Hello ladies and gentlemen. Thank you for standing by for GDS Holdings Limited's Second Quarter 2020 Earnings Conference Call. At this time, all participants are in listen-only mode. After management's prepared remarks, there will be a question-and-answer session. Today’s conference call is being recorded. I will now turn the call over to your host, Ms. Laura Chen, Head of Investor Relations for the Company. Please go ahead, Laura. Thank you.

Thank you. Hello, everyone. Welcome to the 2Q20 earnings conference call of GDS Holdings Limited. The Company's results were issued via newswire services earlier today and are posted online. A summary presentation, which we refer to during this conference call, can be viewed and downloaded from our IR website at investors.gds-services.com. Leading today's call is Mr. William Huang, GDS's Founder, Chairman, and CEO, who will provide an overview of our business strategy and performance. Mr. Dan Newman, GDS's CFO, will then review the financial and operating results. Ms. Jamie Khoo, our COO, is also available to answer questions. Before we continue, please note that today's discussion will contain forward-looking statements made under the Safe Harbor provisions of the US Private Securities Litigation Reform Act of 1995. Forward-looking statements involve inherent risks and uncertainties. As such, the Company's results may be materially different from the views expressed today. Further information regarding these and other risks and uncertainties is included in the Company's prospectus as filed with the US SEC. The Company does not assume any obligation to update any forward-looking statements except as required under applicable law. Please also note that GDS earnings press release and this conference call include discussions of unaudited GAAP financial information as well as unaudited non-GAAP financial measures. GDS conference -- GDS press release contains a reconciliation of the unaudited non-GAAP measures to the unaudited most directly comparable GAAP measures. I will now turn the call over to GDS's Founder, Chairman, and CEO, William Huang. Please go ahead, William.

Hello everyone. This is William. Thank you for joining us on today’s call. I am pleased to report that we had another very strong quarter. We achieved record organic sales, with over 26,000 square meters of 60 megawatt of new customer commitments. We stepped up our development activities. We now have 17 data centers under construction, our largest ever. Area utilized increased by over 14,000 square meters, nearly double the move-in for the prior quarter. We grew adjusted EBITDA by 48% year-on-year and our EBITDA margin crossed 47% for the first time. The operating environment in China is now almost back to normal. And despite the escalating tensions between China and the U.S., we see no adverse impact on our business. Turning to our sales achievement on Slide 4. In the first half of 2020, our organic sales totaled over 48,000 square meters. Our original target for the full year was 80,000 square meters organic. But we are clearly on track for a higher number. As things stand, I am confident that we can deliver over 90,000 square meters organic in FY20, which is a big step up from last year. Demand has clearly gone to a higher level. This is not a temporary blip. It’s the continuation of a strong digitalization trend. China is already far advanced in some areas of the digital economy. But there are many other areas which are just getting started. New technologies – such as 5G, AI, blockchain, IoT, and digital currency – can be demand multipliers. Chinese government policy is giving a strong push. Leading Chinese companies are increasing their focus on the domestic market. For all of these reasons, we believe that demand will be sustained long into the future. Turning to Slide 5. In 2Q20, we obtained 4 organic hyperscale orders.…

Excuse me presenters, the line for our speaker is disconnected. Please take over Dan Newman. Excuse me presenters, the line for Mr. William has disconnected. The other presenter can please take a word, meanwhile I will connect them. Thank you.

Okay. Sorry. The lines are just -- I come back again. Okay. Sorry.

We have the speakers back.

It's okay? Okay. Sorry. So let's -- turning to Slide 15. Lastly, on Slide 15, I would like to say a few words about our -- about what makes GDS fundamentally different from other players in China. As of today, GDS has nearly 60 interconnected data centers, clustered in strategic locations in all of China’s major economic, financial, and network hubs. Within these data centers, we host all the major public clouds, which are accessible over all the major telecom networks. The scale of our facilities, expansion capacity, and market presence, together with software-defined connectivity and a multi-cloud ecosystem, add up to a platform which is unique and far ahead of the pack. There is obviously heightened interest in the data center opportunity in China. But from a competitive perspective, nothing has changed for us. Our customers are looking for a total-solution provider to address all their needs in an integrated way. This is exactly what we offer. We have established our market position over many years and are clearly differentiated by our value proposition. We are confident that we will continue to build on our competitive advantages and further extend our leadership. With that, I will hand over to Dan for the financial and operating review.

Thank you, William. Starting on Slide 18, where we strip out the contribution from equipment sales and the effect of FX changes. In 2Q20, our service revenue grew by 8.3%, Underlying Adjusted NOI grew by 8.5%, and Underlying Adjusted EBITDA grew by 9.1% quarter over quarter. Our Underlying Adjusted EBITDA margin was 47.8%. Turning to Slide 19. Service revenue growth is driven mainly by delivery of the committed backlog. Net additional area utilized during 2Q20 was 29,324 square meters, including organic move-in of 14,336 square meters. On the last earnings call, I said that we expected organic move-in during 2Q of 10,000 to 11,000 square meters. The recovery in 2Q exceeded our expectations. We are now at a level of move-in which we expect to sustain over the next two quarters. Our MSR per square meter was down by 3% quarter-over-quarter. But if we exclude the revenue and area utilized for BJ10/11/12, which closed 25 days before the quarter end, the MSR per square meter was down by 1%. We expect the MSR per square meter to remain at a similar level in the second half of the year. Slides 20 and 21 show the quarterly margin trends. Due to seasonally high power consumption, utility cost was 1.8 percentage points higher than 1Q20. Nonetheless, we were able to sustain our Adjusted NOI margin. Our adjusted EBITDA margin improved much faster than expected during the first half of the year. Some of this was due to government concessions and reduced corporate expenses. In the second of the year, the government concessions will be less and, with the recovery, we expect a step up in our corporate activities. Taking all of this in to account, we expect our adjusted EBITDA margin to remain around 47% in the second half of the year.…

[Operator Instructions] Thank you. We have the first question from the line of Jon Atkin. Please go ahead.

Thanks very much, I wanted to ask first about the slide that shows a lot of your hyperscale orders, slide 5. It would appear that a lot of the demand is in North China and I wondered if you would have any kind of comment about where you expect most of the demand in the market to come from. Is a lot of it also happening elsewhere, but it just happens to be that the wins that you got were weighted towards the North? And then maybe related to that, is the overall pace of demand notably different now, market demands, not just your share compared to say 6 or 12 months ago? Thanks.

Hi Jon, it's Dan here. I'll take the first question. I think third-party market research gives some very different indications, but by our estimate, I think Beijing and Shanghai together account for around 60% of the market opportunity in Tier 1 markets. If you look at the websites of the leading cloud players in China, you'll see that they have most of their availability zones and most of their cloud plots in both of those two places. So, I don't think it's that surprising given that we try to be strategically positioned across where the market opportunity is that Beijing should be such a significant part of our new business. It also so happens that the Beijing market is probably the most constrained in China, and our early move, our first move to Langfang on the scale that we set up there has been very instrumental, enabling us to win so much business. We've significantly increased our market share in Beijing. To be clear, when I talk about Beijing and Shanghai, that would include the surrounding areas. Once again, if you look at the cloud websites, you'll see that there are regions defined around Beijing and Shanghai which include availability zones which could be on the outskirts of those cities. So, places like Langfang, Kunshan, and Changshu in relation to Beijing and Shanghai are within that same low latency to qualify as being within the same region. A place like Jiangbei or Huailai further away from Beijing, they are in a different region. John can you please repeat the second part of your question?

Yeah, I was just interested in the pace of demand, the volume of demand in the market. Is that similar to what it was 6 or 12 months ago, or has that picked up, market demand for data center space?

William, would you like to answer that? Unfortunately, today, William is in Shanghai and I'm in Hong Kong, so a little more difficult to coordinate.

I cannot describe the exact number of the market size right now, because it always changes. But the change is a good thing, because we can see the trend compared with 12 months ago. Obviously in the next 3 or 5 years, the demand is definitely accelerating. That's why I think it's very natural we still catch up with this trend. Compared with last year, we stepped up. The new normal is 700 square meters per year by organic. So, this year looks like we can definitely deliver over 90,000 square meter new sales. That's by organic. It's obvious.

My last question, any update on Hong Kong construction, market demands and what you're expecting there from that project.

We bought two pieces of land in the last 12 months, and we already started to construct. Everything is on schedule right now. The first data center we plan to launch by the middle of 2022. Another one may be eight months later. This is our current schedule. It looks like the demand is very active in the Hong Kong market, especially from our customer installed base. We are not expecting any other new customers in Hong Kong. Definitely we will get anchor customers from our existing customers right now. We hope and we're confident to get most of the commitments before our first data center official launch.

Thank you very much.

Thank you. We have our next question from the line of Yang Liu from Morgan Stanley. Please go ahead.

Thank you for the opportunity to ask questions. I ask two questions here. The first one is, this will be a big upward addition for your CapEx which is even bigger than your sales upward revision. Could management elaborate more about where the biggest delta comes from, especially given the CapEx increase is pretty big? The second question is, could management update in terms of the investment return profile in places like Langfang and Changshu, et cetera, especially given the supplies there are growing very fast? Could management update in terms of the overall at the level of [Inaudible] places? Thank you.

Yes. Yang, it's Dan here. The increase in CapEx, I itemized it in terms of land and buildings acquisitions and organic growth. The land and buildings, to some extent that is pulling forward CapEx, because we're paying for large sites seen in case of Pujiang, Changshu and so on, where development is scheduled to take place over 2 or 3 years, but we incurred the cost of acquiring those sites upfront. The same also applies to some extent to the acquisitions. For example, Beijing 13, we bought out a shareholder from a project and injected capital to take a controlling interest in the land. But that would be a 2021 or even 2022 project in terms of when it comes into service. I don't think you can look at our CapEx exactly in relation to the current run rate of sales. What you can look at is the pre-commitment rates for the projects that we're initiating, which is consistently in that 60% to 70% range. You can also look at the pipeline that we build up, which is now 350,000 square meters, equal to more than three years' new business at the current run rate. I think we're very happy to allocate capital for that purpose. Your question about customer return profile, this is key because this is what we target. It's not the selling price, it's the return on investment. What we've seen so far, we've done a lot of business in Langfang. We've got a large order in Changshu. The return on those edge of town sites is well up to the level of return in the downtown or urban areas. The development cost, the unit CapEx is lower. Hard to be precise, but let's say maybe around 10% lower, not due to the cost of real estate but more due to the efficiency of building greenfield, building large scale on a single site and spreading infrastructure cost across more capacity. That means that the selling price is also correspondingly lower, but from all the deals we've done so far, I can assure you that the project returns are well up to our historic levels.

Thank you.

Thank you. We have our next question from the line of Colby Synesael. Please go ahead.

Okay. Thank you. Dan, I think in your comments you said that net installs should be similar in the third and fourth quarter to what we just saw in the second quarter, and I think that was around 14,000 organic. I'm just curious why it wouldn't even be higher, particularly in the fourth quarter, given the amount of capacity you have coming online. Could you just remind us how much capacity you have coming online in the third and the fourth quarter to help us there? Also, I think in William's comment he mentioned that you guys got 30% of the incremental demand from your top two customers in the quarter. I wondered if you could unpack a little bit. I don't know if I exactly understand what that meant, and maybe giving a sense of market share. One of the questions I always get from investors who might not be as familiar with the Chinese market is, what is the market share of somebody like GDS in maybe third-party data centers et cetera? Anything on that could be helpful too. Thank you.

Sure. So on the net installs, we were expecting this year to be a series of step-ups. One factor behind that was planning new data center capacity to come into service in each quarter. That was according to the original development timelines. What has happened is that the first quarter was around 2,000 or 3,000 square meters lower move-in than we had in our budget. In second quarter I said we were expecting 10,000 to 11,000. It came out at 14,000. Now I'm indicating that around 14,000 or 15,000 should be the case in the third and fourth quarter. One methodology that you can use to predict this, although it's not totally reliable, is to look at the backlog for area and service at the beginning of the quarter. At the beginning of the third quarter, it was 57,000 square meters. Typically, we deliver 20% to 25% of that number in a particular quarter, so 25% of that number would be around 14,000. That's the logic there. If you go back to the second quarter and look to that metric, you'll see that we delivered 30% of the backlog there in service that was existing at the beginning of the second quarter. In terms of the timing of new data centers coming into service, there's 49,000 square meters of new capacity scheduled to come into service during the second half of the year. I didn't break it down by quarter, but I can tell you that a significant majority of that is going to be in the fourth quarter. Normally, we don't see much move-in in the same quarter as when a data center comes into service, so that's arising too late to really drive increase in installs this year, but it's a very good base for when we start to look at next year. Your question about market share, we have our own internal market research. I believe it's a lot more reliable than what we see from third parties. There are issues with market research about whether it actually maps the data center business or whether it includes a lot of other telecom value-added services. We think there's a lot of double-counting, because some carrier neutrals are reselling carrier data centers and some carriers are reselling carrier neutral data centers. Also, I think the data which comes from the telecom carriers typically, they would include a lot of network revenue, so connectivity revenue for data centers in their numbers. Yes, we will endeavor, in the near future, to publish some market research. I think it will show that our market share on the carrier neutral side is in the 20s. Maybe in the Tier 1 markets, it's a little higher than that. As we indicated, with the very largest hyperscale customers, it's more like 30%.

Thank you.

Thank you. We have our next question from the line of Gokul Hariharan. Please go ahead.

Thanks for taking my question. First question, could you talk a little bit about the engagements that you're having with some of the new customers that you've added? Especially ByteDance and Pinduoduo? Obviously, ByteDance, the type of their business, data center needs, compute needs, et cetera. are quite sizeable, comparable to your two biggest customers, so how do you see that relationship progress over the next couple of years? Second, and maybe related question as well, I think, William, you mentioned that we have stepped up from that 70K sales or area committed organic basis over the last couple of years, to definitely 90K this year. Over and above that you're also making some acquisitions. So, if we think about some of these new customers that you've added, plus growth from your existing customers, as well, is that 90K number per year the new normal, or we should actually think that it could be higher than that number, as well, from an organic sales? Acquisition, obviously, or if you can talk about maybe what could the full year number, including acquisitions, be, as we look into the next couple of years, given some of these new customers and growth in your existing customer base.

So, Gokul, this is William. I will take your first question on how we engage the new customer. I think that we are working very hard. It's our key strategy, since our IPO, how to diversify our customer. But in the last couple of years, the leading demands from a couple of big cloud, so that's what we already catched up. But what we realized is currently, in the last couple of years, the new internet giant, grows so fast. And we start to engage them there, at least 2 years ago. But unfortunately, we didn't get any opportunity in the past 2 years, but we still keep engaged our customer. And now, I think our customer start to set up their new criteria. When it was a baby company, they maybe had the different criteria to select a vendor, but now, I think that they're criteria are more close to their standard as a giant. So, I think that we have the ability. We have the chance to get the deal. And so far, I think our customers still focus on to deploy their data center needs, requirement, in the Tier 1 market, plus remote. But everybody knows GDS, our core asset, and our focus, is the Tier 1 market. So, I think this is the first time for our customer to deploy their servers in the Tier 1 market, which we think is our strength and our capability. So that's why we can get the deal. Looking forward, I think those customer will, in the future, we believe will grow very fast and they will be our next generation of the hyperscale contributors. And as I just mentioned, their number, if you look at what the total server they procured, in this year and the next couple of years, the level of the server procurement is start to close to their traditional cloud player. So, we are happy to see that. That means GDS anchor customer, in the future, will be more diversified. This is our key strategy and we will continue on that. This is the first question.

Question is on 90,000.

Oh. Okay, 90,000. We are happy that we will continue to catch up with the acceleration of the market demand. We don't want to put a big number right now. I think the 90K, this year, obviously, we cannot achieve. But I think next year, in terms of 90K, I think we are confident, but maybe we can do more, but we are not just pursue the number. We pursue the growth quality. So, we will see. I say next year, 90,000 square meters. I believe we can continue, but we don't want to put a big number right now for the next couple of years.

Okay. Got it. Just one follow-up, are you seeing some of these customers, you mentioned that they are building Tier 1 data centers for the first time. Do you feel that strategy for some of these customers is still to keep their existing data center providers in remote sites and embrace some of the newer data center providers, like you? Or established and Tier 1 data center providers like you in Tier 1 cities? Or do you feel like they're also changing the strategy in terms of pulling demand from public cloud and starting to move towards their own compute and other services hosted on their own data centers?

I'd say it's not conflict. I don't know what kind of application they deploy in a direct to our data center. But we still see the channel. They still use the cloud vary in a significant way. They also have another purpose. I believe they have a different purpose to deploy some dedicated IT infrastructure by themselves. I think this is also the chance that in my view, they use the public cloud, still grows, and they also start build their own hyper cloud strategy right now. So, we have benefit on both.

Okay.

In terms of remote, everybody know remote site is one of our new product, just what we did with GIC. It's not our core business, but for some strategic reason, we still keep doing it, but it's not our core business. But I would like to say, we are ready to do more and what we can tell is our customer will give us more business. We have more opportunity on this part.

Understood. Thank you.

Thank you. We have the next question from the line of Tina Hou from Goldman Sachs. Please go ahead.

Hi. Thank you very much, management, for taking my questions. I have two. The first one is for Dan, could you help me understand, because we're raising our guidance in terms of our floor area sales, in terms of organic, from 80,000 to 90,000square meters, however, we're keeping our annual revenue guidance unchanged. So, is this because the acquisition side of things is lower than what we were expecting previously, or are we just being more conservative on this side? Then the second question is, we wanted to understand, in terms of the big cloud customers' data center vendor strategy, do you see any difference between, say, Ali Cloud or Tencent Cloud? Normally, how many data center vendors do they usually pick? Thank you.

Tina, it's Dan here. On your first question, there's a cycle from sales, most of which is pre-commitment during the time the data center is under construction, and when the data center comes into service, and then over maybe 18 months or even 24 months, there's a move in. So, our sales is a great lead indicator for revenue growth in the future, but it wouldn't affect revenue in the current year. In fact, when we give guidance, if we look forward 12 months, at any time, pretty much all the revenue growth, looking 12 months forward, would come from contracts which were already signed, prior to the beginning of the year. So, talking about raising revenue guidance, if I was giving revenue guidance for 2022, maybe I'd be raising that on account of the higher sales. That's how it works. William, would you like to answer the question about different strategies of the big cloud customers?

Yes, I think as we talked about that we already build our platform in all the key markets, which we have very significant advantage, compared with other. So, I think this is very important for our customer. I think our customers, the criteria for them is continue high visibility of the resource in each region in your platform, in all core location. These are very important. And continue, of course, the operation skill is also very important for them and the cost is also in the key criteria. Typically, I think the historically, the top customer, they use a lot of the different data center service provider in historically. But in the recent couple of years, they shrink the name, and they narrowed the vendor list, after many years' business relationship. So, I think typically, the major customers, they will have three or four service vendors. So, this is a current situation we faced. And we used to talk about it, we signed the strategical vendor agreement with our top three customers so far. So I think we believe GDS is the major vendor in terms of the top three key vendors right. Is that answer your question?

Yes. Thank you very much.

Thank you. [Operator Instructions] We have the next question from the line of Frank Louthan from Raymond James. Please go ahead.

Great. Thank you very much. Can you comment on the trend in MSR per square meter declined a little bit more this quarter than we'd modelled, just what are the thoughts and the trends on that for the rest of the year? My second question, can you give us an idea just for those of us not in the region as often about some of the current steps that the government has taken with regard to any future outbreaks of COVID-19 and how has that policy has evolved? How do you think going forward reactions to the virus could impact your business, either ability to construct or get labor employees at different locations and so forth? Just talk to us a little bit about what the current policies are and are they better or worse than they've been over the last six months or so, thanks?

Yes. Hi, Frank. Dan here. You know the MSR for -- MSR per square meter for the second quarter it was down by 3%. The way we calculate is we take the average of the opening and closing area utilized. So the Beijing 10/11/12 acquisition closed on June 5th. So, we had 25 days revenue contribution but we include the whole of the area utilized in the quarter end number. So if we strip that out then the MSR decline was just about 1% which is more less. I indicated this is kind of long-term trend line. In the third quarter we will have a full quarter's contribution from Beijing 10/11/12, and I think as some analysts investors calculated when we announced the deal the MSR of that data center is, for those three data centers is lower than our average so there will be some dilutive effects in the third quarter and going forward. But I think the MSR will be around that level that is was before adjustments in the second quarter. I talked earlier about the economics of the edge of town sites, the MSR there is lower, the unit CapEx is lower, I said the returns were very acceptable. So there will be continuing gradual decline in the MSR and from time to time we will try to highlight what's happened with our return on investment, which is really what we target to sustain and I think we are indeed doing that. William, do you want to talk about the operating environment in China and what happens if there's a resurgence of the virus?

Yes, fortunately I'm in Shanghai up to 14 days quarantine I feel that everything is going back to the normal. So it looks like in terms of the daily life, retail, restaurant, cinema, entertainment, everything is going to the normal. So almost 95% I think is going to the normal. So I think from the supply chain point of view and the working permission point of view everything is go back. So I think it comes to the normal and we didn't see any impact in the next few quarters.

Okay, great. Thank you.

Thank you. The next question comes from the line of James Wang from UBS. Please go ahead.

Good evening management. I've got two questions. The first question is on supply. So we are hearing that, for example, steel mills in China which have very cheap access to electricity are being converted into data centers. So I just want to get your thoughts on how do you assess the risk of potential capacity other supply and how would you cope with another supply situation should it occur? The next question is just in terms of contracts being renewed so far this year. What are you seeing in terms of pricing, is it broadly similar or lower versus prior term, thank you?

Okay. I take the first question. I think data center actually, in my view, is a high barrier industry. Of course, in recently we see a lot of the new players jump to the market but in terms of our customers' profile, our customers need a reliable vendor. So I think a lot of the new players cannot catch up in a shortened time. So I think it is they are way behind us. It's not direct impact us. On the other hand GDS uses 19 years build up our value proposition and our position. It's not easy to change our position. So I think in my view we're already there and obviously we are in a better position and our customers are smart, very sophisticated, so it will not change our position in our customer. But, of course in terms of the competition it's more concentrated on the Tier 2 players even Tier 3 right. So it will not change any of our position in my view.

Yes, James, Dan here. On your question about pricing on contract renewals let me answer it in a bigger picture way. Yes, we did our first business with our top two customers in 2014 and 2015 so those contracts have come up for renewal. We'll start to see in the next second half of this year and then the next few years more of the Cloud and large internet business come up for renewal. If you look at the contract renewal schedule, which is on page 42 of our earnings presentation, roughly say in second half of 2020, 2021, 2022, it's somewhere around 50%, 60% plus of that area relates to Cloud and large internet customers. So we've already had quite a number of conversations. We think overall the outcome is going to be flat relative to the pricing of the existing contracts. There may be some isolated cases where it comes down and there may be some where it goes up but overall I think it's flat. I've commented before this is not actually a reflection of the market, particularly not for these first few contracts because those early orders were for downtown data centers close to the CBD where the current market price is definitely higher than what it was. So we can't reset these to market we have to reset it to where we are doing business with those customers and the overall relationship. We still have a standard price for our larger customers. It does change from deal to deal, from place to place, data center to data center, time to time. But, yes, when we do these contract renewals we have to take cognizant of what is the price that we are agreeing with those customers for a similar data center in similar areas if it's a completely new piece of business today. So, yes, I think flat overall and we leave something on the table in terms of not extracting the full market price.

Thank you very much.

Thank you. As there are no further questions, I would like now to hand the call back over to the company for any closing remarks.

Thank you all once again for joining us today. If you have further questions, please feel free to contact GDS Investor Relations through the contact information on our website or the Piacente Group Investor Relations. Bye for now.

Thank you. This concludes this conference call. You may disconnect your line. Thank you. Bye-bye.