CEO Shalini Lagrutta, CFO Laila Green, and COO John Stevens joining. I'll now pass to Shalini to start the presentation.
Thank you, Mel. Good morning, everybody. Mel, can I ask you to just swap to the next slide? Thank you for joining the investor webinar. John Stevens, Laila, and I will run through the deck, and we'll take questions at the very end. During the course of the presentation, feel free to send questions through. Just to sort of recap DXN and who we are, and what we do. If you sort of look at this deck here that you're seeing, the slide here, they're saying we're a vertically integrated manufacturer. We manufacture and operate modular data centers. In terms of core design, we design, engineer, we build, we deploy, we own, and operate as well. It's three major divisions that we talk about when we talk about revenue.
If you look at our FY 2026 financials, 73% of our revenues have been from the modular division, which is essentially a projects-based business covering design, engineering, and deployment of our modular data centers globally. These are basically flexible and scalable alternatives to traditional data center builds. The second part of our business is the data center operations, which is about 23% of our revenue from the last financial year. The final piece is the Data Center as a Service business, and that's us owning and operating a capital-light facility as a service model. We have a lot of deployments around the world, and we continue to grow. We are hitting over 120 sites across the world now, over the last five years. In the last year, it has been the most that we've actually deployed in one year. Next slide.
I thought it would be useful to sort of work through what has changed quite a bit. There's quite a lot of information that we've previously presented to the market around where we sit as an ecosystem. This sort of gives you a view of where we sit in the operator and OEM landscape. We always get questions around where we sit versus, say, a vendor solution, or where we sit versus an operator self-build solution. We tend to be the guys that engineer. Half of the time in a sales process is very much understanding what a customer wants, build around the customer's choice of equipment. The understanding of the design and equipment is typically the customer. DXN owns background IP. The customer owns developed IP. In terms of actually solutions, product and solutions, the choice of the solution is always vendor-agnostic.
It could very well be a choice of different OEMs that you've got in the market, Vertiv, Schneider, Stulz. Depending on what their lead times are, what their pricing is, how that fits in a modular ecosystem, we get to pick and choose together with the partnership that we've got with the client. Our market tends to be mid-sized operators, sometimes very large operators who are deploying mid-sized deployments, neo-cloud operators, telco, defense, enterprise, even global internet companies who are building cable landing stations around the world. We sit in that ecosystem in between the end customer, which is the operator, and the OEM. Just a bit of a sort of reminder of where we sit in that ecosystem. Next one.
Over the past few months, there's been a lot of change and a lot of consolidation and thinking that has gone into the product focus of the business. We had a strategic session a couple of weeks ago with our team, with our CTO leading the product development of our future strategic plans, not just for sort of the next three months, one year, but what it looks like over the next three years, for instance, and what we're seeing in the market. If you kind of look at where we have been successful and where we're adding a lot of value and where, in the ecosystem, we've actually got a product market fit, these are the three market segments. There is a lot of sub-segments in here. There are different type of customers that end up in these three market segments.
But we decided to sort of really consolidate into what this looks like, and this is where we kind of landed. We've got the traditional steady foundation over the last five years that we've been operational as a modular data center company. Cable landing stations, these are global internet companies, customers that we've got. Their growth is quite significant compared to, say, three or four years ago because of AI. A lot more fiber builds around the world, which necessitates cable infrastructure being built and terminated across multiple jurisdictions. Our typical contract sizes in the past used to be AUD 1 million- AUD 2 million because we were doing spurs or repeater solutions as an additive to what we used to do. Now they're getting bigger and bigger because they are a combination of not just the cable landing station and the termination, but also some AI and edge inference requirements.
The deals are now moving into sort of the AUD 1 million- AUD 10 million territory, some a little bit more than that, depending on who the client is. This would typically be the 3 MW- 20 MW for subsea and satellite operators, and we've announced quite a few of these over the last few years, and we will continue to do that in the pipeline. We've got a steady foundation here, and a lot of these opportunities come to us today. The second portion is the prefabricated AI factories, and that's the two contracts that we have signed, and the latest contract being the AUD 12.2 million that we signed in August for 2 MW. These are AI prefabricated modular data centers. These are extremely high growth, and there is a lot of inbound, and there's a lot of conversations with the ecosystem around this, including NVIDIA.
When we talk about prefabricated AI factories, they are typically 1 MW or 2 MW designs. In terms of standardization, we are starting to see a lot more standardization thanks to reference guides from the NVIDIA reference architectures that is out there. We are starting to see a lot of standardization that is coming into the ecosystem, which is fantastic. As the roadmaps of the chipsets that are being released to market consolidates, and there is a lot more information around this, with our ecosystem, we are able to standardize these products as well. These are typically liquid cool modules, so the cooling solution is direct-to-chip cooling. It can be up to 150 kW per rack. In some instances, it is up to 250 kW per rack, depending on if it is Vera Rubin or the GB300 chipsets, in the case of the NVIDIA GPUs.
These can scale from 5 MW- 10 MW clusters, and they are short cycles in large step-ups. We are in that phase of this part of the business where we are starting to do the 1 MW- 2 MW to maybe 5 MW, and that is going to scale in a big way, not with everybody, but with certain parties who are within the NVIDIA ecosystem. That is this segment. The third segment is the hyperscale shortcut hack, and that is DXN's product. We have actually developed it together with our customers, so in terms of what it is that they want, we have come back, gone to the drawing board, and decided these are data center companies that will need prefabrication no matter what. We see that coming, and that is on our roadmap now. We have not actually signed a contract in this space. We have announced this as a product segment before.
We have plenty of conversations with customers who are looking at data halls being completely prefabbed from day one, dollar one, and that those are the new deployments that they have got, like greenfield applications that they have got. This market segment is getting to be more and more important for us. First of all, it will give us a standard product and a recurring revenue base with some of these existing hyperscalers out there who are very well capitalized and desperately need solutions quickly. These will be volume as well, and the way these customers are thinking about how the contracts would land would be reserving capacity in the business.
They would be looking at standardizing a product with DXN, targeting a certain market segment that they would want us to build for, and building that in volume and reserve capacity with us for the future, for X number of years. That is how that market looks like. It is an evolving market. It has evolved, and we have been working on this for two years. We have actually got one contract that we have announced last year with Ventia for a Telstra InfraCo site. That was a much smaller scale of this, but this is the hyperscale version of that product. These are the three sort of bedrock products that we have standardized in our ecosystem. Next slide. Now let us sort of go into what the industry dynamics look like. This is probably just a little bit of backstop before we go into the updates on Q1.
There's a couple of things that's been gaining us over the last couple of months, and we've started to see that play out. The capital markets obviously want execution, minimizing execution risk as much as possible. The conversations we're having with our clients are always about, how quickly can you build, and what's the least path of resistance to get the first token to site, especially in the AI market segment that we've got? It's no different in the hyperscale side as well, but how quickly is it from energy, from power to token? This is the exact NVIDIA terminology that's being used, energy to token. Obviously, there are going to be lots of winners and losers in the industry when it comes to the neocloud play. The NVIDIA ecosystem is evolving to cater for that.
What DXN's been managed to work out over the last three months, and team has worked really hard on this, is to actually be in that ecosystem and understand exactly what that is, what that looks like. So there will be some players who will only co-locate, and some players who would have wanted to self-build. For instance, our first U.S. NeoCloud customer that we won, initially looking at self-building, but they've changed their play to co-location. Cost of capital is a big part of it. Being part of the NVIDIA ecosystem is a big part of it as well. So, co-location versus self-build is a thing with NeoClouds, and you're going to start seeing this play out in the industry. NeoClouds generally sit within an ecosystem of partnership with the chip manufacturers, just like we do.
We're all part of one ecosystem that we are part of. There is design consultants that are pre-approved by, say, NVIDIA, and there are other chips manufacturers out there. NVIDIA's way ahead of everyone, but there are other chip manufacturers out there who are already building this ecosystem of network as well. So it's a bit of a rush at the moment, but there is a lot at play, and that ecosystem is being very well funded by the chipset manufacturers. So, what is now at play is who is the winner is going to be the one that can execute effectively. That is one of the biggest things that we're starting to see, and that's what the expectation is in terms of where the money is going to flow.
This is probably one of the most important key points here that's changed in the last few months that we are observing. So, when we talk about this with our clients, first of all, what do we want to know? We want to know that our clients are in the NVIDIA ecosystem, first of all, or if it's AMD or if it's the different chips, that they are well-funded and properly entrenched in that. That is one of the key lessons and key learnings in this last quarter for us as a business. Nothing has changed in terms of factory delivery, resources, working in a controlled environment rather on site, giving customer the faster deployment from energy to token. All of that still remains the forefront of why every NeoCloud operator who is going to be out there and successful needs to execute on.
They need to actually focus on getting from energy to token as quickly as possible, as safely as possible, in the highest quality as possible. There are not many players out there who can do this at the speed at which the market is expecting them to. That is our opportunity. Number two, power is the gating item. Everywhere we are starting to see, in the last couple of months, that there are delays in power. A lot of the customers who are coming to us who historically, especially in both the two quadrants that I mentioned, the AI and the structural hack product, are looking at delivering, say, much larger sites, 300 MW-500 MW, which could have been a self-build site, are looking at a distributed architecture where power is in regional sites, in locations where power is either stranded or they are bringing power to these sites.
New conversations are popping up around this exact gating factor, and that is an opportunity as well. The third thing is, there is a lot of expectation around ensuring that you bring the local regional councils, wherever we deploy, along on the journey of delivery and deployment. What tends to happen is we are in that sort of conversation where a prefabricated modular data center that is non-discrete, bespoke, and not over a certain size and over a certain build within the compound that is secure, safe, and fenced is ideal for data center delivery in multiple locations in a distributed architecture versus one big site that everyone is worried about with power and water. That is the social license, and that is how that is evolving for us in these conversations. That is the third. The fourth is construction capacity.
By default, everyone is thinking prefab, and Steve coming on board has assisted us to actually look at how we can actually improve construction costs and improve our ability and efficiency within the business to turn around modules quicker within the factory itself. Construction hours, we now have the opportunity to reduce the number of construction hours and modular builds effectively. We know for a fact that cuts our build times significantly. That has been sort of there for the last year, but this is definitely there, and all of our data center clients as well, hyperscale clients as well, are all going towards 100% prefab. I have said this before, I think previously in other webinars, hyperscale customers in the past would come to us and look at, say, power train units or generators or certain things that are standardized that are prefabricated. That has all changed now.
Everything from the data hall onwards needs to be prefabricated. That is how they are thinking, and very few people outside of the U.S. today do that. That is the conversation, and that is our opportunity for that market segment. Next slide. Yes. We announced our maiden AI HPC contract in June, and it was for AUD 8.8 million with a U.S. NeoCloud customer, and that was for 1.4 MW. We have received a notification to look at cancellation of this contract for that particular site, and the follow-on campus, because this NeoCloud operator has decided that they need to co-locate as in the U.S. market as opposed to self-build. They do have a subsidiary outside of the U.S. that is looking at self-building, and there are conversations around redeployment in that respect.
There is conversations on possibly canceling this and redeployment in a completely new contract for that deployment in one way, shape, or form, but it's not finalized yet. What's important is that there was a disclosure of this contract, and it is important that we disclose to the market that that particular site and that particular contract is no longer viable, and that we would need to look at that cancellation with this customer. Our preference is to cancel this particular contract as well. Effectively, we have always maintained a cash flow positive contract milestone with all our clients, and this one as well. It's a pretty solid contract, so all cancellation costs, cost recovery, and committed right now is at AUD 4.2 million, which is half the contract, close to half the contract, and we are looking at terminating it at this point.
It hasn't been done yet, and when it is, we will inform the market. Should we decide that we're going to do a different arrangement of some sort for a different site, and use this as credit, we don't know how it's all going to land, but it is important that we share that that particular customer in that particular ecosystem originally had looked at self-build in the U.S. market, and as we know, in the U.S. market, things have evolved tremendously. Nasdaq-listed NeoCloud operators are either looking at consult, as I mentioned in the earlier slide, they're either looking at co-location or self-build, or both. A lot of NeoClouds are doing both. It depends on where they are in the ecosystem of the NVIDIA GPU relationship.
A lot of them are co-locating, in which case, and a lot of them who are announced by NVIDIA who are partners are going to be self-building. Those are the conversations we're in, and that's the difference between a customer like this three months ago or four months ago when we signed the contract versus what they are doing now. It doesn't mean that this is lost. It means that we might have to redeploy. But that particular contract and that particular site is not valid anymore, and it is important that we update the market on that. I think the cash has been collected. We have remained along the way through the contract. We have remained at AUD 1.1 million. That's expected in Q2, and we are expecting to recognize some of the We've already recognized some of the revenue, and we'll recognize revenue over Q1 FY 2027.
That's essentially what this contract update is about. What has this contract done for us? I think it's important that investors understand it. First of all, it's opened up the U.S. market and the U.S. ecosystem to us, and that ecosystem is now open to us globally. We have worked very closely with the ecosystem of the chipset manufacturers like NVIDIA, the consultants around who are part of that ecosystem that verify designs. Our solution and our design is very much within the NVIDIA architectural recommendation. We're working very closely with chipset manufacturers for variations of these designs in other parts of the world, including Australia. That ecosystem was not open to us until we signed this contract and had that exposure for our entire business and our entire team to understand and develop this product as we went.
It is a massive learning curve for our team, and the IP that we have got out of this is massive. Besides that, to me, that is a big plus. Besides that, of course, we are not losing money out of it. We get to keep the modules. We get to keep the product that we have purchased to date, and either redeploy it for this client in a different site or redeploy it to a different client, some of which we are quite in advanced talks with for other customers in the U.S., which allows for us to deliver at a much faster rate than we would otherwise have, because everyone is all about power, energy to token. That cuts the critical time for some of the UPSs and cooling systems that we already have on order for this project.
Those are two major things that we are taking away out of this. I will answer any questions that are up later. Essentially, from my perspective and the team's perspective, there has been a great opportunity for us to learn, first of all, and there is a level of confidence here that we can always redeploy this. If not, we will have the opportunity to be able to negotiate, close it out, and be able to resell this to many customers who want exactly what we are talking, what we have built. Next slide. The second AI HPC contract, what I want to update the team is that this is a 2 MW site. This is in region. It is not in the U.S. This is going well in terms of delivery. Design is complete with 2 MW. When I say design complete, design is signed off.
We are talking about already manufacturing. Steve can talk about it later and what we are doing with the build phase of it. This is a single high-density data hall. Again, a solution that is highly unique. Again, this is just back to what I said two slides ago. The customer is in the NVIDIA architectural ecosystem for delivery, as in for self-build, and that has evolved over the last few months. This particular customer is there, and we are right in the heart of it. That has been highly valuable as we move forward into much larger expansion for this customer. This site is definitely potent. There is potential for a much larger deployment as well. Next slide. Steve-O, I am going to hand over to you to present this and the next slide.
Beautiful. Thank you, Shalini, and good morning, everybody. I am joining you from Singapore, where I have been over here talking to the High Commission about our new facility in Johor. Before we get to that, I will launch into this one. It has been an exciting time for me joining DXN, coming from a rather comfortable job to this great new challenge and a great new industry that we are seeing evolve. One of the things that we realized that we need to do here at DXN is to accelerate the process of our delivery and to refine that. The best way to do that was to optimize what we are doing, standardize where we can, outsource our non-core activities, leverage our IP, and accelerate to market, as well as expanding our footprint.
You will see that when I first joined the organization, we had our primary or our center of excellence at Welshpool in Perth. Whilst that has been a very, very effective part of our business, it is not big enough for us to meet the scale that we need to meet the challenges and the demand that is on our services. The very first thing, as I said, was to look at optimizing what we do, and that would be to look at how we construct our modules and what they look like, shipping containers on the outside, but we all know there is a lot more inside them and actually are built around the structure. A lot of infrastructure is heavy. You cannot just use a standard container, so you need to have one that has been reinforced on the outside, the framework, and purpose-built, as well as having a metal floor.
They get built, and there is a number of components on those that are the same on every different module we do. The very first thing I wanted to do was to look at how we offshore or outsource our metal fabrication, because that is one of the longest time frames. What we need to do is create a chain of metal containers or modules that are coming to us that we can then apply our IP and fit out, and that is where I am heading. That is that optimization piece. The other part then started is the standardization. A lot of customers, and we have been very good at creating bespoke solutions for customers, and there is nothing wrong with that because that is what has made us successful to date.
But where we need to get better as we scale up is to have a metal module that is actually the same or very similar each time we deploy. If you think about some of those customers that we have got coming our way, they want to scale and they want to make sure that they can modularize and then expand. They land at 2 MW, but then they want to get bigger and bigger and bigger. I want to make sure that we can scale by using the same sort of framework. They will not all be the same, but I would love to have a collection of four or five modules that we can then pick from, and then we can add.
Once we have done that standardization, that gives us, and using that reference guide that we have, we want to accelerate that, and that is part of what we have been doing. If you look at the map that is on the presentation right now, you will see that we have gone from Welshpool. We also have Seven Hills up and running right now. We have got eight modules in that location right now that are being built and fit out. We are in the process of opening our Southeast facility, and I will be taking a road trip to hopefully sign the lease on that new building this week. But we have already started fitting that out. We have got all of the equipment that we need, and what we are going with a philosophy of what I would like to call a pop-up factory.
We build our factories and our facilities close to where our customers need to be deployed. We shorten that chain of delivery, and we make that more effective. If you look up to the left-hand side of the image, you will see Johor Bahru. As I said, I was there yesterday with the Assistant High Commissioner, opening up that facility, which has been fantastic. We have people on the ground now. We have a level of management in there, and we will be starting to operate out of that center probably in the next month or so, which is great. We also have our Jakarta facility, where we focus primarily on our CLS. If I was to take a snapshot of what we are doing, Johor will become a fit-out and minor fabrication factory. Jakarta will be a CLS fit-out organization.
Our Welshpool facility will continue to be our research development, anything that is bespoke, developing what we do as technology changes, and then we deploy that out to our other factories. Seven Hills will close down at the end of the year because those eight modules will be gone by then, and we will not need that facility because we will have our much larger Southeast facility operational by that stage. The aim is to open the door, get the facilities open, and then we will ramp up into each of those facilities. By the end of Q3, we should be fully operational right across the region, which is great. I really want to increase our customer intimacy as well. I think that what we are doing requires a lot of interaction and interfacing with customers, right from the design aspect to the scoping, to the creation of what the customers want.
That is where we need to go, and I think Shalini said that energy to token space, we really need to understand that, but then consolidate it and reduce it, because it is all about speed at the moment. I think that is sort of where that leads us to. As I said, the ability to scale to meet demand is the absolute number one target for us. As I said, that ability to have what I call a pop-up factory. We know what is required because we have already tried and tested and proven that in Welshpool. We have just put that same template into Johor. We are doing the same in Southeast New South Wales. We have done the same in Seven Hills. That enables us to move quickly, get boxes in, fit them out, and push them down the road. That is really where we are heading. Next slide, please.
Just talking a little bit about our facilities. As I said, Welshpool, fully operational. We are delivering four or so modules out of there at a monthly pace. Seven Hills, as I said, we have eight, and they will be out of there by the end of the year. Southern New South Wales, as I said, we are sourcing a new site there at the moment. We have all but put ink on the paper. That will happen hopefully on Monday when I get back to Australia. Johor, as I said, we had the High Commission out there, Simon Fellows, and a number of people from DFAT and Austrade. They have been extremely supportive of what we are doing in Malaysia.
Just to sort of put a picture around that, the reason why Johor, Malaysia, we want to be able to If you look at the maps of where we are deploying, we have a large deployment that comes out of Asia and around Asia. It is nice to have the facilities in different parts of the world so that if we get bogged down by storms or any particular thing like that, we have a capability to move around, and also to pick up overflow from some of the other factories. Malaysia will certainly help us serve into the Asian region. Jakarta, as we said, we have a really well-made and established organization where we are delivering cable landing stations, and that will continue to happen. We are looking at a new factory there as well, and we will continue to drive that through.
I am fully confident that we will be able to get to six times what we have been able to produce so far, and we will have a very, very effective production line that comes through. The key aim for me is, as I said before, optimize what we do, standardize our fabrication, look at the long lead items and outsource them where they are not our core business, focus on the IP that we bring, and that is what is inside the container, and that is where we should be investing our time and money, and then to accelerate that. I would like to think that we have already started accelerating by expanding our footprint and our operational capability across the region. I think we are in a really good spot. It is an exciting time, and I am glad to be part of that team. Thank you.
Thanks, Steve-O. Laila?
Thanks, Shalini. I will walk you through the numbers now. The revenue for the quarter was approximately AUD 3.4 million, and that was up around 380% at the same time last year. This is a like-for-like basis. Most of it came from the modular business, which was AUD 2.9 million. Darwin produced just under half a million with AUD 400,000, and DCaaS, AUD 100,000. On cash, we finished September quarter at AUD 6.4 million, and by the end of October, we were back up to AUD 11.1 million. This was slightly ahead of where we were in June. The dip was purely timing. Some of the customer payments that were due in September came in a couple of weeks late in October, and that is why our operating cash flow for the quarter also was AUD -4.7 million. Had those receipts come in as they were scheduled, our operating cash flow would have been a AUD +2.2 million.
In the quarter, we announced AUD 19 million of new contracts and variations, and that lifted our backlog to AUD 30.8 million, which is up from the AUD 23.5 million in June. This figure is adjusted for the canceled AI HPC contract. So, we just wanted to give you the full picture of what the backlog would be. Thanks, Shalini. The revenue for the quarter. So, this chart here shows Q1 was AUD 3.4 million compared to the AUD 0.7 million of last year, and this is adjusted to remove Hobart DC. The dotted line for FY 2027 is showing us where we would expect to land for the first half of 2027, which is around AUD 11 million. This figure includes roughly AUD 4.2 million of revenue that we are expecting to recognize from the NeoCloud customer, and most of that is going to come in Q2.
Delivering AUD 11 million in the first half of 2027 already exceeds our full year of FY 2026 revenue of AUD 10.7 million, which reflects the team's ability to execute our contracted work and convert into revenue. Shalini, can you Thank you. This bridge shows the backlog movement through the quarter. So, we started off with AUD 23.5 million, and like I mentioned, the AUD 19.3 million new orders and variations. The big one was the turnkey 2 MW AI contract. And also, we secured the Solomon Islands and American Samoa. We delivered in the quarter AUD 3.4 million of revenue, and we've adjusted for the AUD 8.6 million of the NeoCloud contract. The full contract was AUD 8.8 million, but we have already recognized roughly around AUD 200,000. So the quarter finishes off, like I mentioned, at AUD 30.8 million. Sorry, that AUD 30.8 million and our cash position currently supports us delivering it.
Cash and revenue won't always match in a given quarter because our contracts are based on milestones. However, we ensure our contracts are cash flow positive throughout their life, which we'll show in this slide here. So, this shows the cumulative cash over the life of projects, and it goes back to Q1 in FY 2024. And you can see the two lines move together. And we've actually collected a little bit more cash, and that's what we expect because we do try to maintain all our projects to be cash flow positive. We bill as we move through the project Sorry, we bill as the project progresses. And, like I mentioned, although this quarter was AUD -4.7 million compared to the AUD 800,000 in June, we did spend AUD 5.4 million on the manufacturing and operating costs and AUD 1.5 million on staff, and that is the ramp-up and the execution of the backlog.
And if the delayed payments did arrive as expected, we would have actually ended up, for the quarter, AUD +2.2 million. And like I mentioned, we're back up at AUD 11.1 million at the end of October, which fully supports us going forward. Back to you, Shalini.
Thank you. I am going to run through the outlook of the business, the outlook section of the slide. Thank you, Laila. Our current pipeline currently consists of 93 projects as of today. Several projects obviously were won in the past quarter, and we also have consolidated a few projects into one. As an example, if someone says, "We are looking at a much larger deployment," we typically split that up into multiple contracts or multiple counts of a project within the pipeline. If that changes to, say, one large deployment, then we consolidate that into one. There is a little bit of that going on with the 93 versus 99. I just want to qualify, though, that the number of projects are starting to become a bit of a moot point in trying to present this to the market.
Mel and I are going to strategize a little bit to give a little bit more color on, I guess, a different metric of some sort to the market in terms of pipeline. That is primarily because a lot of our deals are significantly larger than it used to be. When I first started at DXN, our modules business was between AUD 25 million, the entire unweighted pipeline. With AI now, it is over AUD 1 billion. The number of deals is not reflective of the opportunity that we are starting to see, and it will go up and down depending on how many variations and expansions there are for one customer that is in the pipeline that we have closed, for instance, and so on. We will have to think about how to better present this in the future. But for all intents and purposes, we are not going to take this away.
Just to give people perspective, 28% of the pipeline previously was 23%. Sorry, it was, I think, maybe 21%. 28% of the pipeline is AI compute, and this is still relating to number of projects as opposed to volume of projects. But, the one thing that I would say is that if I just to give some perspective of where we are this quarter versus last quarter, and we have got another executive who is recently joined the business, Darren McSorley, and he comes from NDY. He is a Chief Development Officer. His background is in data centers, in the data center segment. He is very well-known in the data center industry, which gives us a lot of new coverage within markets that previously were not open to us, as from a relationship perspective.
With a lot of what we are starting to see now with not just the NeoClouds, but the data center segment as well, there has been 115% growth in pipeline value, so not numbers, in pipeline value, and that is just one quarter, Q1 FY 2027 versus Q4. It is very much supported by the growing AI opportunities. That is just a bit of a highlight there. Next slide. In terms of our immediate near-term priorities, execution is everything. We have a backlog that we are going to deliver, and conversion of that into revenue and cash and ramping that up has taken a lot of energy and time off the executive team over the last quarter. People have worked really hard to get some of our facilities online, working with different partners, delivery, execution, logistics, just working through that and setting ourselves up for Q2 and Q3 and so on.
That has been really important. Our backlog has grown tremendously from past years, and converting that into revenue is absolutely critical, even in this first half. In terms of continuing on our Asia Pacific strategy, we have the short-term demand that we have got being met with our Seven Hills, New South Wales facility. As Steve-O has mentioned, we have got a new facility, a permanent facility that we have got on the East Coast, imminently about to be signed on now. Basically, at the end of that, we will be able to have three permanent sites to have production of standard fit-out. Again, that is a big focus for us to make sure that we are able to turn energy to token as quickly as possible in the most efficient way. Delivering scale across the business.
We have never, as a business, had the opportunity to get into that part where all our systems have all been bespoke for each client. We have now the golden opportunity to build this at scale. Standardizing engineering packs for each AI module, that is a big part of our strategic initiative for FY 2027. Standardizing this across the ecosystem, right from the supply chain to delivery, including contract that we have got, forecasting that we have got with our supply chain to get better pricing, get better margins. All of that comes under that category of standardizing engineering packs. Productizing our rapidly deployable modular platform. The productization part of it is also a big part of our work. This includes working with stakeholders like NVIDIA, being involved and being completely immersed in that ecosystem is 100% part of all of our KPIs within this financial year.
We are right there, and we have got that golden opportunity to do that. Of course, the momentum across the multiple businesses, the core business, we have got strong repeat business from our existing customer base that we get repeat orders from both landing stations, so CLS delivery, as well as the AI platform that we have set. I should not at all dismiss the DCaaS opportunities continue to grow. We are hopeful and we are confident we are going to continue expanding that platform. There is a lot of opportunities out there for not just satellite operators, but also smaller DCaaS, Data Center as a Service opportunities, even with AI clients who need small inference sites that we can actually then do on AI as a Service opportunity in regional remote sites.
That is a platform that we have always known is going to be available for us, which is why we set up the DCaaS business in the first place. That is something that we will have and definitely an eye on for FY 2027 for revenue. Next slide. I think that was the one. Oh, that was it. Yes, and that is the end of that. We have got a couple of other standard slides that we have put into the appendix. Thank you. Mel, do we have a few questions?
We do, Shalini. First, let's kick off with the maiden AI contract update. There are a few questions there. Hugh is asking in regards to the U.S. contract loss and second contract in Australia, which market are you seeing the most demand from for customers? Second to that, also noting the press overnight in the AFR on the government banning metro data centers in Australia. Would this impact your existing second customer, or do you see this as giving DXN some benefit, as in a shift to regional sites you touched on?
Yeah. Okay. Where are we seeing most demand on? Definitely, there is a lot of demand in the U.S., and not just with one chipset, but plenty of other opportunities. I am talking about everything from labs, which are multiple, 5 MW to 10 MW to 20 MW, right up to neoclouds and data center companies as well. There is plenty of opportunity in the U.S. market. The U.S. market is the biggest market. The fact that we have demonstrated that we are in that ecosystem with the chipset suppliers as well, and that ecosystem allows for those opportunities to come in. That is not to say that Australia market is not there. 100% Australia market and APAC market is also significant for us, which is where we have got our base, which is why we have got a facility being built in this part of the world for these clients.
The market in Australia is a lot more captive, and it is somewhat easier, maybe, to some degree, in terms of who are the players, because NVIDIA has come out and made sure that that is all super clear. As I said earlier, there are the opportunities, say, with neoclouds who are self-building and the ones who are doing both colocation and self-building. This is all becoming a lot clearer now with announcements by all parties, who NVIDIA is partnering with and so on. Definitely, it is a lot more clearer outside of the U.S., but in the U.S., is where the knowledge and the hub and the know-how is.
I do not think we would ever play outside of the U.S. if we were not learning and growing in the U.S. within that ecosystem, if that makes sense, because that is where the learning and the development and the relationships are built. The second part of that question was about the announcement overnight. Yeah, this is exactly why a lot of data center companies and neoclouds have been looking at regional sites where power is captive or power is stranded or power is needing to be generated as part of their ecosystem. No, it does not affect our second site because it is a regional site, and it is not a metro site.
There is going to be a lot more of those type of applications that we will see in regional areas where councils are hugely supportive in terms of utilizing technology that is going to be, as I think as I mentioned earlier, closed water loop systems, power generated on site, modular data centers built prefab and brought to site, creating jobs within the community and so on. That is where the market is shifting, and that is how we see a lot of the players are already thinking about it. This is not unexpected at all, and that is where the direction of the industry is moving across the world, definitely in Australia. The social license is really important, having that opportunity to work with the councils and work with some of the key stakeholders as well.
I am talking about the off-takers themselves directly starting to think of there is no way they can build 500 MW of data center in one location in the city anymore. There needs to be thinking and planning around multiple hubs and distributed architecture of how these GPUs work. That has been there in the conversations in the background for a while now. Hope that answers that question here.
If we switch to the pipeline, you mentioned in terms of the number of deals that you have consolidated internally, larger opportunities. Can you just talk to some of those consolidated opportunities? How advanced are they in the pipeline?
Yeah. If you can swap over to the pipeline slide. Yeah. If you sort of look at the pipeline as it sits right now, you can see 13% is in final negotiation and 5% is in verbal win. When we say verbal win, it is done deal, except we are waiting on contracts or we are waiting on payment terms. We are negotiating to make sure that we are cash flow positive at all times. These things take a lot of time at the end, and that is where we protect ourselves and make sure that we have got the right terms. It takes a little longer than we would like, but that is the discipline that we have to have in the business to make sure that we are always ahead of the game.
The process of working through the pipeline in terms of advanced, I think the question specifically relates to the volume of opportunities. Where I am excited about and our team is excited about is the fact that the growth in our pipeline, especially in the last quarter, has been tremendous opportunities coming in because of our conversations and our work that we have done with the first and the second AI opportunity that we have won. A lot of that has been inbound. A lot of that has been developed through the pipeline. So the growth of that pipeline revenue as we go through the first half and the second half and the bottom half of the pipeline has been significant. This is referring to the volume of the value of the pipeline as opposed to the number of deals.
Thanks, Shalini. Sticking with the pipeline, can you talk to the future pipeline? Is there any tie-ins with NVIDIA or other modular data center builders you can talk to?
Yes, absolutely. So NVIDIA runs the DGX Pod platform. DXN is obviously heavily involved in that. Our CTO engages in the GTC meetings, and over the last couple of quarters, we are going to be heavily involved much more in the next few quarters. It is definitely a recommended architecture. I think what I think is important is to remember that for the first time in history, NVIDIA has validated the modular architecture, the prefabricated factory-integrated approach, something we have been doing for years, and that is a significant step to where it needs to be. Within that framework, there is a whole raft of opportunities, OEMs, ODMs, vendors, ecosystem providers, partners, and we are right in that space where we have distributed factories in multiple locations being able to deliver and solve domestic localization requirements.
Example, building in Johor Bahru enables us to build and customize for the Asia-Pacific region. Building in Australia, as opposed to importing a prefab modular data center from a different part of the world, allows us to standardize firefighting and all of the local compliances that are small, you would think small problems, but basically kill a business at the end. This whole idea of having distributed local capital investment in multiple locations has always been our plan, and that is heavily supported by partners like NVIDIA. I think the question is, how do I think it is going to reshape the opportunity? Is that right, Mel?
Yeah, but also they've seconded that, Shalini. They've asked if you can just speak more broadly to the AI HPC opportunities within the pipeline.
Right. Okay, so with NVIDIA. Yeah. So, to clarify, most of the opportunities today are based on B300s and plans, GB300s and B300s. So, it's either air-cooled or water-cooled, and that's the chipsets that are available. So, it's really clear who has access to the GPUs and who doesn't, who has funding and who doesn't, who has the backing of NVIDIA and who doesn't. So, for us, that's the play. We work with partners who have that, and they have the balance sheet, offtake balance sheet to support that because a huge part of their capital investment is the GPUs. And so, without that, there is no business for them, and that enables them to ensure that they are funded and we are funded. And that ecosystem is extremely tight. So, if we sort of look at the pipeline, we only work in that environment today.
There are a few new chipsets, like labs in San Francisco and locations that we are working on, and we should always keep an eye on these developments, and new products and new chips that are out there. But by and large, most of the pipeline that is AI-related is NVIDIA GPU-related.
Thanks, Shalini. And just sticking to that, you mentioned NVIDIA in the Q&A, but Nick from Moelis has added, "With NVIDIA's recent launch of the DSX pod, seems to be a significant validation of the modular factory-integrated approach to AI infrastructure. How do you see this reshaping the opportunity set for DXN? In particular, is there a pathway for DXN to move further up the value chain by becoming a strategic partner for manufacturing or integration within the NVIDIA ecosystem? And have you had any discussions about what that could look like commercially for DXN?
Yes, let me answer that one by one. Yes, absolutely, it's a significant validation, as I said before. It was a great piece of update, and NVIDIA was, and a lot of the ecosystem were under embargo until that came out. How does it reshape the opportunity set? Very much so. We expect that we should be getting highly qualified leads coming through for us. Highly qualified, not qualified, and rather than us working through the qualification processes ourselves. That is going to make a big difference to us, and working in that ecosystem, because we're already in that ecosystem, really helps us do that. Is there a path for us to move further up the value chain? Yes, absolutely. How's that going to pan out?
That is all in discussions, and how we're going to work through the integration part of it with the NVIDIA ecosystem, all part of it. We are heavily in that mindset around strategic partnership. How would that look like practically and commercially? It would be very much working with partners that they've already announced and selected, meaning NeoClouds that they've already announced as partners, because that's all openly announced out there. We're already working with a lot of them. Our ability to provide solutions independently to all of these parties is pretty good. We don't have There's no sort of one We are working with one NeoCloud operator only. The entire NeoCloud operator system within the NVIDIA system works with each other anyhow, and people share and learn. It is just making sure that we're part of that ecosystem. Commercially, how does that work?
I think currently, Nick, it's still individual contracts and agreements, but I see that evolving over time because you have an opportunity to consolidate and improve the commercial model, and go up the value chain. I do foresee that that's going to evolve over time. But right now, it's all individual contracts with each suppliers because they have a balance sheet that is supported with NVIDIA GPU, with perhaps loans, that agreement. So, individually, we would have individual contracts with these customers. But over time, as we get more and more entrenched into the ecosystem, I foresee that there is an opportunity to standardize that.
Thanks, Shalini. I guess finally, we've had a couple of questions on the timing of the discussions with your maiden AI HPC customer. When did you find out about their strategic pivot?
Yeah. I think we need to qualify that they have not yet canceled, as in we have not agreed on a signing of the cancellation. I think it is incumbent on us to inform the market of the change in strategy and direction. We have taken legal advice on this, and we are at a quarterly, and we want to provide market updates. It is incumbent on us to provide market updates on this important contract. We are disclosing that we are working towards cancellation. It is not quite there yet. It can always flip around tomorrow, and we might come and tell you that they have decided to proceed. There is always this conversation that if we decide to proceed and we are going to go ahead, how long do you need to ramp up and restart? Our answer is two weeks.
That conversation is not completed yet, but it is incumbent on us, and it is important that we disclose this, whatever the outcome, rather than holding that at the end of a quarter. We can focus on actually the important things in the business, including making sure that we have got the execution of all of our contracts in play. That is currently still ongoing, and when the cancellation is finalized, if it is finalized, we will update the market.
Shalini, that brings us to the end of the Q&A, so I will pass back to you for final comments.
Thank you. Thanks, everybody. Look, I appreciate everyone dialing in. I know there is a lot of interest in the business in the last three months, in the last four months since we won our first and second AI contract. Obviously, it is disappointing for us with the pivot of our customer and having to deal with that. All of us are, if you sort of look around the room in DXN, within the team, everyone is super excited about the future. We are going to win some, we are going to lose some, but that is part of the game. The ones we win are going to win big, and I think that is how we all think. This has never been a straight line for DXN from this time we started in the business, in the modular data center business, when we were working on the fringes of the construction ecosystem.
To be validated by the most valuable company in the world, NVIDIA, as part of an ecosystem to build and deploy is a huge deal for us. That is a significant step change that we foresee over the next few years that is going to make us who we are. Thank you for all the support today, and appreciate everyone dialing in.