Hello, ladies and gentlemen. Thank you for standing by for GDS Holdings Limited fourth quarter and full year 2019 earnings conference call. At this time, all participants are in listen-only mode. After management preferred 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.
Hello, everyone. Welcome to the full Q 2019 and full year 2019 earnings conference call of GDS Holdings Limited. We are deeply sorry to keep you guys waiting for so long. We just had some technical issues last minute to file our release with the SEC. We assure you that everything is fine. The company's results were issued via Newswire Services earlier today and are posted online. A summary presentation, which we'll refer to during this conference call, can be viewed and downloaded probably soon from our IR website at investors.gdsservices.com. Leading today's call is Mr. William Huang, GDS Founder, Chairman, and CEO, who will provide an overview of our business strategy and performance. Mr. Dan Newman, GDS CFO, will then review the financial and operating results. Ms. Jamie Koo, 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 U.S. 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 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 GAAP financial information as well as unaudited non-GAAP financial measures. 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 Founder, Chairman, and CEO, William Huang.
Please go ahead, William.
Hello, everyone. This is William. I'm here in Hong Kong with Dan and Jamie. Thank you for joining us on today's call. With all the recent developments, 2019 feels like a long time ago. Please allow me to begin by talking about our great achievement last year. First of all, we hit our sales target, adding over 81,000 sq m or 174 MW of new customer commitments. When fully delivered, this will add RMB 2.4 billion or US$345 million of annual recurring revenue. We expand the data center capacity in line with sales, adding over 90,000 sq m in service and under construction. In addition, we expanded our development pipeline. We currently have over 320,000 sq m procured for future development, which is far more than anybody else in the market. It's very valuable and key to our continuing success.
All of our sales and all of our capacity additions were in Tier 1 markets. Our financial results were impressive. We grew revenue by 47.6% and adjusted EBITDA by 74.3% year-over-year. We beat guidance on both metrics. Our adjusted EBITDA margin came out nearly seven percentage points higher. We raised $900 million of equity to ensure that we can keep on growing at the current pace or faster. Furthermore, we established an innovative strategic partnership with GIC, which expands our addressable market and gives us access to an alternative source of equity. Let's turning to slide five. Demand was consistently strong throughout 2019. What is driving this? First and foremost is cloud adoption. Alibaba, the market leader, reported over 60% revenue growth for the last quarter. Tencent Cloud just reported nearly 90% growth. The cloud in China is still at an early stage in terms of penetrating large enterprises.
In addition to cloud, we have recently started to see our customer gearing up in anticipation of 5G takeoff. Let's turn into our customer franchise on slide six. The major highlight is expansion of our hyperscale customer base. On the one hand, demand from our top two customers was very well sustained, and continues to drive around 50% of our sales. On the other hand, we made breakthroughs with several key accounts. As a result of which we are now a significant service provider to almost all of the hyperscale customers in China. The winning factors are our multi-market platform, continuous supply, long-term track record, reputation for operational excellence, transparency, and financial capability. These are the differentiators which have taken years to develop and are not easily matched. We believe that our market share has increased in Tier 1 markets.
With our current customer mix, we are pledged into growth across the digital economy. In addition to hyperscale, we added some highly prestigious new logos. In the last quarter, we signed a master sales agreement with Apple and won our first business from Apple. We entered this year with great sales momentum. As a result, we have raised our sales target for 2020 to 100, 000 square meter net add, made up of 80,000 square meter organic growth and 20,000 square meter from the pending acquisition in Beijing. This target does not reflect any flow through from increased usage of digital services in the current period. In Q1 20, we are on track to achieve comfortably over 20,000 square meter net add, and the next quarter also looks very strong. This demonstrates that customers are not holding back.
We expect to make a lot of progress towards the 100K target by the middle of the year. Turning to slide seven. Not only is demand strong, it is also very noticeable that customers have changed their approach. They are pre-committing earlier to secure their supply. This is reflected in the upward trend in our pre-commitment rates. In reality, almost everything that we do is driven by specific customer requirements. On slide nine. As we have been saying for a while, the biggest challenge is keeping up with demand in Tier 1 markets. To deal with this challenge, we have evolved our approach to project sourcing in three major ways. First, because of the restriction on data center development in urban areas, we have established a supplementary presence at the edge of town, such as Langfang to serve Beijing, and Kunshan and Changshu to serve Shanghai.
Second, we have increased our property ownership, existing buildings in urban areas and greenfield land at the edge of town. We now own over 50% of our entire capacity, including the development pipeline. As compared with around 20% at the end of 2018, increased ownership give us much more flexibility and the certainty of supply. Third, we have put tremendous effort into building up our pipeline of future projects. We aim to have at least three years supply in each market and have made great progress towards this goal. The change of approach is already yielding great results. Let's move to page 10. Take Langfang as an example on slide 10. It's 50 kilometers from Beijing and a viable edge of town location due to the existing concentration of carrier data centers.
We selected Langfang with the endorsement of our top customers and spent a long time working with the local government on a framework agreement for power, land, and investment. One year ago, we had nothing in Langfang. As of today, we have 30,000 square meter of capacity in service and under construction across five data centers, all of which are 100% committed by our top customers. We have secured another 83,000 square meter of developable capacity. We aim to repeat this success in other Tier 1 markets. Another way in which we have evolved our approach is with regard to acquisitions. We started off a few years ago viewing M&A as a means of adding to our supply. Now we also view it as a way of increasing our presence in key locations, expanding our relationship with strategic customers, and accelerating our growth on very accretive terms.
We have stepped up our M&A efforts. If the opportunity allows, we aim to do more deals. We are actively pursuing several deals, targets. Before I hand over to Dan, I would like to say a few words about the current situation. From the outset of the COVID-19, our top priorities have been to ensure, number one, the safety and the well-being of our employees and of the people we interact with. The second, incident-free operations. So far, I'm pleased to say we have achieved both of our goals with zero infections and zero SLA breach. It has not been easy. We made many changes to our policies, procedures, and communications. Business continuity is where we come from. It's part of our DNA. The steps that we have taken have been very much appreciated by our customers. We received a lot of positive feedback.
It's at a time like this that you get tested. That the quality of our operations sets us apart. That customers remember why they do business with us. Our reputation has been enhanced. Coming into 2020, we felt that our market position and capabilities had got a lot stronger over the past year, while the opportunity in front of us keeps getting bigger. The virus epidemic is a tragedy, and our thoughts and prayers are with all those who have been affected. During this tough time, digital service has played a critical role. We have all had to change our behavior, and this may result in a structural shift in how we live and work. The importance of the underlying infrastructure has been recognized at the highest level of the Chinese government, and may result in favorable new policies. We are waiting to see the specifics.
However long it takes to get through this period, we believe that the fundamental of our market position and opportunity will remain intact, if not stronger. With that, I will hand over to Dan for the financial and operating review. Thank you.
Thank you, William. Starting on slide 15, where we strip out the contribution from equipment sales and the effect of FX changes. FY2019 finished strongly, and I'm pleased to say that we beat our guidance for revenue and adjusted EBITDA. In full 4Q2019, our service revenue grew by 9.5%. Underlying adjusted NOI grew by 7.4%, and underlying adjusted EBITDA grew by 8.8% in consecutive quarters. Our underlying adjusted EBITDA margin was fractionally down in the last quarter at 45.6%. For the full year, our underlying adjusted EBITDA margin was substantially higher at 44.7%, compared with 37% in FY2018. Turning to slide 16. Service revenue growth is driven mainly by customers moving in to the space which they previously committed. Move-in during 4Q2019 was 18,000 square meters, including 7,800 square meters from BJ9. BJ9 is an acquisition which we entered into last year.
It is not closed yet, pending a final CP. As an intermediate step, the existing customers have entered into new contracts directly with us, and we have taken over operation of the data center under a management contract. Our MSR was quite stable over the course of 2019. We expect a slight drop in 2020, mainly due to customer and location mix, acquisitions, and the timing of moving. Slide 17 shows the quarterly trend in margins. In 4Q2019, underlying adjusted NOI margin decreased by one percentage point, mainly due to 45,000 sq m of new capacity coming into service in the last two quarters. Under the BJ9 arrangement, we are getting a low double-digit profit margin until the deal closes, which is a slight drag.
The decrease was partially offset by leverage on SGA, which went down to 7.8% of service revenue, compared with 8.4% in the prior quarter. Adding it all up, our 4Q underlying adjusted EBITDA margin was just 0.3 percentage points lower. For FY20, we expect around a one percentage point improvement at the NOI level and a further 1% from leverage on SGA, but the quarterly trend could be a bit up and down. Turning to slide 20. Our total CapEx in FY2019 was RMB 5.3 billion, including RMB 1.5 billion related to acquisitions of data centers, property, and land. In 4Q2019, we paid for the Hong Kong two site, the Shanghai 14 building, and most of the consideration for the Guangzhou six acquisition.
Up to the end of last year, we had also paid out RMB 270 million for build to suit joint venture projects, which will be reversed when we sell the 90% equity interest to GIC. The majority of our CapEx consists of plant and equipment, which is essentially the same in each data center, and the cost is easily benchmarkable. We have been able to reduce our unit CapEx for P&E by 3%-4% per annum over the last few years and expect to continue doing so. The remainder of the CapEx relates to the building, which can be leased or owned, and to the external power infrastructure. The unit CapEx for this part can vary depending on the specifics of each project, but on average has stayed at around the same level.
On slide 22, we ended 2019 with gross debt of RMB 16.2 billion or USD 2.3 billion, around 80% of which was in the form of mostly local currency denominated project term loans and finance leases. This debt is structured to fit around the project cash flows and is substantially covered by our multi-year contracts with investment grade customers. The term loans are covenant light or have no covenant at all. The remaining 20% of our debt is made up at the CB at HoldCo level, which is unsecured and has a remaining term of over five years, and of working capital facilities which we have rolled over numerous times. In 2020, we're guiding for RMB 7.5 billion of CapEx, which is elevated due to payments for pending data center and property acquisitions.
Assuming a conservative financing ratio of 40/60 equity to debt, we will need RMB 3 billion of equity and RMB 4.5 billion of debt to finance our CapEx. For the equity part, we are sitting on RMB 5.8 billion of cash, thus we expect positive operating cash flow this year. For the debt part, most of the facilities are already in place. We have RMB 2.5 billion committed but undrawn, leaving about RMB 2 billion, which we are working on right now across seven facilities with local and foreign banks. To put this remaining requirement into perspective, last year we secured nearly RMB 6.5 billion of new debt facilities. The banking market in China is very supportive. We have a great track record as a borrower and have developed great banking relationships. I can see no reason at all why this should not continue.
Finally, we established partnerships with Ping An and GIC to ensure that we have access to diverse funding sources and are not reliant on the public markets. Regardless of the current situation, we are always considering alternative funding options with these and other potential partners to optimize our capital structure and cost. Turning to slide 23. Our contract backlog has been increasing each quarter. We ended FY 2019 with 108,000 sq m, equivalent to 70% of our revenue generating area. Part of the backlog relates to data centers in service. The amount has remained in the 40,000-50,000 sq m range over the past five quarters, driving organic move-in of around 10,000 sq m per quarter, which implies about a four to five quarter move-in period. The remaining part of the backlog relates to data centers under construction.
The amount has increased significantly from just over 31,000 square meters at the end of 2018 to just over 57,000 square meters at the end of last year. There are two reasons for this. One, as William mentioned, customers are pre-committing earlier and to a much greater extent. Two, as you can see from the table on page 21, we're undertaking more greenfield projects where the construction period is around six months longer. As the backlog related to data centers in service increases, we would expect the quarterly move-in to increase. However, this is subject to the timing of project completion and other factors in the current uncertain operating environment. To finish on slide 24 with our guidance. We are nearly at the end of the first quarter of 2020. Before I talk about our outlook, I should mention what we have already seen in the year to date.
The COVID-19 epidemic is affecting us in two main ways, construction and move-in. At the end of January, construction across our 16 self-developed and build-to-suit projects came to a halt for Chinese New Year and did not resume until recently due to government restrictions. We're not experiencing significant problems with our supply chain as we had placed orders well in advance. Nonetheless, we've lost a couple of months, which we will try to make up. The kind of delay which we've experienced will not materially impact our financial results in the current year. Move-in is a much more material issue in the short term, as it's the primary driver of our revenue and profit growth. The first quarter is usually a seasonal low for our business, and this was reflected in our original forecast assumptions for the current year.
As of today, it looks like our 1Q 2020 move-in will end up a few or several thousand square meters short of our original target. Nonetheless, we should still be able to achieve 1Q 2020 revenue and EBITDA growth in the mid to high single digits quarter-on-quarter. Looking forward, the situation is that we have a large amount of capacity and service ready and waiting for our customers to move in. Our customers want to move in, there are still many operational limitations and uncertainties in their supply chains, particularly with regard to IT hardware. China appears to be on a path to recovery, this will be affected by what is happening with suppliers inside and outside the country.
Given the lack of visibility about the pace of recovery, we took the view that we should revise down our move-in assumption by several thousand square meters incrementally per quarter. We've not changed any other assumptions in our forecast. Putting this into our model, we're guiding for revenue of RMB 5.63 billion at the midpoint, implying 36.6% growth year-over-year, and adjusted EBITDA of RMB 2.61 billion at the midpoint, implying a 43.1% growth year-over-year. These growth rates are around five percentage points lower than what we originally intended to guide. Our sense is that this guidance is conservative but appropriate in the circumstances. We believe that the risk to the upside is greater than the risk to the downside.
I mentioned already our CapEx guidance of around RMB 7.5 billion, of which RMB 2.5 billion mainly relates to the pending acquisitions of BJ9, BJ10, 11, 12, and a building in Minhang District, Shanghai. I'd like to reiterate that all of our fundamentals remain intact. Customers have not changed their plans, and despite the tough conditions, we expect our business performance to be highly resilient. We are as confident as ever about our medium and long-term growth prospects. With that, I'll end the formal part of my presentation, and we'd now like to open the floor to questions. Operator?
Certainly, sir. Ladies and gentlemen, we will now begin the question and answer session. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. If you wish to cancel the request, please press the pound or hash key. For the benefit of all participants on today's call, please limit yourself to a question. Your first question comes from the lines of Yang Liu from Morgan Stanley. Please ask your question.
Hello, good evening. This is Yang from Morgan Stanley. I have two questions. The first one is for William, on government policy. We noticed that recently, Central Government in China encouraged the data center as a new infrastructure for the first time. What are you expecting in term of the future policy support? Can we see more power quota or more funding support, and whether this kind of policy will change the investment return profile in this industry? The second question is for the new booking target. I'm not sure if the acquisition announced in December last year, was this split to 2019 and 2020, or is it all of the around 20,000 sq m will fall to 2020 new booking? Thank you.
I answer the first question. Yeah, you are right. The first data center was categorized by the Central Government as a new strategic infrastructure in China right now. It's mainly driven by the 5G strategy. I think the effect to us, now it's still early to say some what happened, what the new policy will launch specifically. We were invited by the Central Government to discuss how to help you guys. How to give them more advice, how to develop the growth, this industry. What I can tell is that there's a broad topic that we have discussed with the Central Government. I have to say, we are the only data center vendor being invited. Number 1, I think what we talk about is how to release some carbon quota in Tier 1 market to release the supply a little bit.
Government still have the concern about the total carbon quota deployment. We have discussed how to appropriately to release some carbon quota to data center industry. Number two is, we talk about some topics around how to reduce the power cost in the future, and how to reduce the loan interest rate. This is all discussed. In general, I think it's positive for us, and the Central Government want to help to develop this industry. So far, I still say now the stage is too early to say something right now. That's my view. Yeah. The second question.
Yeah. Hi, Yang. Last year we entered into three acquisitions. Two were included as part of last year's new business. 106 closed Beijing 9, we took over the way I described. The third acquisition, which we call Beijing 10, 11, 12, announced in December, we're working hard to close that hopefully by the middle of this year. When we talk about our 100,000 square meters sales target for this year, it will be 80,000 square meters organic and 20,000 square meters through that specific M&A deal. 80,000 square meters organic is significantly more than we've done before organic, and we believe that 80,000 square meters organic is sustainable. That's the new normal. Beyond 80,000 square meters is either M&A.
New acquisitions.
Maybe some upside. Yeah.
Yeah. I added one point. We also talk about it to the Central Government, how to release, how to close some inefficiency small data center and in-house data center to force the use of professional data center like us.
Got it. Thanks a lot.
Your next question comes from the line of Jonathan Atkin from RBC. Please ask your question.
Thank you. Dan, you just sort of answered one of my questions, which was about the 100,000 being the new norm, and I wondered, do you think you could potentially do more than that given the boost in demand that you're seeing from the current environment, either organically or through M&A? I wondered if we could also pivot a little bit to the changes that have been taking place at CyrusOne and any impacts on board membership and just the overall relationship. I do know that Tesh, their new CEO, has spent a lot of time in China, but if you could maybe comment on CyrusOne, that would be interesting. Thank you.
Two questions. Can we do more? First part.
Yeah, I think we didn't change our view. Actually, last quarter, when we talk about the market demand, we still sit on that view. The China data center market is accelerated. It was not impacted by the virus. We believe the virus stuff is a short-term impact. From the midterm, long-term, we still think the total market will accelerate. This is number one. That means we have the chance to do more in the future. This first half year, maybe a little bit of tough, but it will not change our view for the future market. We believe the digitalization is a overwhelming trend in China, even in global market, right?
Yeah. Second one about CyrusOne.
The CyrusOne, I think, number one, we still maintain the relationship with the CyrusOne. We still have some deal talk about it together. It will not change that Gary stepped down or anyone who We deal with the institution, not deal with the individual, right? I think it will not change our relationship with CyrusOne. We still help each other. Tesh and Jonathan called me after their announcement. We had the conversation with Gary. I think number one, we will not change the board seat right now. Gary is a respected professional in the industry, always bring that variable opinion for GDS. We appreciate that. On that hand, as I said, we deal with the institution, not personal. I think the CyrusOne and GDS separation will be not changed.
Thank you. I wanted to maybe talk about or ask you, have you seen any differences in the pace of deliveries and construction by the rest of the industry? You talked about how the virus has affected you from essentially a labor standpoint and slowdowns related to that. Has that been affecting the competitors equally or more so? I'd be interested in your perspective on that. You also mentioned, again, the increase in demand, and how does that influence your thinking about entering new markets? In the past, you've sort of alluded to a couple of new metros in China that you would think about investing in. Is the appetite for that equally as strong now, or has customer demand trends changed your thinking on that? Thank you.
Just to be clear, William.
Yeah.
The first question was whether competitors have been affected like us or to a lesser or greater extent.
You mean our competitor?
Yeah, affected. Have you been affected by it in the current situation? Have they been affected, in terms of their construction timelines, them and so on?
I think in general, the current situation is equal for all of our competitor, right, and us. I think this is number one. The advantage for GDS is we have the scale. We are more well-managed internally, in my view, right. In terms of other construction point of view, I believe we manage better than the other competitor. This is my view. Yeah.
Yeah. The second question from Jonathan was sort of whether we see increasing demand in other markets, new markets, and whether we have appetite to go into those places.
I think we are ready to go any new Tier 1 market, right? We didn't change our view. We will go to some new market. As we mentioned last couple quarter, Hong Kong market, Chongqing market, and something new, like, Nanjing and Hangzhou is our target in the future. On that hand, we also think about based on our current install-based customer requirement, we are serious think about how to go to the Southeast Asia.
Thank you very much.
Ladies and gentlemen, once again, for the benefit of all participants on today's call, please limit yourself to two questions. Thank you. Your next question comes from the line of Colby Synesael from Cowen. Please ask your question.
From Cowen.
Great. Thank you. From Cowen, yeah. Two, if I may. Number one, I think you guys said in your prepared remarks, you gave some color on what leasing was looking like. I couldn't tell if you were talking specific to the first quarter or the second quarter, both. I was hoping you could kind of just dig a little bit deeper into what you're currently seeing. Secondly, as it relates to the move-in rate, and I appreciate that you're being more conservative in that number right now, but would you expect at some future point, and I appreciate you're not gonna tell us what quarter that is, or you might not know what quarter this is, but would you expect for all this to kind of catch up?
In other words, would you expect at some point we're gonna see a very sizable quarter or two to kind of make up, if you will, for the lost ground, considering these developments are actually still going on, and at some point everything's gonna get completed, and also we get to a point where all the installs are kind of back on track? If so, if that's the case, would you expect then to see a notable impact on outer year expectations, or is this really kind of focused on a slower 2020, by 2021, we're kind of back on the trajectory that we may have previously been assuming?
Good, thanks, Colby Synesael. On the first one, as we're nearing the end of the first quarter, we already know what we've done from a sales point of view in the first quarter, we just said comfortably over 20,000 sq m organic. We already have, I would say, a very good idea of what we're going to be able to do in the second quarter, on top of which, hopefully, the BJ 10, 11, 12 acquisition will close in the second quarter, that's 20,000 sq m there as well. If you add all that up, versus a full year target of 100,000 sq m, we said we're going to be a long way, which I certainly mean more than halfway. Maybe quite a bit more than halfway towards that target.
I just mentioned that because we know that as a fact and to give some confidence in what we're saying in terms of our sales targets. The second part, what you said could quite possibly be the case. We looked at different scenarios. We spoke to our customers. I listened to the other earnings calls, people not predicting what the shape of the recovery is. We could have assumed there would be very little movement in the second quarter and an enormous ramp-up in the third and fourth quarter. We decided in the end just to take a haircut to the numbers for each quarter. It's only a few thousand sq m. The resulting reduction in revenue and EBITDA was kind of the same, whether you're looking at it V-shaped, U-shaped, or whatever. Interesting statistic that I'll throw out.
The last, I guess, five or six weeks, from just before Chinese New Year until a few weeks ago, there was no move-in. The amount of capacity that was being utilized in our data centers was static. During that time period, our customers' power usage went up by nearly four percentage points, which means, in simple terms, that they're running their servers at higher utilization rate, higher than normal and higher than they would normally do given their operational parameters. That's indicative of requirements to deploy more capacity. That's why we said customers want to move in. It's really a question of whether they can. I believe that most of the current inventory of servers has already been deployed. The next wave of move-in is dependent on the production and the supply.
If that comes through quite quickly or in size, yeah, I would actually expect quite a sharp ramp-up in move-in. Yeah, that might be what we describe as risk to the upside.
Great. Thank you.
Your next question comes from the line of Gokul Hariharan from JPMorgan. Your line is open. Please ask your question.
Thanks for taking my question. Hope everybody's safe. Just first question on, could you talk a little bit about, maybe William, on how you're seeing the dynamic of better-than-expected demand due to work from home and more digital consumption, et cetera, versus the relative lack of ability to execute on move-in or not having enough components, like companies like Tencent also indicated yesterday that they are facing some degree of tightness in terms of some capacity, some hardware or the other. Could you talk about how your customers and you are dealing with this, and what are they doing to kind of mitigate that? Are there any kind of near-term measures that they're able to mitigate that? What are you expecting? How is it going to manifest over the next couple of months?
Are we kind of past the peak of that, or are we still going to be in that kind of a phase in the next couple of months as well? Second question I had was on some of your remote sites. I think if you can take Langfang as a case study. You already have one data center under service. Could you talk a little bit about how the dynamics are shaping up in terms of operating a data center in a remote location, and how you are able to manage capacity ramp-up, and how the dynamics are for this kind of a data center compared to a city center kind of data center? One last question, if I may. On the financing side, I think you clearly explained where you stand in terms of availability of financing, both equity and debt.
This is an industry which saw a lot of financing come in over the last couple of years for your competitors as well, especially private equity as well as other sources of financing. Can you talk a little bit about industry-wide in terms of what you're seeing on financing pipelines? Are they still intact? Do we see some degree of compression on financing on an industry-wide basis? Thanks.
William? Yeah. First question. Sorry, Gokul. First question was, what can be done about the issues in the supply chain? Have customers got any solutions? We got any solutions to-
Yes. The customers are giving us so much.
Yeah.
What's my question?
Gokul's asking, is there any solution to that problem? Is there anything that we can do?
Yeah. We can do nothing, frankly speaking. We are not in this industry, right? Server industry. I think what I heard about it, our customer is try to get more server as much possible in the current market, right? I think they push a lot of the supplier to get some inventory in other country or other market to try to mitigate the impact of the supply. That's what they are doing. So far, we don't know what's the percentage of the target they can achieve, right? In general, they are doing their best right now with what our customer told us. Looks like Q2 maybe will catch up the revenue a little bit. That's my current view.
Okay. Second question is about operation Langfang. Maybe I'll go first.
Yeah
You can top from it. Yeah. Gokul, the Beijing municipal government started to introduce restrictions on new data center approvals in the end of 2016, and it became apparent in 2017 that it would not be possible to maintain sufficient supply to fulfill demand within the urban area. We started then to work on the backup plan. I'd say it took about two years of discussion, interactions, and so on with the Langfang government, resulting in a framework agreement which addresses allocation of substantial amount of power, in fact, substantial amount of power which is available in that area, and the sale of greenfield land and investment. We chose Langfang because the telecom carriers had already established major data center hubs in that market. Therefore, the connectivity, at least from that part of Langfang, not the whole of Langfang, but that part into Beijing is very good.
In terms of latency, it's only a small drop off versus the latency within Beijing. In order to proceed, we really had to cover two bases. One was the government, and one was our customers. We had to convince our largest customers because a place like this, you're not looking for just one order from a customer, you're looking for a customer to deploy a major amount of their own capacity. At the end of the framework agreement, we're in a position to acquire the greenfield land and start the development. We'd already got our customers wound up. We found that we needed to accelerate our time to market. In actual fact, the first thing we did was acquire the land which is for Langfang 3, 4 and 5. It should've been Langfang 1, 2, and 3
Because of the time to market requirement, we leased a number of buildings. We leased Langfang 1, we leased Langfang 2, we leased Langfang 6, we leased Langfang 7. That's why we have five data centers in that area, all 100% committed. Going forward, the intention and the better approach would be to have all of our development on the greenfield sites on the campuses. That here was just out of necessity. Your third question about financing. Yeah, I'm clear about our own position. We have access to the public market from time to time. We also have access through partners and some very significant institutions, PRC institutions, who have also expressed an interest in working with us. I think there's a lot of scope in that approach to sourcing equity and other value add.
As far as the competition is concerned, no one is anywhere remotely close to our scale. There's been a limited amount of private equity participation by foreign PE and domestic PE. You know from the case histories of our acquisitions, that there's quite a few projects that are being undertaken with no equity and frankly, no formal debt either, just a reliance on credit from suppliers. I don't think that on the whole, our competition is particularly well-capitalized or financed, but I don't mean to damn them all. I'm sure that there are good companies amongst them.
Yeah, Gokul, I added more color to your second question. In original, our plan is to supplement supply to the Beijing supply, right? I think the people, because a couple of years ago, we realized the carbon footprint is getting more tight in the urban city, and the demand is huge, right? In the first original plans, we tried to convince our customer to move a little bit, shift a little bit of today demand to the edge of town. Now customer and us realize those product is different. They have the same latency sensitive criteria from our customer. Our customer need a more big scale and a more big power capacity. They want to deploy more big power density. Now we realize this type of the hyperscale campus is an independent product.
They fulfill our customer latency sensitives, a high visibility for the future supply and a hyperscale development and a hyperscale power capacity. This cannot be replaced in the urban town. That's my view. Now it become a new product in our view, right?
Thanks, William.
Your next question comes from the line of Dan Wang from Macquarie. Please ask your question.
Hi, management. Congratulations to the strong results. I guess my first question is, in the past GDS is doing a very excellent job in managing the financial leverages. I do see you guys are guiding a very strong CapEx spending in 2020. I guess I want to follow up on the previous question on, do we funding this CapEx purely from the bank loan side or we need additional equity raising this year? My second question is, can management share some colors on whether the pandemic is actually helping or slowing down the utilization ramp-up in the first quarter? Thanks.
Okay. Dan, on the first question, our approach is to raise equity capital ahead of requirement so that as we initiate new projects, we can allocate from the capital we have in hand to capitalize new projects on an individual basis. We try to maintain around sufficient capital to capitalize around two years' worth of new projects. The end of last year, we ended last year with RMB 5.8 billion of cash, most of that effectively is the equity for future projects. I said that we would need about RMB 3 billion of equity for the new projects in 2020. That means we need to use about half of that. There's also operating cash flow coming through. Yeah, it would look like we have sufficient equity to get us through two years. The debt side, in normal circumstances, is just about execution.
It's about having relationships and a track record, and sound project fundamentals and so on. The situation we're in actually is that we require about RMB 4.5 billion of additional debt to finance our CapEx this year. We already have RMB 2.5 billion of it in committed but undrawn facilities. The remaining RMB 2 billion, actually, we're working on RMB 3 billion of new debt facilities right now, some of which are almost done. There really isn't any financing risk to what we're planning to do, at least for the next one to two years. The second question is whether the virus impact is actually slowing things down or speeding things up. I suppose, I think probably you could say there's a difference between slowing things down in terms of moving, but maybe speeding things up in terms of demand.
Right. I think it's neutral. They're moving a little bit. Stay on the current point. What we can say is the server supply chains looks like uncertain, right? It will impact our customer moving this quarter or maybe a little bit next quarter. What we can tell is in China, inside China, the manufacturer is recovered right now. I think it is a little bit positive for the Q2, as I mentioned just before. For the demand side, I think everybody know this current situation will let a lot of the internet SaaS player, a lot of internet player, and a lot of e-commerce player, a lot of the application were well educated introduced to the market. I think in the midterm, long-term, maybe it will drive more demand in the future.
Great, thanks.
Your next question comes from Arthur Lai from Citi. Please ask your question.
Hi. Good morning. Good evening. This is Arthur Lai from Citi. I have two quick question maybe to Dan. The first question is, can you talk about the contract renewal schedule? The reason we ask this question is, we recall on IPO stage, you talk about contracts sometimes go with four years or even longer. Are we in the middle of the negotiation with the clients? Dive into the detail. In the Tier 1 city or in the retailer client, was the pricing trending up or can we get the better EBITDA margin for the new contract? That was my question. Thank you.
Thanks, Arthur. On page 36 of our earnings presentation is a summary of the amount of capacity which we have coming up for renewal in each year. In the current year, it's 23,000 sq m of capacity. It's 8.9% of our total committed area. It's around the similar level in each of the next few years. If you delve into the individual contracts and who are the customers, there's very little cloud or internet business coming up for renewal. It's mostly enterprise business. Those contracts are one, three, five years, so they renew more often. They renew automatically quite typically. I think this question arises from analysts and investors. I know what you'd like to get at, which is to see some benchmarks for what the pricing will be when we do get to have significant renewals with the large cloud and internet customers.
That's not going to happen this year. I won't be able to give you any empirical evidence on that. Going back to what William said about our layout now or configuration in terms of downtown and edge of town. If customers don't want to pay the price for downtown, we have an alternative option, the edge of town. Certainly, we believe the downtown capacity has increased in value a lot.
Yeah.
Maybe our ability to achieve higher selling price is there. That is something which we can certainly offer to customers.
Yeah.
I think the second question was if you talk about the price, actually, you mentioned EBITDA margin. We have to talk about the price together with unit CapEx, and then in terms of a return on investment. For us, we look at it in a very fundamental way. We talk about IRR. We can't really talk about it in terms of EBITDA margin, because that doesn't really tell you what's happening in terms of the project returns. The unit CapEx has been coming down. I said the majority of it, the P&E, has been coming down by about 3%-4% per annum. The MSR, revenue per square meter, at least in particular over two or three years, has come down, I think, by about 5% per annum. You can see that the degree of decline is quite close.
What that tells you is that actually our returns must be pretty well sustained if the yield and the investment costs are moving in line with each other. That is indeed the case. From IRR, from an NOI yield point of view, we are still achieving the same kind of returns over the last two to three years, which is exactly what we target to do.
Yeah. Arthur, I can answer your question. I mean, add more color on that. Number one, our customer now, how they look at the data center right now is the same view with us. Number one, they needed a urban town data center, because in the future, it will adopt to the edge data center. It will benefit for the edge data center demand. On the other hand, edge town always fulfill the mission-critical system, as we mentioned before. I think in the future, edge town data center will be more valuable. If our customer likes to low latency, can tolerate a little bit latency issue, and they want a big scale close to the Tier 1 market, we can offer the edge uptown product to them.
If they want to just pursue the cost-effective, I think we already set a model to build, to suit for our customer in a remote area, right? This is three different product, which we well-structured to our customer.
Thank you.
Your next question comes from Frank Louthan from Raymond James. Please ask your question.
Great. Thank you very much. Looking back to your comments earlier on the government's position on data centers as critical infrastructure, will that new position make the industry more competitive? Is it going to encourage new entrants or relax foreign companies' ability to own and operate data centers? Then to your comment you just made sort of on pricing and your cost inputs, do you think your costs go up in the short term if labor is in short supply? Thank you.
Yeah. The first one is whether government policies could lead to more competition.
I think our view is if government are more flexible to release the carbon quota in the Tier 1 market. If you look at it in GDS in the Beijing market, for example, in Beijing market, we lost a lot of deal last couple of years because we constrain the carbon quota. If the government release more carbon quotas, and we believe we will do more business. We get more market share in the Tier 1 market. If the carbon quota barrier is equal, that means customer will more focus on the value of the service provider. Our major customer, they are now looking at the service vendor as not just a capacity. They have a lot of different criteria. That means if the carbon quota barrier getting lower, that means the other criteria will be more focused on from our customer.
We are well-positioned in our other value, right? Daniel, you want to say more?
Yeah. Frank, on the labor cost, I'm not sure if you're talking about the cost of revenue or how it affected construction costs. On the cost of revenue side, most of our staff, headcount by number, is in data center operations. It's mid to high single-digit percentage of our revenue, so it's not the biggest cost item. It is one of the parts of our cost structure in which we're getting quite a bit of operating leverage, because whilst there's a certain number of people who have to be dedicated to each individual data center, there's also quite an amount that can be centralized. We don't see anything out of the ordinary in terms of inflation there. Most of our people are back at work, actually. I think our data centers are fully staffed. On the construction side, the initial delays were caused by government restricting activities.
Once construction resumed, construction workers who came from other parts of the country had to go through quarantines. It took some time for the number of workers on site to reach the full complement. It's still not there. I think it's probably like 5,000-6,000 construction workers employed by our contractors. Across our 16 sites, maybe 75% are back in place. I don't think this is a fundamental shortage. It's just a transitory thing as people come back to their place of work.
Okay, great. Thank you very much.
As there are no further questions, I'd like to now turn the call back over to Laura for 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 the website or the Piacente Group Investor Relations. Thanks all. Bye-bye.
This concludes this conference call. You may now disconnect your line. Thank you.