Thank you for standing by, welcome to the NEXTDC Limited first half 2019 results announcement conference call. All participants are in a listen-only mode. There will be a presentation followed by a question and answer session. If you do wish to ask a question, you will need to press the star key followed by the number one on your telephone keypad. I would now like to hand the conference over to Mr. Craig Scroggie, CEO. Thank you. Please go ahead.
Thanks, Jody. Good morning, ladies and gentlemen. Welcome to the NEXTDC results presentation for the first half of the 2019 financial year. I'm joined here today in Sydney with our CFO, Oskar Tomaszewski. We begin on slide two. We're very pleased to present another set of record results with total revenue over AUD 90 million and underlying EBITDA of more than AUD 42 million for the half. Contracted utilization grew strongly on the back of what was our largest ever sales half, finished December with more than 50 MW under contract. Our ecosystem continues to expand. We added more than 2,500 interconnections over the past 12 months to finish the first half with almost 10,000. This is up 34% from the same period a year ago. Our Self-Serve go-to-market platform also continued to evolve.
We're now close to 1,100 customers and more than 500 partners, of which more than 60 are connectivity and network service providers. S2 in Sydney opened for early customer access in the first half, with development ongoing. Our P2 microsite and connectivity hub also opened to facilitate early access to the INDIGO submarine cable. On slide three, our strong performance in the first half was highlighted by robust key operating metrics. Revenue from operations increased by AUD 13.3 million. Contracted utilization increased by a record 11.1 MW or 28%, and interconnection revenue has grown to 7.7% of recurring revenue, which was up from 6.2% in the same period. Our results continue to demonstrate the benefits of the company's inherent operating leverage. Underlying EBITDA increased to 26% AUD 42.2 million.
Operating cash flow was AUD 15 million, after approximately AUD 20 million of net interest paid and AUD 6 million in one-off payments related to the acquisition of APDC. We remain well-capitalized to support the company's growth plans. Total liquidity at 31 December was AUD 644 million, inclusive of about AUD 300 million senior debt facility, which remains undrawn. Our balance sheet positions has never been stronger, which is now underpinned by over AUD 1.6 billion of total assets. At 31 December, we held property with a carrying value of AUD 581 million and plant equipment with a carrying value of AUD 594 million. Our network continued to expand at a rapid pace.
Next two was obviously opened in the first half for our early customer access. The P2 microsite was open to facilitate early access to the submarine cable system, as well as other important telecommunications and cloud infrastructure providers in the W.A. market, which is central to our retail colocation strategy. Finally, we completed the acquisition of the underlying land and buildings at P1, M1, and S1. We also acquired B1 in that time, consistent with NEXTDC's long-term strategy to now own the underlying properties. I will hand over to Oskar to discuss our financial results. Thanks, everybody.
Thank you, Craig. Let's now turn to slide six, a summary of our full year profit and loss. The statutory results reflect data center services revenue of AUD 84.1 million, an increase of 15% on the corresponding period last year. Net loss after tax of AUD 3.1 million, a result which includes the impact of increased depreciation costs for our newer facilities, increased finance costs, and AUD 8.5 million in one-off costs related to the acquisition of APDC. As previously advised, NEXTDC adopted the new accounting standards, AASB 9, 15, and 16, from 1st of July 2018. Our non-statutory highlights include underlying EBITDA of AUD 42.2 million, an increase of 26% on the corresponding period last year. This underlying result excludes distribution income from our previous holding in APDC, transaction costs including landholder duty related to the acquisition and wind-up of APDC, as well as gains on the extinguishment of property leases.
Direct costs of AUD 16.9 million, which rose in line with contracted customer capacity and power costs. The net impact of direct costs relating to energy prices after net increases in power consumption was approximately 15% of total direct costs. Facility costs decreased to AUD 8.3 million from AUD 14.9 million, primarily relating to rental cost savings as a result of both the adoption of new lease accounting standards AASB 16 and the acquisition of the underlying land and buildings at P1, M1, S1, and B1. Corporate overhead increased to AUD 17.2 million from AUD 13.6 million, primarily relating to investing in the staffing, support, and early operations at M2, S2, and P2, reflecting the new investment and growing of our second generation of assets, as well as the associated increase in centralization as we continue with our network and cloud expansion. On to slide seven.
Revenue generated from racks, suites, Cross Connects, and other recurring sources accounted for 96% of total data center services revenue, an increase from 90% in first half 2018. Note that a key driver of this increase is the adoption of AASB 15, the new revenue accounting standard, according to which project revenue is now mostly deferred and recognized over the term of the underlying customer contract. Rather than recognized upfront as was the previous policy. The underlying EBITDA performance highlights NEXTDC's inherent operating leverage, as demonstrated by the continuing strong earnings growth in excess of revenue growth. Slide eight sets out our revenue per unit metrics. Both of our annualized revenue metrics have grown strongly during first half 2019, benefiting from contracted price escalation and increased connectivity, power density, as well as power recharge revenues.
It's also worth noting that revenues from larger ecosystem enhancing customer deployments increase over time due to higher usage of contracted power capacity, increased demand for interconnection, and the use of them to services over time. Slide nine summarizes our balance sheet position and cash flows. At 31 December, NEXTDC held property with a carrying value of AUD 581 million, as well as plant and equipment with a carrying value of AUD 594 million. Our net assets stood at AUD 882 million. Finally, we remain well capitalized to continue our growth trajectory, with total liquidity comprising cash and underlying debt facilities of AUD 644 million. I'll now hand you back across to Craig to go through our business performance and outlook for the 2019 financial year.
Thanks, Oskar. On slide 11, our non-financial metrics are set out. Total number of customers up 25% year-over-year to 1,090. Total interconnections grew 34%, and total Cross Connects per customer grew 7% over the same period to 9.2. On slide 12, further insight into the diversity of our business. Customers by industry shows strong representation from cloud and connectivity, with continued solid growth from the enterprise. There's a shift towards higher density deployments, and that reflects the growth in the hyper-converged infrastructure and what we're seeing in the emerging development of hybrid clouds. On slide 13, at 31 December, 90% of our billed capacity was sold, and 73% of the sold capacity was billing. The increase in unbilled but contracted capacity further underpins our confidence in the forward revenue and earnings outlook.
Both charts highlight the strong growth experienced by the company in recent periods, reflecting high levels of utilization with further operating leverage still to come. On slide 14, our capacity and utilization. In Sydney with S2, we already sold close to 15 MW of capacity, and the ongoing strength of demand in that market has provided us with the confidence to pull forward total capacity to 22 MW of the planned 30 MW in S2. In Perth, we completed the land acquisition for P2 and immediately commenced construction. The P2 microsite and connectivity hub was opened, and we now provide early customer access at this site, not only for the INDIGO subsea cable, but to continue to develop the diverse retail ecosystem.
You would note that we recently announced not only Amazon connectivity services in the Perth market, but we've taken Microsoft to Perth as well, which is a great win for the company. In P1, we also opened the fourth and final data hall. No longer an offspring to Perth. In Brisbane, we opened the second data hall of P2, a facility which recently achieved accreditation for the Tier IV Gold Certification of Operational Sustainability. That's a great achievement for the team. It's a first in the Southern Hemisphere. In Melbourne, M1 also achieved a first, a first for the Australian data center industry, which was a NABERS five-star rating accreditation. A great result for the team from an engineering and efficiency point of view.
On slide 16, revenue guidance of AUD 180 million to AUD 184 million is underpinned by strong growth in recurring revenue and our long-term customer contracts. Obviously, our rich ecosystem continues to develop, which is driving very strong demand in our connectivity solutions, and you can see that reflected in our interconnection numbers approaching now 10,000. Note that our revenue guidance no longer includes distributions from APDC for the second half, as we've now fully acquired, finally, and consolidated the properties. We expect underlying EBITDA between AUD 83 million to AUD 87 million, with scale and earnings growth now driven by our Generation 2 facilities, and obviously you'll see that continue to scale. Total CapEx is expected to be between AUD 430 million and AUD 470 million, but that excludes the acquisition of APDC and the B1 properties. Our guidance for underlying EBITDA and CapEx remains unchanged.
The only change that you'll see is the change in the revenue number as we no longer receive APDC distributions. In summary, it was certainly a massively exciting half for the company. It was our biggest half, biggest sales half, biggest half, obviously, in interconnections and revenue growth. From a performance point of view, very, very pleased with where the company is today and obviously continuing to prepare ourselves to take advantage of that very large shift, not only to retail colocation as customers continue to move out of legacy on-premise facilities, but obviously the massive growth that we're seeing in the hyperscale cloud. That continues to give us great confidence to invest in the future for the business and for our shareholders. Ladies and gentlemen, that is the presentation today. I'm now going to ask Jody to open the line for questions. Thanks.
Thank you very much. If you do wish to ask a question, please press the star key, then one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press the star key, then two. If you are on a speakerphone, please pick up the handset to ask your question. Thank you. Your first question today is from Kane Hannan from Goldman Sachs. Go ahead. Thank you.
Good morning, guys, just three from me, please. Firstly, in terms of that revenue guidance, can you confirm all those revisions are purely relating to the transactions and the interest impact? Secondly, the nine megawatt Sydney contract you announced in November. Can you just provide us with a bit more color around the terms and pricing of that contract, if you can, and I suppose the returns expectations for those megawatts. Finally, just cash conversion during the first half, it looked a bit softer than I was sort of expecting, given I thought you were going to unwind some of the issues in the second half of last year. Just comment around more the impact on the cash conversion in the first half, please.
Thanks, Kane. I'll take those three questions. Firstly, on revenue guidance, I can confirm that the change to the guidance is lower expected distribution income. Obviously, we're not getting any more distribution income from APDC, as well as lower expected interest income because of the funds that we paid for Core APDC. The second question related to the contract or contracts that we announced towards the end of the first half. I can't comment beyond what was in the ASX announcement. I can't go into any commercial terms. Those are commercially sensitive. In terms of operating cash flow conversion, I would note that our operating cash flow is after the expensing of approximately AUD 5.8 or paying approximately AUD 5.8 of transaction costs related to the acquisition of APDC.
That was a one-off in nature, as well as after the incurring of approximately AUD 20 million of interest payments related to our debt facilities.
Perfect. Thank you, [Shaf].
Thank you. Our next question is from John Aiken from ABC Capital Markets. Go ahead, thank you.
Thank you. I was interested in maybe hearing you describe your outlook in terms of the sales pipeline and how does that look in kind of the hyperscale segment as well as in your enterprise segment. The difference in Perth versus Melbourne versus Sydney, maybe kind of a geographic and then segmentation, a couple comments on that in terms of sales pipeline. On the strategic front, just interested in kind of an update on whether there's line of sight towards land, power, and development projects in places like Hong Kong, Japan, Singapore. Thank you.
Thanks, John. I'll take those ones. Sales pipeline continues to, obviously, be at a record level. Both in terms of retail. If I start with core to our business strategy and continue to be focused on the development of those, the strategic ecosystems of customers. We're not exclusively looking to build a hyperscale wholesale business. That's obviously an important segment of the market that's lower rate of returns, if you look at the global benchmarks. We can favor and look favorably at Equinix's business model as a benchmark to NEXTDC's development for our ecosystem. They've certainly built an extremely diversified and massively successful business on a global scale, pursuing that blended business model.
We believe the same is true, continue to focus on retail colocation, focus on internet connectivity and cloud connectivity services, and the investments that we continue to make in the AXON cloud on-ramps to Microsoft, Amazon. You'll note that obviously we continue to take those cloud providers into new regions in Australia. Yeah, retail and wholesale pipeline is at record all-time highs. Certainly, we can't build data centers in every location and what we're seeing in the market, particularly in Sydney, is that the multiple availability zones continue to emerge. We will do our best to keep up with demand in the availability zones that we operate in. It's difficult to do it everywhere, and obviously we don't have unlimited capital. We're focused on our business model and what makes NEXTDC unique.
In a similar regard, obviously, we continue to look closely at markets like Singapore, Hong Kong, and Japan, and there are great opportunities in those markets. The dynamics are very similar, both from a client requirement point of view in hyperscale and in retail colocation, and in the type and style of product that those customers are looking for. There's some great opportunities there. It's really just a question of time and prioritization of our capital. As I've said before, I hope at some point in time we are capable of expanding the company internationally when we're ready to do so.
If I can follow on briefly on the first question, are you noticing any change in lengthening or contraction in decision cycles on the part of some of the larger customers as they seek to kind of locate additional capacity in Australia? On the enterprise side, is there any kind of change in your distribution mix between partners and direct sales?
We're obviously just recently in the U.S., John, as you would know, working with all of the larger customers in the planning period. I don't think the cycle necessarily has changed. Certainly, the amount of capital that we're deploying has changed. It's materially increased, and we're seeing an increase in the total pipeline size as a result of what used to be a one megawatt deal now being a five megawatt deal or a five megawatt deal, which would've historically been extremely large, now morphing into 10+ megawatt type transaction opportunities. For a relatively small market like Australia, they are very, very big numbers. We remain enormously excited about the potential opportunity to continue to grow with those partners.
We continue to be very disciplined in context of our return expectations, and we don't expect that we'll have 100% of the market, and we'll continue to look to invest with customers where we can bring value, grow our particular business model, and benefit from a blended retail and wholesale colocation business. On the partner side, partners is very, very important to our strategy. It was early on, and the partner strategy continued to evolve. Now with more than 50 partners that are reselling and i ntegrating NEXTDC services into their go-to-market strategy.
That productization processes that we work through with each individual partner is critically important. Our partner team plays a very important role in differentiating how NEXTDC go to market and why the value that we offer to partners is different to other operators in the data center industry. Frankly, the continued investment and development of that strategy will see the company continue to diversify its product portfolio and offer a deeper range of services to customers as we build out our footprint over the coming 12, 24 months. Yeah, the partners is critical to our success, and will continue to be.
Thank you.
Thank you. Your next question is from Tim Plumbe from UBS Investment Bank. Go ahead, thank you.
Hi, guys. Just a couple of questions from me, if that's all right. Craig, firstly, on pricing revenue per meter squared, revenue per megawatt, up about 3%, half on half. Looking good there. Can you talk a little bit around pricing in the market at the moment, what sort of pricing impact you're seeing when you're rolling over contracts, et cetera?
Thanks, Tim. On contract growth year-on-year, obviously, we see standardized CPI plus type escalators every year for the install base, and that's fairly common, obviously, across the data center industry. In terms of retail pricing, obviously our benchmark is Equinix, and we continue to view Equinix as the global benchmark for excellence in retail data center. I just met with the Equinix local managing director, Jeremy, last week. Had a good conversation. Their business, by all accounts, seems to be continuing to shoot the lights out in retail. They've been very pleased with their acquisition of Metronode and continued investment in that platform. On the retail side, that's fantastic because we tend to find that in a lot of enterprise deals, if you look, say, at the Sydney market where Equinix are in the Mascot campus, we're in Macquarie Park.
A lot of transactions will be us and Equinix together on either side of the customer's infrastructure when we're building a diversified IT strategy. We're getting both ends from both suppliers. We tend to see and continue to focus on the value in that segment and see them as a very price-disciplined, quality operator of data centers. They're also very focused on the value of interconnection as we continue to be, because we see the importance of the role that we play in working closely with customers and partners to advise clients on how they will adopt and utilize both the public and private cloud platforms. The combination of those public and private cloud platforms with their legacy computing infrastructure. Obviously, that's what today drives the fastest-growing segment for hyper-converged infrastructure, which is really that blended ecosystem that we focus on, which is hybrid computing.
Yeah, Tim, I think pricing we've been pleased with. We continue to see a rational discipline in relation to the deployment of capital. On the wholesale side, obviously, we don't play exclusively in the ultra-cheap wholesale category. We tend to pick and choose when we participate in that segment. Global pricing seems to still be benchmarked. We look at the global pricing that's put out by [Diversity] and RBC, and we tend to see the similar pricing in the Australian market in the last half has continued to be reflected on what's going on from the top data center providers in the U.S. markets. Particularly Digital Realty, CoreSite, and probably CyrusOne are the key benchmark drivers for pricing in the U.S. We think that the Australian market probably reflects that in the last half.
Great. Just secondly on S2, can you give us a sense for how you would expect a major customer to roll out their new capacity into the second half of 2019 and into 2020? Can you give us a sense for what sort of EBITDA contribution you've got incorporated within your current guidance for S2 in the second half or how we should be thinking about average building megawatts in the second half for S2?
Yeah, sure. There's a fair amount of detail in that, Tim. Without picking the model apart, the high-level commentary I'd make is obviously, if you have a look, part of the reason or the key driver for going with our continuous development methodology, and it's quite a challenge to be honest, S2 was without doubt the most challenging development we've ever taken on. Building a multi-story high-rise data center and opening it while it's still under development was certainly one of the most challenging engineering feats that we've ever undertaken. As a company, I'm incredibly proud of what the engineering team has been able to do to build a multi-story, high-rise, hyperscale data center, open a portion of that while the ongoing development is done.
That was critically important to support our customers' needs, because they needed the capacity early, and as it turns out, they need it often. We are building as fast as we can possibly build to support those customers' continued growth requirements. You'll see the revenue come on during the course of the latter 2019 and early 2020 years. Essentially, it'll be in line with as quickly as we can continue to develop. If the development continues to go to plan as it is currently, or we can even potentially speed up. Largely, obviously, we've taken all of those factors into account in our guidance for the full year in the second half. Tim, in terms of very specific line item model-related things, they're probably questions that I think Greg would be capable of working through with you in more detail with your model.
Got it. Is it fair looking at first half EBITDA, I mean, 51% at the lower end of your guidance range, 49% at the top end of your guidance range, plus in the second half, you should be getting some sort of benefit coming through from S2?
Yeah, not a lot. The reason for that obviously is that we always give customers time to move in and ramp up in the early days. As you have with every new data center development, we give customers time to move in and get up to billing capacity themselves. That's reflected also not only in the commercial pricing, but in the tenure of the contract. Those larger longer-term contracts allow us more flexibility to support the customer to deploy their infrastructure and get some time to get their own billing up and running before they hit too high payments from a rent point of view.
Got it. Just last question, any update that you can provide in terms of the Melbourne hyperscaler market? Are we starting to see any signs of life there, or is it bit of a slow burn and kind of 12 months away?
Melbourne continues to be, I think, a city that's got massive potential. Clearly, the Sydney market is 5+ years in terms of growth in Melbourne, and you're seeing both second and third-generation deployments from infrastructure providers right across the board. Everyone's benefiting from those massive scale investments and the bigger platforms that are in their second or third generation are, as I said earlier, they're not deploying 1 megawatt, they're deploying 5 or 10 or even more in a number of cases. The Sydney market unquestionably is of extraordinary size, massive growth, and we certainly see well in excess of another 100 megawatts capable of being sold into the market in short- to medium-term here in Sydney.
Compared to Melbourne, which is probably in the first and maybe moving towards the second generation in size and scale, we support an appropriate amount of infrastructure in M2. If you have a look at the open, the first and second data halls, we sold those out largely in line with our just-in-time continuous development methodology. We've been building the third and fourth data halls. We'll be well sold into those by the time they're open shortly. We don't overbuild inventory. Again, remaining capital disciplined is the primary objective for us to ensure that we don't deploy our capital before it's required. For large, we build for retail capacity, 1 or 2 data halls at a time.
You see that when we obviously are building multiple megawatts and investing in a market like Sydney with additional capacity, it means that we've got commitments from customers. We will continue to be capital disciplined, and we'll deploy relative to the rate of growth for the market. I think it's probably fair to say that Melbourne has great potential, and we're in a position to take advantage of that as soon as the hyperscale platforms will grow to a larger extent in that market.
Great. Thanks, guys.
Thanks, Tim.
Thank you. Your next question is from Paul Mason from Evans and Partners. Go ahead, thank you.
Hey, guys. Just a couple from me. The first one, just wanted to clarify your comments about S2 in your guidance about a pull forward of 16 megawatts. Should I be reading that as essentially like pulling forward the completion of the facility? I think at your last contract announcement, you'd announced a pull forward of eight, which was going to be like 22 or so. Is this saying you're going to 30 now?
I'll take that one, Paul. We have six megawatts of capacity in phase one. Essentially, we're just re-emphasizing that we're pulling forward an additional 16 megawatts on top of that, which is consistent with our most recent capacity update, where we said we're building towards 22 megawatts. That's where that number sits.
Okay, great. Just on maybe the contracts in New South Wales, it looks like you're up to about 30.4 megawatts contracted now, which implies a bit of extra retail. I wanted to get a sense of that. Is that going into S1, or is some of that actually already contracted S2 retail as well?
We're not able to give any detailed splits by facilities, unfortunately, for some specific reasons in relation to what's commonly known as access disclosure.
Okay. Just on Melbourne, maybe. I'm taking your case study and then your accounts, and it looks like you've got about 0.8 of a megawatt contracted at M2, but it's putting out about just under AUD 2 million of EBITDA already. Historically, that would kind of indicate that you've got much more than that contracted. Should I just be looking at that as you've got effectively a lot more floor space contracted and it's just on really low power, or what else is going into that, meaning that the profits kind of coming in a lot higher than the megawatt contracted reported?
Hey, Paul. Great. Thanks. Yeah, I guess that sort of reflects the strong performance of the retail colocation business. Yeah, what you see largely deployed in M2 today is all enterprise. Enterprise obviously reflects a generally lower density to higher rate of return, where the hyperscale would be a higher density to lower rate of return. That's why we love the retail colocation business, continuing to focus as much as it's hard and it's disciplined and it's long sales cycle. It adds enormous value to the ecosystem, diversity, customer consulting and support services as they develop their hybrid cloud solutions. What's reflected in M2 today is pretty much a very dense retail colocation business, and that's why I seen those numbers. Pretty happy with it.
Okay. Just a last one, just on M3 and S3 timing. Are you expecting, say, the development approval process and stuff will pick up after the election? Or are you able to give any color on that sort of activity?
Yeah, sure. I'll take that one as well. S3, the Gore Hill site that we acquired, pre-DA, the design for the 80 MW data center is largely complete. We've been working with the council, getting our pre-DA approvals and other things, power, lockdown. We're pretty much ready to go on S3. It just comes down to working with customers on forecast commitments and other things. As we plan that, obviously, it's a very large data center, getting up to 80 MW to 100 MW type footprint sizes are massive scale, and so too are the type of customer commitments that we're making when we're building facilities of that size and scale. Yeah, we're well down the track there, and I'd expect probably not too far in the future, we'll be ready to get going on S3.
Yeah, we continue not only to work with the council on the DA approvals and design, but also work with the customers on commitments. In Melbourne, you are right. The Victorian election, obviously with the change of government and the development approval process in the Victorian market there was at election time. Now that we're clear of the election, we expect that we'll be able to get the DA approval and site acquisition process complete. We like to know that obviously we're going to have approval to develop the site before we acquire it. Once we get that certainty and now that the election is locked down and ministers have been appointed, we're now moving as quickly as we can to secure and sign the M3 site and hopefully produce something similar to what we're in the process of doing for S3.
Okay. Just one last follow-on for that, though. Can I just ask you, with your CapEx guidance, does that still have a component for the M3 land in it? I think it was flagged as having an undisclosed component when we last spoke. Yeah, just want to check if there's any change there at all.
No change there, Nick.
Okay, great. Thank you very much.
Thank you. Your next question is from Mitch Rivett from Macquarie Group. Go ahead, thank you.
Yeah, good morning, guys. Just a quick question. I couldn't see anywhere. Are you able to give me the actual impact of the accounting changes on the first half 2019 result in terms of revenue, EBITDA, and NPAT, please?
No, we're not.
Okay. Well, looking back at your FY 2018 presentation, we had some guidance there that that was only made at the time. I just want to know whether those estimates have changed or is that an expected change that we can flow through and just maybe the, I guess, contribution to the first half versus second half.
Mitch, if you turn back to page 22 of our FY 2018 results, we gave guidance on the goals for the accounting standards. We put a disclosure around a slight adjustment to the revenue guidance for reasons that we've already explained. Other than that, the guidance is unchanged. Beyond that, for any more detailed or specific questions, I believe you've got a session with Greg lined up a bit later on today.
Okay, thanks, mate. Maybe Craig, just trying to get a feel for the activity in the Melbourne hyperscale market, I guess over the last 12 to 18 months, it looks like, as Paul sort of said before, there's probably only been about 0.1 MW added to M2 over the last six months. Can you give me a sort of sense of how many hyperscale tenders that you might have worked on in terms of volume and what you think might have gone to the market over that time to competitors?
I'm not sure that number makes any sense, but in terms of activity in the local market, there's really only two zones. We created a third within two. Obviously between ourselves having one or two big pieces of the hyperscale puzzle in Melbourne today, and then the other piece of the hyperscale puzzle being out in Western Sydney. Still see plenty of activity there, but again, in similar fashion to what I described earlier, it's probably a generation 2 style of deployment. The size and scale is in that 1-5 MW range versus in Sydney, we're in the 5-10+ MW range. I would expect that Melbourne is an opportunity and will continue to grow. As the next refresh of infrastructure comes around, those larger, that initial deployment or first generation of deployments will grow materially in size.
At this point in time, we build enough capacity to serve the retail market, and once we have a high degree of confidence on the customer commitments in Melbourne, we'll commit to build more. That's obviously part of our M3 planning process. Clearly, size and scale footprint for M3 anticipates some larger deployments, and we continue to work closely with customers from a planning point of view. You're talking about 10-plus year planning windows on the size of the M3 and S3 deployments, and those customers are really getting into 10 or 15-year tenant commitments. They're quite large and long and detailed negotiation processes because that's a lifetime, 10 years in investment in our industry, and certainly securing contracts in the order of hundreds of millions of AUD for decade-plus long commitments takes some time.
We continue to be disciplined and patient and work closely with the clients on building something that's unique to support our business growth.
Okay. Thanks, Matt. I guess I was looking from FY 2018, there was 14 MW contracted in M1 and 0.7 in M2, so 14.7. The first half was up to 14.8 in Melbourne. Just wondering about how do you think we should expect to see M2 going in terms of contracted utilization maybe over the next 12 to 24 months?
Difficult question to answer, simply because on a retail point of view, retail tends to be fairly simple to forecast. We've got a base underlying solid run rate. We continue to perform, hyperscale does what hyperscale will do, and that is it comes in very large lumps and therefore difficult to forecast. Unfortunately, I can be sure of one thing, and that is it will grow. I just can't be definite and sure of when and to what extent. If I said it was 5 megawatts, it could be 10 or it could be 20. Unfortunately, it's just not something that we're capable of knowing with certainty. We are planning, importantly, to be able to take advantage of those size opportunities, because those size opportunities do exist in the market. The longer-term question will be, do they meet our return expectations?
We remain disciplined in how we deploy our capital, we remain disciplined in our rate of return.
Okay. Thanks. That's all from me. Thank you.
Thank you. Your next question is from Nick Harris from Morgans. Go ahead, thank you.
Thanks. Good morning, guys. I'm just interested in Perth. There's a few interesting things going on there at the moment. You've got one and nearly two submarine cables live there. I'm just wondering, do you have any sort of visibility on what the hyperscalers are doing, and is there the potential for Perth to actually pick up as an availability zone like Sydney? My second question was just also on Perth. Obviously, you've turned on a micro site really quickly. I'm just trying to understand, is that really a timing thing to obviously buy you time as you build P2, or is this a potential for NEXTDC to, I guess, get ahead of the next wave, which is that sort of 5G edge computing stuff? Is there a little bit more to that micro site?
Hey, Nick. Craig will take that. Great question. Let me try and work through that. There's a lot of detail in it. First of all, the cables. Yeah, look, we were thrilled, obviously, to be able to work with [Dirk Halleday] and his team at [CircleID] to secure the INDIGO cable, and obviously with people like Amnet and a bunch of other key telco players in there, being able to connect Singapore to Perth to Sydney on the lowest latency route, massively exciting for the team. That initially drove us accelerating the program of works with the micro site. What you'll see if you visit Perth and visit P2, the site is just an absolute cracker. Just a short walk over the bridge from the Perth Mall. It's beautifully positioned. Great location.
Couldn't be more excited about what's actually happening in Perth in the enterprise market. Having just recently announced that we signed a contract with Microsoft to take Microsoft into Perth after having, not long earlier than that, announced that we had signed a contract with Amazon to take Amazon into Perth. Building that retail colo business that's very network-centric with those big global cloud platforms for their on-ramps was a huge win for us. Fantastic result there. The teams have done an incredible job to continue to build out that value, and that drives our ecosystem that's being reflected in the growth of our connectivity numbers. For all the guys that work in the engineering team that build the AXON product, that's why we continue to focus on our differentiated business model, which we're passionate about, is solving that problem for the customer.
It's a real critical and complicated thing to resolve, and if we keep doing a good job there, we should see continued very strong growth both in numbers and in price reflected in our retail strategy. The cable piece is really important. As you know, Nick, because you cover the telco industry, submarine cables in the data center industry are data center fairy dust. When we get a submarine cable, it's a really great win. The opportunities that that opens up with international customers coming into the Perth market and then making their way to Sydney is pretty exciting. Having that opportunity is critical. As you alluded to, the work and development that we've done on the micro site strategy, so rapid access connectivity hubs, we call them.
The 5G and emerging component of the 5G, as we work closely with partners to sort of design how we can solve this problem, deploying micro sites into a larger number of locations may very well end up being an important requirement for data centers in a regional context. We've spent the best part of 12-18 months, the research and development team, specifically on the engineering side, in partnership with the company that we selected, a global leader in the development of those called Flexenclosure in Sweden. The Flexenclosure team worked closely with our team for 12-18 months to develop that product, and we're now capable of being able to deploy that product into any region or any location as a full rapid access connectivity micro site. We do see that as quite an interesting emerging opportunity. We think that that'll be client-led.
Some of the investments that we're making in that area will allow us to work more closely with regional areas. You'll see that yesterday the New South Wales government made an announcement for AUD 100 million allocated to regional data centers. We've been planning and expected to see not only a larger number of connectivity points to support the growth of the digital economy being made in regional areas around Australia, but regional areas right across Asia. We see. The microsite strategy and having a product that's differentiated and network-centric, and obviously tier 3 or tier 4 resilient to the standard that we've become renowned for building an industry, as a really important piece of our strategy. Yeah, Nick, you're right.
There'll be more that we'll share on the microsite strategy coming, because we see it as an interesting opportunity to support the edge growth of data centers, the IoT deployment into regional areas, whether it's been supporting autonomous vehicles and other things. Compute continues to need to go close to the user, both for not only the download of data, but the creation and the upload of it. Last point, Nick, on the hyperscale side, obviously the total size and scale of the P2 development, it will be a continuous development methodology site. We'll build the site in stages. We won't deploy all the capital again. We do see an opportunity for many in the countries in Asia, whether that be Singapore or Hong Kong. Land prices are very high. Data center prices are high relative to what you could secure in the Australian market.
Therefore, it stands to reason that it would make sense if you had diverse cable routes and that latency wasn't an issue, that you would be in a politically stable, comfortable environment that had reasonable energy prices and had all the connectivity both in and out of Australia to Asia and the U.S. Hyperscale is an opportunity there. We just saw Equinix recently announced further investment in the Perth market as well. Yeah, that's continued to also give us confidence that there may be some hyperscale development in Perth in the coming periods.
Great. Thank you, Craig.
Thank you. Your next question is from Bob Chen from Deutsche Bank. Go ahead. Thank you.
Morning, guys. Just a question on the enterprise market. Are you seeing any sort of uptick in activity in that part of the market? Just looking at sort of the run rate enterprise deals that you're winning. It's sort of tracking at that two, three megawatts a year, and it's been tracking at that for a little while.
Yeah. The enterprise market obviously is a long sales cycle from the time, if you imagine that from the time that you first engage with a customer, what drives the moving the data center largely is if they're moving office. A lot of the legacy infrastructure will be located in an office building in their CBD. They'll be moving offices, so that generally is a sort of 18 to 24 month planning window. We start working with a customer design.
The data center in of itself is not massively complicated, but where the real value comes is the customer being able to then connect to the public and private cloud infrastructure that we host, which is why the retail colocation data centers that have diverse ecosystems and cloud on-ramps become very compelling from the client's point of view, because they're able to get material cost savings, both in a telco context, metro fiber is cheaper in the data center. Connecting to the cloud platform is cheaper in the data center as well, because you're not then paying for metro fiber out to the cloud as well. You're actually buying a direct cross connect, either an elastic cross connect or a physical cross connect in the data center.
When you've got the cloud in the data center, that's going to be not only the lowest latency, but the lowest cost path to connect to the cloud. Enterprise takes time. The run rate and size of the enterprise business has been consistent. As I mentioned earlier, there's plenty cheaper alternatives, but we continue to see the benchmark standard in the retail colocation industry with Equinix and really ourselves and Equinix tend to be on both ends of the deal. It's been consistent. It will continue to grow. Very few companies wanting to build and operate their own data centers, certainly in buildings when you've got NABERS 5 and 6 star ratings and green energy requirements, ESG and other sustainability standards. That's also why we've recently made announcements on our carbon neutral. We're now NCOS accredited carbon- neutral as a company for NEXTDC.
We'll offer our customers the ability to be able to offset their carbon as well for their data center footprint, then extending out our investments from our solar arrays into the renewable energy space for more PPA style of partnership agreements as we've done with the Melbourne Renewable Energy Project in partnership with NAB and Australia Post and others. Expect more of those. They seem to be very important to the enterprise customer when they're selecting the right long-term data center partner, because they want to know that not only are they getting a world-class product, but they're getting world-class sustainability and operational credentials to support them for the long term as well.
Sure. Then just in terms of the overall competitive environment, especially in the Sydney and Melbourne markets, I mean, we're seeing quite a lot of investment from some of your competitors, including Equinix just recently. Has that had any sort of impact on sort of your business?
I thought, obviously recently caught up with the Equinix guys and caught up with Digital Realty just recently. The key players all continue to invest. I see that investment as a good thing. From a competitive point of view, people tend to sort of see investment and think, "Oh, that means that you're going to be up against more price competition or you're losing an opportunity." The one thing that's important to continue to remember about data centers is it's not a zero-sum game. We're not selling widgets. Where the data center is located is critically important. Customers generally need two or three or more locations. In a lot of cases, when we're working with a client, we are working collaboratively with the client, but they're also selecting one or even two competitors at the same time.
Price does play a role, and obviously we are not a price-led company. We are an operational services excellence led organization. We focus on Tier IV data center products, the highest standard in the industry, the highest levels of operational excellence and customer service, and we are premium priced. Customers will choose multiple locations, they'll choose multiple providers, and in the majority of cases, we tend to see good competitors as continuing to keep us focused on building a better quality product and being more disciplined about how we run our company.
Sure. Thanks, Craig. Then maybe for you, Oskar, just looking at the sort of guidance for the full year, it seems to imply there's a sort of margin compression in the second half. I would've thought acquiring APDC and saving some facility and rental costs there would have sort of flowing through into better margins.
Thanks, Bob. We're tracking pretty much in line with expectations. There can be some seasonality in some of our operating costs. There's a whole range of different factors underpinning that. That could be software costs, that could be data hall cleaning costs, engine maintenance, and so on and so forth. We're pretty much tracking exactly in line with where we thought we'd be at this time.
Okay, great. Thanks, guys.
Thank you. Your next question is from Wayne Arthur from Monaro Superannuation. Go ahead. Thank you.
Good morning. What are you going to do to get there? Every year, every presentation, we see figures on utilization. For six months, the utilization is 90%. It's always been in the high 80s, and that's pretty good. Even if you got to 100% utilization, that wouldn't significantly increase the profit. We've got assets of AUD 1.7 billion, we've got no profit, and we've got no return on equity. Now, I noticed there's been references to your competitor, Equinix, and I noticed in the past year, they've recently recorded profit of 7% of revenue. What's NEXTDC going to do to finally generate some decent profits for shareholders?
Yeah, thanks for the question. The answer is a long one, but in summary form, we're essentially continuing to invest. If you look at Equinix's historical performance at a similar stage of development as NEXTDC finds itself in at the moment, they too were investing a lot of capital for future growth. We are building in the platform for future expansion, there is no other way of taking advantage of market opportunity other than investing that fund. We could slow down. We could stop growing. We could start generating nice profits, we miss out on all the growth that's happening out there in the market.
Is the company actually undercharging?
No. If you have a look at what's driving a lot of our costs, the low EBITDA, it's essentially costs as we take on more debt to fund the ongoing development of future data centers, as well as the associated depreciation. Each new incremental data center that we build is larger than the previous one. A case in point in Sydney, the first generation data center is a 16 megawatt facility. The second generation is 30 megawatts. When we get around to the third generation, it's an 80 megawatt facility. The market is expanding, we are investing to take advantage of that opportunity, unfortunately, in the business, you have to invest up front and start incurring the cost before the benefits are coming through.
If you look at the case studies that we include in our results every six months, you can see the trends historically through the first generation sites. We'll experience very similar trends in our second generation facilities and third generation facilities.
Is there any plan when the company might actually be paying dividends?
That's ultimately a matter for the board. At this stage, there are no specific plans to start paying dividends.
Okay, thank you.
Thank you. We have a further question from Tim Plumbe from UBS Investment Bank. Go ahead. Thank you.
Hi, guys. Sorry, just one follow-up from me. Craig, Cross Connects at 7.7% of revenues at the moment. Can you give us a sense of how you're thinking about that a little bit longer term, given the mix of the customer base you've got and maybe if you can split it out kind of against your old generation data centers compared to your new generation data centers? It's obviously going to have a different customer or megawatt mix there.
Thanks, Tim. Obviously that growth in Cross Connects, if you go back a few years, when we started obviously, very low percentage. It has continued to grow. I expect it will continue to grow. Can it be 10 or maybe even 15% over time? Potentially, yes. We are very focused on that retail colocation strategy. We are also focused strategically on building the most diverse ecosystem of public and private cloud providers and on-ramps. Also the content delivery networks. As you attract more network service providers to the ecosystem, and they can be your traditional telcos like Telstra and Optus and Vocus are all good partners and continue to do great work with our teams today. They can be the new digital age born in the cloud network service providers like Megaport and PacketFabric and others. We continue to see that growth.
We look at all of those providers as a great opportunity to continue to diversify and grow the breadth of our services ecosystem. The more of those we have in our data centers, the more choice there are for our enterprise customers on network connectivity services. That includes everything from just buying your cloud Cross Connect through to who provides your metro fiber, your interconnects services, or right out to your submarine cable capacity. Over time, I can not give you an exact number, Tim, but I hope that it is certainly in the teens. It continues to grow, and it will reflect our focus on being an enterprise colocation provider that builds a lot of value in public and private cloud on-ramp services.
Great. Thanks, Craig.
Thank you. Once again, if you do wish to ask a question, please press the star key then one on your telephone and wait for your name to be announced. Thank you. There are no further questions at this time. I will hand back to Mr. Scroggie for any closing remarks.
Thanks, Jody. Ladies and gents, thanks for joining the call today. Obviously, like to thank all of our investors for their continued support, but it would be remiss of me not to sincerely thank all of our team members at NEXTDC who are working very hard to build this amazing, extraordinary platform. Very proud of what the company has achieved and the efforts that all of them have continued to make. Not only was it our largest sales half ever in the company's history, but obviously our largest operational half, biggest number of cloud connects, largest number of customers, serving more than 1,000 enterprises today. Very pleased with where the business is as we continue to grow and scale and see enormous opportunity in front of us to continue to take advantage of. Thanks to our team.
Thanks to all of you for your continued support and interest in the company. Bye for now.
Thank you. That does conclude our conference for today. Thank you all for participating. You may now disconnect.