Yet Another Value Podcast - Randy Baron's bull case on $GDS

Episode Date: April 22, 2022

One of the people's favorite guests, Randy Baron, returns to the podcast to talk about his thesis for GDS. GDS is a Chinese data center company, and Randy goes in depth into the company, including why their tier 1 footprint gives them a moat, how GDS is trading at a discount to private market values, and why ADR and VIE fears may be overblown.My notes on GDS: https://twitter.com/YetAnotherValue/status/1516805266565644289?s=20&t=_7FcaMp4YYgZ1S4htJXz6gRandy's first podcast on AMRS: https://twitter.com/AndrewRangeley/status/1407669290241777666?s=20Randy's second podcast on RNLX: https://twitter.com/AndrewRangeley/status/1448604500009242632?s=20Tubes (data center book mentioned in the pod): https://amzn.to/3rKUxRqChapters0:00 Intro2:20 Data Center Overview9:10 Chinese Data Center overview11:20 Randy's history with GDS17:45 GDS valuation versus private market comps24:00 What does Randy see in GDS that the market is missing?28:45 GDS's international expansion plans34:00 Why does GDS's tier 1 footprint give them a moat?39:50 GDS's M&A and growth ambitions47:30 GDS's recent raise from Sequoia China and peer VNET's acquisition offer52:15 The China ADR and VIE issues58:55 Wrapping up the GDS thesis

Hello, welcome to yet another value podcast. I'm your host, Andrew Walker, and with me today, I'm excited to have on for the third time, one of my favorite guests, one of the people's favorite guests, my friend, Randy Barron. Randy, how's it going? Andrew, how are you? Doing good, doing good. You know, as I was saying that, I was just thinking, like, I can't believe, Jacob Rubin's the one who introduced us, I think, and, you know, I can't believe we formed a friendship.

It's the power of the podcast, but let me. What you refer to as the yet another value empire, which I always laugh when you're right about that is really amazing, right? And I have this broader theory that like everyone knows Buffett and Munger and like the next generation was like a belly and all that. But like after that, these guys in their 70s, Lee Cooperman, all of them like there is no great known kind of investor class. And what you've kind of created here in a format, even if Jeremy Raper gets disproportionate representation, is kind of the next generation of up and coming. thinkers, right? I mean, we all look for these idiosyncratic things where we say, you know what, active management has a future, the market's off sides. And let me tell you why I think the market's wrong. And that's a pretty brave thing to be the point of the spirit. So thanks for letting me come back on

again. Hey, no, look, I appreciate that. And that's the thing. The people who want to do, hopefully people know this podcast as people come on, they do deep dive, serious work. And the fact is they're just the people who love it, do it and they love it. But there aren't that many of us left. And it's just fun to be able to connect with so many of them. But let me start this podcast after that. Let me start this podcast the way I do every podcast. First, to disclaimers remind everyone, nothing on this podcast is investing in advice. We'll be talking about a Chinese ADR today.

So people should, you know, nothing on this podcast is ever investing advice. But that carries an extra level of risk. So people should just be aware of that. And then second, with the pitch for you, my guess, you know, third time on at this point, I don't feel like I have to pitch. but Randy is anyone who's listened to the first two podcasts knows, Randy's a super thoughtful guy, and I think this podcast is going to be great. That's why he's one of the people's favorite guests.

So that all out the way, I think we're going to do a bonus episode with a little bit of update on some prior ideas later, they'll post later, so people can look for that. But the stock we're going to talk about today is GDS. This is a Chinese data center company, but I shouldn't say anything else. I'm just going to turn over to you. What is GDS and why are we so interested in them? Okay. GDS holdings trades on the NASDAQ under the symbol GDS. It also trades on Hong Kong, and that's going to be important. We'll come back to that minute under the symbol 9698. GDS is the largest carrier neutral data center in China. It's also, in my opinion, the best run data center in the world. And certainly in the public comps, it's the fastest growing data center in the world. It's EBITDA margins. are best in class, and quietly it's begun expanding

to the south and to the east from this kind of position of China history. Yet, despite all of that, it trades today at 10 turns of EBITDA less than its industry peers, despite being the fastest growing, despite being the best. And for a lot of reasons we'll get into, which involving the political backdrop, the macroeconomics stuff is happening in Russia,

But I thought what would be really useful kind of before we dive into the nuances is to take a step back for those in your audience who may not know what a data center is. Yep. I think that's great. I could use a refresher, to be honest. So that's perfect. Well, it's interesting. I came to data centers.

So, okay, the first two podcasts you mentioned, one was on synthetic biology. One was on a healthcare name called Rinalytics. Both have great updates, which we'll do later. but what's fascinating is that the dirty little secret that I don't really brag about is that I'm a former and reformed telecom analyst. So for the first 10 plus years of my life, I was in the TMT land that, you know, when you write your charter stuff, it catches my eyes that I know this language. You know, I don't know if I knew that about you, Randy.

I'm surprised you. I don't brag about it. I keep that kind of close to the best. When I was thinking about it vying Altis at 30 last summer, you couldn't have parachuted in and said, Andrew, come on, I followed this forever. Those guys are awful. And it makes stuff like Netflix, which reported last night, really kind of fun to watch. Like, you know enough to be dangerous.

But I first kind of got introduced to the space, maybe 12, data center space 10 or 12 years ago when a company out of Cincinnati, Cincinnati Bell bought Cirrus 1. And at that stage, like, none of us kind of knew what data centers were. So what is a data set? A data center can be thought of as the brains of the Internet, right? The role of these buildings is to process, store. and communicate the data behind the myriad information services we use every day. All the stuff that the yet another value empire kind of is about, right?

Social media, online collaboration, emails, your streaming video. You got to think about these buildings. They've got racks of servers, right, which deal with the kind of computational logic response to, like, request. And you've got storage drives that house the files and the data that is needed for those requests. You have network devices that connect that data center to the internet, both for enabling inbound and outbound files. Then you have a ton of electricity, which generates a lot of heat, so you need something to deal the cooling often or just a way to offload the heat.

It's useful to think of a data center like a hotel, right? Data centers make money by leasing power and by leasing space. And to some degree, it's a very virtuous cycle where the ecosystem. system that it's in determines the price. In other words, the more carriers, the more networks that are present, the more valuable becomes to have connectivity for what they call collocation in that data center and therefore it costs more to least space in the more successful data centers. I often think about, and I know this is a bit macabre, but I think about terrorism, right? Like if al-Qaeda was to strike again, what's the best target? And none of us,

No, it's a black swan event, hopefully. But I often think, because you and I both know, you can give up food for a couple of days or weeks, but you know, we can't give up. Twitter for two seconds. This thing right here, right? And we can't give up our phone. It's like if the federallies want to find you, it used to be they tracked your cable address. That's actually how they would track the FBI. Now they track your mobile phone.

And so my point is when you go to these buildings, you got to imagine these huge multi-football field size warehouses with no branding on the side. because they're trying to keep it, you know, you go to Secaucus, you can see them. There are these huge white buildings. And to get into them is tougher than getting to the Pentagon. And when you walk down the bays, there may be, you know, 150, 200 foot tall. They have big trades up in the sky that have in the U.S., it's a little different in China. We'll get to that. But in the U.S., you've got copper on one tray, and then above that, you've got fiber.

And because of the legacy of AT&T here in the U.S., ever since Maubel broke, up in the 80s, you know, you had all the competitive local exchange carriers, the rural local exchange carriers. You had 300, 400 different phone companies funneling their phone cable, whether it was fiber or copper, into this warehouse called a data center. And then in the data center, you have cages. So when people say, what is the internet? Data centers are where the internet physically connects. You have the Netflix cage, near the AWS cage, near the Facebook cage. And from those servers, they go up into those trays I talked out and go down to the next bay or the next window. So that's kind of a little primer on what a data center is and why it matters.

Can I just add one thing in there, Randy? I think that was great. And I think the one thing that will come back to that I think makes sense. If I remember correctly, there is a little bit of a network effect where you mentioned the Netflix trays next to the Google trays. And I think there is a little bit of a network effect, if I remember correctly, because what's going to happen is Netflix and Google might want to connect their trade. So if you're a data center that has Google, the Netflix kind of wants to be in you, again, co-location because they can connect to that. Whereas if you're a small startup data center, you get a little bit of a chicken or the egg problem where Netflix says, why would I, at least with you? Like, I can't connect to

Google, Facebook, charter, AT&T. I'll just go connect Equinix, I believe, is the biggest player in the U.S. I'll just go use Equinix because they've already got everyone else that I want to connect to, so you get this really nice network effect where everyone's at your location, so everyone wants to be at your location, so everyone comes to your location. Am I thinking about that correctly? Yes, and I want to take it one step further, which is there's a word called hyperscale scale. Yeah. These are the major, you just mentioned a lot of them, but add Alibaba to the list, AWS, Microsoft. Like there's about 2024 of these hyperscale players in the world, and that just refer to, I think I've used it before with you, you know,

Infinities are placed it's so far away for wanting farther, it wouldn't really matter. The idea that such computational power comes with mass, right? But what that also neglects is kind of not the hyperscale player, but this is called the microscale players, the guys that are looking to kind of come up and nudge into that framework, they're going to want to come to the place for the hyperscaler's termining and they want to co-locate. And the best example that you can give in the U.S. is Netflix, which was on. AWS and still is, but over time wants to control more of its own destiny, so all of a sudden

starts doing more and more of its own data centers. I think that's a really revealing revelation. China, and I will come to my history of GDS in a minute, is a little different because you didn't have all the diversity after the mob valve breakup. In other words, you didn't have 300 different phone companies, which have since been rolled up in the U.S. So when you go to those data centers, you essentially have fiber coming. You skipped a big part of that copper interface so much so that that company I mentioned

Cirrus I, which was the purchase data center from CBB from Cincinnati Bell, maybe five or six years ago and no longer has it, did a minority investment into GDS, because GDS at that point was not a true biblical seven-year generation, but generationally, lowercase G ahead for what they're doing with their data center campuses. I think that's really revealing. The other thing in China, of course, worth noting is that 50% of the data center market in China is the incumbent. So it's China Telecom, it's a little bit China mobile. And then there's 50% of what I again described with carrier neutral data centers,

in words, they don't care if the phone company is, GDS is 30% of that other bucket. So call it 15% of the total China market is GDS. Perfect. So I think we've done a nice job framing what a data center is. We can obviously add it on to background on the data center if you want. But if you're ready, why don't we switch over to, we've covered the industry. So let's switch over to GDS in particular in your background with them. So let me start, if you don't mind, a little history in terms of how I got to know GDS because I think it's revealing.

GDS and now we're only talking the U.S. listing. They did their Hong Kong listening at the end of 2020. But for the sake of this history, let's go back to 2016 when they IPOed at $10 in ADS. And by the way, for your audience, ADS or ADR is American Depository Receit. It's essentially U.S. vehicle to own a foreign stock. Very common. This was, and I was laughing when I referred earlier to our friend Jacob, but also Jeremy Raper, all three of us, kind of look for idiosyncratic ideas, right? And I had a phone call from the RBC analyst as a friend, John Ackin, the TMT analyst, who said,

Randy, I got one for your look at it. And every once in a while I get that call. And, you know, more often than not, it's a wild goose chase. But sometimes it really works out. And at that point, this $10 IPO was a broken IPO. It was trading at $7. No one wanted to own anything in China. That's right around the time that Baba is starting to get build traction, but it's a little

before them in the U.S. market. and I went to lunch. And I met these guys, the CEO, William, who's the founder, still owns five or six percent, and the CFO, Dan Newman, who's one of my favorite CFOs in the business. And if I tell you the mood of the table, it was lamentation. It was pure sadness. Why does the market not realize what we are doing? And at that point, people forget this about China. The cloud did not start in China until 2014. So at that point, 2016, we're talking, China was basically 2% of its total IT spend

was on the cloud. It's double digits now, but that's part of the argument, which is they're growing. And so therefore, the two main customers that GDS had were Alibaba and Tencent. Two-thirds of the revenue

were those two hyperscale players use the term we just introduced. And what was fascinating was this relationship with Alibaba, which continues through today And as part of the thesis, it's why I love it, which was if Alibaba was going to grow, GDS was the contracted data center to grow with them. So if you believed that the data, that the internet in China was going to have a future,

this was a very, very easy way to play it. And by the way, because of the way that those contracts were structured, you could pencil out the P&L three or four years and pencil in 50, 5.0% growth top line with EBITDA margins at that point and the 20% going up to what they are now, 50%. So I sat there and I was like, guys, if you just cover the company, cover the geography, which is part of my process, right? Like I'm totally agnostic to geography. I'm just looking for great ideas. I said, guys, don't lament. This is a wonderful opportunity buy more shares. And they did. They bought more shares personally. Like I said, William, the founder, now 6% owner of this multi-billion dollar enterprise, the market cap for your audience

in US dollars, five and a half billion dollars, another four or five billion in debt, 10 billion enterprise value. So I sat there and went in, went back to my office with my team, and we kind of did the work. And I was like, this is amazing. We're going to go. It became a top three position for us. We bought it about $9 a share. And then by the end of 2017 had sold it, had doubled. It was a win. It was great. But as you and I know in this business, relationships matter. Right. And so even though we were out of the stock at that point, we had maintained a dialogue. And I remember, I'm out of the business. I'm out of the business. I'm out of lunch. This is in the end of July 2018. And all of a sudden I get like the flashing like,

you know, the bat phone goes on, you know, get to the office type thing. And there was a new at that point nascent hedge fund called Blue Orca, still around a short selling fund. Yep. Put out a 50 page short thesis on GDS. It was actually in retrospect, the best executed short thesis I've ever seen because they did it in the afternoon, which was kind of middle of the night morning time in China and Shanghai. And at that point, there's 25 publishing analysts today, but at that point there was four or five. Frank Laothan was in, who's the Raymond James guy, was in the Unity courtroom in Virginia or D.C. John Atkin, I mentioned, was in Australia on a plane home. No one in China could respond. And in the period of three hours, the stock lost

47 percent of its value, lost almost $2 billion. It was one of the best executed, and actually Bill Alpert and Barron's talked about it. This was like the best debut of a short thesis ever. And part of the reason, by the way, this goes back to knowing infrastructure, was in the report on like the fifth or six page, they had kind of one of those surreptitious shots from the hip of a cage that had wires hanging out of it basically in Frank, fraud. This is LinkedIn coffee. This is fraud. You need to run. Well, you and I know as apartment dwellers that it doesn't matter if you're in the apartment or not. You got to pay for that apartment. Buy rent, whatever. You got to pay for it. And so in the case of GDS, because we know data centers,

I don't care if those cages are occupied or not. I know they're being paid for the cage. So whatever the technician may be doing there doesn't matter, but it paints a really, you know, poignant picture for the short. Anyway, we went back in in a real way in the low 20s. The socket dropped that day from 40 to 20. And then we wrote it up for the next three years. We exited in the COVID year 2020 at 80. The stock last year, February, so 2021, hit $114 a share. And then when you and I started talking about it last November. November, yep.

That's when the macro, you know, the China regulatory landscape, all the stuff we'll talk about today had begun to shift. And then in February of this year, when Russia invaded Ukraine, and the inference was, is this World War III? In other words, is China going to take this moment to seize Taiwan and make the land grab, which would effectively become World War III. All of the China Internet stocks got hit, and we're able to really build this. So now this is the top three position for us.

You know the other two. They've been podcasted, about 7% of both. We're going to need you to get a couple new positions on so that we've got future podcasts. But that's great. There's lots of places I want to dig in there. I think listeners probably, you know, you know from my notes, You can tell where I'm going to start. But first question I'm trying to ask going forward after we've got the background and everything.

You know, GDS, the stock, as I look, it's down, you mentioned. It's down a lot over the past year, you know, probably 60% or so over the past year, a little bit more from its peak in kind of the November time frame. Or, no, it's not a peak from probably the October time frame. But, you know, the market's down 67%. it's still got a pretty high multiple, you know, I think it's around 16 times EBDA, which isn't crazy for a data center, but it's not super cheap. But when you look at the stock, top three position for you, what do you see that you think the market is missing? Where, you know,

top three position, you obviously think it's going to generate risk adjusted alpha. What are you seeing that the market's missing? Well, I want to, I want to disagree with one sentence you slipped in there. This is very like, you know, Andrew Walker, he's a smiley guy, gets it in there. But the one sentence I don't agree with you is the data centers at 16 times is expensive. So I'm going to give you a little bit of statistics, which I wrote out, so forgive that. But I think it's relevant. And I'm going to give you the three most recent public transaction multiples, okay? Because the argument I'm making here is, guys, we know stuff's going on in China, right?

There should be a country risk discount. Absolutely. What I'm arguing is it shouldn't be 10 points. So June of 24, Blackstone acquired QTS Realty Trust 26 times EBITDA. That's treating CapEx's debt, but 26 times EBITDA. Cirrus I, that company we mentioned, the spin-off from CBB, KKR, Global Infrastructure Partners, acquiring them for 23 times 2022 EBITDA. Corset acquired from American Tower, 27 times. 22 EBITDA. And Switch, which is the latest one that's rumored to be the next one to go, is trading at 26 times today. Right? I can look at my grid here. You mentioned Equinix, so I should just probably bring that up as well. And Equinix is trading today at 24 times.

Right? So just to give you some frame of reference, I want to finish framing that out. Hold on a second. Just going back, the percent change in revenue, this is really relevant because GDS today is trading it under 15 times, okay? I mentioned QTS acquired. Year-on-year revenue growth, 12%. 55% EBITDA margins. Serus 1, 8%. High single digits, revenue growth, 48% EBITDA margins. Core site, 8%. Again, the point is top line growth, high single digits. GDS, this year is its low year, for all the reasons we're going to talk about, including COVID, 21%, 22% this year, historically 30 to 40% top line growth with the same EBITDA profile 51%. Look, I agree with you.

I guess in my 16 next EBDA, like with Charter, I've been saying since the stock was at 12 times ago, I was like, this is too cheap. Go find me a cable company that's been taken out for 15, for less than 15 times EBDA in the past five years or something, you know? And the stock's gone from 12 to 10 and people have been saying, oh, it's expensive. say no versus every other deal I can find. It's actually cheap. It might look expensive versus, you know, a retailer trading at two times EBDA. So my 16x was more in the absolute sense. But I am 100% with you. Like one of the key thesis is with data centers and everything,

and we'll probably get into this a little bit more as well as these are extremely popular infrastructure plays, right? When debt is at two or three percent, go buy a data center. You can put tons of debt on it. It's going to grow. And you get double leverage on it. So, I'm definitely with you. Even on that point, even in this increasing rate environment, on the private side, and by the way, the private multiples are even higher than the public multiples, called 30 plus, cap rates right now, last week, 4%. So the point is, like, you can still do a lot of that.

And in fact, another metric that they use for the space is growth adjusted EBITDA. It's essentially the same concept of a peg ratio, which is PE adjusted for growth of EPS over time. And because a lot of the companies just talked about charter included, you don't use PE, use EBDA, right? Because it's depreciation in there and throws off the map. But if you look at the kind of so enterprise value divided by EBDA divided by growth in EBITDA, you know, GDS today is at one, one point one time. The US peers two and a half to four and a half times. Yep. So that that point that you just said about charter is important. But my point is you don't get this kind of growth with, by the way, 70% of their revenue.

revenue in the backlog. So I can add 70% to revenue just in converting backlog. And churn in a data center, we probably should have led with this, is nil. Because to what we're saying, the network effect before, once you're in, yes, you may need to replace a server, right? Things wear out over time, et cetera. But you don't, you don't give up your footprint. You add to it over time. And theoretically, the thesis basically boils down to, are you going to have more internet usage or not. Is AI going to have more or not? The great example I can give you in China, I remember even the New York Times maybe two or three months ago, they talked about how COVID in China and what they call the health code has created this techno authoritarian tool,

meaning they used to know location data from your phone, but now because you need to proactively put in more data, they have even more and more data tracking you. And, you know, the zero COVID policy, you could have a whole political conversation of it, that sound or not, from an epipelagogeological perspective. But the idea is, you know, I'm of the opinion that we're going to use more and more data over time. And especially like we said before, the cloud in China is still in its nascent days. It is going to surpass the U.S. over time. We're just still in the kind of call it third inning of that. Let me just go back to the question. So again, I am with you on the valuation, everything. But I just want to go back to the question.

Today, you know, the stock's down 60%. Obviously, you think this is valuable, like, what do you think the market is most missing when you look at the stock? Do you think the market, there are lots of options. A few could be the market is over-discounting the China risk here. The market is, there's concern on financing, and we'll probably talk about that later. The market's too concerned about the financing needs of the business, all sorts of other ones. But what do you think, like, your biggest variant perception on the stock versus the market is?

So I think the issue is the market is viewing today. And by the way, this is not just China. I think a lot of the market has thrown out growth altogether and said, you know, interest are going up. Growth is dead. I don't agree with that. And that's a whole different conversation we could have. But in the case of the China data centers, and by the way, there are three. Chin data is one simple CD, 21 Binet, which we'll talk about how to bid, VNT is the ticker and GDS. The market has basically said there is a structural issue with these companies. And that's what I disagree with. What's happening with GDS today is we are at a cyclical low, which happens in infrastructure companies. Then you kind of just flavor it a little bit with COVID and, oh, maybe add a little bit of geopolitical Russia backing, you know, saying what's going to happen with China? And don't forget that ever since the education stuff in the summer of last year, the regulatory

regime in China has made it so that today the rate of growth of internet companies in China is the slow as it's been, despite the fact that the Chinese government, at least in its words, is saying we're trying to support a digital economy. Like that never the two shall meet, right? We can't keep stop having internet companies stop them from growing in China and also say we want to be a digital company in the future. I think that's going to resolve itself over time. I clearly don't know when COVID is going to resolve. Right now, today, recording this in April of 2022, there are people in GDS data centers who are sleeping in cots in the data center. They're not leaving. Zero COVID policy. They're not letting it in.

And that's wonderful because it keeps it going, but it's also very revealing that we're not in normal times. I don't know when Russia comes to its senses over Ukraine. I'm hopeful, but I don't know. So what I'm saying is that we may have one, two, three quarters of slower growth. By the way, slower growth is still 20%. Let's just be clear. on that. It's not 40, it's 20. I'm saying that as I look for for ideas that are going to compound over time, and I guess this is, this goes back to the amorous and realics things, but I don't, you only care about price if you're selling at a given moment, right? Otherwise, it's a great buying opportunity and you're looking for things almost like a private investor

in public form, private equity investor. You're looking for things that are going to compound generationally. And I don't want anyone to leave this podcast. Often I feel, when any of us get up and put our hands up, you know, we get slapped for, oh, you're talking your book or, oh, you know, you're pumping a stock. I honestly, today is the 20th of April. I don't care what the price is in 21st. I may care at the end of the month when I get marked market. I make care at the end of the quarter, honestly. But truthfully, generationally, what do I care? Where is it three, five years ago? That's how I try to have a worldview. And I do believe that three, five years from all these issues we should talk about, the ADRD listing.

You know, China, COVID, I think it gets resolved over time. And my broader thesis, and this is a long-winded answer to your question. Nope, love it. That A, China has an internet in the future and be specific to why do I think this is misvalued? I'll come back to this at the end. This is an excellent management with a founder-led company, great track record, focused on execution, got great relations with the customers. Alibaba loves them, has them build stuff. So that's super important.

they've had STT, so the Singapore Telecom guys have come in, which are among the toughest due diligence investors in the telecom space and have owned 32% of the company. I mean, they've clearly said what we want to have VR-China data center play. That's all positive. We've got a huge customer base and ways to grow in multiple ways, including the fact that they're just now entering Singapore and leaving China. So up until this point in the narrative, it has been a 100% intra-China story.

And what I'm saying is we're not picking up on. You mentioned equinex, is this could become the next equinix, but it's just coming from a different geography. So it kind of gets dismissed because it's not in the Occidental world. That's perfect. That's great. This is a smaller point in the overall thesis, but it was something that was curious to me. You know, most Chinese companies that I look at tech companies don't really have a lot of success. outside of China because in China they're basically behind the wall, right? China's not going to let

anyone else compete with them. I'm sure Google would dominate in China if Google was allowed to really run a search engine in China, but the fact is they're simply not. The only Chinese internet company that I can think of that really had success outside of China is TikTok. So there right now, I think it's two countries. It's Singapore and one other country who's escaping me, but they're in the process of expanding, really expanding outside of China for the first time. And You know, I don't think it makes a huge difference to the downside of the story, but it could impact the upside where I look at then and I say, hey, did they only really win in China because they had that protection and they were, you know, they're really in a no competition zone? Or do you really think these guys are operationally at a point where they could go to a different country and kind of win, despite I haven't seen it a lot from Chinese companies doing that in the past? Okay, a couple things that unpacked there.

it's not just China that stops tech companies coming in. And there's a company in Russia called the Andex, which is effectively the Google of Russia and beat Uber out because Uber wasn't allowed to come in. Another podcast guest of yours, Adam Lindsay, who came on and did TripAdvisor. He went through the bond indentures. It turns out that there was in the Yandex paper. If you didn't list for five days, it was never disclosed in a 20 app. It was just in the bond indenture.

If you didn't publicly have a stock for five days, you got to claim the debt. I mean, it was like crazy stuff is happening in the world. So I don't want us to say this is just China. There's other countries that don't allow tech companies in. Yeah, yeah. To your geographic point, let me get there, but let's start with where is GDS today. Okay. GDS is in only tier one cities.

About 65% of the countries, and it's not just GDF, 65% of the country's data centers are on the east coast. So just kind of coming down to east coast, you've got the Anzi River Delta, that's Beijing, basically. That's one. Shanghai is two, the West Pearl River corridor. So think Hong Kong, Macau in the south. That's the third major area. And by the way, when I say those cities, they're also in like 150 to 200 kilometers, so about 90 miles in radius on that. So I'm just using these because we would know these major cities. And then the fourth market in the middle of the country is Cheng Yu and Changcheng, which are two $20 to $30 million population centers. And that's kind of all these tier ones that are at.

What's changed in the story since I first invested in five years ago is today they are expanding to Singapore. So if you look at, again, the map of China, you go south, there's Thailand, Vietnam, I'm working my way down, Malaysia, which is one of the countries they're in, and Singapore, the little the island, they're the size of like Chicago, right? And then Indonesia, the island archipelago below it. Singapore has a moratorium. Again, small island, not a lot of space, moratorium on data centers right now. They are putting to bid a 60 megahertz deal that there's going to be a category that if you don't have a data center in Singapore, you're allowed to come in. But short of that,

the way that GDS is positioning it is from Malaysia on the southern shore and from Indonesia, which is a little bit more like our Alaska, you know, they're coming in. You can see it from Singapore across a bridge or a ferry right across the bay. I think you're absolutely right that this is option value, but I think it also goes to speak that they are going where their customers are asking them to go, right?

So Alibaba and kind of that hyperscalor we talked about is the one that said, hey, let's go to Singapore. And GDS is so committed to this idea of going outside of China that their COO, Jamie, is now based in Singapore. I mean, they're very, very fully committed. So then the question becomes, I mentioned STT is a 30% owner. Well, they've got data center assets in Thailand, Vietnam. Is there something over time where GDS maybe combines assets, does a merger,

Georgia joint venture? I mean, these guys are wonderful capital allocators and we'll get into the economics of their kind of per unit data center. But I think Singapore is the canary and the coal mine. that you're going to start seeing this Chinese company start moving. And by the way, it's not a surprise. If you talk to people in Hong Kong today with the crackdown of the PRC in Hong Kong, there's no free press right now in Hong Kong for all intents and purposes.

A lot of people are choosing to exit, especially in the business community, to go to the Singapore, to go. So my point is you will find the next global city of the future, but it's not going to be equitics. It's going to be someone that has a relationship with the local cloud player. like GDS. Let me ask you a quick question. So my experience with this, I was looking at a company internet a couple years ago.

I'm trying to find the book, but I read a book about data centers and the history and everything there, which was very interesting. But, you know, if I was somebody who hadn't looked at a data center before, I think my question would be, it's all the internet, right? So, you know, one of GDS's claims is, hey, we've got these great data centers in Tier one city, Shanghai. You mentioned them all, but they say it's really hard at this point, maybe impossible to go find enough land in a suitable location to build a data center in a tier one city.

And I do think somebody who doesn't have a lot of experience here would say, why does that matter? Why can I build a thousand miles from a tier one city, build a data center, and we'll just lay fiber connecting it from a thousand miles to the city and, you know, it'll be a lot cheaper because it's a thousand miles from the city, lands much cheaper and everything. So why does being able to build and why does being able to geographically expand like this? Why does that matter? And because you mentioned INAP, which is dangerous when you drop that kind of stuff, we're going to talk about why data center only matters because INAP's a little closer to like VNet and some of the other players in that they had managed hosting as well.

GDS is a pure play data center, which is really important. The book I read, by the way, just for anyone who's interested, it's called Tubes, A Journey to the Center of the Internet. I'll include a link in the share note if everyone's interested, but sorry. But I think you've touched on a really, really important thing, which is the least. latency matters, and specifically low latency. In our business, you would say high frequency trading, right? The reason in the data centers in Socaucus is that Wall Street is, you know, as a crow flies, whatever that is, five miles, six miles. And so just the more distance you put in, the more latency you put into the network. And if you're a CTO at a company that's especially

trying to build your own cloud, you don't want latency. It's as simple as that. It's a redundancy. real estate in the core cities has gotten very expensive. And in the case of China, because they're still, it's kind of like we talk about the environmental issues of the world. Well, guess what? China uses coal. And the price of thermal coal was up at 1.40% last year. I think it closed the year up 20%. But the cost to generate power was totally skewed from what all these companies needed. And as you go more and more towards cities, I remember before the Olympics, because they shut down all the factories because they didn't want pollution. Well, the reason is because that's the burning coal, right? And so what's fascinating in the case of GDS to that story

is, A, they have enough power quotas for the rest of their bill. I mentioned the 70% backlog. They're fine. But it's a moat. Because to your point, if someone came and said, oh, why can you just build? You can't just show up in China and build an inner center without power. These are massive, massive warehouses, maybe instead of the Cummins or Caterpillar equipment, it's the Chinese equivalent, right, of the generator. But you need power. You need to be on the grid. And that's where China's regulating to say, hey. And like I said, in Singapore, total moratorium, right? So I think GDS is in a pretty good place in that regard. And I want to also mention because I do think, and I feel like every time we do one of these, we talk about socially responsibly

investing or environmental and social government's ESG concerns. GDS today is 30% renewable power. So they're taking, to your point, power generating the West through wind, through solar, and they're piping it in to do their data centers. Equinix is a leader in. So I want to totally tip my hat. I think 80% of their global numbers, renewable. That's wonderful. Others like DLR, 20-something percent. So GDS, again, as I'm positing it within the peer group, understands, This is their goal. By 2030, they want to be carbon neutral. And they're doing this ahead of their customers because the internet companies haven't

come yet and said, oh, we need to be carbon neutral. But they see the writing on the wall. And again, these guys, they have vision. They totally see it. They say, we want to do what's right for the world. And by the way, if you're an internet company, and this is not specific to China, although it is the case here, 60 to 70% of your power usage as an internet company is your data center. So if you can cut that to be carbon neutral, we've done a good thing for the planet.

And I don't think that should be very. Well, and it's not just a good thing for the planet, but you're seeing it right now, right? Like, Nat gas prices. A year ago, you and I would have been talking, and Nat Gas would have been priced at $3 per MMBTU or whatever it is. And today, as we're talking to this domestically, it's way worse in Europe. You know, it would be $7 or $8 for MMBT to you. And, you know, if you're a data center and your power is getting supplied by a Nat gas,

but guess what? Your power cost just tripled. Whereas solar, yeah, it only runs. runs during the day, but a lot of the renewables, at least once you do that fixed cost of getting them installed, you're no longer kind of subject to the whimstitude. You're no longer subject to the isolations and power prices. So, you know, not only is it an ESG thing, I think it's just a smart business thing to say, hey, we can get a lot of our power locked up at basically no costs and have good visibility into getting power. And by the way, shame on us. Before the podcast, we were talking

about writing our letters, our quarterly letters. Like neither you nor I have like, you know, exploration companies and oil and gas, like it's been a total win, right, since Russia invaded Ukraine. But that's for me a very conscious decision. I want to, every time I come on, I feel like we talk about this, but I want to invest in things that are going to leave the world a better place for my kids. And I don't mean that as like, oh, I'm a woohoo SRI guy. Like that's not my North Star, but just the idea of you want to back ideas that are constructive and you want to back companies that have the same ethos and vision that they see that this is a change that you can make. I think it's wonderful.

That's generally what I do, though. I do have one, one oil and gas company. It's Amplify Energy who had the oil spill off the coast of California, which was not their fault. But I felt bad when you were saying that because I was like, oh, Randy says this. And of course, I own an oil spill company, but not their finger there. And you can refer to that podcast. I listen to that one, too. Tim's great. I love Tim. But let's talk to GDS. Another thing. They say, hey, they're very clear M&A, they want to grow, right? And one of the ways they want to grow is M&A. And I want to talk about M&A in two forms.

A, is there MNA out there for them at this point? Because you mentioned how most of the market is controlled by either China Telecom or they control a big piece of the non-China telecom market. So is there kind of accretive bolt on M&A for them at this point? And then B, we can use that to go into a discussion on their financing because I think financing is a, is both. a almost many catalyst for them what's happened recently with them and it's a source of concern. Okay. So yes, there is M&A. That's very simple. That's part of their strategy. It's why in,

I think March or February of this year, 2022, STT and also Sequoia, China put in more money for them to be able to go forth. And liquidity is not an issue with this company. People will come to you and say, well, this is a very levered company. And I think when we talk about this and that's part. We'll come back to why I think that's a false flag. But when China cracked down on the education companies last year, June, July, August, and the summer, whatever it was, what people don't realize is two things happen that are good for GDS. One, the hot money left the market. Private equity has been looking at China for a long time. But basically, if people are like, we can't get our money out, we're not going to put money in. So all of a sudden, GDS, which was

competing against all this hot money coming in from other places, for the mom and pops went away. Second, GDS and stock price we mentioned had been $115. When that rolled over, that was no longer a comp or a benchmark that people were able to use. So it's kind of almost a victim of its own success where the private guys would say, well, you're getting 20-some-odd times in the public market. We want you to buy us for third versus these guys who seek to have an IRA and they execute it, by the way, all the time. Their goal is between 10 and 13 percent for any project that they're doing. How do they do it? They basically go out and they get bank funding to pay for 60% of their new build spend. So when you look at the net debt to EBAA

today, which is roughly seven times, a little inflated, it's not a fair number because you have the debt and none of the denominator. Yep. Right. So you get the money, project finance. They're all ring-fenced. The banks structure it in this, you know, 15-year contracts so that the payback period, because they understand the cash flows of data centers, takes you two years to bill. So you don't have, you have the payments are back-weighted, right? And you basically borrow 60% of the data center. You have a guaranteed customer because Alibaba's come to and said, we want to be in this geography. So I can borrow, it's a, you know, good paper, if you will. And they go to the banks and they borrow on that. And that's why for that moment in time, on a normalized basis,

once a data center is up and running, leverage on a net basis about three to three and a half times. And that's kind of where I look at it on a constructive basis. So if this was a pro, you know, pro forma, all the buildup. Now that having been said, they're building all the time. It's about 65 fully owned GDS data centers plus another 15 that they call BOT, but basically building for other people. And it's a fascinating conversation. that the banks never push back on this issue of, you know, are you two lever? It is entirely an equity holder conversation. It's one of the things, A, it's one of my favorite catalyst, but it's one of the things I've seen so many companies that do, you know, basically what

something similar to what GDS does. They build big things that cost tens, hundreds of millions of dollars, takes two or three years to come online. You know, you've got this huge asset. But I love it because the screen really poorly because, you know, when that asset's about to turn on is right when the leverage on the thing looks crazy and particularly something here where it's project financing. So even if for some reason it couldn't start, you know, you could just hand the keys over to the bank. But I also think it speaks to kind of the moat here where these guys want to go build a new data center. Guess what? They've got Alibaba in hand. And as we talked about once you've got Alibaba, you're going to get a lot of other people who want to co-locate

and connect to Alibaba, whereas if you and I started Randy and Andrew's data center business, right, we wouldn't have Alibaba there. So we couldn't get the financing because we don't have that key customer. We couldn't get anybody else to come in because we don't have Alibaba. So I do think it just shows. And Alibaba today, because this is important, about 40% of the data center project footprint, whatever you want to call it, 25% of revenue, 10 cent, the other big customer, I think, about 25% of revenue. So it's important to note. But remember when I said, four or five years when I first looked at it, those two were 65, 70%, right? So as the other player, so cloud computing was 70%, like I just said five years ago. Now it's about 50%. And what's

been interesting is the large internet guys are now call it 30%, and the financial players are about 20%. And what's happening there is essentially, so we all know the Chinese government's been cracking down regulatory and all this stuff. But in part, one of the things that benefits GDS is the Chinese government is trying to stop to what the Andrew and Randy data center would be, you and I in our little like closets with servers, right, we wouldn't have a huge footprint, not yet, but it being a server, they want to be able to control the data if they need to access it. So they are forcing the closing of the mama pops and the consolidations into the bigger players like GDS. So that's actually a virtuous wind.

behind them. It's also interesting because the banks, the Chinese banks, have made a decision that they want to have more things instead of on the mainframe in the house on the cloud. And we talked about how cloud is lagging in China. Well, the Chinese financial institutions are the tip of the spear to say, we're going to be more and more in the cloud. So that's the part that's been really growing. Ironically, that was also how GDS started its business with the financial institutions years ago. We should probably talk a little bit about. VNet and some of the comparative stuff, but what's fascinating is that both GDS and 21 Vionet started almost as retail guys. And GDS saw the writing on the wall where the enterprises

wanted to go to the cloud. And they said, we will help facilitate that. 21 Vionette made a different path. They said, we're going to be essentially the white label for IBM and for Microsoft. So if you want Azure in China, it's VNET, okay? But what that means is they basically turn their back on the Chinese cloud players, and now they're playing catch up because people say, well, why is the multiple, even at this 50% unsolicited bid premium, why is the multiple for VNET half of GDS? Well, that's part of the reason. The VNET business, not that it's a bad business, it's essentially an infill business,

what we said before. Once you're in those tier one cities, you're there. you're not leaving. And Alibaba is there and everyone wants to be there. And so it's a very constructive process versus infill, which, yeah, you can do great on project financing and project building, but it's not, it's not the headline, right? It's very much as filling in the punctuation where it needs to be. Yep. So you mentioned 21 and I think I said earlier, there are two financial things here that I think are like many catalysts or many reaffirmations of the business. And one, Again, you mentioned it, Sequoia China put in a pretty big check into GDS in very, very low cost.

I think it's 0.5% convertible bonds. 25 basis points, 0.25.25 basis points in convertible debt, which gives them cash to go play with now that multiples have come down. And it gives them, Sequoia China is super right here a little firm. It kind of gives them more blessing, right, on top of everything else we talked about. So that's number one. And the number two, probably a little more interesting timely. is VNet got an offer a couple weeks ago to buy, to buy the whole company for $8 per share. And at the time, this was like right at the Nadir of the Russia invasion, everything, Russia, China,

just getting sold indiscriminately. I think the stock was $4.50 or four. But as you and I speak, VNet is trading for about $6 per share. And there's an $8 per share offer on there. And, you know, the stocks when hammered very much like GDS over the past year. So I guess the question is like, do you read anything into the VN? B-Net offer? Do you read anything into the China's the Sequoia China convertible financing and we'll build out from there. So two different topics. I think you're absolutely right. The Sequoia

China, a feather in the cap. I think STT, the Singapore technologies telemedia, which by the way, just for people that aren't as familiar with this region, that's the TMT investment vehicle of Tamasic Holdings. Tomasik Holdings is the Singapore state-owned company that owns like the airline. It owns everything is founded in the mid-70s. And they first invested in GDS back in 2014 with a 40% stake, did a follow-up listing at $65 in ADS in 2020. And then like you just said, did the $620 million convert note with what's the Koi. The point is, for anyone that says, how can you trust the numbers, right? Okay. Our job is determined who can, you and me, Andrews, determine who can be a good allocator capital, right? And the only way you know that is with time.

Right. I've seen these guys now for six or seven years allocating capital really, really well. So I don't worry about that. But then if I like had to say, okay, well, do I trust the auditors and we'll come to the ADR issue of like auditing the auditors, STT and Sequoia, they're kicking the ties. They're doing the real work. And so for me, sitting in the U.S., right, I'm not sitting in Hong Kong as we're talking to your audience today, that gives me a lot of comfort that what I'm seeing and what I'm seeing delivered are real, right? So I think that's important. I forget what your second topic was. That was, so you just covered how SETT and Sequoia blessed them. And then the second one was the bid for VNET. Yeah. So I don't know enough about the private equity firm that bid for them,

if that's a real bid or not. I mean, obviously the ARB guys are discounting it. I do think as well, and maybe this is a good transition to talk about the issues. There's two major issues in, in China today. one is the ADR issue and one is the VIE issue. And I think for the purposes of painting a full picture, we should just mention both really quickly for your audience. I'll start with VIE just because I'm thinking about it. So VIE is a structure that was created in 2000 for the Internet companies. And unfortunately, you may not remember the company that used VIEs the most and the best.

I remember it. I know. Right. So Enron kind of gave the structure a really bad name for a good reason. But the reason, and the VIA stands for variable interest entity, right? And the reason that this exists is that in China, the government comes and says in certain sectors, education and internet, most importantly, you can have zero percent foreign ownership. So what that means is they create a special purpose vehicle. This SPV is based in the Caymans. And you, so if you want to Alibaba or JD.com or 10 cent today, you do not own a proportionate and old stake in a business like you would if we bought AT&T or Verizon. If you own 10 shares of

that, you own whatever the fractional percentage of the company, your minority shareholder where you own. With a VIE, you own a pro rata proportion of the P&L. They pass it through, but in the case of chapter, you have no rights. The most famous kind of story here was Yahoo Japan owned a big piece of Alibaba. And Jack Ma came in and said, no, we're going to pay. I forget what the number was. It was something crazy, like, $6 million or $60 million. It was like a tiny amount.

CEO of Yahoo got fired over that, by the way. Like, it was not a success story. But the point is there's no recourse. And this is the concern with the VIE is like there's no dissenters rights, any of the Delaware stuff that you used to. And just anyone who's listening to, I'm sure most of our listeners aware, but this has been one of the bare cases on China stocks for that has kept people away from China for a long time. What Alibaba, Google Alibaba shortcase, and the first thing that's going to pop up, A, there's

going to be some funkiness with their numbers that people are going to point out. But the main thing is people are going to say, it's a VIE structure. If the insiders there decide that they don't want you to have money, they want that money to go to themselves, they can take it for themselves. And there's really no recourse for that. Right. And so the other, if we're going to do acronyms, the other one that I want to introduce your audience is WFOE, which is wholly foreign-owned enterprise. And that's where the sectors,

data centers are one, where you are allowed foreign ownership. In the case of data centers, you're allowed 50% foreign ownership versus internet zero. And if you look at the actual things that we own, GDS owns its data centers. GDS owns its generators, its turbines, its cooling towers, there's no limit on foreign ownership of those things. So I think that's a really important thing to stress here is this is not Alibaba and JD where we have no proportion of ownership of the underlying assets. Yes, 50% restriction, but we still own something. So that's important. The second topic that's certainly topical is the ADR issue, which is in December of 2020, in the midst of all the saber rattling, the U.S. Congress can't.

what became law something called Beholding Foreign Companies Accountable Act, which is HFCAA. And what that basically said, I don't know why it always reminds me of that JFK quote, who's going to spy on the spies. But the idea is who's going to audit the auditors? So if you want to be listed in the U.S., and by the way, you can go in the SEC and look at the list of companies. It's public. It's only about 30 names right now, but it's going to grow. And you can go in and say, how do we audit the auditor's working papers? If you missed that checklist for three consecutive years, you have to do list. GDS has not filed a annual report yet. It's 20F. That will happen by the end of this month. We fully expect they will be on that list a few days after that, starting in May. So that would start the clock.

It's worth noting just as an asterisk that currently there's talks in June of 2021. So six months later, the Senate passed the accelerating the holding foreign companies accountable act. So H-F-C-A-A, which is to say instead of a three-year put up or shut-up window, it would be a two-year window. That is not past the House, but it's something that should be on the radar of people here. It's an easy political win, right? Like it feels like it's probably nobody's out here saying, hey, we need to give Chinese companies trade in the U.S. We need to give them more rights and more power. Yep. But when you look at kind of what, there's two different things, right? So this is fundamentally a technical issue that's been created by a

geopolitical problem, right? Like, if this was actually technically created by a technical problem, you can fix it very simply. This is geopolitics, right? And you and I can't invest knowing how geopolitics play out. But what I can do is position myself to know that there's a runway or an outlet if this goes sideways. Can I, can I put, so just to clarify, what would happen here is the, the, worst case scenario, and what you're about to talk to is GDS and several other Chinese stocks could be kicked off to NYSE, kicked off the NASDA if they don't meet these requirements to the worst case. And what you're about to say is if that happened for GDS, particularly, there is an out tier. Right. And the out, I mentioned the very beginning is Hong Kong. So when you look at that list

of 30 names, only about 15% of them have a secondary listed. And at the end of 2020, GDS raised $2 billion U.S. huge rates in Hong Kong. And I know that it's easy in our seats to say, oh, maybe that's a secondary market because it's a secondary listing. The onerous process of going public, it's not fun going public in any market, by the way. And in Hong Kong, it is equally painful. So this is as valid of a market. Secondly, the actual transaction process to switch your ADR. So 190 million ADRs, roughly speaking, at IPO for Hong Kong. There was, I think, 20 million of those were in Hong Kong. Today, 35. So another 15 million ADRs have switched over to Hong Kong. And then you say, why? Well, maybe it's people like you and me that send emails at

ridiculous hours. We won't tell the audience, like, we're workers, right? We're grinders. And so maybe there's people who want to be able to trade longer. Maybe there's some structural reason. Maybe there's some parakeasyu trading issues. I don't. The point is you pay de minimis fee. We were talking like $20. It's not a big number. De minimis fee and about 48 hours later, you are up and running. So the risk of delisting for me, for GDS, is making sure my clients can settle it.

I don't worry about there being a fungible security that I can trade versus the other 85% need to start the process. And we mentioned the other two Chinese data centers. We should say Chin Data, CD, for example, is actively talking about when can we start our Hong Kong process because we see the writing on the wall. We need to know, but you need two years a run-up. The reason that, so then the natural question becomes, why wouldn't GDS just drop its U.S. listing today and do Hong Kong today? Chinese data center kind of makes sense. The answer is while it's pretty frictionless, the only real difference is they need to do an IFRS versus GAP reconciliation, and they need to reveal on their directors, certain holdings, take over a lot,

Like much of the countries outside of the U.S. Takeover rules in Hong Kong are much more onerous than they are here. And I mentioned before they're in Singapore, China, and STPs in the middle, and maybe something, I don't know, but maybe something will happen with some shared data centers. That would be a lot harder if their primary listing was Hong Kong. But when I talk about kind of the potential risk, and we've got a lot of work on it because we have clients that own GDS, So we need to make sure that we have something that trades. This seems to me, much like the VIE example with GDS, as a non-issue, more of a headline thing than anything else.

Perfect. So look, I think we, this is the tough thing with you, Randy. We've already hit the hour mark. But, you know, I think we've done a really nice job of covering everything that we talked before this. You and I kind of wanted to hit on this podcast. But I just want to make sure we wrap it all up in a nice bow, right? We've covered a lot of the risks. We've covered the multiple arbitravers.

But, you know, just real quick, why should people, there's China risk here, this is trading cheap, there's big growth, you know, what do you think that people should be thinking about when they're jumping in here to look at it? Why should people be getting excited about the potential to look at the stock and maybe invest in it? And why I was flagging and saying, hey, Andrew, it's time, right? This is the moment. And the answer to that is pretty simple. It seems to me that a lot of the things that have impacted the stock down, and I'm not saying it's going to go back to $150 tomorrow, but I'm saying it's gone too far down, are things that will resolve themselves in the next couple of quarters

or certainly the next couple of years, which is to say the geopolitical risk. COVID at some point will become endemic versus pandemic. The regulatory environment, China, I'm not convinced China doesn't fold on the ADR issue, for example. By the way, if China says we want to have you know, an internet. We want to have capital market access to the rest of the world. You know, China, I think this is important to remember. The worst decade for China was in the 1960s to 1976 before Nixon opened it when they had no friends in the world. That is, and they remember the Communist Party, we forget, while there is an ideology, they're pragmatists, right? And if you look at the lesson of Russia, I pulled my, my quarterly letter.

better here, because this is what I think it's relevant, which is a Chinese confrontation with the West, economic or military will be wildly irrational, but so was Russia's invasion of Ukraine. Tellingly, the Ukraine war led in March to a large-scale capital flight from, dot, dot, dot, China, right? And so, like, we have put all of this onus onto them. And if you view the likelihood of a land grab in China, Taiwan as low, you today now have a very large margin of safety with a proven management team that has executed wonderfully project by project with great backers, including their customers like Alibaba and Tencent and their investors. Like we talked about STT and we talked about Sequoia.

I don't see as I look at the space another company with the potential growth profile that these guys have for the next three to five years. The only other thing I'll add, I thought that was wonderful. The only other thing all that is, this is just me personally, I just love, this is an infrastructure play. I love infrastructure plays like cable and data center stuff where it's big, hard assets that basically a couple people can, but it's really hard to recreate, like you can look and see those assets in the ground, and they get network effects, right? Where a competitor who says, oh, I want to go buy 100 acres right next door to you and build the same thing. Well, guess what? Yours isn't as valuable as mine, because mine already has all the network effects and connections and everything.

there. And I just love those things. And, you know, this is, this plays right into that, right? You mentioned the multiple arbitrage versus what a lot of really sophisticated private equity investors play. You get this big thing and you just know it's going to be there, right? If you believe data usage is going to stay there, there's a big tail one for this thing. And the phrase that comes to mind as you wrap up is God only made so much waterfront. We're on the waterfront. And so I can't tell you third quarter 22 is that, no, what I can tell you is, is if you think China Internet's going to grow over time, this is the play. Perfect, perfect.

Well, we are going to do a separate video real quick right after this. We're going to update Amaris and maybe a little bit of renaletics. But we'll stop the video right here. People can go check out that separate video. But Randy, thank you so much for coming on and talking GDS. And I will stay on and we'll talk in one second. Okay.