Good morning, everybody, and welcome at Rio Tinto. Good evening for those of you listening on the webcast. My name is Menno Sanderse , Rio Tinto Investor Relations. Before we kick off, a couple of housekeeping points from my side. First, can please everybody turn off their phones or at least turn them to silent. Secondly, and very importantly, because quite a busy room, there are no planned emergency drills. If you hear the alarm and it's a broken tone, then please stay in your seat, but be very alert. If you hear a continuous tone interspaced with a voice message, then please follow the fire warden and evacuate immediately. They'll come through those two doors. Leave via the staircase, as you would usually do. The muster point is in front of the King George statue on Number 1 Carlton Street. Then to start, the agenda.
We will start today with J.S., followed by Vivek on markets fundamentals, Simon on commercial and our connection with customers, Chris on iron ore, followed by the first of two Q&A sessions. Post the break, we will continue with Steve on technology, Arnaud and Steve on Oyu Tolgoi, and Jakob on the financials. I'll provide strict instructions for the Q&A process at the start of each session. During the break and afterwards, please, please try out the very cool virtual reality goggles where you can take a look at the OT open pit mine and the OT underground project. In the room opposite the virtual reality goggles, you have a display of the Mine Automation System, which look for a financial nerd is very, very cool. After all that, it's my pleasure to introduce J.S.
Thank you, Menno. I don't have a phone. Steve, have you stopped your phone? All right. Thank you, Menno. Good morning, all. I'm absolutely delighted together with the team to welcome you to our 2019 Investor Day. This morning, we will share our views on the macro forces shaping our industry, detail our strategy on how we will improve performance, and explain why Rio Tinto's many advantages position us to thrive in the short, in the medium, and the long term. To start, let's reflect on the last three years, which give us a strong foundation for ongoing success. Since 2016, we have outperformed the market on cash generation and shareholder returns, delivering $53 billion of cash, including $23 billion of free cash flow and $32 billion of cash return to our shareholders. We have deepened our relationship with our customers and sometimes with the customers of our customers.
For example, our breakthrough partnerships with Baowu in China and Apple. We have strengthened our portfolio and balance sheet, raising over $12 billion through divestment and reducing our net debt by $9 billion. We have improved our performance of our world-class assets, delivering an average return on capital employed of 17% and an average EBITDA margin of 42%. We have invested in growth, starting with exploration and embedded capital allocation discipline with $16 billion of capital invested. This performance demonstrates the deep partnerships with our customers, the quality of our asset portfolio, the strength of our cash flows, a capable team, and a commitment to reward shareholders. All in all, we have delivered 135% of TSR from 2016 to now. The last few years have not just been about cash returns.
We have also used our profits to advance our sustainability efforts, developing pioneering new value-enhancing partnerships, and delivering on our commitment as a responsible company. In 2016, we developed our purpose. As pioneers in mining and metals, we produce materials essential to human progress. In 2018, we released our refreshed ESG approach, which is core to the company strategy. This reflects our view that in the 21st century, business must play a bigger role to build trust and stay relevant. Sustainability is, as I've said over the last few years, a make or break for our industry. Key to this approach is understanding all material risk and how we are managing them. Sustainability has to start with profitability. Only a profitable business can generate benefits for all our partners, starting with our communities and customers. It is all about actions.
We know we will be judged by what we do, not just what we say. We have three pillars. One, running a safe, responsible, and profitable business. Two, delivering sustainable economic benefits. Three, pioneering materials for human progress. We are delivering in each of these areas. I will cover safety and operational performance shortly, but I would like to highlight other areas where we have made significant progress in 2019. We are proud that Rio Tinto now ranks number two globally in the Corporate Human Rights Benchmark, just behind Adidas. Our Canadian aluminum sites are all ASI certified. We were the first in the industry to issue a climate change report under the TCFD framework. By the way, the best way to tackle climate change is by working across the entire value chain.
Therefore, we have put in place breakthrough partnerships with world leaders such as Tsinghua University, the World Bank, Apple, and Vale, with more to come. Indeed, managing our partnership ecosystem very well is a key focus, not just in climate change. We are placing our customers and understanding of our market as absolutely core to decision-making, which is key to our value over volume execution and long-term success. There is no doubt sustainability is a complex area. We do not have all the answers today, but we are working hard on pragmatic solutions. We have a strong track record of performance and confidence we can continue to deliver. We know to thrive in this new era, we will need to do more. What can you expect from Rio Tinto?
Key to our approach is an understanding of, and planning for, the forces impacting the world and our industry. Let me share a few thoughts on this. Most of you have heard me say that I believe our industry is at a crossroads. There is absolutely no doubt in my mind, as we approach another decade, that we will face even greater complexity. This is due to three main forces at play globally: geopolitics, society, technology. Of course, these forces do not exist in isolation, but we believe in the immediate future, geopolitics will dominate, and all business needs to be resilient in this environment. The ability to manage the three together will define success in our industry in the medium and long term. Let me cover each of them quickly. Geopolitics is important as it does influence the two key drivers of the mining business, trade and GDP growth.
This is not new, but in the world of growing political fragmentation and nationalism, the impact of geopolitics is far, far greater. Just look at the impact of the U.S.-China trade wars on global GDP. Oxford Economics predict a figure of 2.5% GDP growth in 2020. Society is about a world where climate change, environment, as well as inclusive growth and being a good corporate citizen really matters. Companies can no longer be spectators on the sideline. We must be part of the solution. We are also seeing technology, automation, data, AI drive improved performance. Technology can disrupt and offer a new solution, which has been the case in many industries, but in some ways, mining is lagging behind others. As an industry, we need to embrace a different digital operating environment. All these forces will impact global economic development. Where is Vivek?
Vivek will talk more about this shortly. In this context, to create superior value for our shareholders, we need world-class assets, balance sheet strength, innovative partnerships, and disciplined capital allocation. We believe the Four P Strategy we developed in 2016 is the right framework. Underpinning it is our value over volume approach. It's about portfolio, performance, people, partners. The team will cover different aspects of this today. Let me share some high-level thoughts. Fantastic music. That was not in the script, this one, I can tell you. All right, let's carry on anyway. Starting with portfolio. We have a portfolio of high-quality assets. Each asset class has a different strategy in line with the Four Ps framework. In iron ore, it's about optimize and flex. As I said in August, we experienced operational challenges this year. We took clear actions.
We have positive momentum in the mines in Q3, and the team is working hard to fully optimize our entire iron ore system. Chris, where is Chris? Chris is here, will update you accordingly. In aluminum, it's about protect and fix in a challenging market environment. You will have seen last week's announcement about our NZ smelters in New Zealand, where we have started a strategy preview. Alf is here as well, will answer any question you may have. For copper, it's about unlocking growth. Arnaud and Steve, say hello. Thank you, guys. They are not twin brothers, by the way. Arnaud and Steve will provide you with an update on Oyu Tolgoi as well as answer questions on our copper portfolio. In minerals, where is Bold? It's about developing new opportunities. Bold is also here for questions.
We will continue to strengthen our portfolio, primarily through organic growth. As you would expect, we will maintain a watching brief for attractive M&A opportunities. Rest assured, we'll be absolutely disciplined in this area. We will also maintain our industry-leading investment in exploration. We have 69 programs in seven communities across 18 countries, and we'll spend around $350 million in 2019. Copper remains the main focus, and we have a number of exciting opportunities, including Winu in W.A., Western Australia. Our exploration program is a differentiator for Rio Tinto. We believe it is essential to keep our pipeline full and option-rich. Looking at our performance, let me start with safety. Safety is the fundamental building block of our operational excellence. In 2019, we have had no fatalities, and as you can see from the slide, we have reduced our process safety incidents in a material way.
Our financial performance has been very strong, and we are stepping up our operational performance across our entire business. External factors such as underlying cost inflation make this even more important. We will work hard to protect our margins. Technology will play an important role in this phase. Our commercial operation is a key part of this. It is helping us to strengthen our relationship with our customers to realize full value from our products in the market. Simon. Where is Simon? Simon Trott will explain his plans in details. Turning to people, we've got Vera as well. Our people are a competitive differentiator, and we are doing a lot to further develop technical and commercial capability. We have set up center of excellence in open pit, underground mining, processing, and last but not least, energy and climate change.
These centers bring experts together and are a way to develop skills, share knowledge, and deploy expertise to improve performance. Our investment in the commercial team is delivering additional value by providing insight and developing relationships that maximize the value of our products. More broadly, our focus on employee engagement and values and integrity continues. The trend is positive. Lastly, we are investing in the skills of future. Let me give you a few examples. In Australia, we invest around AUD 40 million in skills development in STEM and other education programs with university. We have started a course with TAFE on automation skills, the first of its kind. In Mongolia, we have a workforce of 15,000 people that is 93% Mongolian. We are very proud to have developed a generation of underground miners there. The last piece is about partnership. This is essential for strength and resilience.
It is a core building block of our sustainability approach and it drives value. We have a number of partnerships beyond our asset joint ventures across environment and climate change, across skills development and communities, across the supply chain. Our ability to extend our partnering philosophy to different players in the years ahead will be even more important. We need to partner better with technology players. We need to partner better with the customers of our customers. Both Steve and Simon will tell you more about this. In closing, we have the strength and resilience to thrive in the years ahead. In our business, there will always be new challenges. We are constantly assessing the prevailing winds and prepare for any changes and opportunities they may bring.
Our ability to create value in the short, medium, and long term is down to the quality of our assets, the capability of our people, our operational performance, innovative partnerships, and our disciplined capital allocation. Across the group, our EBITDA margins and return on capital employed have been resilient throughout the last 10 years, despite the cyclical nature of our industry. Jakob will talk more on this shortly. Our history prove we can deliver in different pricing, economic, and geopolitical conditions. To remind you, in the last three and a half years, we have returned $32 billion in cash to our shareholders. Although current market conditions are changing, our estimated 2019 free cash flow is $10 billion as for pricing. This compares with $7 billion in 2018 and an average of $7.5 billion over the last three and a half years.
This strong performance and cash generation is why we are confident we will continue to deliver in the short, in the medium, and the long term. We are not complacent. We have the ingredients for success, and we must execute with excellence and continue to allocate cash with discipline. On this note, let's turn to the team. Who's the first one? Come on, Vivek. How is China doing, Vivek?
Great. Okay, thanks very much, JS. Today, I will take you through our global outlook. The top line is that for the near term, we see global growth slowing. Rio Tinto is well-placed to manage in this environment, but we're not complacent about the risks. In the longer term, each of the markets in which we operate will face different structural features, and I'll actually spend the bulk of my presentation talking to you about these. Following several years of unprecedented economic stimulus, global trade, investment, and manufacturing have now entered a long-expected cyclical trough. This situation has been exacerbated by trade tensions and other geopolitical concerns, and all of that has led to elevated global risk aversion. In this context, economic activity has obviously weakened, and global GDP growth this year is expected to be 2.5% in market exchange rates.
JS. has talked about next year's growth about the same. In the meantime, Chinese growth has fallen to 6%, and this slowing trend is expected to continue as that economy approaches high-income status almost inevitably. In recognition of all these bearish conditions, policymakers around the world have started to stimulate their economies, and the consensus is that such efforts will lead to a stabilization of growth in 2020. We nevertheless remain alive to the risk of headwinds facing our sector. Looking beyond these near-term uncertainties to the medium and long run, we expect global commodity demand will continue to be driven by the fundamental mega-trend of income growth in emerging markets, along with some of the forces that J.S. has mentioned in his presentation. In this process, India and ASEAN will become increasingly important actors in this story.
They have large growing populations that are expected to move increasingly into towns and cities. With commodity utilization rates, as you can see on this chart, several times below those in industrialized economies. We expect that commodity demand in these regions will escalate rapidly. For instance, we expect steel demand in India and ASEAN to grow in high single-digit % numbers over the next decade. By contrast, China is entering a new era of economic development, and that was outlined by President Xi at the 19th Party Congress. This shift will see an increasingly wealthy and urbanized population working in more value-added jobs and expecting a cleaner environment. What does that mean? The use of materials for new buildings and infrastructure is expected to slow.
Additional demand will come increasingly from a large and growing manufacturing sector, including electric vehicles, and the need to replace buildings that have reached the end of their useful lives. This means, in net terms, that China will continue to provide a sustained baseload of demand for the commodities that we produce. I'll now turn to the individual commodities. An important idea within iron ore is what we call contestable demand. Some of you will be familiar with it, others may not. This is the market in which our iron ore competes, and it is made up of a number of geographies, and that includes Europe, Japan, Korea, Taiwan, and ASEAN. Of course, there are two others, China, which is the largest by far, and India, which could in the future become an important market for seaborne iron ore.
Pulling all that together, demand for iron units in that contestable region is expected to grow by 1%-2% in line with steel demand. For contestable iron ore, however, access to this underlying growth will be limited by two things. First is the increasing use of scrap in China especially, and second, whether India will be able to meet its own demands from its own resources. Turning to scrap first. You can see on this chart that China will certainly produce more scrap in the coming decade. The key question is how much of that will it use? The answer will depend on a number of factors, and key among those is the cost of actually collecting the scrap. For example, a significant proportion of scrap in China will come from demolished buildings.
It's actually expensive to recover that scrap because much of it is embedded in concrete. Turning to India. Publicly available forecasts for Indian iron ore imports can range from very low numbers up to a quarter of its overall demand. That will be well in excess of 300 million tonnes, that total demand. This represents upside potential for the contestable market in the future. We presented a range on the previous chart that shows what we're thinking in that space. As the iron ore market transitions from the exceptional growth of the past decade, we believe it will nevertheless continue to provide strong, sustainable returns for producers with low costs. There are two connected parts to this conclusion. Let me take you through those. First, prices, we believe, will be supported by high Chinese operating costs.
Those costs are expected to increase in the future as more Chinese production shifts to underground mining. Second part of the proposition is that given substantial reserve depletion over the coming decade, the industry will need significant investment just to meet the demand that's going to be there. These two factors together, we believe, will be enduring sources of support for return in this industry for incumbents. I'll now turn to aluminum. Demand growth in that industry will be led by automakers as automakers seek to reduce the weight of their vehicles to improve fuel efficiency. This has been an ongoing trend for years and years, we expect it to continue. Estimates from our main target market, that's for our smelters in North America, are healthy. North American demand, we expect to grow between 1% and 2% per year over the coming decade.
Of course, as with all the other industrial metals, scrap will play an increasingly important role in meeting demand growth, as substantial quantities of products containing aluminium reach the end of their life. The profitability of aluminium continues to be challenged by the capacity for people around the world to bring supply on quickly and cheaply to meet the demand growth that's there. I think that this underlines the value of Rio Tinto's position in Canada. With operating costs in the bottom decile of the cost curve, supported by hydro, which will obviously become increasingly important and valuable in a carbon-constrained world. We also see very significant potential in the demand for electric vehicles in the coming decade. First in China, then in Europe, and ultimately in the United States.
Battery costs have fallen almost exponentially in recent years, and as a result, EVs have become increasingly competitive against fossil fuel alternatives, even as the subsidies that governments have put in place to support EVs, even as they have rolled off. This will drive rapid growth in demand for battery minerals, including lithium, nickel, cobalt, and it will be positive for copper. Lithium-ion batteries are expected to dominate the market, with nickel-rich, higher-density chemistries increasing their market share. In terms of copper, demand for primary metal is expected to grow by 1.5% to 2.5% per year, and that's going to be supported by EVs as well as the increased uptake for renewables and power generation. Renewables need more copper now. With growing demand and depletion at existing copper mines, the market will require 6 million-9 million tons of additional mine supply by 2030.
To conclude, global growth is slowing, but obviously off a much bigger base than we've seen in the past. We expect the long-term structural trends that I've outlined to continue to support demand for our products. Thank you on that note, and hand over to Simon.
Thanks, Vivek, and good morning all, and great to be here. Against the market context outlined just now by Vivek, I'll talk today about how commercial is working to improve the way we market our products, work with our customers, and purchase the material and equipment we need in our operations. Let me take you through our plans and progress. Commercial puts the company's value over volume approach into practice. We link our customers and markets with our operations, informing production and future investment decisions to ensure both the amount and the type of products we produce meet customer needs and manage the trade-off between volume, quality, cost, and CapEx. Our strategy is built on four key pillars. Firstly, deepening our understanding of the value chain.
Improving how we collect, organize, and monetize information ultimately is about solving our customer challenges so that we generate value for them and for us. Secondly, building commercial excellence. We are taking what we do today and doing it better, and we aim to sell every ton we produce to the customer that values it the most, as well as rigorously measure and improve our performance. Thirdly, expanding our commercial activities into new areas, and that's really about moving from simple risk avoidance to an approach where we better identify, quantify, and then manage our risks. Finally, as J.S. touched on, achieving full value requires an integrated system responding dynamically to changes in the market and our operations. Commercial is ensuring that the needs of the market inform decision-making in real time. How are we organized?
Across these overlapping areas, there is one unifying goal, and that is to maximize the value of our physical flows. Commercial team is now strategically located closer to our key customers and suppliers in both Asia Pacific and North America. Our hub is in Singapore with key offices in China, Japan, South Korea, and Chicago. We have a vast network of customers and suppliers. Around 2,000 customers across 96 countries and 37,000 suppliers. This platform gives us tremendous insight and perspective on key markets. Our focus is on turning these insights into value, improving our business and those of our customers. Let me now cover some of our products, and I'll start with Pilbara Blend. Our PB Blend product is the single largest, most liquid and consistent product in the market.
We've positioned our PB grade around steel mills' average requirements. While steel mills take a range of factors into account, from phosphorus to particle size, our PB product is baseload for many China mills. PB quality and its consistency offer our customers significant flexibility, helping them more efficiently run their business, including managing inventory, price risk, and working capital. For these reasons, as you can see on the right-hand chart, our PB products trade at a consistent premium. Maintaining product quality maximizes value to both our customers and ourselves. PB is a core part of our mix and it's designed to maximize the value of our resource, as well as meeting the baseload needs of the China steel industry. PB is only one part of our portfolio, and we also have a number of other key products.
Within the high-grade segment, we produce IOC concentrate and pellet products, together with Pilbara Blend lump, which is increasingly valued by mills as a direct charge. These products target mills who have tailored their operations to extract the greatest value. HIY fines produce a high iron center, lower alumina and phosphorus. With an iron content of 58%, Yandi fines achieve a price similar to our 62% Pilbara Blend fines. In 2014, we introduced our SP10 product, we're targeting the segment of smaller mills, particularly in North China, who are less sensitive to phosphorus levels and more sensitive to input costs. SP10 has an iron content above 60%, as well as moderate silica and alumina, it is better than many competitive products and has performed well, particularly in times of compressed product differentials. Our road products are lower in phosphorus, we target the customer segment that produces high-quality steels.
By combining our understanding of our products, our competitors' products, and their value to specific mills, we can tailor our offering to the customer who values them the most. This allows for both short-term value optimization, as well as shaping investment decisions, so we align with our customer requirements. The added benefit of this work is that it helps us to build stronger long-term relationships with our customers. Let me now talk about how customers are at the center of our commercial activities. We have a number of customer partnerships, such as efforts to lower carbon emissions throughout the value chain and working jointly to improve the handling and screening of our products through their system. As J.S. has touched upon earlier, you probably saw recently that we signed an agreement with our customer, Baowu, China's largest steel producer.
We are also looking to maximize the value of our physical flows by extending our supply chain and building further optionality. By building a presence in Chinese ports, we are better placed to work with a broader range of customer base of mills and respond quickly to market conditions as conditions change. It brings us much closer to our customers. As an example of potential options, through portside trading, we are now trialing a blend of SP10 and IOC concentrate to meet customer needs in new ways and open additional optionality for our business. Newer technologies also have great potential to improve how we engage with our markets. For example, our China portside customers will be able to order via a mobile app.
You'll be able to have a go during the coffee break, perhaps also order a few tons, in the same way as you would place an order on Amazon. Earlier this year, we also piloted the first fully integrated blockchain paperless transaction in the industry. Turning now to aluminium. In bauxite, we have combined the strong technical skills developed in our refineries together with deep customer relationships to build a market for our bauxite within China, and we're now taking those insights to further expand our customer base. In alumina, we require over 6 million tons for use in our own business and have built a traded book of 11 million tons using third-party purchases, swaps, and our own production to balance our requirements between the Pacific and the Atlantic, as well as improving returns.
In aluminium, we are focused on optimizing our portfolio of products by arbitraging across 3 dimensions: the type of product, the most attractive end markets and customers, and the right geography. For example, value-added products are around half of our portfolio and provide incremental premiums of $240 a ton above remelt. As society expectations evolve, our customers are becoming more and more concerned about how their products are produced. We're working with others in the industry on the Aluminium Stewardship Initiative, leveraging the low CO2 emissions of our Canadian assets to deliver a product that is valued by our customers, including the auto industry. With increased volatility, we have improved our ability to take advantage of relative differences in regions and reroute production to where it best satisfies customer demand.
In titanium dioxide, we are the market leader, and a diversified product suite and product development enables us to meet the needs of customers and optimize our resource base. For example, monetizing previous waste streams with new products such as monazite. By making better use of our data and with more sophisticated tools, we are improving the way we respond to market conditions, tailoring the amount we produce to meet customer needs. Our copper supply is uniquely positioned to take advantage of the two key demand regions of the U.S. and China. KUC is one of only three copper smelters in a region low on concentrate but short smelter capacity. We're leveraging this position to maximize returns through the smelter with use of third-party material, also improving our ability to exploit various arbitrage opportunities. At OT, we have a long concentrate position located next to the world's largest market.
As we move to the next phase of OT's development, we will provide Chinese smelters with a high-quality source of concentrate against the trend of generally declining grade. In summary, the commercial teams are better harnessing the skills, knowledge, and insights we generate from everything we buy, sell, and move around the world. We're centered around our customers and suppliers, bringing better market insights into our operational investment and production decisions. Ultimately, our success as commercial will be determined by two factors, people and data. Securing value at every opportunity means having the right people with the right skills and with the right mindset. The more we do, the more we see opportunities to further improve our business and those of our customers. Thank you. Over to you, Chris.
Good morning, everyone, thank you, Simon. It's great to be here in London to update you on our iron ore business. It's also great to see some glimpses of blue sky, although I might say not quite as bright as the Pilbara at this time of year. Just to remind you what our iron ore business is made up of. We have a fully integrated system of 16 mines, 1,700 km of rail, four ports, and supporting infrastructure. Value over volume drives our business. Our strategy is built on four pillars, which distinguish us in the industry. Firstly, our portfolio of world-class assets, including a resource base of more than 23 billion tons and multiple low-cost development options. Secondly, a highly valued product suite. Our flagship Pilbara Blend is supported by a suite of other products which we place with customers who value them the most.
Thirdly, our fully integrated system, which gives us great capacity or great flexibility to deploy capacity in a way that is consistent with market conditions and is responsive to customer needs. Importantly, people and partners who all play a role in delivering shareholder value. The strength of our business is reflected in its financial results. Between January 2016 and June 2019, we have delivered around AUD 24 billion in free cash flow, an average ROCE of 43%, and importantly, an average FOB EBITDA margin of 68%. Whilst operating costs are important, and we will continue to manage these, our focus is on maximizing margin. On an EBITDA per ton basis, our performance is very strong when compared with our competitors. This is outstanding, consistent financial performance by any measure, equivalent to that of a blue-chip company.
However, I know what really counts is not what we've done, but what we will deliver into the future. As we look ahead, I'm confident that we can continue to deliver outstanding financial performance. We are not complacent and have more to do. Our key areas over the next five years are optimizing our product strategy to meet customer needs, with Pilbara Blend continuing to be our flagship product, sustaining operational excellence, renewal of the existing mine network and assets right across the supply chain, driving productivity with an emphasis on leveraging technology, and progressing project commissioning and study work on mine-sustaining projects to deliver value into the future. As part of this, we are also exploring renewable energy solutions. I will speak to each of these areas in more detail. Let's start with our mines. We are focused on operational excellence after our challenges earlier this year.
In the third quarter, we produced 87.3 million tons with a quarterly average run rate of 347 million tons per annum. This is a significant step up from the first half. We ran the mines at an annualized rate of 360 million tons or more for five weeks during the quarter. I should note that in this quarter, there was lower plant scheduled downtime, and of course, we tend to have good weather in the Pilbara in the third quarter. As you will recall, when faced with weather and operational challenges in the first half, we took immediate action to address the issues and chose to protect Pilbara Blend quality, and we have. Product quality has remained consistently high throughout 2019. We've increased contractors and introduced additional fleet, and these steps are working.
We are tracking well against our estimate of around $18 million spend on recovery actions, we had record, I should say, total material movement in Q3. We will continue to focus on mine development and operational excellence into 2020 and beyond. There has been much talk about our decision to produce SP10. As Simon has already explained, this is a valuable alternate product which contributes to our strong margin, I should note we have sold it periodically since 2014. Producing SP10 supports the high consistency of Pilbara Blend. It also increases resource recovery and improves mine productivity. SP10 is also lower cost than Pilbara Blend on average. It provides us with a greater ability to optimize products for our customers, the overall Pilbara system, and in fact, the entire Rio Tinto iron ore portfolio.
Importantly, we can turn it on or off depending on market conditions. For these reasons, if, in addition to producing Pilbara Blend, we can extract high margin over operating costs from alternate products, we will pursue those opportunities. Regardless of the product we are producing, we know we need to drive performance from our mines to ensure optionality and resilience. This means driving productivity in an environment where some of our costs are increasing. With aging brownfield mines, we face higher work indexes. Our overall 2020 mine work index will increase by 12% over 2019. We also expect our proportion of the low water table mining to increase from around 26% to 33% over the next five years. I will talk more about sustaining our low-cost position shortly. First, let me highlight recent progress on productivity across our mines.
As you can see, the effective utilization of our manned and automated haul trucks continues to improve. The use of our autonomous trucks is delivering significant productivity and cost benefits. By the end of this year, 50% of our truck fleet will be autonomous, and we have a pathway that will see the large majority automated by the end of 2022. High utilization means that 20 autonomous trucks now do the work of 23 manned trucks at 15% lower cost. The reliability of our excavators is also increasing. A key part of the mines is the performance of our fixed plant. We can, and are, making progress with continuous improvement in overall equipment effectiveness as you can see from the reduction in the unscheduled loss from failure of conveyors. Turning now to rail, which also had a strong quarter.
Our average run rate, excluding the period of the recent major rail shut, was 345 million tons per annum, I should say. We had an annualized capacity rate of 360 million tons or more for seven weeks during the quarter. Far, we've unlocked capacity and de-risked our rail network through our productivity and maintenance programs. We are achieving much higher utilization of our installed track than our peers. This means we are making good use of the significant capital cost of this asset. With increased utilization comes increased wear on the network. We have transformed excuse me. We have transformed maintenance tactics, redesigned our rail maintenance organization, and purchased specialist equipment. The major rail shut noted at our half-year results was safely completed in early October. It involved rail assets across an area of 50 km, and we mobilized more than 800 contractors.
Having been on site during the shut, I saw firsthand what a major logistical exercise it was, and the efforts of our team have paid off with a reduction of 20 minutes in cycle time coming from this shut alone. We will continue to focus on rail maintenance in 2020 and beyond. How we do this, for example, whether we undertake further super shuts, will be determined on a whole of system basis. This is one of our advantages of our fully integrated system. The other lever for unlocking rail capacity, of course, is productivity. We have increased consist numbers and train payloads, and we are focused on initiatives to reduce cycle time and increase utilization. AutoHaul has played a significant role in unlocking capacity and reducing rail cycle times. To date, AutoHaul has increased capacity by 10 million tons, and we expect this to increase with further optimization.
It has also reduced cycle times by around 72 minutes. A key measure of rail asset health is the impact of temporary speed restrictions or TSRs. These are applied to areas of track needing repairs, being repaired, and require trains to slow to predetermined speed levels. This impacts cycle time. Since the first quarter, the average TSR impact on cycle time has reduced by 32%. Rio Tinto collects an increasing amount of data as a result of the implementation of AutoHaul and other technology across the rail network. You might be surprised to hear we collect some 90 gigabytes of data each day. A multidisciplinary team has been put together to help us apply that rail data. The team uses artificial intelligence and a technique known as random forests to prioritize rail maintenance, and the results are impressive.
We are currently predicting the optimum removal of existing defects to greater than 90% accuracy. Even more impressive is the ability to predict the location of future defects before they manifest, to greater than 80% accuracy. This work enables us to adopt a preventative approach to maintenance. It unlocks value by proactively repairing track and prioritizing our resources for maximum impact. Productivity improvements, including the use of data, will be key in an environment of increasing cost pressure in 2020. Let me cover them. There is a significant project pipeline in Western Australia, which is leading to a market tightening for some specialist contractor and technical skills. Our ongoing maintenance requirements for assets are increasing. The longer haul distances, the low water table mining, and further development of brownfield pits are pushing up the mine work index.
There will also be additional costs needed for the exploration, evaluation, and approvals work required to support major renewals, which I'll come to later. On the other hand, we are seeing considerable cost benefits from our productivity improvements, some of which I've already covered. We will extend our automation program and have a strong pipeline of numerous productivity initiatives, including the ever-increasing use of data. Given our cost base is highly geared to the Australian dollar, we are also experiencing relief through foreign exchange. Of course, we will work hard on cost performance regardless of the headwinds. Our primary focus is, as I've said, on margin and value. In addition to driving the productivity of our assets, future developments are critical to maintaining our production rate and providing options for future capacity. We have a number of projects in execution.
Koodaideri phase one will have an annual capacity of 43 million tons. The mine will make an important contribution to the Pilbara Blend and, subject to market conditions, will increase the lump-to-fines ratio of the entire portfolio from the current average of 35% to around 38%. It is expected to deliver an internal rate of return of 20% at around a capital intensity of $60 per ton of annual capacity. This is highly competitive for a new mine with additional infrastructure such as a rail spur, airport, and camp. Koodaideri is progressing to plan with first ore expected in late 2021. Our Robe River sustaining projects, West Angelas and Robe Valley, are also on track, again, for first ore in 2021. West Angelas and Mesas B and C have now received all approvals and construction has started. Mesa H has experienced some environmental approval delays.
Sensitivities around water drawdown mean that it is appropriately receiving careful consideration by regulators. As Robe Valley is developed as a hub, there are mitigation options. Mesa H is a good example of the increasing complexity of the approvals in the Pilbara. More projects are developed close to environmentally sensitive areas. More development is below water table, and the cumulative impacts of multiple projects requires consideration. We are working closely with all regulatory agencies. We have a large amount of renewal ahead. We have six major projects in the conceptual or order of magnitude phase, and a further eight in the pre-feasibility and feasibility stage. This is on par with the number we had during 2010. We have a number of brownfield developments at or below a capital intensity of AUD 30 per ton. These multiple developments make use of our existing operating hubs, leveraging infrastructure, reducing capital, and creating optionality.
In addition to a large number of projects in study, we also have great flexibility within potential projects. For example, we have multiple project scopes currently under study for Koodaideri Phase 2. There is a potential to increase the capacity for the Koodaideri hub to 70 million tons and beyond. It has various options. For example, wet, dry, or concentrator processing. These options, of course, will have varying capital intensities and will be designed to maximize overall value. We're also looking at how we can reduce emissions, including options for renewable energy. Year to date, we have also undertaken around 650 km of drilling to assist in maintaining our eight to 10 years of reserve cover. We have resources to continue to underpin production of Pilbara Blend for decades. Now looking ahead. Let me turn to system outlook and guidance.
First and foremost, our objective is to optimize the entire system end to end. As mentioned earlier, on a short-term basis, we are capable of running both our mines and rail at rates consistent with our port nameplate capacity of 360 million tons. To achieve this on a consistent basis across the system, taking into account weather, maintenance cycles, and materials handling variability, greater capacity and resilience in the entire system is needed. A step change in capacity will come with Gudai-Darri phase 1. With its high output mine, dry processing plant, rail arrangement, and proximity to the port, Gudai-Darri will provide 360 million tons per annum of system capacity once fully commissioned. As I mentioned earlier, we have various deployable capacity options under study for Gudai-Darri phase 2. Ultimately, the capacity of the Gudai-Darri hub could be 70 million tons.
Of course, volumes will always be set by balancing various factors, including market demand, quality, production cost, and capital. Any decision will be made with value over volume in mind. Our ultimate aim is always to meet customer needs and optimize EBITDA margin and cash flow. Our 2019 production guidance remains at 320-330 million tons, and our cost guidance of between $14 and $15 per ton. In terms of shipments for next year, we believe that we can achieve an increase of up to 5% on 2019 guidance. We will come back to you in mid-January with specific production ranges and cost guidance for 2020. Our 2020-2022 guidance for sustaining capital is between $1 billion and $1.5 billion, as compared to existing guidance of around $1 billion per year.
The need for increasing sustaining capital is due to continued automation of equipment, expanding our HME fleet to manage increased workload, as well as, of course, replacement of equipment installed as part of original investments. Ongoing maintenance and replacement of assets across the integrated system. For example, the stackers at East Intercourse Island and our Tom Price concentrator structural upgrade, as well as accommodation. Finally, and importantly, IS&T upgrades integral to supporting our ongoing digital transformation to drive productivity. Our iron ore business is well-positioned to continue to deliver superior value to our shareholders. As you've heard, we are taking a number of steps to further optimize and improve. For us, continued delivery of outstanding margin and ROCE is key. Our focus on operational excellence in the mines is delivering results. I promise you, the entire team is very focused on strong operational performance across the entire system.
We will always work hard to ensure that our assets deliver their full potential, but not at the expense of asset health. Our focus on productivity is delivering good results, this will continue in an environment of increasing cost headwinds. This will include further leveraging technology, especially the use of data. We have an extensive resource portfolio and a strong project pipeline of replacement mines. This will enable us to underpin the production of Pilbara Blend well into the future. It also gives us greater flexibility in our development sequence. Our brownfield options are focused around operating hubs, which means that we can use existing infrastructure and reduce capital. A step change in system capacity will come with Koodaideri. We know that value comes from a system that is flexible to respond to external conditions, delivering the right product to the right customer at the right time.
By continuing to seek to generate and prioritize value in all that we do, we are laying foundations for ongoing success throughout the cycle. Thank you. Back to you, Menno.
Great. The speakers will sit up front here, please, and then we'll start the first Q&A session. Jakob and J.S. as well, please. We're a little bit ahead of schedule, which is good news because that means there's slightly more time to look at those very exciting virtual reality goggles and the Mine Automation System. We're sticking to 30 minutes. A couple of things on process, please. For those of you on the phone, the operator will take you through the procedure now. For those of you following us by webcast, you can also pose questions via email. The email address is at the bottom left of that first page. The speakers here up front are joined by Jakob, our CFO. We also have Bold Baatar, Chief Executive of Energy and Minerals, ready to take questions.
Alf Barrios, Chief Executive, Aluminium, and Simone Niven, our Group Executive of Corporate Relations. Now, this is one very important point. Please limit yourself to questions on the presentations that you've heard on aluminum, mineral, and corporate and government relations. Questions on copper, on technology, on OT, and on very broad financial subjects will probably be discussed in the presentations afterwards. As you can imagine, I'm going to enforce that rule very, very tightly.
Start now.
Jason, you are warned. On that note, Jason. Please. Oh, sorry, one question. One minute for the mic. Yeah.
Actually, it's a question for Vivek. I was fascinated that you think we need new greenfield iron ore mines. Could you talk a little bit more about that?
Sure.
Yeah.
Yeah, look, depletion in this industry is going to happen at quite a pace. For example, we've got about 1.9 billion tons of production of consumption taking place today. That's 1.9 billion tons of less reserves in the ground every year. Let's say for the next 10 years. That's 19 billion tons fewer reserves, and that has to be replaced. Over time, of course, when we look at the project pipeline that's out there, and when I showed you the Woodmac project pipeline, for example, you'll need to start to see some greenfield come in.
Just to push you a little bit on this. Greenfield mine's fine, but do we need new greenfield systems, or do you think the existing systems can supply?
That's a question I think that we'll need to see pan out. For example, Rio Tinto, Vale, BHP, FMG have very substantial projects available, and there is, of course, Africa. There is enormous scope. I think what you have to look at is the range of costs in that system. You start off with some development costs down at, say, less than AUD 100 a ton, all the way up to AUD 200 a ton for those systems, when you need to develop an entirely new system or open a new greenfield mine in a part of an existing system and develop a lot of extra infrastructure. Where that will go in the future will depend entirely on the amount of demand that's out there.
It's very hard to predict precisely, but yeah, if there's a lot of demand, then you'll see prices a lot higher, and that will incentivize some of those higher-cost systems to come in.
Okay, thanks.
Dominic.
Morning. Dominic, J.P. Morgan. Two questions on some of the iron ore commentary and guidance. If we could dig into some of the details on higher costs we're seeing in OpEx and sustaining CapEx. To what extent is that permanent, specifically around the sustaining CapEx? Is it related to this ongoing higher intensity of maintenance that you mentioned? The second question is around how do you consider a 70 million ton phase 2 for Koodaideri in the context of port capacity? Should we think about an increase in port capacity, or does it fit within the overall value of a volume strategy?
You want that?
Yeah.
You got this?
Yeah. Okay. Thank you. Look, in terms of the sustaining capital and the ongoing nature of it, we are approaching a phase where we have a generation of assets that needs renewing. I wouldn't say that's permanent, but for the next couple of years, we have increased guidance from existing guidance of around $1 billion-$1.5 billion. We'll continue to update that guidance as we go forward. It is certainly cyclic in nature. As I said, if you think about some of our assets, the Tom Price concentrator has been with us for 50 years. It needs a bit of a birthday. You need to think it in that terms. Sorry, second question was on.
Phase 2 Koodaideri.
Phase 2 Koodaideri. Sorry. Yes. Look, phase 2, we do have options for deployment up to 70 million tons. We'll work that through. In terms of, though, is that additional capacity, you should think about that potentially as part of it will be sustaining because we have other mines expiring and some of it would be part of it. We haven't yet determined exactly what scope we'll go to. That will be a value over volume decision. I think really what I was trying to magnify there is how important the Koodaideri hub and the flexibility of that hub once established gives us and will be a very long-term asset for the future with lots of options.
Specifically, does that phase II also consider a potential increase in port capacity?
Yes. Well, look, the port capacity, we have actually run the ports already at 360 million tons. We think there is actually optionality for greater through some minor debottlenecking, but we're not studying additional port capacity today.
We'll take one more in the room before we go to the phones. Liam. Mike.
Right here.
Morning. Liam Fitzpatrick from Deutsche Bank. I have two questions on iron ore, one on the market and then one on the operations. Just on the market, it's been the one area of upside surprise in terms of Chinese demand. Do you believe the figures? If not, what do you think the real demand is for this year? Can you give us an outlook for 2020? Just on operations on that replacement CapEx intensity, what does it look like after 2022? Is $1.5 billion-$2 billion the go-forward rate for that? Thank you.
Thanks for the question. Firstly, on the market, as you say, 2019, we've seen very strong underlying steel demand, and that's translated into a very solid pricing environment through this year. Look, we're continuing to see that into the back end of the year. You've seen some moderation, obviously, in prices. Some of that supply side has normalized. Conditions in China in terms of underlying steel demand remain strong. We're not seeing unusual build-up in stocks and certainly engaging with customers. We continue to see that demand picture supported into the back end of the year. I'm not going to give a forecast in terms of pricing, et cetera, into next year. Certainly, in terms of our business and our engagement with the customer, we're continuing to see that underlying demand remain strong.
There's some moderation in some sectors and in other sectors, we're getting some of the, I guess, the tailwinds of some of that stimulus and additional both infrastructure and construction projects start to come through. Into next year, we're continuing to see pretty solid demand. Certainly, that's the feed we're getting from the market and our customers.
If I may add, I think what is important for us is, I think Vivek explained very well, there will be increase of scrap usage in China and so on and so forth. There is a level of uncertainty about what demand is going to be in China going forward, and we fully acknowledge it. What is important for us to make sure we have the best product, the best quality of service, the best relationship with customer, to make sure that if the market was to soften, we could protect or even increase our share wallet in China, protect our market share, and so on and so forth. That's why we have started to change, I think, as Simon gave us your example, inventory at the port.
We're doing some blending, we're doing some partnerships with some customers to make sure that under any kind of market conditions, we will be very well-placed and extract full value from our product into China and so on and so forth. At the end of the day, for us, it's going to be about having the best product, the best relationship with the customer, the best supply chain, including with some of the small mills in the north of China, in order to make sure that we maximize the value of our production out of the Pilbara or out of Canada going forward. Did you give the example? About IOC the blending you've done or not? Yeah, I think you should tell, because that's a good example of what we're doing. Sorry, Simon.
Yeah. I touched on the presentation, but one of the things we're trialing, and it is a trial, is that blend of SP10 and IOC. Look, I think the broader point is just with a diversified product portfolio and really deep relationships with customers, it's about the optionality that you can build within your book. Clearly, regardless of what the demand levels are, customers' businesses continue to evolve, and we need to evolve our business, as well. By furthering that understanding with customers, engaging with them in different ways, like some of the technology examples that I spoke about earlier, you can really make sure that we're setting up our business to really meet their needs in different ways. As I said, that generates value for us and for them.
Was there a second part of your question, Liam, on capital intensity?
Just replacement CapEx beyond 2022. Is $1.5 billion-$2 billion the new number?
Yes. You're asking for guidance beyond our guidance. Basically, we are not giving guidance beyond 2022, what I can tell you is that we're not aware of any material changes further out there. Obviously, those CapEx are subject to individual investments. The sustained CapEx is the more stable factor.
Yeah, Jakob, just a comment. You also need to think we have got some renewal ahead. When you do renewal, you build a new mine, you actually get the trucks and the assets, if you like, for free. Then that gives you another 10 years of asset life. It is quite cyclic.
I'm not sure for free is the right word. You and I may have a discussion at the coffee break. Forget what he said here.
Localism.
Any questions from the call? Okay. Operator, please. Question from the call.
Yes. Thank you. As a reminder, if you wish to ask a question, please press star and one. The first question is coming from the line of Ian Rossouw from Barclays. Please go ahead.
Hi, guys. Morning. Just a question on this SP10 product. If you can give us a sense of your expectations of the 2020 volumes of that. Maybe just to give us a sense of what the sort of EBITDA per ton, the slide you show that sort of average for first half of 2019, what that would be for SP10 on its own. Just a second question. Just sort of curious on the slide you talk about electrification. You obviously list nickel and cobalt as part of that, and you sort of saying you're looking at opportunities to explore in this market. I was just curious, does that include sort of nickel and cobalt options? Maybe just an update on what's going on with Jadar, if you can.
All right. Maybe we're going to start with the second part of the question, Bold. If you can pick it up, this one, and then we'll come back to yourself, Simon, and Chris, if that's okay.
Thank you. Thank you for the question. As part of Rio Tinto Ventures, we are evaluating battery materials and are screening opportunities out in the market. It does include nickel. Obviously, it's very difficult to find. At the end of the day, it is about creating value. We need to make sure that all the projects meet our return threshold. We're out there, we're screening them, but at the moment, we're not pursuing any aggressive acquisitions. Secondly, we're the pre-feasibility study, which means that we are studying a range of options. We're going to be moving into a feasibility study in the second half of next year, at which point after that, the board will decide whether to invest in the project, after the feasibility study is complete.
Thank you. Bold?
In relation to the SP10 question, it is a product we introduced in 2014. We have sold it periodically through that period. It is one of the products that we're placing through our port side trading capability, in fact, reaching out to customers that otherwise aren't customers of our iron ore business. That gives us some additional optionality within our book. We are targeting those customers, particularly in the north of China, that are less sensitive to phosphorus levels. It is a product that we bring into the market and take out of the market, depending both on market conditions, and also on operational factors. It gives us additional optionality there. We don't give forward guidance in terms of the volume of particular products.
You would have seen we've done a little bit over 10 million tons year to date of SP10. We'll continue to place that into the market, depending on market conditions, and where we see value for it, both for ourselves and for our customers.
Maybe just to pressure on that. Chris was saying it obviously helps to sustain the Pilbara Blend. Is that the implication that over time you will probably see that share of the SP10 grow and Pilbara Blend decline over time?
I think the point Chris was really underlying was Pilbara Blend is the flagship product in the customer. It's the baseload of the China steel industry, and we've deliberately set it up so that it meets the average mill requirements in terms of its specification. There's great value in making sure that we maintain both the quality and the consistency of Pilbara Blend, into the future. That benefits our customer, and it benefits us. I think that's the point that Chris was really underlying, was just the value of the Pilbara Blend product.
Okay. Again, from the web.
Here we have two. One on the recent MoU we signed with Baosteel, Chinese university. Can you talk us through a bit more about what that's about? Should we expect more similar partnerships with Chinese SOEs or others going forward?
Simone, if you can pick it up. You were in the picture assigning the-
MoU, you should know all details.
Good morning, everyone. Thanks very much for the question. Look, we're really excited about the partnership. It is a breakthrough partnership between China Baowu Steel Group, who is our largest iron ore customer, also Tsinghua University. I'm not sure how many people in the room know of Tsinghua. 50,000 students are in Tsinghua, leaders in STEM, but also most importantly in the context of this partnership, they're also leaders in climate change. They're experts in climate change research, but also policy. Us joining forces with Tsinghua, China Baowu Steel Group, and ourselves, but also importantly, CISA are also sponsoring, which is the peak industry association, Steel Industry Association in China. Breakthrough opportunity for us. It's early days. We just signed the MoU in September.
We have a dinner on Monday.
We have a dinner with them on Monday and a steering committee on Tuesday. The opportunity really is to look at sharing the technology opportunities, really look at carbon reduction across our supply chain, which is one of the key parts of our climate change strategy overall. Great opportunity for us. As I said, early days and we're hoping to advance that in the next few weeks.
Yeah. Especially what we want is to connect the dots between the iron ore in the Pilbara or in Canada and potentially the automaker in China. That's what we're trying to do, is across the entire value chain. It's not only about emissions, but the entire environmental footprint of the system. Okay. That's what we are looking at. We know it's going to be a combination of implementing existing technology, developing new technology and the policy framework. Maybe just to explain how Tsinghua is important in China is some people would regard it as the Harvard of China. That's what it is. If you were to do any kind of benchmarking, you'll see most of the leaders in China are coming from Tsinghua, and so on and so forth.
For us to be able to get Tsinghua on the MoU was a massive breakthrough that give us some confidence that we have a strong position in China that we need to the next phase. You have another one or?
Yeah, there's one more. You recently announced a strategic review of your smelter in New Zealand. Can you give us an update on your plans there?
The review is underway. Alf, I hope.
It's on the way. Good morning. Thank you for the question. Yes, last week we announced that we were putting our asset in New Zealand strategic review. Before I go into a bit more detail, I just want to say that it is an asset that is very well run, as our Pacific assets are. It's one which has operational metrics which are among the best in the world, and it produces low carbon, high purity aluminium. Unfortunately, it lacks internationally competitive, both power and transmission costs. We've been working now after the announcement with both the government and the power supply to find pathways towards making the smelter a viable ongoing business. The review will include all options. We're looking at curtailment and closure as well. The review will be completed by the end of Q1.
I'm not going to speculate on the outcome of the review, but I must say that the current situation is not sustainable.
Let's be clear, we took with Alf and the team this decision very seriously. We will fight hard to protect this asset. Maybe some of you have some Apple Watch I don't have, but if you have an Apple Watch, you may have some aluminum coming from NZAS. I'm not joking. There are people, lots of people, lots of communities relying on this one. With Alf, with Kellie Parker in Australia, we will fight hard, very hard to find a sustainable solution to this problem. We have, as Alf said very nicely, is we have a problem and we are working hard on this one. Doug.
Back there.
Thanks very much. It is Doug Upton with the Capital Group. I have a question for Vivek, if I could. Steel demand growth in China looks like it is somewhere in the 5%-10% range this year, depending upon which figures you want to believe. If you believe the 5%, last year was even stronger than the 10% that we kind of mostly have in mind. Steel has been very strong
Yet copper and aluminium demand has been slowing. This year you've got maybe steel is 8%, copper and aluminium 1% or 2%, so directionally that's unusual and the gap is very unusual. I'm interested in what you think might be going on there.
Okay. Chris, steel production in China approaching 1 billion tons. I think you may have heard somebody say that in the past. Don't know.
I don't know who that is, by the way.
There are two parts to that. Obviously, Simon did talk about the growth in demand in the construction sector and the infrastructure sector. That's one important factor. The other important factor is that, and this is perhaps a slightly more data-oriented factor, is that the reforms that have taken place in China have brought a lot of production that was once, let's call it off the books or unreported. It's now reported because that capacity has now been transformed, shut down in many instances, and has been taken over by legal producers. It's now moved into the reported category. The increase in Chinese crude steel production is partly this phenomenon that Simon mentioned, but also partly a data factor, which has led to an increase in the underlying level of reported crude steel production.
The question you're asking is as well about why are we seeing increases in steel but not in the other. The first part of it was this data issue. There is a point about the baseline that's important. The second part is that we've seen a lot of construction taking place. The first phase of it is very steel intensive. It's the subsequent phases that are more copper intensive as the lines go in. At the same time, the aluminum windows go in, the tiles go onto the wall, which requires titanium dioxide. We would expect to see some of that demand come through a bit later on.
I think, Simon, you may say a few words about the impact of the automotive industry on aluminium, so that we have been impacted across all geographies. We see it in through all the books as well, which should explain the difference between iron ore and Doug and aluminium and copper as well.
Yeah, sure. Doug, in relation to autos, as touched on, we've really seen that soft both in China but also elsewhere. A bit of a mixture of underlying demand conditions, but also specific policy responses of some of those subsidies and other government measures had been withdrawn. That's exacerbated what was a bit of a cyclical slowdown anyway. We'll also see a bit of a recovery, both as some of that policy measures go back in and purchases begin to resume. That'll have a bit of upward pressure, both in terms of aluminum, but also the other commodities that go into it as well. Clearly, that underlying trend around lightweighting, and the greater use of aluminum in auto continues.
It's certainly one of the things we're really focused on in our aluminium business and with VAP, is around really making sure that we continue to meet those customer needs and innovate in terms of our product to be able to place with those customers.
Next question. Myles, please. Yeah. Just in here. Can you do for me? Just a moment. He's behind you, yeah.
Great. Yeah. Myles Allsop, UBS . A couple of questions. Maybe just for Chris and iron ore, just going back to January this year before the disruption, you had guidance for 2019 of, was it 338 to 353? It feels that the guidance for 2020 is not fully recovering the disruption that we've seen this year. I was just wondering, is it that there are still some lingering issues that are holding back the iron ore shipments? Obviously the SP10 coming to the fore, which hasn't really been talked about before. Is it more value over volume and a kind of view on where prices are, sort of volumes are likely to be next year?
Maybe secondly, for Vivek as well, I was thinking about if we are in an oversupplied iron ore market next year, you mentioned about sort of Chinese domestic producers and their cash costs sort of supporting the price. Just a sense as to where you think that support would kick in.
Okay. Myles, thanks for the question. If you remember, take a step back, we also had a target of achieving 360 million ton run rate through the system by the end of this year, in fact. What I experienced after the recovery for the first half, we had a very strong third quarter. We were able to push the system reasonably hard and just really test that assumption. Despite the fact we have peaking capacity around 360 million tons, we do have a tail of production that's lower than that. Where is the main source of that constraint? It's really through the plants themselves, and that's just a variety of factors, cyclic maintenance, material handling variability, and that's what's really caused us to reassess the capacity. It's not a lingering issue actually from the past. It's actually just as our ability to test the system.
Look, we haven't stopped trying. Of course, we in fact crept capacity through the plants this year, and I gave some examples, a couple of million tons. We'll continue to creep productivity. We've got a record of doing that, and our guidance next year, we flagged that up to 5% increase on our existing guidance, 320 million-330 million tons from the midpoint. Really what we're now saying to achieve that step change in capacity.
We think we're going to need an additional change, and that will come with Gudai-Darri. We're already building the plant, if you like. We don't see a lot of point sinking a lot of capital into de-bottleneck plants. I'm sure if there's minor capital, we'll look at those things on their merits, but that's really the thinking behind all of this.
Chris, do you want to say a few words about dry versus wet products?
Yeah. Of course. As we continue to test our plants with more and more wet material, in particular, and we do find the nature of the variability of the low water table mining material is one of the things that we need to be able to increase the robustness of our plants to manage. At times, we can run very, very fast rates, but then you'll go to an area, say, where there's a lot of clay, and it will slow the system down. That's part of that tail that we're seeing. Some of the work, and actually Steve will talk to some examples later about the work we're doing to use data and so on to improve that. We still think we're going to need a step change to achieve the 360 nameplate.
Thank you, Chris. Vivek. You can't give a price, by the way, Vivek. Hello.
No, look, I don't think we said the market is going to be oversupplied next year. That's certainly not something we're flagging. Look, I think you are right. It's the Chinese high-cost producers that will ultimately set the marginal cost and along with, of course, there's some high-cost Australian producers and others. Where the market ultimately sits will depend on a whole range of factors, on certain macroeconomic conditions, and on a whole range of supply-side factors. As J.S. said, I think you question very cunningly. The question of what is next year's price? I don't think we can comment on that.
Can you say, I don't know if it's Vivek or someone, do you want to say a few words on how we see the capacity in China reacting this year? Because we've seen a pickup.
Yeah, sure. We've seen around about 20, 25 million tons in terms of that domestic capacity. I think one of the interesting points this year has been it being a bit less responsive than we've seen in the past, and that's a range of different factors. Obviously, some of the environmental restrictions, some of the permitting points, I think have flowed into that. As we look forward, and as I think Vivek touched on in his presentation, longer term, some of those costs continue to elevate, particularly as some of those mines turn underground. That really goes to your long-term positioning in the market, and obviously as an incumbent producer on the left-hand side of the cost curve. That's going to drive pretty healthy margins as we go forward.
I think that's a very important point. In the current annual prices we enjoy in Q2, lots of people would have expected a significant increase of production in China. Remember, if you go back five years ago, they did produce up to 400 million tons. We didn't see it. That give us a sense of the cost position of most of those mines, who couldn't even with a pretty favorable annual pricing environment, couldn't restart and so on and so forth above and beyond the environmental issues.
Great. Thank you very much. This rounds up the first Q&A session. Don't worry, there's another half an hour afterwards. I know who wants to ask a question, who hasn't. You're not going to be forgotten. 10:05 back here. Please, please, virtual reality goggles and Mine Automation System. Have a look.
Okay.
It's a pleasure to introduce Stephen McIntosh, Group Executive, Growth & Innovation, with some fascinating insights into what we are doing in the coming period.
Thank you, Menno, welcome back, everyone. I hope you had a chance to look at some of the tech that we had on display. Rio Tinto may have one of the oldest heritages in the mining industry, it's our ambition to lead the sector in adopting new technologies to deliver the most efficient, productive operations. As an industry, we tackle some of the greatest engineering challenges on the planet. Rio Tinto has a long and proud history of meeting those challenges head-on. Our ongoing success will depend on us embracing technology and data to find new and more efficient ways of mining at lower cost. We began this technology journey over two decades ago with automation, where we now lead the industry. Today, our strategy and our aspirations stretch far beyond automation through value chain optimization and the emerging areas of data science and artificial intelligence.
At the core of that strategy is the view that technology is a potential disruptor as well as an opportunity in areas such as digital and energy transitions, to name but two. To be clear, we plan to lead. How are we developing our technology to give real competitive advantage? We actively scan many industries looking for insights and opportunities to replicate or adapt. We have a long-term commitment to development of proprietary technologies and spend more than AUD 200 million each year on very early-stage R&D. We spend far more each year on the scaling and deployment of new tech into our business. We have a strong innovation culture, world-class technical talent, and an extensive partner ecosystem that we engage with. We've established centers of excellence in areas such as surface mining and underground mining, processing, automation, ore body knowledge, and energy and climate change.
These centers provide centralized technical assurance, service, and support to our operations and are a pathway to nurture and grow our technical talent. In exploration, we've taken an advanced approach to data and technology to improve our targeting, enabling us to uncover opportunities in areas that have been well explored by others. In studies and construction, innovation and digital design is helping us to achieve improvements in safety and costs as seen at Amrun. We're also looking at more agile ways to build new mines, starting smaller, building quickly and safely with embedded optionality for growth. Technology has also an important role to play in helping us tackle critical industry challenges, such as tailings, energy and carbon reduction initiatives.
Technology and innovation achievements can at times provide easy headlines, but being successful in the execution and integration of new tech at scale is difficult, and that's where our industry-leading track record sets us apart. We've been using automation at scale in the Pilbara for over a decade. By the end of this year, as Chris has noted, we will have 183 autonomous trucks and 26 autonomous drills in service globally. Our autonomous drill fleet is the largest in the world and set to grow further. Our automated trucks operate at 15% lower cost than an equivalent manned one, and our drills have unlocked a 25% increase in productivity and a 40% improvement in equipment utilization. Today, our flagship remote operation center in Perth is the central nervous system of our Pilbara business.
As well as using technology for better performance, we've been working on new ways to reduce our carbon footprints. For decades, we've been pioneers in technology developments in the aluminium sector. Our wholly owned AP Technology have the lowest emission levels in the industry. The AP60 platform delivers 40% more metal per pot at a lower cost than any previous smelting technology. We're not resting on our laurels. We're actively pursuing greenhouse gas-free aluminium through our ELYSIS partnership with Alcoa, Apple, and the governments of Canada and Quebec. Rio Tinto also has been a bit nerdy in something that we're proud of. Before data science was the buzzword it is today, we'd already moved beyond traditional programming into this area. We started in early 2014 when we deployed our first machine learning-based tool called Predicta, which optimizes predictive maintenance on critical assets.
We've now invested in a world-class data science function, and I'll run through some examples on what has been delivered so far. Another component of our digital play is our open data environment, which allows us to expand our use of modern analytic technologies. This fit-for-purpose platform gives us the ability to bring the best minds in the markets to our doorstep. It will drive improvements in terms of both scale and time to market for new digital deployments. All of this is only the beginning of what can be achieved and a small insight into our technology strategy and ambitions. Digital transformation will be a game changer for the industry, and we're very well-positioned. There is enormous value to be had in moving beyond automation to digitize and then integrate across the entire value chain.
To do this, we need to redefine what a future mining operation could look like when every step of the value chain is connected and optimized in real-time. We're very nearly there thanks to the technology foundations we laid in the mid-2000s when we began working on a system called MAS, the Mine Automation System, which some of you got to look at earlier. MAS captures our mining and technical domain knowledge and combines it with a broad array of operational and ore body data to provide us with insights to improve our business. The intelligence we get from MAS includes real-time operational insights. We track nearly everything across our mines, down to the content of every bucket of material, to stockpile models, to blending parameters, this allows us real-time ore body optimization at scale.
We are also digitizing our business and putting tools in the hands of our people and our customers. Let me give you a few examples. TrueView, a mobile app which provides frontline operators with real-time decision support. Paperless Maintainer, an efficient digital workflow to ensure our teams are spending more time on the tools. Pioneer Portal, a new collaboration platform that helps us leverage the best minds to solve critical challenges. Edison, our own Google that helps us navigate our vast internal knowledge base. Open data environment was mentioned earlier, portside trading was mentioned by Simon. We're taking productivity to the next level through the application of data analytics and artificial intelligence. At Kennecott, we're using a predictive model to increase copper recoveries. Normally, operators rely on a daily view of ore quality coming into the concentrator to set the reagent dosing strategy.
The problem with this is when the ore feed differs from the daily plan, the strategy does not match, and we miss either yield or throughput. Through the development of a machine learning model, we can now accurately set the optimal dosing strategy in real time using over 4 million data points. We're also using data science to improve materials handling, as Chris mentioned earlier, and identify ore that is difficult to process before it gets to the crusher. This information is fed in real time to the operators, enabling them to design blending tactics to reduce downtime in the plant and at the port. We've seen a 40% reduction in material handling problems in the processing plants at our pilot site in the Pilbara. The final example is from our Port Alfred operation in the Saguenay, where we're deploying an analytical model to better predict ship arrival times.
We expect this to reduce demurrage costs by 20%. All three examples have the potential, and actually are being replicated across our global business. Let me turn to exploration. Exploration is a competitive advantage for us and does differentiate us from our peers. We have invested significantly in our people and in R&D, developing a range of proprietary tools to accelerate discovery. We take a sophisticated approach to data, combining public and proprietary data with advances in assessment techniques to improve our targeting. Our copper-gold discovery at Winu in Western Australia proves this capability. It's an area that has been extensively explored by others, but Winu was hidden under 60-100 meters of cover. This is one of those rare and exhilarating stories where the very first drill hole was the discovery hole.
There's still much to be done to understand exactly what we have at Winu, but we're encouraged by the results to date. As I mentioned earlier, at Winu, we're taking a different approach to development and assessing a smaller starter case as an option to improve our agility in executing growth. At our Falcon project in Saskatchewan, we're working with Star Diamond Corporation to reevaluate a known cluster of kimberlites using a custom-built modified Bauer trench cutter. Previous drilling methods caused breakage of larger stones and likely under-sampled diamonds. This is a great example of adapting existing technology for a novel use. This is no small drill rig. Given the importance of recovering the large, high-quality diamonds intact, we had to go for something that could deliver very large samples to a depth of around 250 meters.
We've made good progress with trenching and are assembling the bulk sampling plant, which will operate during the Canadian winter. This is still an early-stage exploration project, and we will continue to work through 2020. The pioneering spirit is strong at Rio Tinto, and our people excel in applying innovation across all parts of our business, including our new projects. The modular construction of the 1-kilometer-long Chith export facility in Amrun, using 13 prefabricated pieces, was the first of its type in the world. It was constructed safer and faster than by traditional construction methods and delivered improved capital intensity. In recognition of this innovation, the Chith wharf has won multiple awards, including the internationally recognized Brunel Award here in the U.K. Even more importantly, this construction method eliminated more than 300,000 high-risk work hours, and we only recorded a single injury during construction.
To innovate, you need the right people to constantly challenge conventional practices. Nothing is sacred. One project we have in development is seeking to decouple the activities of dig units and haul trucks. The fully mobile surge loader is being piloted at Kennecott. This unit breaks the dependency between diggers and haul trucks by putting a surge bin with a conveyor in between. Instead of diggers waiting for haul trucks to get in place for the next bucket of material, the dig unit can continuously load the surge bin, and the trucks can drive by and be loaded autonomously by the conveyor. In addition to the safety benefits of separating trucks and diggers, we expect a 50% increase in digger productivity and payload accuracy of 98% or better. Our U.S. Borax business in California has been at the forefront of innovation since it was founded.
It produces around one-third of the world's supply of refined borate products used in glass manufacturing, ceramics, and agriculture. I'll talk to two examples of the latest innovations at U.S. Borax. First, our teams are trialing an approach where they're combining recovered tailings with mined ore to deliver a blended feed into the processing plant. To date, this trial has shown a 7% increase in overall recoveries. Second, as we announced last week, we're also working to further develop our pilot to generate battery-grade lithium carbonates from existing waste rock at U.S. Borax. The ore type the team has been testing is challenging to process, so we're very pleased the initial small-scale trial has delivered positive results. We're now running a pilot plant focused on further trials to optimize the process. We've seen a structural shift in our industry with much greater expectations from society and a new sustainability age.
We have an opportunity to use technology for more than safe, productive, environmentally efficient operations. We can use it to tackle critical industry challenges. The Brumadinho dam failure in Brazil was a human and environmental tragedy and a stark reminder that all industry players must have a well-structured approach to managing major hazard risks. At Rio Tinto, we introduced a global tailing standard in 2015. Today this is supported by a team of leading technical experts and our Surface Mining Center of Excellence. We're actively seeking new and better ways to minimize waste from operations and to improve management of tailings. One possibility we're looking at is for the potential reuse of material by reprocessing tailings to recover byproducts, as we saw in the U.S. Borax example. We're also looking at limiting water in tailings. At our Vaudreuil alumina refinery, we're investing AUD 188 million in filter press technology.
This will remove the water from the tailings to produce a safer and more stable waste product. Water use and recycling is another area where we have made great progress. We use water to process ore, to manage dust, for drinking, and in some cases, to generate hydroelectric power. We have a set of water targets for the business that focus on improving performance, and we use technology to help minimize our consumption and maximize recycling. At Oyu Tolgoi in Mongolia, we have invested in innovative recycling and conservation processes. We currently recycle more than 80% of the water we use at OT, which is more than double the global average of similar copper mines. We're also committed to decarbonizing our business, reduce our Scope 1 emissions.
At our Diavik mine in Canada, about 15% of the power comes from wind. We use solar power at our Weipa and Amrun operations in Australia. In Canada, we exclusively use hydropower for our aluminum business. We're also working with our OEM suppliers to identify opportunities for hybrid and electric equipment. Finally, since 2018, we've been working with our partners, Alcoa, Apple, and the governments of Canada and Quebec on ELYSIS. I've shared how we're using technology and data to improve our business today and our ambition to continue to pioneer for tomorrow. With Gudai-Darri in development, our goal of a more digital future is almost a reality. An intelligent mine that goes beyond automation to enable an agile, integrated, and optimized value chain. We're focused on evolving our business to master the challenges and opportunities ahead.
We're very well positioned to continue to lead the next frontier in mining. Thank you. Okay. Arnaud.
Thank you, Steve. Hi, everyone. I'm going to discuss today about Oyu Tolgoi, our copper business in Mongolia. I will give you an overview of the strengths, the performance, and the future challenges of the business, and then Steve will talk about the underground project. Let's start by considering OT's many strengths. There are five key points to emphasize those strengths. First of all, OT is a world-class ore body, a high-grade copper and gold deposit supporting a multi-decade life of mine. Secondly, the operation is well situated to come into full production just as copper demand is predicted to reach a high over the decade, driven by electric vehicles and related infrastructure. The operation is close to a key customer market in China.
The underground development will unlock 80% of the value of the project, and we are seeing a solid current performance from the open pit with outstanding safety and production. In the third quarter of 2019, Oyu Tolgoi achieved an all-injury frequency rate of 0.18 per 200,000 hours of work. This is the lowest rate across all of our copper and diamonds operations and is industry-leading in many respects. Safety is our top priority and critical to our continued success. A safe operation is a well-run operation. We have delivered free cash of $1 billion since 2013 from our current open pit operation. We are on track to meet our 2019 guidance for copper with increased gold production. My team is very focused on continuing to improve productivity and further optimizing our performance.
Our relentless focus on safety and production and productivity means we have also been able to significantly contribute to Mongolia's economic and social development over the 10 years since the investment agreement was signed in 2009. To date, OT has directly contributed more than AUD 9.5 billion to the Mongolian economy, which includes approximately 15,000 national employees, over AUD 2.5 billion in taxes and royalties to the government, and AUD 3 billion in payments to suppliers just from our open cut operations. Indeed, it's worth noting also that our OT underground project has committed AUD 2.8 billion or 75% of the direct project contract and procurement packages to domestic Mongolian companies. OT has a number of strengths and is already significantly contributing to Mongolia's economy. Like every other huge copper project in the world, there are some challenges as well.
This is not unusual for a project of this scale, size, and complexity, and we are working hard to mitigate and manage all of the risk. This requires regular engagement with key local stakeholders, including from the highest level of my management team. Let me take you through some of the complexities. It's important to understand the context and operating environment, which is often fluid and unpredictable. Mongolia is a young democracy with a high degree of political uncertainty.
Over the last few years, there have been a number of new governments. We have been working hard to develop relationships as key stakeholders change, as well as manage our own leadership continuity in a country that had low mining experience when we began construction in 2010. Oyu Tolgoi is a major contributor to the Mongolian economy. The underground development will account for almost a third of Mongolia's GDP in the future. I have also shared some of the economic contributions of the business that we've made as I described earlier. I have no doubt that we'll continue to play a strong role in the economic and social development of the country in the decades to come. Rio Tinto fully supports the diversification of the Mongolian economy. OT has been playing its part in developing small and medium businesses and encouraging other industrial and economic activities.
As you know, the government of Mongolia is a 34% equity partner in OT, and so therefore is contractually committed to as a shareholder of the business. Since the agreement was signed in 2009, there have been questions about the value-sharing model and the distribution of wealth. Again, this is not unusual in resource development, and we have been open to discussing value sharing with the government and people of Mongolia. At the same time as doing more to communicate and share the contribution the business is currently making. In 2018, the parliament decided to form a working group to review the contribution of Oyu Tolgoi to Mongolia and the investment agreements. The report of this group was presented at the Economic Standing Committee in early May 2019.
Following this, a new working group was formed, consisting of nine members of the parliament to finalize the report and present a resolution of recommendations to the Cabinet of Mongolia. This work is still not complete, and the conclusions have not yet been published. We have been very clear that OT's foundational agreements must be honored as they underpin the current and the future investments in Oyu Tolgoi, including the $4.4 billion invested by 20 financial institutions in the underground development project. We continue to work with the government of Mongolia and TRQ to find ways for all shareholders to receive greater benefits within our existing agreements. Steve will now provide an update on the underground development project.
Great. Thanks, Arnaud. The progress that we've made at OT in the last few months has been better than expected. I'm going to talk you through what we've been doing, the impact it's had on our productivity, and then explain where we are regarding updating on the mine design in the next phase. The most significant change since August is that we have completed the construction of shaft 2, and are now in the final stages of commissioning. The shaft itself is a significant piece of engineering. It's just over 1.3 km deep and some 10 meters in diameter. Above ground, there's a 98-meter high tower housing two of the world's largest friction hoists. Why is shaft 2 important? It helps overall productivity as those two hoists enable us to move more material, equipment, and people between the surface and underground.
We've already moved more than 1,000 tons of material to surface with the production hoist as we work through commissioning. I'm pleased to tell you that today we received our permit for the service hoist from the Mongolian authorities. This allows us to transport 300 people at a time, compared with the previous maximum of 60 through shaft 1. We've completed several major components this year, including shaft 5 and the surface discharge conveyor. The shaft 2 jaw crusher, ore bin, and transfer station are now also complete. We've also finished extensions to the central heating plant that will set us up well to work safely through the winter. We've completed the 50-meter-high primary crusher 1 excavation, and construction of the supporting infrastructure is underway. I'll talk to productivity.
In many ways, we're building something akin to an underground city with approximately 200 kilometers of tunnels and supporting infrastructure to deliver Hugo North Lift 1. To put this in local terms, we have more tunnels than London Underground. Our tunnels are much deeper. The deepest London Underground station is Hampstead, some 55 meters deep. The base of shaft 2 at 1,300 meters depth is around 25 times deeper. In September, we achieved record productivity of 1,385 equivalent meters of lateral development, the best months yet. Our teams now understand how to develop through the ground conditions they are facing. Before I move on to the mine design, I want to show you a short animation which brings to life how the mine works and gives you a better sense of its scale. Here we're looking at the to-be completed primary crusher 1.
The chamber is finished, and we're now constructing up from the base of what you see here, this supporting infrastructure. We're basically coming out of the grounds at this time. Clearly what we have here is from the base of the crusher, the flow into the conveyor system. The majority of the tunnels that you see here have been complete already. Not all of the infrastructure is fitted out yet, but we're well on our way, as I said earlier. Again, let's just give you some perspective of the scale and quantity of equipment, tunnels, and the scale of what is being built here. As we come along the conveyor system, we're heading, in this case, towards shaft 2, up through essentially a tensioner in the conveyor system.
Once we move past this in the animation, we get to the juncture where we basically can split our production between what the two primary crushers that will be installed underground. At this point, we can decide where the ore moves. The material can either go to the skip, to shaft two, or up the conveyor to surface infrastructure. Here we see one of the skips being loaded. As I said, we are already operating these skips, having moved over 1,000 tons so far to surface. These skips operate at 16.3 meters per second and take 138 seconds to go from load to unload. Here we're seeing the upper sections of the 98-meter high shaft two. Again, two largest friction winders ever constructed.
We're now seeing the skip coming up to its unload position, automatically unloads into a hopper feed system, and then places this material onto the surface conveyor system. Now we're rising out of the ground, finally above ground. Through this conveyor system, the system has all been commissioned. We're able to move material onto the overland conveyor system, the one that today takes the ore from the open pit mine up to the ore barn that you can see in the distance. Again, a large part of what you saw there has been constructed already. As you know, we're working on reviewing our mine design to cater for the geotechnical conditions that we've found. We're still evaluating a number of mine design options, and the decisions we take need to deliver for decades to come.
This work includes reviewing the location of access drives, the ore handling system, and options for panel sequencing. Of course, we cannot give you full estimates until that work is done. Let me sum up. The above-ground infrastructure is substantially complete. This includes ventilation systems, the mine control center, quarry and batch plants in a 5,500-person camp. Shaft 2 is complete and well into commissioning, and was by far the most complex element within the project's construction scope. The below-ground infrastructure, supporting infrastructure, is also well underway. We're now moving our focus to key packages such as primary crusher 1, conveyor to surface, and to shafts 3 and 4. We're working relentlessly on improving lateral development rates both on and off the footprints. We have the world's best underground mining experts working on delivering the mine design in the first half of 2020.
Once the design has been finished, we can progress the definitive estimate. We've made very good progress in the past few months and are building on this momentum. Thank you.
Thank you, Steve. Ladies and gentlemen, good morning. You have this morning heard from JS about Rio Tinto's strategy and direction. A number of my colleagues have provided a detailed update on our opportunities and our challenges and how we are addressing these. I will now, from a CFO perspective, describe the key attributes and fundamentals of Rio Tinto and how we generate value. In short, our investment proposition. Ultimately, it is, of course, all about performance. Performance today, performance tomorrow, and consistent performance over decades. Let me take a step back and share with you the elements that underpin our strong financial performance. We believe that these elements will continue to lead to superior performance and value creation, and it gives us confidence about the future. We have a portfolio of very high-quality assets. They are one, long life, two, competitive, three, expandable, and four, sustainable.
In addition to our physical assets, our very strong balance sheet is actually also a key asset for us. It gives us resilience and de-risks the overall enterprise. Moving on to our approach, the way we run our business. It is based upon one, sustainability, and above all, safety in everything we do. Two, operational excellence, driven by the right capabilities and people. Three, our value over volume philosophy. Four, our disciplined capital allocation framework. Five, our ability to act countercyclically. The combination of these strengths has led to strong financial performance. JS showed this earlier, and let me reinforce that. Firstly, our return on capital employed during the period expressed as underlying earnings after tax over capital employed, was an industry-leading average of 18%.
Secondly, and even more importantly, we have effectively converted earnings into cash. Our underlying free cash flow of AUD 23 billion over the last three years is equivalent to a return of 17%. Thirdly, we have demonstrated that we are not shy of returning cash to our shareholders, paying out AUD 31 billion of dividends and share buybacks. This is equivalent to 18% return on capital. Finally, over that same period, we have actually strengthened our balance sheet significantly. We have a strong track record, which is a good indicator of future success, but we are not complacent. We do believe that our unique strengths will support strong performance in the future and offer resilient cash flow and returns. Let me turn to our assets. First of all, we are blessed with very long-life assets and vast resources.
If you look at our capital, you can see that approximately 40% is deployed in our processing assets. These assets require maintenance, but they do not deplete, and generally, you are able to creep production. The remaining 60% of capital is in mining assets. These assets deplete, but at a very modest rate. Certainly, when I compare to oil and gas, the industry where I have spent most of my career in, it is a much lower decline curve. On the slide, you can see the numbers for bulk commodities, and in our annual report, there are actually 11 pages of extensive disclosures of our vast resources and reserves. We have reserve lives for decades for most of our products. Sorry, our resource base is multiple times of our reserves.
We have ample of opportunities to mature resources in order to also maintain a long reserve life in the future. Of course, having long life assets is only the starting point. In order to be truly resilient against volatility, our assets need to be well-positioned, and that means on the lower half of the cost curve. The good news is that over 80% of our assets are exactly there. Our iron ore, titanium slag, and borax assets are on the left side of the cost curve, and even better, those businesses have a leading position in well-structured markets. Other assets, like our copper assets, benefit from attractive industry supply fundamentals. Our portfolio of assets provides the diversification. However, the key point is the competitiveness and market position of our asset will allow us to consistently generate attractive margins and cash flow through the cycle.
Now, good assets need to be continuously maintained, renewed, and supplemented in order to have a portfolio that performs and remains relevant to the market demands of today and tomorrow. I would argue that Rio Tinto is in a unique position to do that. We are creating opportunities by consistently applying our technical skills and experience to our world-class assets. We have been disciplined, but we have kept investing and have spent AUD 15.5 billion in the last three years throughout our pipeline. Importantly, we are one of the only majors that has continued to invest in exploration, spending AUD 700 million over the last three years. You have just heard both J.S. and Steve express their optimism about our exploration pipeline, in particular about the Winu opportunity. I just want to say I'm equally excited.
We're also spending on evaluation projects, such as the Resolution project in Arizona, where earlier this year, a further AUD 302 million was approved to fund additional drilling and ore body studies. We spent AUD 7.3 billion on development capital over the last three years, which includes expenditure on replacement projects such as Amrun and Gudai-Darri and growth options, as you just have heard about the exciting Oyu Tolgoi Underground project. I've earlier said that the health of our asset is a priority. Investing in sustaining them is always our first use of cash. Over the three years period, we have invested AUD 6.7 billion in sustaining capital. In addition to having long-life assets and competitive assets with many strong growth opportunities, our portfolio is well-placed for transition to a low-carbon future.
Indeed, we are the only major miner not involved in extraction of fossil fuels, and we have reduced our own emissions by 18% in the last five years. The materials that we produce from recyclable aluminum, copper used in electrification, and our higher grade iron ore products all play a part in the transition to a low-carbon future. Over 71% of our electricity usage is already generated from renewable sources. We're not standing still, and we will communicate new targets for our emissions in the first quarter of 2020. Our world-class assets, combined with a strong balance sheet, supports the ability to provide superior cash return to our shareholders. It also enables us to better manage the business through cycles, enabling us to act countercyclically, and provides us with optionality. In summary, we have great assets.
They are long life, they have an outstanding competitive position, they provide us well with opportunities, and they're well-positioned for a low-carbon world. Our balance sheet strength provides resilience. Now let's move on to talk about what I call our approach, how we run our business. Let's start with sustainability. You heard JS cover the topic, and let me elaborate a bit further. Understanding and managing risks are central to our ability to generate value to our shareholders. That actually places sustainability at the heart of doing business. Our sustainability work is core to our strategy and strategic processes. Our climate change strategy is led by our strategy team and is central to how we look at our portfolio now and in the future. We also consider sustainability matters in all our decision-making processes, including in our investment committee. Transparency is crucial to create trust.
We were one of the first companies to publish a taxes paid report back in 2010 and have continued to expand our disclosures in this area in line with increasing government and community expectations. This year, we published our first climate change report under the recommendations of the Task Force for Climate-Related Financial Disclosures. A signature of Rio Tinto has always been our operational excellence, and now we are also building our commercial excellence and deepening our relationship with our customers. When we face operational challenges, like we did in iron ore in Q2, we take it very seriously. We are transparent about the problems. You notice it, and we do address the issues immediately. Chris has already explained our actions in iron ore, and we remain focused on excellence in our operations across all parts of the business, starting with safety. We are continuing to pursue productivity opportunities relentlessly.
They are our first lever to offset inflation, protect margins, and ultimately deliver superior return. We believe this is well reflected in our unit costs, margins, and cash flow metrics. This is how we want to continue to progress on productivity and discuss productivity progress in the future. We will stop now reporting the mine-to-market results, which I receive very few questions from. Nonetheless, we're still targeting $1 billion-$1.5 billion of additional free cash flow each year from 2021. Moving on to value over volume. You already heard Simon talking about our value over volume philosophy. It really drives our daily commercial decisions. It's important to say as well, our investment decisions are not driven by volume ambitions, but firmly based on value creation. We are a company which is growing, and we have exciting growth opportunities from our development pipeline and productivity opportunities.
Whether we will grow our production volume at 2% per year will always be subject to value over volume. Our capital allocation framework, which you know well, is here to stay. We will continue to invest in safely managing our assets and improving their performance. This means that sufficiently spending on sustaining CapEx is always a priority. The next priority for allocating capital is our commitment to our shareholders through our ordinary dividend. We carefully consider allocating to growth opportunities, balance sheet strengths, and further return to our shareholders. Our investment decisions are carried out with incredible rigor using, I must say, the most diligent process I have experienced in my career. I believe, though, that this is the best assurance for our shareholders that we will only invest in opportunities that create value. Let's turn to our investment plans.
Overall, our slow and disciplined capital ramp-up and total investment guidance has been very consistent over the last few years. Since the half-year presentation in August, the only change we have made is a slight phasing of spend from 2019 into 2020. Furthermore, we are now giving guidance for 2022 of AUD 6.5 billion, AUD 2.5 billion of which is in sustaining capital. As Chris discussed, we will increase our sustaining capital expenditure in the Pilbara as we move away from the initial expansion and as our asset base requires more maintenance and replacement. Among the equipment. Of course, we also continue to automate our fleet. Our key growth project, Oyu Tolgoi Underground is also ramping up. You've just heard about that. With Rio Tinto's long-term focus and strong balance sheet, we had the means to act countercyclically.
Our actions over the last three years, I would argue, demonstrate that we have done just exactly this. We have sold assets for good value in strong markets while maintaining discipline in our CapEx. Since the start of 2016, we have divested 18 assets and generated AUD 12.3 billion of additional pre-tax cash flow, which we have returned to shareholders. We keep a watching brief on M&A. As we already have heard, the focus is to organically grow our business. It is, though, clear that buying asset also needs to be done for value at the right time in the cycle. We've now covered the 10 elements which support continued strong performance. Let me now turn to the performance measures. The ROCE of our assets of 22% in the first half of 2019 continues to outperform our key mining peers.
This is not a short-term trend, as you can see. Over the last 10 years, our return on capital employed has on average been around 16%. Going back another decade shows an impressive average return of 22%. Mining is cyclical. We are exposed to volatile commodity prices, but the evidence suggests that our diversification and performance has created stable, high profitability over decades. In fact, I also checked in the 1990s, and here our average profitability was 14%. Another way of looking at our resilience is that during the last two decades, we have only experienced one year with single-digit return. Since we implemented our disciplined capital allocation framework back in 2013, we have demonstrated an ability to turn strong cash generations from our assets into a strong free cash flow. The chart here even excludes the cash flow from divestments. The cash flow should continue.
On the graph, the cash flow for 2019 is included. This is not a target, but simply a calculation based upon today's spot prices and consensus estimates for volumes here in the second half, in the fourth quarter of 2019. In 2016, we implemented a new dividend policy. Since then, we have paid out AUD 31 billion of dividends and share buybacks. The policy commits to a total cash return to shareholders of 40% to 60% of underlying earnings. We have, though, since its inception, consistently paid out well above the range in each year, 70%, 83%, and 72%. We have also returned the proceeds from divestments. In total, we have returned in some years in excess of the total underlying earnings of the company to our shareholders.
Given our strong balance sheet and our resilient assets, we are well-positioned into the future to continue to provide superior cash returns to our shareholders. In closing, our high-quality portfolio of long-life, competitive assets consistently perform strongly and has generated superior returns and cash flows over decades. We know, though, this is not enough. We need to run our assets with excellence, and we believe our approach to do this is the right approach. You can expect from us the same disciplined allocation of capital, a commitment to value or volume, and sustainability. We are, as the management team of Rio Tinto, along with our 45,000 employees, committed to continue to deliver superior value to shareholders in the short, medium, and long term. Thank you. Now, before J.S. closes out today, we will have our final Q&A session. Let me hand over to you, Menno.
Great. The speakers can come up front, please. Arnaud as well. Thank you. Great. Just a quick reminder for those on the phone, please register yourself if you want to ask a question. For those of you on the webcast, please remember there's an email address that you can use to send questions as well. We'll take this for 30 minutes, and after that, J.S. will wrap up. Let's start here in the room. Sam, you haven't had a question yet.
Where is Sam?
Sam.
Oh, there. Good.
Thank you. Good morning. Sam Catalano from Credit Suisse. Two questions on innovation, so probably for Steve. How do you sort of deal with the challenge of incentivizing mine managers to trial and adopt some of the new technologies, given they've obviously got targets and productivity, targets and such to meet? Then the second question is, you've outlined your gains for things like downtime and digger productivity. At what point do you think this innovation push will actually lead to changes in overall output, overall unit cost guidance? Or is it more just offsetting some of the inflation from external and internal sources? Thank you.
All right. Sam, thanks for that. I think the first one is, there's two sides to it. One is that nearly all of our operations are set tough targets that they have to meet. They're incentivized to go and look for whatever can actually help them deliver their targets, improvements in their business. Second thing is we use global metrics across the group. Everybody gets to see everybody else's. Nobody likes to be the laggard on the wrong side of the graph. I think that internal competition within the company around actually having full transparency and key data is a key one to get the competitive tension high. What we have to make sure is that what we bring to the business are options that are scalable and will work.
I think in the past, many across the industry ourselves tried a lot of stuff, didn't always work. We do a lot more piloting these days and go through a very structured rollout and scaling process, and then it's about replicating that. A lot of changes on that side of the business. I think the second one is-
If you can pause. Maybe you should, Bold, can you give an example on how some of your people, for example, in Boron, are testing some new ideas?
Yeah, sure.
Arnaud, if you want to tell us a few example from Kennecott as well?
Yeah, sure.
Yeah, maybe I'll give two examples, if that's okay.
Okay.
The first one is, we set up a Pioneering Pitch, which is effectively a sharks tank. It's run by actually our GMs and lower bands, and it's across the group in the southern and northern hemispheres, and that's driven a lot of innovation across the board and obviously improvement in safety performance. The other one is on Pioneering Pitch, sorry, on the Boron specific example. This has been going on actually for a number of years, but I must say, sorry, we asked for forgiveness, which is the team found a trial roaster on eBay. They found it on a Friday, bought it on Monday, and it was at the site in two weeks. We do try things in a bit of a different way. I think now we're obviously moving into a more scientific engineering of that.
That's not what you were supposed to say, Bold. Anyway, thanks to be on screen for once. Arnaud, if you want to go for it?
I'll come back to Bold after on the scandium, what you're doing on scandium, because that's very. Go for it.
The two examples from Kennecott, one on the ESG side, which has got impact on cost, and the other one on productivity. On ESG, you may have read a few months ago that we decided to change the power supply at Kennecott. We used to have a coal power station, and Kennecott is based in Salt Lake City. In wintertime, because Salt Lake City is surrounded by mountains, there is an inversion phenomenon, and therefore reducing the emission of dust and fine particles is extremely important to the community. What we decided to do is to shut down our coal power station completely and to source a fully renewable energy power. That is contributing extremely favorably to reducing our carbon footprint. We've reduced with this, the carbon footprint at Kennecott, by 60%.
The next step is actually to look at what do we do with our fuel usage in trucks, because this is now the biggest source of greenhouse gas emissions. The second example is the one that Steve has described on the surge loader. This is something that is being tested at Kennecott and is going to be very helpful in increasing the productivity of our fleet. As Steve said, there is a healthy competition within the different sites about productivity. Kennecott is on the top of the league ladder, and so with the surge loader, that will be a step forward again.
Scandium, I think that's a good story on how we can extract more from our resource base.
Yeah. Again, I'll steal a thunder. Two examples. The first one is obviously our ore body at Havre-Saint-Pierre in Quebec is very rich, and besides titanium and iron, it does contain scandium. We've been working our technology and R&D team to find a breakthrough of how to extract it at low cost. I must say it was an excellent collaboration with our aluminum team. I think we cracked the super scandium alloy with aluminum that hopefully makes the Boeing wings a bit lighter and more reliable. We all fly.
I think there are a few other issues besides Boeing.
It's a great example because we're working with our commercial team to make sure that we find new markets for it.
Right
if we start producing at scale, it will double the global scandium market. It has a material impact on the market, but we need to make sure, obviously, we receive new customers and new end market use. The second example is actually the same with the commercial team. They have found a customer for our Madagascar monazite. If you're not aware, but monazite contains high percentage of neodymium, which is used in heavy magnets, that is used in wind power generation. We're naturally a bit in the EV space.
Great. I think the final piece probably is a reference back to what Chris talked about. We're going right across the systems of ours, looking at what we call constraint utilization. How do we unlock all of the potential in the system? Chris gave ones around improved performance in the rail network, predicting maintenance into the rail. We've got examples that I showed for the processing plant. Remember, Chris said, we're moving more and more to below-water table ores. We've got the predictive algorithms now being designed to basically characterize the plant performance. As I said, really in the iron ore system, our constraint really is in the plants. Getting the availability of those as high as we can, but also working out how can we actually improve them beyond their nameplates and doing sophisticated work there.
I think us saying something specifically in the market around that, I'm not quite sure we're ready for that yet. Obviously, we are targeting that entire digitization of the value chain. Again, for those that had a chance to look at the MAS example, the Mine Automation System, you will see that we're digitizing everything from the ore body, and then Simon gave the example of portside trading. We absolutely intend to connect all of those pieces in near real-time to look for opportunities.
Can I just follow up to ask on this?
There's a very long question. Let somebody else get a chance. Paul.
There's Paul. Oh, yeah. There we are.
Hi. Thanks very much, indeed. Paul Gait from Bernstein. Just two quick questions. The first of which was around technology. Obviously, you're not the only company that's making a sort of showcase of the technological prowess of your company. I'm just wondering how easy is this for your competitors to replicate over time? Does this just end up with essentially a cost curve that in all the commodities ends up moving downwards, just creating deflation that is, to be honest, of no actual benefit? Is this actually something that is unique to Rio Tinto that cannot be replicated elsewhere and therefore leads to a rotation, a steepening of the cost curve and then rents? The second question is around OT, and I appreciate the sort of caveats that the definitive study only by H2 2020.
Is there anything more that you can say around from that point on time horizon that it would take before we start seeing production or is it just far too early to even give any kind of indication there? Thanks very much.
I'll start with the second question first. We can't say anymore at this time. What we need is that mine plan. Really everything hangs off the mine plan. As I said, we're working to have that with us by the first half of next year. Once we have it, that allows us to complete definitive estimate. The definitive estimate will give us those times, the cost, the schedule impacts, everything. We're not in a position to say any more than that today. I think for the first one is we absolutely believe we have a competitive advantage. That advantage manifests in a couple of different ways. One is we've been at this for a long time. We've kept some components of the core systems and platform technologies to ourselves. We have not pushed that outside the company.
The Mine Automation System and the visualization engines that sit behind that are core to that, but MAS itself is really at the heart of that. We're able now, because of the models that we're building, to ingest data at a rate that we think will be very difficult for our competitors to catch up anytime soon, and we plan to stay ahead. There's two parts to it. Do we change the cost curve of the whole business? No, because not all ore bodies are built the same. Notwithstanding you applied automation on it, they still have a characteristic, whether it's grade or it's geospatial size or it's location or whatever, that is going to give it some form of a disadvantage. I think we'll see the cost curves continue.
What we want to be able to do is have that value chain integration, which is really hard to do for most. We believe we've assembled ourselves in a way that we think is quite unique to get at that part of the problem, because that's really the sum of the parts. That gives us something greater than the sum of the parts, ultimately.
I think you should give more details on AutoHaul, the upside we have on AutoHaul, because autonomous trucks, you can buy from Caterpillar, Komatsu. AutoHaul, I don't know if you have six, seven, nine, 10 years ahead of everybody, but I think there are lots of upside, and I think it would be good if you can explain how the next wave of benefit from AutoHaul. For example, the safety block, moving from physical block to virtual blocks. That will free up further capacity.
I'll get Chris to talk to the upside because the business is looking long and hard at all of those opportunities. I think I would just remind you how much it took to deliver AutoHaul. This is not a trivial day out at the races to deliver a system of that size and scale and complexity. We underestimated it up front. We now understand profoundly what it takes to do something like that. Nobody else has done this in the world. Again, we understand that, what are those barriers, and we also now understand increasingly what the opportunities are.
Yeah. The heart of AutoHaul, the Driver Strategy Engine, DSE, is actually AI. That's not something that you can buy off the shelf. That's at the core of it.
The reason we can do now the data analytics on our rail network is because we know everything about the way the trains perform, because that came from AutoHaul. That wasn't an objective, but that's because now we collect the data, and again, that's a unique characteristic. The third thing, of course, building on the platform for the future. Should we require further rail capacity, then we can move to what's called virtual block. In very simple terms, every train now knows where every other train is on the network. You can then squeeze up the trains' spacing and allow virtual block rather than have trains parked at sidings. You can actually build on the technology at its core.
Great. Questions? Alain.
Hi, this is Alain Gabriel from Morgan Stanley. A quick question on your capital allocation of the growth component of your CapEx. If you were to think about the medium to long term, what excites you the most in terms of commodities or products outside of iron ore, and how do you tie that into your portfolio mix in five or 10 years' time?
Everything excites me. All right? As long as it's profitable. No, I think it's an important point here. I'll turn to Bold after. We believe in diversification. Okay? We have believed in diversification for a long, long time. Okay? We found a document. Louise is not here. We found a document from the 1970s or even earlier, 1960s. I don't think it's on the web, this one. What was the turnover of Rio Tinto in the 1960s? There were only two commodities that did account then for 80% of the revenue and the cash flow of Rio Tinto. Any idea what they were? No?
Uranium.
Uranium and copper.
Zinc.
Well, no. No zinc, my friend. Jason, you have to go back to the file on this one. The point is the following, is we don't know what perfect product mix will be in 20 years, 50 years, 100 years from now. Okay? We believe that if we have a portfolio of options, then we'll be much better placed. We fundamentally believe in diversification from that perspective. However, I think what was pretty clear from today and from a year ago, 2 years ago, 3 years ago, 5 years ago, 10 years ago, what we believe is, if you put it aside, safety is making money. Okay? We don't have an allocation of capital based on commodities. We allocate the entire capital pool on the back of profitability. That's a very important point.
We don't allocate on the basis of what we want the product mix to be in five, 10 years from now. It's all about making money. Having said that is, we acknowledge that commodity mix may change. We are excited with the caveat I just said about minerals for the future green economy, aluminum, bauxite, copper, and some of minerals for the battery technology from that perspective. We are looking at options. The job of everyone in the team is really to find options. At the end of the day, we will crystallize or trigger the options only if there is a profitability case from them. I don't know, Bold, you want to add a few bits and pieces?
Yeah, sure. The number one criteria for us is, of course, low cost, long life, expandable. We're not pursuing large, and that actually means looking for partnerships where we can provide value, whether it's an underground technology, whether it's a processing technology, whether it is in the geographic proximity of our existing operations. Those are the types of angles that would create value. It's very important to, of course, look at it on a risk-adjusted basis. A lot of screening going on, a lot of reviews going on, but at the end of the day, it has to match those criteria.
Just to wrap it up is, lots of people are asking always, are you comfortable with having lots of iron ore in your portfolio? Let me put it this way, and you heard it before. We've got an iron ore business which has generated more than 50% EBITDA on average for the last 20 years. You don't have many businesses anywhere in the world, and don't tell me Google or Facebook, okay? Who have generated more than 50% EBITDA in a consistent way for the last 20 years. Therefore, I don't know what the portfolio will be in 10 years from now, but don't be surprised if we still have a big share of iron ore and the target will be more than 50% EBITDA increase. Let's be clear. That's how we look at it.
I can only describe the philosophy, the way we look at it, and the portfolio will be whatever it is at the end. However, we make some choices, and the exit of coal, thermal coal, was a clear choice in the sense of we believe that this industry, the thermal coal industry, is a sunset industry. Therefore, you have two choices. Either you monetize your position today or two years ago, or you keep it for long term, and you try to harvest it like the tobacco industry. We made a choice to exit and to reallocate the capital accordingly. That's where we are.
Let's take one question from the web. David.
It's another question on PacAl. What is the solution to the power issues and the costs? Would you actually close or curtail any of these operations?
As I mentioned before, when we look at our sites, I talked about New Zealand smelter, but the same challenge we're facing are in our smelters in Australia. They are very well-run smelters, among the most efficient smelters in the world, but they lack internationally competitive power prices. The focus really at the moment is, as J.S. said, is to protect our position as one of the leading aluminum companies in this industry, sorry, with a 27% margin in the first half of this year, but can we fix those assets which are not profitable in the current market conditions. That's what we're doing. We're working hard with both the governments and the power suppliers to find viable solutions to make these assets profitable in the long run. That's a work which is underway now, and it includes all options.
That will be analyzed at the time, and we will obviously communicate once a decision is made regarding any specifics.
Let's go back in the room. Anybody back there? Yeah, Sergey. Sorry, come back here. We have time.
Thank you. Sergey Donskoy again. Two questions for Arnaud, I think. First, just to understand how achieving the ESG goals can affect your financial performance. This 60% reduction in CO2 emissions at Kennecott was very impressive. Was it accompanied by higher or lower cost of power? Second question on Resolution. Just give some update on the timeline ahead. I understand that giving any precise estimates at this point must be very difficult, but taking into account all the steps that you are going to take, what is the best time or the minimum time it will take to reach a point of where you will begin construction?
Okay, great. Thanks a lot. In terms of the cost of power at Kennecott, we haven't disclosed the details on this. However, you've seen that we've been doing the same thing at Escondida with our partners in Escondida. There it was disclosed that the contract, which is currently supplying electricity from burning coal, is going to be changed into a solar energy contract. If you look at the market in Chile, actually solar energy power contracts are extremely competitive. This is one way to create value. The other way to create value, based on my own experience in aluminium, when I was in aluminium, we did a lot of work to reduce our carbon footprints.
When you have a product like at Kennecott, copper, gold, silver, low carbon footprint product, you can start discussing with long-term customers and look at how can we create value together through partnerships. Part of the value creation is also in those long-term partnerships with either your customers or sometimes even the customers of your customers. In terms of Resolution is another incredible project, okay, with Oyu Tolgoi. As Steve was talking about the size of the deposit at Oyu Tolgoi. If you want to picture it underground, where 80% of the value is, the ore body is the size of the island of Manhattan. Okay? No wonder why in the first phase we are building 200 km of lateral development. That will require even more in the future. Resolution is also in the same league, big deposits and complex deposit as well. Okay?
Two kilometers underground is going to also be developed through block caving, therefore it's a complex project. To give you a better idea of the timing, as you've asked. The first step is completing the permitting process. That has been evolving extremely well, has progressed very well. Over the past three and a half years, we've met every single of our milestones, either on time or ahead of time. The latest milestone that we've met was the publication of the first draft of the environmental impact statement. That is very important because now it's giving an opportunity to the community to commence. This period of commence is going to be completed early November, then we'll be working with the regulator, the U.S. Forest Service, to address those comments.
We think that sometimes towards the middle of next year, we should be able to have the EIS final environmental impact statement published. There will be the land exchange. In parallel, we are progressing with the engineering and progressing with sinking the shaft. We've got already one shaft in operation, and the second shaft is two-third of the way down, and we are now sinking the remaining third. It's very interesting to see that we are transferring the learnings and the best practice from Oyu Tolgoi to Resolution. There is huge value in having the portfolio of growth projects that we have in copper and transfer best practices and people from one project to the other. In terms of the beginning of the production, at this stage, we think it will be in the late '20s.
Yeah. I may ask Peter Toth to say a few words on the ESG and the cost of the ESG, because Peter, as you know, is the head of strategy and is leading the charge on this one. We are working on the abatement curves, so we start to have for all class of assets what those abatement curves. I think if you can say a few words, Peter, to explain where we are.
Back to your question about the cost of power, it varies a lot from one region to another. In Chile, for example, it's easy to shift from one to the other, and the cost of solar power is cheaper. In some of the regions, if you have to support the cost of the firming of the grid, the direct cost could be cheaper, but the full cost could be very different. It varies a lot location by location, but Tofty, if you want to say a few words.
Thanks, yes.
Come here.
Yeah, sure.
It'll be easier.
Probably the most important first point to make is that we are in the process of obviously setting new targets, as you mentioned, that we'll publish in the first quarter of next year. Part of that exercise is obviously building abatement curves bottom-up for every one of our assets. Those abatement curves will then look at our existing carbon footprint, apply various project initiatives in terms of energy efficiency, switching out of coal-fired power into renewable power, looking at electrification opportunities and then various pieces of sort of technology such as LSS into reducing the carbon footprints over a period of time.
In terms of what Arnaud is saying, is I was mentioning a very big piece of that carbon reduction sort of journey will be looking at every piece of coal-fired power that we're using at our operations and looking at what opportunities there are to switch them into renewables. We expect significant reductions in power costs during that transition as part of that process.
Thank you, Tofty. From an energy standpoint, if I'm selfish as far as Rio is concerned, there could be lots of opportunities as well. At the end of the day, you're going to need to have more copper, more aluminum, more minerals for batteries, and so on and so forth. I think it's important. We want to be part of the solution, and we will be part of the solution. We've been very clear on this one. Okay. We have been using carbon pricing for the last 20 years. I think it's important to understand the cost, and that's why we disclosed for the first time the TCFD report in March.
There are lots of opportunities from a commercial standpoint, and that's why, and maybe we didn't make the point that clearly, the partnership with our customer and the partnership with the customer of our customers is so important. If we understand properly how it's going to work across the value chain, back to my partnership with Tsinghua and Vale, if we can connect the dots between the high-quality iron ore from Canada, from Pilbara to the automaker assembly lines in China, or doing the same in Japan and so on and so forth, then we work through this value chain. That's how we will be part of the solution. That's work in progress, but we are starting, and we're making progress.
One last question. Fawzi had a
Thanks, Tofty.
Thank you. Fawzi Hanano at HSBC. I have a couple of questions around CapEx allocation for Jakob. Firstly, when you look at the 2022 $6.5 billion CapEx that you put out there, to what extent does this number already have an allocation for still unapproved projects, both in terms of replacement and development? Secondly, with regards to debt management, do you have some sort of target net debt range? Or more appropriately here, do you have a minimum level of pro forma net debt that you prefer not to go below?
Thank you. Look, when we disclose this future CapEx guidance, we do take into consideration all our plans. We make kind of a risks assumptions around future plans such as, for example, Jadar, we haven't taken a decision on, but we risk those things in. I got the question already in the foyer. Otherwise, if you only look at the approved one, why is it not tailoring off? That's actually because we think we have another 100 years to go. It's a kind of a 50/50 guidance at this point in time from our side. The second point, thanks for that question. I like that very much. We have absolutely no debt target, because if you really fundamentally think about it, if you set yourself a target for your debt, you are almost per se acting a procyclical.
As soon as there's pressure, for example, on the earnings, you have to cut your CapEx and in a way, you should do the opposite thing. No. The key thing about a strong balance sheet is we are very comfortable even if we go to net debt 0, because it gives us optionality. If you think about the value of leverage compared to how profitable our business is, it doesn't make the mark. It's the right thing for us to have the optionality and the spend. We will see how it develops, but there's no specific points where we get very concerned.
Great. Thank you very much. Last-minute wrap-up from J.S., please.
You are throwing it. Okay. All right. You heard from the team, we have clear plans and a strong track record, but we are not becoming complacent. I mean, AUD 32 billion of cash returned to our shareholders in the last three and a half years. We are focused on operational excellence, starting with safety. I'm going to remake this point. The partnership with our customers in order to provide the right quality of product and the right quality of service is absolutely essential going forward. From that perspective, we have a good momentum. This will be the focus for 2023, 2020 and beyond. Our portfolio and our financial performance give us strength and resilience. We will continue to generate superior cash flow, and we will continue to allocate the cash flow with discipline.
I'm going to finish the same way I've done for quite some time now. For Rio Tinto, it's all about delivering value day in and day out. On this note, thank you very much. See you soon.