GDS Holdings Limited

Hello, ladies and gentlemen. Thank you for standing by for GDS Holdings Limited third quarter 2024 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. Hello, everyone. Welcome to the third quarter 2024 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 will refer to during this conference call, can be viewed and downloaded from our IR website at investorsgdsservices.com. Leading today's call is Mr. William Wong. GDS founder, chairman, and CEO will provide an overview of a business strategy and performance. Mr. Dan Newman, GDS CFO, will then review the financial and operating results. Before we continue, please note that today's discussion will contain forward-looking statements made under the safe harbor provisions of the U.S. Private Securities Litigation Reform Act of 1995. Forward-looking statements involve inherent risks and certainties. As such, the company's results may be materially different from reviews expressed today. Further information regarding these and other risks and certainties is included in a company's prospectus as filed with the U.S. 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 gap financial information, as well as unaudited non-gap financial measures. GDS press release contains a reconciliation of the unaudited non-gap measures to the unaudited most directly comparable gap measures. I'll now turn the call over to GDS founder, chairman, and CEO. William, please go ahead, William.

Okay, thank you. Hello, everyone. This is William. Thank you for joining us on today's call. In 3Q24, we achieved a revenue growth of 18% and adjusted EBITDA growth of 15% year-on-year. This growth rate is quite remarkable in current market conditions. It reflects the progress we have made in turning around our China business, plus the uplift from the successful execution of our international strategy. From the beginning of this year, the moving rate has clearly stepped up. In 3Q24, gross additional area utilized was over 25,000 square meters. It is the highest level we have ever achieved. It is all organic. and all interior markets. The main reason for this improvement is AI demand. About half of the moving is coming from the recent new orders and half from orders which have been in the backlog for longer time. For the full year of 2024, we expect net additional area utilized of around 60,000 square meters, which is similar to our peak year of 2020. Looking forward, we have better visibility on the deployment plans of our customers, which gives us a high level of confidence. Therefore, we expect this level of moving to be sustained. The first wave of AI demand is for machine learning. For this use case, customers can deploy in remote area where land and power are available. The second wave of AI demand is for inferencing. For this use case, customers must deploy in the tier one markets. This has been confirmed with our customers in multiple communications. So far, the volume of AI new orders in tier one markets is quite small. Customers are typically looking for moving ready capacity. This fits with our current sales strategy, which is target order, which match our inventory. In the coming year, we expect AI deployments in tier one markets to increase in scale. This will help bring the market back into balance. Some of the new demand will be for hyperscale campuses. We are uniquely well positioned for this. We have large blocks of land and power at diverse sites around Beijing, Shanghai, and Shenzhen, Guangzhou. This kind of resource is scarce. In order to capture the AI wave in tier one markets, We will carefully consider sales opportunities, which require us to initiate new deployment. This is a more expansionary business plan with higher investment and higher growth. It goes hand in hand with the execution of our China REITs strategy. We made a decision to establish our international business as a standalone entity. We believe that this approach will maximize value for our shareholders. The vision for GDSI is to become a leading international data center platform. We pioneered the creation of new markets to fulfill AI demands. As of 3Q24, we have 431 megawatts of total customer commitments. The sales pipeline is very strong. During the third quarter, we signed a contract with a new customer, which is a leading global tech company, for capacity at our campus in Bhutan. 34 megawatts is committed and the 38 megawatts is reserved. This is the first data center of its kind in Asia, designed for the most advanced AI technology. The ramp-up is very fast. We already delivered the first phase in the past few days. It's an incredible achievement by our management and the perfect illustration of how we are able to create new markets. During the current quarter, we officially entered into another new market, Thailand. We are the first data center, hyperscale, we are the first data center operator to undertake the first hyperscale project in Thailand. We received strong support from the Thailand government. We think this is the next big opportunity in Southeast Asia. In the past few months, all the global tech leaders have announced large investment plans in Thailand. We have acquired the land for a data center campus in Chamburi Province near Bangkok, which is the focus area for hyperscale deployment. We plan to develop around 120 megawatts of total IT power capacity on our first land site. We expect to serve the domestic demand for Thailand's fast-growing digital economy. We are very confident to make Thailand another success story. I will now pass on to Dan for the financial and operating review.

Thank you, William. I'll run through the China segment first, followed by international. Starting on slide 16, in 3Q24, GDSH segment revenue increased by 6.1% and adjusted EBITDA increased by 3.6% year on year. GDSH revenue growth was mainly driven by an increase in total area utilized of 11.6% year over year. MSR per square meter has declined moderately over the past four quarters. Adjusted EBITDA margin for 3Q24 versus 3Q23 was down by one percentage point. The main reason for this is the increase in power tariffs, which occurred in the second half of last year. Looking forward, we expect net additional area utilized to continue at around 15,000 square meters per quarter. We expect MSR to decline slightly over the next year and assume that power tariffs remain at current levels. For the first nine months of 2024, our China capex totaled 2.6 billion RMBs. We originally guided for around 2.5 billion RMB for the full year. However, as we had higher move-in, we had to bring forward a small amount of capex. As a result, we are revising up our China capex guidance to around 3 billion RMB for 2024. In our base case business plan for next year, we envisage China capex within the 2 to 3 billion RMB range. For the first nine months of 2024, cash flow before financing for GDSH is negative 293 million RMB. The fourth quarter of the year is usually a good quarter for collections. Therefore, we expect the cash flow before financing for the full year to be positive. This is in line with our financial target. This remains a key financial objective to deliver steady growth while generating positive cash flow before financing. As shown on slide 20, at the end of 3Q24, the cash balance of GDSH decreased to 7.8 billion RMB, and the net debt for last quarter annualized adjusted EBITDA multiple was 7.4 times. Over 60% of our gross debt is RMB denominated project term loans. The interest on these loans is floating rate linked to the over five-year LPR. This benchmark came down by 25 basis points at the end of 3Q24 to see some benefit as our interest cost resets. As William mentioned, we are fully committed to establishing a China REIT. This strategy for a number of reasons.

it will give us continuous to the office with a lower cost of equity.

Two, it will increase our liquidity and enable us to accelerate deleveraging. Three, it will provide us with an additional source of capital to capture attractive new sales opportunities as AI demand takes off. And four, It will benchmark the valuation of stabilized data centers in China and give us the means to recycle capital on an accretive basis. We are approaching China REITs with two transactions in parallel. The first involves the creation and IPO of a listed CREIT, which holds stabilized data centers. The second involves the creation of asset-backed securities, ABS, which holds ramping up data centers. The ABS is the form of security which can be injected into a CREIT once the underlying data centers are stabilized. Hence, we call the ABS a pre-REIT. For the CREIT, we have already completed a very thorough review process at provincial level. Our application is now subject to a national level review by NDRC, CSRC, and the Shanghai Stock Exchange. We are aiming to obtain all the necessary approvals by the middle of next year. This would create the possibility of launching the CREIT in the second half of 2025. For the purposes of the listing, we selected one large site which fits the typical CREIT offering size of around 2 billion RMB. However, once established, we would intend to eject additional data centers if it creates shareholder value. Currently, CREITs are trading on significantly higher multiples than the current implied multiple for our China business. For the pre-REIT, we are at the final stage of obtaining all the necessary approvals. The pre-REIT delivers many of the same benefits as the CREIT, but it's a one-off transaction. We're trying to get it done as soon as possible. Turning to the international segment on slide 23, in 3Q24, GDSI revenue increased by 42%, and adjusted EBITDA increased by 15% quarter-on-quarter. Currently, we have around 130 megawatts of IT power utilized. As the delivery schedule for most of the backlog is very short and customers undertake to move in quickly, we expect to have around 400 megawatts of utilized capacity within the next 18 months. For the first nine months of 2024, international CapEx was around 4 billion RMB. We originally guided 4 billion RMB for the full year. But due to the extraordinary ramp up, we now expect full year capex for international to be around 8 billion RMB. Turning to slide 29, on 29th of October, we announced that GDSI has raised 1 billion US dollars from the issue of Series B convertible preferred shares. The issue is expected to close by the end of the year. Post-closing and on an as-converted basis, GDSH will own approximately 37.6% of GDSI. The value of our equity interest implied by the Series B subscription price is approximately 1.3 billion U.S. dollars, representing a 75% premium over the Series A valuation and equivalent to approximately $6.75, per GDS ADS. The private equity valuation of data centers is typically based on the sum of installed capacity, contracted backlog, and near-term sales pipeline. As GDSI ramps up its installed capacity, wins new orders, and adds to its supply of highly remarkable capacity, we expect the equity valuation to increase significantly. In this round, we aimed to secure enough capital to build and deliver over one gigawatt of capacity. Including the proceeds of Series B, we will have $2.1 billion of equity. With the addition of a moderate amount of mezzanine debt, we should be able to achieve our one gigawatt target. Post-closing, GDSH will deconsolidate GDSI. However, we will continue to disclose enough key information for you to track GDSI's performance. To ensure that the value of our equity investment accrues to GDSH shareholders, our plan A is to IPO and spin off GDSI. We will give you a better sense of the timing on our net sterling score. Finishing on slide 30, we are maintaining our formal guidance for FY24 consolidated revenue and adjusted EBITDA. However, as I mentioned before, we are raising our CAPEX guidance to 11 billion RMB, which includes 3 billion RMB for China to support the faster move-in and 8 billion RMB for international, reflecting its accelerated expansion. We would now like to open the call to questions. Operator?

Thank you. If you would like to ask a question, you'll need to press star 1 and 1 on your telephone and wait for your name to be announced. To withdraw your question, please press star 1 and 1 again. For the benefit of all participants on today's call, please limit yourself to one question. If you have more questions, please re-enter the queue. Thank you. We will now take our first question. First question is from the line of Yang Lu from Morgan Stanley. Please go ahead.

Thanks for the opportunity. I have one question about the China strategy. First, we saw a pretty strong demand, good moving and good booking. At the same time, we saw the upper revision of the capex. Is that a pull forward of the previous plan 2025 or 2026 China capex or is that new? And William just mentioned that the China future investment will be linked to the REIT strategy in China. Now they mean that if you deliver the REIT listing then you can invest more aggressively in China

without a risk uh you're still in a tight control mode of the capex does it mean this kind of thank you yeah i think yeah i think the number one we still stick on our previous strategy try to stabilize our uh uh business in china and uh maintain the the like let's say sort of the growth organically, right? And digest our inventory. This is our first priority. But in the meanwhile, we always talk about we selected to do the new business. That means very high quality business. But Of course, we are currently trying to recycle capital in an efficient way and then create more value for our shareholders. In the meanwhile, if we can do recycling capital in a sufficient way in the future, then we will start to be more aggressive to take some new orders, seeking the growth again. Yeah, this is, I think, we try to make the growth and stabilize more balance in the future. Thank you.

Thank you. We'll now take our next question. This is from the line of Jonathan Atkin from RBC. Please go ahead.

Thanks. A couple of questions. If you could maybe talk a little bit about the development of the compound market in happening more rapidly than I would have expected. Do you expect there to be a lot of smaller internet enterprise demand, or will it largely be large deployments of the type that you've recently gotten commitments for? And then secondly, on GDSI, can you talk about the ongoing financial commitments that GDS would have?

John, your line is not that clear.

I think the first question was about the kind of demand. Was John talking about Batam? Yeah, I think so. John, what kind of customers interested in Batam? I think John was asking.

Yeah, okay. Let me answer the first question. In Batam, with your view, it's a security data center hub. I think that this, whatever, in Johor or Singapore or Batam, all the customers, hyperscale customers, whatever from the U.S. or domestic or China, all very interested in Patent. So far, we have a very strong pipeline from the different countries.

Yeah, ongoing commitment to GDS International and GDS Holdings.

Yes, okay. On that one, I think we can be clear that the decision was made, I think at the beginning of last year, that we would cap our allocation of capital to international. At the level then was around $400 million. And in the course of the Series A capital raising, actually part of the use of proceeds was to repay some shareholder loans from GDSH to GDSI to bring that number back down to the $400 million level that we had established as our limit in terms of capital allocation. So from that point forward, our motivation is to make our $400 million investment as valuable as possible. Based on the Series B price, we've turned that $400 million into $1.3 billion of value. And I personally think that will continue to increase at a rapid rate. Every shareholder or international has preemption rights, but we don't expect that GDS Holdings will exercise preemption rights in the future. So GDS Holdings portion will be available to allocation to other investors, particularly investors who can add value to the international business.

And then lastly, the Singapore CSA process, you know, ECHO would have just announced within the past hour or so in progress on the air what they're going to be doing with their build. What is your visibility around timing, magnitude, location of where you're going to be building in Singapore?

Yeah, I think, yeah, location.

yeah I think it Singapore location which is a site a very good location which is another cost data center cluster in the in a in a western part of the Singapore I think in the future we will link them link link them to that what you call in a button data center campus as well and so we will create a very good platform in a security area and which was our customer very very like And this is number one. I think we target, of course, we target by the end of the 2026, we will complete the construction. This is our current schedule.

The size of that building is quite large relative to the power quota which we have as of now. And that was a deliberate decision because with a relatively small amount of incremental investment, to give us the option and the potential to be able to expand in a number of different ways on that same site.

Right. Thank you.

Thank you. We'll now take our next question. This is from Louis Tsang from Citi. Please go ahead.

Hello. Thank you for taking my question. Congratulations on the stronger supplement I think overall, the domestic movement's real results are very positive and encouraging. So today, I have two questions. The first one is regarding the GDSI. Could you please further elaborate on the rationale of entering the talent market? Or let me put it this way. How should we think about the future demand and the availability of the resources like connectivity, power, land, water? And do you think that if there's any possibility that Thailand could be another DC hub similar to Juho. That's my first question. Maybe I should go one by one. Thank you.

Okay. I think the rational to enter into Thailand, we always think about Thailand's 80 million population, and the economy grows very healthy, and the If you know that recently, this year, a lot of the hyperscalers announced investment in Thailand. And we also understand that digitalization, developed digital economy is the key driver. It's a national, let's say national strategy to drive the Thailand economy. So the government very, very support all the foreign investment to Thailand. So we are very proud. We are the first hyperscale player into this market. We always... we all enter into a market more early than anyone else. And we're definitely confident there's a lot of demand will come in. So I think this is the rationale. In terms of the, I think the, currently we are more targeted, our customer, they mainly service, serve local domestic market. I think the, but Thailand has the potential to grow up another data center hub in Asia Pacific in terms of the the local economy plus their influence for the country around them, right? So I think in the In Asia Pacific, we believe there's not only one data center hub, right? In the future, I think maybe there's another two or three data center hubs will grow up. So I think this is our next rational to think about.

Okay. Thank you so much. Thank you so much. And the second one is about the domestic one, because I think the domestic moving target that I just heard, was really good. And, um, I just want to know, like, how should we, like, we calibrate the outlook for the domestic movie, like next year, because it's not that the demand is going to, um, sustain, but just like, if there's any chance, like that, the movie we'll see like a further acceleration next year, or which I think that like, maybe it would maintain like a similar level, because to me, like, based on my understanding, right, the campus of cycle of the domestic giants, similarly, will you accelerate in next year, the first half, maybe? So I just want to know more about the outlook for that. Thank you so much.

Yeah.

Yeah. To clarify, because I think Lu Yang asked earlier this, The slightly higher than guided capex this year is bring forward. It doesn't indicate that that's the kind of level of next year. In fact, I'd say if we were giving guidance for next year now, it would probably be the same as our original guidance for this year, around 2.5 billion RMB. For the move-in, we're talking about net additional area utilized. the gross level is the 25,000 square meters in the third quarter and then net of churn on an annual basis. We expect 60,000 square meters this year and we're confident enough at this time to say that we expect that level next year as well. This rate of growth, when he was talking about our strategy, is really the result of continuing to be very selective in terms of new business and focusing on delivery of the backlog and minimizing capex to prioritize cost to complete and increase asset utilization rate we expect to increase to the high 70s from low 70s at the moment to high 70s by the end of next year. That's our base case. What we can't be certain about is the opportunity to undertake new campus development, because undoubtedly there will be demand from hyperscale coming to Tier 1 markets, which requires new development. And we have a land bank, land with power quotas. multiple sites around Beijing, Shanghai, Shenzhen, Guangzhou. And that kind of resource is scarce and it will be in demand. It's our choice to proceed with those kinds of opportunities if they're attractive enough and we choose to do so. But ideally, that would go in sync with having established a mechanism and vehicle for monetizing stabilized assets. I think that's ideal from the point of view of our shareholders. The timing may not be exactly matched, but I think both are, you know, we expect to see those kind of new development opportunities next year. And we talked about the progress we expect to make with our asset monetization, both the pre-REIT and the CREIT over the next few quarters.

Okay. Thank you so much. Looking forward to the further development. Thank you so much, Manish. Thank you. I have no further questions. Thank you.

Thank you. The next question is from the line of Frank Luthan from Raymond James. Please go ahead.

Great, thank you. On the Vatan design, can you give us some idea of what makes that so advanced? And when you do a build like that, how much of that capex are you paying for on your book that's sort of at risk versus customers that are building in, what you charge them to pay for, what I assume is additional cooling or other things like that? Thanks.

What kind of design? What kind of design?

this is a liquid cooling net design based design i think yeah this is a definitely is all the new new design for the air we call the air ready data center right obviously yeah I think, of course, a lot of the designs are mainly customized by the customer. They pay everything, right? In print, yeah.

Okay. And what sort of power density are you able to get with this design? Is there a limit to it, or how should we think about that?

It's around, I think it's around the, let's say, averages around the can take a bit per graph if I maybe, yeah, it's roughly, I think, yeah.

Okay, great. Thank you.

Thank you. The next question is from the line of Edison Lee from Jefferies. Please go ahead.

Hi. Thank you for taking my questions and congrats again on a very good I think I want to focus on China because, well, I'm very interested by, but when I looked at, you know, that area utilized, it actually grew by 11, 12% on a year-on-year basis, but your China revenue only grew by 6% on a year-on-year basis. So can you help us understand the trend of that unit pricing going forward?

Yeah, the reason for the disparity between increase in area utilized and revenue is that there's always a lag, right? Because the move-in happens over a period of time. I think from 4Q23 to 4, sorry, 3Q23 to 3Q24, there was a 4% decrease in the MSR. If you do that comparison, on a like-for-like basis. The decline over the past year has been less than that in most quarters. But third quarter to third quarter, it was 4%. And it was also during the third quarter of 2023 that power tariffs, both the tariff charged by the generator and by the grid, both went up in the third quarter of 2023. And that resulted in like one percentage point reduction in the adjusted EBITDA margin when we compare 3Q24 versus 3Q23. So I think those two together explain the disparity between the increase in the area utilized in the revenue and EBITDA growth rates. Going forward, I think over the next four quarters, you can expect the MSR to decline by something like 2% if you compare the same quarter of each year.

So does it mean that, assuming that, as I said earlier, the net additional utilized next year would be, let's say, 60,000 square meters, and then you have price erosion of 2%, final revenue can actually grow at a high single digit by next year.

That's right. I mean, aim for 10%.

Aim for 10%. Okay. And then a quick follow-up. I mean, given plenty of high-capacity on China, and then you're talking about maybe investing in some high-potential new projects, does it change your objective or previous target to be free cash flow breakeven in 2025?

No, it doesn't. For a start, that's not reflected in the numbers. It's an upside opportunity that we're just flagging, but it would require additional investment. I think we will, in our face case, be constantly positive in terms of cash flow before financing next year, as indeed we will this year, if we incur additional investment and the capex is higher than, say, the 2.5 billion RMB number that I mentioned, hopefully that will be in conjunction with having completed some asset monetization. And that net, it should result in us still being positive.

Yeah. I see. Okay. Got it. Okay. Great. Thanks.

Thank you. The next question is from the line of Sarah Wang from UBS. Please go ahead.

Hi. Congratulations on the solid set of results. I have two questions. First is on China. So would you please elaborate more on the potential outlook of China business. Meaning, I think as management just mentioned, if we want to grab the AI opportunity in tier one markets that involves new investments, but at the same time, we are also having quite a decent backlog. So previously we were talking about cost to complete existing backlog. Is there any change in the cost to complete assumptions given Now the demand is driven by AI, but maybe some of the backlog is entered one or two years ago in the late cloud cycle. And then my second question is international business. Given the solid new orders signed this year, how shall we think about the new orders into next year, for example, in terms of volume? And then it's glad to see that we're entering into new markets, but given we are working in multiple overseas markets, which one is our top priority? Thank you.

Let me start with the cost to complete. So at the end of the third quarter of 24, we had 120,000 square meters under construction. And the cost to complete for that capacity is 6 billion RMB. Some small part of it actually relates to some data centers which are already in service because sometimes there is some cost to complete, meaning fitting out some space which is not yet utilized where we're able to phase the timing of the installation of M&E and so on. But across the in-service and under-construction portfolio, the cost to complete in aggregate is 6 billion RMB. If we were to undertake a completely new project, we would be utilizing land and energy quota, which we already have. That's why we put into our earnings presentation, again, a disclosure about how much capacity we have held for future development, because this is becoming relevant again. And this is scarce resource. There are not that many options for customers as they're looking for somewhere to be able to deploy 50 megawatts or more in the edge of town around Beijing, Shanghai, Shenzhen, Guangzhou. And we have a number of solutions for that. But you know our unit development costs in China, 50 megawatts, we're probably talking about incremental one to 1.5 billion RMB. But one project like that could add four or five percentage points to our growth rate. So that's how we think about it, but we're just talking about it now because it's a possibility for us next year, but it's our choice whether we go ahead with that. It really depends a lot on how attractive the commercial terms are.

Okay. Yeah, I think here we still, yeah, in terms of the international market, I think we definitely, Sejuri is still our main focus, Sejuri area, and we're well positioned on that, and we have very good location and custom base already, and we see that a lot of demand in our pipeline, and I think to monetize our position, secure the power in the land is definitely our first priority. But in terms of the new market, in terms of new sales, Yeah, I think the next year we aim to, let's say, 200 megawatt. And as I last quarter, I said, in three years, we will double our, we try to reach to the one gigawatt cells in next three years. So I think this is our target. And I think we are very confident on this target, which set up already. and in terms of new market i think there's definitely a new market so so another strategy because this demand wave is global demand it's so if you look at it the ai driven the demand is not only in such as asia even japan and the middle east europe and the us as well so we try to uh we try to Maintain our let's say long-term growth. So we should look at today the new market opportunity very closely So telemarkets are one of our new market, but I think that is still in the sausage Asia But I think the we also target from the new market like in Japan like we already started a new project already But it says it's a it's a test of what the project is We have our team very closely to follow up the market trend in Japan. And even we have some study, very deep study in the European market. So we try to set up another couple of the growth engine in the next few years.

Got it. Very clear. Thank you.

Thank you. As there are no further questions, I'd like to now turn the call back over to the company for closing remarks.

Thank you all for attending today, and we'll see you next time. Bye.

Thank you. This concludes the conference call. You may now disconnect your line. Thank you.