to X percent in you know, tech and real estate and you know, mining and blah blah blah. That's right. Um Yeah, so in your in your case, they've all they've also got uh international and overseas, they also have private. So they also have vantages, oh no, is it public?

SPEAKER_00 47:10

Data's private. Um they've invested in a number of other companies. They're invested in uh Syria and they're invested in uh data bank. Data bank, yeah, I saw that as well. They're invested in quite a number, um, and they've made you know that very sharp investment team. Yeah. Um, you know, spend spend a lot of time with those guys. They're good people too. Yeah, they're good. They're good, yeah. Aussies. Yeah, of course. They're Aussies. Goodbye. Exactly. They're Aussies. I don't know, good man. So yeah. No, and I look, I think uh, you know, it's exciting to talk about the industry. Right. I mean, we're in an industry that I don't see it slowing down. As we talked about earlier, you know. Um, the one thing I, you know, I think we're all gonna be interested, and you know, to to put another plug in for Brian and team at PTC. I think PTC is you sort of come there in January and figure out kind of what the tech is for the rest of the yeah. Yeah. It's kind of interesting to see, hey, what's going on, you know?

SPEAKER_01 48:03

It's like an uh yeah, you uh I think it was invalid, maybe you can tell it. It was it I think it was originally invented for all the carriers when they first started to start to actually total each other.

SPEAKER_00 48:12

Yeah, minute was it's time, was it it was trading minutes, so for voice 50 years ago, it was almost 48 years ago, it was started where East meets west and the middle point was in Hawaii. Yeah. And you go and you do minute exchanges for voice tracks. That's hilarious. And you talk about accounting rates, and that's it.

SPEAKER_01 48:30

Well, nothing's changed. We're just changing bits or whatever. Maybe now it's um space, but the yeah, it's a great this is the bit of and COVID, you know, we all went home, everyone worked on video. Humans still need to get together. Um and it's in that um in those discussions, you you hear the the little the it's like the water cooler chat between the organizations. That's where the the discussions that that are are not shared if you get on a video call. It's like that example I told you about just you start hearing a few stories come out about these data centers that actually are not built to spec or not built appropriately, and then you're inheriting these spaces. And these companies that are buying them assume that they're gonna be delivered at quality spec. Like when we land in a data center, we expect we just built well. Exactly. And then they've got their entire business writing on the fact that this is a time that it's gonna be delivered, and we're gonna put our infrastructure in and then publish that infrastructure to our customers, and all of a sudden it comes along and it's actually not delivered to specs, which means you can't deliver it. Uh, and then to the whole company gets like there's a big challenge through it. But that's that I've never heard I hadn't heard that last year. Last year all I heard was we're out of space and power. That's right. And the year before that was like things are starting to get like move, but last year was like we've sold out completely. Um, and then this year was just like actually the builds that are coming through are random in some instances. Now, there'll be plenty of good ones, but it's just like an interesting sort of anecdote that I picked up.

SPEAKER_00 49:56

What's interesting now is to try and find a megawatt or two is is almost impossible in most markets. It's so I mean that's you know, you think about that where, you know, because we're sourcing basically for 100, 200 gigawatt. Totally. But you can't find a megawatt.

SPEAKER_01 50:10

You can't find a meg, yeah. It's crazy. There's so many gigawatt builds going on or whatever it is. So if I just want a meg. I want one thousandths. Can I have one cabinet? Yeah, pretty much, yeah. It's it's really interesting. And so then uh someone was telling me that they're looking, I met this company that jumped on, they're finding power and space in like um commercial buildings. Yeah, sort of like stranded power. If there's if there's five megawatts or something, I don't know what it would be in a building, but let's say it's two. I want it's only you can use one, we can put one in here and you could do that. Are you seeing that as well?

SPEAKER_00 50:44

Or do you think that is space or folks that are looking at you know industrial builds today that if you can repurpose that industrial build to make a data center that has What about office buildings or something? Like is less This is what I heard. I think there's some you know, there's a bunch of people that are looking at taking malls and converting them into data space. Yeah, because you know, there's been we've had a lot of challenges with the malls here in the US where in certain areas where they're you know, you know, we've we've basically, you know, Amazon has you know has become irrelevant. And so can you can you repurpose that into a data center or something else, you know, industrial building? And so I think we've seen that. I think you are I would say less so in New York City will you take an office building because New York City power is expensive. Um we have we are a single point of failure here as well. Um just because the power distribution uh just because bridges and you know it's it's easier uh to probably be in New Jersey. Yeah. Um and pop uh actually send days a little in. Yeah, well that and you know, you know, if you know, if there's a geopolitical event or something like that, or you know, some type of activity that you know prohibits you know entry and access.

PTC, Boards, And Industry Community

SPEAKER_00 52:00

Um so I think that that that's more relevant. But yes, I d I think that we are in a point where we need to find stranded power wherever we can.

SPEAKER_01 52:11

And then the other side is that people are choosing who takes their power. Like that's the other piece.

SPEAKER_00 52:16

It's you know, it's it's you know, it's basically uh, you know, if if you're a credit worthy tenant, you get it.

SPEAKER_01 52:22

Yeah, that's that's the interesting piece. That's the funny piece. Like some of our customers who have scaled so fast are actually startups that are huge. Yep. Um, you know, raising huge amounts of capitals. They're very well funded, they're worth, you know, billions of dollars of market cap, but they're still not credit worthy in the or they don't have a credit component. They're credit worthy, I suppose, but they're they're so new, they're so uh the banks don't have a rating for them or what have you. So yeah, so it's hard um for them to now figure out what they do in that space.

SPEAKER_00 52:51

It's tricky. Yeah, it's tricky. I mean it's it's kind of like you know, I would not I don't think I'd compare it to the 2000, you know. Yeah, that's everyone's insane meet stuff like that. I like it's I don't agree with that. Yeah, I think it's quite different. I think Well, there's revenue in this. That's probably different. Yeah, I mean, people are making money.

SPEAKER_01 53:08

Yeah. Yeah, versus just market cap as it with no money. That's exactly right.

SPEAKER_00 53:13

Right. And you're basing it off of the future, right? I mean, there is actual income coming in. I think that you know that's what we need to realize. I think the challenge today is everybody wants a uh you know, investment grade tenant, right?

SPEAKER_01 53:25

Which is they all want AAA Microsoft to work. Yeah, that's that's that's it's but the Yeah, the um the interesting thing about the fiber is most of it was unlit back in the 2000s. These data centers are getting utilized. Like the power is getting the chips are turned on. So they are sucking, you know, power out and they're doing something.

SPEAKER_00 53:45

Right.

SPEAKER_01 53:45

So you're not turning them on and not doing something with them. They have to be doing something, yeah.

SPEAKER_00 53:50

I think one of the things that, you know, is unrealized today is that we've sort of there was this thought of data center, data center, data center. There is a huge amount of emphasis that we need to put on in the future of how important fiber is. Yeah. Right. I mean, there is no wireless alternative from a connectivity standpoint today. Correct. And so I think that the point that, you know, your company and others are becoming so much more relevant today with the massive usage and passive. Yeah, right. Exactly. I mean, it's, you know, not only is it, hey, I need, you know, my I need my servers, I need my X, I need my power, I need my HVAC, et cetera, I need my fiber too. You know, and I think that's really important to kind of think through.

SPEAKER_01 54:38

Do you when you're building your DCs, how do you figure that piece out?

SPEAKER_00 54:41

We make sure we when we send a site location to the the specific hyperclot that we work with, it has to pass through their fiber team. Okay. So their fiber team has to approve.

SPEAKER_01 54:50

So either figure out they could run, they run to it or could get some or there's or there's three or four local providers or five providers.

SPEAKER_00 54:58

There might be a build as well. Yeah.

SPEAKER_01 55:00

So why don't you keep doing that then? Because I you know what I feel like's coming is inference. And I and inference, this is one thing we've found with our customers. Is we had this statement is AI, you know I don't know if you've heard this from NVIDIA, they're like, AI doesn't care where it goes to school, which was this statement about a year ago or two years ago, where training was so big and they needed all these gigawatt scale data centers, and then they put all these chips and have them super dense inside these data centers. And then the statement was it doesn't care where it goes to school. And the meaning behind that was it didn't need to be located in New York from a latency perspective. It could be in Australia, in Tasmania, or wherever it is, in some um location a long way away, which was true or is true. Yep. That's right. But what's different now is inference is actually another step. Not only does it not care where it goes to school, it actually doesn't care if the chips are all together at the same school. And that means that inference can have every chip operating independently as long as it's connected to the internet. And so in theory, you could stick all these chips in every in, you know, you could have 124 chips, uh servers with eight chips in each one, 1024, 128 servers, sorry. And then you got you

Stranded Power, Fiber, And Inference

SPEAKER_01 56:10

could split them to 124 different data centers in theory and run that. Now, what I think is coming is instead of just having to have a gigawatt campus where everything's so dense and all right next to each other for training, but I think the 10 megawatts or 5 megawatt campuses constantly appearing all over the place is enough to keep running enough inference. Listen, you could probably get, I don't know, um, call it 124 uh B300 um servers would probably be about 1.7 megawatts. You multiply that out, you could get X in in five, whatever that number works out. Yeah. That's a lot to to do what you need to do. So I think if we could I think there's a space to keep doing that.

SPEAKER_00 56:50

Um I completely agree. I mean, look, I think we're gonna see more and more of these. I think we started with edge, you know, and that didn't it didn't pan out as much as the other thing.

SPEAKER_01 57:00

Yeah, well, edge was this it was more around trying to solve for latency and people probably don't care that much. That's right. That's right. And I think what we've done is But it will it will for la for inference where they just don't care. No, no, it will.

SPEAKER_00 57:12

And but I think what we've done is the edge dat the development of the current data center providers, Equinox, Digital, TierPoint, their data centers have become edge data centers already because they they brought in a number of fiber, et cetera. And so And actually the density is not high. No, no, no, then that's it's not high. I mean you're you're you're talking about a you know eight to ten KVA per cabinet. But I do think as you you touch on this, we are going to see deployments of five to ten megawatts across the globe to to handle this inference that that's coming. And we're not there yet. Yeah. But that is a huge opportunity.

SPEAKER_01 57:46

I feel like that's the next wave. It is. And I feel like everyone's just focused on the large build. And it makes sense. You go after the big moisters to start with, and the funding comes from it because it's triple A, all the rest of it. But we're seeing huge demand for just just give me one megawatt, give me two megawatts.

SPEAKER_00 58:03

And we're gonna continue to see that.

SPEAKER_01 58:04

Yeah.

SPEAKER_00 58:05

And I think that we're in the initial stages of that, right? Um incubating that process over the next 12 to 18 that it's probably like maybe 36 months, it's definitely gonna happen.

SPEAKER_01 58:19

Uh DC is building that, or they're all just going for the big end. Like I actually don't hear from any of them. Not not yet.

SPEAKER_00 58:25

No. Not yet. I mean, I think I think there's such a huge focus.

SPEAKER_01 58:28

Like are you odd building 10? Is that like a weird thing?

SPEAKER_00 58:32

I mean, it's definitely unique. Yeah. Um it wouldn't have been like a few years ago, but now it's sort of a few. I mean, yeah. 10 megawatts a few years ago was like a spot. That was a giant. Yeah, that was a giant, you know. But I do think it will change. You know, we'll see that development going more and more as you talk about inference. I don't think that that's there yet. There's a lot of discussion around it, but it's happening. It'll happen. It's coming. Yeah, yeah. Yeah, yeah, yeah. Yeah, it's definitely. You're gonna keep building or you're gonna be. Yeah. I mean, look, I think I think we are we are very, very happy with our partner in Opp and the hyperscaler we're working with, but uh, you know, I think we want to focus on this project right. Yeah.

SPEAKER_01 59:09

It's a big one, it's a big one. It's a big one. It's huge.

SPEAKER_00 59:12

It's a big one. It's only 19D. It's a lot of work. Oh, congrats, man. Thank you so much. Appreciate it. It's been great to catch up. You uh I'm in awe of what you guys are doing. I mean, I think congrats on your acquisition because it's uh incredible. It was perfectly. I mean, it couldn't have been better, and I think you got a

Final Thoughts And Wrap

SPEAKER_00 59:30

great team in latitude. Yeah, that's a good idea. You have an awesome megaport team. Yeah, and I think you know, I'm I I love to see good people successful. And uh, you're doing a great job running this company, and I think you know, uh I'm I'm the old dinosaur here and I like to watch the young guys actually kick ass.

SPEAKER_01 59:49

Perfect. Thank you.

SPEAKER_00 59:50

Thanks for having me. Thank you.