Virtual queue.
That slide is the entire group made up of both the computer network business. If you see this at the top right is compute, and this is network. Let's just move the slide here to break it down so that we can share how each business is performing. Today, we will be going through company highlights, financial results, strategic update, and then a guidance update. Let's just pause for a second and look back. I think it is over the past 10 months from today. In November, we actually could not fit enough on the slide for what we did throughout the year. Otherwise, it would be too big. On 26th of November, we announced the acquisition of Latitude.sh. That was our entrance into CPU and now GPU as a service platforms, totally automated.
We also announced an acquisition of Extreme IX, which is our entrance into India with 40 different data centers landing. It is the largest internet exchange in India. On 27th of April, this has happened in a pretty short period of time. Between 27th of April and today, we have announced AUD 1.3 billion of TCV. The first one was AUD 35 million on 27th of April, and then we announced a AUD 254 million GPU, CPU, network, and storage contract on 14th of May. Then on 3rd of June, we announced AUD 458.9 million of TCV. We also launched our GPU on-demand pool. We did a capital raise via entitlement offer, which was AUD 827 million. The day after, we launched storage, so it has been pretty busy. Today, we announced another AUD 506 million of TCV contracts, which were announced today.
That is obviously forward-looking, so that is not reflected in the FY 2026 numbers, and we have got an announcement in the ASX on that. We have also been incredibly busy, so a huge shout-out to the finance team. I think a lot of the investors, as we have spoken in the past, have asked us about how we are looking at this from a debt perspective, how we are looking at all of our capital management. We have shared that we have been in a process for a period of time. We are super pleased to announce we have got AUD 825 million of new debt facility, which has been announced as of today. So that is fantastic news, which sets us up for the future. Again, the highlight here is AUD 1.3 billion of TCV contracts signed since April 27 to today. How does that look looking backwards? Let's talk about the FY 2026 ARR chart.
You can see, quick note at the top right, this is the group business. You can see broken down is the network and then the compute business ARR as at June 26. AUD 395 million of ARR landed in that year, which is an astounding achievement. If you break it down from a group revenue perspective, we are at AUD 312 million, which is a 37% year-on-year increase of group revenue for FY 2026. Networking ARR is up 27% in constant currency to AUD 289.6 million. An incredible achievement, and I am very, very proud of what the team has achieved here. On network net retention, we break down the net revenue retention effect for the network business specifically, separating out the compute business. I think this would be appropriate to say that that little thing does not represent that. This represents the network side.
We are up 5 percentage points year on year. That is 114% net revenue retention. Probably one of the most important metrics that we have been focused on in the business over the last three years. This is an astounding achievement, and we are going to walk through some of those pieces coming up. The compute ARR. Since the acquisition, which we announced late November, we are up 72% as at June to AUD 105 million of ARR. Just an incredible achievement for the Latitude business and a massive shout-out not only to the founders that have joined us, but to the entire team and to all the new folks that are joining the business and helping scale. It is a story of massive execution coming out for the next year and continued growth. Let us look at the annual recurring revenue breakdown for the network. There is the network logo up the top right.
This is representing the network annual recurring revenue throughout the year. You can see the breakdown between Americas, obviously the largest component of that, AUD 290 million at the top, AUD 166 million in Americas, AUD 77 million in Asia Pac, AUD 47 million in EMEA, all regions growing. But I think I want to draw your attention really specifically to the ARR incremental addition. This is the net increase in ARR on top of the start of the year to the finish of the year. You can see in history, we have been performing well, but look at FY 2026. This is 103% increase. It is on the largest ever year-on-year growth that we have ever seen. So the largest year that we had ever seen, we have over doubled that in net and ARR incremental additions to AUD 62.3 million in FY 2026.
An incredible proof point that the investment that we made actually this time last year back into the business and all those pieces we have been setting up from a product standpoint is paying off in dividends. Really exciting. This is the compute business and just a reflection of how the ARR has grown inside that business. Right now, we are hitting sort of a little skate park here as you come up this side. You can see since the acquisition in November 26, you can see the extreme growth that we have had here just in that very short period of time. This is not even taking into account all of the TCV contracts that we have got in place. I think there is a very small amount of ARR from them in there. So the future is looking incredibly bright. Again, congratulations to the Latitude team.
What a magic partnership between us, and that is showing right there. We have shared our DC cohort's ARR contribution over time before. It is a little bit of an eye chart, but it certainly reflects how important it is for us to continue to invest in building out new data centers. I will just sort of explain what you are looking at here. Each group here represents a cohort of data centers that we have added. If we go back to this very first one, in FY 2022, we added 26 different data centers, and it started in that year with this level of revenue. What was really important is that the following year, that is when you start to see the revenue come to fruition when you invest in building out a data center, and you can see its compound interest from there on after. We sort of shared this back here.
When we first came in, there were very few data centers added, and we started to say we need to continue to add. What we saw in FY 2025 was we added 115 data centers. You could see how that landed from a revenue perspective, but have a look at the following year from those data centers. Astronomical. The point of this is every time we add these new DCs and get new revenue, it expands our TAM and helps us grow the business, but it is the ongoing growth of that. Have a look at what we have done in FY 2026. The team is on fire. They have added 155 data centers, landed the revenue at this component, and if you play this forward, you can start to see the impact of compound interest of building out the TAM.
This is a really key part of the strategy as we constantly continue to invest in building and showcasing why we are doing that. You can see we are up 35% year on year in terms of data centers. What is really interesting is we are landing new data centers with much higher ARR because we have got so many more products that we are selling into that customer base. So it is a real key example of the strategy coming together. Probably one of my favorite charts. This sort of represents what that compound interest around net retention and how it plays out in the customer cohorts. Just to sort of quickly look at it. Whilst it looks very pretty, and it is very pretty, each color represents a cohort of customers, i.e. let us pick a particular color.
So in June 2022 to June 2023, so FY 2023, this very small piece of revenue is the customers that landed. Then what happens to them the next year? They expand, and then the year after, they expand, and the year after, they expand. So that is actually showcasing it is incredibly sticky but also expansive, which is your net retention number. But what we have done right up the top here is we keep stacking these on top of each other. The FY 2026 cohort ARR has landed 55% higher than our previous record. It is showing an incredible result from the go-to-market team. Again, back to all those new products that we are landing and innovating. We get more revenue, more wallet share, more TAM per customer, and we are proving that out as to how we land.
It is always great when we tell you what we were going to do, and we can show you historically that we actually did what we said we were going to do. This slide, there is a lot of data in here. We are putting this in the deck more for when we go through the investor roadshow, so I am not going to walk through every one of these. But in short, the execution against the strategy throughout the financial year was nothing short of astounding for what was a pretty small team. We added 155 data centers. We have added so many different products. If you think about the strategy of the company, there are three pieces. We build, and that is how we expand and add different total addressable market into existing products. So that is if you look at it, Megaport infrastructure across 40 sites.
We build out storage hardware. We build out all this infrastructure, all these different components. We upgrade speed, 400 G, you name it, et cetera. On top of that, there are probably a few things worth calling out. We had three times the expansion in 100 G internet locations, which is a 200% increase in market coverage. Just one little example of going big on the products that we are seeing massive success in. We are constantly testing products and all sorts of things inside the business. When we find success, we double down behind them and continue to expand. 67% increase in 100 G-enabled locations, for example. 58% increase in bare metal servers. Huge investment. The innovate component is about us bringing out new products. So this is our engineering teams actually building cool stuff. We launched Cloud Storage, DDoS Protection. We launched MCP.
This is all the AI components that are sitting inside that, and virtual machines, Kubernetes, Agent Hub. There are so many things that came through. Cloud Router, IPsec, Packet Filtering, Ping, Traceroute. The team has delivered that in that period of time. Each one of those steps us further away from anyone that would try and do what we do and actually adds a TAM to the component. The last piece is the investment component. We talked through a few of those. We have acquired companies. We have added capital to it. We have invested in all the different folks. We are diversifying our capital base. All those different components. We will take the investor roadshow if we need to go through that. All right. The financial results. This is the exciting part, folks. I am going to hand you over to Leticia Dorman, our CFO. Cheers.
Thanks, Michael. The FY 2026 performance, we just have a slide here really to touch back on the guidance that we provided, and where we have landed compared to that guidance. So reported group revenue is AUD 312 million. That is made up of the network revenue, which has been since experienced significant growth over the last year, landing at AUD 268 million. And then the compute revenue, which we talk extensively about, and what we have landed a number of strategic contracts that largely will land into FY 2027 for the impact onto revenue. However, that compute revenue against there is around AUD 44 million. EBITDA is above the top end of guidance at 25%, made up of a variety of components, being revenue and a range of costs in there, which we will talk through on the next slide.
CapEx around the guidance that we had AUD 90 million -AUD 100 million, which excluded any strategic contracts or initiatives, landing at AUD 98 million. Here we have a fair bit of text, which is helpful for post results for a few of the folks on the call. The call-outs here, revenue up 37%. That is, we did acquire two companies, with revenue in those companies, Latitude and the India acquisition. That, but also the underlying network business itself, with significant NRR growth and new logo acquisition during the year has added to that result. The direct network costs, it does reflect the ongoing investment we have talked about extensively previously. 155 net new DCs, and we continue the 400 G core backbone upgrades throughout the globe. Our employee costs are up substantially as planned and as communicated a number of times.
As the business has been rebuilding, that cost has increased. We also have incorporated the two entities, and in much bigger group, and we continue to invest in the business, particularly to support the scale of growth and opportunity and revenue. Other operating costs, there are some higher non-recurring expenditure associated with the two acquisitions in particular. We are now located heavily in Brazil, India, and the U.S., so there is a significant amount of travel, as we start to merge the three entities together, and some heavy marketing activity to support that as well. EBITDA at AUD 77 million, which showcases that 25% margin there, is also a call-out there for us. Employee costs. This is some insight into where we have invested for employees and continue to invest, and you can see that insight there largely around even the LTV to CAC of 5.8.
That really just reflects even with the go-to-market investment, which continues to scale, that supports the revenue growth in line with that. We are scaling the customer support and deployment functions to enable the compute build-out. That is a really critical component. We continue to invest in the product and engineering, as we continue to scale out the onion rings. I think Michael will talk to a little bit later in this, around the continuing expansion of products. Cash flow. That is up pretty materially. We did do an entitlement offer in the month of June, and so the cash had landed in the balance sheet. We continue to place deposits on significant orders of GPUs, and as we start to have those be delivered, we will continue to utilize that cash, in order to start to support the rollout of the strategic contracts as planned.
You can see here the operating cash flows of the combined entities. That is driven by the inclusion of the two entities as well. This is a consolidated cash flow, as with EBITDA, so some pretty significant changes in the business overall, across the P&L cash flow and balance sheet that you will see. Michael has talked to this previously around the new debt facility where we have commitment letters signed by a syndicate of leading domestic and international banks. That has been a significant process that we have undertaken over the last few months, and so we are really excited to announce that as part of the overall funding strategy that we have undertaken to support growth for the consolidated group. We do continue to explore further funding strategies in line with our capital management plan.
This is really designed to highlight that where we see the growth, then we will continue to invest our cash in where we want to build and innovate to enable growth for the future of Megaport. That debt facility will form a core part of the overall funding strategy for that. Back to you, Michael, on your strategic update.
Fantastic, and congratulations for a huge amount of work from the team. Just acknowledging the finance and HR and the legal team that have made two acquisitions come to fruition. The finance team has done all the multiple announcements, since April, plus the end of the financial year results, and on top of that, a debt facility process, which has been astounding. Tish, you are allowed to have a little break, maybe a couple of days after this. Strategic update. For those of you who have not been following, we are now what is a globally distributed automated infrastructure company. People are like, "Well, what is Megaport and what do we do?" We do three things. We deploy compute, network, and storage, which are physical infrastructure assets. We deploy them in data centers all around the world. We own that infrastructure. We use our capital to procure that.
We physically install it in the data center that we lease. We then stitch it all together with fiber, and then we run that. We run a software automation layer that we have built and manage in-house. That is the crown jewels of the business that automates the deployment in seconds for all compute and GPU, for all network globally, and then for all the storage elements as well. The three pieces make up the trifecta to what is all IT infrastructure service in any application that you have ever used in your life. You can see how we are distributed and broken down around the globe. We are in 31 different countries. We have 330,000 pieces of physical fiber that we lease and stitch together 3,000 network devices, 13,000 plus GPUs and CPUs, and scaling at a rate of knots across 1,100 data centers.
This is an incredible business with an impenetrable moat that is just scaling at a rate of knots. Look at this thing, it is incredible. We are super proud of what has been built there. As I say, a picture is worth a thousand words, and a demo is a thousand PowerPoints. Instead of putting 1,000 PowerPoints in here, I am just going to do a very quick demo to show you what that actually means, assuming I can switch over here. Just checking everyone can see my screen. Is that coming through? Yay. All right. What you can see here is planet Earth. We have two options. We have the flat Earth version, which Leticia is fond of, and we have also got the spherical version. I am a spherical guy. Let us zoom in.
What you have got here is every single data center that Megaport has rolled out across the globe from a network standpoint. This is specifically looking at the network part of the business. We are compute, network, and storage, and this is the network element. If we just take a look at the United States, this is me logging into the portal as a customer. I can see all the different connectivity and different products that we have deployed. We will pick a clean location, and we will just zoom in on that. We will take 13 data centers clustered over in this particular location. We will keep zooming in. There is 12. You can see all these different data centers sitting in this particular location. We have got NTT Global, we have got QTS Portland, Flexential, EdgeConneX, and we keep scaling. We go right down here to two. And there is Digital Realty Portland.
What you have got in this very small location in Portland is a whole range of data center operators that we are currently live and active in. This is when we talk about deploying data center locations. This is what we do. What happens is a customer has their own compute hardware sitting in this particular data center, and they need to go and connect. Let us pick EdgeConneX here. We click on there, and we can choose what product we want to create. We want to create a port. We want to get access from our server that we have got in that data center, and we want to access connectivity to the outside world. We have picked the data center. We click Next. We scroll down, and we choose what site. It is a port. This is the Megaport, the 100 G location. We will give it a name.
We will choose whether you want to add a cross-connect, which is automated for you. We can actually deploy that. Month-to-month terms, you name it. You can add all sorts of technical details. We will just click Next and click Add the port. What we have done, as you will see, it has gone blue. We now have deployed physically a physical 100 G port in that particular location that I am going to connect my data center infrastructure to. I could be a bank, and I am running a server in there with all my applications, and I need to connect somewhere. We take the port, and we go and add a connection, and we say, I want to connect to a cloud, to a private data center, to Megaport Internet, to an internet exchange, to a marketplace, to a service key, to Megaport Storage, and we keep adding more products.
Let us say we want to connect to cloud, and we will scroll down. These are all the different cloud locations. We will choose Amazon Web Services, a hosted connection. Every one of these locations is physical infrastructure that we have built out around the world, over 300 different cloud on-ramps. You can see, these are all the different connections. A blue and a red connection adds diversity. It is that simple. Let us choose a location. We will choose up to Seattle. You can see how many milliseconds delay. We will choose a blue. We click Next, and then we give it a name. We give it a speed. Let us say 25 G, and we just click Next. All you need to know is your AWS account ID.
The platform automatically stitches into the AWS portal, gives you a piece of fiber, in effect, that is yours, that you can access through that at 25 G from your data center. We click Add, then we are done. You will see that connectivity is now live. It is that simple. We go back to that particular location. Every time we go and add products, we grab that same product, and we say we would like to connect that to, say, a Global WAN platform, and we want to go and, say, connect into Brisbane, Australia. There is NEXTDC. Let us go in. It is 148 milliseconds all the way from the U.S. to that particular location. We will click Next. We will give it a name. We will give it a speed. Scroll down. You have all this cool technical stuff you can add in. Click Next, and we have deployed.
What we have done is we are traversing across many, many different subsea cable that is already built. You are now stitched in and connecting across to Australia. That becomes the network component of this piece, and the network stitches it together. Then we go into the compute side of the business, and so the exact same thing. We look at live demo to shareholders, and we go in here, and we have this beautiful thing that says Create a server. When we click Create a server, first of all, we have created network. Now we are going to create compute. You have options. You have bare metal, bare metal GPU, virtual machines, and there is a whole range of other cool stuff that the team keep building. These are the locations that we have physically deployed infrastructure in.
Imagine walking into a data center and seeing racks and racks and racks of servers all spinning, lights on, and you name it. Let us say 70% of those servers are customers and 30% are waiting to be deployed. Let us take North America. Here is Ashburn. That is where the home of the clouds. We will scroll down. You have all these different types of servers that you can deploy on demand at the click of a button. Core optimize, memory optimize, storage. Let us just choose this one, rs4.metal. You can see how much it costs per hour. You can scroll down. We can even pre-deploy an operating system. The platform is searching through all the servers and deploying operating systems that make sense that we expect a customer to utilize. Why is this important?
Imagine pulling an iPhone out of the box, and it did not have an operating system on it. It would now take you 20, 30, maybe two hours to download and then get it up and running. We can actually pre-deploy the operating system on you, which means it deploys in five seconds. We can scroll down, choose hourly. This is live. I click Deploy hourly. In effect, what it is doing is taking a server in that particular location, it is deploying it, and in less than five seconds, it is on. On with an operating system live in Ashburn. We can click into that. I cannot tell you how complicated the software is to allow you to physically do that. Mind-boggling. Massive shout-out to Edu and his team that actually made that happen. If you have a look down here, you can see all the different options.
This is a live server. It is available. You can ping it. You can put your own software on, do whatever you want to do. Then you can manage it. You can say down here, look, I want to out of band remote access, move it, install it, delete it, you name it. You can scroll down and add a database, add Kubernetes, add a network, or add storage. This is the next pillar. What we have talked about is network, we have talked about compute that we have now deployed. I would like to add a storage bucket onto the side of it. If you think of it, in your laptop that is sitting in front of you have a compute, little bit of Intel chip or something sitting in there.
You have some memory, you have a storage, which is a little bit of a hard drive, and then you have a Wi-Fi device. That is your compute network and storage in terms of a laptop. It is what we do, too. It is really that simple. We click Storage, we click Object, Block. We have different options. Let us take Object Storage, and you can start to create a bucket. It is really simple. I am going to create a bucket, and I get to choose super high-performance storage delivered in seconds on our VAST platform. Then we look at our standard cost-optimized storage that crosses between VAST, Wasabi, and a few other components inside there. Now you can start to see where we have deployed that. Let us look at a very low-cost storage platform that you can deploy from anywhere.
Let us just say you wanted to create a backup out of your cloud to make sure that if there was a problem with the cloud, you had a full backup that is actually protected. If, say, you were attacked from a cyber perspective, you have complete control. There is North America, all the different locations, Europe, Asia Pacific. This is a truly global platform. Let us say Ashburn again. All you need to do, imagine that you are trying to protect your data. If you have ever heard of these sort of CryptoLocker and so forth, where they break into your laptop and then they encrypt all your data and ask you for Bitcoin. 70% of the time that that occurs, it actually has already encrypted your backups. Your IT team says, "Don't worry, I have got a backup." They start downloading it, only to realize that is also encrypted.
How do you protect that? One button from Megaport. You click this, you go right, and you have object locked it. You click Retention, how long you want to keep it for. For governance, you click 30 days. You have an immutable backup for 30 days. You click Create the bucket. That is compute, network, and storage, all delivered via software, all on demand. The only thing I am going to go back to is I am going to delete this server because what it shows you is the fact that it is on demand. I actually want to sell this server to someone, so it is live. Let us just delete this. I will have to copy and paste this piece so I do not fat finger it, and I have deleted it. Now what has happened is that server has been deleted. It gets wiped three times, made available back in the portal.
Okay, hopefully that gives you a perspective of what compute, network, and storage really is without 700 slides. We go back to this. I think this is probably the, I would say, almost strategy on a page. These are the investments in strategic pillars that unlock TAM, total addressable market. If we look here, I will just get my laser pointer. Each one of these rings, Tish called them the onion rings. Each one of these onion rings or rings, in effect, represents a product that opens up total addressable market.
One of the key pieces to continued net retention, one of the most important factors if you are investing in a tech business, is to ensure the technology is awesome, customers love it, pricing, and all those components are covered off, but the company is continuing to invest in new products and making things that solve customer problems and that they love. If you do not do that, you do not earn the right to get your net retention. Your growth starts to slow. This is this sort of continued investment. You can see in the past three years, we have gone from what was a cloud connectivity company in Virtual Edge to adding Global WAN, Data Center Interconnect, DC internet, NAT Gateway, and a whole range of security that we have just launched in the last however long, including DDoS and IPsec, and you name it.
That represents the network part of the business. The compute component, which is CPU and GPU elements, came with the Latitude acquisition, and that team innovated incredibly fast and added storage to that pillar. You can do two things. You can continue to add rings, and you can also expand the ring. The expansion of the ring is building. That is when we added new data centers that I talked about before, new markets, new countries, taking the same product set and landing in new locations, or even just expanding capacity, going from 10 G- 100 G, from 100 G- 400 G. That will actually expand the ring. That is the build-out. The innovator, all the developers and engineering teams sitting there building awesome compute products, network products, storage products, AI products, and security that you will see continually invested in that space.
You should expect from us constant updates around what we are bringing to market to solve customer problems. The last one is where we invest, and so we expand product and engineering to constantly invest in that space around innovation. We expand the go-to-market because the last piece that you have is when you build beautiful core products, you need to go and tell the world about them and take them to your customers. The last piece is you can build that either internally with your innovation team, or you can look to acquire strategic acquisitions. Latitude and Extreme are great examples of that. Actually, they represent two things. Latitude adds a new ring, whereas Extreme expands the existing ring. You are actually seeing acquisitions that hit both expanding a ring and adding a new ring.
If you had to ask what our strategy is, we are adding ring and expanding them. Then you would say, "Well, what constitutes a focus from a product standpoint?" These are the guiding principles specifically for product or an acquisition that we are looking at. We are very disciplined around what we look to acquire. If you look at Latitude, we acquired a company the same as Megaport. Total automation, super important. That gives you the ability to be instantaneous. You saw us deliver that server in less than five seconds. If you can automate the infrastructure, you can build it at global scale. If you can make it incredibly resilient, I can do something that you cannot do as a customer. I can do it much more resilient than you could do yourself.
If I make it flexible, so you can turn it on, turn it off, add it, change it month to month, per hour, whatever it is. You can make it super flexible. If you can make it self-service, you can add this magical thing called product-led growth, where people jump on the platform and start to add products, and you reduce the amount of team that you need to service it. If you make it super easy, just that alone is what you have actual entire companies existing just because they are easier than something else. Add the best support, make sure the pricing is super disruptive, and make sure that the whole thing is profitable, and you have the ingredients for an incredible company. That is Megaport. That is Latitude that we have brought into the business, and everything else will scale from there, and every single product lives and breathes off of that.
We have done some very, very large strategic deals, whatever it was, 1 point something billion dollars that we have gone and announced. A lot of folks are looking to say, "Well, how does that play out? What is the timeline? What does it look like? It is different to what you have ever done." What we have sort of broken down is just so that you can have an understanding of the process that we go through. There is sort of sales, finance, and legal procurement, and then operations. It is pretty obvious, you could play this through, but I think we get a lot of questions about how it works. First of all, we would have a customer that we would talk to.
They either come to us or introduced to us, they are existing customer, and they would say, "We want to do a whole range of CPU, a whole range of GPU," or whatever it may be. Now, we do not have it in the portal or platform, so these are custom strategic deals. We would work on pricing, which is pretty tricky because pricing is only available for a very short period of time because of what is happening from a market perspective around supply-demand, particularly memory if you are across it. So two weeks and pricing changes, so you have got to be very, very fast here. Contract negotiation can happen in parallel or very quickly. This is a very quick process. That can happen incredibly quickly too, or for traditional enterprise, a bit longer.
You need to operate very quickly in this environment because by the time you've gone through your contract negotiation, your pricing could have gone up. Everyone moves fast. We need to ensure we're funded, so we can't announce or do a land a deal that's not fully funded, which is what we've gone and put from a debt perspective. Before that was the capital raise, before that was the debt, et cetera. We then land signature, and we announce that obviously, if it's material to the market, which is why you've seen those announcements. Separate to that, we look at DC contracts. We have to procure the hardware, and then we also, at the same moment, need to make sure it's got a home to live in with power.
That is a very fast process to land, actually, the data centers that we're going to land in. In many cases, we're pre-talking to hundreds and hundreds of data centers anyway because we live there, but that's a very fast process to land that contract. Then we would say there's an equipment delivery period of time. This is where I think the market, it's just worthwhile being aware. This can take time, and I think we've called it out on the next slide, six to nine months for delivery and all these different components can play out. To be clear, you don't deploy the CapEx until it's sent, but you do have deposits that you pay prior, and we still have to announce the full thing fully funded is how we've played that. So we've got GPU, CPU gets ordered, network gets ordered, and storage.
These are the pieces of the puzzle that make up the AI, data centers that we're building out, mainly for inference and also some, you can look at that for training and so forth for smaller pieces. This becomes the inference cloud in effect that our customers procure. The team then builds out the data center. The data center actually has to get delivered on their site from a data center perspective, which is the build-out, the ready-for-service components, putting all the power, making sure it's available, getting the racks actually physically installed. The team then rack and stack. It's a very fast process because we're only talking about small numbers of servers. It's a fast process. They call it a week to rack and stack, and then a month to get the whole thing set. It depends upon the size, the scale, the complexity.
That would take time. We test, make sure it's up and running. We hand over and give delivery to a customer. From that moment, it instantaneously starts billing. That's why you'll see step-change jumps in our ARR throughout the year as we turn on the take-or-pay, which is the entire cluster, then it just starts billing. Then we have a live customer, and we continue to manage from an operations perspective. I spent a bit of time on it because I think it's important. These are pretty sizable deals that we're doing, and we're rolling them out in many different locations and executing many different contracts to many different customers. The team is on fire. So quick update on the strategic contracts and the GPU pool. Strategic contracts, what we have, as we pointed out very specifically, you need three things.
You need the power, you need the network, you need the compute element. You also need storage. You need to procure all the infrastructure to get that power. Lots of questions around whether we have access to that, et cetera. Just clearing the market here. We have equipment, space, and power has been procured for all contracts prior to today's announcements. Delivery and deployment is progressing through FY 2027. The addition of these new long-term contracts brings AUD 1.3 billion of TCV and a total ARR of these long-term contracts. Once deployed, remember, as we deploy them, and it will take time, we will add AUD 435 million of annual recurring revenue, and then we have supported that with AUD 826 million of capital expenditure. That sort of breaks down the strategic contracts. On the right side, we are building out a GPU pool, which we spent a lot of time taking the market through.
It is basically exactly the same infrastructure in clusters that we build out. It takes six to nine months to procure, get that deployed, and then you have got to ramp. So once we have physically deployed it, we ramp that to the market. That could be three to six months from deployment, and then you look at your average paybacks sort of distributed through that for the on-demand platform. We are one of the very few companies in the world that I think will deliver true on-demand B200s as an example to the market. We have shared this slide in the past, and this is sort of like, yeah, okay, well, geez, Megaport is changing and what is happening and where have you come from? Where have you been? It is funny because when we look back, we have had so much transformation in this year.
But if we bring ourselves back to this moment last year, I asked, or in effect, we told the market that we were going to actually invest more in building out the go-to-market teams and start to grow. We went through this period, we were building the company. There was that period where the world caved in and we said, "Look, let us prioritize profitability." We saw a decline in net retention. The entire world sort of tightened post-COVID. We then did this massive transformation where we said, "All right, we are going to start to invest in the company because we are seeing all the net retention stabilize. We are seeing an opportunity for growth." We invested in the last year to continue to scale that and to see that revenue to start to scale. Now, we said that that is what we would expect.
We had a transform and reset, and in the FY 2027 to 2029, we will see acceleration in revenue, where you can see that we are ahead of it. We are already seeing incredible growth from those investments that you allowed us to do as our shareholders, and we are returning that to you. Next year is continue to accelerate revenue, capitalize on all those prior investments, continue to expand the TAM, execute all these contracts, aggressively grow that market share, and we are going to continue the investment with revenue growing faster than costs, as we promised. We have accelerated the revenue through the reinvestment as well. So what you have seen is actually our revenue has been accelerating, and I showcased that by the incremental ARR, the net retention, the growth rates. Everything is to plan. Okay, cool. What is in the future? What is beyond that?
FY 2030 and beyond, if we continue at the rate of scale that we're at, which you can see how we're performing just in the last, since April, you'll see us continue to build out this business. We're going to be significant scale at FY 2030 and beyond. At that point, your growth off that base is so large. We'll be a global leader in automated infrastructure as a service, and you're sustainable 20%+ growth. When you grow at that rate with infrastructure, you become highly profitable, and you convert a huge amount of that scale into free cash flow. We're still in an investment phase to get ahead of it, but as we get up to that point, it starts to spit out cash, which is the future on that side. Here's the guidance update.
I know you're all waiting for this. We get to talk through the success of what we've built and get really proud about that. You've probably all just flicked to the last page. I'm still excited about the previous. Now let's give you the full guidance. This is pretty astounding, actually. We're going to call group revenue for FY 2027 between AUD 620 million - AUD 730 million. That is 100% - 130% year-on-year growth from a revenue standpoint and guidance. You can see the breakdown between the two businesses. The network business is AUD 315 million -AUD 325 million, and the compute business is AUD 305 million -AUD 405 million, is the contribution of those two businesses in there. Our EBITDA is moving to 38% - 40% of the revenue that lands.
Then you can see the CapEx that we'll be deploying inside that year to service a lot of those existing contracts that we've just announced. That's where we're at to date. That's a pretty exciting finish. Then we'll hand over to questions. Tish is going to join us, and I look forward to taking some questions from the audience or our analysts.
We will now begin the Q&A session. As a reminder, if you would like to ask a question, please use the Raise Hand feature to be placed in the virtual queue. Attendees are limited to one question before you can rejoin the queue to ask a follow-up question. We ask that you keep your questions short and sharp. Given the time constraints today, unanswered questions will be responded to post-event. Our first question comes from Siraj Ahmed with Citi. Please unmute your line and ask your question.
Morning, Michael and Tish. Can you hear me okay?
Yeah.
Yes.
Great. Just a two-part question. You were not able to procure kit for two contracts. Can you just expand on that? Looking at the FY 2027 CapEx guide, it does seem like it's a one-off issue because, correct me if I'm wrong, that AUD 1.3 billion would mean you are essentially procuring all the kit for all contracts and the GPU pool. Can you just confirm that? Thanks.
Yes. You are correct on the CapEx. That's the plan for the CapEx piece. You'll see that we didn't have the only time it triggers CapEx is actually when the equipment is delivered. That's when you have to pay a deposit up front, and then once it's procured, it then converts into CapEx, and you pay the remainder. Say, for example, you might pay 20% upfront deposit, and then once you deliver the kit, your CapEx actually triggers at that point. I'll just answer the second part. In terms of the equipment, the contract announcement we made on the previous ones, I'll probably leave that to Michael to describe, because I think there's a little bit nuance to that. But yeah, it's largely around technical specs.
Yeah. We're rolling out lots of different technology for our customers. We've got lots and lots of different contracts that we've been rolling out. One particular contract was associated for a particular hardware that we weren't comfortable with the spread of the different hardware for the automation system. We changed that with multiple customers, in effect, to a larger TCV and a much better contract for a specific infrastructure that we would deploy for the automation system globally. That was done very quickly. We've announced that, in effect, inside these pieces. We continue to deliver more ARR, more TCV, and then we've got more contracts on top of that as we continue to deploy.
It just shows the robustness of the market at the moment, and the fact that actually we have this unbelievable amount of demand in the industry at the moment across not only just CPU, GPU, and storage, but also the network elements that stitch it all together. It is an incredible time to be in the industry.
Got it. If I may, Tish, just follow up on this? If you are spending all the CapEx, that would mean that FY 2028, when you enter FY 2028, you are essentially run rating the whole ARR from the contracts, which is AUD 540 million, or compute ARR of AUD 540 million and the GPU pool, right? Is that fair, or is that too aggressive?
The GPU pool, you deploy the CapEx for that, and there is a period of ramp. That is why I think this is important that we shared this slide, and we will just go back to this because I think there are a lot of questions that come out.
Two is helpful.
Yes. I think the point is, if it's a GPU pool, once you've got deployment, let's say we hand over here, this is the GPU handover, but the handover is to the pool, and from that moment forward, you would have a ramp period as all the GPUs start to get taken up, which is different to a strategic deal. The strategic deal is instantaneous, which is what I was saying. The strategic deal is take-or-pay the entire piece. As soon as it's delivered, it instantaneously starts revenuing. There's just a nuance between the two pieces of that. Everything in these strategic contracts, as soon as they're handed over, the ARR will hit at that moment in time. Just reminding folks, ARR is different to revenue, i.e.
ARR, let's say ARR hits at December mark, we'll only get six months of that booked in revenue. Timing is a huge piece to this puzzle, and there's lots and lots of pieces that make up that timing, which is why we're sharing this component. By the way, to get to this handover, every single one of these components must have occurred to hand it over. If one of these delays out, then you can see how that changes and moves, which is why it's tricky to land on the exact moment of time that everything starts revenuing, which is also why we've got a range from a guidance perspective. Hopefully. Obviously.
Super helpful. Thank you.
Thanks, Raj.
Your next question comes from the line of Eric Choi with Barrenjoey. Please unmute your line and ask your question.
Hey, Michael. Hey, Tish. I am so sorry, Michael, going to have to ask on guidance, if that is all right.
Oh, yeah.
Could I just confirm the bottom end essentially assumes minimal GPU pool? The very, very dirty math is you are guiding to network revenues of AUD 320, your base compute ARR is AUD 91. If we assume your new contract is only come in the second half, that is still another AUD 190 million. So AUD 320 million plus AUD 90 million plus AUD 190 million is pretty close to that bottom end revenue guidance, and I have essentially assumed no GPU pool.
We need an AI bot to respond to you as you are throwing all these numbers at us. I think we will. Let's not specifically answer to your numbers because I don't want Tish to do that math, but let's just sort of announce the general piece of your question. Maybe we will announce that piece.
I think it's, Eric, timing is key in this, so I would just build out a broad range of assumptions on that. The timing elements, and this is not just a Megaport, it's for the industry wide, is everything has to align perfectly to then deploy to happen perfectly from a GPU pool standpoint. It's really just around timing, when it's deployed and when it's taken up, and the cost of that and the price that customers are paying at that point.
It's tricky for us to give you guidance within the year specifically, because let's say one month delay makes a material difference to-
Material.
Also the thing that we also can't control is exactly the ramp of a pool, which is why we give this guidance here, or insight here, to say it takes six to nine months to procure and deploy. Then once we've deployed it, there is a three to six -month ramp, which as you can see, if you rolled that through, we announced those in this financial year. Did we? Just at the end of last financial year. If you play that forward, yeah, you are going, I don't know what your exact maths were, but in short, we're not expecting a whole range of that pool to be revenue in this financial year.
I think just look at the range within the guidance. We've built in some timing elements to that, and that's the real driver of that range.
Super helpful. Can I do a really, really quick follow-on, which is, if you look at the difference between the bottom and top end of your guidance range, it's about AUD 50 million of EBITDA. If the GPU pool might be the bulk of the difference, it sort of suggests the bottom and your top end has got three months of different GPU pool utilization, if you like. Sorry, one other follow-up is, FY 2027 is a bit of a weird year because you've got partial contributions from your new contracts, and you've got partial contracts from the GPU. Can I just check, if we hypothetically assumed 12 months for all of your new contracts and all of your GPU pool, I reckon you're run rating above AUD 600 million of EBITDA now on a run rate basis. Could you help with that, please, Tish?
Eric, I would just focus on what guidance we've provided here. What we've provided is really the timing on the range for revenue is driven by timing of strategic contract delivery, and then the EBITDA is a fall through of that. Don't forget, we're also building the business, so we have to keep reinvesting in the new group business. You don't just drop every single dollar to the bottom line. That EBITDA we feel pretty comfortable with, in terms of a range, and it is dependent on revenue.
Yeah, just you can see this breakdown. The network revenue is tight.
Yeah.
There is a broader range for compute revenue because of the sheer difference between how it ramps. The network business is like all this recurring revenue, very, very tight thing that you can predict. It really changes between that retention component. Whereas the compute is really a mixture of when we deploy. That is deploy and handover or in the pool, when we deploy and then adoption. I think we have given some pretty good insight to the business here in covering those ranges so that we are conservative in how we approach both ends of that. We do not want to set the market up for something that is not achievable, and that is what I think is where we see it at the moment.
Remember, there is risk associated to all those components because as we remind you all, each one of these components needs to be delivered in order to turn on the paying component. If any of that delays, which as you know, is a massively supply chain challenged world at the moment across all these elements, it can move it. That is why we go back to, we feel comfortable at this guidance and all these components that break it down. I think we just keep pointing back to that.
Makes sense. Thanks, guys.
Pleasure.
Next up is Tim Plumbe with UBS. Please unmute your line and ask your question.
Hi, guys. Sorry. For me, it is another question around the new contracts that are there. Just apologies if I have missed it, but can you talk about payback period, or the GP margin that you are expecting out of that? If I think about that AUD 435 million of long-term contracts, and you sit there and you say average GP or average incremental margin at 80%, gives you AUD 350 million. Then if you assume reinvestment, every 10% is kind of AUD 35 million. So I appreciate your comments around reinvestment back into the business, but you are starting to talk about some pretty large numbers. So how do we think about what is a reasonable expectation for reinvestment of incremental contracts back into the business, please?
Just reminding you that this is FY 2027 guidance, and so yes, you are saying AUD 435 million of ARR, which we have said here. But that depends on when you deploy that. So if we are talking specifically about the guidance for FY 2027, it goes back to the same statement we.
Not the guidance for FY 2027, just in general, how should we be thinking about reinvestment?
You mean into the future beyond FY 2027 or like what?
Well, just reinvestment within the AUD 435 million of ARR that has been generated to date through the strategic contracts.
Yeah. I think Tish can just quickly hit on that.
Yeah.
I think we have shared this component.
In terms of margins there, Tim, I would not necessarily want in terms of the future for those, the business is changing and evolving. Yes, you get to a quantum where it does not make sense to kind of, you would start to see a high margin, but I would not necessarily be baking that in at this stage. The guidance for FY 2027, Megaport will continue to reinvest where we see opportunity to build innovate, and continue to build out the product set, go to market, and also making sure that the customer support is there as well across the board for these new product sets. I would not necessarily just start baking in huge dropping down to the bottom line, but we will. We will continue to be profitable. This is with a higher CapEx in business, you would expect to see higher margins over time, but-
And-
Give us time to grow.
We've shared that again consistently for the past, I think, year around we accelerate revenue through reinvestment, yes, but we will continue investment with revenue growing faster than costs. We also say, what is the long-term future, FY 2030 and beyond? Very sustainable 20% growth after we've had sort of these injections of incredible growth. We're going to do 100%+ in revenue next year. Then you get to a point where, okay, when that winds down, you're very highly profitable, then you convert all of that, or a huge amount, not all, but converting a huge scale into free cash flow at that point.
That is, we're not going to give forward-looking guidance on those numbers, but that is, you can look through the business and do some math and definitely catch up with the team to talk through your modeling after. Very hard for us to do modeling with all the numbers just on a call. We do not want to give the wrong impression around it. I think definitely spend some time with the detail with the team after.
No, that's useful guys. Thank you. Sorry, just one last, like back to the point in terms of payback period for the new contract. Is there a way for us to think about that?
They're consistent with what we've previously announced, and they meet our internal thresholds.
I think we've stated that in the-
Yeah.
in the announcement.
Okay, great. Thank you very much.
Thanks, Tim.
Your next question comes from the line of Roger Samuel with Jefferies. Please unmute and ask your question.
Oh, hi. Good morning. I have got a question about your funding requirements. Obviously, you have been using a mix of debt and equity to fund future growth. But if I look at your upsized debt facility, plus the cash on the balance sheet, plus the operating cash flow that you generate in FY 2027, it looks like you just have enough to cover the CapEx requirement in FY 2027 of AUD 1.3 billion. I am just wondering what other options you have to fund future contracts in Compute, and whether you are prepared to increase your gearing ratio.
Roger?
Yep.
Roger?
Yeah, this is for-
We did the capital raise right in June. Which is-
Yep.
slightly inconvenient timing. Yes, you have that as at 30 June, but then we have the following retail entitlement offer that landed on 2nd of July . You will see that in the actual annual accounts as well. I know it is-
Yeah-
been a little while and it is reporting, but yeah.
So it's-
Yeah.
missing a big portion-
Yeah.
of that capital raise.
We are comfortable with our funding that we are covered, for what we have announced. Look, we are not necessarily going to go, this is funded exact thing by debt for this contract. That is not how it is. We have got a general corporate facility for those purposes.
Yeah, I suppose my question now is that, are you prepared to increase your gearing ratio, going forward? Maybe to something like 2 x-
You mean from a debt facility? Is that what you are saying?
Debt, yeah.
No, well, we have said-
Oh, yeah. Using more debt as opposed to raising equity.
We are well funded, so we will continue to manage within thresholds. We have only just got the commitment letters at this stage. That is what we have announced today.
We will adhere to the strict governance associated with that, then we will assess strategic deals as we go forward. We have now a very flexible mix of capital that we can access, which is something I think the market was asking us a lot of, was getting this debt facility. We obviously did a capital raise most recently and will continue to assess strategic opportunities based upon what we can service.
Roger, I think the question that I probably hear from that is, we are going to grow into the debt rather than just from day one, so-
Oh, yeah. But at the end-
As we continue to expand.
Yeah. But I suppose at the end of FY 2027, you will, yeah, you have more debt on your balance sheet.
We will currently, we don't have any-
If you were to win more contracts, if you were to win more major contracts, then you probably need to-
That's-
Fund new, yeah.
Depending upon new contracts that we haven't yet shared here, obviously, we would need to-
But-
Constantly look into that.
and we are. This is just step one in, or step feels like 100, but step one in our capital management framework plan. That is what we have been building over the last few months. I think we have shared that a few times since we have come to market in the last few months.
But that debt facility is sizable as well. So it is, you know we have got runway here. Yeah. I am not sure what your math was saying that we are out. So I think maybe catch up with the team after and just go through it, and it might be misleading around the AUD 435 that landed in June 30, but there is a remainder that came from the retail offering that landed after that we cannot report on. So maybe it is worth just sort of ducking through those components.
Okay, thanks.
Your next question comes from the line of Ware Kuo with Bank of America. Please unmute and ask your question.
Morning, team. Just on the compute side, just wondering who you're mostly coming up against in tenders today. The reason why I ask is I think we've seen a number of CDN players, edge infrastructure providers all pivot into the AI compute space. In tenders, are you seeing this bucket increasingly come up, or is it still other new cloud providers? Just wondering, how do you see yourselves positioned versus some of the CDN players that also have a very distributed network globally?
I haven't come across any of the CDN players that I'm aware of. I think the space is just an unbelievable demand scenario. Deals just constructed incredibly fast because of the pressure around pricing. I think if you think through a traditional IT enterprise procuring IT, where they run a six-month process, they run an RFP, and they get pricing, and they spend their time going through it, getting board approval, whatever, and they finally go and deliver upon it. Every two weeks, the pricing changes. The whole world has had to change, and materially. If you map just memory pricing between November last year and today, every few weeks, the pricing changes. Vendors have gone away from long pricing. They'll give you about a two-week window, and it still doesn't even guarantee it, so it can move.
Customers have had to change because imagine you're running a process, you get some budget, you say, "Let's set AUD 1 million." By the time you go through the process, it's now AUD 1.2 million. Your budget's not approved. You go back to get the approval. You go and run the process again, only to find that it's now AUD 1.4 million, and then it just continues in this cycle. What happens is it's significantly faster because of the environment. A, there's no tenders, so to speak. You're dealing with customers very quickly to land what they need when they need it. It's a timing game, and then it's how quickly you can get access to the capital, to the data center space, and the delivery of the chips and get them running operationally. Then that depends on the location that they're looking for.
I would say there are multiple factors, and I would say the pie is expanding so rapidly and so large that there is plenty of pie for anyone and everyone in this space at this point in time and for the foreseeable future. It is less around competition. It is more about what you can actually execute against as a company, and the same for everyone.
Great. Thank you.
Which is why the depth capabilities in these components help us project forward opportunities to continue. That is why it is like all these pieces of that puzzle is really, really important.
Next up is Fraser McLeish with MST. Please unmute your line and ask your question.
Hi. Great. Just a huge year for you. I just want to say congrats on all the progress you've made. My question's really just on the acquisition that you've kind of looking at, that you've called out today. Just wondering why, when you say it's early stage, why you've chosen to call it at the moment, and whether it's in the network or the compute space. You also announced a partnership, I think, on storage recently, whether that gives you everything you need in storage or whether that would still be an area you'd potentially be looking for acquisitions if something comes up. Thanks.
Yeah. I think I'll draw you back to this slide. We're constantly and forever looking to innovate, to add rings, and expand rings, and we can do that via acquisition or via internal build, and we're constantly exploring a huge range of options that would look for that internally. We make decisions on whether we can build this style of product set and how quickly we could deploy that. So in effect, from a storage element, I would say we are comfortable with what we have today for everything, but uncomfortable in that we need to constantly add more and more innovation to it. So we will be always looking to add more rings or even differentiate the rings that we have in there, and it's just this constant process that you'll see from us.
So that's why I think we're sharing the component around these three pieces, and we're constantly exploring options in that space, and we will forever be exploring options in that space. We'll announce more detail if and when any of those come to fruition. By the way-
Just on the-
It will be across each group. This is the network element, this is the compute, and this is the storage, and we will be constantly exploring each one of those elements.
The acquisition you have highlighted in the release today?
Which area?
We are currently looking, I think it would be the network and security space.
Great. Thanks.
We are continuing to build each part, so it is not just a focus on GPUs, for example. It is important to note that this is not just a compute business here. This is a giant network business that we will continue to invest in, a giant CPU business and a GPU business, and then there will be the storage elements that come with it as well.
Great. Thank you.
Your next question comes from Paul Mason with E&P. Please unmute your line and ask your question.
Hey, team. Just one for me on just technology for cooling and stuff. Next year it looks like it is going to be harder to deploy a GPU into an air-cooled environment because I think Vera Rubin is going to require liquid cooling. Just like maybe how you are thinking about the plans for trying to grow Latitude from there. Are you going to have to maybe start procuring data center space on spec a little bit in order to support growth and get liquid cooling built in time, or are you thinking about a longer lead time, or are you thinking about just sticking with B300 for the foreseeable future, which can do air conditioning?
It will be a mix. I think as we have sort of progressed into this space, we have started progressing appropriately for the opportunities that we have had. All those, at the moment, if you look at the market, there is very little liquid cool data centers that have been physically built as you are sort of alluding to. There are a number of reasons for that. They actually need to change the CDUs in there, the chillers, and all these components. The actual data centers themselves, I do not know, whatever it is, 99.1% of data centers are not built for liquid cooling. There is this sort of transformation happening in the industry. NVIDIA is pushing that by releasing more and more liquid cooled components. Yes, if you want to deliver Vera Rubin, you will need liquid cooled elements just because of the sheer heat that they create in that space.
But what we are starting to see is that data centers are catching up. We are seeing that more and more are coming with liquid cooled capabilities. We will constantly procure and deploy upon what our customers are requesting of us, what the market is demanding, and so we will cross all of those different areas. There will be elements of liquid cooled, there will be elements of air cooled, distributed all around the globe, depending upon where the requests are. The answer is yes, to deploying liquid cooled. It is sort of the obvious path. You can also deploy B300 as liquid cooled. You can take a data center that is liquid cooled, deploy all the stack in there. But yes, Vera Rubin specifically requires more. They are quite dense in terms of what they deliver. But you can look at Rubin. You can also look at that component as well.
If you look at it today, GB300 do not sit in an air cooled facilities, but B300 sit in there. They are still B300. One has the Grace Blackwell on 72 in one rack. The other, you can put eight in a server and you can distribute it from an air. We build depending upon what makes sense.
And just a quick follow-up one. It hasn't been mentioned for a while, but could you make some comments on just like the rollout of the new compute platform? I think that was meant to be like a multi-year program of upgrading data centers in order to have the new software in your historical install base. How's that sort of gone?
Sorry, which, the Latitude side? Or are you saying-
No, no. No.
Oh, Megaport. You're saying that-
The Megaport new compute platform.
the MVE platform? I think that's two years ago. We rolled all of that out. I think we have 100 and something plus sites or locations that deploy virtual edge. If anything, that equipment's probably getting old and in the future, we'll look to replace that as we scale that through. In effect, that is the compute platform to support all the network edge, and that's been performing incredibly well. You can see actually insight, well, I don't know if the metrics are in there, but basically that business is continuing to scale. That supports our MVE and our MCR platforms and all the different countries and locations at the edge. Yes, I think we told the market we're investing in delivering that, and that's been done.
Great. Thank you.
That gives us that 400 G connectivity up and then 100 G opportunity to run. It was like a, I don't know, a significant increase in what we can deliver at the edge. Super important.
Oh, thanks.
Your final question comes from the line of Andrew Gillies with Macquarie. Please unmute and ask your question.
G'day, Chub. G'day, Tish. Can you hear me?
Yeah. Gotcha.
Thanks very much. Just one final one, if I may, around the customers for new contracts, just feedback from clients around concerns that potentially the larger side of enterprise is the long-term riskier space for customers. Can you maybe share some of the points that give you confidence that the customers you are signing up in Latitude on these compute projects are quality counterparties? Also underneath that, how you are seeing that take-up of initial testing on pooled capacity, how the rollout is going on the initial compute contracts as well. Are the customers happy? Have they got any feedback?
Yeah. So customers are very happy because they keep asking us for more. We are being disciplined around how we accept more, because in effect, you have got to be very careful you could just sell everything to one customer and you end up with this concentration risk. We also want to make sure we are growing and servicing our customers here, so having these two components play out. So we are already, as you can see in some of the slight amounts of ARR before June 30, but you can start to see that we are deploying a whole range of GPUs across, and CPUs and storage for these contracts as we press out. We are hitting the time frames that we are telling our customers, and I think that the industry is full of folks that are telling people things that they cannot execute against.
One of the important things to remember is that Latitude and Megaport combined, we run 26 data centers plus that we keep adding for compute infrastructure, as well as 1,100 physical data centers all around the globe with automated platforms. That is what we do for a living. When we add all these components in there, I think other folks out there in the market haven't been able to deploy as quickly as they have promised or they have had challenges around how to operate and run that. Remember, we have a huge automation software platform that delivers that.
It is not just as easy as installing a thing and handing the keys over and just saying, "Good luck to you," which I think a lot of entrants in this space have done, which has created a really bad taste in the mouth of many, many customers. So when they see how we deliver and we deploy, it is the good old-fashioned, you tell someone, you commit to doing something and you deliver against it. In this world right now, it is so important, and that is what we are seeing. I think you asked the next question around counterparty risks and so forth. We have a vast range of customers, and I think the important part about adding the pool is to constantly diversify our customer base and then continue to scale with those customers.
We also spend time, great time, looking into the customers and the details that we have. There is a mix. Tish actually looks through financials. We look through, I do all the checks. I think you could probably articulate that component.
We do a mix of credit checks. Obviously, speak to the customer, understand the financials. Then we also just make sure in terms of the contract, it will be a mix of terms, but a mix of upfront payment and then payment in advance.
How they are funded and all those components. Then you do multiple things. You diversify the customer base, and then you actually diversify the risk associated as well. Remember, everything that we deploy is deployed fully automated, which is what is so important. When you deploy the automation platform, you can instantaneous make it available to the market or to a pool in seconds, rather than having to send humans back to go and do an installation or a handover. We have a giant market that services that. If you think of a lot of folks that are entering this space that are really doing a finance model to go and deploy some GPUs without the tech side, they do not have that ability to deliver that.
I think there are multiple elements there, which is why I think showing you that demo that I showed before hopefully gives you the perspective of the diversification of how the business operates and the strength and the power of having an automated software platform across the top. So hard to physically automate, to automate via software, physical infrastructure. It is very difficult.
Thanks, guys.
That brings our Q&A session to a close. I will now hand back to Michael for closing remarks.
Well, team, what an outstanding result. I just want to take the moment to thank our shareholders for continuing with us on this incredible ride. The company has transformed significantly. I will draw you back to the very first slide. How often do you get Well, firstly, there is the first slide. How often do you get to see a front page like that? I think let's focus here. In the last 10 months, the execution of this company has been unparalleled. Look at where we have come from to where we are. It is a monumental change in the business, and we are incredibly excited about taking advantage of what is the hottest market I have ever seen in my history. From an AI perspective, it is out of control, the growth.
The opportunity is long-term, because what you have is natural rate limitings on physical infrastructure, physical data centers, so that actually this will continue for the foreseeable future, and you can look through all the different fabs. At the end of that growth piece, you end up with consistent percentage growth on top, no different to how the cloud market built out. We are in a really unique position because of the platform that we have built globally and the skill set that we have, along with the software infrastructure, and you can already see that. We just announced another AUD 500 million of TCV today, and in that period of time, AUD 1.3 billion from where we were in April, where we had not looked at any of those components, and it was just the on-demand pieces. So important to keep the on-demand continuing and these contracts.
You blend them together, and you have this incredible business for the future. Super excited. I want to spend one second to shout out to the team. The amount of effort and energy from the team that's gone in over those past nine months has been unbelievable. We were firing fast before, but just to look at what this team has achieved is phenomenal, and I want to thank each and every individual. When you run fast, you work long hours. It's tough, all those components, but everyone here leans in, and that's what's so exciting to be in this business. Thank you for all the support, and we're hiring. If you know people, send them our way. It's an exciting place to be. Thanks, team. We'll see you on the roadshow.