Sumco Corporation (OTCPK:SUMCF) Q4 2022 Earnings Conference Call February 9, 2023 1:00 AM ET
Company Participants
Hiroshi Shibaya – Managing Executive Officer
Mayuki Hashimoto – Chairman and Chief Executive Officer
Michiharu Takii – Vice President and Chief Financial Officer
Hiroshi Itoh – General Manager, Accounting
Conference Call Participants
Takashi Enomoto – BofA Securities
Atsushi Ikeda – Goldman Sachs
Takato Watabe – Morgan Stanley MUFG Securities
Go Miyamoto – SMBC Nikko
Shigeki Okazaki – Nomura Securities
Shuuichi Nakahara – Tokai Tokyo Securities
Yuta Nishiyama – Citi Group Securities
Yu Yoshida – CLSA Securities
Ryoichi Watanabe – Mitsubishi UFJ Securities
Mikiya Yamada – Mizuho Securities
Hiroshi Shibaya
I am Hiroshi Shibaya of Sumco Corporation. Thank you for your participation today. This is the results briefing for Q4 of the fiscal year ending December 2022.
Before starting the presentation, allow me to confirm today’s materials, which consist of three items. The brief statement on consolidated financial results for Q4 fiscal 2022, the announcement regarding revision to dividend forecast and the presentation deck entitled results for fiscal year 2022. This will be a roughly 1-hour briefing, which will end at 5:00 p.m.
Next, a disclaimer. The estimates, expectations, forecasts and other future information discussed here and shown in today’s materials were prepared based on the information available to the company as of today and on certain assumptions and qualifications, including our subjective judgment. Actual financial performance or results may differ substantially from the future information contained in this material due to risk factors, including domestic and global economic conditions, trends in the semiconductor market and foreign exchange rates.
We will have presentations today from Representative Director, Chairman and CEO, Mayuki Hashimoto, and Vice President, CFO, Michiharu Takii. Hiroshi Itoh, General Manager of Accounting is also on hand. Chairman and CEO, Hashimoto will discuss our forecasts and operating environment, to be followed by an explanation of the financial results by CFO, Takii. We have set aside time for a Q&A session as well.
I will now hand over to Chairman, Hashimoto.
Mayuki Hashimoto
I am Chairman Hashimoto. I will start with the overview on Slide 5 of the presentation. These are the Q4 results. Q4 sales were ¥117.4 billion, operating income, ¥29.7 billion, ordinary profit, ¥27.9 billion, and net profit attributable to the owners of the parent was ¥18.5 billion. The results were well ahead of our forecasts at all levels. The only divergence versus our forecast was the dollar-yen assumption. The yen was slightly stronger than we had expected.
Turning to our Q1 forecast, we are guiding for sales of ¥105 billion versus Q4’s ¥117.4 billion. We expect operating income of ¥23 billion and ordinary income of ¥25 billion. For net profit attributable to owners of the parent, we are guiding for ¥45 billion, reflecting extraordinary gains of ¥30 billion in the form of negative goodwill we expect to incur related to the acquisition of Mitsubishi Materials semiconductor polysilicon business. CFO, Takii will go into more detail on this later.
Our guidance implies a ¥6.7 billion Q-on-Q decline in operating profit. The background is first of all a ¥5.2 billion impact from the ForEx assumption. This accounts for the majority of the decline. In addition, we have factored in higher costs. Overall volume is expected to decline slightly as a result of the absence of the Leap Year Day in February this year as well as lower production levels owing to regular maintenance at Nagasaki. The market correction is challenging, especially in PW. PW is the main product for a subsidiary FST. So, volumes at FST are likely to be down slightly. Having said that, the FST volume decline will not be that large. These factors should be offset by higher prices, which will have a positive impact of ¥5 billion. The combination of the various factors is projected to lead to a Q-on-Q drop in profits of ¥6.7 billion.
Next slide please. Given the strong performance in fiscal 2022, we have chosen to raise the annual dividend guidance to ¥81 per share. This represents a dividend payout ratio of 40.4% versus our previously stated target of around 40%.
Next slide please. This is the market environment for silicon wafers. First, looking back to Q4, demand for 300-millimeter wafers varied quite significantly by customer. I have been involved in this business for a very long-time, but one key feature of the current correction is this variance by customer. Typically, when the market had been weak, demand from all players had been similarly weak. And when the market was strong, all of the players showed strong demand. This time, however, the variance between customers has been marked. Some players have been relatively resilient, while others fell sharply resulting in a large gap between individual companies. For 200-millimeter, automotive demand has remained strong. Although there are industries where there are demand corrections, Sumco’s wafer production remained at virtually full capacity utilization and we have continued to be able to sell everything we produce. In contrast, for 150-millimeter and smaller diameters, there has been a strong correction in consumer electronics applications slightly depressing demand. On Q4, prices for both 300 and 200-millimeter LTA prices are being firmly maintained throughout the industry. There is not much change in spot prices either, although Sumco does very little spot business.
Turning to the outlook for Q1 of fiscal 2023, on 300-millimeter, I commented earlier on this in talking about the situation at FST, but we are seeing a strong correction in memory. On the logic side, however, we had already been seeing a variance in demand by customer. That said, overall, we expect the correction to be relatively mild for logic. If we look at automotive, demand coming from chipmakers focused on auto applications has remained firm. For 200-millimeter in Q1 smartphone demand is of course weak, but demand from automotive and industrial applications remain strong. We expect the trend to be relatively flat in Q1. For 150-millimeter and smaller diameters, the correction in end demand in consumer applications is ongoing. We expect the correction to continue. On price trends for Q1, as appears to be the case for our peers, we expect LTA prices to continue to be maintained. However, there might be some pockets of slight weakness for spot prices. However, Sumco doesn’t have much in the way of spot to begin with. So, spot trends are unlikely to move the needle for Sumco.
Looking at the outlook going forward, for 300-millimeter, we expect that first half will be the bottom for the correction in memory, but we expect second half to be relatively lackluster as well. For logic, the situation is very different from customer to customer. We expect that the correction will be relatively short in duration lasting only through the first half and subsequently showing a relatively smooth recovery. Automotive and industrial application demand should maintain the firm trends seen so far. Overall, we expect Sumco’s production to remain firm given solid demand. Automotive and industrial application demand in 200-millimeter should also remain firm. The correction in 150-millimeters and smaller diameters is relatively large. We expect this to persist for some time.
Next slide please. Slide 8 shows the wafer trend for 200-millimeter by quarter. The red line is the trend for last year. As you can see industry production remains capped at the ceiling. I believe this level of 6 million wafers per month probably represents the limit. The fact that volumes are largely unchanged from last year suggests the industry is running at full capacity utilization with the potential for further improvements in productivity having reached their limit. Given this situation, the industry is not able to increase production volume in response to demand, so conditions are likely to remain stable.
Next slide please. This is the same chart for 300-millimeter wafers. On this chart, you can see that there was a step up in volumes in Q1 2022, but since then, you can see that volumes have remained largely flat at the ceiling. This reflects the fact that the industry has reached the limits of brownfield expansion. The next increase can only come from greenfield expansion. So, until greenfield capacity comes online in 2024, it continues to be the case that volume is unlikely to increase significantly even if demand picks up.
Next slide please. Recently, we have seen many countries indicate a strong interest in developing local leading-edge semiconductor manufacturing capacity. What are we talking about when we say leading-edge? The current leading-edge is 2-nanometer and the current mainstay technology is FinFET, which is the second from the left. But going forward, gate-all-around is considered to be necessary to make progress beyond 2-nanometers. This is a very challenging technology. At this point, only the top three are able to do this and even within the top three, there were differences in technological capabilities. The technology to achieve LDP miniaturization exists, but for gate-all-around, debris or crystal defects are much smaller and the contamination tolerance is quite restricted. Front and back surface flatness is also viewed as very challenging. However, for gate-all-around crystal orientation is yet another consideration. So, achieving this will be a high hurdle indeed. My impression is that things will change dramatically.
Next slide please. This shows demand for server use 300-millimeter wafers based on server unit volumes. Servers are also being used in data centers, so unit volumes should show relatively solid growth. What is weak now is PCs and smartphones. Data centers are growing. Near-term wafer demand from servers is around 1.7 to 1.8 million wafers per month. In 4 years time, demand is estimated to be at around 2.5 million wafers, a hefty increase of 700k or 800k.
Next slide please. This shows demand forecasts for smartphone use 300-millimeter wafers. In contrast to servers, smartphones are seeing a relatively large correction. At the same time, there is the ongoing transition to 5G. 5G silicon consumption is 1.7x larger than 4G. These two opposing trends cancel each other out. So the decline in terms of wafers is likely to be relatively mild. 2023 is likely to be down year-on-year, but there should be a recovery in 2024 with unit volume showing steady growth thereafter in our view.
Next slide please. This slide shows auto sales forecasts broken out by vehicle type. Automotive remains firm, but if we look at EVs, they are expected to account for more than 50% of auto sales by 2026. ADAS adoption is also similarly expected to pickup.
Next slide please. So, what is the difference in silicon consumption between ICE and XEV? Roughly speaking, an ICE vehicle currently consumes about 10 square inches. This compares to silicon consumption of around 3x higher for XEV combined with ADAS. As noted earlier, XEV sales and ADAS adoption are expected to increase. So this growth trend should remain unchanged.
Next slide please. So how much wafer demand is represented by automotive devices? Currently, automotive demand for 300-millimeter wafers is estimated at around 500k to 600k wafers per month. In 4 years time, this is expected to rise to 1 million wafers per month. So, automotive will drive an increase in wafer demand of 400k to 500k. So if we think about overall wafer demand, it is likely to increase by around 1.5 to 2 million wafers per month over the next 4 years.
Next slide please. This shows current customer wafer inventory levels. As you can see, there has been a sharp increase in inventories. Although I can’t really say too much about it, if you look at inventory levels for each chipmaker, there are companies where inventory is up sharply and others where the increase is not that large. There is significant variance in inventory levels between individual companies.
Next slide please. This shows customers 300-millimeter wafer inventory broken out into logic and memory. The bar charts seem to suggest that the increase in wafer inventory and memory makers is more moderate than logic with inventories rising sharply at logic players. However, we are seeing a clear divergence between the winners and losers in logic. Conditions at individual logic makers differ widely. This makes it tough to generalize. This is where logic and memory are at this time.
I think it is better to explain the situation using the graph on the next slide. So how should we analyze what is happening with inventory? This chart plots trends for inventory price and demand for 300-millimeter logic use epi wafers. The orange line is inventory volume on a monthly basis. The rising blue line reflects trends in consumption or demand. What this chart tells us is that for logic, even in the event of a correction, the magnitude of a correction tends to be relatively shallow and generally only lasts for four quarters before a recovery. Near-term, we are entering a slight correction phase although there are differences between individual companies. But even if it is a correction, the level of inventory compared to historical levels is not that high. So even if there is a correction, we would expect the correction to be of relatively short duration and modest in terms of the magnitude of the downturn.
Next slide please. This is the same chart but for polished wafers. If we look at historical trends for polished wafers, when there is a correction, the duration is typically longer at around six quarters. As well, generally, a recovery is very sharp. If we look at the current correction, we can see that inventory levels compared to historical levels are quite elevated. This also suggests that a correction could be more prolonged than logic. We believe that first half 2023 is likely to be the bottom, with the market starting to pickup sometime in the second half. However, our view would be that it would be tough to have high expectations for this year. That said, I would expect to see a solid rebound sometime in Q4. One feature of PW is that any rebound tends to be quite sharp, which means that it is necessary to be prepared near-term the market is in the midst of a relatively tough correction.
Next slide please. This is the outlook for supply demand including historical data, which we always share with you. We expect 2023 to be flat reflecting current brownfield capacity. The industry will start to ramp up greenfield capacity, but the new capacity may be slightly surplus to requirements this year. However, delaying CapEx now would mean that the capacity would not be ready if demand surges in 2024. I believe given this that everyone in the industry will execute on CapEx as planned. This completes my section of the presentation. I will hand over to CFO Takii to talk about details of our earnings.
Michiharu Takii
I, Takii, will present the earnings in more detail. Next, slide please. As touched upon at the beginning of the presentation, both sales and operating profit were up year-on-year. Full year sales rose ¥100 billion year-on-year to ¥441 billion, largely in-line with our expectations. OP doubled year-on-year to ¥109.6 billion. Ordinary profit also roughly doubled to ¥110 billion. As discussed previously, the corporate tax amount increased in the absence of offsets, as a result of the depletion of cumulative loss carry forwards. Also, the net loss attributable to non-controlling interests widen as a result of the strong profitability of FST, which led to an increase in deductions. CapEx was ¥130.8 billion also doubling year-on-year mainly related to the start of greenfield investments, such as utility facilities and the construction of a physical shell, which came up for verification during the year. Depreciation also increased. OPM for the full year was 24.9%, EBITDA margin was 38.4% and ROE was 13.9%. The average ForEx rate for 2022 was ¥131 to the dollar.
Please turn to the next Slide. This is the analysis of changes in operating income. First, the sequential analysis, sales and operating profit were virtually unchanged Q-on-Q so no major moves. Costs rose ¥1.7 billion, mainly the result of increases of around ¥0.5 billion each for labor costs, maintenance costs, materials costs and electric power costs. Depreciation was also up Q-on-Q. Sales related variance was impacted by regular maintenance conducted in November at the mainstay 300-millimeter plant, which depressed volume. However, on the back of the weekend, there was a positive contribution from ForEx, in particular in Q4, when the yen weaken beyond ¥140 to the dollar, resulting in the ¥3.1 billion positive.
On the right, we showed the year-on-year change for the full year. Sales increased ¥100 billion year-on-year. The large increase can in part be attributed to ForEx, which contributed ¥40 billion to the top line. There was also a contribution of more than ¥30 billion from higher selling prices. Volume also increased for an impact of around ¥30 billion year-on-year. We show the analysis of change in operating income in the bar chart below. OP increased ¥58.1 billion year-on-year. Costs increased ¥20 billion. Labor cost increases were a major contributor at ¥8 billion. This is due to frontloaded hiring for greenfield expansion. We increased headcount to expand production for an impact of ¥5 billion. And per employee costs also increase on higher bonus payments as a result of improved profitability, pushing up labor cost by ¥3 billion.
The next largest element was electric power costs which rows ¥6 billion. Surging materials prices raise costs ¥2 billion, and maintenance expenses were up ¥4 billion. Effectively costs related to increased production and unit prices for electric power are rising and materials costs were also up. Depreciation was up ¥8.2 billion. On sales related variance the contribution from prices was ¥33 billion as mentioned before, with volume increases contributing another ¥20 billion, a combination of both 300 and 200-millimeter ForEx had a sizable positive impact of ¥33 billion.
Next slide please. Slide 25 shows the balance sheet and cash flow statement. As a result of higher profits, cash and time deposits on the balance sheet increased ¥34.7 billion. Accounts receivable finished product and work in progress also increased on the back of higher sales and increased sales volume. Raw materials and supplies rose ¥0.5 billion. The reduction in polysilicon inventory within this was ¥9 billion, which was offset by an increase of around ¥10 billion in materials inventories, which we considered important from BCP standpoint, given we’re running at full capacity.
Tangible assets rose ¥73.5 billion on the back of higher CapEx. Other liabilities increased ¥58.8 billion while interest bearing debt was flat. We discussed this previously, about our joint venture subsidiary in Taiwan FORMOSA SUMCO TECHNOLOGY took in advance payments and deposits on LTAs of roughly ¥26 billion. Also, on tax expenses. We discussed this when talking about the profit/loss statement, but the actual tax payment will happen in the next fiscal year. So the increase in other includes hefty ¥15 billion increase in income tax payable.
If we look at the capital account, retained earnings increased ¥49.3 billion, reflecting the impact of ¥70 billion in net profits after deducting the payment of dividends. Down below optically the shareholders equity ratio appears to have fallen slightly, but equity per share improved to around ¥1500. The DE ratio is as shown here. On the cash flow statement, operating cash flow was ¥179.4 billion, while investment cash flow was an outflow of around ¥130 billion. Free cash flow was ¥53.1 billion. After dividends paid and other adjustments. We get to the net increase in cash and time deposits highlighted earlier of ¥34.7 billion.
Next slide please. These are the forecasts for Q1 on a Q-on-Q basis. We expect sales to decline around ¥12 billion. We also projected drop in OP to ¥23 billion. The projection for non-operating income and expenses is positive. We had put on ForEx contracts through Q1 which will generate currency gain. We project ordinary profit of ¥25 billion. Under extraordinary gains we expect to incur negative goodwill on the acquisition of Mitsubishi Material’s polysilicon business. We estimate a difference of ¥30 billion between our acquisition price and the NAV of the business we are buying, which we will recognize in Q1 as an extraordinary gain. After deducting for corporate tax, we forecast net profit of ¥45 billion. We project depreciation of ¥14.5 billion down slightly year-on-year. This reflects the fact that we are into a new year and have made progress on depreciation. We show a number of management metrics at the bottom of the table.
Next slide please. This is the analysis of change in operating profit. Sales are projected to fall ¥12.4 billion Q-on-Q. This breaks down into ¥10 billion negative impact from lower volumes. A ¥7 billion negative from ForEx impact, but a positive ¥5 billion impact from higher selling prices. The result of a step up in LTA prices with the start of a new fiscal year. On the back of this OP is expected to fall ¥6.7 billion Q-on-Q, the main factor is higher costs, we expect cost increases to persist into this year. Inflation continues to push up raw materials prices for an impact of around ¥1.5 billion. We also expect the unit price for electric power to also continue to rise and have factored in ¥1 billion increase. We also expect to see increases in R&D and labor costs.
Base wages are expected to rise but that will happen in April so not Q1. In total, we therefore expect costs to rise ¥3.7 billion, depreciation is expected to improve ¥2 billion as a result of progress on depreciation and the start of a new fiscal year. Under sales related variance the positive impact of higher selling prices of ¥5 billion due to 300-millimeter LTAs is expected to be offset by ¥5 billion resulting from lower volumes. The drop in volume is partly a function of a lower number of days in February, as well as the impact of regular maintenance. Project a negative ¥5.2 billion ForEx impact as a result of yen appreciation due to setting our ForEx assumption at ¥131 to the dollar.
Looking at the analysis of year-on-year change for Q1 on the right, the year-on-year negative impact of costs is expected to be a sizable ¥9.8 billion. Materials and supply costs are expected to rise ¥3 billion year-on-year. We expect electric power unit prices to rise for an impact of ¥3.5 billion. Labor costs will raise ¥1.5 billion year-on-year. Repair and maintenance costs are expected to increase ¥1 billion year-on-year. Effectively, all costs are rising. Depreciation is also expected to increase.
For sales related variance we expect ¥4 billion positive reflecting the positive ¥5 billion impact of higher selling prices with the start of a new fiscal year. We expect a ¥6.6 billion year-on-year positive from ForEx for Q1, based on an assumption of ¥131 to the dollar. All of the above factors combined contribute to largely unchanged OP year-on-year for Q1.
Next slide please. As you can see on Slide 31, you can see the trend has shifted to a slight decline in sales and profits for Q1 that said EBITDA remains elevated. This completes my remarks.
Hiroshi Shibaya
Thank you. We will now open the floor to questions.
Question-and-Answer Session
A – Hiroshi Shibaya
[Operator Instructions] We will start with Mr. Enomoto of BofA Securities. Mr. Enomoto.
Takashi Enomoto
I would like to know if the 300-millimeter LTAs are functioning properly, with demand weaker in the March quarter, particularly for memory are the LTAs ensuring the customers respect their volume commitments? For example, our customers merely extending delivery to sometime later this year, I suspect the customers will respect the LTA prices, but please comment on how well the LTAs are working?
Mayuki Hashimoto
With regard to volume commitments under LTAs, we are seeing some slight adjustments from customers, particularly memory players. We aren’t really seeing customers ask for lower prices, but what they are asking for is to push out the shipment timing for a portion of the volume. In our case, FST is being impacted by a decline in volumes from Chinese customers, but this is unrelated to the LTAs. A customer is now unable to buy even replacement parts for fabrication equipment. As a result, they have reduced production levels significantly, which has led to a decline in wafer demand from this customer.
Apart from this particular customer, however, most appear to expect a recovery at some point. Some have pushed out deliveries to later this year. There are others that have asked to push out volume by effectively adding to the term of the LTA, there have been a variety of responses. However, customers are respecting the overall volume of the LTAs although at the fringe, they appear to have been some shifts between customers. It appears that this is also the case for our peers. Under the terms of the contracts, it is possible to force customers to take volume even if they have cut production levels and have full warehouses, but we prefer to engage with our customers to find common ground. In Q1, we intend to maintain production levels without reductions. To-date wafer makers inventory levels have been very low, so there is a need to raise inventories back to more normal levels as well, so as you can see, that has been the extent of the correction to this point.
Takashi Enomoto
Understood. Thank you.
Hiroshi Shibaya
Thank you, Mr. Enomoto. Next is Mr. Ikeda of Goldman Sachs.
Atsushi Ikeda
In looking at your Q1 forecast, I believe that you are assuming a decline of around 10% for 300-millimeter wafers in-line with your peers. If we look at past history, in the March quarter of 2019, there was an overall decline of around 4%, followed by a more moderate correction. This time, I believe the overall decline to be quite large. However, Sumco’s significant business with the chipmakers considered to be amongst the winners. So relative to your peers, I think you are better positioned. Compared to the last downturn, the March quarter drop is quite large. How are you thinking about the slope of recovery in Q2 and beyond? Also, do you still expect customers to feel that wafers are in short supply in 2024 and beyond? Please talk about how to think about a recovery?
Mayuki Hashimoto
As you know, we have very good relationships with chip makers that are strong in automotive. These players are actually saying that they want to increase wafer intake rather than reduce it. However, the situation appears to be quite different depending on sector, the situation for memory appears to be very tough, with many announcing production cuts, we are seeing some impact from this. That said, the proportion of memory use at Sumco is lower than at our peers, so we aren’t seeing a large correction, but we are impacted by China. For logic, the magnitude of the correction is not that large, with the exception of one customer. What is happening with this customer is not so much the impact of a correction. But instead, this customer may be falling behind. I think what determines winners versus losers, is a subset of each player in the market. Overall, for logic, I think the correction will be mild, and that there will be a recovery in second half. From that perspective, I think the 10% decline expectation is for the total market, including 150-millimeter and smaller diameters, where the decline is likely to be acute. If you look just at 300 millimeters, I don’t think the correction is that large. I think that there isn’t much of a correction in 200 millimeter either. 150 millimeter, on the other hand, is very tough. In terms of the outlook for 2024, I don’t really have any concerns as far as logic goes, but memory may need more time to work down inventories. I suspect memory may be relatively weak for 2023 only really rebounding in first half 2024.
That said, I am not that worried about memory either.
Atsushi Ikeda
Understood. I have one follow-up. On Page 22, the demand forecasts for 2023 is flat, should we interpret this to mean that you expect a decline in first half, but a recovery in second half?
Mayuki Hashimoto
I think first half will be the bottom. I expect logic to rebound in second half. Memory should also improve in second half, but I don’t expect memory to show as strong a recovery is logic. That’s because inventory levels for memory are elevated.
Atsushi Ikeda
Understood. Thank you.
Hiroshi Shibaya
Mr. Ikeda, thank you. Next is Mr. Watabe of Morgan Stanley MUFG Securities.
Takato Watabe
My question is about the grey line for price trends that you show on Pages 20 and 21. When you look at the last 5 years, although demand was down, the price trend on this chart suggests prices did not fall. In actual fact, I suspect there were times when prices did fall. This time, should we expect price trends to be similar because of the LTAs? Please comment on the actual price trends.
Mayuki Hashimoto
When you look at the chart, it clearly shows the prices prior to 2013 were volatile, but with the introduction of LTAs prices remained stable and were not impacted by supply-demand changes. My sense is that customers are respecting the LTAs. I say this because if you look at the peak in 2018, there were customers that signed LTAs. Although it was rare, subsequently there was a customer that did not respect the LTAs. However, at the next peak, this customer was unable to source wafers. We understand that this customer went on an apology tour, but I believe this experience was a lesson to customers that they need to respect the LTAs in times of volatility. I think there has been a broad acceptance that this is the case.
Takato Watabe
Understood. Thank you.
Hiroshi Shibaya
Thank you, Mr. Watabe. Next is Mr. Miyamoto of SMBC Nikko.
Go Miyamoto
With regard to Greenfield investments, you mentioned that a delay in investing would mean you could not keep up if demand in 2024 recovers. Given this, I expect that you will continue to invest in line with your initial plan. Are you still expecting a gradual stepwise ramp up from Q3? If so, will depreciation increase from Q3? If possible, can you talk about the scale of CapEx, you expect next year.
Mayuki Hashimoto
We are executing on CapEx spending exactly in line with our plan. However, FST is lagging slightly as a result of issues in procuring materials. Japan CapEx is exactly on schedule and I have no intention to slow down. On CapEx spend, I will defer to CFO Takii.
Michiharu Takii
As you can imagine CapEx activity will be at a high level with facilities and equipment being brought on site. As such, total CapEx spend for 2023 is likely to be high. There will be a wave of investment in 2023. Japan CapEx spend for this year is likely to be around ¥200 billion. Taiwan is likely to be ¥100 billion this year.
Go Miyamoto
Thank you. Will depreciation increase from Q3?
Michiharu Takii
It will increase of course.
Go Miyamoto
Understood. Thank you.
Hiroshi Shibaya
Thank you. Next is Mr. Okazaki of Nomura Securities.
Shigeki Okazaki
Referring to Page 25, how should we think about costs? You indicated your intention to raise base salaries. For this fiscal year, can you provide a range for how much you expect costs to increase, I do understand that there will be some variables that may change. You also just touched upon depreciation, but can you give us a sense for the scale of depreciation on a full year basis for the new fiscal year? Also in Sumco’s case, I understand that it is your policy not to pass on increases in fuel and raw materials costs, even with LTAs, which is different than your peers, can I confirm that your policy in this area is unchanged?
Michiharu Takii
Obviously, given the expected installation of significant facilities and equipment, CapEx levels on a verification basis will also be relatively high. Depreciation last fiscal year was around ¥60 billion, I would expect to see an increase of around ¥20 billion year-on-year. Obviously, this will depend on utilization levels.
Mayuki Hashimoto
On the question of the degree to which cost volatility is reflected in prices, I think it is the same for our peers. We don’t have commitments down to the level where we agree in advance to tie a certain increase in prices to a fixed scale of cost increases for items such as electric power. That said, if there is a huge surge in costs for items such as transportation costs, we can and do engage with our customers and make adjustments.
Shigeki Okazaki
Understood. Thank you.
Hiroshi Shibaya
Thank you, Mr. Okazaki. Next is Mr. Nakahara of Tokai Tokyo Securities.
Shuuichi Nakahara
With regard to the acquisition of the polysilicon business, does this mean that you will see polysilicon prices decline, or given you will need to invest in R&D, should we temper our expectations? Please talk about the impact that the polysilicon business will have on Sumco.
Mayuki Hashimoto
If we look at the historical trends for polysilicon, once every 12 years to 13 years, the market experiences shortages. Polysilicon is a process industry, so when shortage is hit, everyone rushes out to expand capacity. The last time was in 2009. At that time, some cobalt very large amounts of polysilicon ending up with a huge inventory which was a bad experience for us. We are starting to see shortages in polysilicon now. For instance, for relatively simpler solar grade polysilicon prices used to be around $10 to $15. But at the peak last year, it was over $40. So, we are starting to see shortages on an overall basis. So, first of all, although the acquisition does not fully cover all of our needs, it does ensure that we have access to a steady supply. It also gives us a good idea of costs and allows us to make plans. When the market gets tight, polysilicon tends to see dramatic price action. In 2009 prices got as high as $80 or $90. As prices rise, it can lead to panic buying. By having internal capacity, this means some code does not need to engage in such behavior. Also, when dealing with leading edge, the requirements for polysilicon are very demanding, so quality is a major issue. Rather than having to make detailed explanations to vendors every time and pay a premium for product or run the risk of losing knowhow to external parties, having in-house capacity is a positive. Going forward, we will be able to make high purity polysilicon in-house. Obviously, we have an understanding of the business so costs would not get that high, although vendors would want to charge a premium. Having this capacity in-house will be an advantage in R&D. Also, the situation for trichlorosilane is precarious. Only two companies in Japan can supply it. One is Mitsubishi and the other is Tokuyama. Because it is a hazardous material, it’s almost impossible to import from overseas. If Mitsubishi were to exit that means that Japan would only have one domestic supplier Tokuyama. Given that a significant proportion of what we produce is epi, a situation where there is only a single source would be a problem. So, we also wanted to lock in a supply of trichlorosilane. I expect that there will be moments when we will be glad to have the capacity and other moments where we might feel that having the capacity in-house actually leads to higher costs. When the market has an oversupply, suppliers will say that they are willing to sell cheap. When the market is tight, they press for higher prices. The polysilicon makers could have their customers over a barrel. Without polysilicon we can’t even make a single wafer. Given our experiences, although it doesn’t fully cover all of our requirements, having this capacity in-house is a good thing. There will be better stability in prices and supply. I believe there will be big benefits to having this operation in-house.
Shuuichi Nakahara
Thank you.
Hiroshi Shibaya
Thank you. Next is Mr. Nishiyama of Citi Group Securities.
Yuta Nishiyama
My question is about 300 millimeter wafer demand. Volume is expected to decline 10% Q1 Q in the March quarter, but your demand forecast for 2023 is flat. Hypothetically, if March quarter volumes fell 10% and you assumed that there will be a recovery in second half, that suggests that you would need to see growth of around 20% in second half on and a half-on-half basis to achieve flat volumes on a full year basis. Based on what you know of customer inventory, do you believe that this is achievable? If we look back at the previous cycle, the bottom for semiconductors was early 2019, but the end of wafer inventory adjustments only came in 2020, so there was a gap in timing. This time should we expect that there won’t be much of a time lag between the recovery in semiconductors and a reduction of inventories.
Mayuki Hashimoto
If you have a 10% decline in the first half it is possible to achieve flat growth on an annual basis if second half grows 10%. Our view of 300 millimeter is that the correction is not as large as 10%. If you are looking at PW, then the magnitude of the correction could be larger. But actually PW for discrete components for automotive applications, or IGBT are actually up. So, not everything is down, what’s down is memory used wafers. Within logic PC and smartphone use is down, but demand for industrial applications and datacenter use is strong. The current situation is not one that lends itself to generalization. My sense is that the decline is not as large as 10%, if we assume that it will be flat on an annual basis. In Sumco’s case, the proportion of logic is high. We don’t necessarily have full visibility for the overall market, but that is my view. Customers are also saying they expect a rebound in second half, although this doesn’t apply to memory. Also, only a few logic customers are reducing production. Putting it all together suggests that volume in 2023 will be flat.
Yuta Nishiyama
When you say flat year-on-year, I would like to confirm that you are not saying that it will be flat versus the fiscal year end level, but instead flat compared to a monthly average for the last 12 months.
Mayuki Hashimoto
Yes. But when we get into the second half, we should see positive year-on-year growth.
Yuta Nishiyama
Understood. Thank you.
Hiroshi Shibaya
Next is Mr. Yoshida of CLSA Securities.
Yu Yoshida
Can you provide an image of the direction of quarterly profits this fiscal year based on what you have said so far, you are expecting volume to rise in second half, but have also said that spot prices may come under slight pressure and depreciation will rise in second half. I am assuming that the Q1 ForEx assumption remains constant.
Mayuki Hashimoto
Probably depreciation will rise significantly, so I don’t think we can hope for a performance similar to last year. The reason why we don’t expect sales to fall that significantly is because we benefit from the impact of higher selling prices, given that LTA prices step up in the New Year. So, even if there are slight push outs, I don’t think sales would fall that much. That said, given that we have been executing on our investments, depreciation will definitely be up obviously. The increase in depreciation will weigh down profits, which is likely to mean, we can’t match last year figures.
Michiharu Takii
We expect to see a positive contribution of ¥5 billion per quarter as a result of selling prices rising for ¥20 billion contribution for the full year. But this is likely to be offset by an increase in depreciation of ¥15 billion to ¥20 billion. The remaining question is what happens with cost increases. But if volume were to increase in second half that might mitigate the impact of cost increases.
Yu Yoshida
Thank you.
Hiroshi Shibaya
Next is Mr. Watanabe of Mitsubishi UFJ Securities.
Ryoichi Watanabe
I would like more color on the acquisition of the polysilicon business. On Page 11, when you refer to gate all around you mentioned crystal orientation. Is there any relationship between this and your acquisition of the polysilicon business?
Mayuki Hashimoto
Not really.
Ryoichi Watanabe
Thank you.
Hiroshi Shibaya
Next is Mr. Yamada of Mizuho Securities.
Mikiya Yamada
On Page 11, you discussed gate all around and the transition to 2 nanometer. In addition to crystal orientation, I think there is also added value in layering. Given that Sumco has strengths in epi layers, I believe that you can achieve significant added value. Can you comment on this, or is it still too early to say?
Mayuki Hashimoto
Wafers for gate all around are technologically very challenging. Only two companies are able to supply wafers to the one company that is furthest ahead on gate all around. Other players aren’t even capable of providing samples. This suggests that the transition will represent a very significant change. We are certainly strong in epi and have been able to provide wafers to the customer. We understand that the customer is adding further layers to the leading edge wafers we provide. So, it is necessary to supply layers to support this process. There is also significant process in 3D driving demand for carrier wafers or interposers, so we are also seeing diversification of wafers. Gate all around is very difficult. We struggled, but were ultimately able to develop a wafer. This will be a very challenging era.
Mikiya Yamada
I have high expectations for Sumco. Thank you.
Hiroshi Shibaya
Thank you, Mr. Yamada. Thank you to everyone. We are grateful for your participation today. We will end the meeting here.
Be the first to comment