Randy Baron's bull case on $GDS (Podcast #104)

Randy Baron makes his third podcast appearance to talk about his thesis on GDS. You can find my notes on GDS here, Randy’s first pod appearance on AMRS here, Randy’s second appearance on RNLX here, and Tubes (the data center book mentioned in the pod) here.

Please follow the podcast on Spotify, iTunes, or most other podcast players, as well as on YouTube if you prefer video! And please be sure to rate / review the podcast if you enjoy it, or share it with someone else who would enjoy it (more listeners is a critical part of the flywheel that keeps this Substack and podcast going!).

Disclaimer: Nothing on this podcast or on this blog is investing or financial advice; please see our full disclaimer here. The transcript below is from a third party transcription service; it’s entirely possible there are some errors in the transcript!

Transcript begins below

Andrew Walker: Hello and 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 Baron. Randy, how's it going?

Randy Baron: Andrew, how are you?

Andrew: Doing good. As I was saying that, I was just thinking like I can't believe Jacob Rubens, the one who introduced us, I think. And I can't believe we formed a friendship. It's the power of the podcast...

Randy: What you refer to as the Yet Another Value Empire, which I always laugh at when you write about that, is really amazing. I have this broader theory that everyone knows Buffett & Munger and the next generation was like [inaudible] and all that. But after that, these guys in their 70s, Lee Cooperman, all of them, there is no great known investor class and what you've created here in a format, even if Jeremy Raper gets disproportionate representation, it's the next generation of up and coming thinkers.

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. That's a pretty brave thing to be the point of the spear. So thanks for letting me come back on again.

Andrew: 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 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. Let me start this podcast after that. Let me start this podcast the way I do every podcast.

The first two disclaimers are to remind everyone, that nothing on this podcast is investing in advice. We'll be talking about a Chinese ADR today. Nothing on this podcast is ever invested in advice but that carries an extra level of risk so people should just be aware of that. And then the second with a pitch for you, my guest… 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 is a super thoughtful guy, and I think this podcast's 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 that'll post later so people can look for that. But the stuff 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 it over to you. What is GDS and why are we so interested in them?

Randy: 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 in a 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. Its EBITDA margins are best in class. And quietly, it's begun expanding to the south and to the east from this kind of position in Chinese 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 involves the political backdrop, the macroeconomic stuff that's happening in Russia. But I thought what will be really useful 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.

Andrew: Yeah, I think that's great. I could use a refresher, to be honest so that's perfect.

Randy: Well, it's interesting. Okay, the first two podcasts you mentioned, one was on synthetic biology, one was on healthcare [inaudible], 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 when you write your charter stuff, it catches my eyes and I know this language.

Andrew: I don't know if I knew that about you, Randy. I'm surprised.

Randy: I don't brag about it. I keep [inaudible]

Andrew: When I was thinking about vinyl tease[?] at 30 last summer, you couldn't have parachuted in and said, "Andrew, come on, I've followed this forever that those guys are [inaudible]

Randy: It makes stuff like Netflix which was reported last night, really kind of fun to watch, like you enough to be dangerous. But I first kind of got introduced to the data center space ten or twelve years ago when a company out of Cincinnati Bell bought series one. And at that stage, none of us kind of knew what data centers were.

So what is a data center? A data center can be thought of as the brain of the internet. The role of these buildings is to process, store, and communicate the data behind the myriad of information services we use every day. All the stuff that the Yet Another Value Empire kind of is about, right? The social media, online collaboration, emails, your streaming video.

You've got to think about these buildings. They've got racks of servers, which deal with the computational logic response to requests 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. And then you have a ton of electricity, which generates a lot of heat. So, you need something to do the cooling often or just a way to offload the heat.

It's useful to think of a data center like a hotel. Data centers make money by leasing power and by leasing space. And to some degree, it's a very virtuous cycle where the ecosystem that it's in determines the price. In other words, the more carriers, the more networks that are present, the more valuable it becomes to have connectivity, or what they call colocation in that data center, and therefore costs more to lease space in the more successful data centers. [crosstalk]

Politically, I often think about… and I know this is a bit macabre, but I think about like terrorism. If Al Qaeda was to strike again, what's the best target? And none of us knows, 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 what we can't give up?

Andrew: Twitter for two seconds.

Randy: This thing right here, right? We can't give up our phones. If the Federalist wants to find you, it used to be they track your cable address. That's actually how they would track at FBI. Now, they track your mobile phone. And so my point is, when you go to these buildings, you've got to imagine these huge, multi football-field-sized warehouses with no branding on the side. You go to Secaucus, you can see them, they're these huge white buildings. And to get into them is tougher than getting into the Pentagon. And when you walk down the bays, they may be 150-200 feet tall. They have big trays up in the sky that have- in the US, it's a little different in China, we'll get to that. But in the US, you've got copper on one tray and above that, you've got fiber.

And because of the legacy of AT&T here in the US ever since Ma Bell[?] broke up in the 80s, you had all the competitive local exchange carriers, the [inaudible] local chain carriers. You had three hundred, four hundred 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 that boot 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.

Andrew: Can I just add one thing in there, Randy? I think that was great. And I think the one thing that I think makes sense, if I remember correctly, is 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 because what's going to happen is Netflix and Google might want to connect their trays.

So if you're a data center that has Google, the Netflix wants to be in you, and again colocation, because they can connect to that. Whereas if you're a small startup data center, you get a little bit of a chicken and egg problem where Netflix says, "Why would I lease 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 US. 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, everyone wants to be at your location, and everyone comes to your location. Am I thinking about that correctly?

Randy: Yes. And I want to take it one step further, which is there's a word called 'hyper scale'. These are the major- you just mentioned a lot of them but add Alibaba to the list, AWS, Microsoft. There are about 20-24 of these hyper-scale players in the world and not just refer to- I think I've used it before. Infinity is a place that's so far away that wouldn't really matter. The idea that such computational power comes with mass.

But what that also neglects is kind of not the hyper-scale players, it's called the micro-scale players. The guys that are looking to come up and nudge into that framework. They're going to want to come to the place where the hyper scalers terminate, and they want to colocate.

The best example that you can give in the US is Netflix, which was on AWS and still is. But over time, it wants to control more of its own destiny. So, all of a sudden it starts doing more and more of its own data centers. I think that's a really revealing revelation in China.

I will come to my history of GDS in a minute, it's a little different because you didn't have all the diversity after the mob[?] outbreak. You didn't have 300 different phone companies that have since been rolled up in the US. 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, CyrusOne, which was the purchased data center from CBB, from Cincinnati Bell, maybe five or six years ago, and no longer has it, did a minority investment into to GDS because GDS at that point was general. Not a true biblical seven-year generation but generational, 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 Unicom, it's a little bit China Mobile. And then there's 50% described as carrier-neutral data centers. In other words, they don't care who the phone company is. GDS is 30% of that other bucket. So, call it 15% of the total China market is GDS.

Andrew: Perfect. I think we've done a nice job framing what a data center is. We can obviously add on to the background on the data center if you want. But if you're ready, we've covered the industry, so let's switch over to GDS in particular and your background with them.

Randy: Okay. So, let me start if you don't mind, with 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 about the US listing. They did their Hong Kong listing at the end of 2020, but for the sake of this history, let's go back to 2016 when the IPO was at $10 in ADS.

And by the way, for your audience, ADS or ADR is American Depository Receipt. It's essentially a U.S. vehicle to own a foreign stock. Very common.

I was laughing when I referred earlier to our friend Jacob, but also Jeremy Raper, all three of us kind of looked for idiosyncratic ideas. I had a phone call from the RBC analyst, he's a friend, John Atkin, the TMT analyst, who said, "Randy, I got one for you to look at."

Every once in a while I get that call, and 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 build traction, but it's a little before them and in the US market.

I went to lunch and I met these guys, the CEO, William, who's the founder, still owns 5 or 6 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 a lamentation. It was pure sadness. Why does the market not realize what we're 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[?] it 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 for Alibaba and Tencent- two-thirds of the revenue where those two hyper-scale players use the term we just introduced.

And what was fascinating was this relationship with Alibaba, which continues through today and is part of the thesis as to 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 internet and China were 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 PnL 3 or 4 years [inaudible] at 50% growth top line with EBITDA margins at that point in the 20% going up to what they are now, 50%.

So, I sat there and I was like, "Guys, have you discovered the company code of geography, which is part of my process, 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, as I said, William, the founder now, 6% owner of this multibillion-dollar enterprise. The market cap for your audience in US dollars, five-and-a-half billion dollars, another 4 or 5 billion in debt, 10 billion enterprise value.

So I sat there, and I went back to my office with my team, we kind of did the work and I was like, "This is amazing, we're going to go and it became a top-three position for us. We bought it at about $9 a share. And then by the end of 2017, had sold it. It had doubled, it was a win, it was great. But as you and I know, in this business, relationships matter. And so even though we were out of the stock at that point, we had maintained a dialogue.

And I remember, I'm out at lunch- this is at the end of July 2018. And all of a sudden, I get the flashing, the bat phone goes on, get to the office type of thing. And there was a new- at that point, a nascent hedge fund called 'Blue Orca' still around, it's a short-selling fund. It 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, in Shanghai.

And at that point, there are 25 Publishing analysts today, but at that point, there were four or five. Frank Louthan, who's the Raymond James guy, was in the unity courtroom in Virginia DC. John Atkin, I mentioned, was in Australia on a plane home. No one in China could respond. And in the period of 3 hours, the stock lost 47% of its value. It lost almost $2 billion. It was one of the best executed and actually billed out [inaudible] the Barron's talked about it. This was 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, in the fifth or sixth page, they had kind of one of those surreptitious shots from the hip[?] of a cage that had wires hanging out of it, basically inferring fraud. This is LinkedIn copy[?] this is a 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've 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 poignant picture for the short.

Anyway, we went back in a real way, in the low 20s, so I could drop 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 of 2021, hit $114 per share. And then when you and I started talking about it last November [crosstalk] that's when that macro, the China regulatory landscape, all the stuff we talk about today had begun to shift.

Randy: 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 a top-three position for us. You know the other two, they'd been podcasted, about 7% of the book.

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

GDS, the stock, you mentioned it's down a lot over the past year, probably 60% or so over the past year, a little bit more from its peak from probably in the October timeframe. The market's down 67%, it's I think it's around 16 times, EBITDA, which isn't crazy for a data center but it's not super cheap. But when you look at the stock, the top three positions for you, what do you see that you think the market is missing? Where the top three positions, you obviously think it's going to generate risk-adjusted alpha. What are you seeing that the market's missing?

Randy: Well, I want to disagree with one sentence you slipped in there. This is very like Andrew Walker, smiley guy, gets it in there. But the one sense 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. I'm going to give you the three most recent public transaction multiples. Because the argument I'm making here is, guys, we know the stuff's going on in China. 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 cap ex[?] as debt, but 26 times EBITDA. CyrusOne, that company we mentioned, the spin-off from CBB, KKR global infrastructure partners, acquiring them for 23 times 2022 EBITDA. CoreSite 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 to, is trading at 26 times today.

I can look at my grid here, you mentioned Equinix. I should just probably bring that up as well. Equinix is trading today at 24 times today. So just to give you some frame of reference, I want to finish framing that out. Hold on one second.

Just going back, the percent change in revenue. This is really relevant because GDS today is trading at under 15 times. Okay? I mentioned QTS acquired year-on-year revenue growth, 12%, 55% EBITDA margins. CyrusOne, 8%, high single digits, revenue growth 48%, EBITDA margins. Coresight 8%. again, the point is top-line growths, 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 of 51%.

Andrew: Look, I agree with you. I guess in my 16x EBITDA, like with Charter, I've been saying since the stock was at 12x EBITDA. I was like, "This is too cheap. Go find me a cable company that's been taken out for less than 15 times EBITDA in the past five years or something." And the stock has gone from 12 to 10 and people will say, "Oh, it's expensive." I say no versus every other deal I can find, it's actually cheap. It might look expensive versus a retailer trading at two times EBITDA.

So my 16x was more in the absolute sense but I am 100% with you. One of the key thesis with data centers and everything, and we'll probably get into this a little bit more as well. These are extremely popular infrastructure or place[?], right? When debt is at 2 or 3 %, go by a data center, you can put tons of debt on it, it's going to grow, and you can double leverage on it. So, I'm definitely with you [crosstalk] just to defend myself.

Randy: Even in this increasing rate environment, on the private side and by the way, the private multiples are even higher than the public multiples, it's called 30 plus. HAP rates right now, last week, 4%. So, the point is you can still do a lot of that and in fact, that other metric that they use for the space is growth-adjusted EBITDA. So it's essentially the same concept of the PEG ratio, which is PE adjusted for growth of EPS over time. And because a lot of companies we just talked about Charter included, you don't use the PE, you use EB to EBITDA because the depreciation in there throws off the math. So, enterprise value divided by EBITDA divided by growth and EBITDA, GDS today is at 11.1 times. The US peers[?], two and a half to four and a half times.

So, 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 in the backlog. So, I can add 70% to the revenue just in converting backlog and churn in a data center. We probably should have led with this, is no[?] because to what we're saying, the network effect before, once you're in, yes, you may need to replace a server; things were out over time, etc. but you don't give up your footprint, you add to it over time. 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 in 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.

The zero COVID policy, you could have a whole political conversation if that's sound or not from a geological perspective. But the idea is, I am of the opinion that we're going to use more and more data over time. And especially as 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 on the third inning of that.

Andrew: Let me just go back to the question. So, again, I'm with you on the devaluation and everything but I just want to go back to the question. Today the stocks are down 60% obviously, you think this is valuable. What do you think the market is most missing when you look at the stock? There are lots of options, a few could be: the market is over discounting the China risk here, there's concern on financing, and we'll probably talk about that later. The market is too concerned about the financing needs of the business, all sorts of other ones. But what do you think your biggest variant perception on the stock versus the market is?

Randy: 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, "It's risks[?] 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, there are three. Chindata's one simple CD, 21Vianet, which we'll talk about how to bid, VNet as 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're at a cyclical low, which happens in infrastructure companies, then you kind of just flavor it a little bit with COVID and maybe add a little bit of geopolitical Russia backing, 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 slowest 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. We can't 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, we're 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. 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 1, 2, 3 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 ideas that are going to compound over time, and I guess this goes back to the amorous [inaudible] You only care about price if you're selling at a given moment. Otherwise, it's a great buying opportunity and you're looking for things almost like a private investor in a public forum, a 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, we get slapped for, "Oh, you're talking your book, or, oh you're pumping a stock."

Today's the 20th of April. I don't care what the price is on the 21st. I may care at the end of the month when I get the mark to mark. I may care at the end of the quarter, honestly. But truthfully, generationally, what do I care? Where is it in 3 to 5 years? That's how I try to have a worldview. I do believe that 3 or 5 years from now, all these issues we should talk about. The ADR delisting, China COVID. I think it gets resolved over time. My broader thesis and this is a long-winded answer to your question [crosstalk] is that A, China has an internet in the future, and B, 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, a great track record, focused on execution, got great relationships with customers. Alibaba loves them and has them built 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 own 32% of the company. I mean, they've clearly said, what we want to have via China data center play. That's all positive. We've got a huge customer base and ways to grow multiple ways, including the fact that they are just now entering Singapore and leaving China.

So, up till this point, in the narrative, there's been a 100% [inaudible] in China's story. And what I'm saying is we're not picking up on- you mentioned Equinix. This could become the next Equinix but it's just coming from a different geography, so it gets dismissed because it's not in the occidental world.

Andrew: That's perfect. That's great. This is a smaller point in the overall thesis, but it was something that was curious to me. 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. 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 Tik-Tok. Right now, I think it's two countries. It's Singapore and one other country that's escaping me, but they're in the process of really expanding outside of China for the first time. I don't think it makes a huge difference to the downside of the story but it could impact the upside, where I look and then I say, hey, did they only really win in China because they had that protection and 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 a lot of Chinese companies doing that in the past?

Randy: Okay. A couple of things to unpack there. It's not just China that stops tech companies from coming in.

Andrew: That's a good point.

Randy: There's a company in Russia called Yandex, which is effectively the Google of Russia, and beat Uber out because Uber was allowed to come in. Another podcast guest of yours, Adam Lindsey, came on to TripAdvisor. He went through the bond indentures, it turns out that there was in Yandex paper, if you didn't list for five days, it was never disclosed in the [inaudible] 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 has happened in the world. So, I don't want us to say, this is just China. There are other countries that don't allow tech companies in.

To your geographic point, let me get there but let's start with where is GDS today? GDS is in only tier-one cities. About 65% of the country's- not just the U.S. 65% of the country's data centers are on the east coast. So just kind of coming down the east coast, you've got the ANZ River Delta, that's Beijing, basically. That's one. Shanghai is two. The West Pearl River Corridor, so I think Hong Kong and Macau in the South. That's the third major area. And by the way, when I say those cities, they're also in 150 to 200 kilometers so about 90 miles in radius. 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 Chanyu[?] and Chonqing, which are 20 to 30 million dollars population centers. And that's kind of where all these tier-one cities are at. What's changed in the story, since I first invested five years ago, is today, they're expanding to Singapore. So, if you look at the map of China, you go south, there's Thailand, Vietnam. I'm working my way down. Malaysia, which is one of the countries therein. And Singapore, the little dot island there, the size of Chicago. And then Indonesia, the island archipelago below it.

Singapore, is a small island, with not a lot of space, is a 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 Alaska, they're coming and you can see it from Singapore across a bridge or across a ferry ride 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.

So, Alibaba and the hyper-scale 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, doesn't murder Georgia Joint Venture. These guys are wonderful capital allocators, and we'll get into the economics of their per unit data center.

I think Singapore is the canary in the coal mine, and 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 on 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 Singapore. So, my point is you will find the next global city in the future but it's not going to be Equinix. It's going to be someone that has a relationship with the local cloud players like GDS.

Andrew: Let me ask you a quick question. So, my experience with this… I was looking at a company internet a couple of 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 if I was somebody who hadn't looked at data centers before, I think my question would be, it's all the internet, right?

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 and 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't I build 1000 miles from a tier-one city?" Build a data center and we'll just lay fiber connecting it from 1000 miles to the city, and it'd be a lot cheaper because it's 1000 miles from the city, land its much cheaper, and everything. Why is being able to build and why does being able to geographically expand like this? Why does that matter?

Randy: 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 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 [inaudible]

Andrew: The book I read, by the way, just for anyone who's interested, is called, 'Tubes: A Journey to the Center of the Internet.' I'll include a link in the show notes if anyone's interested. Sorry about that.

Randy: But I think you've touched on a really, really important thing, which is latency matters, and specifically low latency. In our business, you would say high-frequency trading. The reason the data center is in Secaucus is that Wall Street is a crow[?] flies, whatever that is, 5 miles, 6 miles. And so 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 talked about the environmental issues of the world. Well, guess what? China uses coal, and the price of thermal coal was up at one point 40% last year. I think it closed the year up to 20%. But the cost to generate power was totally skewed from what all these companies need. And as you go more and more towards cities… I remember before the Olympics, they shut down all the factories because they didn't want pollution. Well, the reason is, that they're burning coal.

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't you just build?" You can't just show up in China and build a data center without power. These are massive, massive warehouses. Maybe instead of the Cummins or Caterpillar equipment, it's the Chinese equivalent 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 then like I said in Singapore, total moratorium. So, I think GDS is in a pretty good place in that regard.

And I want to also mention it because I do think… I feel like every time we do one of these, we talk about social responsibility, investing or environmental social governance, 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 this. I want to totally tip my hat. I think 80% of their global number is renewable. That's wonderful. Others like DLR, 20 something percent.

So GDS as I'm posing[?] it within the peer group, understands, and 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 stay on the vision, they totally see it. They say, "We want to do what's right for the world..." If you're an internet company, and this is not specific to China, although it is the case here. 60 to 70 percent of your power usage as an internet company, is your data center. So, if you can cut that to be carbon neutral, you've done a good thing for the planet, and I don't think that should be buried.

Andrew: And it's not just a good thing for the planet but you're seeing it right now, 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, domestically, it's way worse in Europe. It would be seven or eight dollars per MMBtu. And if you're a data center and your power is getting supplied by a nat gas bank, guess what? Your power costs just tripled, whereas solar, yeah, it only runs during the day but a lot of the renewables, at least once you do that fixed costs of getting them installed, you're no longer kind of subject [crosstalk] to the vacillations in power prices.

So, 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 cost and have good visibility into getting power.

Randy: 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 exploration companies in oil or gas. It's been a total win since Russia invaded Ukraine. But that's, for me, a very conscious decision. 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 woo-hoo, SRI guy." That's not my North Star but just the idea that you want to back ideas that are constructive and you wanted 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.

Andrew: That's generally what I do, though. I do have 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 [inaudible] [crosstalk] who were there"

Randy: And you can refer to that podcast. I listened to that one too.

Andrew: Tim's great. I love Tim. So GDS, another thing, M&A, wants to grow. And one of the ways they want to grow is M&A. I want to talk about M&A in two forms: A, is there M&A 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 a creative 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 both almost [inaudible] callous for what's happened recently with them, and it's a source of concern.

Randy: Okay. So, yes, there is M&A, that's very simple. That's part of their strategy. 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 next 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 summer, whatever it was, what people don't realize is that two things happened 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 were 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' 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 will say, "Well, you're getting 20s a month, times in the public market. We want you to buy us for a third." versus these guys who seek to have an IRR, and they execute it, by the way, all the time. Their goal is between 10 and 13% 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 EBITDA 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. So, you get the money, project finance, they're all ring-fenced. The banks structure it in this fifteen-year contract so that the payback period because they understand the cash flows of data centers takes you two years to build.

So, the payments are back-weighted and you basically borrow 60% on the data center. You have a guaranteed customer because Alibaba has come to you and said, "We want to be in this geography." It's a good paper if you will. And they go to the banks and 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 is about three to three and a half times. And that's kind of where I look at it on a constructive basis. So like, if this was a pro forma, all the build up[?] 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- Are you [inaudible]?" It is entirely an equity holder conversation.

Andrew: It's one of my favorite catalysts, but it's one of the things that I've seen- so many companies that basically do something similar to what GDS does. They build big things that cost tens, hundreds of millions of dollars, it takes two or three years to come online. You've got this huge asset but I love it because they screen[?] really poorly because you know when that asset is about to turn on, it's 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 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 colocate and connect to Alibaba whereas, if you and I started Randy and Andrew's data center business, 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 [crosstalk] [inaudible]

Randy: And Alibaba today- because this is important. About 40% of the data center or project, footprint, whatever you want to call it, 25% of revenue, 10 cent of the other big customer, I think about 25% of revenue. So it's important to note. But remember, when I said four, five years, when I first looked at it, those two were 65 to 70 percent. 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 called 30% and the financial players are about 20%.

And what's happening there is essentially... we all know the Chinese government's been cracking down, regulating 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 Randy data center is, you and I in our little closets with servers. We wouldn't have a huge footprint, not yet, but it'd be in 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 mom and pops in the consolidations into the bigger players, like GDS. So, that's actually a virtuous wind behind them.

It's also interesting because the Chinese banks have made a decision that they want to have more things instead of a 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 VNnet and some of the comparative stuff. But what's fascinating is that both GDS and 21Vianet 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." 21Vianet 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. But what that means is they've basically turned their back on the Chinese cloud players and now they're playing catch up because people say, "Well, why is the multiple- EBITDA's 50%, unsolicited bid premium, why is the multiple for VAET 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. Alibaba's there, everyone wants to be there, and so it's a very constructive process versus infill, which you can do great on project financing and project building, but it's not the headline. It's very much as filling in the punctuation where it needs to be.

Andrew: You mentioned 21Vianet. And I think I said earlier, there are two financial things here that I think are like mini catalysts or are mini reaffirmation of the business. And one again, you mentioned it, Sequoia China put in a pretty big check into GDS at a very, very low cost I think it's 0.5% convertible bonds that can-

Randy: The 25 basis point, 0.25.

Andrew: Yeah, 25 basis points in convertible debt, which gives them cash to go play with now that multiples have come down. Sequoia China is a super regular firm. It kind of gives them more blessing right on top of everything else we talked about. So that's number one. And then number two, probably a little more interesting, timely, is VNet got an offer a couple of weeks ago to buy the whole company for $8 per share. This was like right at the nadir of the Russia invasion, everything Russia, China just getting sold indiscriminately. I think the stock was 450 or 4. 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 the stock's been hammered very much like GDS over the past year.

So, I guess the question is, do you read anything into the VNet offer? Do you read anything into the Sequoia China convertible financing? And we'll build out from there.

Randy: Two different topics. I think you're absolutely right. The Sequoia China is a feather in the cap. I think STT, the Singapore technologies Telemedia, just for people that aren't as familiar with this region, that's the TMT investment vehicle of Temasek Holdings. Temasek Holdings is a Singapore state-owned company that owns the airline and owns everything. It was founded in the mid-70s.

They first invested in GDS back in 2014 with a 40% stake and did a follow-up listing at $65 in ADS in 2020. And then like you just said, they did the $620 million convert note with Sequoia. The point is, for anyone that says, "How can we trust the numbers?" Our job, you and me, Andrew is to determine who can be a good allocator of capital, and the only way you know that is with time.

I've seen these guys now for 6 or 7 years allocating capital really, really well. So, I don't worry about that. But then if I 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 are kicking the tires. They're doing the real work. And so for me, sitting in the U.S. 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 is real. So, I think that's important. I forget what your second topic was.

Andrew: So, you just covered how STT and Sequoia [inaudible] and then the second one was the bid for VNet.

Randy: 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 Arab guys are discounting it. I do think as well maybe this is a good transition to talk about the issues.

There are two major issues in China today. One is the ADR issue and one is the VIE issue. I think, for the purposes of painting a full picture, we should just mention both really quickly for your [crosstalk] [inaudible]

I'll start with VIE just because I'm thinking about it. VIE is a structure that was created in 2000 for the internet companies. Unfortunately, you may not remember the company that used VIEs the most and the best.

Andrew: I remember it. I know.

Randy: Right. So Enron kind of gave the structure a really bad name for a good reason. VIE stands for Variable Interest Entity, and the reason that this exists is that in China, the government comes and says in certain sectors, education and the internet, most importantly, you can have 0% foreign ownership. So what that means is they create a special purpose vehicle, this SPV is based in the Caymans.

So, if you own Alibaba or jd.com or Tencent today, you do not own a proportional stake in the 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 is. You're a minority shareholder but you own.

With the VIE, you own a pro-rata proportion of the PnL that passes through. But in the case of Charter[?] 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 going crazy, like 6 million or 60 million. It was like a tiny amount.

CEO of Yahoo got fired over that, by the way. It was not a success story. But the point is there's no recourse. And this is the concern with the VIE, there's no dissenter's rights, any of the Delaware stuff that you're used to.

Andrew: Anyone who's listening… I'm sure most of our listeners are aware but this has been one of the bear cases on China stocks that has kept people away from China for a long time. Google, Alibaba [inaudible] case, 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.

Randy: Right. If we're going to do acronyms, the other one that I want to introduce to 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 the internet, zero.

And if you look at the actual things that we own, GDS owns its data centers, GDS owns its generators, its turbines, and 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 proportionate 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 2020, in the midst of all of the saber-rattling, the U.S. Congress came [inaudible] something called, the 'Holding Foreign Companies Accountable Act' which is HFCAA. 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., you can go on the SEC and look at the list of companies. It's public. There are only about thirty names right now, but it's going to grow. You can go in and say, "How do we audit the auditors working papers?" If you miss that checklist for three consecutive years, you have to delist.

GDS has not filed its annual report yet. It's the 20th. 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 are talks in June of 2021, so six months later, the Senate passed the accelerating, the Holding Foreign Companies Accountable Act. So, HFCAA, which is to say, instead of a three-year put up or shut up window, it would be a two-year window. That has not passed the House but it's something that should be on the radar of people here. It's an easy political win. It feels like it's probably [inaudible]

Andrew: 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.

Randy: There are two different things. This is fundamentally a technical issue that's been created by a geopolitical problem. If this was an actual technical issue created by a technical problem, you could fix it very simply. This is geopolitics, 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.

Andrew: Just to clarify, the worst case scenario and what you're about to talk to is, GDS and several other Chinese stocks could be kicked off the NYSC, kicked off the NASDAQ if they don't meet these requirements in the worst case. And what you're about to say is, if that happened for GDS, particularly, there is an out here.

Randy: Right. And the out I mentioned at the very beginning is Hong Kong. So when you look at that list of thirty names, only about 15% of them have a secondary listing. At the end of 2020, GDS raised $2 billion dollars U.S., a huge raise 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 ADR is 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 sent emails at ridiculous hours, we won't tell the audience. We're worse, we're grinders. And so maybe there are people that want to be able to trade longer. Maybe there are some structural reasons, maybe there are some [inaudible] with trading issues. The point is, you pay de minimis fee. We were talking like $20. It's not a big number. De minimis fee, in 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, Chindata, CD, for example, is actively talking about when we can start a Hong Kong process because we see the writing on the wall. We need to know but you need two years to run them.

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 GAAP reconciliation and they need to reveal to their directors, certain holdings.

Much of the countries outside of the U.S. takeover rules in Hong Kong are much more onerous than they are here. As I mentioned before, they're in Singapore, China, and STTs. 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 the potential risk and we've done a lot of work on it because we have clients who don't own GDS. 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.

Andrew: This is the tough thing with you, Randy. We've already hit the hour mark. 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. We've covered a lot of the risks, we've covered the multiple arbitrages.

Just real quick. There's China risk here, this is trading sheet, and there's big growth. What do you think that people should be thinking about when they're jumping in here to look at it? Why should people get excited about the potential to look at the stock and maybe invest in it?

Randy: And why I was flagging and saying, "Hey, Andrew, it's time. This is the moment." 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 $115 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 risks. COVID, at some point, will become endemic versus pandemic. The regulatory environment. I'm not convinced China doesn't fold on the ADR issue, for example.

By the way, if China says, "We want to have capital market access to the rest of the world…" I think this is important to remember. The worst decade for China was in the 1960 to 1976 before Nixon opened it when they had no friends in the world, and they remember. The Communist Party, we forget why there is an ideology, they're pragmatists.

And if you look at the lesson of Russia… I pulled my quarterly letter here because 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… China.

And so we have put all of this onus onto them. And if you view the likelihood of a land grab in China of Taiwan as low, 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.

Andrew: I thought that was wonderful. The only thing I'll add is- this is just me, personally. I love infrastructure plays like cable and data centers stuff where it's big hard assets that basically, a couple of 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 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. This plays right into that.

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. If you believe data usage is going to stay there, there's a big tailwind for this thing.

Randy: And the phrase that comes to mind as you wrap up is 'God only made so much waterfront.' And we're on a waterfront. And so I can't tell you the third quarter 22 is that... No. What I can tell you is if you think China's internet is going to grow over time, this is the play.

Andrew: Perfect. Well, we are going to do a separate video real quick right after this. We're going to update Emerson[?] and maybe a little bit of re-analytics 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 [crosstalk] and you and I will stay on and talk in one second.

Randy: Okay.

[END]