SMSMY (SMSMY) Q4 2026 Earnings Report, Transcript and Summary

Thank you for standing by, and welcome to the Sims Limited FY '26 Results Call. [Operator Instructions] Today's presentation has been lodged with the ASX along with the results release. It may contain forward-looking statements, including statements about financial conditions, results of operations, earnings outlook and prospects for Sims Limited. These forward-looking statements are subject to assumptions and uncertainties. Actual results may differ materially from those experienced or implied by these forward-looking statements. Those risk factors can also be found on the company's website, www.simsltd.com. As a reminder, Sims Limited is domiciled in Australia, and all references to currency are in Australian dollars, unless otherwise noted. I would now like to hand the conference over to Stephen Mikkelsen, Group CEO and Managing Director of Sims Limited. Please go ahead.

Thank you, and good morning from Sydney. Presenting with me today is Warrick Ranson, our CFO. Rob Thompson, our Global President for the Metal business, is also in the room, and I'm sure we will hear from Rob during the Q&A session. We will follow our normal run-through with me providing an initial overview of the results and market conditions. Warrick will take us through the numbers in more detail, and then I will return to talk about our strategic position and outlook. That should leave us plenty of time for questions. I will turn straight to Slide 5, which looks at our strategy and strategic priorities. The fundamentals of our strategy have not changed to repurpose and recycle. We have, however, updated the left-hand part of this slide to reflect the significant growth and importance of SLS to our overall business. A couple of examples. Under customers, we now specifically call out deepen hyperscale partnerships. And under the heading of Innovative & Agile, we now include the importance of R&D for SLS to capitalize on opportunities from the next generation of technologies. Consistently delivering on this strategy over the last 3 years has underpinned the EBITDA performance shown on this slide. All divisions, with the exception of ANZ, which has been impacted by factors external to its control, are up significantly on last year, and I'm very pleased to report a return on invested capital above our cost of capital. Turning to Slide 6. Another very good year of safety performance with our total recordable injury frequency rate at 1.09. These results could not be achieved without the buy-in from our employees, and you can see this in the excellent completion rates of our lead indicator safety metrics. A couple of other points to note on this slide. Firstly, it can be easy to forget the significant contribution we make to reducing greenhouse gases. As the global economy transitions toward net zero, circular solutions such as recycling are becoming increasingly important. By recycling 5.9 million tonnes of ferrous, we helped avoid more than 9.5 million tonnes of greenhouse gas emissions measured in CO2 equivalent. That is approximately 1.6x our combined Scope 1, 2 and 3 emissions. The final point I'll highlight is that our pay equity gap percentage has fallen to under 2%. 5 years ago, we had a pay equity gap of over 9%. So it's very pleasing to see the progress. Moving briefly on to Slide 7 as the detail behind this slide is covered later on. In our metal business, we grew unprocessed scrap by 2 percentage points, which helped lift our shredder utilization by 5 percentage points. SLS had significant growth across the entire business, including a 91% growth in repurposed units. Warrick is going to cover financial performance in detail, so I am only going to make a couple of overall comments on Slide 8. This has been a very good year for Sims. Underlying EBIT is up nearly 170% to $468 million. And this has fed through to our return on invested capital reaching 11.7%, which is comfortably above our cost of capital. I know return on capital has been a focus for our shareholders for a number of years, so it is pleasing to see our strategy delivering. This significant lift in performance has been delivered through positioning SLS as a world leader in data center refresh and decommissioning services, focusing on buying unprocessed scrap and upgrading it to more valuable products through excellent operating capabilities, selling into the best available markets and managing costs. The next four slides look at the state of the markets in which we operate, starting with the memory market on Slide 9. These two very simple charts tell the story. Firstly, on the left, you can see the decline in new production of DDR4s. At the same time, you can see the rapid rise in DDR5 production as manufacturers have switched from DDR4 to higher-margin DDR5 chips in an effort to meet the insatiable appetite for memory driven by AI. This has resulted in two very positive implications for SLS. Firstly, as DDR4 production has declined, the price of new DDR4s has increased significantly, which is clearly shown in the chart on the right. The demand for DDR4 hasn't gone away. So the requirement for used DDR4s has also increased dramatically, driving those prices up as well. Secondly, the enormous amount of installed DDR5s will need to be refreshed over the coming years and also repurposed, either back in the data centers or sold into the used market. This provides many years of repurposing opportunities for us, and we are well positioned to capture those opportunities. While the medium and long-term implications of massive data center expansions are clearly positive for SLS, we are seeing a short-term variability in the refresh cycles, and this is presented on Slide 10. Firstly, on the right-hand side, there is a bar chart showing the annual investment in data centers. Importantly, around 60% of total spend is on compute infrastructure, including servers, chips, networking equipment and related technology. Right now, the infrastructure capacity to build these data centers is under significant pressure. There are a few bottlenecks causing delays, with a significant one being the availability of both the electrical connection and then the electricity itself. This, in turn, has caused variability in when existing data centers are ready to be refreshed, particularly when they also need electrical capacity upgrades. It is important to note that this impacts timing of refresh cycles, not whether the refresh will actually happen. Moving on to the metal markets and firstly, non-ferrous on Slide 11. The structural drivers have been strong for 2 years now, and our view is that these drivers will remain. Just to highlight two or three of those drivers. AI infrastructure requires copper and aluminum, driving those commodity prices up, which also benefits silver prices. The electrification of the world continues at pace, requiring conductive materials such as copper and aluminum. Looking at the chart on the right, you can see the growing gap between U.S. aluminum prices and the LME, largely driven by U.S. tariffs. We are nicely exposed to these higher prices through our U.S. metal businesses. And it is also worth noting that on a number of occasions due to our market expertise, we have sold non-ferrous from ANZ into the U.S. and have been able to absorb the tariff. Turning now to ferrous on Slide 12. The headline captures the market dynamics. In the U.S., we have tariffs protecting the steel industry and encouraging the reshoring of manufacturing. This has lifted margins in steel manufacturing and is driving investment in EAFs. Data center construction is also boosting demand for steel. Both NAM and SAR are well positioned for this market structure and the increased demand for ferrous material. ANZ continues to be impacted by Chinese exports, although there has been an overall modest improvement. At a very local level, the shuttering of Whyalla has increased domestic demand for ferrous scrap. I'll hand over to Warrick now to take us through the financials.

Good morning, everyone. So as Stephen mentioned, global scrap markets for the 2026 financial year reflected a number of dynamics. While the ongoing shift toward electric arc furnaces fueled steady buying in a number of regions, broader commercial construction activity remains soft outside of data center development. Regional restrictions and stricter trade controls, tightened cross-border supply chains, however, elevated Chinese steel exports despite some production rationalization, continued to dampen Asian and Middle East buying, with demand from Turkiye remaining soft as buyers switched to cheap Chinese and Russian billet late in the year. At the same time, we saw copper prices surge, driven by a relentless demand for AI-related activities, green energy grids and EV infrastructure, and hit record highs in the year. Tight primary ore supplies and tariff expectations further amplified bidding for secondary copper. Similarly, geopolitical conflict in the Middle East created primary aluminum supply crunches, pushing global buyers towards aluminum scrap substitution and boosting values despite regional trade friction. Zorba pricing peaked across May and June as a result, adding significantly to our overall financial performance for the year. As we noted at the half, with both export and domestic markets exposed to global scrap dynamics, we've continued to leverage the arbitrage in our key domestic and international markets and sold volume proactively between the two to maximize margins, again, reflecting the significant agility and flexibility embedded within both our inbound and outbound logistic chains. Concurrently, our total repurposed units handled this year was nearly double the prior year's volume. Prices for new DDR4 memory continued to increase exponentially, with our market reference price finishing the year over 1,000% above the prior year as demand continued to increase against diminished supply with manufacturing shortfalls and a focus on new generation cards continuing to uplift repurposing and resale activities. Across the business, we continue to deliver disciplined cost efficiency initiatives. Current activities such as moving to a global shared services platform and the operational changes now implemented for our Houston operations will continue to drive cost and performance improvements in the business. Our average metal fixed cost per intake tonne fell as we capitalize further on existing infrastructure and improved material flows. I'll come back and talk further about our cost performance shortly. Our statutory result reflects those targeted restructuring initiatives and a slightly lower number than what we had at the half for the write-down of the U.K. metal receivable. We've continued to pursue partial recovery options there where they exist, recouping around $17 million over the last 6 months. Pleasingly, I think we've just about stabilized our statutory to underlying position now and expect to see some consistency in this going forward. Speaking of underlying and moving to Slide 15. I've touched on the principal drivers of most of these already. June was a particularly strong month, surprising us on the upside, and we were able to move additional volume at attractive spreads. Importantly, that outcome reflected not only favorable market conditions, but also the capability we have built to respond quickly, manage logistics effectively and place material into the highest value channel available at the time. While market conditions clearly provided support in a number of areas, the more important point for us is that the business is demonstrating a stronger structural earnings base. Lower unit costs, better network utilization, greater market optionality and more disciplined capital allocation leave us better positioned to capture upside in favorable markets while maintaining resilience through commodity cycles. Focusing in on the individual businesses then and strong performances by both the NAM and SAR businesses absorbed the impact of the continuing market pressures on ANZ. Global trade reverted to its previous levels as broker tonnage reduced following the winding up of Unimetals in the U.K. This year's result effectively represents the cost base of our trading activities to the business. June itself was an exceptional month for the metal business, surpassing initial expectations from early in the month as ferrous margins strengthened from favorable market dynamics and non-ferrous volumes and Zorba pricing maintained their highs. Similarly, secondhand memory pricing achieved its highest level in the year on a gigabyte basis, albeit on a lower ratio to new prices given the mix. As Stephen has mentioned, we see some variability in inbound volumes as data center construction and decommissioning pipelines are consistently challenged by a range of external factors. However, we have deliberately built a flexible operating model, allowing us to adjust cost and activity levels with the inbound volumes. I'll expand on some of the other factors driving these various movements in subsequent slides. Moving to the metal business more specifically. And in North America, total intake volumes increased by 240,000 tonnes over the prior year as we again prioritized unprocessed material, increased shredder utilization and improved margins. Intake volumes were also supported by stronger domestic steel demand and higher domestic ferrous prices. Even though we increased the level of domestic shipments in the U.S., we continue to maintain full optionality of material placement for best value. While intake levels also added to comparative costs, the team were able to generate a number of offsets through further restructuring and productivity initiatives. Having TCT in Houston is also now giving us the opportunity to better manage spreads in that region and lower the run rate cost base further. In ANZ, ferrous margins were, again, impacted by the subdued international market, which also flowed on to domestic pricing, although we did see some demand benefit from that prolonged outage at Whyalla. Favorable non-ferrous prices provided overall revenue support and helped offset shredder downtime at our St Marys operation in the first quarter. Notwithstanding elevated consumable input costs, particularly in the areas of fuel and waste disposal, which we felt right across the business, net operating costs continue to be well controlled here with most of the increase over the prior year related to trading currency losses, which, for accounting purposes, are classified into operating costs. Non-ferrous and, particularly, Zorba pricing provided our SAR joint venture with a significantly elevated financial performance versus early June expectations. While ferrous intake reflected a record year following further small-scale acquisitions, the U.S. tariff war and the surging Zorba price ran through to the bottom line, enabling that business to close out the year extremely well. Our global trading platform was also able to keep its costs relatively flat. They saw reduced broker revenue following the cessation of trading activities for Unimetals in the U.K. early in the year, as I mentioned. Moving to SLS now, and Stephen has covered several of the drivers here already. As we've noted, the business has experienced significant growth in the number of repurposed units, demonstrating the broader strength of the market as well as specifically benefiting from the dynamics of memory chip prices, with memory averaging around 30% of hyperscaler spend. We saw that pick up even further in the second half as the impact of uplifted prices filtered through and repurposed volumes increased despite the industry's growing pains and planning volatility. Total memory sold on a gigabyte basis fell from prior year levels as DDR3 volumes reduced and we repurposed more 16 gigabyte cards in the second half. Improved unit costs reflected both volume gains and expansion activities, and the team continues to look at additional opportunities around automation and robotics to support its cost management program. On Slide 18, I want to quickly touch on the ongoing strength of the SLS business for us. While memory pricing has certainly been a primary contributor this year, the business is evolving into much more than that. The structural shift in demand that we are seeing with both hyperscale and enterprise clients in response to this phase of what is effectively the fourth industrial revolution is being matched by both the current need to source an array of components for growth, but also their recognition of the associated circular and economic benefits. Our deep relationships and proven scalability to respond to this demand in a secure, trusted and certified manner provides us with confidence about the role that SLS can play in our earnings base going forward. Touching briefly on central and functional costs now. We continue to look for cost-out efforts in this area. This year, we relocated our corporate office to further reduce costs as well as beginning the transition to a new global shared services hub as part of a more extensive shared services model being progressed over the next few years. Following stabilization of the company's SAP platform implementation, project costs fell by nearly $5 million, noting that we continue to incur costs in developing our new yard management software for metal, which we are aiming to commence the rollout of in Q2 this year. All of these initiatives are expected to contribute to lowering the ongoing cost base and improve consistency of execution. As previously advised, we elected to cease work on the development and commercialization of the plasma-assisted gasification technology that was being undertaken by Sims Resource Renewal during last year. This further reduced the central cost pool by some $10 million to $12 million per year on a full year basis. Just a heads up that in this area, commencing in the current financial year, we intend to allocate costs for centrally provided services that are currently unallocated out to the business in order to provide a more comprehensive and focused approach to their management. This will, of course, result in changes in the comparative performance for the business segments, and we'll provide additional color across this area as we approach the results for the half in the new year. At a group level, we are, once again, able to keep total costs relatively flat over the period, limiting the increase to around 5% before variable costs and off that rebased comparative prior year. Waste management costs continue to be a major contributor to our cost uplift each year, and we are progressing a number of targeted initiatives at extracting the residual metal in this waste and how we reduce volumes to landfill into the future. Variable operating costs increased in line with the increased volume of unprocessed material and higher repurposed units at SLS. We also experienced higher fuel costs as a result of those Middle East tensions. Labor, of course, remains our largest cost element at around 50% of operating costs and ongoing labor cost efficiency initiatives continue to provide significant benefits in this area and in line with our previous cost-out commitments. While we remain focused on all cost opportunities, we maintain the view that our best way to drive further efficiencies in the business is through volume productivity gains and infilling our existing network. We progressed some initial opportunities in this area over the last 6 months in both ANZ and NAM, and expect to progress additional opportunities in this area during FY '27, further improving returns from assets already in the portfolio. Capital expenditure was significantly higher in the second half as we completed a number of planned initiatives across the business. Redevelopment of the Pinkenba site in Queensland continued, with activities focused on site infrastructure and an extension of the wharf. We also progressed new fines and metal recovery plants across ANZ, including at Pinkenba and in Auckland, and expect to see the benefits of this flow through to the ANZ result in the current year following commissioning. We also completed our dredging program at Claremont at the beginning of the year as well as several other productivity initiatives at that site. Other growth and productivity projects include extensions to rail capacity and network efficiency, together with small yard infill opportunities in both the East and West United States as well as in Australia to ensure we get more out of the network we already own. And in February this year, outside of those smaller organic growth opportunities, we acquired the operations of Tri Coastal Trading in Houston to better position ourselves in that market. Total group depreciation and amortization, inclusive of leased assets, is currently forecast to be around $260 million in FY '27, consistent with the current year. The group completed the year with net book assets of $2.7 billion at balance date, reflecting a stronger comparative Australian dollar at period end, dividend payments and removing the Unimetals receivable. We recorded some $130 million in foreign currency translation differences this year from the stronger dollar, reducing our reported net asset backing in Australian dollar terms. Of note, this includes a $200 million uplift from non-ferrous prices, impacting both our inventory and receivable values. Despite this increase, we were able to retain overall trade working capital at a comparative level to the prior year. And following stabilization of copper pricing at its high levels, reduced broker deposits related to our derivative trading activities over what we had reported at the half. Intangibles uplifted by $64 million, principally because of the favorable infrastructure services contract associated with the TCT acquisition, and this will be amortized over the life of that contract. Post the sale of our Houston properties, we expect gearing levels to revert to be more in line with our target range, and we remain deliberate in focusing our growth activities to where we see efficient through-the-cycle returns while protecting balance sheet flexibility. Pleasingly, our strong earnings and capital discipline uplifted the group's ROIC to 11.7% and together with our positive free cash flow performance, the Board has determined a final dividend of $0.20 per share fully franked and payable in October. This brings the total full year dividend for 2026 to $0.34 per share, but noting that the availability of future franking credits will become limited going forward as our earnings base becomes more U.S.-centric. So a little bit more on our working capital movement and the group's focus. Here, we've again isolated some of the movements to show the impact of those higher non-ferrous prices on the business, which continue to be quite significant. Following a relative stabilization in the copper price since the September run-up, we've been able to reduce the amount of restricted cash sitting in margin deposits at June, which, if you recall, was some $95 million at the half. While our total physical year-end metal inventory increased over prior year levels, we continue to align inventory holdings with scheduled sales and are focused on our conversion of receivables and the management of payables to match cash movements, keeping our overall working capital levels steady. All that summarizes into our overall cash movement for the last 12 months. I've talked about most of these already. We converted over 70% of our EBITDA performance to operating cash and invested some $488 million back into the business through capital and acquisitions. Funding for the purchase of Tri Coastal is still expected to be covered by the sale of our Houston properties. The Mayo Shell property remains under contract as the preferred purchaser completes its due diligence and concludes legal requirements. This is now likely to be a Q2 transaction for us. In addition, we have recently signed a letter of intent to sell our two other Houston properties, subject to due diligence. They are targeted to close early in Q3. In October, we made our final FY '25 dividend payment of $25 million and a further $27 million for the FY '26 interim in March. As previously noted, the Board has also determined a final dividend of $0.20 per share fully franked for 2026, in line with our capital management framework. And with that, it's back to you, Stephen.

Thanks, Warrick. The next few slides will look at our strategic position and opportunities in both the SLS and metal businesses. Turning first to SLS on Slide 26. What this slide demonstrates is that the technical infrastructure required for AI is significantly more compute-intensive and expensive. If you look at the right-hand bottom chart, you can see the growing price for GPUs as they become more and more sophisticated to deal with the increasing demands of AI compute. We resell some GPUs today, but they're not overly sophisticated and the price is measured in hundreds of dollars. This is about to change. The complexity of repurposing GPUs that will start coming to the market in the next 12 or so months is an order of magnitude higher than what we do today. We are well positioned to capture this and have already commenced R&D to prove we can test and certify these GPUs in a real-world environment. And I'll cover this in a bit more detail on Slide 27. The left-hand graphic compares the more traditional infrastructure we have been repurposing for the last few years, with what is on the way from AI infrastructure. A couple of points to note here. Firstly, the significant expansion in the technical complexity and value of GPUs that I've already mentioned. And secondly, the density of the racks we will be dealing with. Right now, a rack we repurpose weighs up to 1,000 kilograms. Soon, these racks will be 5 tonnes. We have already been sent samples of what is coming from a major customer in order for us to assess the opportunities. We are well positioned to capture these opportunities. We have strong global and embedded relationships. We have proven and global leading technical expertise in the current equipment and are actively pursuing R&D to ensure we are ready for the next wave. Slide 28 brings this all together to look at the medium-term growth drivers for SLS. Firstly, with all the talk focusing on the rapid growth in AI infrastructure, it is easy to forget that market commentary indicates higher volumes of DDR4 repurposing will last beyond 2028. DDR5 will be the next major memory repurposing opportunity, likely emerging over the next 12 to 24 months. I've already covered off GPUs, advanced processing and AI infrastructure expansion, but it's worth noting that certification and aftermarket services will be an increasing part of our business as it relates to these activities. We see deeper customer integration as key, driven by our R&D and specialist engineering. This leads to the final point where it is becoming increasingly self-evident that the long-term pipeline of recoverable AI infrastructure is very large. Switching businesses now to our metal operations, beginning with North America on Slide 29. The demand for both ferrous and non-ferrous continues to rise. Ferrous is being driven by further commissioning of EAFs. Our estimate is that another 7 million tonnes of high-quality ferrous scrap will be required by 2029. The demand for non-ferrous is being driven by many things, including the building of data centers. The table at the bottom shows that an estimated 11 tonnes of aluminum and 11 tonnes of copper are required per megawatt of new data center capacity. This growth plays nicely into our North American businesses, as shown on Slide 30. The left and middle charts show just how important non-ferrous is to our North American businesses, with over half the revenue now coming from non-ferrous and over 1/4 of that coming from NFSR, which helps underpin our investment in shredders and downstream processing and recovery technology. NAM and SAR operate complementary footprints across the United States, but with different network profiles. SAR has a greater concentration of sites in dense regional markets, while NAM is more weighted towards major metropolitan areas. SAR's network comprises 153 sites and 23 auto shredders, giving it almost twice the number of yards and shredders as NAM. This density allows SAR to buy more non-ferrous material at source, purchase more unprepared ferrous scrap and produce more NFSR. This operating mix was an important contributor to its relative performance during the period, particularly the second half. Turning to ANZ on Slide 31. There is no denying that the ferrous business for ANZ has been tough over the last 2 to 3 years with exports from China. This will be somewhat alleviated over the medium term with planned mill upgrades in Australia, the Glenbrook EAF in New Zealand and the likely development of one, but maybe two EAFs in Australia. A reasonable scenario presented in the chart shows that surplus scrap available for export could fall to under 1 million tonnes by 2029. This will be quite pronounced by state with some in surplus and others deficit. Our national coverage and superior logistics will be an advantage as this scenario unfolds. Like our North American businesses, ANZ has a strong non-ferrous operation, and this is shown on Slide 32. Nearly 60% of ANZ sales revenue comes from non-ferrous and around 14% of that comes from NFSR. This strong non-ferrous position has enabled ANZ to navigate the particularly tough ferrous market conditions it has experienced. ANZ is investing in more advanced metal recovery plants and fine plants to ensure valuable non-ferrous is recovered and not sent to landfill. ANZ has an extensive national network of collection yards and is growing at source volumes, which will drive further non-ferrous growth. What this all means for ANZ and North American medium-term growth prospects is covered on Slide 33. The North American market is likely to see further consolidation, benefiting both NAM and SAR. Some of this will be acquisition of mid-tier businesses with shredders and feeder yard networks. Others will be bolt-on feed yards to expand our network supplying existing shredders. This will provide more unprocessed outsourced material, including non-ferrous. There is still room to improve downstream processing through our existing operational excellence and further technology. The demand for ferrous scrap is a medium- to long-term tailwind as tariffs and EAF expansion support demand foreseeable future. In ANZ, the commissioning of Glenbrook and potential FID for Alter will provide medium- to long-term support for ferrous scrap demand and prices. As with North America, there are opportunities for bolt-on acquisitions to support increased processing capability and capacity. Finally, non-ferrous is a near-term, medium-term and long-term growth driver for ANZ. My final slide before going to Q&A is Slide 34, which is more short term by focusing on the outlook for FY '27. Starting with SLS. We expect the first half to produce an underlying EBIT between $75 million and $90 million. Fundamentals remain strong, but as we have discussed on a number of slides, variability in decommissioning in the very short term will be a feature of this market. We are not expecting as much high-speed DDR4 volume in the first half of FY '27 as we had in the second half of FY '26. We expect that the factors supporting a strong non-ferrous contribution will continue in FY '27 for all our metal businesses. We expect the ferrous contribution from our North American metal businesses in FY '27 to be supported by tariffs, EAF growth and steel demand from data center growth. We are not expecting a material reduction in Chinese steel exports in FY '27, and this will continue to impact ANZ's ferrous business. Before we open for Q&A, as always, I want to thank our employees for their drive and commitment in delivering on our purpose and most importantly, doing that safely. Back to you, operator.

[Operator Instructions] Your first question comes from Peter Steyn from Macquarie.

Just on SLS, you mentioned the need to engender yourself with your customers. In my mind, what's happening in your services revenue is probably the key one there. And it's lost a little bit of momentum in the second half relative to where you were in the first half, obviously, still growing, but not showing the type of growth that certainly your repurposed units would demonstrate. Could you give us a bit of a sense of how you focus the team on that, what you're seeing from a services perspective, whether that is the correct way to think about this business and the strength of it and depth of its customer relationships?

Yes. Sure, Peter. So when we -- with those services, the predominant service we provide there with the most valuable service is where we'll take a DDR4 out and repurpose it back into the business. So we're -- so that continued at [indiscernible] pace. There was no issues with that in the second half versus the first half. What happened is that the customer can get to choose whether or not it gets repurposed back into them or do they want it to be resold into the market. And I think particularly in that second half, a number of customers made a decision, well, we may not need that as urgently back into the center, let's take advantage of some higher selling prices and resell it into the market. So it's more what was driving the sales. We are optimizing the sales. It wasn't a falloff in the amount of that sort of core service that we do around taking DDR4 out of the call center and the data center and deciding what to do with it from that point.

So I mean, maybe just coming to the strategic thrust of the question, are you actually focusing more attention there than ultimately trying to maximize e-commerce revenue?

So I guess what we focused on is maximizing inflow. And so that is about -- I mean, an inflow comes from strong relationships, embedded relationships, our systems tying in with their systems. What -- the process that we then do on what we inflow is very, very similar, whether it's going to be resold or repurposed back into the data center. If it's repurposed back into the data center, there's more around, I guess, inventory management and logistics to get it back into the data center at the right place. But we focus on inflow. Ultimately, whether it is resold or goes back into the data center is the core of the data center itself.

Sure. And then maybe just a change of focus to SAR. Could you give us a bit of a sense of how to think about the annualization impact of acquisitions in SAR just rolling into '27?

Yes. I think -- and I might get Rob to have a think about this question as well, but let me give me my initial thoughts and Rob spends a lot of time with SA Recycling up in the U.S. So SA Recycling, its acquisitions in FY '27 were almost entirely good quality bolt-ons. I cannot recall off the top of my head whether they bought another shredder. I think they did bought another shredder down in Florida. So what that does for SA Recycling is it provides a sort of further in-feed into its shredders, which have also got extra capacity in them. So what would I think -- what does that mean for the run rate leading into sort of leading out of FY '26 into FY '27? I mean I think a combination of strong non-ferrous markets, which is really driving Zorba and non-ferrous retail and justifying these feeder yards hugely for more unprocessed material. And I think broadly speaking, I think the run rate in the second half is not bad as we go into the first half. But Rob, I'd like to get your views on that as well.

Yes. The only thing I could add, Lee, is that in our slide deck, you see the incoming, I guess, demand curve, we've been talking about for the last 3 or 4 years on the ferrous side. And what SAR and NAM are enjoying right now, and we've invested in our capabilities to get product to market is I won't call it exactly an insatiable demand for shredded product. So we have a very good demand for ferrous, which leads well into the non-ferrous story that we've depicted here today as well. A very good demand curve for aluminum products, for copper products, all of which we liberate from a ferrous intake. So incredibly good margin uplift there and capacity, still, that we have in our network. We've invested back into those networks, the SAR and NAM both have capacity to do spare as well.

Your next question comes from Owen Birrell from RBC Capital Markets.

I just want to -- probably just going to extend on some of Lee's questions around the volume assumptions rolling into that first half '27 guidance. Your guidance is from an EBIT perspective, down, call it, 1/3 half-on-half. And I just want to understand, should we be thinking about ongoing growth in, say, the input volumes but it's purely a mix or should we be assuming both a decline in the input volumes and then also the mix on top of that? Because I know last time you gave us a memory sold guide to align with the EBIT guide. This time, you haven't given us the memory sold guide, but it sounds like you have some reasonably good visibility into that. Just wondering if you can give us a sense on when that memory sold number, so we have a comparable...

Yes. I think in the medium term -- so a couple of points in there. As this market has matured, it's really interesting that gigabyte memory sold is still a nice measure, but then it's the mix within there, about speed. So it's not just DDR4, it's the speed of the DDR4s. I think what you should be thinking beyond this, beyond the first half, and like I said, we do have quite a lot of visibility into the first half. We don't have as much visibility into the second half. But what we do know is that the volume is going to increase. There's -- in the market as a whole, there's not a list coming out from December on. There's more coming out from December on. The timing of that is still -- like is it going to come out in the second half of '27? Is it going to come out in the first half of '28? That's not up to us. That's up to the timing of the data center. So more volume more volume will be coming out because more volume is sitting there than has been in there. The second point I'd make is that the mix of that volume should improve as well because as we're getting more through the refresh cycles, we're getting higher and higher quality DDR4s that are faster and faster, and those are the ones that are worth more. So I'm sitting here feeling pretty comfortable about the medium-term outlook. I just, unfortunately, can't provide you a half-by-half assessment of exactly when that will come out. But it's not least volume and it's not least quality. It's the -- I think the main point is more volume, more quality than I would make.

Yes. I understand, I guess, the quality of comments that you had provided. But the challenge that we're facing is we don't have a baseline. And it's all good and fun to talk about growth in different metrics of the baseline, but we don't have the baseline. And I guess that's what I'm trying to get at. So is there any better color you can provide around, I guess, how the mix currently stands so that we can move beyond that, it would be appreciated.

Yes. If it -- let me take that question on notice because I don't want to -- I'll be really frank, I don't want to blow out a whole lot of numbers now that we haven't thought through as to what -- as to how you would interpret those because there is some complexity going on here. So it's a good question. Let me take that on notice and Warrick and I will talk to Anna, and we'll see what else we can maybe look at for you -- well, not for you, for the market.

For the market. Sure. And can I just ask -- I mean you mentioned comments around the broader market and [ volume ] has kind of increase in the market as a whole. Can I just ask how you feel SLS is positioned within that market because we're constantly hearing about data center growth coming at pace. I know you sort of talked about the decommissioning cycle kind of pausing for a moment or different assets being decommissioned. But we do know that the growth in the platform is continuing to grow across multiple different clients and across the U.S. predominantly. But what I want to understand, I guess, is how is SLS positioned within that? Do you -- has SLS grown its customer base as quickly as the market has grown during, say, the last 12 months? And do you feel like you're growing or losing share as a proportion of the total market? I know it's very, very vague, but I just keen to get your thoughts.

Yes. So I feel we -- so if I look at -- 2 aspects to that question, existing customers and potential new customers. We -- I feel we are growing our share of existing customers. Our relationships are strong. We have a good idea of what they are producing in total. We know what we are getting, and I feel very comfortable with our existing customer base that we do. And we're actually doing very well in the existing customer base. We are -- so, I feel very comfortable about that, and those customers are big within the market. We're not talking a small players, they're big customers within that market. And it really is -- I know I'm going on about this a bit, but it really is the short-term 3 months, 6 months is just not the way that these people think and so, therefore, they'll decommission and refresh however they want to. So -- but I feel we're growing in there. In terms of new customers, we definitely had some success in gaining new customers. But when you gain a new customer, it's a foot in the door. It's not being -- you're going to get the same volumes as your existing customers from day 1. But I -- when I look at our competitors in the market, I feel we are holding our own at the very least in terms of market expansion, and I could argue growing it.

And can I ask, in terms of the volumes that are coming through, I know you mentioned the sort of differentials between high-speed, low-speed, DDR4s, as I mentioned, DDR3 in here and at some point, DRR5s are coming through. Are you able to give us a sense of your -- the rough splits around the memory that's currently coming through the -- your sales base more particularly?

Yes. Right now, it's virtually all DDR4s. DDR3s probably came to an end through the second half of last -- of FY '26. Most of it would be. I mean, for all intents and purposes, you should think of it as DDR4s right now, and the odd DDR5 is starting to arrive. Some of it is from our customers who want us to check exactly how we're going to handle DDR5s, that there's some slight technical difficulties, nothing -- difference, not difficulties, nothing we can't handle. And then so think for this first stop, it should be DDR4 is almost exclusively. And then DDR5 will start to come in, I would say, sometime in the next 12 months. But let me be really clear not at the expense of DDR4s. DDR4 will continue to -- there's a lot of DDR4s installed in data centers that need to come out.

The repurpose units numbers are pretty strong in '26. Of your resale revenues, how much is from DDR4 sales versus sales of other repurposed units. Is it like 90%, 95%?

No, it's not as high as that. So repurposed unit is -- can be anything that comes in from -- it could be -- like, for example, a cable. I mean just to use a simple example, cable that comes in and resold is considered a repurposed unit. I do know what that number is. I don't know if we have disclosed it. So...

You haven't. That's why I'm asking.

Yes. So. It's certainly not 90%, but it's more than 50%. Let me -- again, I wasn't -- let me take that. I'm sorry, I see -- let me take questions on notice for you. Let me just double check that, that's not commercially sensitive, and I'll put that into the -- and is taking notes. So I'll put that into the list of things that do we need to disclose that or our commercial guys' saying were made to be disclosing that.

You next question comes from Brook Campbell-Crawford from Barrenjoey.

I just had 2. Firstly, just around new contracts that you're signing recently, the revenue share terms similar to your existing kind of longer-term contracts? Just want to check if -- in case there's any slippage there given, obviously, it's a really profitable business there relative to how it's been in [ used online ].

Yes. So I mean, firstly, Brook, we definitely won't disclose revenue share terms on specific contracts. And I think I know what you're talking about. What I would say is the contracts, moving forward, will have higher volumes because there's more coming out. But we definitely won't disclose what our revenue share terms are on those contracts.

Yes. No, that's fair enough. I guess I wasn't sort of looking for specifics about customers. Maybe just a broad trend. Is it sort of staying stable in terms of your share versus customer in general? Or is it changing at all?

I mean I think as we're rolling contract, it varies by contract. We always trade-off -- I mean, as we're negotiating revenue share, we trade off volume with that rev share to see because we're looking for total margin. I'm not sitting here -- if your question, am I just sitting here worried about it? No, I'm not. I'm very comfortable with the deals we're negotiating at the moment and the rev share split and the volume split and what type of volume we're getting out. Actually, maybe that is a very, very good point that I will make. And I want to talk about GPUs because the rev share on a GPU will absolutely be lower in percentage terms than the rev share on a DDR4. The GPUs that are going to be coming out over the next 12 months and beyond are extraordinarily valuable. And we will do a lot of testing on them in the real-world situation for them to be either repurposed or resold. And clearly, we're not going to get a 25%, 30% rev share on those because the value of them is just extraordinarily higher. But the opportunity for absolute margin is very, very strong.

Yes. That's really helpful. And maybe just digging into that slightly more. If we think about the opportunity for EBIT in SLS, how you think about GPU that the new product focus for your business? And how could that compare to memory? I guess you could think out medium term, is there some sort of rough split...

Yes, it's a good question. I see GPUs as a bigger opportunity than DDR4s for a couple of reasons. One is -- 3 reasons, actually. One is the complexity of repurposing them is an order of magnitude higher than what we do today. So therefore, the relationships you have with customers, your -- the R&D work you've put in creates an even bigger moat around it. So I see that as very, very important. Secondly, the volume of GPUs that needs to come out is extraordinarily large over the next 1, 2, 3, 4 years. And thirdly, the value of those is extremely high as well. Fourthly, and this one, I think we're generally still thinking about it. There may be a bias in these new GPUs for them to be resold, we'll have to see because whether or not I want them back, and are they going to be good enough for the compute, that really fast compute they need, I'm not sure for massive hyperscalers but there's certainly going to be good enough for a huge number of potential customers globally. So I see GPUs opportunity is larger than the DDR4 opportunity and the DDR5 opportunity.

Your next question comes from Scott Ryall from Rimor Equity Research.

Stephen, the first one, probably, very quick. Corporate costs. You had good decline this year. Is there further declines that you expect you can deliver over the next couple of years, please?

I will flick that one to Warrick who's been working hard on this.

Yes. I mean, I think it's -- we've continued to chip away there, Scott. So I'd hope to see some further reduction. We do have project costs in there. So they will come out at the -- in this half. We start -- not totally, but partially come out in this half as we move to implementation of that new metal software. So yes, you'll continue to see a reduction in that. And we continue to look at opportunities across the board.

Okay. So can I just confirm then, you would expect to see -- so, if you've got a negative $114 million of EBITDA, that number should be closer to 0 than what it is -- like it will reduce as a loss in '27. Is that fair?

Correct. That would be our intention.

Okay. Great. And then my second question, probably for Rob. I'd just be interested now you've got the North American business in a -- obviously, it looks a lot healthier. And you've talked about the tailwinds from EAFs and the strength of non-tariff pricing. I'm just wondering now, does this give you a little bit of time now or an opportunity perhaps to think about positioning the business for the next 3 to 5 years, maybe it's taking out some of the cyclicality, maybe it's improving what you think is trough earnings. But how do you -- how are you thinking on the kind of medium to long-term of positioning yourself for both tariffs in the context of EAF and the opportunities that come with non-ferrous volumes attached to that, but also, there's got to be opportunities outside of just collection of scrap metal for the non-tariff as well I would have thought.

Correct. Yes, I think NAM is definitely in a position now, foundationally, I think, solid results with the investments we've made, with the efforts we've made to diversify from largely an export-based company from a ferrous perspective to having the supply chain and logistics capabilities that we've invested in. Our diversification, I guess, has grown beyond my expectations in terms of our ability to optimize on a monthly or quarterly cadence. So we'll continue to leverage that. The demand in the U.S., we will not ignore the customer's customer, as Stephen has explained. It's a good environment, good economy, but the data center environment is just putting it on hyperdrive for us. So now that we've stabilized and I'd say, earned the right to kind of grow again, much along the lines of SAR. I think the right opportunities at the right time for tuck-ins to continue to utilize the deployed capital we already have in the market, in North America with our shredding aluminum capacity with Alumisource and the granulates that we have on site. We'll continue to look for those opportunities and grow in the market.

Your next question comes from Harry Saunders from E&P.

Just on SLS again, looking at it a slightly different way. If we look at the moving parts, getting to that guide in the first half there's volume, there's other factors. And it looks like the memory price that you've referenced previously, if anything, has strengthened into the half. So just trying to figure out what those moving parts are in broad buckets? And then maybe what is a more normalized figure do you think for this business? Can you then factor in the uplift from Ireland because presumably, Ireland is kind of contributing there?

Yes. So the biggest driver for the first half of FY '27 is the DDR4s that have come out that we're selling are of a lower speed. And so therefore, the discount that they are to the new is a higher discount. So when we -- in the chart that we've traditionally shown, I'm now doing this off the tip. I think the speed of those are 3,300 or 3,600 -- 3,000-something. 100? 3,200. Okay. So the DDR4s that we're selling in the first half, the ones that have come out they are at lower speed, which is kind of intuitively correct because if you're refresh cycle is being delayed because of various blockages or your new sites are being delayed because of various blockages, you're going to be pulling out the less valuable stuff first and that will continue to come out. And so that's the biggest driver, I think, on first half '27 versus second half 2026. Nothing else material has changed other than that. And that's why, I guess, I get some confidence you talk about what's normalized going forward. I think I think it's very hard to do normalize because we're in a high-growth business. And so where I get confidence from that is the high-speed stuff, the DDR5s and the DDR4s all that's still to come out. And that will come out sometime after December 2026. So Harry, unfortunately, I can't say it's a hugely normalized base because we just need to see those decommissioning schedules, and as they get into the higher quality -- back into the high-quality material that they've done in FY '26 as that starts coming out, again, the prices are still -- as you've noted, prices are still very, very strong. And there's been no indication that those prices are falling. There's been no indication that demand for DDR4 are falling. So haven't given you a specific answer of what's the normalized earnings, but I think normalized earnings in SLS is a very hard concept because it's -- what's going to drive the future is higher quality, more sophisticated material coming out.

I guess we're just trying to understand that because there's a huge swing from the second half to the first half. Sort of whether the second half was overrunning perhaps. I appreciate you're referencing a different speed, but maybe you were selling more to the higher speed than in the normal kind of level of sale. And then also just trying to figure out what the Ireland contribution...

So what I would say now is we're selling -- what we're selling now is less than normal of -- in terms of mix because we are selling more of the low-speed stuff. So I'd say it's a less, not more. And Ireland is not contributing much at all to the first half because it was being -- it was set up and is up and really and running more for the high-speed DDR4s that were expected to come out. They will come out. I mean, it's not it's not they're not coming out, but -- so, Ireland is not contributing a lot in the first half at all because the DDR4 was that it was going to repurpose resell haven't come out yet.

And just to follow up on SAR. Given that enormous step-up in the second half versus first. I mean, is that a sensible run rate going forward? Or is there anything else to sort of bear in mind in that number that we need to normalize for?

So volume-wise, I don't think there was anything particularly special in the second half of -- I think it's the type of volumes that they can -- that SAR can absolutely do. So it really is price, and I would go -- and this be really clear, non-ferrous zorba prices, copper aluminum are the drivers. And I guess what we're seeing -- and again, this is an opinion because it's kind of like a global opinion. It's do you see anything driving down the value of non-ferrous? We don't see anything driving down the value of non-ferrous, but I'm sure you and your research teams will have your own view. I guess what I'm saying is there was nothing special about volume. It's about non-ferrous pricing. What do you see as non-ferrous pricing over FY '27 is going to be the key as to whether or not they can repeat that result. And frankly, it's the same with NAM. And frankly, it's the same with ANZ. None of them are different. There was -- maybe ANZ is slightly different. Do you see something happening with China in ferrous, that would impact ANZ. But NAMs volume -- the volume that NAM did in that second half, very sustainable, Rob. I don't see there's nothing particularly -- like there was no one-offs that you think that can't repeat itself. So it's about the margins that we're making on non-ferrous. Zorba I guess, in particular. And I think we don't see anything driving that down on FY '27.

Your next question comes from Chen Jiang from Bank of America.

Most the questions have been asked. Just again, a follow-up on the SLS to give us, I guess, more conviction. Your earnings growth from SLS is deferred further rather than disappeared over in the near term. So to summarize what -- from your answer, it seems like -- so over the next 6 to 12 months, you will be selling less DDR4 high margin and your customers are taking more in-house for repurposing because for repurposing units volume continued to grow and the price is strong. Is that how I should read it from all your answers?

Yes. So there's a few -- the one thing I would say is that -- we haven't said that DDR4 high volume will be less in the second half. We've seen it will be less in the first half and we have a line of sight for that. How quickly it recovers after that, it may well recover in the second half. It may take 2028 before it recovers. That will largely depend -- no, that will entirely depend on how our various customers -- do they secure connections, do they secure the DDR5s that they need so that they can refresh and take out DDR4s? I'm not saying -- making it really clear. We're not saying that's not going to happen in the second half. Then the mix of it, I don't think that's going to particularly change in FY '27. I think what I've seen is beyond FY '27, beyond FY '27, I think more of the DDR4s -- and this is a view here. This is -- more of the DDR4s rolling up in the resale market because it will be DDR5s that are getting more repurposed back into the data centers. Believe me, the world needs DDR4s in the resale market because they -- all the market commentary is that DDR4s are needed right through to beyond 2028, 2029.

Sure, sure. So basically, are you saying your customers are taking more in-house for the -- because your repurposing units are growing, whereas you are selling less. So does the...

Yes. So think about -- yes, so a lot of our repurposing unit growth was for non-memory parts. There's lots of parts that get repurposed and add to that DDR3s as well. So between those 2, I mean, there was a lot of DDR3s in FY '26 they're not going to -- I mean DDR3s are not going to come out in any significant way. From now on, it's DDR4s and DDR5. So I don't think you can say -- no, let me make it really clear. You cannot conclude that, therefore, our customers are taking more back into the business as opposed to reselling. We're not seeing any particular -- right now, we're not seeing any particular variation between those 2 from what we've seen historically.

Yes. So that DDR4 sales, is that more like a timing thing deferred further into your like whatever medium term you mentioned?

Yes, DDR4s are differently timing. There hasn't -- 100% agree. There hasn't suddenly there's no DDR4s in the world that need to be taken out of data centers. In fact, it's -- that installation happened 2 or 3 years ago. So it's not list DDR4s, it's simply the timing of when they're going to come out.

Okay. Got it. So deferred further. Okay. And then if I can ask about your North American metals recycling business because comparing them, your FY '26 EBIT versus the market being consensus, I think it's weaker than expected. But looking at your sales volume seems okay because U.S. steel production have been, I guess, strong year-over-year because of the tariff. And then now looking at non-ferrous price, copper and zorba, which is the secondary aluminum price, is still strong, which reflected your JV . SRs EBIT, everyone's asked about it. It's very strong, that's how the EBIT should be. And then you continue to implement your turnaround strategy. So I'm just try to understand what happened to the North American metals for the FY '26? And -- I mean, everyone has their own assumption for non-ferrous, but how should we think about it like over the next 6 to 12 months especially your turnaround strategy.

So broadly speaking, I mean, both -- there's one slide in particular that Warrick showed, which showed the contribution from non-ferrous versus ferrous between FY '25 and FY '26. I cannot remember what slide it was, but what it shows is the 2 big stories in FY '26 where the contribution from non- ferrous and the contribution from with a big growth drivers for the business. How you should think about North America and SAR relatively speaking? SAR is benefiting more at the moment than North America because it produces more zorba. And as I said, zorba comes -- it's not a 0 marginal cost, but zorba becomes at a very low marginal cost. So as the price rises, it tends to fall straight to the bottom line from an EBITDA and an EBIT perspective. NAM is proportionately doing just as well out of non-ferrous as SA Recycling. It's just SA Recycling produces more of it, on the zorba side. And then the second thing I'd add because it's got roughly twice the number of feeder yards that we have in NAM, they also do very well out of non-ferrous retail. And non-ferrous retail is just plumbers bringing in copper guttering and electrician dropping off some cable and some copper cabling that they picked up, and then we process that into high-quality non-ferrous products. So I wouldn't be thinking that SA Recycling has got something that NAM does it. What SA Recycling is twice as ministries and twice as many yards providing this non-ferrous product.

Right. So comparing if your [ JV ] operated SAR with your NAM, so basically, from your ounces, there's not much difference, but you believe they have more high-margin like zorba non-ferrous various metals, which give them higher margins. Is that [ a good ] conclusion?

That is correct. They shred more, therefore, produce more zorba and zorba has a very low [indiscernible] run marginal cost and so that has impacted them in a positive way. Don't get -- it's impacted there in exactly the same positive way, just lease volume.

Right. And you are benefiting from higher U.S. domestic steel production, which we've seen year-to-date because of tariff as well as strong early prices and the copper prices in the U.S.. All right.

Yes. That's correct. Now...

Your next question comes from Daniel Sykes from Jarden.

I was just wondering if you could give us a bit of a time line around some of the other hardware components you're talking about in SLS, particularly the GPUs and the market opportunity there, just in terms of any significant milestones we should expect in terms of kind of testing whether they can be resold any contracts and -- will hit the financials as well.

So we have some of them in our HQ facility right now, some of these high-quality ones, and we are working on them at the moment, testing them in a real-world environment. So -- that's a very important step. We've already developed 3 or 4 tests and have proven that based on those tests, we can certify these things to a certain level. In the next, I'd say, around about from -- 12 months from now, there's going to be another step up in the quality and sophistication of what's coming out. So it really is -- for me, it's starting now and it will ramp up in the next 12 months or so. But what I would say is we are very well-positioned to be sitting in that market, testing and certifying GPUs. And that's what this market is going to be about. GPUs are not like DIMMs, if a DIMM fails, it's not the end of the world. And the testing that you do on DIMMs is relatively simple. DDR5s and neither of those things, they are more fragile, the tests are more complicated. That's why they're worth more and that's why you need to have the skill to do it. And I think we are -- I do believe we're really well-positioned. It's starting from now ramping up over a 12-month period is the way I'm seeing it.

Great. And do you see any one, I mean, in the market doing this already? I know on the memory side, you kind of talked to hyperscale as being the #1 competitor there and what they do in-house? Is that the same on this side? Or is there anyone you would point to?

This is new. This is new and we've recently written quite a good white paper on some of this stuff as well. I mean, I'm not going to say there's no one out there doing it because it's hard to know what people are doing behind closed doors. But I do believe that we are at the forefront of it because we've been -- particularly with one of our customers, we've been working on this for a while. And that customer has a particularly strong focus on sustainability and making sure that the stuff is not going to landfill. So we've been working on it for quite a -- for a while, and I feel we're in a good position. I mean, we're never going to have a world monopoly on it that would be lovely, but highly unlikely. But it's -- this takes it to another level of sophistication versus DDR4s and DDR5s.

Your next question comes from Ramoun Lazar from Jefferies.

Just a couple of questions for me. Just with the SAR, Stephen, you mentioned that assuming non-ferrous prices stay where they are, that earnings base is sustainable into '27, I guess, just with all the investments SARs made over the last little while. Are you able to sort of give us a bridge on what sort of volume uplift we should expect from those acquisitions? So any sort of led it would be helpful to try to frame that into '27?

The volume outlook won't be huge Ramoun, and it's not about volume. It's -- I think it's about shoring up their market position. So making sure that they preserve what they've got and they're preserving their margins. So I wouldn't be assuming a huge volume uplift relative to the size of recycling, which is a -- it's got huge volumes down, don't assume a significant -- don't assume a material volume uplift. It's really around shoring up the market position. It's probably the best way to say it.

Okay. Great. And on SLS, it sounds like the second half '26 was a bit of an anomaly in terms of customers just trying to look to monetize a high DDR4 prices. Am I wrong in thinking that using second half as any kind of baseline for SLS is probably the wrong approach to forecasting this business going forward?

I think over the medium term, FY '26 is perfectly fine. I don't think the second half of FY '26 was an anomaly. I think the anomaly is what's happening right now, is the absolute crunch that's happening with everybody wanting to build data centers, everybody looking for electrical connections, everybody looking for land. I think that's the anomaly that's hit everybody. I -- look, and the reason why I'm comfortable in saying that is there is a hell of a lot more DDR4s sitting out in the data center land than what's been repurposed to-date. So I guess what my argument here, Ramoun, would be 6 months -- I think 6 months is a blinkered eye relative to this market as what -- how quickly it changes. Once those bottlenecks are opened up, which they'll have to be, the material will flow just as strongly, probably more strongly than the second half of FY '26 because there's just simply more out there.

Yes, yes, yes. No, I take your point. Do you have -- I mean how much visibility does SLS or Sims to have with regard to that, Stephen? Like can you see to the June half of next year yet? Or is that just based on the scheduled time lines of your customers or not yet?

We have a pretty good line of sight for the first half, which is what we've got to have a reasonable basis before we can put things out. So we have a reasonable line of sight for the first half. There's some initial forecasts come from our customers for the second half, but they become heavily caveated. So no, we -- it won't be until we get into the second half that they really firm up when the material is actually going to come out in that period.

Okay. And there's no sort of lead you can provide us in terms of an assumed seasonality in that business this year?

No, No, the seasonality, if there will be commissioning seasonality, which is -- it's just entirely based on when they get hold of electrical connections when they get hold of land, when the construction happens. If they've got a data center they're refreshing, obviously, if they're refreshing it with DDR5s and HBM and high-power GPUs, it's going to need more electricity. So they're getting the electric grid connections sorted out. Once that sorted then bang, all the DDR4s that's sitting in there will come out ready for the new DDR5s and HBM and GPUs to go in. So that's -- that doesn't have a seasonality to it. That just has when is the electrical connection approved and put in place. Those are what causes the variability between reporting periods.

Got it. Okay. And just one final one on ANZ. Do you think that division is now stabilized in terms of the step down given what's going on with Chinese steel exports being partly offset by some of the additions like Glenbrook?

Yes, I do. I do. On the assumption that non-ferrous stays where it is, and we're confident in that. But it's fine. I hope this is not a famous last words, but it's hard to see the ferrous market getting much worse for the impact that China is having on it. So yes, in that sense, I feel that this is the bottom earnings for ANZ.

Thank you. There are no further questions at this time. I'll now hand back to Stephen Mikkelsen for any closing remarks.

Okay. Well, thanks, everybody, for the questions. Very -- some very good questions there. We will see -you - we will see all of you over the next couple of days, and I look forward to catching up. Thanks very much for dialing in.

And that does conclude our conference for today. Thank you for participating. You may now disconnect.