3436.T3436.TJPX
Loading
Takayuki Komori: Thank you for your participation today. This is the results briefing for the fourth quarter of the fiscal year ended December 2025. Before starting the presentation, allow me to confirm today's materials, which consists of 3 items: the consolidated financial results for the fiscal year ended December 31, 2025, the announcement concerning difference between forecast and actual figures for the fiscal year ended December 31, 2025, and the presentation deck entitled Results for Fiscal 2025, which we will use now. Next, a disclaimer. The estimates, expectations, forecasts and other future information discussed here and shown in today's materials were prepared based on 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 Representative Director and Vice President, CFO, Shinichi Kubozoe. Chairman and CEO, Hashimoto, will discuss our forecast and operating environment to be followed by an explanation of the financial results by CFO, Kubozoe. 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 an overview of the results. I often get scolded for missing forecast, but this time, we overshot quite significantly. With regard to the JPY 5.5 billion overshoot in operating profit, major factors were JPY 2.1 billion from cost reductions, a JPY 1.7 billion impact from ForEx and JPY 1.2 billion as a result of delays to depreciation. These 3 items alone totaled JPY 5 billion. However, despite the increase in sales, the contribution of volume and product mix to profit was only around JPY 0.5 billion. This is because there was a relatively higher level of polished wafer or PW sales. Marginal profitability for epitaxial wafers is actually slightly higher than PW. So despite top line growth, the shift in product mix had an impact on profits. With regard to Q1 earnings forecast, the actual content for the quarter is largely unchanged from fourth quarter. The reason for the JPY 1.5 billion Q-on-Q widening of the loss is the expected negative impact on production volume owing to periodic maintenance at a key plant slated for March as well as one-off maintenance expenses related to the periodic maintenance. This is what is behind the Q-on-Q widening of operating losses. We expect no other major changes from Q4. Next page, please. On dividends, we take into account factors such as free cash flow in deciding dividends per share. While we were in the red, factoring in cash flow and other considerations and given that we have significant retained earnings, we set the fiscal year-end dividend level at JPY 10 per share. Next page, please. This is the trend for 200-millimeter wafers. As you can see, there have been significant declines to this point. The fall is structural, and as such, we do not expect to see a rebound. We must consider countermeasures. 200-millimeter fell 21% in 2023, 13% in 2024 and a further 4% in 2025. Optically, it may appear that there was an uptick in Q4, but this is probably a reflection of a very slight pickup in power management IC, MOSFET for AI. We think that the current conditions reflect the structural reality for 200-millimeter now. Next page, please. This is the trend for 300-millimeter wafers. The overall trend is not as strong as the very favorable trends we are seeing in AI, but 300-millimeter is definitely recovering. 300-millimeter was down 11% in 2023, but was up 2% in 2024 and up 9% in 2025. This is, of course, due to the continued strength in AI-related demand. Other sources of demand are sluggish. However, we are starting to see a pickup in AI-related demand, particularly for memory. While DRAM had already been a beneficiary, NAND has now become necessary for AI as well. I will go into more detail later, but NAND is needed for inference. That said, there are still uncertainties about whether this will drive strong growth going forward. This is because DRAM, specifically HBM, is already in short supply. Because tack time for HBM is significant, it is a significant consumer of fab capacity. Taking capacity away from other types of DRAM has led to a shortage of conventional DRAM. This is on top of the shortages in HBM. So DRAM as a whole is experiencing shortages. For memory players who produce both DRAM and NAND, the higher profitability of DRAM has meant such players have been converting NAND capacity over to DRAM and reducing capacity for NAND. With demand now emerging for NAND, the situation is starting to change. With NAND capacity also falling and challenges in increasing capacity, we have a situation where DRAM customers are actively expanding capacity, which should eventually lead to improved wafer consumption. Given the pickup in NAND, we should start to see changes. That said, it is likely to take around 12 months for the capacity increases to come through. So even at the earliest, I believe it will be late this year or early next year before the gradual increases in capacity kick in. So I expect wafer consumption to grow solidly from this year into next year. Next page, please. The Q4 results were as discussed earlier. However, although 300-millimeter is improving, the overall gains are moderate. Last year, 300-millimeter grew 9%, which is a fair recovery, but it still only gets us back to the peak levels of 2022. Current 200-millimeter wafer conditions reflect the structural reality. On prices, LTA prices are, by and large, being respected. So prices are not bad and are generally flattish. We don't have many 200-millimeter LTAs. So reflecting supply/demand, 200-millimeter prices were generally softer. Setting aside specialty products, commodity 200-millimeter wafer prices declined. The outlook for Q1 is for continued solid volumes in 300-millimeter, particularly for leading edge. Demand for legacy node wafers remains lackluster. You may have seen that TSMC has shifted to using the Kumamoto plant to produce 3-nanometer. This reflects the continued weakness even in design rules that are legacy adjacent, such as 16-nanometer and 28-nanometer. When it cannot keep capacity utilization in Taiwan for these design rules at 100%, it doesn't make sense to produce at these design rules outside of Taiwan. 3-nanometer production in Taiwan, on the other hand, is running at a high utilization. They are working very hard to expand existing capacity, but it's still not enough to keep up with demand, which only further suggests that there is an excess in legacy node products. So the customer is keen to get inventory back to normal levels as quickly as possible. I have already described the situation for 200-millimeter. While there appears to be a slight pickup in power management ICs or MOSFETs for AI, all other applications are similarly weak. Of course, we expect LTA prices will continue to be respected. For spot, 200-millimeter prices are falling with commodity prices particularly under pressure, as mentioned earlier. This is our outlook for Q1. Looking further out into the longer term, there are those that suggest that AI applications are frothy, but I don't believe that, that is the case. The current situation is one where supply is failing to keep up with rising demand. So while our customers' selling prices are rising, I don't think that we are going to see a spike in wafer volumes. I don't know how long the very pricey AI chips remain at these levels, but I don't think that demand is likely to come off much. Until recently, customers were keen to rapidly expand existing capacity on the back of very strong demand for leading edge logic and HBM, but we are now seeing demand for NAND grow as well. As mentioned earlier, this is being driven by an increase in chips used for inference. However, with the glut of legacy products, customers are now seriously making plans to normalize legacy inventory levels this year. Customers are likely to adjust their wafer purchase volumes. However, given that these are customers that have been very respectful of the LTA conditions, we understand their situations. Suppliers, ourselves included, are likely to have little choice but to cooperate with the customers to a certain extent. For 200-millimeter and smaller diameters, as noted earlier, demand is likely to remain at current levels. Next page, please. So what are we talking about when we say AI server? You may well already be familiar with much of this, but we show here user numbers for well-known generative AI services like ChatGPT and Gemini. I use them as well, but as you can see, growth is strong. This level of growth is understandable. That said, up to now, AI has primarily been in the training phase, which allows it to answer queries like what is this? However, as questions become more complex or detailed, it requires the model to use inference. There is a shift underway. Next page, please. This shows the number of AI servers. This shows the split between training use and inference use servers. We are currently seeing significantly stronger growth in servers used for inference. Next page, please. There is a lot on this slide, but simply stated, when you show a model in the training phase, a picture of a human and ask, is this a human, you get an immediate answer of yes. To answer this question, the model has learned from a huge volume of images of dogs, monkeys, orangutans and humans. So when presented with an image of a human, it is able to respond immediately. To achieve this, what the server needs is high-speed DRAM and GPU or in other words, temporary storage. However, if you ask how old is the person in the image, the model cannot respond immediately, but would need to refer to a significant accumulated database. If each also needs to save individual histories, there is a need for high-capacity NAND memory and within NAND, ESSD, which is relatively fast. We are seeing rapidly increasing demand for these now as well as for ASICs. Next page, please. So how much DRAM is used by such servers? Training use servers use significantly more DRAM. If we look at the chart on the left for memory capacity, training requires 3 to 4x more DRAM per server. In total, AI use DRAM is currently around 500,000 to 600,000 wafers per month, but is expected to rise to 1.5 million over the next 3 to 4 years for an increase of 1 million. Next page, please. This shows NAND capacity for AI servers. Unlike the previous chart, inference use requires a significant volume of NAND per server as shown on the left. The chart on the right shows inference, which consumes a significant volume of NAND use wafers. That said, we are still only talking about volumes of the order of 200,000 wafers per month rather than 1 million plus for DRAM as covered on the previous page. However, given solid declines in NAND capacity, this increase in volume would be sufficient to drive a shortage. This is our image. Next page, please. So how much of a shortage in memory use wafers will result from AI servers? I believe that we are talking about demand growth of the order of 1.5 million wafers per month. So there is a need to increase capacity, particularly NAND capacity. NAND has generally been considered a nice to have and is usually the first thing that gets cut when money gets tight for devices like smartphone handsets. Photos can be stored in the cloud. You could argue that you don't need a huge memory of, say, 1 terabyte. If that's the case, NAND would be the first to go. Because of this, they are somewhat reluctant to invest in more capacity. While players recognize there are shortages, there isn't a huge shift towards investing yet. This is different from DRAM where players are keen to invest. This is why many suggest that there will be shortages in NAND memory for PCs, smartphone handsets and in particular, automotive applications. For us, we haven't seen a sudden surge in NAND memory use wafers and our customers' capacity is not increasing, so it will probably take some time for wafer demand to pick up. Perhaps it will take until next year. I think that the market may be okay with some shortages with the more difficult NAND for now. I will say that this doesn't particularly have a major impact on us this year. Next year, on the back of capacity expansion by customers, we expect a favorable environment with wafer volumes increasing for logic, NAND and DRAM. We hope that wafer increases for logic will kick in from the second half of this year when the new plant comes online. Next page, please. This is the situation for customer inventories, which is of keen interest to all of you. Inventories are not coming down much. I would like to show you why on the next page. Logic inventory is significant. We don't show the number of inventory months, but the white bars are purchase volumes and the blue bars are wafer inputs. We continue to see purchase volumes outweigh wafer inputs. The reason for this is that while leading-edge wafers are selling like hot cakes, and we have seen TSMC shift to 3-nanometer at Kumamoto, their original plan was for legacy products at 12-,16- or 28-nanometer. However, even the parent fabs are not seeing favorable levels of capacity utilization, so it doesn't make sense for them to be producing at these design rules overseas. 3-nanometer, on the other hand, is seeing extreme shortages. My point is that there is a significant accumulated inventory in non-leading edge. So there is a pressing need to significantly normalize inventories, and we will need to cooperate with inventory adjustments in many locations this year. I don't expect large inventory adjustments in memory, but dealing with the significant inventory in logic must be addressed. It is unavoidable but a one-off. This completes my section of the presentation. I will hand over to CFO, Kubozoe, to talk about details of our Q4 earnings.
Shinichi Kubozoe: The results for fourth quarter fiscal 2025 are shown in the third column from the right, as highlighted earlier by Chairman Hashimoto. Sales were JPY 105.2 billion, operating profit was minus JPY 4.5 billion, ordinary profit was minus JPY 5.9 billion and profit attributable to owners of the parent was minus JPY 10.8 billion. To the right, for the full year, sales were JPY 409.6 billion and operating profit was JPY 1.3 billion. We were able to be in the black on a full year basis. Ordinary profit was minus JPY 3.8 billion and losses attributable to owners of the parent were JPY 11.7 billion. Total CapEx for the year was JPY 79.9 billion. We show 2024 CapEx on the far left at JPY 214.9 billion. Compared to this time last year, CapEx is down a substantial JPY 135 billion year-on-year. If you look at the quarterly progression, while there was still some CapEx on an acceptance basis in Q1, subsequently, there was a sequential decline over the course of the year. In contrast, if you look at depreciation expense, it was up JPY 36.7 billion year-on-year to JPY 115.6 billion for the full year. In terms of quarterly progression, Q1 was the bottom with depreciation rising sequentially in each quarter to hit JPY 35.6 billion in Q4. EBITDA was JPY 112.4 billion, largely unchanged from the 2024 level. We show the key metrics based on the above results in the lower half of the table. This is the analysis of change to operating profit. Starting on the left, in the analysis of sequential change to quarterly operating profit, Q4 sales rose JPY 6.1 billion Q-on-Q to JPY 105.2 billion from JPY 99.1 billion. We also beat our forecast by JPY 5 billion. The overshoot in sales was due to the arrival of more goods than expected at the end of Q4 as well as ForEx impact. Our standard of revenue recognition is arrival of goods. Operating losses widened from JPY 1.6 billion to JPY 4.9 billion, a Q-on-Q deterioration of JPY 2.9 billion. As you can see from the chart below, depreciation increased a hefty JPY 4.5 billion. In addition, production was down Q-on-Q as a result of periodic maintenance at a mainstay plant in Q4 as well as closures for the end of year holiday season, which depressed production levels versus Q3. Also, as mentioned earlier by Chairman Hashimoto, there was a negative impact from changes in mix, partially offsetting the impact of higher sales. However, this was offset by steady production activity and utilization for an improvement on costs and by a positive ForEx impact. The net Q-on-Q decline in profits was JPY 2.9 billion. On the right, we show the year-on-year change for the full year. Sales rose JPY 13 billion, while OP fell JPY 35.6 billion. The increase in depreciation accounted for the vast majority of the year-on-year profit drop with cost, ForEx and production reporting only small changes year-on-year. On sales variance, as mentioned earlier, last year, there was an increase in PW relative to epi. The resulting product mix impact as well as a small impact from price responses on spot product is why there wasn't a significant contribution in either direction from sales variance. Next page, please. On this slide, I will cover the balance sheet and cash flow. Looking at the middle of the balance sheet, total assets as of the end of December were JPY 1,127.9 billion, down JPY 44.7 billion compared to the end of December 2024. The major changes were a JPY 20.4 billion decline in cash and deposits and a JPY 29 billion drop in tangible and intangible assets. I will discuss the change in cash and deposits in covering cash flow on the right in a moment. Tangible and intangible assets fell as of the end of December with depreciation outweighing CapEx. Raw materials and supplies rose particularly on a slight increase in polysilicon inventory. However, in terms of impact on total assets, what is more significant is the decreases in cash and deposits and tangible assets. Liabilities declined JPY 35.2 billion to JPY 480.2 billion, but interest-bearing debt was largely unchanged from a year ago. We have kept our balance of borrowings unchanged. Under other liabilities, there is a JPY 35 billion negative. This is related to the fact that actual payments for CapEx were quite significant relative to CapEx acceptance in 2024, reducing unpaid liabilities. Under net assets, I highlight retained earnings. As a result of the net loss at the end of the fiscal year and dividend payments, there was a decline of JPY 17.4 billion in retained earnings. Based on this, the equity-to-asset ratio was 51.3% and the D/E ratio on a gross basis was 0.61x as of the end of December. Both are largely unchanged from the levels as of the end of December 2024. On the right, we show cash flow. Operating cash flow was a positive JPY 100 billion for the year. Cash flow for investment activities was an outflow of JPY 111.4 billion, the combination of CapEx acceptance for 2025 and net changes in unpaid liabilities related to facilities. As a result, free cash flow was a negative JPY 11.4 billion for the full year. At the time of Q3 results, I indicated that free cash flow for third quarter became positive, but Q4 free cash flow was also positive for positive free cash flow for second half in total. After factoring in dividends paid, cash and deposits declined JPY 20.4 billion. With regard to meeting our cash obligations in 2025, we tapped into cash and deposits. Jumping forward to Page 23, I will now discuss our earnings forecast. The projections for Q1 are as shown on the third column from the right. We project sales of JPY 100 billion and an operating loss of JPY 6 billion. We project an ordinary loss of JPY 10 billion and net loss attributable to owners of the parent of JPY 10 billion, given that corporate taxes and profit attributable to noncontrolling interest offset each other. Reflecting the start of a new fiscal year, the roll-off of existing depreciation outweighed new depreciation, resulting in a decline of JPY 4.3 billion Q-on-Q. We expect a Q-on-Q decline in depreciation. Our ForEx assumption is JPY 155 to the dollar. Next slide, please. On this slide, we show the analysis of changes in operating income. On the left, we show the sequential changes. Q1 sales are projected to fall to JPY 100 billion, reflecting the timing differences in arrival of goods, which pushed up Q4 sales. Operating losses are projected to widen by JPY 1.5 billion. If you look at the waterfall chart below, you can see that while there are positives from depreciation and ForEx, sales variance is expected to have a negative impact. We are expecting to undertake periodic maintenance again in Q1 at another mainstay plant. Also, we have reflected a lower number of operating days, taking into account the fewer number of days in February. Both will be negative for sales. On costs, we expect an increase of JPY 2.1 billion, some due to pushouts from Q4 and to seasonal factors reflecting payments in Q1. As a result, we expect costs to rise Q-on-Q. On the right, we show the year-on-year change for Q1. For OP, we are guiding for an JPY 11.9 billion deterioration. The major factors are an increase in depreciation and a negative impact of product mix on sales variance with PW volumes rising relative to epi. Next slide, please. We have provided reference material at the end of the presentation with historical trends for earnings and EBITDA. This completes my section of the presentation.
Takayuki Komori: Thank you. We will now open the floor to questions. Mr. Enomoto, please go ahead.
Takashi Enomoto: I am Enomoto of BofA Securities. I believe you are implementing major changes to the senior management team. I think you will be handing over to President Ryuta. Please comment on the background to this major management change, how you selected the new President and what your expectations are for the new team given the significant change? I would also like to take this opportunity to thank you for all of your efforts since I believe this is the last time you will be presenting the results. When I think back to when I was covering Sumco in the past, I think Sumco has changed dramatically since you joined.
Mayuki Hashimoto: With regard to the change, I have been in this position for 14 years now. I had been thinking that it might be time to hand over, but we had the pandemic and then we made large-scale investments, which pushed cash flow significantly into negative territory, effectively putting myself in a position where I couldn't step down. Now cash flow is back in positive territory, so there isn't a need to worry about cash. Typically, Sumco generates operating cash flow of around JPY 100 billion and in a steady environment, invests around JPY 50 billion, including modernization investments. Under current conditions, we should be able to generate profits of around JPY 50 billion. Because we have made major investments, there is depreciation, but this is a noncash expense, and we have already paid for our investments, so there won't be cash outflows for the investments. With depreciation declining and the market recovering, I felt this was a good time to step down. My successor is someone who has spent a long time in the U.S., like me. I was in the U.S. for more than 10 years. He is someone who has significant international business experience. 80% of our business is overseas, so the ability to engage directly with customers overseas is very important. At my level, directly engaging gives you access to very high-quality market intelligence. Typically, my counterparts have a very broad outlook and good visibility. I have learned a lot from my interactions with them. Therefore, I felt it was important that my successor was someone who could engage with our customers. Also, I think he is very skilled at managing people, which is another point in his favor. With regard to the team, as I have been in this role for 14 years, the entire team consists of executives that I handpicked and appointed, and they have worked alongside me for the last 14 years. They are all very talented and highly capable individuals. When I joined, retained earnings were minus JPY 80 billion and the effective equity ratio was in the teens. Many people said that the company was going to go under, but together with this team, we were able to rebuild the company. I have a deep trust in this team, and I'm extremely grateful to have been able to work with them. I am confident that they will take the company forward. I see my role as laying the foundation so that the company can function without me. I would take no pleasure if my departure were to lead to serious challenges. My role and the true role of senior management should be to put into place a framework that will allow the company to develop regardless of whether I am there or not. This is why I created the Sumco vision with the aspiration to become #1 in technology. Over these 14 years, and it may well be that anyone can do this, but I never turned down a request for funding for R&D. In fact, if R&D did not fully use up their budgets, I would suggest that they were being complacent. This is an industry where you must have technological capabilities. When I joined, our technological expertise wasn't necessarily as high as it is now. We took on many challenges together. It was a very meaningful and fruitful time for me. Also, we have seen a solid increase in the number of women in our workplaces. In addition, people that I hired after we restarted hiring are now getting promoted to section chief level. We have superior technology in 300-millimeter, and our customers rate our leading-edge product very highly. From the second year after I joined Sumco, we have been consistently recognized by TSMC, winning awards now for 12 consecutive years. So I do think we are well recognized by our customers. Also seeing our employees be motivated is inspiring. Progress in R&D doesn't happen because I pushed. It happens because the employees are motivated. I did a lot to cultivate this culture, creating programs to recognize excellence. As a result, our turnover is very low. I think morale at Sumco is very high. If I have to sum up how I feel in one word, it would be gratitude to the employees. Mr. Enomoto, thank you for everything. I hold high hopes for Sumco's future.
Atsushi Ikeda: I am Ikeda of Goldman Sachs. I would like to ask about 300-millimeter inventory levels and longer-term customer concerns about supply. I suspect there is a significant polarization between legacy products and leading edge. The situation with inventory is really the result of elevated inventory for legacy products over a prolonged period of time. I am concerned that there could be shortages in leading edge. What is your strategy as you think about 2027 and 2028? And with regard to LTAs, how are you thinking about the next round of contracts with customers? If you can comment from a longer-term perspective about when you think supply/demand becomes balanced or we start to see shortages, that would be helpful.
Mayuki Hashimoto: We have invested heavily in increasing capacity, but almost all of our investments were for leading edge logic at 7-nanometer, 5-nanometer and below. Going forward, I think legacy capacity utilization will drop and there will be a transition to leading edge. However, when we entered into contracts with customers, there was no specification of volumes for legacy or leading edge because no one knew and the contracts were simply for total volume. There are only 2 companies doing leading edge. There isn't a third player. Because of this, we have contracts with everyone and our customers have no choice but to buy legacy wafers because that is what the contracts stipulate and there is a shortage of leading-edge wafers. With legacy use wafer inventory rising, customers want to reduce inventory in one fell swoop. This is why our legacy use wafers have been significantly impacted. So we want to get this over with in a short time frame. Once that's done, the picture is much more favorable in my view. Our leading-edge wafers are very highly rated by the 2 new customers that are fabricating leading-edge chips. Given this, I expect our new plant's capacity will be filled up first with utilization rates at our older plants dropping off. We have a plan to modernize our older plants swiftly, which is already in motion. We are steadily replacing facilities. That's my thinking. Does this answer your question?
Atsushi Ikeda: When do you think the current LTAs roll over?
Mayuki Hashimoto: It's a long way off.
Atsushi Ikeda: Understood. Do you have any major concerns? Are prices okay?
Mayuki Hashimoto: I think prices will be fine.
Takayuki Komori: Next is Mr. Yoshida.
Yu Yoshida: I am Yoshida of CLSA Securities. With regard to 2026, you provided information about trends in 2025 earlier, but is it possible to provide volume forecast for 300-millimeter and 200-millimeter for 2026? Also, what will be the percentage decline in wafer demand as a result of the adjustment to mature node inventory? When do you think that the adjustment will be complete?
Mayuki Hashimoto: First, I think the correction will last until the end of this year. It is a very dramatic correction. In terms of what will happen, as I said earlier, I think we will see very strong growth in leading edge, so I expect to see some gradual increases from the second half of this year. With regard to 300-millimeter wafers, while there will be cuts to purchase volume, we are also seeing an increase in PW. So I think overall 300-millimeter wafers this year should be in line with last year, even if the customers make cuts. So volumes may be flat year-on-year to growing slightly in the second half of the year. This suggests that there may be a slight recovery this year. Last year was 9% market growth for 300-millimeter. Sumco was fortunately able to do better than this. 2026 might not get that high, but there are some offsets, so I don't think volumes will fall significantly year-on-year.
Yu Yoshida: I'm sorry, can I confirm? When you say in line with last year, are you talking about matching the growth rate achieved last year?
Mayuki Hashimoto: No, I don't know that.
Yu Yoshida: So you are saying that absolute volume could be flat to higher year-on-year?
Mayuki Hashimoto: I do think that absolute volume this year should be able to match last year. I can't say how many percent it might grow, but my sense is that there should be some growth. If you were asking whether growth will be higher than last year, I can't really say. I think we might see some growth given the strength in PW.
Yu Yoshida: That's very helpful.
Takayuki Komori: Next is Mr. Miyamoto.
Go Miyamoto: I am Miyamoto of SMBC Nikko Securities. This is a high-level question. Currently, Sumco and Siltronic are in the red. At the same time, if we look at memory makers' OPMs, in the most recent quarter, South Korean makers reported 58% and the U.S. maker was at 45%. I feel that wafer prices are unfairly low. Is it possible to use the gap in margins between wafer makers and memory makers to negotiate for higher prices in future? Is it not possible to link the earnings of wafer makers to the margins of memory players given you support them? I understand, of course, that you have LTAs, but can you comment on your thoughts?
Mayuki Hashimoto: The higher memory chip prices are the consequence of supply/demand. So in our world, the business world, it's all about supply-demand balance. It's not possible to ask for higher prices just because the customer is making good profits. In fact, there have been instances when the customers are loss-making, but we asked for price hikes when there were shortfalls due to supply/demand. However, although we can't generalize at a minimum, I think that higher customer margins do mean less downward price pressure on suppliers. I think that is true. I also think that you might be able to raise prices to a certain extent. But while we expanded epi capacity, many wafer makers expanded PW capacity. So the capacity increases in PW may mean that even if customer margins have improved, it might be challenging to win a price hike. Unless there is a supply shortfall, it isn't easy to raise prices. I do think you make a good point. But unfortunately, economic rationality is not that forgiving. I have been here for 10-plus years and 40 years in the industry. This is not that kind of industry. I have never seen price hikes go through because a customer is making solid profits.
Go Miyamoto: Over the 10-plus years at Sumco, I think there must have been cycles where once profits improve, you see downstream investments, which drive up wafer demand. And as wafer supply tightens, it becomes possible to raise prices. Can we look forward to this going forward?
Mayuki Hashimoto: I do think that there have been such cycles. While I think this could apply to leading edge, when you look at commodity products, Chinese wafer makers now produce test wafers. Also, if you don't mind poor yields, Chinese players could use prime wafers. For instance, yield is not an issue for Chinese chip makers. They use wafers regardless of yields because they are instructed to do so by the government. That makes for a challenging environment.
Takayuki Komori: Thank you. We will end the meeting here. Thank you to everyone for joining the Q4 fiscal 2025 results briefing. We are grateful for your participation today.