NEXTDC Limited (NXT.AX) Earnings Call Transcript & Summary

Thank you for standing by, and welcome to the NEXTDC FY '20 Results Announcement. [Operator Instructions] I'd now like to hand the conference over to Mr. Craig Scroggie, CEO. Please go ahead.

Thank you. And good morning, ladies and gentlemen, and welcome to the NEXTDC results presentation for FY '20. I'm joined today by our CFO, Oskar Tomaszewski. We're very pleased to present another set of record results, beginning on Slide 2, total revenue of over $205 million and underlying EBITDA of $104.6 million. A record new sales year saw utilization grow strongly to finish FY '20 with under 70 megawatts under contract. Our network continued to expand, adding more than 2,000 interconnections to finish the year with over 13, was up 19%. And our ecosystem continues to evolve with over 1,300 customers and more than 640 partners, including 70 connectivity service providers. Turning to Slide 3. Our strong performance in FY '20 is highlighted by robust key operating metrics. And revenue from data center services increased by $31 million to $201 million. Contracted utilization increased by a net 17.4 megawatts or 33% to 70 megawatts. And interconnection revenue has grown to 8.1% of recurring revenue, up from 7.7% last year. Our results continue to demonstrate the company's operating leverage with underlying EBITDA increasing 23% to $104.6 million. Operating cash flow grew 37% to $54 million. And billing utilization stood at 75% of contracted capacity at the 30th of June, indicating significant revenue to flow into FY '21 and beyond as contracted capacity is brought online. We remain well capitalized to support our growth plans, with total liquidity at 30 June of approximately $1.2 billion, inclusive of our $300 million senior debt facility, which remains undrawn. Our balance sheet position has never been stronger, now underpinned by $2.7 billion of total assets. And at the 30th of June, we held property with a carrying value of over $850 million as well as plant and equipment with a carrying value of over $700 million. Our data center fleet continues to evolve at a rapid pace. Our CapEx came in at $418 million versus guidance of $340 million to $380 million due to building completion coming in slightly earlier than we projected as well as the settlement of M3 Melbourne land upon satisfactory completion of due diligence in the last week of FY '20. Our P2 facility construction was completed, and that facility is now open to customers with 2 megawatts of installed capacity. M2 capacity expansion continues with 15 megawatts of new capacity currently being fitted out. Importantly, M2's total target capacity has increased by 50% to 60 megawatts. S2 development continued with 4 new data halls being opened, taking total installed capacity to 22 megawatts. And S3 earthworks are in progress with practical completion of Phase I expected in the second half of FY '22. I'll now hand over to Oskar to discuss the financial results in greater detail.

Thank you, Craig. Let's now turn to Slide 6, a summary of our profit and loss for the year. The statutory results reflect data center services revenue of $200.8 million, an increase of 18% on last year. Net loss before tax of $18.7 million, and it's worth noting that the net loss after tax includes the recognition of $33.5 million of carryforward tax losses. Despite this de-recognition, these tax losses can be carried forward indefinitely and have no expiry date. Our nonstatutory highlights include: Underlying EBITDA of $104.6 million, an increase of 23% on last year; direct costs of $38.1 million, which rose in line with contracted customer capacity, offset by improved power efficiency and energy costs; facility costs, which increased to $21.9 million as we've ramped up operations across our facilities, particularly S2, P2 and M2 as well as corporate costs, which increased modestly to $35.8 million from $34 million a year ago, reflecting the operating leverage of our business. On to Slide 7. Revenue generated from racks, suites, cross-connects and other recurring sources accounted for 96% of total data center services revenue, with project revenues representing a lower proportion of revenue each year. The underlying EBITDA performance highlights NEXTDC's operating leverage as demonstrated by the 23% growth in earnings relative to 18% growth in data center services revenue. This trend is further demonstrated by the longer-term revenue and earnings CAGR figures on this page. Slide 8 sets out our revenue per unit metrics. Annualized revenue per square meter continued to grow during FY '20, benefiting from contracted price escalation and increased connectivity, power density and power recharge revenues. Annualized revenue per megawatt reflects some new, larger customer deployments coming online during the year. And it's 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 ancillary services over time. Slide 9 summarizes our balance sheet position and cash flows. At 30 June, NEXTDC held property with carrying value of $854 million as well as plant and equipment with a carrying value of $704 million. Our net assets at the end of the year stood at $1.7 billion. And finally, we remain well capitalized to continue our growth trajectory with total liquidity comprising cash and undrawn debt facilities of close to $1.2 billion. I'll now hand you back across to Craig to go through our business performance and outlook for the 2021 financial year.

Thanks, Oskar. On Slide 11, our key nonfinancial metrics are set out. Total customers are up 15% year-on-year to 1,364. Total interconnections up 19% year-on-year to 13,051. And total cross-connects per customer, up 3% over the same period to 9.6 per customer. On Slide 12, further insight into the diversity of our business. Customers by industry shows strong growth in enterprise, supported by cloud system integrators and connectivity partners, and the skew towards high-density deployments reflects the growth of the hyper-converged infrastructure and hybrid clouds. Slide 13. At the 30th of June, 89% of installed capacity was under contract and 75% of contracted capacity was billing and generating revenue. FY '20 saw a 40% increase in billing capacity, a record of over 20 megawatts, with 17 megawatts of contracted capacity at year-end still to commence billing. There is scale for continued improvement in operating leverage and for strong revenue and earnings growth. On Slide 14, our capacity and utilization. In Sydney, we've already sold over 20 megawatts capacity S2 and have installed 22 megawatts to date. The ongoing strength of demand in this market has provided us with the confidence to commit -- to fitting out the remainder of S2 and commencing building works for S3. We've also committed to fitting out the first 12 megawatts of that facility. In Perth, the initial construction of P2 is now complete with Phase 1 open to customers. And in Melbourne, M2 capacity expansion is progressing with 15 megawatts of new capacity being fitted out. The total planned capacity has been increased 50% to 60 megawatts in response to strong customer demand. Further, our land parcel for M3 has been purchased and early works relating to design, and approvals are now underway. We also continue to prepare for our Tier IV certifications of constructed facility and the Gold Certification for the (sic) [ of ] Operational Sustainability at S2 and P2. On Slide 16, sets out our outlook. Data center services revenue guidance of $242 million to $250 million is underpinned by strong growth in recurring revenue and our long-term customer contracts. Over 17 megawatts of contracted sales are yet to commence billing at the 30th of June and will do so progressively during FY '21 and beyond. We expect underlying EBITDA of $125 million to $130 million, with scale and earnings growth to continue to be driven by our generation 2 facilities. Total CapEx for the year is expected to be between $380 million and $400 million. FY '21 is going to be a year of continuing acceleration for growth at NEXTDC. The foundations we have put in place over many years will see the company continue to scale rapidly through FY '21 and beyond. Now Jesse, if we could please open the line for questions now.

[Operator Instructions] The first question comes from Jonathan Atkin with RBC Capital Markets.

So a couple of questions. I was interested in the FY '21 guidance and the revenue midpoint that you're showing. How much of that is based on contracted but not yet commenced utilization versus, say, new bookings that might happen between now and year-end '21? And then I had a second question regarding just the return profile. Is there any kind of an update you could provide on the current returns you're seeing on M2 and S2?

Thanks, Jon. Appreciate that. And good evening to you in San Francisco. Thanks for joining. The first question in relation to how much of the FY '21 current revenue and EBITDA ultimately is currently contracted. The majority of our forecast for FY '21 is already under contract. As you know, Jon, in the course of the year, really, as we sign new customers and bring them online, it can take 3 to 6 months in the enterprise space and 6 to 9 months in the larger hyperscale space in order for contracts that we signed to actually be delivered and activate its revenue. So the majority of what we're seeing certainly in the forecast today is already under contract. Now the second question, just in relation to returns, pricing overall, we continue to be very pleased with meeting our return on capital expectations. It's quite an interesting environment at the moment, obviously, with people both looking to accelerate the move to the enterprise data center out of on-premise. So we've continued to see in the new COVID environment further demand for that move out of office environments, and we're also seeing, obviously, material growth in demand for the network and cloud connectivity elements as well. So our overall return expectations continue to be in line with what we've delivered previously. And we see good, strong performance at the pricing level overall, particularly given quite an interesting environment that we're all living in.

The next question comes from Kane Hannan with Goldman Sachs.

Just 3 quick ones for me, please. Firstly, just the contracted megawatt. Just confirming that 70 megawatt of the 4 megawatts you guys announced on July 1. And then if that's the case, can you just give us a sense of, I suppose, split of that 4.6 megawatt incremental for the 12 megawatts [indiscernible] across hyperscale and retail? Secondly, just interested if you could comment on how sales have been trending in July and August and whether there's been much of a slowdown as the economy emerge a little bit from COVID. And finally, just the M2 comment and that big increase in capacity. Can you explain a little bit more about what's happened there and how we should think about the incremental CapEx to hit that 60-megawatt target?

Thanks, Kane. I'll break that down into 3, I think, questions in there; so the retail, wholesale or hyperscale mix. As we normally do, the retail business generally tracks at around about 2,000 kilowatts a year. So we're seeing similar numbers. Obviously, with 2 key forces driving that one is the move-out of on-premise data centers. We still see about 70% of enterprise customers operating on-prem infrastructure. So that's driving the first transition, which is the move to enterprise colocation. The second component being the move to cloud, both public and private for commoditized infrastructure. So that's a similar performance year-on-year than what we've experienced previously. Network-related demand is trending strongly, and that won't be a surprise to anyone given the underlying redesign of networks, particularly in a decentralized sense as we all went to work and school from home. The second component in question in relation to -- into capacity change. The key driver there is that we have spent time redesigning the overall facility in order to be able to improve the total megawatts delivered that required us to do engineering redesign works to both increase the total amount of power available and increase the building size. We obviously sold more capacity at M2, both orders and options than we originally designed the site for. So key driver of the upgrade by 50% to 60 megawatts was the redesign of the site to deliver all of those enormous options that sit on the back of those announced contracts at M2. So very pleased that we've been able to get M2 to deliver a design capacity in excess of 60 megawatts. In the context of modeling, you can use similar CapEx assumptions to those for all of the underlying dollars per megawatt. Today, we don't expect any change in the forecast for dollars per megawatt to get from 40 to 60 megawatts, just the overall site redesign from an engineering perspective. And the last item, Kane, was what we've seen in July and August. And obviously, in the environment, particularly given Melbourne being in stage 4 lockdowns, we continue to build and construct on-site in order to be delivering, obviously, all of that material capacity that we have contracted. So no change in the wholesale context of delivering all of those hyperscale contracts. And certainly, we're changing the way we're selling. So demand in July and August is, in an enterprise context, has been similar. We've seen a few hundred kilowatts transact as we would ordinarily at this point in time. We've won some very important network-related points of presence for major global expansion capacity that will go to drive further demand in the Melbourne market. So I think given the environment, particularly the challenges in many enterprises, the business continues to actually perform well. And we've seen a record pipeline coming out of FY '20 and into FY '21. Thanks, Kane.

Sorry, could you just confirm that 4-megawatt New South Wales contract? Is that in the 70 megawatts on June 30, just given how close it was to sort of the start of FY '21?

Yes.

The next question comes from Paul Mason with E&P.

Just a couple from me. So the first one is just on M2. Obviously, you absolutely flew this half. But you've got 2 hyperscalers in there, and there's more than 2 that are in the market in Melbourne. And so just wondered if you can make any comments on views around making further room again. Because obviously the 60 fits the 2 contracts with options, but if you want another big player there, you might need to expand further. The second one, just on the M3 land purchase. I'm sure, given you haven't made comments, you won't be able to go into too much detail, but it sort of looks like potentially that's either an amazing price for a Port Melbourne block or you might be looking elsewhere in the city. I don't -- if you can make comments on that or not. And then the third one is just on Asia. There's obviously like some costs for business development in Asia that have been in the accounts. I think in February, you guys were saying that, that had sort of been put on hold. But maybe if you can sort of comment on whether there's any ongoing activity there or whether that's actually still on hold.

Thanks very much, Paul. So again, break those down. The first one in relation to hyperscale contracts. So you said that we had 2. We have 3, obviously, major global deployments of material size in Melbourne across M1 and M2. We expect, obviously, that with the M3 development that, that will expand out even further. So work with those customers to expand and essentially activate further capacity into M3. We'll provide more information, as we always do, once we've secured the site. We try and keep the information related to the overall development location and other things specific to us and the planners, the appropriate government officials and others, to avoid people getting in the way of the work that's going on. So once we've got the development approvals moved through shortly, we've made very good progress there both in design and planning and support from the government is very strong. Obviously, the backdrop of a challenging environment in Victoria and just overall economic development, investment, job creation is critically important at this time. So very good support from the government despite the circumstances down there. And I would expect that it won't be too long before we're in a position to disclose the location, the overall design itself and talk a little bit more about some of the customer commitments that we'll have there at M3. So more on that, but just to reinforce, obviously, 3 major existing deployments. There are more than that, that we have in Melbourne today with others that I expect will grow over time as that new region starts to develop. It's quite exciting opportunity for us to see the Melbourne region overall start to resemble something similar to what Sydney is today. So it has huge potential to come through in the next 6 to 12 months. And the last question, Paul, was in relation to Asia. We have an office in Singapore and one in Japan. We've continued to look very carefully at those markets. We still see, obviously, very strong demand from customers who'd like to see us take our business into those geographies. Obviously, in the backdrop of the COVID environment, it's more difficult to travel and move around and other things. So look, I don't anticipate that, that -- or this new world that we're living in will stop us from advancing the work that we're doing. But I think it's fair to say it will certainly slow down travel and other things. But we do intend to continue to look at where it makes sense to expand in the region. Overall, most of our larger customers would like to see us operating in a number of those key markets, and we continue to look very closely for the right entry opportunities. It's not our goal to want to buy someone else's legacy data center business. We want to develop a world class-leading quality business like it's taken us a decade to do in Australia. And I would anticipate that, over the course of the next year or so, we will begin that 10-year journey of building something across the region that's as successful as what we've spent time building here in Australia. So a little bit more on that in the not-too-distant future. But Paul, our plans haven't changed. We still desire to want to see the company grow and take advantage of the platform that we've built. And we think that we can do that across the region. Thank you, Paul.

The next question comes from Tim Plumbe with UBS.

Just 2 questions from me. Firstly, you guys added a fair bit of incremental capacity in the second half and a fair bit of that coming in at S2, 22 megawatts now. Can you give us a feel in terms of how many of those megawatts have been fitted out and are ready for delivery to the customer, please?

Sure. Thanks, Tim. So obviously, we are in a position now where we'll have -- for the first time for quite some time, actually have some inventory to sell. So whilst we don't have a lot left between options and other things that we're working through with customers, I would anticipate that the balance of S2's capacity will soon be gone. That certainly puts more pressure on us to deliver S3. Our current forecast for delivering S3 is going to be March 22. And at this point in time, I'd expect that we're certainly trying to avoid going dark in a market. We have a relatively small amount of inventory compared to what we see in the context of demand. We want to continue to focus on building a great ecosystem, supporting our customers and importantly, supporting our partners in the context of that diverse enterprise and government business. And I think we've got enough inventory, hopefully, to get close to seeing us through to the opening of S3. But based on current pipeline and other things, it is highly likely that we'll sell-through the balance of what we have in S2 in relatively short order. But we are doing everything that we can to bring S3 online as quickly as we can. And you can see from the pack, the early works that is already underway there. Shortly, we'll announce the appointment of the main contractor for the development of S3, and we'll be working to bring that online in a relatively short time frame.

The next question comes from Nick Harris with Morgans.

Great results. Just 2 questions from me. First one, just historically, it's taken maybe 2 to 3 years for the hyperscale deals to ramp up to full billing. Obviously, the world's changed in the last 6 months. Just wondering, should we still be thinking in 2 to 3 years? Or could that ramp-up profile get shortened to 2 years just given the huge amount of demand? And then the second question is just on options. That's a massive amount of options you have provided in the last 6 months. It's almost a decade of sales. So it's really, really significant. So I just wondered if maybe you could give us, from your perspective, Craig, what you think needs to happen for them to get exercised. I know it's the client's call, but I'm sure you've got some sort of feel for what would need to happen for them to want to trigger or exercise those options?

Thanks very much, Nick. In relation to the ramp. So certainly, in the past, the time for customers to move in, if you just think about the volume of infrastructure, the thousands of racks and tens of thousands of servers and other infrastructure that goes into deploy on hyperscale, we would certainly anticipate there's still a very physical element to that fit out. So the time to development, the building itself, the base building from site selection through to construction, continues to be in the 12- to 18-month period. And then the time for customers to deploy the capacity with this contractor tends to be within 12 to 18 months. Now we have certainly seen that we have an acceleration of customer demand. So we have customers asking us what we could possibly do to bring capacity forward. And that is certainly a positive trend. It's not a surprise to anyone, obviously, that the volume of demand for all of these platforms is accelerating. But I sort of tend to see that, Nick, as a shift that really represents the transition to a new working environment. So I don't feel it's a short-term thing. I just see this as a really pull forward of capacity that would have come over time. And the 2 to 3 years that it would have taken to ramp is probably 18 to 24 months. And that then leads into the second question in relation to the options and the scale of the options and what needs to happen in order for those to be activated. Particularly in a new region like Melbourne, I think it is as simple as that -- those first tranches of new regions being consumed by enterprise customers and the capacity and demand forecast teams within our key clients we work very closely with. Obviously, you can't activate enormous amounts of capacity on very short notice. So we need to have regular and detailed planning work with our customers to be in a position to bring that forward. So I would say that we are going to see an acceleration. We are certainly already planning for bringing forward additional capacity both in Sydney and in Melbourne. And I do think the backdrop of the environment that we've been living in more recently will simply just accelerate the adoption of compute and drive acceleration into more network depth and breadth generally. So I can't give an exact time frame, Nick, but I think it's fair to say that we will see an acceleration across hyperscale consumption both in the Sydney and Melbourne market.

The next question comes from Roger Samuel with Jefferies.

Very strong result. I've got 2 questions. Can you tell us about the medium-term outlook beyond FY '21? Yes, so it sounds like there is some pull forward of demand because of remote working. And I'm just wondering if you're concerned that growth in contracted capacity will slow down after FY '21 when life goes back to normal. Or do you think that there is a next wave of demand coming? Second question is on the plant capacity. So you've obviously increased the plant capacity at M2 by 20 megawatts. And I wonder if there's any upside to S2, which is currently planned for 30 megawatts, given that -- yes, 30 megawatts, it's half the size of M2. Yes. So I'm just wondering if you can increase the capacity at S2 as well.

Thanks, Roger. I appreciate that. So in relation to the medium-term forecast, I would say that, again, not dissimilar to the previous couple of questions, it's an acceleration. There's no question that there is a pull forward of demand in public and private platforms. But I don't, sort of, look at that and get concerned about growth. I see the growth environment as a positive tailwind for us. And the next wave or our ability to be able to continue to meet those demands, clearly, we are 10% to 12% of the market domestically in Australia. There are a number of others that will also be in a position, I think, to benefit from the growth that we're seeing. Customers themselves, obviously, will continue to work to try and meet some of the capacity requirements as well. But the growth issue is a quality problem. The next wave is continuing, I think, to come. We've got better visibility to forward requirements that we've ever had any other time in history. The fact that we've got contracted options of reasonable scale just sort of go to demonstrate the maturity of the planning work both the customers are doing on their side and we're doing in terms of capacity planning. So giving a really high degree of confidence that not only are we planning, to a greater extent, to the next wave than we ever have before, but we've got much better visibility to know 2 or 3 years out what we need to be planning for in size and scale. And the answer to your second question, Roger, is that, yes, we obviously were able to redesign M2 because the whole building wasn't fitted out. It's a large block. Whereas in the case of S2, if you have a look on Page 17 of the results presentation, at the first page of the appendices, there is a photo of the fully completed S2 building from the exterior. And what you'll see is that the building now is fully fitted out. So in the case of S2, 30 megawatts, it's an enormous building, it's a huge facility. It's extremely high density, given the number of hyperscale customers that we're serving there. And clearly, the fact that we've already started building S3 means that we will have pressure on us again to deliver S3 in record time, getting S3 underway, while we really just -- having completed the delivery of S2 gives me a high degree of confidence that we'll be able to continue to support our customers' growth requirements. But yes, at 30-megawatt, S2 is fully fitted out and will be fully sold. And again, will represent very high levels of utilization given the nature of the customers that we're supporting there.

The next question comes from Mitch Sonogan with Macquarie.

Just the first one, on the outlook for '21, just a little bit more detail on the drivers. Just wondering what you've assumed in the guidance for upfront project fees versus recurring. Is there anything else that could be impacting that guidance, e.g., electricity costs? And can you also just tell us what the expected costs are in '21 for the Asian market review that will be excluded from adjusted EBITDA, those $1.2 million in '20?

Yes. Thanks, Mitchell. I'll answer the latter and then come back to the former. But in relation to Asia cost, the team is not changing current size that are working on site selection reviews and engagement with customers on building design and other things is unchanged. So it will look pretty much exactly like it does at the moment. That cost to continue to review locations for greenfield development in each of the key markets will be the same pretty much year-on-year. The earlier question, I think, that whilst I don't anticipate a lot of change in the mix year-on-year, obviously, the bulk is recurring revenue. But just to comment on power, specifically, obviously, we're seeing power prices change and move down, which is a great positive impact. So whilst that may reduce revenue marginally, it will improve overall returns. So as you know, a very high percentage of our power is passed through. So the fact that power prices may come down a little, improve our margins, is a good thing right across the board. And at the moment, we've built in estimates for what those reductions in power prices would look like. But if you want a slightly more -- a detailed level of conversation on power, mix and forecast prices and where we think that will go, you're more than welcome to spend a little time chatting with Greg on our forward power contracting and where we see power prices going overall.

Your next question comes from Siraj Ahmed with Citi.

Just 2 questions, Craig. Just first thing on the pipeline. You've mentioned record pipeline in your accounts in your other call as well. So just, I mean, FY '20 is a big year, especially -- I mean, calendar '20 has been a really big year. Just can you tell us on how you think about new contracts this year? Is it more about delivery this year? Or do you think there could be strong bookings again from a whole [indiscernible].

Thanks, Siraj. Yes, yes. Thank you for the question. Look, I think it's probably going to be both. It's going to be our -- look, FY '20 was our biggest sales year ever. Team did a great job. Proud of the contribution, not just in the sales and marketing context, but everybody in the organization. I mean, that pressure flows right through the team to deliver. Obviously, with S2, it was, without question, our largest and most complicated build. It is a stunningly beautiful facility of extraordinary quality. Customers are absolutely thrilled with what was delivered. So there's no question that FY '20 was a big year in terms of delivery, a big year for the company maturing materially in systems and processes. You've got tens of thousands of active services, thousands of active cross-connect and network and cloud on-ramp support services, a large delivery and support team there day in and day out running a national business. So FY '20 was big, FY '21 will be even bigger because of the record sales here in '20. And I would expect that this, again, could be a record sales year. We certainly have enough options contracted. Just in options alone, just the conversion of options alone, and obviously the planning work that we're doing with customers with multiples of megawatts, both at S3 and at M3. So I'm reasonably confident that, given the backdrop, the demand we're seeing, migration for enterprise and government colocation and then the combination of the network density driving more cloud, both public and private, and a combination of all of those factors could see FY '21, again, our biggest sales year ever. So a big delivery year coming up, enormous construction projects with M2, M3 and S3, not to mention the great work that's been done to deliver additional capacity in all the other markets. So thanks, mate, for the question. And yes, both in the delivery context and in a new sales context and in closing out options, there's no question that FY '21 will probably be like nothing we've ever seen before.

The next question comes from James Bales with Morgan Stanley.

I've got a couple of questions. Firstly, I wanted to touch on the liquidity and CapEx commitments. And you guys seem to have a really strong demand environment and seem somewhat constrained by your ability to deploy capital fast enough. And when you compare the CapEx versus the liquidity on the balance sheet, there's not a lot of runway there. How should we think about the mix of capital incrementally from here?

Thanks, James. Look, I don't think we're doing a bad job. I don't think we're constrained in terms of deploying it. We've got $400 million out the door in FY '20 and into the ground. There is a physical limitation on obviously the volume that you can deploy. But as it relates to the forecast for FY '21, we're reasonably confident between the big projects and the big projects being M2, M3 in planning and then obviously S3. But we still have plenty of headroom. So I guess 2 things. You've got the best part of over $1 billion worth of cash yet still to be deployed. The $300 million senior is undrawn still, so you've got more than $800 million in cash and $300 million in senior. But on top of that, and this is something that we'll talk a little bit more about soon, Oskar and the team, our advisers, have been doing a material amount of work in preparing for the next major step-up in our debt packages. So there is a very large piece of debt work coming through that will give us an enormous amount of additional firepower in the future. So I won't preannounce what that is in detail at this point in time. There's a number of banks globally engaged with the team right now. Very well advanced through the process, and that's looking really good. And I think, again, in the not-too-distant future, we'll be in a position to announce what that looks like. But I'm very excited about what additional flexibility that will give us to fund the business's growth over the next few years.

Your next question comes from Entcho Raykovski with Crédit Suisse.

Craig, Oskar, a couple for me. Firstly, I was wondering if you can provide us with some color on the difference in pricing you might be seeing for the Melbourne versus Sydney markets, particularly for the contracted commitments you've announced over the past 6 months. And just secondly, a follow-up to the funding question. You've spoken in the past about looking at partnerships with sovereign wealth funds to develop parts -- data centers. Just interested in whether that's something you're still exploring, particularly now that you've raised capital.

Yes. Thanks, mate. I appreciate the question. So the first one on pricing is relatively straightforward. So again, if you look at the historical pricing on what we've signed previously, so pretty happy with the pricing mix. And certainly, the revenue per unit metrics that we've shared on Slide 8, that $10,714 per square meter continues to trend up. And then the average revenue per megawatt coming out of FY '20 at $4.4 million per meg. So overall, we're very pleased with pricing and returns. Again, we are not the largest, cheapest provider of data center services in the country. We don't aspire to be the largest lease returning company. It's not our role in life. We see our focus on building world-class quality Tier 4 data centers and building a diversified, highly diversified business that serves enterprise and government and hyperscale, focuses on network-sensitive locations. So again, our focus is a little bit different to many of those companies that you would say are very price sensitive. The second comment in relation to pricing, really, between the markets, that is, Melbourne-Sydney, is it doesn't really matter so much if you look at the markets on a global scale. The majority of hyperscale pricing across most competitors is relatively similar. Pricing is published. I know Jon Atkin is on the call, publishes global pricing for all of the cloud platform providers and gives great insight into the pricing metrics in each market. And what we tend to find is that the prices in the U.S. are reasonably similar right across Asia. So that's something that we continue to watch. But again, no particular surprises or big differences between what we're seeing in Sydney and Melbourne. And then the second question in relation to funding and partnerships. Look, the big piece of work on -- for the team right now is working to close out the next big step-up in our debt funding. So as I mentioned earlier, it's a piece of work we'll share fairly soon. But team are really well advanced in doing some great work there and appreciate all support from banks and others that are working closely with us to support the company's growth plans over the next few years. They are quite amazing numbers when you start to step back and look at where the business is today and where it's going. So some great stuff coming there. And the second element in relation to considering partnerships. Yes, look, some of the largest sovereign wealth funds in the world would love to be obviously continuing to get access to the data center market and key players. Our business is in a great position because we have the luxury of choice with those partners. So if I go back as little as 5 years ago, it was quite difficult to find partners that understood the data center business. But given the volume of sovereign wealth funds and others that really desperately want to be involved with data center operators universally, we've got a great universe of people to work with. And we continue to look at what that -- what makes sense in that area, both in terms of the legacy or the existing install base of assets, but also for future development. So as we start to consider some of the opportunities for building new developments on a large-scale in key markets, Singapore, Japan, Philippines, Thailand, there's a number of different areas that are all growing today across Asia. And partnerships certainly may make sense in some of those. So we're looking and still talking with people about what that may look like. Nothing to announce at this point in time, but certainly engaged in a number of conversations at quite a detailed level.

The next question comes from Sameer Chopra with Bank of America.

Just had 2 questions, please. One, just on CapEx trends beyond '21. Should we start thinking about a tail off in CapEx? I'm just thinking, does CapEx start to over-index on fit out rather than build? And therefore, we should start to see that CapEx starts to moderate? That's on question 1. And then question 2, you raised capital now, there's potential there for refinancing from the banks. When should we start thinking about dividend? Is that even in consideration right now in your discussions with the Board?

Thanks, Sameer. So the first question in relation to do we see CapEx trailing off at some point. Look, I think that is really very simple. I guess the growth forecast is what drives the businesses' need to continue to fund growth. We only spend money when we have -- largely, we're a customer-led business. We're an investment-led business off the back of customer commitments. And we're extremely fortunate, obviously, to run a highly diversified business. We've got a network business. We've got enterprise and government. We've got hyperscale. Hyperscale clearly drives the demand for large volumes of capital, but they're offset by 5-, 10-, 15-, 20-year commitments from some of the largest trillion-dollar companies in the world that are all AAA quality credit. And that's why we don't have any issue supporting that with extremely large -- larger volumes of debt support. So what we're doing with the teams at the moment to support that future growth really does reflect the maturity of our customer commitments, and it reflects the diversity and margin performance in our business. So I think the answer to your question, Sameer, is that do I see CapEx trailing off beyond '21? I don't simply because I see more pipeline and more customer contracts in FY '20 to deliver in '21. And I've seen more opportunity for conversion to growth through options than we've ever had at any other point in time in our history. So the fact that we'll need to continue to work closely with our new debt providers, our existing and new ones, over time to support that material growth is a positive trend. And I guess that will also be a by-product of the second question. Look, the Board talks about and is -- continues to be very focused on return on invested capital and return on equity and performance for our shareholders. I mean, at the end of the day, we're a customer-focused business, but we are here to serve our shareholders and build a quality business over time. As I said before, I don't aspire to be the largest least returning operator. If we will, we can all go and do something else if that's the case. I think dividends really come down to reviewing something that's discussed regularly by the Board. But at the end of the day, we come back to that one important point, and that is if we can continue to generate great returns, we can build a quality business that is a long term, multi-generational asset that will be capable of producing extraordinary dividends in the future. At some point in time, if we're not happy with returns, we can always stop investing and start returning that capital to shareholders. But the decision point that we can continue to come to at this point in time is that the Board are very happy in what our return on equity will look like for shareholders that causes them to want to continue to take advantage of the opportunity and the market while it is strong. And at this point in time, whilst we discuss dividends and returns regularly, there is not a plan per se for stopping investing in the growth of the company. And I would feel reasonably confident to say when you are in a position to be able to continue to grow business at 20% to 30% compound a year, maybe for the next decade, our greatest days are in front of us, not behind us. So a lot of growth still yet to come.

The next question comes from Bob Chen with JPMorgan.

Just a couple of questions from me. Just firstly, very strong year in terms of contract wins for the team. Can you talk a little bit about the contract process? How long you might have been in discussions with these contracts? And was there a sort of competitive process for these contracts? And then the second question is really just around your international expansion ambitions. Could you give a little bit of color on your, sort of, broad strategy there? Is it going to be more of the same? And what sort of size deployment would you be looking for?

Thanks, Bob. First question in relation to contracts for process competitors. I referenced this a little earlier. I mean, obviously, given the size and scale of developments that we're working on now, when you're planning an 80- to 100-megawatt data center, it's something that, obviously, customers are deeply involved in. And I think as these larger clients are maturing materially, their process, their systems and processes and planning are maturing. But the rate of growth in the industry, they're all struggling to keep up with serving their customers. We continue to see maturity in clients' businesses. But the rate of growth that they are all trying to support means that when you're looking 2 or 3 years out, just keeping up with today is a challenge, let alone knowing that you've adequately planned for the next 2 or 3 years. And again, I classify those as really great quality problems that we are incredibly fortunate to have. But we are working more deeply with our larger clients than we ever have at any other point in time before on forward planning. Because you can't just wake up one day and make a call and say, "Hey, listen, I need 10 megawatts, and I didn't tell you about it before. Or I need 20." Or in some cases, we're getting to the point where customers are planning for multiples of 20 in multiple locations. So that means that our level of forward planning has to be in a place that has never been at any other point in time before. The future land acquisitions and other things that we spend time working on today are far further out than any other point in time in the company's history. I spend a lot of my time really just thinking and working with customers on what the next 5 to 10 years' worth of capacity planning looks like. And we would never have been working on that time frame before. They do still take time, and they are still competitive processes. Obviously, when you're in a business that -- in an industry that is growing at the rate we are, everybody wants to be taking a piece. And it's only because when we reflect on this 5 or almost 10 years ago, when we were in an industry that the jury was still out on, people looked at it and said, well, that could go any particular way. Maybe the cloud will take off, maybe it won't. There were still plenty of people saying that cloud was a fad. You fast forward a decade, my favorite quote from Bill Gates, "People almost always overestimate what can be done in a year and underestimate what can be achieved in a decade." And I think we'll see a similar acceleration again over the course of the next 5 to 10 years, where the numbers, in a exponential sense, will just be difficult for people in a linear mindset to appreciate. It's extremely difficult to look at the exponential rate of growth over the last -- even just the last 5 years and try and think what that means in the next 10. But we are in a position now with the 3s, S3, M3; and the 4s, we're already planning for what S4 and M4 will look like and the size and scale of those. So yes, I'm confident that we're doing and have a better handle on what planning looks like in the next 5 to 10 years than we have at any other point in time. And Bob, the last question is probably similar to the response I gave on Asia earlier. What do we see the model looking like, I think, is the important question. And that is, is that, again, we want to build a business that's unique to the NEXTDC quality product and brand. That means that we focus on enterprise and government. We build a highly diverse network businesses with our physical and elastic cross-connect products, and we combine those with the best of hyperscale platform. And we see the combination of all of those as something that continues to see customers want us to expand into the region. So we will certainly want to build a business, not buy a legacy one. We'll want to build a quality business like we have over the last decade in Asia. It's not a simple thing to do. Each of the markets are unique. The emerging ones are always -- come with more risk than the mature ones. But the mature ones come with more competition. But I'm highly confident that the quality of product that we build and the business, the platform, the people -- the people, process and technology investments that we've made and we continue to make in scaling the company will allow us to build out successfully into those markets. But we will do it in a disciplined way. And I think, again, we'll share a little bit more on that over the course of '21. But we will be disciplined in the actions that we take in expanding the business in Asia.

The last question is a follow-up from Tim Plumbe with UBS.

Sorry. Thought I'd just squeeze one more in there, if possible. Just -- you mentioned really down to how you can deliver in those core markets. What sort of run rate do you think you can do in terms of a, building out additional capacity; but probably more importantly now, b, fitting out that capacity within the S2 and M2 locations? And is there anything that you can do differently to kind of accelerate that over the next 12 to 18 months?

Yes. Thank you, Tim. I would have been disappointed if you hadn't tried to sneak a second question in. So look, that's a really great question. I think I'll start with, I guess, just a reflection on how much the build process have changed and what we've learned through the development of S2 and M2. Without question, the S2 development, building that massive hyperscale building, doing it on a relatively small amount of land and building it up the way it is built today, the engineering team have just done an extraordinary job. I couldn't be prouder of Jeff Van Zetten and Simon Cooper, our Chief Operating Officer and our Chief Engineer, and their teams of so many people that have done just extraordinary work in delivering these new buildings. We went from a generation 1 design where we increased -- look at the performance of those buildings today. You've got $150 million, $160 million worth of capital invested and almost $50 million worth of run rate EBITDA. Those generation 1 buildings are almost returning their capital every 3 years. And their performance continues to improve. The level of energy efficiency and other things that the team has designed into those buildings, to get NABERS 5-star certifications on those, they are the most highly efficient certified data center buildings in the country today. But the team haven't stopped there. Not only have they raised the bar in engineering and design and gone to more modularized construction and delivery. If you look at the way the buildings are built, constructed and delivered today, a lot of the lessons in the first generation were applied to the second gen. The second-generation builds are now informing the third generation. So speed to delivery, construction and assembly of larger volume of components off-site in the factories, and then having them to deliver the site and installed is a key element of speeding up the overall supply chain. I think the key lesson for us is not only do we want to be able to reduce the amount of people on-site in construction and allow us to deliver a better quality product on a shorter time frame. The other element is the search that comes with being able to build and deliver a data center from off-site. And living in a COVID type of world, which I think we all have to adjust to and accept that this is the new normal, the stop-start of the global community, the stop-start of the ability to be able to travel and have people working freely in the community means that it puts a lot more pressure on organizations to do things differently. We had already put ourselves in a very strong position. So in Melbourne today, as an example, there's really very little impact to the overall delivery program as a result of the whole state being shut down. And that program is still going beautifully today. And I think many of those lessons that we've learned, that are now going into generation 2 and will go into our generation 3 construction means that we can probably, Tim, materially improve. And if I said the team might be capable of doubling the amount of capacity that we've delivered in previous years without putting that pressure on them, I know they'd all be -- they've worked incredibly hard to get to the point that we're at today. They've done amazing work, I couldn't be prouder. And it's not just -- again, it's not just the design, engineering and delivery components. It's the operational teams: Nathan McBride that leads our facility management team and his facility leaders around the country do extraordinary work; and Brett Ridley, our Head of Central Operations, and the guys that maintain our data center fleet and keep them operating to world-class standards and allow us to get Uptime Institute gold operational certifications at the Tier IV level. I mean, these are things that are world-class in the industry and just represent the quality of what it is that we want to do in the data center sector. So I do think, Tim, we can materially continue to increase the speed to deliver. We are proving it in the second-generation designs now with what is being delivered in M2, the Super T infrastructure for delivery of the substructure and then the containerized delivery of our long lead time items. So I think as we watch this space, but yes, could have another 20-, 30-megawatt year coming up.

The next question comes from Mitch Sonogan with Macquarie.

Craig, just sneak in one more quick one. Just on the interconnections business. So it was up 19%, so it was a solid result. But actually by number, you had slowed to 2,079 versus 2,333 in FY '19. So I was just thinking about your commentary around a big uptick in demand for interconnections in the fourth quarter and on the call today. Should we expect an acceleration in '21? And just following on, it's at 8.1% of revenue. Do you see that trending towards maybe Equinix levels towards 20% over time?

Thanks, Mitch. Yes, look, I'd love to see it trend towards 20%. I think it's certainly capable. It's a reflection of the byproduct of the maturity of the installed base and the size of your enterprise, clients. And as we see more demand in public and private cloud, I mean, we've got more than 600 partners today, 70 connectivity service providers. I mean a deep ecosystem. It takes a long time to build. And the more we focus on continuing to help our enterprise customers not only transition out of their office, but make the transition into that hybrid multicloud, where they're connected to a multiplicity of service provider and networks, interconnections will naturally continue to grow. So it's a by-product of that, and it's a by-product of the maturity of the facility. So year-on-year numbers, whilst it was up 19%, you've got to remember, despite those amazing numbers, FY '20 was a year where we actually didn't have a lot of inventory to deliver. It was one of those years where we really -- we're hustling hard to play catch-up to get inventory to sell. So being in a position now where, for the first time, actually, for almost a couple of years, we're going to have inventory to sell, we'll see an even greater acceleration in the number of interconnections. So I think it's core to our strategy. It's critically important to serving enterprise and government customers. It's another key decision criteria for why we keep winning the cloud on-ramps. Cloud on-ramps go where the enterprise customers are. That's a really key factor for determining where you want to put a cloud on-ramp. Might put your cloud on-ramp in the place that's most convenient to access enterprise clients. It's not only the enterprise client accessing the cloud, it's the cloud deciding where the enterprise customer is and going there. And for us, fortunately, in almost every location, we have been the choice for that reason. So yes, I think interconnections will continue to grow. I'd love to see them get to that Equinix level that is reflective of an enormously successful, quite mature enterprise business. And Equinix have built one of the best in the world. And we continue to see them as a fantastic builder and operator of data centers that have built something extraordinary over a couple of decades.

Thank you. That does conclude the question-and-answer session. I'll now hand back to Mr. Scroggie for closing remarks.

Thanks. Appreciate everyone taking the time to join the call. It's been an extraordinary year, a challenging year for all of you. I just want to wrap up by really sincerely thanking the team at NEXTDC. It has been obviously a challenging year for everyone, particularly those who have had to work and school from home. I know everybody has had to adjust to that new normal. But we've got an extraordinary business. We continue to do great work, and we couldn't do that without the tireless efforts of the team, and I'm eternally grateful for how passionate, engaged they are and continue to be. It's a great pleasure and my greatest honor to be working in this business and supporting all of those team members today. And lastly, to thank all of you for your support, our shareholders. We wouldn't have the opportunity to do this if it wasn't for the support that we've had for you over the last decade. But I passionately believe that whilst we've built something great. It really genuinely is only the beginning. The network is playing out, supporting public and private cloud to build what will be the platform for our digital future. All COVID-19 has done is accelerate many of the laggards that wouldn't have made change to the way they work and live to this new world order. And we're in incredibly great position to take advantage of that. I'm very proud of the team. I hope you're proud of the work that they've done to deliver another fantastic result. FY '20 was a wonderful year, but FY '21 is going to be something extraordinary. So thank you to everyone in our shareholder base that has continued to support us. And I hope that many of the most exciting years for the company are the ones that are yet to come. So thank you.