REC Silicon ASA

OSE:RECSI

Earnings Call Transcript 2019-Q3

[Tore Torvund, CEO & President]

Okay. Welcome to this REC Silicon presentation for the third quarter of 2019. The agenda today will be that we go through the highlights. James May will, of course, give you the update on the financials. And then today, we also have invited our Vice President for our Semiconductor business, Kurt Levens, which will give, let's say, a more comprehensive review of our semiconductor business since that is now our main activity in the company since we shut down our PV capacity in Moses Lake. I will come back to the -- an update on the China -- or the U.S.-China trade dispute and also talk a little bit about our Yulin JV and how we foresee the company's future, given the circumstances we are operating under.In terms of revenues, it's down approximately $10 million, from $46 million down to $36 million in revenue. We report a loss of $5.6 million during this quarter, and James will come back to the details behind this. The bright spot this time is the cash balance. We are having $46.2 million as of September 30, so we are somewhat better than what was forecasted in terms of cash.Silicon gas sales, 860 compared to guidance of some 900 metric tons. While we sold 900 -- 194 metric tons of semiconductor polysilicon, this is lower than what was guided. And again, I will let then Kurt and James go through all these details.As you know, the Moses Lake facility was shut down by 15th of May. We laid off approximately 100 of our employees in Moses Lake by mid-July. And we have then paid out $2.2 million in severance during that layoff in Moses Lake. We are now down to a handful of people working in Moses Lake to keep the facility ready to start if and when this will be appropriate.And we have already informed you that we have started the divestment of the Butte, Montana plant, but it will only be done if we get an acceptable bid, and we will also come back to that during our presentation. As I said, we are now not into the solar market. But to give you a short update still on the market development, it seems that 2019 is tracking approximately where we have expected all the year, approximately now 118 gigawatts will be installed. What is happening is that China has become much less important in terms of installation. In this graphic, this had 30 gigawatts as of end of September. In fact, China had only installed 16 gigawatts of solar panels. And there is now some forecast saying that China will only install between 5 and 10 gigawatts in the fourth quarter. That means at least that China is less than half of what it was back in 2018. If this is, let's say, the policy going forward or not, it's very hard to understand, but basically, China was the largest market. Now it is 50% of what it used to be, and the forecast for the upcoming years is definitely uncertain. And this forecast from PV Infolink next year is expected to be 130, 131. In terms of quarterly installation, Q4 was expected to be a very strong quarter. According to this one, it's about 30 in total. But as I said, it came out yesterday, our forecast for China, which is just half of what is in this forecast, that it will only be installed between 5 and 6 gigawatts in Q4. And definitely, that's part of the reason why there is a tremendous oversupply along the PV value chain. So in hindsight, it was definitely right on our side to shut down Moses Lake as long as we don't have access to this market anyway. We should, in hindsight, have done it earlier, but that's very easy to acknowledge today. It was not that easy during the beginning of this year. Okay. Then I will give the podium to James, so he will go through the financials here.

[James A. May, Chief Financial Officer]

Good morning. As Tore said, our revenues for the quarter decreased from $47 million in the second quarter to $36.4 million in the third quarter this year. That's a $10.6 million decrease. This decrease is largely a result of lower granular polysilicon sales due to the shutdown of the FBR facility. I'll talk a little bit more about that in a few minutes. The EBITDA for the quarter was a $5.6 million loss in the second -- in the third quarter compared to $500,000 of income in the second quarter. This decrease in earnings is largely a result of a production interruption that we experienced at the Butte facility, higher electricity prices at the Butte facility and then the severance payments that Tore mentioned associated with the shutdown of FBR in Moses Lake. And I'll provide a little more detail on these items in just a few moments. In terms of the Semiconductor Materials business, revenues were $29.7 million for the third quarter compared to $34.8 million during the second quarter. Lower revenues were a result of, again, lower total polysilicon sales which decreased by some 45% to 142 metric tons from 352 metric tons in the prior quarter. A large portion of this decrease was due to higher-than-normal sales of solar grade polysilicon from this facility in the second quarter, which was 151 metric tons in the second quarter compared to only 52 metric tons in this quarter.However, semiconductor sales volumes also declined by 57 metric tons to the 142 metric tons that Tore mentioned during the third quarter. In particular, a large portion of this decline was due to lower sales volume of the higher-margin FZ polysilicon, which also had an adverse impact on EBITDA contributed by the segment. Silicon gas sales came in at 860 metric tons for the quarter, which was about 40 metric tons below our guidance. However, it was an increase of 3.1% compared to the prior quarter. Average silane gas prices were approximately the same. They decreased by 0.5% during the quarter. EBITDA contributed by the Semiconductor Materials segment was $8.2 million, about $7 million lower than the EBITDA that we reported in the second quarter. As I alluded to a moment ago, a large portion of this decrease was due to higher electricity prices at the Butte facility. The average cost of electricity for the quarter was approximately $37 per megawatt compared to $19 per megawatt during the second quarter, which was a historical low. So price change results in higher manufacturing cost of approximately $2.8 million compared to the second quarter. The remaining $4 million in negative impact on EBITDA contribution can be attributed to lower manufacturing efficiencies caused by the production interruption, which forced the shutdown and the restart of all polysilicon reactors at the Butte facility; lower prices, primarily for semiconductor grade polysilicon; and then also the lower volumes of FZ grade polysilicon. The Solar Materials segment, I indicated a few moments ago that the revenue declined sharply because of lower sales of granular polysilicon as we sell out the remaining inventory subsequent to the shutdown of the facility on July 15. Revenues for the segment were $6.6 million for the third quarter compared to $12.2 million in the second quarter. At the end of the quarter, we had approximately 400 metric tons of granular polysilicon inventory remaining, for which we expect to report the last sales revenues during the fourth quarter. The EBITDA loss for Solar Materials was $6.1 million. The loss represents the declining expenses as operations in the facility are completed and the equipment and the facility is placed in a condition to be safely idled for an extended period of time. Beginning in the first quarter, there will be no sales revenue. We expect to report expenses only for the Solar Materials segment of approximately $5 million for each quarter going forward. Cost represents the remaining workforce and expenditures necessary to maintain the facility safely in a nonoperating status. Other and Eliminations were a net cost of $7.7 million for the quarter compared to $5.3 million in the prior quarter. The increase in expenses is almost exclusively associated with the $2.2 million in severance payments that Tore mentioned a few minutes ago. On a positive note, our cash increased by $7.8 million during the quarter. The increase in cash was primarily a result of cash inflows from operating activities, which was $9 million for the quarter, and a large -- well, more than all (sic) [ half ] of that was due to a decrease in working capital investment of $17.4 million. Change in working capital is a result of the reduced activities and turnover associated with the shutdown of FBR as inventories are converted to cash and the remaining accounts receivable are collected. The decrease in working capital was offset by the EBITDA loss of $5.6 million, interest of $1.5 million associated with long-term lease obligations and a currency loss of $1.1 million due to the impact of a stronger U.S. dollar on cash deposits in Norwegian kroner. Cash flows from investing activities were a positive $400,000. This was due to a decrease in restricted cash balances of $500,000, which was offset by about $100,000 in capital expenditures. Cash outflows from financing activities were exclusively associated with the payment -- pay-down of long-term lease liabilities associated with the adoption of IFRS 16 that we've discussed before. As said a moment ago, cash balances increased by $7.8 million in total to $46.2 million on September 30. In general, our cash balances are higher than the cash balances that we predicted with the private placement of equity in April. At that time, we expected a cash balance of about $36 million at the end of the third quarter. The difference is primarily due to a more rapid conversion of working capital to cash than we expected at that time. In terms of our debt, there were no underlying changes to the debt during the third quarter. The changes can be attributed to the payment of long-term lease liabilities of $1.5 million that I -- or $1.2 million that I spoke about a moment ago, the impact of a stronger U.S. dollar on the indemnity loan and the -- no, just the indemnity loan and then the $7.8 million increase in cash balances.The strength of the U.S. dollar also contributed to a decline in the amounts of contingent liabilities that are faced by the company. The reassessment of tax in Norway declined from $30 million last quarter to $28.2 million this quarter, and the indemnity loan declined from $23.5 million to $22 million, and that is all associated with the change in the currency exchange rates. However, once again, there are no significant changes with respect to our expectations regarding the contingent liabilities. We believe that the statuses of neither the tax issues or the indemnity loan are likely to change before next year. With that, I'll turn the microphone over to Kurt to discuss the market for the semiconductor segment.

[Kurt Levens]

Good morning. My name is Kurt Levens, as Tore had introduced me. I'm responsible for our electronic polysilicon and silicon gases business. So I'll just have a short and brief dive into our business. First of all, our business is driven by same macro drivers that drive a lot of other things in the world right now when you look at the future. It's about 5G, data transmission, Internet of Things, display ubiquity, increasing automotive content and also PV, renewable energy and energy optimization. How we create value within our -- within this particular business is it all starts with silane. From silane, we get intermediates of that process as well as derivatives of that process. We make gases that we send in silane to make polysilicon and the other gases that we sell out into the market through global-leading channel partners, large industrial gas companies who we've had relationships with for many years and work with the customers to provide our product as part of a package that they provide to the consumers. On the polysilicon side, we sell primarily to large semiconductor wafer producers, which is -- that industry is dominated by 5 large producers, of which we are customers of 4 of the 5. Gases. So when we talk about how we get these gases to our customers, to the end users, to the distributors, out into the market, we do it through these packages. We have 4 gases: silane, dichlorosilane, monochlorosilane and disilane. Now these packages are not just something that anyone can go and procure. They are engineered packages for safety, reliability, passivation of the interior surface and they're relatively long lead items.To be able to invest in a fleet like this in order to get material out into the market to end users is expensive and it's timely. It takes time to do this and you have to have the know-how. I mean a great example of that is that if you look at our largest competitor, after 5 years of restarting their facility, they're still only at 50% utilization. They have nowhere near the fleet that we have that we've built over 20 years plus of consistent supply, reliable supply, quality supply and safe supply into the marketplace. We have a dominant position in silane gas. This is due, of course -- I mean in part due to our cost position, it's a function of scale, but it also comes down to you can't just build a plant and then from there expect that you're going to have a customer base outside of an internal-consuming customer. It takes time to build fleet. You have reputation, quality, reliability, consistency, safety with the end users in this market. These companies spend $5 billion, $9 billion to invest in a new facility to build advanced semiconductor devices.The cost stack -- where we sit in the cost stack is minimal. So for them, well, price is, of course, always important. What's more important is reliable quality supply because without silane, which is the workhorse in semiconductor deposition, you would have a whole entire facility come down.So in that marketplace, we are a dominant player, 70% market share and we have this large fleet. We have secure distribution channels through contracts as well as interaction with our end users and the industry. I mean we're known for our ability to train the industry on safety. We've been a market leader in terms of reliable supply. We have over -- we've made over 40,000 deliveries of these containers to over 4 -- well, to 4 continents, 14 countries and more than 80 bulk locations, which is -- no other company can touch in either breadth, depth, length, longevity, in any metric.Now when you look at our growth, it's been steady, consistent growth. We forecast growth to continue on. We -- like the rest of the industry, 2019 has been a tough year. So there's been some inventory correction. We expect that growth will restart again next year based upon discussions with end users, based upon discussions with our channel partners. Again, when we look at future growth and what it's driven by, it's driven by the same megatrends that I pointed out in the beginning of the presentation. In addition to that, we have some accelerated growth opportunities due to the other gases that I talked about: dichlorosilane, monochlorosilane, disilane. These are being adopted as advanced precursors in next-generation semiconductor devices. And as such, they have a higher growth rate than we find with our baseline silane demand. So there are opportunities that we have to be able to expand that capacity in order to fulfill the demand that we have for that product, and we're in discussions with partners on how that may happen. Something that's not included in those numbers, outside of what we already supply into the marketplace, is the silicon anode opportunity. Silicon in lithium-ion battery anodes has been shown to increase the capacity by approximately 30%. It increases energy density, increases the charging speed, and there's a significant amount of work and optimism around silane as a source of that silicon in the battery. There's more than 30 companies that we've been talking with, that we know of that are looking at how they're going to get silicon into the battery space, and there is a segment of that who are looking at silane as the source. We're currently supplying into that market space to some of these companies. And we expect next year, based upon the forecasts, that those volumes will increase. And of course, in my model that I just showed you in terms of our forecasts, we don't have the large, expected increases should the industry adopt that platform, but we are monitoring it. And one of the reasons why we are in a very good position in that particular matter is that we have available silane capacity. We're going to allow them to be able to ramp up relatively quickly. They don't need to worry about feedstock as a bottleneck in their ambition. Polysilicon. We produce 4 main types of polysilicon. In reality, we target 2 main types and we have 1 preferred type. So our preferred product is high-purity float zone polysilicon. This is a very specialized polysilicon. It's the purest in the world. It's used for energy applications, high-power applications. There are some other esoteric applications for it in other industries as well. But it is the most pure in terms of resistivity, in terms of intrinsic quality, surface qualities, mechanical properties, its smoothness, its straightness -- or morphology in terms of smoothness, its straightness and how it's delivered to the customer. So they can just use that directly in their process without any further handling or minimal handling. We also make electronic grade chunk. This is what goes into wafers as well that are used in all of the devices that you would normally find in semiconductor applications: microprocessors, memory devices, whatever it may be. Out of our process, we do make 2 products that are, in fact, not byproducts but secondary products. Those are photovoltaic products. These are not made preferentially. These are made as a result of our yields. So this is just -- there's 2 categories: one goes into the high-purity mono, high-efficiency market; the other one goes into multi, they're high-efficiency and they are standard markets as well. These are a very small part of our production. However, we do sell into that only as needed. Ideally, if we were to have perfect yields, we wouldn't sell anything into that, but that's not achievable. 100% yield would not be an achievable target. You always make some amount of loss in the product -- process. So float zone, talked about it a little bit. Again, applications are high-voltage grid, renewable energy, high-speed trains, electric vehicles; in some models of electric vehicles, IGBTs are made with float zone polysilicon. One of the attributes of this market is that the product is supplied via contracts, there's ownership relationships in some cases with customers and it's very long-term relationships in supply. It's not something that is easy for somebody to break into. There are only 2 producers globally, of which we are one. Now electronic grade polysilicon, again, is a product that we make. It's not the preferential product. We'll fill out first how much float zone we're going to make. And then after that, we'll fill out our capacity with electronic grade polysilicon. That's simply because there are at least 4 other producers of that globally. And for us, it's important but not our primary product. It is dominated, of course, by the top 5 wafer companies in the world who represent more than 90% of the market in terms of semiconductor silicon wafers. Regarding our shipments, 2018 was a better year due to general conditions in the semiconductor space. In addition to that, we had opportunities, since we're in a growth mode, to qualify materials. So we moved material forward out into the market for qualification. Followed by 2019, where the market weakness became apparent and then started lasting longer than end users or silicon wafer producers had predicted. So we will have down year in terms of our supplies. That is forecasted right now based upon our customer forecast to pull back up in 2020, mainly towards the second half. And then we'll continue to supply somewhere in the range that is indicated by this forecast, plus/minus due to optimization of our reactors because our capacity will vary slightly based upon what products we're running, again, preference on our high-quality float zone then on our semiconductor polysilicon. And I'll turn it over now to Tore.

[Tore Torvund, CEO & President]

Okay. Thank you, Kurt. As Kurt said, the Butte facility is having -- or using the -- or serving the semiconductor business. What I'm going to discuss now is what is our idea behind as it what to do with Moses Lake. And basically, Moses Lake, first of all, main focus is to reopen the market for, let's say, solar grade polysilicon for the U.S. industry. The second thing, which is an alternative strategy, is basically to create some kind of a value chain outside of China if this trade war with China will continue, and definitely that will be focused on a much lower carbon dioxide footprint. And the third thing is basically to work on these opportunities which seems to emerge in the batteries, particular for electric vehicle. Let's say, the important thing which is also, let's say, in Butte is that we, as a company, we are silane-based making polysilicon. Silane is a prerequisite to make FBR, and silane is also what is of interest for the battery industry. And no other competitor, except for, let us say, one company in China, is silane-based. All the other companies are based upon a different gas when they make their polysilicon. Let me then give a short update on the trade war. As you all know, it has been an agreement, what they call the Phase 1. It was announced on October 11. And it is rumored, it is announced that it will be a signed deal when President Xi and President Trump meet in Santiago on the APEC summit on November 16, November 17. I had, together with Wacker and Hemlock, a meeting with the #2 in USTR early October to discuss particularly the polysilicon issue with USTR. Definitely, on the U.S. side, this is a priority to get resolved. What is the situation on the Chinese side? We don't know. And we have had no, let's say, feedback after the meeting on the 3rd of October concerning REC. And silicon -- polysilicon is part of this Phase 1 agreement. It has been completely tight in terms of communication. And we are hopeful, but we have no more information about it than what you can read in the newspaper. It was announced yesterday that it will be signed something on November 16 and 17. And hopefully, we will be part of it. But as I said, I don't have more information about that than what is out in the news media. The positive thing is at least it starts to be a more normalized discussions between the U.S. and China. And it's also said that there will be further phases in the negotiations beyond this first phase, which is basically focused on purchasing, and China has [ tentatively ] agreed to purchase about $40 billion to $50 billion of goods from the U.S. Let me then go through the non-Chinese PV value chain. And as you know, basically, China has overtaken the value chain from making ingots, wafers, cells and then into modules. In terms of U.S. polysilicon, we have approximately 80,000 capacity -- 80,000 metric ton of capacity in the U.S. That is 11% of the total capacity on a global basis. In the U.S., we have no access to the Chinese market due to this duty of 57%. China makes themselves now 64% of the polysilicon required. Rest of the world, which is mainly Germany, South Korea and Malaysia at about 25%. Both the green and the red both -- all these countries have access to China, except for the U.S. And almost all polysilicon made on a global basis end up in China to customers making wafers. Then you will see on the cells side and the modules side, there is approximately 25% capacity outside of China. Chinese companies have built module capacity in other countries to, from time to time, avoid tariffs which has been put on the modules. But that doesn't change the situation for the polysilicons since they have complete dominance in making wafers, as you will see. When you look to the carbon footprint of making solar panels based in China, and I have hopefully permission from Andreas to use his graph. And as you'll see on the left-hand side there, which is the, let's say, light blue, the expansion of polysilicon in China has been tremendous over the last 2 to 3 years. And they have mainly moved their capacity into Inner Mongolia where it is accessed through subsidized power based upon coal.And as you will see, China is, this is in the middle graph here, is the country which where the CO2 emission increased the most from '17 to '18. It increased by 4.7%. And if you dig into these numbers, it is mainly in Inner Mongolia this happens. The provinces where they not only put polysilicon capacity but also other industry which consume or need a large amount of power has been moved over to these areas. And that makes the CO2 emission increasing in China in spite of all the rhetoric that they are worried about the climate change. If we look then to the right-hand side, and I can take that to a bigger screen, you will see that in terms of gram per CO2 equivalent of per kilowatt hour, modules or polysilicon made in China has 10x more CO2 emission than what is made for REC in Moses Lake. The main reason for this is the efficiency of the FBR. As you know, FBR use only 10% of the power compared to a traditional Siemens reactor. And definitely, it is about 100% coal-fired power in China. While in the U.S., this is -- let's say, we, in Moses Lake, use in fact hydroelectric power from Columbia River. But the average in the U.S. is about 35% of coal-fired in the mix, and that's what is used on this one. And you will see also on the wafer, cell and module, the CO2, let's say, footprint in -- for a module made in China is huge compared to made, for example, in the U.S. So there is somewhat a paradox that solar panels installed in all other part of the world made in China, make a lot of CO2 emission in China, while it is installed mainly to reduce the CO2 emission elsewhere in the world. So this is something we are working on with, definitely, U.S. government. But also there is a lot of initiatives now in Europe to try to put basically some standards for solar panels, and the CO2 emission for a solar panel will be an important criteria for where we are going to install solar panels, particularly in the U.S. and in Europe. In terms of tariffs and, basically, if you see the top line here is that we have polysilicon which is equivalent to 15 gigawatts in the U.S. And we have the end market of about 15 gigawatts. The U.S. is the second largest market for solar panels. So basically, we have the polysilicon in the U.S. and we have the end market, but we don't have what is necessary in between these 2 markets. There is no capacity for making wafers in the U.S. It is only 1 gigawatt of cell capacity in the U.S. But as you can see, in fact, the capacity in modules has increased from 2 to 6 gigawatts over the last 2 years because of the tariff. I'd say a lot of companies have decided to get access to the U.S. market, but they need to have manufacturing capacity in the U.S. And this is, again, something we discuss with the U.S. government, that one alternative if there is no resolution to the trade discussion with China, it will be that basically, we create a value chain not only based in the U.S., but within countries which do not have this kind of tariff and to try to make an alternative to the dominance we have out of China. And then the second option on the -- or the third option for Moses Lake is what Kurt said, is the batteries. This is a graph made by a company called OneD based upon data from EPA, and it shows the efficiency of the batteries using silicon. The batteries in the Tesla, which is made by Panasonic in the U.S., is more efficient than the batteries which the other car companies have access to. And the only batteries using silicon today is the batteries which is then installed in the Teslas. And as you see, the last model of Tesla 3 is by far the most efficient batteries, and they have between 3% and 5% silicon oxide into the anode part of the batteries. And you can see that all the other car manufacturers have problems to find efficient batteries, as indicated in this graph. It is U.S. data, but hopefully, they are reliable and not only, let's say, giving credit to what is happening in the U.S.And Elon Musk, as you have seen here, is the anode side is graphite with silicon oxide. According to what we are told by companies working on this, silicon oxide has its limitations. The future will be to use silane -- pure silane into the anode side of the battery, and I've shown this graph previously. This might increase the efficiency by the batteries by some 30%, and it will be a huge market if this, let's say, is going to materialize. It will not happen overnight. We have -- let's say, there is more than 30 R&D companies working on this, mainly located in the U.S. We are discussing with at least 3 of those as of today. And these companies have an interest to look of the opportunity to start larger pilots in our facility in Moses Lake or in Butte.And the interesting thing concerning silane is that if you are going to do these volumes, you are not going to move them by the modules Kurt talked about. You have to co-locate these facilities adjacent to where you make the silane, and that means that you can move the silane in pipe directly into the facility where you make this anode material, which then is much easier to move to the final destination. So it is a very attractive market. These new companies definitely see that REC is the only location where they can find a huge amount of silane. The question is when and if they are successful. They are very confident, and there is a lot of money spent on developing this part of the battery to increase the efficiency, and that means basically to reduce the cost of a battery per kilowatt hour by some 30%. Short update on the JV. The market for polysilicon, not only for the U.S. but also in China, is difficult. There is no doubt that today's level prices is not sustainable in terms of Chinese production either, let's say, $7, $8 a kg doesn't cover their cash cost given, let's say, what is the overall cash cost making polysilicon. As you know, we have made a huge plant together with our JV partner, Youser Group, in China. We are not running at full capacity. We are mainly now focusing on the quality. We have been able to now produce very high-quality FBR, what is -- let's say, exceed what is necessary for mono production. So basically, we have already demonstrated the quality of the FBR. We are making high-quality Siemens products over there, and we have started to load silane.I know that someone has said that this is going to be a huge competitor to our Butte facility. It is -- let's say, we have invested 400, 500 metric tons of silane. We have only then loaded so far 12 metric tons. And there is no strategy in China to move heavily into the silane market. Let's say, we have invested for 500 metric tons, and definitely this facility will be so competitive in terms of the polysilicon market, that there is no indication, no interest to make the necessary investments to increase the capacity beyond 500 metric tons of silane. In fact, we have only 3 modules acquired in Yulin. And so far, we have decided not to increase the fleet. So it will -- even if we decided to do so, it would take years until we reach the 500 metric tons of silane. So someone I heard had said that this might be a huge competitor to Butte. That is not the case. And also at the REC, we are still on the Board. I'm the Vice Chairman of the Board over there. And we are -- let's say, we have -- there is a need for having REC on Board to make decisions in the JV in China. But I'm very pleased with the development there. As you know, in Butte and Moses Lake, we have the first generation of FBR. In Yulin, we put in the second generation of FBR, and it was meant to be able to produce electronic grade polysilicon. And we have demonstrated that, that is now possible by using our second-generation FBR reactors. Let me then talk a little bit about why we, let's say, intend to divest Butte and what will be REC thereafter. And we have then -- we are now out in the market to check if there is attractive bid for our Butte operations by, hopefully, most likely an American investor since Butte is in the U.S. and it serves a market mainly in the U.S. or outside in Asia, but not in China. The leading arguments for the divestment is, first of all, semiconductor is not the core business for our investors. Let's say, the semiconductor business is probably not very well known in Norway or as most of our investors are still Norwegian. And basically, we feel that this is not the core business. This is not what REC was focusing on. It has been a very good business for REC, but it also has probably a high value for others. It is a mature market with a growth, according to Kurt, approximately 1, 1.5x the global GDP. It's not going to be a tremendous growth in it, but it is growing decently. We do have a very good reputation, and there's definitely opportunities to develop Butte. But that need -- there is a need for capital and this capital is not -- let's say, we don't have that capital. And particularly within the precursor of silane on the gas opportunities, we have seen for many years that we should have invested, but we didn't have the capital to do so. Hopefully someone else will see these opportunities. And finally, definitely, a divestment of Butte will strengthen our balance sheet. The semiconductor market should be attractive to those who are focusing on that part. As I said, it's a mature market, but it is a very stable market. REC, we have been in the semiconductor business. The Butte facility was built back in the '90s. As Kurt said, we make the best polysilicon on a global basis. float zone is only one other company who's able to make it. We have 78% of the global market for silane, for silicon gases, and most of this is outside of China. And so there is a value chain in the semiconductor which does not depend upon the Chinese market. So even if the trade war between China and the U.S. continue, this will be a good business for the foreseeable future.It will -- if we make a decent, let's say -- or if we receive a decent bid on Butte, it will definitely strengthen our balance sheet. We will be a debt-free company. As you know, our major debt is a $110 million bond, which is then associated with the Butte plant. Without -- if we divest Butte, we will be a debt-free company. This will make REC having the opportunity to actively participate in building this PV value chain outside of China, not only REC, but together with the other participants. It will be -- let's say have the capacity to support commercialization of the lithium battery anode opportunities.We might have -- as you know, we have the opportunity to increase our partnership in China from 15% to 49%. And we have also what we call Reactor 25/26 where we already have invested more than 50%. We have to stop that development, then we can finalize 25/26 and we can then produce very high-quality FBR through this Reactor 25/26. There is no doubt that also all the experience we have now gathered in China will be put into the development of Reactor 25/26. And I didn't say that, but we still have about 10 people working in China to help out with the fine-tuning of the Yulin JV facility. All this experience will come back then to the U.S. and then put into 25/26. And if we can then move into the semiconductor market with FBR at the low cost we have on the FBR side, this is a huge opportunity as well. So the short term for REC is basically that we are now working for the divestment of Butte. We will, in the meantime, continue to operate Butte, and we anticipate an annual EBITDA from Butte, $40 million to $50 million. This is also what we then will try to convince a potential investor, that this is the typical EBITDA level for Butte. And we will then continue to operate or to keep Moses Lake in a condition where we can restart Moses Lake when the time is coming to -- or if there is a demand for our FBR capacity. And the future will definitely then be that we will be the leading company using FBR. There is no, let's say, no company really focusing on FBR for the moment. Most companies are now struggling to survive. So we are the only one who knows how to operate and how to build the FBR. We produce silane. Silane is the only -- let's say, seems to be the only way you can increase the efficiency of the anode of the batteries. And we have 25,000 metric tons of silane available in Butte. 20,000 -- let's say, according to some of those companies, if you have 20,000 metric tons of silane, that will support between 3 million and 4 million EV or electric cars in terms of efficient batteries. It was 2 million electric cars produced in 2018, so it's twice that level which was produced then in 2018. But we have, let's say, sufficient capacity for 4 to 5 million cars. And we also then, by our JV in China, we have capacity both in the U.S. and China, which is the 2 biggest markets for EV as such. So that's the idea. A lot of, let's say, discussion about what to do with REC, but we have then decided to divest Butte and to use the, let's say, the proceeds to develop Moses Lake because we think that Moses Lake definitely has the real upside together with what we have done in China. Okay. If there is some questions to Kurt, James or myself, we will be happy to answer those.

[Andreas Bertheussen, Equity Research Analyst]

Andreas Bertheussen, Kepler Cheuvreux. I have a few questions, if I may. So first, on this Butte divestment, you have been looking at opportunities to divest Butte before. There were indications of bids. Can you kind of explain why -- or how serious were those bids at the time and why didn't you sell at the time? And now, in obviously a softer market, can we expect similar valuations? Can you comment on what happened then and why you have kind of reconsidered now and decided to divest more for real?

[Tore Torvund, CEO & President]

It is true that we had some -- let's say, we tried to see what was the value of Butte. That's probably 24 months ago, something like that. We had very serious interested parties at that time. We decided not to -- let's say, the level was definitely attractive. We decided at that time not to do it because we were very hopeful that the trade war was resolved and we could get access to the Chinese markets out of Moses Lake. So we felt that the right thing for REC was to keep those 2 facilities together because size is also of importance in this business. What has changed since then, is basically that this never-ending trade war continue. And then we have to basically to look into what is the options here. Let's say, it is hard for us to continue to ask for new capital into the company as we have been -- we have asked our shareholders for new capital twice since I came to -- into this position. And we will -- definitely, we felt that it is better now to divest Butte and to use these proceeds to develop what we really think is the growth story of REC, which is basically our FBR technology. We have 2 core technologies in our company. That is the FBR, which is unique. Let's say, it is -- let's say, if PV or solar continue to grow and there's no reason not to believe that, let's say, our cash cost is half of what we have in Siemens. And I've told this many times, it is because we consume 10% of the power. It is a continuous process. Our reactors running more than 12 months without turnaround. So that means that we don't -- we have 1/3 of the labor to operate. And what we make can directly be used for making mono.So there is no doubt that FBR is a very complicated technology, but it is really efficient. And that's why we would like to continue to keep Moses Lake, to continue to have the JV, but we want to do it without asking for new, let's say, proceeds from our shareholders.

[Andreas Bertheussen, Equity Research Analyst]

So the timing is kind of, of course, odd in looking at where semiconductor markets are right now. And you could kind of draw the conclusion that you are, in some way, forced to sell now or divest now because of, I don't know, this softer quarter and perhaps more visibility on liabilities, tax bills, et cetera. Can you kind of comment on the timing? Because obviously, if you had opportunity to wait for a recovery in the semiconductor market, you would probably get a better valuation? Or are we looking at industrial buyers perhaps looking through the cycle?

[Tore Torvund, CEO & President]

Let's say, hopefully, those who are going to look into Butte are looking through the cycles. Let's say, as Kurt said, definitely -- let's say, this quarter, next quarter and the next 2 quarters might be difficult, but there is signs out there that semiconductor is picking up. Let's say, 5G is now starting to be rolled out, not in the Western world, but mainly in Asia. And several companies now in Taiwan has reported much stronger Q3 and much stronger outlook for Q4 than 2 quarters ago. Sure, it is -- let's say, we are comfortable with our cash situation as of today. But we also know that, let's say, forecast, predictability, how long can we survive with our cash, let's say, so far we feel comfortable that we are breakeven, more or less, given the situation if Butte is giving approximately the EBITDA we talk about here. But we feel that it is necessary to at least test the market, and we have also told our adviser that we are not selling below a certain level. On the other hand, we have also guaranteed that if we get above a certain level, we will not go -- that we will, in fact, divest Butte. But we have already, without being actively out there, we have already received a lot of interested parties to look to Butte.

[Andreas Bertheussen, Equity Research Analyst]

That's my kind of final question. On Butte, can you compare this process compared to the previous process in terms of interest, what type of potential buyers, categories in industrial, financial and time line?

[Tore Torvund, CEO & President]

No, I wouldn't -- let's say, we have -- let's say, those 3 companies which had -- was part -- let's say, was further in the process last time, has again come back and said that they are interested to look into it. But we have not actively started yet. We have just said that we are going to do this.So what kind of companies has an interest? It's too early to -- hopefully, when I meet with you again in February, we have an update to give you on that. But we are not going to do this, let's say, in a hurry. We are going to -- we have time. Let's say, we have definitely time. So we are going to do this in an orderly fashion.

[Andreas Bertheussen, Equity Research Analyst]

And final question. In terms of the semiconductor demand, you were talking about a pickup perhaps towards second half of 2020. I just wonder more kind of as an industrial point of view, because obviously, you are -- you should be kind of early in the cycle in terms of a semiconductor recovery. And can you comment a bit on your visibility? How far ahead -- like for Q4, you're obviously guiding on volume production. You don't see a pickup in semi in Q4. You're talking about second half. Is this based on what your investors are -- or what your -- the clients are telling you at the moment? How good visibility do they have? Are they actually seeing a pickup at some point?

[Tore Torvund, CEO & President]

Kurt, you have to be here, and you have to bring your mic. Since you are the guy who is always out there in Taiwan and meeting with the customers, you can answer this.

[Kurt Levens]

To answer your question, yes, on both the polysilicon side, the wafer customers that we have who are the large leading wafer customers have indicated that the pull on their inventory should be strong enough, that they believe that when that -- how that translates back to us in the value chain is that we should have a stronger pull in the second half. Of course, it is also -- it's segregated by large 300-millimeter diameter versus smaller 8-millimeter -- I mean 8-inch or 200-millimeter and below. But by and large, that they believe that the second half will be better based upon their conversations with customers.We know on the gas side, based upon when we now are forecast to make deliveries of this because quite often, at a large semiconductor -- new semiconductor site, they use one of these modules directly. So it doesn't necessarily go into a separate container. That container sits on our site. As such, we're very integrated with that startup process. We go out, we audit the site, we train them, we train our channel partners in terms of handling and make sure that everything is set up.And based upon those indications of some very large semiconductor users, when they plan to have startup and then from there when they plan to start ramping, of projects that were previously kind of put on ice due to the issues in semiconductor, that those are coming along as scheduled in 2020, which will then result, at least from our standpoint, in terms of turns on those containers and more sales in gases towards the second half.So we do have some indications. Of course, I will just caution by saying last year, everyone thought it was going to be over by the first half -- or I should say this year. It started Q4 of 2018. Everyone thought it would be done. By everyone, I mean all of the major semiconductor guys. You can go back and read their investor releases and their quarterly presentations, they thought it would clear by the first half and that the second half was going to be good. So I only caution saying now everyone says the same thing, but in 2020, and in fact, we are planning on that based upon our communication with them when we have to have deliveries and need to make sure that we're prepared for their ramp-ups.

[Unknown Executive]

There's a quick question on Butte. If you could say anything more about production and also electric prices in Q4 and Q1 for the Butte facilities. How has the electricity prices developed so far this quarter? And what are your expectations going forward?

[Tore Torvund, CEO & President]

We will not, let's say, disclose what is actual power prices. What we can say is that we are, let's say, we are purchasing the power on the spot market. So we are related to the spot market. We don't have a long-term power contract in Butte.

[Unknown Executive]

And another question on cash burn for the company. Any indications for 2020?

[Tore Torvund, CEO & President]

We should have had 3 microphones put out.

[James A. May, Chief Financial Officer]

As I said this morning, if you look at cash flow, cash flow -- excuse me? Cash flow is dependent upon foreign markets. So as of this morning, we're still very much on track with what we anticipated with the April release. We expect that through Q4, that will be the case with a few items. After that, it's dependent upon the market as we go forward from there.

[Unknown Executive]

And there's quite a few questions on an acceptable price for Butte and what will be kind of necessary to execute on your -- the list of ambitions that you stated in a kind of non-Butte future for REC Silicon?

[Tore Torvund, CEO & President]

We will not disclose what is the, let's say, the target price we have put for Butte. But internally, we have, let's say, we have told our advisers that above a certain level, we are willing to divest. And that's something we had to do more or less because last time, even it was an acceptable bid, we decided not to divest. And it is fair now that at least those who are going to bid that if they are above a certain level, we will be then willing to divest Butte. But that level is a decent level, and we feel that if we get that level, that will be a good transaction for REC and shareholders of REC.

[Unknown Executive]

And if Butte is sold, how will Moses Lake have the capabilities for meeting market needs when it comes to silane gases?

[Tore Torvund, CEO & President]

So we have 25,000 -- let's say, Butte has about 7,000 to 8,000 metric ton of silane capacity. So basically, half of this capacity is going directly to the market as silicon gas. Half is then used for polysilicon. We are now on the polysilicon side Butte, they are focusing on the very high end, and that means basically float zone. What we move into the semiconductor market as EG is more the fallout when we make float zone. So half is used for polysilicon, half is used for the market. There is no doubt that when the market of silicon gas grows, we will continue to serve that market to keep our market share of about 70% as this grows because that's the main focus. And then the remainder of the capacity will be used for polysilicon. In Moses Lake, we have 25,000 metric tons. And so we have, let's say, 3x, 4x more capacity of silane in Moses Lake. So we have plenty of silane for what is the need, either on polysilicon if the trade war is resolved, if we have an outside value chain or if the battery companies require more capacity. So what we retain in the company is the big chunk of silane.

[Unknown Executive]

And one final question on the operation in Butte. Did you build any inventory to float zone during this quarter?

[Kurt Levens]

Yes, there was a slight build of inventory. In general, we'd make the order. So as the orders slow down, you'll see an increase -- a small increase in inventory. We'll work that out as we go forward with lower production or to meet higher sales.

[Unknown Executive]

And maybe one final question here, you maybe touched upon it already, but what kind of type of companies will you see as natural buyers of the Butte plant, more specific than likely American or...

[Tore Torvund, CEO & President]

The limitation, let's say, it might be industrial players, it might be private equity companies. What it will not be is Chinese-based companies because they will not get access or permission to buy these kind of assets in the U.S. due to the political situation. Definitely, the Chinese companies would have been very interested, but they will not be allowed to purchase this kind of company today, due to the high quality and also due to security reasons.Let's say, that might be negative in one way. On the other hand, if there is a continued, let's say, situation between China and the U.S. due to, let's say, what we have heard about all this, let's say, security issues, that might also be of interest to American investors because it means that it will be a complete value chain built outside of China on these devices, things like that, which require that type of product to be kind of manufactured.Okay. If there is no more questions, thank you so much for coming, and have a good day.