Sims Limited (SGM.AX) Earnings Call Transcript & Summary

Perfect. Okay. Thanks. Welcome, everybody to Nashville. I thought the tour yesterday to Atlanta was fantastic. It was great to see you guys there. And I thought the SA Recycling team gave us a really good understanding of SA and its strategy. But we're here in Nashville today. Let me go through a couple of safety things first. The -- if an alarm does sound, the muster point for us is out on that hill over there. At the top of the hill, there's a sign. There is just a notice, we gather there. I would say, again, as of yesterday, there are no planned drills today so that if an alarm does sound, it's genuine. If we're in here, we'll just quietly make our way out of the door at the top. If we're out in the -- on the floor, follow your Sims person and we'll make sure you get to a safe place quickly. Bathrooms are just out of the store here, ladies and gents outside there. I think that's all we need to do from a safety point of view. So today, we're going to have presentation firstly, from Ingrid Sinclair. Ingrid is the President of SLS and reports directly to me; and Lynn Jacobs. And Lynn Jacobs is the Chief Financial and Sustainability Officer here at SLS. On that, I'm going to hand over straight to Ingrid -- actually I will run through. I probably think we'll do it similar to yesterday. So we'll run through the presentations, Q&A. We'll then go out onto the floor, have a good look around. I suspect that may sort of spark some more questions, so we'll come back in here for a bit more Q&A, and then we'll get you on your way. Thanks very much.

Great. Thank you, Stephen. Good morning. So welcome to Nashville. Out of curiosity, who's here for the first time to Nashville? Okay. So some have gone through. Any chance that you were here 2 years ago when we had Investor Day? Anybody? No. Its first time -- well, very pleased to have you and pleased to have the opportunity to present the business to you and give you some more insight into why we're so excited about this. So my team. SLS, as you know, is a wholly owned subsidiary of Sims, which means we have our own executive team. We're based in Irvine, California, which means we get to spend half the year with Stephen when he's over, which is quite nice. But we run independently. So what does that mean? So we have our own HR, our own IT, finance, ops -- actually you know -- and they all report to me. We have global responsibility. We have sites throughout the world in EMEA, APAC and in North America. In the Americas, we have a site in Mexico as well. So we're every -- coverage everywhere. And the reason we have this autonomous team is so we can have very quick decision-making processes, be able to move quickly and agile because that's the business we're in. It's a very quick paced business. The one thing that does go up to group is capital resource allotment. So we all compete for the same capital funds, and that goes up to group. So what is nice about this that if in the future, there are any plans to change ownership structure, it would be very easy to do, right? Very easy to pivot. But -- I'm sorry, I went backwards. So I'm going to introduce a little bit of the team. You've Stephen presented Lynn. So she also has IT under her as well as sustainability and finance. Chris has all the operations globally, and Marie has HR globally as well. So who has heard of AI, right? It is everywhere, right? And it's really the addressable market for SLS. We have at in entertainment, shopping, at work, everywhere. And I know that some of you are integrating AI in your businesses. So you're all very well aware of its power and its expense. I mean it's just -- it's everywhere. And I just pull out some of the headlines that have been out in the press lately. And just the pace of change is incredible, and the amount of investment that's going on. $80 billion, Microsoft; Amazon, $100 billion; Google, $75 billion; Meta up to $65 million, all on AI building capacity. It's just amazing boom and the amount of money that's being spent in this sector. So if you take a -- think of the pace of change, took 75 years for people to adopt the telephone, right, to get to 50 million users. The internet, it took 4 years. ChatGPT had 10 million users in 40 days and 100 million users in 2 months. I mean it's just amazing, the pace of change of AI, it's everywhere. And that's $1 trillion that's coming being spent in this sector. And just this week, NVIDIA announced another $150 billion in partnership with OpenAI. So [indiscernible] looks, we have a couple of more people coming. Sorry, we have a couple more folks joining from back there. So to bring an analogy to you, it's like the gold rush, right? Those who won't dug for gold weren't necessarily making the money, but the ones who are selling, the shovels and the [ pickets ] were. So it's sort of analogous to what we do. We're adjacent. We provide tech services. We're not building AI data centers or working, maintaining it, but we are adjacent. We're providing tech services to them. So as a shareholder of Sims, you probably never thought that by holding Sims' shares, you would have the exposure to the AI sector, but here we are. So what we did, we wanted to kind of show what this is -- the yellow line is the total investment that's going into the AI data center hyperscaler. And we plotted the blue line as SLS revenues. I'd like you to kind of look at the comparison of the slopes. It's quite interesting that there is some direct correlation between the AI sector investment and the performance of SLS. But what you have to note here is that what is installed today will come to us in 4 years because the decommissioning cycle is normally 3 to 5 years, so 4 years on average, which means that here we have a rack and all the servers in it. Every 4 years, it gets refreshed, right. And it -- you might think, okay, but is it slowing down over here where the slope is going? That's still $300 billion -- between $350 billion, $400 billion of investment. So it's still huge. And that doesn't take into account all the recent announcements of the $150 billion that went in with OpenAI and all these things are coming. So I would suspect that, that curve will keep on going up. Also to note here, so since it's a 4-year cycle, it -- what gets installed in 2020, we would see at the end of '23. But throughout that, we're also giving -- providing services -- the life cycle of it. Okay. So because of this, of course, it's a very exciting industry, the growth and it's appealing to others. So what's the competitive landscape look like? We estimate about 95% are local regionally held providers. So they're looking at market sectors that we're not interested in. So it would be your universities, your hospital groups and that type of thing, small regional business. The middle part, the 5% is what Stephen often talks about, the global competitors that we run into, Iron Mountain and SK Tes. Those are our main competitors in that sector. But also, the hyperscalers could be a competitor themselves, right? They either pull it in, do it themselves or they don't do it at all, right? So I'm going to be bold and say that there's only one company that's strategically placed to take advantage of this growing aggressive market, and that's us. So let me kind of break through -- go through why I state that. So we have strategic market positioning. We're aligned with the growth of AI growth, infrastructure. And so we go to where our clients are, where they need the capacity. We're a full service provider, a one-stop shop. So we do from the forward upfront logistics all the way to the decommissioning, all the way through the life cycle. We innovate with our clients. We have strategic partnerships. So we're always coming up with new ideas on how to do things better as the technology changes, we have to change with it. So it's very quick. And we integrate with them. So we get into their data systems and they're in ours. And then we -- scalable capacity is what is very important in this sector because you have to be able to move quickly. And we estimate we have an additional 50% to 60% of extra capacity available in the U.S. because we've been investing in that growth over the last 2 years. We've spent 13 years, the previous CapEx year, over 50% of that was for growth capacity. So it's just being ready and tacked in anticipation of the growth and to position ourselves so we can be ready for it. And as I mentioned before, what is super important in this sector is to be able to keep pace with our clients, it grows so fast. It's very quick moving. So we have to move quickly with it and just to have that scalable capacity. So if I give you a concrete example of what we did successfully to validate what we do and what we can do as far as moving quickly and being able to pivot. In June of '21, we were presented with an opportunity to bring in a new service offering, which is a redeployment of DIMMs. So in 5 months, we stood up a site, and it's not a regular site. This is to hyperscaler standards, which means the floor is [ epoxy's ]. You have the environmental controls, temperature, humidity, antistatic controls, high security throughout the site. It's not just a regular asset management site. It is a higher manufacturing like site, which you'll see in about an hour when we go out on the floor. But we did that in 5 months, which is incredible. And then within a year, we had reprogrammed 1 million DIMMs. So Lynn, put this out, [ the DIMMs ]. So this is when I -- they're memory, memory sticks DIMMs. And I kind of get a little bit technical with you here, there is a group called OCP, which was founded in 2011. It's the open compute project. It was founded by the hyperscalers like Google, Microsoft, Meta. Intel is also in there. And it was -- they came together to try to agree on a generic form of the hardware. So this is OCP hardware, which means it could be used anywhere. You're not with proprietary DIMMs proprietary hard drives, it's all open compute. So they did that so they could be more scalable, efficient, sustainable and cost effective. In 2019, Intel introduced CXL which is Compute Express Link. And this is ideally what these reprogram DIMMs are used for. It's older tech that can go into new tech. And what they do is they build pooling. So it's pooling for memory, storage and compute power because you can imagine, with all these AI with the language learning modules, that you need -- you have these peak periods of compute demand. So CXL is an opportunity. It's a pooling. It's -- so you can meet those peaks, instead of designing for a peak, you can go use the pool. It's a much more cost-effective way. You're not using the new technology. You don't have to have a DDR4. You can 5, you can use the 4, which is right here. So more sustainable, more cost effective and helps data centers meet their challenges with memory constraints, the performance bottlenecks and hardware inoperability. So the old technology can be used in newer technology for pooling effects. And that's the main use for redeployment of DIMMs. And when we go down the floor, you'll meet Isaac Martinez, and he's the expert on this. So if you have any extra questions on it, he's the guy to talk to. So the markets we serve in. I talked a lot about the hyperscalers, the data centers. We have most of them under contract, long-term contracts, 3 to 7 years. Enterprises -- these are global enterprises, so your large banks, your like SAP, Bank of America, all those type of clients and then the OEMs. So the folks that make the equipment, your Ciscos, your Dells, your Oracles, those types. So all into contract. And they all -- in one way or the other, touch the hyperscaler space. And to note here, no single customer contributes more than 20% of our total revenue base. So we try to keep it very diverse. So this is looking inside a data center. I mentioned the 4-year cycle, but we're in all aspects of it from the beginning to the end, not just at the end of the decommissioning the 4 years but we're also upfront. So we do the upfront supply in their supply chain. We have some of our employees that are sitting in data centers now and doing services in a live data center. We do the wiping of hard drives. We do the shredding of hard drives, that is all ongoing throughout the 4-year cycle. So just in a way and to make sure that we're there throughout, not just at the end of the 4-years. What's super important here is to be sticky with the clients. And these are the methodologies that we use to be sticky, right? Sticky, so it's hard to change the high switching costs. So just -- we don't -- we want to make it hard for them to leave us, right? So integrating their data is a big part. We're in their data systems, and they're in ours with the APIs, reporting and so forth. We have our employees sitting in their site. So currently, we have 15 SLS employees sitting in 15 data centers in the U.S. We have some hyperscaler employees that are sitting in our site. So we have clients, hyperscaler clients there -- have their employees sitting in an SLS site. And then we also offer custom programs throughout, be it the backup battery supply in the forward chain or so forth. All of this makes us unique to this market. So I know yesterday, you went through metal yard, scrap yards. So I wanted to bring into focus a little bit of what you saw out on the floor or out on the yard. Yes. So Lynn has my 3 DIMMs. So my 3 DIMS here, new DIMMs are the same revenue value as 1 ton of ferrous scrap. So the size doesn't matter in this case. So with that, I will switch it over to Lynn, and she'll do a deeper dive into financials.

All right. Thank you, Ingrid. Ingrid spoke about how fast pace this industry is. And so I have a tendency to speak at the same rate at which the industry is growing. So I'm going to slow it down. But if I get going too fast, just let me know. So they say the success of a recipe is in the proof is in the pudding, and I'm lucky enough to show you the pudding today. So if we go to our financials, you'll see we are presenting '23 to '25. We've eliminated '21 and '22 as they were COVID years, and so we're going to focus on '23 to '25. Your eyes will automatically be drawn to the underlying EBIT earnings, which have increased 78% from '23 to '25, moving from $8 million to $32 million. How is this possible? Well, this is built on a really robust revenue stream, which outpaces the rate at which costs have increased. So looking at revenue, we started at $325 million and we ended at $427 million and revenue can be looked in different ways. The growth in the revenue here is definitely as a result of the growth in our hyperscaler revenue, which has moved $100 million in the 2-year period. So as hyperscalers have grown, SLS has captured that market and grown with them. If you cast your eyes further down, you will see trading margin or gross margin as we refer to it. That has really increased substantially. And one thing that makes our margins robust and sustainable is the diversification of revenue streams. We have different revenue streams, and we'll get into that in the slides that follow. If you go further down, you'll see our operating costs. They have only increased 15% over the period, which is a lot less than the increase in the revenue. But interesting to note that as a percentage of the overall revenue, our operating costs have fallen 7%. The increase between '24 and '25 is $25 million, of which $17 million is directly related to new revenue streams, which indicate growth. Just like in the metals yard, they have tonnes and they have pounds, we have repurposed units. So these are not a financial metric. It's a metric to indicate market growth. So it's -- sorry, this thing keeps pinging -- anyway. So it's to indicate market growth. So what's really exciting to see here is our repurposed units have doubled. So our revenue is not just price related, it is volume and price, which makes it really robust and sustainable. So I think that is it on the slide. It is really exciting, and now you can see the proof. But I'm going to take you further so you can understand a little bit more about how we generate revenue and how it shows up in our P&L. Right. So SLS has 3 main streams of revenue. We break it down into resale revenue, service fees and commodity. So we'll start with resale. At the end of an asset when we get an assets, for example, a laptop, and we've performed services to it. We've cleaned out the data. We've done data's destruction. We get to resell that asset into the market. This is a revenue share for us, and that line will show up as a cost of sales in our P&L. The second type of revenue is a service revenue. This is an area where we see growing and expanding as we go further and further into the hyperscaler market. Each of these is a service fee per unit. So we have several different types of services. We have decommissioning, we have data destruction. There is a whole range of services. And as we become more and more integrated with the hyperscalers, we're seeing different types of service revenues being generated. And then the last one you'll be pretty familiar with, that's commodity recovery, very similar to metals. So as Ingrid pointed out over there, we have a real life server. And if you have a look at the server, the outside casing is still, that will land up in one of our SA recycling or Sims Limited yards, whichever yard is strategically close to us. And so it goes right back into the Sims family. Every asset comes to an end of its life. So it can no longer be repurposed. And so we actually have to recycle it. So at this point, we have the ability to provide that service, and you'll see a whole lot of hard drives, we will shred. And that shredded hard drive is sold as commodity and so we have some smelter activity and commodity recovery. Pretty sure you're going to ask me what is the percentage of each of these revenue breakouts. So we're going to go to the next slide so that we can go through that. So we've taken the $427 million, which is our revenue, and we've split it into the 2 areas or the 2 ways in which we would look at revenue at SLS. You can look at it by the type of revenue or the client. So if we look at type, you can see a really healthy mix between commodity, resale and services. This makes our margins very robust and very sustainable. If we look at it per client type, you can see these are the 3 types that Ingrid spoke to earlier about, OEMs, enterprise and hyperscalers. Hyperscalers are 47% of our current revenue mix, and that is up from 31% in '23. So you can see the expanding growth in the hyperscaler segment. All right. So strong revenue growth, very much led by the hyperscaler demand. So if you have a look at the pie chart, the dark blue is a representative of the hyperscaler growth. It moved from $100 million to $200 million in 2 years. So really robust growth. An exciting thing about hyperscaler growth is that it touches all 3 types of revenue. So we get service revenue, commodity revenue and we get resale revenue. So as we grow in the hyperscaler space, we will give further strengthening and the diversification of revenue streams. But to take note here is also the really strong enterprise base on which we stand. This is a strong foundation which allows SLS to continue, and it helps us when we have peaks and troughs in decommissioning schedules. So going on to how does it look? So you'll see the 40% revenue growth over the 2 years, but interesting to see or exciting to see is a robust growth in our return on invested capital. So that allows you to see how we scale profitably and also that we have a very controlled and disciplined capital deployment program in SLS, right? If you go to the chart on the left or right, I'm not sure which one now because I'm confused, you'll see we have underlying EBITDA and operating cash flow. There's a strong correlation between the 2, showing that once again that we take high profits and convert them into cash. I'm going to call out the peak in '24. It's -- just because it's really noticeable. In 2024, SLS had a small refinery business, which was called Franklin Park, and was moved under the Sims Metal umbrella as Sims Precious Metals as it was better suited in the metals umbrella than ours. And so that increase just shows the release of the working capital and the inventory in that period. So now you've seen the pudding, I'm going to take -- let Ingrid take your way. Thank you.

Thank you, Lynn. So how are we going to continue taking advantage of this aggressive growth in the sector. We're going to continue being capital light. And I know Stephen likes to say that quite a bit about us. We will continue that way where the last 2 years, we had $13 million last year and the year before, $4 million in capital. So we'll remain that way. Automation, we will automate where it makes sense so we can scale up quickly and again, gain efficiencies. And we'll continue with our geographic expansion. So adding more sites and also growing organically with our clients. So we are excited for the future prospects of SLS. The market is huge, which I showed you, just the investment is incredible that's going on in the sector. We're going to continue with our scalable model to take advantage of this and to continue capturing a meaningful share of the market. So we'll continue taking advantage of the tailwinds that we see in AI in the hyperscaler space. We are uniquely positioned to serve hyperscalers globally. We are where they need us to be. We have diversified revenue across the sectors of resale service and commodity recovery, ultimately at the end when they meet their end of life. We'll remain profitable and take advantage of the accelerated growth and the earnings momentum. We have high returns with our cash conversion supported by capital discipline and effective cost management. And we have proven to be successful in the execution of our strategy. So the hyperscaler relationships that we have and the automation that we've put at scale. So with that, we'll open it up to questions.

Thanks, Ingrid. I think it's the same as you say with the microphone, so that everyone on the call can hear it. So we've got plenty of time for Q&A. We don't have the 4:30 bus problem as yesterday, and so let's get into it.

Listen, just a quick one on the growth in hyperscalers. How should we think about how that might influence margins as that grows faster than the other 2 segments?

Well, the chart -- let me pull up the chart. So I think certainly, the growth is there and we parallel this growth with the margins, with our cost control, efficiency, automation, margin should increase. So we do expect growth, but we're not going to give you any outlook today on '26 or '27.

I think [indiscernible] margin mix.

Do you want the margin mix, sorry?

Margin mix [indiscernible].

Okay. So hang on, you got to slow -- ask the question again.

[indiscernible] a lot of the growth you're saying is coming from hyperscalers as has been the case. But -- and just as that part of your business and that customer grows faster than the others, how should we think about the margin profile? Is it a higher-margin customer?

So I think in the hyperscaler growth, it touches all 3 revenues. So as you grow the hyperscaler, you're actually going to touch all 3 components, but we do see growth in the service revenue as we're going to perform more and more services within that service revenue line item. So that component will definitely increase.

The margin -- do you make more margin out of service revenue than the others or...

It's really sometimes at the customer's discretion at what they want to do. So that when the asset comes into us, they have the ability to decide, do they want us to redeploy it, resell it. So it's a bit of a fluid conversation that I don't think it's really specific.

But I think it's probably -- I mean, I think from my perspective, when I think about it, I've got a mic on something. When I think about it, the hyperscalers is where our good strong margins are. So I mean, I think to answer the question very specifically, we would expect increased margins from hyperscaler growth, not decreased margins.

Well, and if we can point you to, we use TrendForce, which is -- you can subscribe to it, and it does forecasting on DIMM prices, in particular, memory. And $12 -- it was $12 4 months ago, and now it's selling -- a used DIMM is selling at $64. So this is due to the demand. There's not enough manufacturer new DIMMs coming out and folks since they're using CXL, they can use the older technology, which is more cost effective, but still with that, there's a huge demand in that area in the DDR4 sector. So we do see margin increase definitely in the DIMM space.

I 100% agree with that.

And that's information you can get publicly by going to TrendForce.

I think that -- I 100% agree, Ingrid. One other comment I'd add to that is why is the hyperscale? Why are we focusing on it? One is margin. I think Lynn correctly pointed out that its margins across all of what we offer, but it's volumes as well. So the margins in trading margin percentages are good, but it's volumes. I mean that's where the growth is. That's why we're focused SLS on the hyperscaler market for the last 4 years. That's where the growth has come from.

Perhaps just to add some color to the perspective, Stephen, could you, Ingrid, help us understand how you scale. So is it a people? Is it a robotic capacity? Is it land and buildings? And then sort of how does your costs -- how do your cost flex as you scale?

Right. So I had mentioned 50% to 60% capacity in the U.S. that we have available now and that is partially adding a third shift, so that would be people, but also automating where we can. So where we can automate, it has to be a uniform process, and you'll see that when we go out on the floor. Automating is attractive only when you have uniform process through it. So that also helps with the scale. So you can -- decommissionings, they come in big chunks. So what we have to be able to do is scale up and scale down. But to meet that, the differences, we don't want to be adding people, removing people so where the automation helps us is that it can run 24 hours, doesn't take breaks. And then if we have a slow, we turn it down.

I think, Ingrid, I think it's also fair to say on that. I understand, but I think it's also fair to say that as we're growing, and this is certainly what I've noticed, as we're growing, we're getting a more diversified customer base. So your peaks and troughs tend to smooth out because one hyperscale is not decommissioning exactly the same time as another hyperscaler. And so I think that's certainly what I've observed with you guys over the last few years.

And we do have -- we talk a lot with -- we're very ingrained with our hyperscalers. So we're always in conversation. They know their decommissioning schedules that are coming out. They know where they're coming. So we try to place ourselves where we need to be. that might be adding a site. So for instance, we'll be expanding into Europe for one of our hyperscaler partners. And Stephen will probably announce it at AGM or at the half year where we're going. But that is a replication of what we're doing here, and that's to meet their capacity. So that's adding a site and adding people.

I was actually curious on that particular point. How close do you need to be? I presume there's a very active freight market in some of these components and you don't necessarily need to be very close? I am just curious on that element proximity?

To close to the source, do you mean?

To the data center source.

Yes. When we try to be close, honestly, to help with sustainability efforts with timing, and just the movement of material, we tend to go where they need us to be. Geographically, yes.

I've just got a few questions around that chart that you've got there on the screen there. The first one is just that investment profile you've got there, just to confirm that's U.S. only? Or is that global?

So it's global.

In U.S. dollars, billions of dollars. Yes.

But a global footprint?

Correct. Yes.

Okay. And the second question is...

And then our line is in Aussie dollars.

Sure. Yes. Can I get understanding...

Million, that's $1 billion, a difference.

In terms of whose forecasts are those going forward? Are they yours? Or are they third party or?

No [indiscernible] available. The yellow line is publicly available. And the blue is us actual -- SLS actual.

Yes. But the forecasts, are they your forecast? So have you aggregated -- you put those together based on publicly announced information?

Correct.

I guess. So your forecast is not a third party that you've brought in to do that.

Correct. Yes, AI-enabled.

Okay. And just a final question on that yellow line. I mean that's data center investment, but I assume that includes the building of shares, construction costs. Are you able to give a feel for what that yellow line would look like only for the racking equipment and the actual chipsets and the equipment that you would be recycling? Because I'm just wanting to take away construction cost fluctuations and things in those numbers to get a better feel for what your addressable market actually is in terms of that investment profile?

We haven't split out the capital cost from the -- it's data centers themselves don't particularly provide that guidance to us. So this is -- was merely just put up to show the growth in there and the scale.

I'm going to give -- I think I understand your question there, Owen. So let me give you my perspective on that question. A significant proportion of it, though, is the rent costs, which we -- is our addressable market. A good example, I think we're going to see one out there today is there is -- of these new racks that are going is...

GPUs. The GPUs.

The GPU is a good word. One GPU, one rack holds 30 of these things and they are $50,000 each. So that's $1.5 million. Is it roughly the size of the [indiscernible] in that size over there, this new equipment is $1.5 million per rack, not that one, but at that size. So a massive proportion of it is our addressable market. I mean, we could maybe go away and we could...

To the next slide.

Actually, least to be honest, we -- I don't know what the construction cost of the warehouse is or the air conditioning or the electrical. But what we do know is the investment they're putting in the rack is huge.

A couple of questions, if I may. Just on the pie charts here. Just on hyperscalers themselves, their demand and their growth, is it more -- you mentioned AI a few times. Is it more looking forward the growth in training facilities or inferencing at the replacement of racks within the cloud sort of...

Right. Yes. So yes, AI is driving all the refresh cycles and go faster because the technology has to keep up. So as Stephen mentioned, with the GPUs, that's going to be the new technology that's going in. So you have that compute power to go forth. Does that answer your question?

Sort of just a different data center types. So if you've got a big AI training facility getting upgraded or a new build, is that where you're seeing the opportunity with the growth in the hyperscale going forward? Or is it more in the existing cloud infrastructure?

Its both. If you look at today, the capacity in data centers is about 82 gigawatts, and that's going to go to over 200 by 2030. So it means that they're going to build total new data centers in order for this AI computing capacity, but they also take the existing data centers that they have and refresh those and then [ recap ] them. So it's definitely, I think, refreshing what they have, but there's building out new for their growth. So we get to catch both tailwinds.

And then why SLS? If you're a big firm, no names specifically, but why would and why do they choose you over Iron Mountain or someone else?

Because we are publicly traded. We have big Sims that are helping us get some capital allotment, but mainly why? Because we can move quickly, we can move at pace. So what I showed you with our concrete example is because we were able to put a site up in 5 months. So we move very quickly. This space, you have to move fast, and that's really one of the reasons why. And then we integrate with our clients. It's the continuous innovation. Technology is changing so fast. So it's staying lockstep and then what's coming next, what can we do next to increase sustainability, to use the tech again because you don't need to buy the newest, as I mentioned, DDR5, we can use the 4s, which is much more economical, sustainable.

Can I add? We hear from our customers also, it's the level of service that we give and the ability to meet our SLAs and that we consistently deliver. And then there's a whole lot of ancillary services that we provide. So we may help them with their sustainability reporting. We have, as Ingrid said, really automated interfaces, which allow them to pull data, which makes their tracking of their data just much easier. So the fact that we can partner with them and then just worth working with such a strategic relationship we actually are able to put ourselves in the front end. So when there are new services coming up, we're always on the front end of it. So we're definitely, I would say, a bold statement, are the first movers in a whole lot of areas within the space. Yes, It's attractive too.

If I can add. I had mentioned where we're sitting in their data centers. So that gives us also the ability to hear what's going on, what's coming next, what their problem areas are, how can we service it. So really, I would say it has a lot to do with our ability to customize and serve quickly. It's a tech service.

And just one more question, if I may. Just the location we're here in Nashville. There's a lot of data centers in Virginia. In terms of your footprint, could you touch on your footprint? And then why here in Nashville?

Why here? So [indiscernible]. No, originally what we started here with our electronics recycling just over the road because it's 65 and 24. You had UPS up in Louisville, FedEx in Memphis, this logistics location. But as far as some of our clients do their kitting here, so there's [ Qantas ] close by. There's a lot of tech that is starting to come here. So yes, it's not Virginia, where all that you have your kind of all your data centers. But a lot of them are here in this area.

Other locations, Ingrid?

Other locations. So yes, so Roseville, California servicing that area the West Coast. Chicago, kind of we're getting a lot of -- picking up a lot of data center work there as well.

Atlanta now. So are we in [indiscernible].

Yes, Atlanta is our...

Little bit in Tucson.

And Europe is on its way.

Mexico. Mexico, we're just outside of Amsterdam. We're in Eindhoven. We're in Gustavsburg outside of Frankfurt. We're in the U.K., Manchester and Slough, which is right in the data center activity just outside of London.

It's probably fair to say, though, Ingrid, when you and I talk about it, the U.S., I mean, we've got some announcements. We're doing some things in Europe, which are going to be very interesting. But the U.S. is our massive market. And you should expect to see, you should expect to see growth in the U.S. as well. I mean, we'll be opening up new sites in the coming years. But it's -- again, I guess I have an interesting perspective of it. I guess every now and then I go to see these companies that we service, and I agree that it's entirely around service levels is what drives it. Location is what's going to drive it as well. We will be -- we will locate where we need to be. It's still capital light. So I would -- if I would you, I would expect to see significant further growth in both the US and some interesting things in Europe.

[indiscernible].

Well, Virginia, where do we service -- do we service Virginia from anywhere at the moment. So these are great growth opportunity. I mean right now, Owen, I mean, we've got a -- I think we've planned our growth well. If we do need to be in Virginia, we'll be in Virginia. We've stood up -- we stood up a new site in 5 months. It's an expertise we all have now. We'll be wherever the growth needs us to be.

And sorry, if I can also add to that. The new hyperscaler builds that are coming out, you're going to see them for instance, in the primer pack, I said to you, there was a little map of Meadows new site that's going in Northeast Louisiana, the middle of nowhere, right? It's getting the cheap land and they're bringing the power. So you're going to see new builds that are going to be -- it used to be -- you would go over be close to the people. Now it's where you can get the cheap land and they're bringing their own power, be it nuclear power or whatever. So yes. It's going to be diverse.

Can I ask a follow-up question to that? Because for all the reasons for why this business is good for Sims now, capital light, return on invested capital is high, high-growth industry. It seems like we're still fairly early in the piece. But of the reasons why it's good for Sims now? Is there a reason for this business being disrupted by someone else, whether it is the customer or your clients saying, hey, look, there's actually a lot of value on this, so please can I have some of it back or someone else coming in and offering a similar service? Like how do you plan or how do you keep ahead of the competition and make sure that you don't get disrupted by someone else?

Yes. Really, it is integrating with them, getting sticky. So making sure that our systems are integrated. I mean if you're in each other's system, it's really hard to unpick that, right? And it's being there and providing the capacity as they need it. But as far as taking it in-house, there's one in particular that did that and found it very challenging because it's not their core business, and they're taking away valuable real estate from their data center to do decommissioning, which is very lumpy, right? And the reason it works for us is because we have multiple hyperscalers, so the decommissioning -- the schedules will space out.

And a healthy enterprise [indiscernible].

That's right. And we have the enterprise foundation that [indiscernible].

It's interesting question, and I often get asked this. What I think worries us the most is it's -- to be frank, it's not so much the external competitors because I think Ingrid's indeed right. With the service you offer once your systems get integrated, and you're really offering that really safe and secure way of doing this redeployment, that is hard to just rip out and put to someone else. That is hard to just rip out and put to someone else. That's -- that you really are embedded in the customer at that point. I think what worries me the most is either the hyperscale is doing it themselves or not doing it at all. I think those are our 2 biggest competitors, and we've talked a lot about that. I guess our argument would be on hyperscale is doing it themselves is it's not really core to their activity. I think Ingrid is indeed right. They want to and should be focusing on very much the front office getting new data centers up to deploy all this AI that's going to be required. And their back office is our front office. And I always find that's a good model. When someone else's back office, which for them is decommissioning is someone else's front office, which is us. You always do it better when it's your front office because you're focused on service, you're focused on security, you're focused on that full-scale delivery. I think that argument is compelling. I think not doing at all as a risk. We are if they do locate in the middle of nowhere, how can you effectively decommission? I think that is definitely a risk. But I would say, for us, for the foreseeable future, the addressable market in places where the logistics work is huge. I mean if we can only capture that market, the growth is going to be enormous.

And can you help me understand when you talk about systems integration -- so you're going to have to pardon my ignorance here. But what are -- does it actually mean? That mean practically like do you plug into their CapEx plans? Or like how does that all work?

So some of it could be inventory. So they can see -- we can see, well, decommissioning schedules, is what we can also see into their systems. And for -- we do BBU for one of the clients, which means we store backup batteries, bring them in and kit it out once they need it and send it to their data centers. So we hold inventory. We're in their inventory system on that. And then is the information going back and forth in that way.

And presumably that's important because they want to know that they're getting their DIMMs on time at the workplace every time because for it not to be as a disaster for them. And I mean -- and Ingrid said it right, once our system is integrated like that. Is it really -- so we're in their inventory, they are looking at us and we understand where parts are. When we go out and -- you'll see it's a massive logistics operation to make sure that the DIMMs are in the right place at the right service level, whatever you want to choose, it would be very risky to pull apart and give that to someone else. So you're really going to take on that risk of not having those DIMMs where you need them or whatever particular service we're providing.

And every component is serial coated. And so those serial codes, they track it from the time it enters into our door to the time it's finished whatever its purpose is. That information between the 2 systems, they actually use it to correlate their systems and to work out their schedules. So it really is integrated. There's so much information between the 2. It's -- I mean, I came from it also, there was a lot less detail or a lot less trackable information. Everything is tracked here. So that integration is key to them.

So each of those DIMMs, you can track to which server blade it came out of and which rack. So that's a system that we hold for them. And then they can follow that out when it goes back to their data center, just full circle.

So just another one for me. Again, on that chart there. Just looking at that revenue lag by 4 years. Just wanted to get a sense as to what drives that 4-year delay. Is it something that you've just estimated based on the volume flows and general interactions? Or is it basically back down to this whole inventory integration where you know that the customers are wanting to, I guess, refresh those facilities. And the reason I ask that is because conversations I've had with data center operators in Australia are very different with regards to that refresh profile. So I just want to understand, is that a very U.S.-centric dynamic?

Yes. So normally, refresh are 3 to 5 years. And certainly, during COVID, we saw it get pushed more to 7, right? We just do the conditions in the market and all that was going on. But what we're seeing now is the refresh cycles even going faster due to AI, right? It's the tech change that they have to do. So it's the recircling going faster, so honestly getting closer to 3 years. But we used 4 as kind of in the middle average.

No more questions?

Right on time to go for our tour.

[ And a ] plan now. We're going to get on safety gear. The safety gear is available for everybody.

So we have safety vest, safety glasses. If everybody has closed shoes, we're good. So we'll walk down to the other site and through -- we have to go through a metal detector. So you can't take any pictures out on the floor, so I would leave your cell phones here and what have you, you can leave -- everything is safe to leave here. So we'll go over there, take the tour, then we'll come back and have more Q&A. We have goody bags for you to take away with a little bit of Nashville stuff to bring back to Australia. There we go.

Thank you for your [indiscernible].

And honestly, if you have some further technical questions, Isaac is the man to speak to, certainly on CXL. He's very, very knowledgeable on that.