GlobalWafers Co., Ltd. (6488) Earnings Call Transcript & Summary

Good afternoon, everyone. Welcome to today's conference call. And Roland, please begin the call and I'll be standing by for the question-and-answer section.

Thanks, Frank. Ladies and gentlemen, welcome to GlobalWafers' Fourth Quarter 2019 Earnings Conference Call. I'm Roland Shu, Citigroup's semiconductor analyst. With me today are Doris Hsu, Chairperson and the CEO of GlobalWafers; and William Chen, Vice President and the Spokesman of the company. Today's call will include prepared remarks by Doris and William, and both of them will join the question-and-answer section later. The call is expected to last for 1 hour to 6 p.m. Now we are ready to start the call from Mr. William Chen. William, please proceed.

Thanks, Roland. Hello, everyone. Welcome joining GlobalWafers' 2019 Earnings Call. I'm William Chen, GlobalWafers Vice President and the Company Spokesman. We also have Doris Hsu, Chairperson and the CEO of GlobalWafers in this call. Doris will give us the executive comments later, and then I will present the company overview, industry overview and the 2019 full year performance update. The final Q&A session will be hosted by Doris. For today's presentation file, same as usual, we have just uploaded on company website around 2 hours ago. If you do not have the file on hand, please access into our website to get the most updated file. Please note that some information during our discussion today will consist of forward-looking statements, which are applied throughout the call and this presentation. These are subject to significant risks and uncertainties. Actual results or trends could differ materially from our forecast. Please refer to the safe harbor notice in our presentation, Page 2, disclaimer. Now I'd like to hand over the call to Doris for the Page 3, executive comments. Doris, please.

Thank you very much, William. Good afternoon, ladies and gentlemen. Thanks a lot for joining GlobalWafers' earnings call for 2019 annual financial result today. First of all, during this unprecedented time of the spread of coronavirus, I wanted to provide an update for GlobalWafers in the impact in light of this epidemic. We have implemented travel constraints from mid of January early -- very early this year. And then from end of June, we strictly implement travel ban. Now all of our employees who are able to work remotely from their homes are encouraged doing so. So so far, expect -- except our fab in Kunshan, China, the operation of all our other 14 fabs does not have any impact from COVID-19. The fab in Kunshan has resumed its operation from mid-February. Its production now has already recovered to around 80% now, and it will keep improving week by week. However, at 10 p.m. last night, Malaysia Prime Minister just announced a national quarantine policy for 14 days, starting from tomorrow to end of March. As you know that we have a fab in KL, so definitely, we have to follow the national quarantine as well. But as up to this moment, we are still trying to clarify some key detailed regulations and rules of this urgent quarantine -- national quarantine policy. We haven't really clarified all details yet. We will have a better picture later this afternoon. So no matter what the result is, what the policy is in Malaysia, GWC will very flexibly adjust our worldwide capacity allocation plan to minimize the impact from all related disruptions. So far, up to now, I think the COVID-19 has been spreading for more than 2 months. But so far, there is no order cancellation or shipment pushout requested by any of our customers. Although all of the operation of GWC group is doing okay so far, we never underestimate the potential impact of the outbreak. We think that the outbreak of COVID-19 does bring quite significant uncertainties to the world economics. If COVID-19 outbreak cannot be contained effectively, it is assumed to dampen global economics significantly. In near term, we think that end-market demand will be affected. Global economics will be further dampened. And what's even more complicated is that FX rate will be very volatile and will be very hard to manage. But for long term, the deployment of advanced semiconductor technologies such as 5G, cloud, server, memory, AI, CIS, all of these advanced technologies will significantly increase and accelerate because these advanced technologies will play a very important role to minimize or diminish the impact of this kind of black swan. We never know in the next several years. So we believe that the demand for long term for -- after the disease is over, I think the demand for semiconductor silicon wafer will keep increasing. Advanced wafer process technology development will even be accelerated. That's our view for COVID-19. Now I would like to brief some key overall performance of GWC in 2019 and some key projects, the status of some key projects. First of all, GlobalWafers' overall financial performance in 2019, I would like to update this part of our overall performance in 2019. Again, 2019 is a record-breaking year for GlobalWafers. Except revenue, GlobalWafers' 2019 reached all-time high on almost all index, including the following: gross margin, gross margin percentage, operating profit, operating profit percentage, net income and net income percentage, EPS and EBITDA percentage and also EBITDA amount. All these index are all-time high, in 2019 are all-time high, except revenue. Our revenue in 2019 was the second highest in our record. The highest was 2019's revenue, which was 1.6% higher than our 2019. So all -- so I think all -- our overall performance, 2019 GlobalWafers' overall financial performance is really not bad. I'm very pleased to -- especially, I would like to highlight that I'm very pleased to report that GWC 2019 EPS reached as high as TWD 31.35 per share, which is definitely our all-time high. And today, we just had our Board meeting and the dividend -- our plan is to -- the dividend will be TWD 25 per common share. And this payout ratio is as high as 79.7%. I think after our AGM approval, this dividend will be paid out in Q3 this year. Another schedule thing I would like to highlight as well is that starting from 2020, we will make the dividend payout once every 6 months. That means that we will make 2 times payout every year. That means the dividend for the first half of 2020 net income will be paid in Q1 2021, and dividend for the second half of 2020 net income will be paid in Q3 2021. That's an update for our overall EPS and our dividend payout amount and schedule. Okay. Next item I would like to update to everyone is our new fab in Korea. We finished our expansion in Korea in -- at the end of October last year. And starting from November 22, we started our tool moving. So the construction has already completely completed. And the tool moving is still ongoing. Our plan is to start the whole ramping up of the new fab from second half of '20. Now it's in the tool-moving stage. Due to COVID-19, actually, the tool-moving schedule, tool installation schedule will be a little bit delayed because many of the tool engineers are from our suppliers, meaning that from a lot of countries outside of Korea. And because of the travel ban and a lot of special prevention as containing measures taken by MKC, we decided to add those foreign engineers to delay their schedule to Korea. So I think our tool installation in our Fab2 will delay a little bit, not much, but a little bit, depends on the disease, the outbreak, the status of the outbreak. So our forecast is that maybe starting from end of Q2, we will be able to start preparing some sample for customer qualification. And we hope that starting from second half this year, we will see some significant ramp-up. So that's our Fab2 in Korea, the status of our Fab2 in Korea. And the third item I would like to update is the status of our repatriation of our offshore funds. In Q4 last year, we announced that we will wire around USD 300 million back to Taiwan. Now we have already wired -- has already confirmed this one that we have already wired USD 355 million back to Taiwan. And this fund will be used for 4 main categories. Number one is to expand the capacity for advanced process. This is mainly for 300-millimeter advanced products. And the second area -- second category is to use the fund to increase our compound wafer capacity, which include silicon carbide, semi-insulating silicon carbide and some gem nitron silicon as well. And the third category for the fund is to enhance our Taiwan R&D resources for product differentiation and also for advanced automation as well. The fourth category we are going to use the fund for is to invest in renewable energy for CO2 emission cut. This is very important for the whole group's overall commitment to the society and also for the industry as well. So all these tasks, the above 4 tasks will be completed within 5 years. So it means that by end of 2024, all those 4 categories will be completed. And let me move to another very important project, 300-millimeter SOI. We just signed an MOU with an RF leading company, GlobalFoundries, for a long-term strategic supply for 300-millimeter SOI. In light of the market evolution toward next-generation RF applications, we are very pleased with this opportunity to develop 300-millimeter SOI process with our partner for these very promising applications. Last but not least, inventory correct -- inventory -- industry inventory status is what I want to update to everyone. The inventory correction continued in the second half of 2019. Most of our customers Q-on-Q in the DOI level are slightly down in Q4, slightly down or flat. Inventory level is improving. I think the whole industry's inventory is improved. It's not as bad as the first half of 2019. It's improved, and this will keep improving in 2020 as well. The proliferation of new innovations of our new application, new product will further work off the inventory in 2020. Of course, there is some uncertainty, including the potential disruption or impact from COVID-19, but we think that in general, 2020 inventory -- industry inventory will be further improved. Okay. So that's all I would like to highlight and share with all the -- all of you. Thank you again for joining us this afternoon. I will hand the microphone to William now, and he will present further details with more slides. Thank you very much.

Okay. Thank you, Doris. Let me continue the company and the industry overview, also 2019 full year operational performance results. Page 6, company overview. GlobalWafers was a spin-off from our company SAS, Sino-American Silicon Products Inc. in 2011. We are #3 silicon wafer supplier in the world. There are a total 16 operational sites in 9 countries to produce full-scale products, headquarter locating in Hsinchu, Taiwan. For 300 millimeter, we have total 4 mass-production fabs now in Taiwan, Korea and Japan to provide multicountry, multisite supplying to our worldwide customers. Besides, from this 3 -- quarter 3, our new fab in Korea will also join the production and expand GlobalWafers' total 300-millimeter wafer capacity more than 1 million pieces per month. At the same time, we are continuously improving operational efficiency. As of February 2020, we had managed the worldwide employee head count down to 6,700. Now we just shared some pages of industry overview. However, please note, all these market research information were all collected by 2020 February that were not considering the impact of COVID-19 revelatory. Please turn to Page 8, worldwide semiconductor revenue. After the super cycle courses in 2017 and '18 with 21.6% and 13.7% Y-o-Y growth, respectively, 2019 was dropping around 12.8%. In 2020, all regions are forecasted to grow again with the overall market up 5.9%. For 2021, it is also estimated another 6.3% market growth in 2020. However, again, this 2020 and 2021 growth were estimated before COVID-19 impacting. Page 9, semiconductor capital spending rebound history. A stable semiconductor capital spending as the coefficient of semiconductor industry cycle will be depicting an upturn in near future. Page 10, semiconductor capital spending resume climb. Capital spending is anticipated to resume its climb in 2021 and keep growing for 3 years to 2023, mainly because of rising complexity in downstream applications increases this capital investment for manufacturers to accommodate such innovations. Page 11, market CAGRs of major product categories. For the market CAGRs, top 3 product categories in 2019 to 2024 are: #1, MOS Logic with 6.6%; #2, Analog with 6.5%; and #3, Memory with 6.1%. Page 12, positive growth in IC products. In 2020, NAND Flash, Logic and the DRAM occupied top 3 fastest-growing IC segments with estimated 19%, 13% and 12% growth, respectively. From the right-hand side table, a total of 26 IC products are forecasted to show growth in 2020, a very strong turnaround from 2019 when only 6 product categories had positive growth. In 2020, 5 products are expected to enjoy double-digit growth, an increase from 4 in 2019. Also 21 product categories are expected to have the single-digit growth; comparing to 2019, only 2 products. Page 13, 5G worldwide coverage. For 5G worldwide coverage, technology advancements enable the rapid coverage increasing up from less than 1% of the worldwide population in 2018 to 65% in 2025. Besides, at the right-hand side of the table, by 2025, it is expected that 45% of total mobile data traffic will be carried by 5G networks. 5G is the fastest deployed mobile communication technology in history. Also believe, because of the COVID-19 impacting, the 5G rolling out will be speeded up after worldwide disease is stabilized. Page 14, global automotive semiconductor market. On the left-hand side table, the global automotive semiconductor market should grow from USD 42.5 billion in 2018 to reach USD 70.6 billion by 2022 at a CAGR of 13.4%. On the right-hand side table, the top 2 fastest growth automotive components are: #1, DSPs, representing digital signal processors, used in infotainment and audio equipment; and #2, optoelectronics used in ADAS, representing advanced driver assistance system, and driver monitoring system. Page 15, attractive opportunities in SOI market. As market research, SOI market is expected to grow from USD 0.9 billion in 2019 to USD 2.2 billion by 2024 at a CAGR of 19.6%, propelled by the growing adoption in mobile technology, infrastructure, IoT and automotive driven by 5G commercial deployment. Now let's move to 2019 financial performance update. Page 17, 2019 financial highlights. About 2019 financial performance, as Doris highlighted earlier, 2019 was a record-breaking year. Below are the brand financial highlights. The first all-time high, in 2019, GWC reached an all-time high record in EBITDA percentage, gross margin and gross margin percentage, operating profit and operating profit percentage, net income and net income percentage, also the EPS. All these best results were reflecting correct operational strategy and supply chain management efficiencies. The second, EPS and the dividend. 2019 EPS amounts to the historical high, TWD 31.35. It's appropriate to TWD 25 as dividends. The payout ratio, 79.7%. The dividend payout, once in every 6 months, estimated schedule in Q1 and Q3 2021. The third, financial leverage. By 2019, we had sufficient net cash position, TWD 32.822 billion. We also kept a very high prepayment on hand, TWD 20.393 billion. Page 18, financial highlights, 2019 versus 2018. As mentioned earlier, GWC reached the second best revenue in 2019, just 1.6% less than the record-high revenue in 2018. So was EBITDA, TWD 22.648 billion, the second best for the company. For the other figures, EBITDA percentage, EBIT, OP and OP percentage, net profit and net profit percentage, also EPS, GWC all reached the record high in 2019. Regarding 2019 ROA percentage lower than 2018, it's mainly because of LTA much increased in 2018 as a result of big increase in assets, then drove down the 2019 ROA percentage. 2019 ROE percentage dropping was also caused by 2018 assets' big increase, which have been explained just right now. Page 24 (sic) [ Page 19 ], revenue and the gross margin. Given the fourth quarter market continued softer with Q-o-Q revenue declined 5.6%, however, 2019 Q4, gross margin percentage still increased 0.4%, from 37.8% in Q3 up to 38.2% in Q4. There are 2 main reasons: the first, we kept higher percentage of LTA execution continuously; the second, we had higher ASP product mixing, especially fully production running for the highest ASP product, 12-inch Epi wafers. In overall, 2019 full year gross margin, GWC achieved the best-ever full year gross margin percentage of 39.3% since establishment. Page 20, EBITDA and the EPS. Regarding EBITDA, despite of headwind and the macro uncertainties through the whole year, we reached the all-time high EBITDA percentage as of 39%. About EPS, they were amount to TWD 31.35, the highest ever. Page 21, financial leverage. 2019 capital spending on cutting-edge R&D and 300-millimeter capacity expansion, that slightly lowered the cash position, yet still maintained high level. The bank loan increased owing to much higher dividend distribution in 2019 and 2018. Prepayment net amount decreased some, yet still keeps at a high position. Page 22, dividend payout. Same as last year, we are planning to have a high dividend payout policy, TWD 25, 79.7% payout ratio to bring out profitable earnings for shareholders. Page 23, multiple dividend distribution. As announced in last con call, GWC is going to change dividend distribution from one per year into twice per year. This dividend payout in every 6 months is a steady income stream for shareholders, beneficial for reinvestment and flexible use of funds. It's also delivering a robust track record of dividend payments reflecting GWC's commitment to share the constant growth with shareholders. The dividend distribution schedule will be 2020 first half dividend to be distributed in 2021 Q1, and the 2020 second half dividend to be distributed in 2021 Q3. Regarding 2019 free cash amount lower than 2018 free cash, there were 3 major reasons: the first, higher dividend payout, TWD 25 in 2019 and TWD 10 in 2018; the second, higher CapEx spend in 2019 and 2018; the third, lower prepayment collected in 2019 and 2018. However, actually, as mentioned earlier, company reached the record high operating profit and the net profit in 2019. Page 24, income statement. In this income statement, just to recap, GWC 2019 all listed items. Revenue and EBITDA were second best in company's history. Gross profit and margin percentage, EBITDA margin percentage, operating profit and margin percentage, profit before tax and margin percentage, net profit and margin percentage, also EPS, were all-time high in company's record. Page 25, balance sheet. We were continuously increasing company assets in 2019, up to TWD 96.6 billion, which was around TWD 6.8 billion increase in then 2018. We also reduced TWD 191 million inventory in 2019 from TWD 7.0 billion in 2018 to TWD 6.8 billion in 2019. For the short-term loan increasing to TWD 9.9 billion, that was caused by higher dividend distributed in 2019 Q3. Above is my presentation for 2019 financial results. Thank you. Next, I would like to hand over to Doris for the Q&A session. Doris, please.

Thank you, William. Okay. Thank you very much for joining us. These are our slides and our key information. So yes, any questions will be welcomed. Thank you.

This is Roland Shu. Yes, I think as Doris said, we are now ready for questions. [Operator Instructions] Operator, please open the line for questions now.

[Operator Instructions] And our first question has come from Haas Liu with Crédit Suisse Taiwan.

So Doris and William, congratulations on the good results. My question would be on your business outlook factoring the impacts from the coronavirus. With your base case still assuming a pretty positive industry outlook beyond the coronavirus, could you quantify your expectation for the shipment pricing and also the long-term agreements in 2020, if you're factoring a scenario with a prolonged coronavirus impact?

Thank you very much, Mr. Liu. I think, first of all, it's very hard to -- for us to gauge how significant the impact from COVID-19 will be to the industry. What we can share for now is that basically, customers are -- they are -- actually, they are eager to get more wafers from us. That's what they are doing. But we know very well that part of the demand is for safety inventory. So that's our view. And I think for short term, maybe from Q2, Q3, maybe the demand from our end customer will be adjusted a little bit because of the virus, the coronavirus. So we don't really know that how big the adjustment will be. So far, we haven't received any clear instruction from any of our customers asking for cancellation or shipment pushout. But we expect that if the situation getting worse, then maybe this kind of request will be coming someday in Q2 or Q3. But for long run, I think, as I said in my message, I think for silicon wafer, especially for advanced technology, advanced process, I think those technology development will be accelerated because of the advancement of the technology will be one of the very important solution for any black -- any disease or any new black swan like this. So for 2020, our overall LTA status and ASP price development, our view now is that based on our current information, our LTA coverage in 2020 will be lower than 2019. This is for sure, but it's not a big drop. If you check our prepayment amount, I think you can see that our prepayment amount dropped not that significantly. It's a little bit lower than 2018's prepayment balance, but it's still high. So that means that our LTA coverage will be a little bit lower in 2020. It will be lower than 2019, that is for sure. So this is the first point I would like to share. And price-wise, as William said, we have the LTA, LTA price remain no change, remains the same. But as I explained that our LTA coverage is getting lower, so that means that our spot orders will be more than last year. And fortunately, spot order price this year is higher than the spot order price last year. But spot order price is still lower than our current LTA price. So I mean LTA price is higher than today's spot price and higher than last year's spot price. That's the current status. So if we put all this -- if we factor in all these price and LTA coverage issue, I think our assumption is that 2020, our ASP will be slightly lower than 2019. Our overall ASP will be slightly lower than 2019, but this gap is very, very small. It's definitely single digit and very low single digit. That's the ASP trend. Of course, it's very difficult for us to really factor in COVID-19's impact. We don't know that how much impact would that be and which customers will have the impact more than the others. The reason I talk about this is that if customers adjust their demand for the second half, then maybe the ASP forecast will be different from what we are expecting now. So that's all I can share for now. Thank you very much, Mr. Liu, for your question.

Okay. Just a quick follow-up question on your comment on the LTA. What is -- with your customers prebuilding roll wafer inventory well, and market demand will probably slow down potentially impacting the semiconductor industry from 2Q, do you think you will need to provide more flexibility on the LTA delivery schedule from current quarterly shipment to biannual shipment?

Yes, I think that is possible and I think that is reasonable. But that doesn't mean that we have already made any conclusion for now. We -- as a long-term, long-standing partner for all of our customers, we definitely will try our best to be as cooperative as possible on the shipment schedule. But so far, no firm plan. As I said earlier that, actually, as of today, none of our customers is asking for any cancellation or pushout, rescheduling the shipment. So many of them are even asking us to pull in our shipment. So we don't know how the things will be going. So even for Q2, right now, we have received orders. Our Q2 orders are still very healthy. So far, it looks okay. But the whole development, the COVID-19 development is very fast. So it seems that the situation is changing every day. We don't know that maybe a couple of days, a couple of weeks later, maybe American customers, European customers, maybe some of our customers will have further discussion with us. But as of today, our Q2 outlook is still quite normal. But it's changing every day, so we don't know yet.

Okay. My second question would be regarding your long-term agreement renewal. As the supply/demand seems to stay unvariable for the overall industry, what should we think about your strategy on the new LTA and renewals beyond the existing ones? Will you consider to accept a lower wafer pricing to lock in more demand?

No, I don't think so. So far, we do have some new LTAs under discussion, but those new LTAs are not ordinary products. We are talking about special products and advanced -- especially for advanced node, advanced products, those are the products we are talking about. But if the price is too low, then -- lower than our expectation, then that doesn't make too much sense for us to lock up any demand by committing very low price. So that's not our strategy. We will not sacrifice our price to lock the quantity because from our viewpoint, 2021 will be very tight again. I'm sorry, Mr. Liu, if there was no COVID-19, actually, I think starting from the second half this year, I think the supply/demand will be very tight. We will start seeing some very tight. A matter of fact, starting from Q4, November, December and Q1 this year, we have already been complained by many of our customers for not being able to ship enough wafers to our customer for some advanced products. So we see the tightness. We know the tightness for some spec already. So if there was no COVID-19, actually, second half should be already very tight. But because of the COVID-19, it's hard for us to gauge the potential impact. So what we can see now is that definitely, this disease will be over maybe in Q2 or a couple of more months. But definitely, the demand will be strong again. 2021 definitely will be very tight again, the whole supply will be tight again. So that doesn't make too much sense for us to lock up -- to come in any low price, lock up the capacity by offering lower price. So that's why I say that we don't have -- that's not our plan to lower our price to lock up future capacity. Thank you, Mr. Liu.

And our next question is coming from Donnie Teng with Nomura Taiwan.

My first question is regarding to capacity expansion and the potential investment plan. So you mentioned about you will try to have some samples in second quarter for the Korea fab. And I'm just wondering, at the current time point, whether do you have any idea whenever you can ramp up to like 150,000 wafer per month capacities like in the first quarter next year or it could take longer? And regarding to investment plan, could you remind us what kind of CapEx plan this year or next year or what kind of depreciation cost would be, if any? And whether it has considered your investment plan by utilizing the offshore funds? This is the first question.

Okay. Thank you very much, Donnie. First of all, our capacity, I think our Korean capacity is -- because, as I explained, that the installation will be delayed a little bit due to some foreign engineers are not allowed to work in our operation, so that will delay a little bit, and we will be able to start submitting some sample for customer qualification. So according to our plan, I think by end of 2020, our capacity should be very close to full capacity. But that depends -- pretty much depends on the qualification status and the impact of COVID-19. If the impact -- if the whole disease is over very soon, then I think customers will accelerate the overall qualification procedure, then we will be able to ramp up very rapidly. So that's our view. But again, this will be affected by COVID-19, slightly affected by COVID-19. And for the CapEx investment, this year, we will have much higher CapEx than last year. But that doesn't mean that we will have new CapEx, new money. No, that most of the CapEx in 2020 are for Korea, MKC. So the CapEx number we put in our finance, that is based on the real investment payout schedule. So that means that we completed construction in Q4 last year. So most of the payment will be made this year and some small part will be made in 2021. So 2020 will be the peak of our CapEx especially for Korea. So this year, according to our forecast, I think this year, the total CapEx will be slightly higher than TWD 10 billion. Compared with last year, I think that's a big increase. And depreciation-wise, it's quite stable. I think we will -- in the past several years, I think our depreciation every year is always around somewhere around 4 point -- or close to TWD 4.5 billion per year. That's our annual depreciation for the whole group. I think this year, it will increase a little bit because of some tools we have already installed, we pay out and we buy off and so we will start doing the depreciation. It's not only for MKC, also for some other debottlenecking. And in 2019, we had some capability improvement, that kind of new tools as well and some advanced process, some special investment, inspection machines invested last year. And those will -- those new tools will be one of the key factors of the increase of the depreciation this year as well. So -- but that's very mild, that's not really very big. Okay, thank you.

My second question is regarding to the SOI MOU with GlobalFoundries. So as the GlobalFoundries operation is not pretty stable, and previously, they plan to build a fab in Chongqing to do SOI wafers, but now like suspended. So is there any risk that we have MOU? Of course, MOU does not mean we are really entering into mass production. That's why we signed a formal contract. But considering this customer's track record was not pretty good, and they probably also asked Soitec to expand capacity in Singapore as well. So I'm not sure what kind of commitment you are seeing right now and what kind of demand outlook you think that will drive necessary to expand the capacity in the future.

Yes. Thank you. That agreement with GlobalFoundries is in MOU. So far, it's still in MOU. And of course, MOU is nonbinding, that's current status. But both parties are very serious for this project because that's a win-win for both party. They need very stable vertical integrated supply. That means that we have InGa, we have wafer, we have SOI process, so they want to have -- they want to secure this long term -- this vertical integrated turnkey supply, 300-millimeter SOI supply. And for us also we want to reenter 300-millimeter SOI industry, and we want to play a more important, more aggressive role for RF applications. So that is -- this project is very good for both parties. And we understand the potential risk of -- that we also of course, we know Chongqing project. We know some cases before of our partner. But for this case, I think both parties are very serious, and we are very confident that we will work out a very good project here. From our position, Donnie, GWC always is very conservative and very cautious on, especially for downside protection. So we definitely have some -- we require some very important commitment as well. Due to the NDA, I'm very sorry that I cannot disclose further details, but I just wanted that you know that we have downside protection. I think risk is well managed.

And our next question is coming from Sunny Lin with UBS Taiwan.

Congratulation for the good results in 2019. So my first question is regarding the long-term supply versus demand outlook. So Doris, you just mentioned that you expect going to second half of the year, the industry supply could be back to a tight situation, assuming a limited impact from COVID-19. So I just wonder, based on the current visibility, if the industry supply is going to be tight again in the second half of the year or 2021, when do you expect to start engaging with customers regarding the potential new brownfield or greenfield investments?

Okay. Thank you very much, Sunny, for your question. I think this year, maybe second half, I think because of the COVID-19 up to this moment, I think it's very hard to ignore the impact from COVID-19. So I don't think that second half will be -- supply/demand will be tight as what we've been -- what we expected earlier. So -- but 2021 should be the year, the timing that the supply/demand will be tight again. So far, GlobalWafers don't have any plan to have any brown -- greenfield expansion in 2021. So far, I think that's not the plan. What the plan we have, we know -- according to our expectation, I think that 2021 supply/demand will be tight. So what our countermeasure for this, our plan is to, first of all, as I explained earlier, that because of the repatriation of the funds back to Taiwan, we are going to have more investment in Taiwan for advanced technology, especially for 300-millimeter and a lot of special technology. So -- and also we will have 300-millimeter SOI. That will be another expansion. And we will have more automation and capability improvement and also productivity improvement. So our solution to deal with the tightness of the supply/demand in 2021 is to improve our overall operation efficiency and also to expand some most advanced technology, some special process or most advanced technology instead of to have a greenfield expansion. So that's where we are. And even for brownfield expansion, I think that so far, we don't have any aggressive plan. Of course, the one in Taisil, in Taiwan Taisil, that is -- we do have to expand some clean room, and that's in our campus, so it's not really a greenfield or brownfield. Yes, thank you.

Got it. Just wanted to follow up on your offshore funds of USD 355 million. So you mentioned that you plan to invest over the next 5 years. So just wonder how much capacity increase should we expect from this investment?

Yes. When I said that in 5 years, that doesn't mean that the money will be used in 5 years. So basically most of the money will be used in year 1. We have to order and then we have to have the tool moving maybe in year 2 and we start ramping up. So 5 years means that in the 5 years odd, we want to see the full output and a lot of things. So that's what I said. And 5 years means that, especially we have some R&D automation process, so before the capacity expansion, actually, we are going to start advanced technology expansion from this year. Most -- almost all of the money for expansion will be used this year. Despite the weakness caused by COVID-19, we will do the expansion. But that's not really increase our capacity, so we will ship more 300-millimeter wafers, no, that's not the plan. The plan is that maybe total shipment quantity-wise, very close, but we will have more advanced Epi instead of more polished wafer, and we will have more advanced SOI wafer instead of polished wafer. So total wafer amount may be very close. Total wafer numbers or square-inch-wise will be close, but the product mix will be very different. It will be migrate from ordinary polished wafer to high -- much higher value-added, very difficult, advanced Epi wafer and also change from ordinary polished wafer to much more expensive SOI wafer. So that's the plan. But the square-inch-wise may be very close. I hope this -- I hope I made myself clear enough. Sorry about that.

No problem. That's very helpful. So very quick for the second one. So just want to quickly follow up on the long-term agreement. So you mentioned for this year, whole year, the LTA coverage may be coming down slightly from 2019. I wonder if you can give us any color regarding first half versus second half. Should we expect maybe flattish coverage for first half of this year or maybe coming down more meaningfully going into second half? That's all for me.

Thank you, Sunny. I think that's a very good question as well. Q1 this year will be -- so far, Q1 this year will be the highest -- LTA coverage-wise, Q1 this year will be the highest this year because starting from April, some LTA -- our LTA with Japanese customer always starting from April to March next year. So some of the LTA will be expired by end of March this month, starting from next year, that will be another fiscal year for Japanese company. That will go down a little bit as well. So Q1 is the highest coverage. Q2 will be slightly lower because just a couple of Japanese LTAs will be expired then, and then second half will be flat. So no big difference. So Q1, a little bit higher than the other 3 quarters. But in general, if you -- I think the easiest way -- because we have a lot of NDA with customers, so it's very hard for me to detail -- to share the details. But if you check our LTA prepayment balance as of the end of 2019 versus the end of 2018, you will see the difference is very mild. So basically, not so big difference, but it's really lower. Okay. I hope that I give you some picture, some costs about the LTA coverage change. Thank you, Sunny.

Okay. Due to the interest of time, now we take the last question from the last caller. Operator, please proceed.

And the question is coming from Sebastian Hou with CLSA Taiwan.

My first one is, so I'm curious about your assumption for the supply tighter for 2021, even if with COVID-19 may potentially lower the demand base for 2020. Because when we enter 2021, we will actually start from the different base in 2020, which is lower. So I'm curious about your assumption for 2021, the demand and also supply situation for the industry.

Okay. Thanks for your question. I think, of course, as I said, that we don't have crystal ball. We don't know that how big the COVID-19 will affect the whole industry or even worldwide macroeconomics, we don't know how big the impact will be. So that's based on very limited information we have already collected. And also we -- what we view is that, of course, the COVID-19 will affect the end-market demand, that's for sure. Automotive, mobile phone, a lot of demand will be -- definitely will be weakened, that's for sure. But next year, we think that it will take about 1 quarter to -- or maximum, maybe, 2 quarters, the whole disease will be over and then demand will be back again. And it will be even much higher than what we expected at the very beginning without the COVID-19. And another thing also we think that because of COVID-19, we believe that some companies who plan to make further expansion in 2020, maybe they delay as well because of the uncertainty and low visibility, so they delay as well. So that was why we think that 2021 demand should be back. This disease already over, demand should be back and -- but the supply, some expansion by some companies that delay a little bit, push out a little bit and the macroeconomics affected. So the existing players who have strong position with customers, these companies should have better position to increase the demand. So that was the base why we think that 2021 should be -- supply/demand should be tight again. That's our view. But as I said, again, that there's no crystal ball for us, so we don't really know that how significant COVID-19 is going to affect the whole global macros -- macroeconomics.

Of course. Just one follow-up on that. So you talk about the industry players may postpone their supply expansion plan into next year, given the uncertainty we're seeing right now. So does that also apply to GlobalWafers, that particularly your Korean Fab2, that whether you still keep some flexibility to pull -- or to start manufacturing or ramp up the capacity depending on how the COVID-19 progresses? Is that the right way to interpret?

No. That would be -- for GWC, as I said earlier, that all of our plan will keep moving on, no matter how much the COVID-19 affects the industry because our view is that the impact will be short term. So tool installation, qualification, all these actions take time. And as I said earlier, we have already finished the construction, and we have already placed the orders. Tools are already in Korea. We have everything ready. And the only delay now is the tool -- the foreign engineers' installation. So we are not going to slow down our installation. We are not going to intentionally slow down our installation due to the COVID-19, the concern of the COVID-19. No, that's not our status. The only issue we will delay is because of the travel ban. So for so many countries, basically, they don't accept foreigners who visit those countries. So that will affect the installation a little bit. So not for us, I think the investment slowing down, that is not for us. For us, we have already made the decision last year, and we have already spent the money, so we'll keep moving on. But some other companies, if they have some plan to make further expansion this year, I think that maybe they will have to -- they will slow down a little bit. That's our guess. We don't really have any solid information or inside intelligence knowing that which company will cancel their CapEx, so that's not the case. We just -- based on our understanding of the industry, that was the assumption from our understanding. Thank you.

My last question is on the very simple one. Just I want to get your thoughts on the -- what's your view on the revenue contribution, maybe, say, in 2, 3 years' time from the non-pure silicon wafers, including SOI, gallium nitride, silicon carbide, all these things?

Yes, that's a very good question. We internally have a lot of discussion on this as well. I think that will be -- if we think that worldwide -- I think silicon wafer will increase as well. The whole silicon means that proton, SOI, all these applications, silicon wafer will increase. So the whole pie will increase, but compound -- CAGR for compound wafer, non-silicon wafer, definitely will be much higher than silicon. So I think that in 5 years, maybe single digit, still single digit of the silicon wafer in 5 years. That's our view.

Okay. Since we already finished all of this Q&A session, so now I back the microphone to Doris.

Yes. Thanks again for all of you to joining us this afternoon. I hope that our 2019 results is okay for everyone. If you have any follow-up questions, please feel free to e-mail us, and we can have more discussion by e-mail. Thank you very much. Have a good day. Thank you.

Thank you, everyone. The conference call has been concluded. Thank you for your participation.