Siltronic AG
XETRA:WAF
US |
Fubotv Inc
NYSE:FUBO
|
Media
|
|
US |
Bank of America Corp
NYSE:BAC
|
Banking
|
|
US |
Palantir Technologies Inc
NYSE:PLTR
|
Technology
|
|
US |
C
|
C3.ai Inc
NYSE:AI
|
Technology
|
US |
Uber Technologies Inc
NYSE:UBER
|
Road & Rail
|
|
CN |
NIO Inc
NYSE:NIO
|
Automobiles
|
|
US |
Fluor Corp
NYSE:FLR
|
Construction
|
|
US |
Jacobs Engineering Group Inc
NYSE:J
|
Professional Services
|
|
US |
TopBuild Corp
NYSE:BLD
|
Consumer products
|
|
US |
Abbott Laboratories
NYSE:ABT
|
Health Care
|
|
US |
Chevron Corp
NYSE:CVX
|
Energy
|
|
US |
Occidental Petroleum Corp
NYSE:OXY
|
Energy
|
|
US |
Matrix Service Co
NASDAQ:MTRX
|
Construction
|
|
US |
Automatic Data Processing Inc
NASDAQ:ADP
|
Technology
|
|
US |
Qualcomm Inc
NASDAQ:QCOM
|
Semiconductors
|
|
US |
Ambarella Inc
NASDAQ:AMBA
|
Semiconductors
|
Utilize notes to systematically review your investment decisions. By reflecting on past outcomes, you can discern effective strategies and identify those that underperformed. This continuous feedback loop enables you to adapt and refine your approach, optimizing for future success.
Each note serves as a learning point, offering insights into your decision-making processes. Over time, you'll accumulate a personalized database of knowledge, enhancing your ability to make informed decisions quickly and effectively.
With a comprehensive record of your investment history at your fingertips, you can compare current opportunities against past experiences. This not only bolsters your confidence but also ensures that each decision is grounded in a well-documented rationale.
Do you really want to delete this note?
This action cannot be undone.
52 Week Range |
46.42
92.75
|
Price Target |
|
We'll email you a reminder when the closing price reaches EUR.
Choose the stock you wish to monitor with a price alert.
Fubotv Inc
NYSE:FUBO
|
US | |
Bank of America Corp
NYSE:BAC
|
US | |
Palantir Technologies Inc
NYSE:PLTR
|
US | |
C
|
C3.ai Inc
NYSE:AI
|
US |
Uber Technologies Inc
NYSE:UBER
|
US | |
NIO Inc
NYSE:NIO
|
CN | |
Fluor Corp
NYSE:FLR
|
US | |
Jacobs Engineering Group Inc
NYSE:J
|
US | |
TopBuild Corp
NYSE:BLD
|
US | |
Abbott Laboratories
NYSE:ABT
|
US | |
Chevron Corp
NYSE:CVX
|
US | |
Occidental Petroleum Corp
NYSE:OXY
|
US | |
Matrix Service Co
NASDAQ:MTRX
|
US | |
Automatic Data Processing Inc
NASDAQ:ADP
|
US | |
Qualcomm Inc
NASDAQ:QCOM
|
US | |
Ambarella Inc
NASDAQ:AMBA
|
US |
This alert will be permanently deleted.
Hello, everyone. And welcome to Siltronic's conference call on the full year results 2018. Please note that this call is being recorded and also streamed on Siltronic's website. The call will be available as an on-demand version later today. Your participation on this call implies your consent with this. At this time, I would like to turn the conference over to Petra Muller, Head of Investor Relations and Communications of Siltronic AG.
Thank you, operator. Welcome, everybody, to Siltronic's presentation of the full year 2018 results. With me are Chris von Plotho, our CEO; and Rainer Irle, our CFO. Both will comment on the market development and our financial performance and will be available for Q&A afterwards.Please note that our presentation contains the usual safe harbor statement and applies throughout this call and presentation. Today, we have published all documents relating to our 2018 reporting. They are available on our website. I'll now turn the call over to Chris for opening remarks.
Good morning, ladies and gentlemen. 2018 was an excellent year for Siltronic. We have seen strong wafer demand throughout the year, full loading and prices increasing each quarter.Due to further price increases and less headwind from the U.S. dollar than originally expected, sales came in at EUR 1,456 million, nearly 24% higher than in 2017. Our EBITDA reached EUR 589 million and our EBITDA margin was 40.5%. EBIT more than doubled at EUR 498 million and ROCE was 58%.Net profit for the period also more than doubled at EUR 400 million. Our net cash flow reached EUR 240 million. We recorded a net financial asset of EUR 691 million at the end of December 2018.Now let's have a look at end-market developments. While we have seen full utilization of our lines during the first 3 quarters of 2018, we saw first a decline in utilization in 150-millimeter in Q4, while 200- and 300-millimeter remain fully loaded. At the end of the year, customers in 300-millimeter still took all upside potential that was available. However, since mid to end of January, we saw utilization in 200- and 300-millimeter coming slightly down. While in 200, this particularly refers to standard products, special products like arsenic or red phosphorus doped wafers are continuing on a high level. And demand is more muted going into 2019 with mobile phones softening, a flattish PC and auto business, while servers as well as electric and hybrid vehicles are up. Since the end of January, we see in 300-millimeter inventories of our raw wafers at our customers going up. We believe that this is partially due to the fact that process times at our customers is rather long, with something like 2 to 3 months, so their reaction to changes in demand takes some time. We still believe in the underlying growth of our market as silicon continues to grow. However, in 2019, this is impacted by declining or flattish unit numbers in some end markets.Content in smartphones in 2019 seems to be rather muted because of lower number of high-end smartphones sold. PC business is impacted by CPU shortage in Q4 2018, and we believe this will extend into Q2 2019. On the other hand, servers are up in units and content. And in cars, the number of units sold is expected to be flattish, but content continues to grow.Content in smartphones is still growing. This is mainly driven by memory and more features. According to smartphone high-end uses, content in future will increase by integrating more memory, more cameras and more sensors.However, 2019 seems to be a difficult year for high-end smartphones, which didn't grow as expected. Apple seems to lose most and also Samsung shows a declining trend. On the other hand, we see names like Huawei, who are gaining.Now let's have a short look at the PC market. As you all know, CPUs were short starting back in Q3 of last year. This shortage still holds on while the demand is basically intact due to the transition to Windows 10. But for the full year 2018, number of PCs, laptops, notebooks was nearly 4% down compared to 2017. On the other hand, servers showed a 6% growth. This increase is fueled by units and content, and content will continue to grow due to the additional content required by artificial intelligence applications. 2018 was a difficult year for the automotive industry, also driven by problems with the new emission standards, and most probably, 2019 won't be much better. However, going forward, we see continued content growth in cars due to more driver assistance systems, more hybrid solutions and sometime in the future, some -- more electronic cars on the street. However, today, the outlook for 2019 is much more muted than it has been in the beginning or even in the beginning of the year or even at the end of 2018. With this first overview, I would like to hand over to Rainer for the detailed presentation of our financials.
Thank you, Chris, and welcome, everybody. Our sales in Q4 2018 were up again by 2% compared to Q3 and reached 338 -- it's EUR 388 million. ASP increased further in Q4. Exchange rate effect had a slightly positive impact quarter-on-quarter as the U.S. dollar averaged $1.14 against the euro in the fourth quarter after $1.16 in the third quarter. Cost of sales increased slightly quarter-on-quarter from EUR 205 million to EUR 208 million, while gross profit in Q4 increased to EUR 280 million compared to EUR 275 million in Q3. Gross margin was basically stable at 46%.In 2018, we saw higher ASPs every quarter. The EBITDA at EUR 161 million in Q4 '18 was roughly stable compared to Q3. However, in Q3, we had a positive onetime effect of a customer pulling forward wafer shipments requesting AR instead of sea freight. The EBITDA margin reached 41.4% in Q4 after 42.2% in Q3.EBIT came in at EUR 139 million also roughly stable compared to Q3 at EUR 138 million. The EBIT margin was about stable at 35.8% after 36.3% in Q3.Net profit more than doubled compared to 2017 due to the ongoing positive business development during 2018. We reached a profit for the period of EUR 400 million. Income tax expense in 2018 were EUR 88 million. The corresponding tax rate was 18%. Our equity increased to EUR 960 million. This is mainly due to the net profit for the period. Our equity ratio remains stable at roughly 50%.While we received the majority of customer prepayments already in the first 9 months of 2018, in Q4 there was also a smaller effect.Pension provisions as of 31st of December 2018 were relatively stable compared to year-end 2017 despite a lot of changes in interest rates and mortality tables, and that's why we added a new page to explain that a bit more. As you know, most of our German pension liability is funded through the back of pension funds, that's the blue portion, and only a smaller part in Germany is unfunded. DBO in Germany increased by roughly EUR 50 million due to new mortality tables and also due to lower interest rates to be used under IFRS. In the U.S., interest rates went up a little bit.Given our strong cash flow in 2018, we decided to increase funding our pension funds. We funded CTA and put EUR 30 million aside for the red portion of our liabilities. And again, as said earlier, in my view, the interest rate of below 2% is far too low and overstates the liabilities by at least EUR 150 million. For comparison, we put in a calculation here using 3% discount rate, and you can see how far the liability goes down.Our net financial assets increased further and reached EUR 691 million, slightly lower than in Q3. In 2018, we received about EUR 270 million in customer prepayments, and we paid back roughly EUR 41 million.Net cash flow increased significantly to EUR 240 million in 2018. CapEx in Q4 was about EUR 150 million. Our CapEx for the full year adds up to EUR 257 million, slightly below the forecast of EUR 260 million to EUR 280 million. Apart from MOB, that mainly relates to the capacity extent in 2019, the new crystal-pulling hall in Singapore and ongoing automation projects. With that, I would like to hand over to Chris again.
Thank you, Rainer. Unfortunately, our view on 2019 is more muted than it has been at the beginning of the year. Demand in 150 and smaller diameters is further slowing down. 200-millimeter is healthy for some special products, but somewhat muted for standard products. And 300-millimeter, many customers do not take any upside beside the contractual volumes anymore. This was still the case in Q4 of last year. As Siltronic's average selling prices higher than the industry average due to strong price increase we realized over the past 2 years, we believe that we will only see a slightly higher ASP in 2019. Customer raw wafer inventories are showing an upward trend. We believe that we will see this for some time -- for some more time as process times at our customers are in the range of 2 to 3 months. Hence, it is not that easy for our customers to hit the brakes if markets slow down. Overall, we assume that wafer volume in H1 2019 will be significantly down compared to H2 2018. However, we expect the market to be in a better shape in the second half of 2019. We expect that our sales in 2019 will be in the region of 2018 and our EBITDA margin will be slightly down. As we see headwind in our cost positions by wage increases as well as EUR 20 million higher energy cost in Germany and higher depreciation, our EBIT should be a good 10% below 2018.Our net cash flow will be clearly positive again. However, approximately EUR 100 million below 2018 due to the increased investments. However, in 2020, it will be up again. Our depreciation will be up to around EUR 110 million due to new equipment installed during 2018. Tax rate should be between 15% and 20%.Our CapEx will be around EUR 350 million. The focus is on capabilities and more automation and also to further increase our 300-millimeter capacity based on LTAs with customers. In 2020, CapEx will be significantly lower. I would like to emphasize that we are still convinced about the underlying growth of our industry, even if 2019 is a bit muted. We see more content in smartphones and servers, more memory, more sensors, more hybrid and at some stage, more electrical cars and so forth. Please bear in mind that wafer manufacturers only have a limited channel capacity right now. So under the assumption that the market will still be growing maybe in some years and -- or in some years less, the supply and demand situation will get tight again, even tighter than what we have seen in 2017 and 2018. However, currently, it is very difficult to predict when exactly this will happen. With this, we will close our presentation and are now available for your questions.
[Operator Instructions] The first question is from Amit Harchandani of Citigroup.
Amit Harchandani from Citi. Three, if I may. Firstly, to begin with, with respect to demand. Clearly, you saw a shift in demand by exiting the Chinese New Year. So I was just trying to wonder if that's clearly down to customers taking less, ordering less? Or do you also believe you may have potentially lost some share to other competitors who may be more aggressive in terms of churning out the wafers? Secondly, with respect to CapEx. Clearly, an elevated number for this year. I note that you have not given us a breakdown compared to last year. Is there anything we should read into that? And does that also mean you are continuing to add capacity to maintain your market share versus competition? And lastly, with respect to pricing, you've talked about slower pricing because you are stronger -- you realized stronger pricing in the previous years. Could you maybe give us a sense for how the spot pricing or current pricing discussions are trending? Or in other words, is this average pricing increase more driven by contracts you've already negotiated rather than ones you are looking to negotiate?
Well, let me take first the last question. Also in 2017 and 2018, there were some little volumes in -- sold on spot basis, but this spot basis did not have a major influence on ASPs, specifically not for Siltronic. Actually, ASP is impacted by the contractual situation. And like I said in my statement regarding the outlook, actually, customers are not taking any spot quantities. So there is no actual spot price. Secondly, you asked regarding CapEx. Of course, we looked into our CapEx expansion plans, like we said during '17 and also during '18, are always linked to additional LTA contracts that we signed. And we've got some prepayments for that, and we do not see a reason to postpone these investments. And last but not least, you asked for share loss. Yes, this is more a best guess than any knowledge. We do not have any reason to believe that we lost share.
The next question is from Rob Sanders, Deutsche Bank.
Yes. My first question is just on the 2019 outlook. It sounds like you're thinking that pricing will be slightly up and volumes slightly down. But when we look across the major customer groups, logic foundry, DRAM, those will all be flat to up. So can we assume that it's the NAND customers that are driving this decrease? Presumably, 2D NAND is a big problem right now, but it's below cost. And I have some follow-ups.
Well, I think looking at the end market, it's probably the right thing to do if you look at mid- to long-term outlook for our industry. But when you want to have a judgment of the actual situation and the outlook for the next 2 or 3 quarters, I think looking at end markets is not a perfect indicator. Like I said during my presentation, the big challenge is that, for us, when we start to put crude silicon metal into the crucible and the day we shift away for produced out of that to the customer, we have something like 15, maximum 20 days. When we look at the next process step at our customers, they do a wafer start until the chip can be sent out to the customer, it's between 2 and 3 months. So for them, reaction -- the reaction time in slowing markets or in increasing markets is very, very difficult. We have good reasons to believe that there were some inventories piled up, not for all wafer at our customers, but finished goods with our customers and raw material at the customers of our customers, and we do believe that this is the major reason for, let's say, a slowdown in memory -- or memory adjustments, not only NAND, also DRAM.
Got it. And then that was the biggest driver? Okay, got it. And on the LTA percentage, I think Shin-Etsu is running at 95% for the 300-millimeter; GlobalWafers, 90%, SUMCO, 80%. What are you running for 300-millimeter LTAs right now? I think I remember that you're being lower than those numbers.
Well, as far as we know, when competition talks about LTA, they refer to 300-millimeter, we refer to the overall activities of Siltronic, and we say slightly over 40%, but we always gave an indication that 300-millimeter is higher than the average of the company.
Okay. So more like 60%, 70%, okay. And the last question would just be on the prepayment side. So you took a lot of prepayments in recent quarters, but I'm just wondering when did those prepayments -- should -- how should we think about the cash flow from prepayments in 2019 and 2020 going forward?
Yes. I mean, you can see in the balance sheet from the current portion of the prepayments that we will pay back some EUR 50 million to EUR 70 million this year. We are also expecting to see some additional prepayments at a smaller extent. And then paying back will be something the rest of the prepayment that will happen over the next 6 or 7 years.
Got it. And just to be clear, that's not included in your definition net cash flow. So the EUR 100 million decline is relative to the 230 or -- sorry, 240-odd you did in 2018, which is excluding prepayments?
Yes, that's perfectly right.
The next question is from Florian Treisch, Commerzbank (sic) [ MainFirst Bank ].
So my question would be if you can add some quality statements on the respective sequential improvements over the coming quarters, i.e., if Q1 really the low point, how much of an improvement are you expecting for Q2 or H2 in general? As a kind of follow-up to it, if you just mentioned the long process time at your clients, in particular when it comes to memory players, are you afraid that you, too, can potentially be even weaker than Q1 given low visibility into the production here?
Well, since we are a public company, we never gave an outlook for quarter-by-quarter. The only thing we said this time, that Q1 '19 will be weaker than Q4 '18 was, and we talked about the first half of this year, which will be, like I mentioned, significantly below the second half of last year, and that we, like all the others in the industry, expect a stronger H2 2019.
The next question is from Martin Jungfleisch, Kepler Cheuvreux.
I have 2. First one is on pricing and in addition to the previous question that was asked before, maybe you can also comment on the smaller diameter pricing, how sustainable that is given the decline in importance and volumes. And the second one is on the 2020 CapEx. I want to understand if this is more capacity or demand constraint. So could you comment on how much share capacity in 300-millimeter would you have left after the deeper post expansion in '19 and '20? And what's your current thoughts on greenfield timing?
Well, first of all, greenfield timing, there is no timing on greenfield. The only thing we know, and this didn't change due to the things that happened in the last 6 months. The industry, including Siltronic, will reach a situation where greenfield investment will be needed. We don't know when that will happen, but we are convinced that it will happen. And I think our customers will be a little bit more concerned about that question than we should. Okay, second one was pricing. Typically, we do not disclose pricing development by diameter, but I understand that the question comes up regarding LPE because there we see a more significant drop in utilization. Up to now, in most cases, they will -- prices are still stable in SD/CLE, but it's also for us a valid question how long will that last or will it last forever. When utilization suffers several quarters around 80%, then it's really questionable whether prices will hold. And then you asked a question regarding CapEx 2020. CapEx 2020 has the typical mix. We have maintenance of business and capability enhancement, which is something like EUR 80 million to EUR 90 million, then we will continue to do our investments in Germany regarding automation. And then last but not least, we have things that we started in '18 that will also create some additional cost in 2020 like the crystal-pulling hall in Singapore.
The next question is from Jürgen Wagner, MainFirst Bank.
Actually, how are your long-term agreements trending? And are customers currently still asking for new long-term agreements? And then a question on Chinese competition. Has the slowdown now changed anything on the ambitions of China to get into wafer manufacturing?
Well, first of all, LTAs. Actually, we do not have discussions on additional LTAs with customers. And I think this is simply due to the actual market environment, but we continue to -- to our argument that one day, the industry will need greenfield, and we'll stick to our point that we need to see additional LTAs in order to justify that. And we continue to say that between the decision point and first wafers out, it's 3 years or more. So like I said before, customers are more concerned about it. Regarding China, I still want to emphasize that the major focus of the central government is on semiconductors and not so much on wafers. We stick to our statement that we have unchanged in the last 3 years that Mainland Chinese 300-millimeter prime players will not play a major role in the next 5 years unless they buy technology. In the meantime, we got 1 wafer out of Mainland Chinese production, which was a test wafer, compared to that kind of test wafer. The performance compared to our test wafer was disappointing, and then we got an iffy wafer in our hands. But this was a technology, which Siltronic produced something like 12 years ago. And when we compare it to the performance that we had 12 years ago, it's also somehow behind. So there is no reason to believe that China will be faster than we anticipated up to now. But I always say, one day, they will have the capability. We never said they never will. One day, they will.
The next question is from Gustav Froberg, Berenberg.
I have 2 as well, if I may. Firstly, just on this recovery that we're all talking about in the second half of this year. Could you maybe give us your view as to what will actually drive this recovery and what will be the reason for the pickup in H2?
Well, I think the major driver will be that in the value chain, starting at the end product of our customers and the end product, which is sold to the consumer, there were some inventories going up, and we will see some inventory adjustments and then we are back to normal again.
Okay. So it's just that the inventory will go down.
I want to mention something on the inventory, because in my speech, I mentioned that we saw inventories for all wafers going up at our customers, and that's perfectly right. But I take to -- I take example of one customer where I won't give you the name, but I will give you the development of inventory. Beginning of December, this inventory was based on our wafers was somewhere between 2 and 4 days. And now it's 15 days. So yes, from 2 to 4 going to 15 is a significant increase. But when you look at the targeted days of inventory of our customers, 15 days is not at all a problem, it's still too low.
All right, perfect. And then just to add, I just want to go back on pricing again. What is it that makes so you confident that ASPs will be higher in '19 versus '18 given that you have 40% on these long-terms contracts and then you renegotiate every quarter and half year for the remainder?
Well, first of all, up to now, we see prices being relatively stable, point number one. Point number two, we have a pricing effect, which is a carryover from 2018 because prices went up, like I said in my speech, basically every quarter in 2018.
Okay. So is it then right to assume that if prices stabilize a little bit maybe this year, then they will go down in 2020? Or...
Well, we are always somehow reluctant to give a precise outlook for pricing for the current year. So if you ask now for 2020 and '21, my answer is simply I don't know. At the end of the day, like I said in 300-millimeter, we see a future where we will have a tighter situation that we had in '17 and '19 -- in '18. In '17 and '18, prices went up significantly. And when we come to a next level tightness in the market, I do not see any reason for prices going down or being stable.
The next question is from Achal Sultania, Credit Suisse.
Just one more clarification on pricing. So I guess like you're saying pricing should be up slightly year-on-year for 2019 full year. Obviously, prices last year went up every quarter on a sequential basis. So does it imply that prices like the average blended price starts to come down at some point during this year, because you're guiding only for a small increase and given we are already at the peak pricing at the end of Q4? So does it imply that prices slightly start to come down? And if it comes down, then is it mainly because of the 200- and 150-millimeter wafers?
When we said we are looking only for -- we expect only to see a slight increase, we were not arguing that price are going down. We are arguing that we already reached, after the significant increase in '17 and '18, a price level, which is significantly above the average of the market.
Okay, that's clear. And then one follow-up. Like when I try and contrast what you are saying versus your competition, clearly, there seems to be some disparity in how you are talking about 2019 and how your customers are talking about. Is it mainly a function that you saw price increases earlier than them? Is that the main reason? What's causing that deviation in outlook commentary?
Well, we have a follow-up on this on a quarterly level. And since the beginning of the price increase, which is end of '16, beginning of '17, we saw the gap becoming bigger. We had a peak and now we still are at a stable, but relatively higher level over the average of the market. But this is not new. We always said that we are under the impression that we increase prices significantly more than competition did, and I always told you, look at the EBITDA development of the players over the last 2 years and ask yourself what are the major influences. There was no currency influence. There was probably no customer mix or product mix significantly. There was no big volume increase over these 2 years. So the major reason is pricing, and I think it's a given that Siltronic outperformed competitors on EBITDA development in the last 2 years.
Okay. Okay. And Rainer, maybe a follow-up on this pension thing. So I'm looking at the slides where you present the pension obligation by country. So is it fair to assume that like depending on what discount rate we take, whether it's 2% or 3%, basically the unfunded part of pension obligation is somewhere around EUR 140 million to EUR 170 million? Is that what you're trying to communicate today?
Yes, if you look at the right side, the second bullet, 3% unfunded portion is roughly EUR 200 million. I mean you could even be more aggressive and say 3% is still too low because the back of pension fund calculates with 3.85% under German regulation that seem to be fully funded. I could argue that EUR 200 million is even too high, but I think the point I was trying to make is simply it is overstating the obligation clearly.
We have a follow-up question from Amit Harchandani, Citigroup.
A few follow-up questions from my side, if I may.Firstly, I again just wanted to go back to the volume assumption for 2019. Your guidance definitely implies volumes being down, which is a sharp contrast to I think the IHS plus 4% estimate referred to in the preliminary release. I note that in the past 20 years, probably volumes have declined 3 times in the industry, I think 2001, 2008 and 2009. But is it fair to say, looking at the situation today doesn't -- it's tough out there, but not probably as tough as those years. So volumes being down year-on-year, is that you being too conservative? That will be my first question.
Well, our future will tell that. We will only know it afterwards. But please keep in mind, in my speech, I said that the situation after, let's say, end of December last year, outlook 2019 and the actual outlook changed significantly. The 3.6% from IHS Markit are from December. So you compare them -- if you compare that to our outlook, then you compare outlooks which were given that's completely different point in time. And we all know that during this period of end of December until now, it's 9 weeks, the environment changed quite a bit.
Okay, okay. Secondly, if I may, with respect to CapEx. I think in the annual Report, you have called out investment in epi field reactors, if I said that correctly. And again looking at EUR 350 million obviously, it's a fairly large outlay. Could you give us a sense for what proportion of this is related to epi field reactors? Just trying to get a feel for the breakdown of the CapEx. And if I may also on CapEx going into 2020, when you talk about a significant step-down, is that down to automation? Is that down to capacity addition? Which components in CapEx are going down in 2020?
Well, first of all, automation does not go down significantly, and maintenance of capability, maintenance of business is relatively a fixed figure. So I think that's a relatively clear answer. And about the epi reactors, we do not disclose these features. And by the way, the -- adding epi reactors is a continuous story that started some times during 2017, where we added 1 epi reactor in Burghausen and then we added another one. And there's a third one to come into Burghausen, and the further steps in the future are to be -- are to happen in Freiburg.
Okay. And lastly, if I may...
And the reason for that is on one side, that the epi market grows, point number one. And point number two is that we always refer to the fact that, let's say, outlook by equipment is in some cases decreasing due to a more demanding specifications, and that's probably more true on epi products than on polished products.
Okay. And lastly, if I may. You talked about an example of a customer where days of inventory have gone up from 4 days to 15 days, which admittedly looks quite low in an absolute sense. So are you thinking about continuing to produce wafers at full utilization right now and build up inventories at your end because you would anticipate this demand to snap back in the second half of this year and you want to be ready with inventories from your side? Or are you trying to just follow how demand is slowing and take down your utilization levels in the first half of this year?
Well, I think the best indicator for our production volume is our contractual obligation.
Okay. But you said contractual obligation, long-term agreement is only 40% or more than 40% of the output.
But every contractual obligation is an LTA.
And we have a follow-up question from Rob Sanders, Deutsche Bank.
Just very briefly, on utilization by wafer diameter, do you disclose that? I mean, I'm sort of working out that you get to the mid-80s in 150, but maybe low 90s in the other diameters. So I'd love any color there. And then just back to this pricing question. I think it was TSMC that said they were looking to renegotiate pricing even on LTAs. Have you seen any behavior or -- that would suggest those companies might come back to you? Obviously, we've got a situation our volumes are not growing, but everyone's adding capacity in 2019 and 2020, which does suggest that they will look to play players off each other. So have you seen anything there? Or are the penalties so severe that they wouldn't go down that path?
I think you referred to the article in the DIGITIMES, where in the headlines, it was said -- on the summary, it was said that TSMC renegotiated with all suppliers pricing down to minus 10% during Q4. When you go to the end of this article, you would realize that there is a sentence that they were not successful at wafers. And by the way, we talked about plenty of stuff to TSMC during Q4, but never about pricing for wafers.
Sorry. And so coming back to the question, would you think the penalties are too severe for them to play you guys off each other and squeeze you on price in future though?
I think when customers are under the impression that they don't need the wafers they signed up for in contracts, they will come to see us and talk about that. But the contract is not only pricing, it's very much -- it's much more than pricing, and we are not willing to talk about prices and we are not willing to talk about total volume. If there is a huge pressure and we might postpone quantities from today into the future, this is something where we might think about, but not changing total volume or pricing.
Got it. And any thoughts on the utilization by diameter?
Yes, I have, but I won't share it with you, I'm sorry.
The next question is from Holger Schmidt, Metzler Capital Markets.
I have 2. The first is how do you see the ramp-up of your new capacity? Can we expect the cost burden from inefficiencies during the ramp-up phase? And secondly, earlier you talked about your most important end markets. Could you give us a more detailed revenue split that you get from these end markets?
Well, we never gave the revenue split by end market. We -- when we split the market, we speak to what is published in quantity area-wise by IHS. And everybody knows that 65% around is related to 300-millimeter, 25% to 200-millimeter and 10% to smaller diameters. This is not revenue. This is area sold. This is the industry, and Siltronic is relatively typical for the industry. It's also known that 2/3 of 300-millimeter is memory. So 2/3 of 2/3 is something like over 40% of the total market in area is memory. And with that, it's the most demanding market. So this was your second question. In the meantime while answering it, I forgot the first one.
The ramp-up costs.
Ramp-up costs. The ramp is going very well, a little bit ahead of schedule. All the investment is -- all the execution was in the budget, and there are basically no ramp-up costs.
And we have another follow-up question from Amit Harchandani, Citigroup.
Just a quick one for Rainer, please. With respect to the tax rate for this year, you've guided us to 15% to 20%. If I remember at the Capital Markets Day previously, you've talked about it going up to 25% in 2019. Could you maybe help us give some guidance in terms of how we should think about tax rate beyond 2019? What would you recommend us to use for the years beyond 2019 cash and P&L?
Yes, sure. Amit, there were basically 3 changes since the Capital Markets Day. The one obviously in the U.S. with the corporate tax rate coming down from 35% to 21%. Number two was that in our -- in one of our locations in Germany, the municipal tax rate was lowered a bit. And the third reason simply being that whenever we extend, we extend the majority in Singapore. And Singapore has a very preferential tax rate or for some parts of our business, even the tax rate, tax break. So going forward for the next years, we definitely see a tax rate in that order of magnitude 15% to 20%.
There are currently no further questions. [Operator Instructions] There are no further questions. I hand back to the speakers.
So thank you very much for participating in today's call. Our next publication will be our Q1 results on the 3rd of May, and we wish you all a good day, and I'll just hand over to Chris for...
I would like to make an additional remark because there was a question of -- regarding loading by diameter, and I was -- and I said yes, I know it, but I won't share it with you. In detail, that's still right. But keep in mind, we are still highly loaded. Thank you.
Thank you very much, Chris. Thank you very much to all participants, and have a nice day. Goodbye.
Ladies and gentlemen, thank you for your attendance. This call has been concluded. You may disconnect.