Sumco Corp

TSE:3436

Earnings Call Transcript 2022-Q2

[Hiroshi Shibaya, executive]

I am Hiroshi Shibaya of Sumco Corporation. Thank you for your participation today. This is the results briefing for Q2 fiscal 2022.

Before starting the presentation, allow me to confirm today's materials, which consist of 3 items: the brief statement on consolidated financial results for Q2 fiscal 2022, the announcement regarding the interim dividend and the presentation deck entitled Results for the Second Quarter of Fiscal 2022.

We will have presentations today from Representative Director, Chairman and CEO, Mayuki Hashimoto; and Vice President, CFO, Michiharu Takii. Hiroshi Ito, General Manager of Accounting is also on hand. We have set aside an hour for the briefing, it 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.

Chairman and CEO, Hashimoto, will discuss our forecast 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, executive]

I am Chairman Hashimoto. I will start with the summary on Slide 5 of the presentation. As you can see on this page, the foreign exchange rate was largely in line with our expectations. We significantly exceeded our guidance for net sales, operating income and ordinary income. The primary reason for the overshoot boils down to the strong production levels. Effectively, we were able to sell more volume than expected, which contributed to higher profits.

Turning to our Q3 forecast. We are guiding for sales of JPY 115 billion. There are no plans for periodic maintenance in Q3, so production volumes should be relatively higher, hence, our forecast for sales growth. Typically, this should lead to solid growth in operating profit, but as will be covered by CFO, Takii, later, we have made progress on CapEx, which is pushing up depreciation. That said, the yen has depreciated significantly. As a result, we are expecting sequential gains in both sales and profits.

Next slide, please. With regard to the dividend per share, reflecting the overshoot of guidance, we plan to pay an interim dividend of JPY 36, which is higher than the JPY 34 previously announced. This represents a payout ratio of 40.2%, in line with our stated target of 40%.

Next slide, please. This is the market environment for silicon wafers. In Q2, demand continued to outstrip supply impacting both prices and volumes. Reflecting this, obviously, LTA prices were maintained and spot prices rose. Prices were higher on transactions with customers that buy on a quarterly or semiannual basis, although Sumco currently doesn't do many of this type of transaction. On our forecast for Q3 in 300-millimeter wafers for both logic and memory, we expect demand will continue to exceed supply capacity. 200-millimeter wafer demand should remain strong, backed by robust demand from automotive and industrial applications.

However, for 150-millimeter wafers, we are expecting to see some easing of tight supply/demand conditions. LTA pricing should track in line with the contract prices. Sumco has virtually no inventory available for sale in the spot market. Even if Sumco does have a few thousand wafers available for sale from time to time, spot prices are up strongly.

In terms of the outlook going forward, although the end markets for smartphones and PCs are softer on a correction, we expect data centers and automotive applications will continue to be the key drivers of demand. This should largely offset the impact of the weaker smartphone and PC markets. In 300-millimeter leading-edge epitaxial wafers, which is Sumco's focus, data centers account for a high proportion of demand. As well on the issue of the smartphone market correction, the migration to 5G is estimated to drive a 1.5-fold increase in wafer consumption, which should mitigate the impact of a weaker handset market even if the smartphone market declines around 10%.

As such, we don't expect a large impact from the smartphone correction. I believe that we will continue to see a situation where supply volume will not catch up to demand, particularly in leading-edge wafers. For 200-millimeter, automotive and industrial applications account for a large part of the market. So I would expect current strong demand conditions to persist. However, for 150-millimeter wafers, we are seeing a sudden return to more normal balanced conditions, so we are no longer seeing egregious imbalances in supply and demand.

Next slide, please. Slide 8 shows the wafer trend for 200-millimeter by quarter. As you can see, the volumes remain consistent at 6 million wafers per month, a genuine reflection of current market capacity levels. Sumco has devoted significant effort to improving productivity, but we believe we are getting close to our limit. Our 200-millimeter facilities are relatively older as well.

I believe that 200-millimeter is likely to remain at around the 6 million wafer per month level. As a result, there will be shortages in the market with some customers unable to fully cover their needs. It's possible that the corrections in smartphone and PC end markets could lead to a slight easing of conditions. That said, for the foreseeable future, I don't believe that a situation where there is an excess of 200-millimeter wafers is at all likely.

Next slide, please. On Page 9, we show the same chart for 300-millimeter wafers. It is rare that we see a flat quarterly progression. However, the flat trend is not the result of a weak market. The market is very strong. So all players are operating at full capacity, but do not have any more room to increase production. The market is at maximum capacity. These conditions have persisted since Q3 of 2021 and are likely to continue through 2023.

None of the wafer makers, including Sumco, have any more room to expand brownfield production capacity. The only alternative is to ramp up greenfield capacity, but we are unlikely to see solid volumes from greenfield capacity until 2024 and beyond. So I believe conditions are likely to be toughest in 2023.

Next slide, please. This chart shows our estimates of customer inventory for 300-millimeter wafers. Inventory levels remain below 1 month with no prospects for a sudden and rapid recovery.

Next slide, please. If we break out customer 300-millimeter wafer inventory between logic and memory, the issue of extremely low inventories is much more acute on the logic side at levels well below 1 month. In contrast, memory inventories are at just below the 1 month level. So even if we were to see a correction, the inventory shortages are very severe for logic. Generally speaking, the logic market has not been prone to corrections. For memory, given that inventory levels are below 1 month, I suspect that memory players would prefer to have a higher level of inventory.

Slide 13, please. With regard to the memory market, in which everyone has a keen interest, it is true that we are seeing a correction. This is a market that is subject to relatively frequent corrections. Since 2008, we have seen 4 market corrections if we include small adjustments. 2 of these 4 corrections spilled over into the wafer market in 2008 and 2017. In both of these years, we did see a decline in wafer shipments.

Next slide, please. In terms of our view of the current correction, if we look at the market forecast published by market data firms in June, when talk of a correction began to emerge, compared to March, as shown in the bar charts, they generally show a downward revision of around 3%. These charts are for the DRAM market. In volume terms, this represents a decline of around 60,000 to 70,000 wafers per month for DRAM. The total DRAM market is typically around 1.5 million to 1.6 million wafers per month currently.

Next slide, please. These bar charts show the same data for NAND flash memory. Here again, there was a downward revision of around 3% if we compare the June forecast to the March forecast of the market data company. The NAND market is generally around 1.5 million wafers per month as well, so a 3% revision represents a decline of about 50,000 or 60,000 to 70,000 per month. Combining the DRAM and NAND forecast suggest a potential correction of between 100,000 to 150,000 wafers per month, I think we are already seeing a correction.

Assuming this to be the case, if we look at the peak shortfall for just Sumco out to 2026, based on what our customers have said versus the LTA volumes, even fully stretched from production, there is a gap of 800,000 wafers per month. The reason for the shortfall is expected delays for delivery of equipment for the greenfield capacity. Although I suspect that the orders may include some double orders, the impact of the memory correction of 150,000 wafers per month would not be enough to fully cover even the shortages just at Sumco. There would still be shortages.

Please turn to Slide 16 for a discussion of the situation for epitaxial wafers. This is Slide 16, which shows the epitaxial wafer market. As you can see from this chart, the epitaxial wafer market has not seen a significant correction in the past. The market has continued to show consistent growth over time with a CAGR of 12.1%. This reflects the diversity of chips produced by foundries. Customers that only produce for PCs are vulnerable to downturns in the NPC market. Foundries produce a wide range of chips. So even if PCs are weak, it can be offset by data center or other demand. Foundries have access to many different market segments. Because of this, the market is relatively less prone to downturns, which is why the epi market has shown generally consistent growth over time.

Next slide, please. This shows our estimates for leading-edge epi wafer demand. We look at demand for design rules of 7-nanometer or lower, which is an area of strength for Sumco. Currently, it is at slightly less than 500,000 wafers per month, but we project demand to rise to around 1.2 million wafers per month by 2026. We have found that our forecast for epi rarely diverged from actual trends. This is because leading edge has shown solid growth. That has been the case to this point.

Our capacity increase plans this time are almost all focused on business with customers that concentrate on leading edge. Therefore, although the scale of investments were large, we are expanding leading-edge epi capacity where we think the likelihood of a correction is extremely low.

Please turn to Slide 19. This is my final slide. So if we look at the green bars, which is the greenfield capacity, some of this may be challenging. Even if we do see a correction as touched upon earlier, I believe we will continue to see wafer shortages up to 2026. This is merely a forecast, but if we look at the various conditions, it seems inevitable.

Some may say that our forecasts were proved wrong in the past, but the magnitude of the shortages we are seeing this time around are significantly larger than the shortfalls experienced in 2017 and 2018. This is because last time, wafer players, including Sumco, were able to expand brownfield production capacity and respond in relatively short order. This time around, the only way to increase production capacity is with greenfield expansion. This has coincided with the emergence of geopolitical developments such as the invasion of Ukraine by Russia or the intensifying trade friction between the U.S. and China, which has led to longer lead times for many items or longer delivery times.

For example, even if we want to ramp up new equipment, it may not be possible for the supplier to send staff to provide on-site support. There are many factors that are a drag on progress. Unlike in the past, it isn't possible to advance the process smoothly. I believe our peers are in exactly the same situation. As a result, it isn't possible to replicate previous lead times. So when we think about adding capacity, we are talking about greenfield. This is being compounded by longer lead times.

This is creating a tough situation in terms of increasing capacity in line with demand. So this is why I believe that it will unfortunately continue to be difficult to fully meet the needs of customers out to 2026, particularly in leading edge. This is my view.

This completes my section of the presentation. I will hand over to CFO, Takii, to talk about details of our earnings.

[Michiharu Takii, executive]

I, Takii, will present the earnings in more detail.

Please turn to Slide 21. I will start with the first half sales and profits. First half net sales rose roughly JPY 50 billion year-on-year. 6-month operating income rose JPY 28 billion year-on-year to approximately JPY 50 billion. Ordinary profit was also around a similar level. As discussed when we announced Q1 results, corporate taxes increased on the back of the depletion of cumulative loss carryforwards and the resulting absence of deferred tax assets. As a result, the effective tax rate has reverted to a more normal level.

Net income attributable to owners of the parent was JPY 31.3 billion, roughly double year-on-year. Below that, CapEx was JPY 45 billion and depreciation was JPY 27.5 billion. Both rose reflecting year-on-year increases in investments in facilities for leading-edge production as well as the start of some greenfield-related investments. The exchange rate for first half was JPY 121 to the dollar. OPM and other management metrics shown here, all improved year-on-year.

Please turn to Slide 22. This is the analysis of changes in operating income. Sequentially, 2Q OP improved JPY 2.9 billion. As shown in the waterfall chart on the lower left, costs were impacted by higher or elevated electric power unit prices, labor costs and shipping costs, which led to a negative sequential impact. Depreciation was also up Q-on-Q.

Sales-related variance was a slight positive of JPY 0.3 billion Q-on-Q. I touched upon this previously, but in April, we conducted periodic maintenance at our main 300-millimeter plant, which limited sales growth due to limited volume growth. The biggest Q-on-Q factor was foreign exchange with a positive impact of JPY 4.5 billion.

On the right, we show the year-on-year change for first half results. First half OP improved JPY 28 billion on a JPY 50 billion increase in sales, as noted earlier. There was a significant year-on-year increase in costs, reflecting higher unit prices for materials and electric power unit prices for an impact of JPY 2 billion year-on-year. Logistics expenses rose by slightly more than JPY 0.5 billion. Labor costs also increased by JPY 4 billion. Roughly half was from increased bonus payments as discussed last time, reflecting higher unit labor costs.

In addition, headcount also increased on higher production levels and some hiring in anticipation of the greenfield capacity. Depreciation increased by JPY 4.5 billion year-on-year. Sales-related variance had a roughly JPY 30 billion year-on-year positive impact. Across all diameters, prices rose around 10% year-on-year for a combined impact of JPY 16 billion as noted previously. The remainder of JPY 14 billion reflects the increase in production and sales volumes. The year-on-year depreciation of the yen to JPY 121 to the dollar, as shown in the table above, had a positive impact of JPY 9.8 billion.

Next slide, please, for a discussion of the balance sheet and cash flows. Slide 23 shows the balance sheet and cash flow statement. Cash and time deposits on the balance sheet increased JPY 54 billion, partly as a result of improved earnings. Raw materials and supplies increased JPY 0.3 billion. The reduction in polysilicon inventory within this was JPY 3 billion, which was offset by an increase in materials inventories, particularly stock of maintenance-related materials, which we consider important from a BCP standpoint.

Tangible and intangible assets rose specifically on an increase in tangible assets, which was up JPY 23.3 billion on the back of higher CapEx. if you see the brief statement on consolidated results for second quarter fiscal 2022, this includes an increase of JPY 18.8 billion in the construction in progress line, reflecting investments for leading-edge as well as some advanced greenfield investments.

Other liabilities increased JPY 42.9 billion. The outstanding balance of interest-bearing debt was flat, but our joint venture subsidiary in Taiwan, Formosa Sumco Technology, or FST, took in advanced payments and deposits on LTAs of roughly JPY 29 billion. This also contributed to the increase in cash. If we look at the capital account, retained earnings increased, reflecting the impact of higher profits. The shareholders' equity ratio was down slightly. Despite the increase in net assets, there was a larger increase in total assets. However, all other metrics improved.

On the cash flow statement, as you can see, the advanced payments received by FST are included under the other changes in working capital of JPY 17.8 billion. Under investment cash flow, you can see the previously mentioned CapEx of JPY 45 billion on an acceptance basis. Free cash flow was JPY 52.4 billion. Cash and time deposits increased by JPY 54 billion.

Moving on to Page 25 and the forecast for Q3 2022. These are our forecasts for Q3. We project sales of JPY 115 billion for Q3 and OP of JPY 27.5 billion. We expect net income attributable to owners of the parent to be JPY 16.5 billion, largely in line with the Q2 level. Our ForEx assumption is JPY 135 to the dollar.

On the next slide, we provide the analysis of sequential change for Q3 and year-on-year change for the 9-month operating income. I will start with the sequential change from Q2 into Q3. We are seeing the impact of inflation and have projected a Q-on-Q increase in cost of JPY 3.7 billion. Of this amount, electric power costs account for JPY 1 billion of the increase, a JPY 1 billion plus rise in materials and supplies and a JPY 0.5 billion rise in repair and maintenance expenses.

Labor costs are also expected to increase. Depreciation is also expected to rise. We project a JPY 3.1 billion positive in sales-related variance of which roughly half is from an increase in volumes, and the remainder will come from slightly higher spot prices for 300-millimeter and 200-millimeter. The positive exchange rate impact of JPY 3.2 billion reflects our assumption of the dollar to yen at JPY 135.

Looking at the analysis of year-on-year change for the 9-month OP. Effectively, it is an extrapolation from the first half results with trends to remain similar. Costs are expected to rise JPY 13.8 billion year-on-year. We expect electric power unit prices to rise for an impact of roughly JPY 3 billion. Repair and maintenance costs are expected to increase JPY 3 billion year-on-year. Logistics expenses are projected to rise JPY 1 billion, while materials and supplies costs are expected to rise JPY 2 billion year-on-year on higher unit prices.

Labor costs, including headcount increases, will rise JPY 5 billion year-on-year. On a year-on-year basis, depreciation is expected to increase on the back of investments. Most of this is investments to increase capacity for leading-edge production. For sales-related variance, we expect a JPY 40 billion plus positive. We have factored in an increase of JPY 25 billion from higher prices. As noted earlier, prices have been rising for all diameters. We expect this trend to continue. The remainder will come from higher production volumes on a year-on-year basis.

We expect a JPY 19.4 billion year-on-year positive from ForEx for the 9-month period, reflecting the JPY 125.6 dollar-yen rate, as shown in the table above.

Slide 28, please. As you can see, revenue is rising sharply, reflecting the impact of rising prices and volumes, as noted earlier. ForEx is also a factor. For EBITDA, you can see the blue bar show a sequential increase of JPY 2.7 billion for Q3. However, as a result of the stronger sales growth, the Q3 EBITDA margin is expected to be largely flat.

This completes my remarks. Mr. Shibaya?

[Hiroshi Shibaya, executive]

[Operator Instructions] We will start with Mr. Enomoto of BofA Securities. Mr. Enomoto.

[Takashi Enomoto, analyst]

I am Enomoto of BofA Securities. Listening to Chairman Hashimoto's presentation, my impression is that the correction in semiconductors is not having much of an impact on the wafer cycle this time. Is that the right understanding? If so, I believe the need to expand capacity in the future, say, from 2026 or 2027, is rising. With regard to future capacity expansion, is your stance unchanged?

Or for example, are you now thinking about raising the magnitude of capacity expansion for epitaxial wafers and perhaps limiting the capacity increase in polished wafers? Please comment on whether there are changes to your future investment stance, if any. I recognize that you haven't actually announced future plans, so talking about changes may be difficult. What I want to know is how you are currently thinking about future capacity expansion.

[Mayuki Hashimoto, executive]

To begin with, Sumco's capacity expansion is almost all focused on leading-edge epitaxial wafers, which account for 90% of the capacity expansion. We are making almost no increases to polished wafer capacity. Actually, probably polished capacity is likely to be allocated to epitaxial going forward. The challenge for us is figuring out how to deal with the shortfall. Our policy has been that we will only expand capacity in leading-edge epi wafers. This policy remains unchanged.

With regard to capacity in 2027, if delivery times remain as long as they are now, we will need to start placing orders now to be in time for the next round of greenfield expansion. In terms of brownfield expansion, we have built a large-scale shell, but populating this space is challenging given the significantly longer lead times. So we will probably need to start initiating conversations with our customers to be in time for the next round.

Customer response is relatively positive on capacity expansion for 2027 and beyond on concern of potential shortages. However, it appears that there is a polarization amongst the customers with a clear gap widening between the winners and losers. Despite the flow of announcements from chip makers, there are some that are seemingly slow in placing wafer orders. That said, the majority continued to show a very strong appetite for wafers. So our stance remains unchanged. Nothing has changed.

[Hiroshi Shibaya, executive]

Next is Mr. Ikeda of Goldman Sachs.

[Atsushi Ikeda, analyst]

I am Ikeda of Goldman Sachs. Congratulations on the solid results. Can you comment on how you were thinking about selling prices? Despite a slight stabilization in the spot market and end market demand, you are expecting Q3 spot prices to rise. I would like to know if the year-on-year increase in LTA prices in fiscal 2023 will be similar to fiscal 2022 levels.

Also, you indicated you expect the tightness in supply demand to be at its most acute in 2023. I believe Sumco won't have much available for sale in the spot market, but do you expect to see further price increases? In addition, you mentioned advanced payments received by FST. How should we think about how such advanced payments will increase going forward.

[Mayuki Hashimoto, executive]

You have asked many different questions. Shall I answer one maybe? I expect supply/demand to be at its tightest in 2023. In terms of price, I do think that some types of wafers will see further price increases. However, there may be some types of wafers where prices will not rise that much. That said, given that fundamentally, there aren't enough wafers to go around, even if there were a correction in the end markets, wafers would still be in short supply.

In this situation, it may be that customers may not be prepared to buy at any price to lock in wafer supply as is the case now. But nonetheless, I would expect a market environment where customers would still be willing to pay a premium. On the advanced payments to FST, capacity has been fully spoken for, so I don't expect to see a further increase.

[Atsushi Ikeda, analyst]

Understood, can you at least provide some hint as to whether we should expect the increase in LTA prices in 2023 to be similar to the scale of increases in 2022?

[Mayuki Hashimoto, executive]

Are you talking about spot prices or LTA prices?

[Atsushi Ikeda, analyst]

LTA prices.

[Mayuki Hashimoto, executive]

For LTAs, 2023 prices will also rise. Generally speaking, our contracts are structured so that the prices rise gradually over the course of the contract. So the magnitude of the increases would be similar to the increase in 2022.

[Hiroshi Shibaya, executive]

Next is Mr. Watabe.

[Takato Watabe, analyst]

I am Watabe of Morgan Stanley. Given that we are seeing what could be called recessionary conditions in the event that wafer demand does not pan out in the way you expect, hypothetically speaking, would you anticipate difficulties in enforcing your LTAs? How should we think about price and volume under such a scenario?

[Mayuki Hashimoto, executive]

In the past, there have been multiple instances where prices and volumes did not trend in the way we expected. The first point I would like to make is that our policy is to only engage in LTAs with customers where we have long-standing relationships of trust. In terms of what we have done in the past when dealing with customers where we have established long-standing relationships, we have not lowered our price point.

However, there have been instances where we have agreed to allow customers to adjust volumes by pushing out the timing of some deliveries. That is probably what we would do. That said, what we are seeing currently is that our customers are pressuring us to increase volumes. One other point is that the customers haven't necessarily been able to fully cover their needs with LTAs.

Hypothetically, if the customers needed to adjust wafer input, they have said that they would suspend all their spot purchases in favor of their LTA commitment.

[Takato Watabe, analyst]

I see. So basically, the message is that the risk going into the next fiscal year are not that large.

[Hiroshi Shibaya, executive]

Next is Mr. Miyamoto.

[Go Miyamoto, analyst]

I am Miyamoto of SMBC Nikko. I would like to ask about the situation for Chinese wafer makers. Given the overall tightness of the market recently, I believe some players have started to buy some silicon wafers from Chinese makers. My understanding is that the Chinese players are mainly focused on test and dummy wafers and that there is a significant technological gap between them and Sumco. Are there signs that they are catching up?

If you look at NSIG's 300-millimeter wafer sales trends in the last few years, sales have been rising slightly. They have also announced proactive CapEx plans. Please share your view of Chinese wafer makers.

[Mayuki Hashimoto, executive]

With regard to Chinese wafer makers, I think what we are doing is the most unattractive business for them. We have significant financial resources and have invested heavily to develop new capacity. If they were to try to do this to make money, there are many things that would be challenging.

First of all, it takes a very long time to get certification from customers for new capacity. For a new entrant, at a minimum, it probably would take at least 2 years. During that period, the manufacturing facilities would be idle. The interest rate and depreciation burden while the facilities are idled is substantial. Beyond that, even if they were able to get certified in 2 years, yields would likely be very low in the early stages of production.

So a new player would need to get over the hurdle of low yield. So even if they have significant financial resources, this phase of the process would be very challenging indeed. I say this because China has been trying to break into the production of semiconductor-grade silicon for 30 years without any success. This means that even if a Chinese player were able to procure semi-grade silicon to actually be able to successfully raise ingots require significant expertise and know-how.

It is technically very challenging. So while Chinese players may be able to raise and fabricate test and dummy waivers, the degree of complexity and difficulty associated with prime wafers is completely different. For instance, there are tough requirements around defect levels and specific ranges that must be achieved for items such as oxygen content while also hitting volume targets. Achieving this was very challenging for us.

If we think about what it takes to overcome all of these challenges and the fact that even after 30 years, China has had no success in manufacturing polysilicon despite the fact that the specs have not changed during this period, it suggests that this remains a significant hurdle for Chinese players, in my opinion.

[Hiroshi Shibaya, executive]

Next is Mr. Okazaki.

[Shigeki Okazaki, analyst]

I am Okazaki of Nomura Securities. I have a question about Slide 22. The cost and depreciation figures you show here are slightly lower than the figures you had shown 3 months ago. Can you explain why? My sense is that depreciation has been relatively controlled despite significant investments. How should we think about this?

On sales-related variance, is the tight capacity what makes it difficult to increase volumes further? I believe that is what you suggested earlier, but I want to confirm that I understood you correctly.

[Mayuki Hashimoto, executive]

I will answer your last question. Your understanding is correct. From last year onward, we have devoted significant efforts to improve productivity. This delivered significant results for us but to achieve an even higher level of productivity will require renewed efforts. Achieving this to get to the next level may take some time, say a year or so. So for now, the Q3 level is probably the upper limit. Note that there won't be periodic maintenance in Q3.

I will hand over to CFO, Takii,to respond to the other questions.

[Michiharu Takii, executive]

On your question about whether the increase in cost is smaller than we had previously projected, production is running very smoothly. We had expected a relatively large increase in costs, but output levels and yields have improved. I believe the smooth production was a contributing factor.

[Shigeki Okazaki, analyst]

What about depreciation?

[Michiharu Takii, executive]

With regard to the differences in our depreciation assumption, depreciation can be impacted by timing differences for the ramp-up of equipment. So it's likely to have been factors like this.

[Shigeki Okazaki, analyst]

can we expect you to be able to continue to control costs going forward?

[Michiharu Takii, executive]

Obviously, we want to keep costs as low as possible, we will do our best.

[Mayuki Hashimoto, executive]

We are doing our best to negotiate with our suppliers.

[Michiharu Takii, executive]

But unit prices for electric power have doubled.

[Shigeki Okazaki, analyst]

My impression is that you have been able to control costs in spite of the backdrop.

[Michiharu Takii, executive]

There is also the impact of ForEx rates. For imported materials we procure prices are likely to rise. So going forward, it may be difficult to mitigate the trend towards higher costs.

[Shigeki Okazaki, analyst]

Yes, today's [ weekly ] had a major article on this issue, but I have high expectations for Sumco.

[Michiharu Takii, executive]

We will do our best to improve productivity.

[Hiroshi Shibaya, executive]

Next is Mr. Nishiyama.

[Yuta Nishiyama, analyst]

I am Nishiyama of Citigroup Securities. Could you comment on progress in improving yields for leading-edge wafers? In the September quarter 2 years ago, you struggled to ramp up leading-edge capacity and incurred some opportunity losses. However, near term, I believe you are now seeing an increase in shipment volumes for the current leading-edge wafers.

Under the new LTAs, I believe your customers agreed to appropriate prices that recognize the improvement in added value for leading edge. If the ramp-up is proceeding smoothly, I would expect the increase in production of leading-edge wafers to lead to higher margins for the overall company. Could you comment on this?

[Mayuki Hashimoto, executive]

In leading edge, for instance, we are currently in mass production for 5-nanometer, and the process has been very smooth in terms of improving yields. While the degree of difficulty is slightly higher than conventional products, yield levels are comparable. What we are doing now is development work for the next generation. The reason why we are so focused on using AI is because of the experience of our customer. Several generations ago at 28-nanometer, the ramp-up of a new generation took 2 to 3 years for our customer.

But more recently, when you look at the customers' ramp-up of 7-nanometer or 5-nanometer, it has been very rapid. The customer adopted AI technology, which has raised their agility. They are able to swiftly apply data analytics to boost their ramp-up capability. They are able to identify issues through data analytics and can immediately address the problem. Their level of expertise has improved significantly. 2 years ago, we were not able to keep up with them, but they are sharing this expertise with us, which has allowed us to improve. So we are not experiencing any yield difficulties on leading edge.

Going forward, we aim to ensure that we don't experience such issues again.

[Yuta Nishiyama, analyst]

So a smooth ramp-up should have a positive impact on overall margins. Is that correct? Also, how have you factored this into your Q3 forecast?

[Mayuki Hashimoto, executive]

It is still early days, so we haven't reflected much of this in the Q3 forecast. However, you should see some impact in terms of product mix. As the proportion of leading edge increases, it should push up the average selling price, which should be beneficial for margins. We haven't done detailed analysis, but at a minimum, there should be a positive benefit from improved product mix.

[Hiroshi Shibaya, executive]

Next is Mr. Omura.

[Shunta Omura, analyst]

I am Omura of UBS. Global Wafers recently announced investments in the U.S. for a capacity of 1.2 million wafers. I believe it will take several years to achieve this level of capacity. Global Wafers have said they are targeting leading edge. I believe they are aiming to win share from players like TSMC, Intel and Samsung in the U.S. How should we view your share of the wafer market in the U.S. going forward?

[Mayuki Hashimoto, executive]

As you know, TSMC dominates in leading edge. Intel seems to be lagging slightly in leading edge, so it is really TSMC and Samsung that are strong in leading edge. In the U.S., we are already one of Intel's top suppliers for leading edge, so we intend to continue to supply Intel as we have up to this point. I would say the same goes for Samsung as well.

It goes without saying that our stance on TSMC is unchanged. That said, we do not have plans to build a wafer plant in the U.S.

[Shunta Omura, analyst]

I see, so in other words, there is a possibility that Sumco will increase supply capacity for the U.S. market in 2026 and beyond at the plants you have in Japan and Taiwan?

[Mayuki Hashimoto, executive]

Yes.

[Hiroshi Shibaya, executive]

Next is Mr. Watanabe.

[渡邉 亮一, analyst]

I am Watanabe of Mitsubishi Securities. I would like to ask for clarification on the relationship between the semiconductor correction and a correction in wafers. Earlier, you said that corrections in the memory market spilled over into the wafer market in 2008 and 2017. What determines whether the wafer market is impacted? Is it the depth of the trough?

You also noted that compared to 3 months ago, market research demand forecast were revised down by around 3%. Are you suggesting that at this level, a downturn can be absorbed by the strength of demand in the epitaxial wafer market, and therefore, it wouldn't have an impact on the overall market. I would like to better understand the relationship between corrections in the memory market and the implications for the wafer market.

[Mayuki Hashimoto, executive]

As our peers have also been saying, the magnitude of the shortage is different this time. In the past, even when there were shortages, customers were not willing to pay a premium without complaint. However, now there has been a lot of talk about customers paying up significantly in the spot market. Unfortunately, we don't have inventory to sell in the spot market, so we aren't benefiting. The magnitude of the shortfall is very large this time.

As I explained earlier, a 3% correction in the memory market only represents 150,000 wafers per month. Even if there is some double booking, Sumco alone is seeing a shortfall of 800,000 wafers per month up to 2026. So even if the memory downturn leads to a reduction in demand of 150,000 wafers per month, given that Sumco alone has a shortfall of 800,000 per month, even if there were some double bookings of, say, around 200,000 or 300,000, if we take into account Sumco's market share, it implies that there is an overall market shortage of around 1 million wafers per month.

Compared to this, my view would be a correction of only 150,000 doesn't change the fact that there are shortages. The reason why I say this is because previously, there was still room for brownfield expansion, so the industry was able to increase output relatively quickly. However, now, expanding production capacity can only be done with greenfield investments, so it will take time for the capacity to come online. This is being exacerbated by the surge in geopolitical risks as well as the ongoing pandemic impact.

This has led to a significant increase in lead times. So it is tough to catch up in terms of increasing supply. At the same time, customers are raising productivity and expanding capacity, although they are probably also behind schedule. All of these announcements of multiple trillion yen investments by customers begs the question of how they are going to find the wafers for all this new capacity, in my view.

So even if there is a correction in memory, given all this, it doesn't seem realistic to expect that it would have an impact on wafer supply demand given the sheer scale of the shortages. So it isn't simply that epi demand will offset a memory correction. Polished wafers are also likely to remain in short supply, in my view.

[渡邉 亮一, analyst]

I have just one follow-up. Is more of the 800,000 shortfall at Sumco in epitaxial or polished wafers?

[Mayuki Hashimoto, executive]

That is the combined total.

[渡邉 亮一, analyst]

Are both in shortfall?

[Mayuki Hashimoto, executive]

The shortfall in polished wafers is larger.

[Hiroshi Shibaya, executive]

Our final question is from Mr. Yamada.

[Mikiya Yamada, analyst]

I am Yamada of Mizuho Securities. I have one question. You have talked about geopolitical risk. If the U.S. were to block sales of SPE to China, then I believe that semiconductor production in China would grind to a halt, which would lead to a shortage in devices. What would be the impact on Sumco? I think that very few of the local Chinese chip makers are using Sumco wafers, although there may be a small amount.

I suspect there may be major chip makers that are using a certain volume of some co-wavers at their fabs in China. Would there be an impact on Sumco? Alternatively, is it too early to be thinking this way?

[Mayuki Hashimoto, executive]

There are some promising chip makers amongst local Chinese players. For example, you have players like SMIC in logic or YMTC in memory. SMIC is on the U.S. entity list, but they are still able to sell conventional non-leading-edge product. They are also currently able to buy non-leading-edge SPE. It would probably be challenging if they are banned from buying all SPE. Probably inspection equipment would be the most difficult since the U.S. is the only source.

They might be able to get around it somehow by using old equipment. However, they would be able to buy equipment like enters from Japan, so they should be able to manage, although they probably will need to produce more low-grade product.

If you look at YMTC, this is a company that has strong technology given the relationship with Tsinghua University. The stacking technology is world-leading technology that everyone else is trying to emulate. In terms of production equipment, they have some capability so they may be able to manage. If the U.S. adopts a very extreme ban on doing business in China, it would have an impact on Sumco. It may not be huge, but there would be an impact as we do have a certain level of sales into China.

The question would be where China would source wafers if the U.S. bans all sales and who would sell to China? I would question whether the U.S. would actually totally prohibit the sale of semiconductor-related products into China. If the approach is too aggressive, it could well lead to war. Unlike the past where semiconductors were just an industrial building block. Now semiconductors are probably more important than oil.

So a total ban would be the equivalent of completely halting all sales of oil into China. I suspect that the U.S. wouldn't go that far, although they probably don't want to support China in the area of leading-edge semiconductors.

[Mikiya Yamada, analyst]

Actually, 82 years ago, the U.S.. Did exactly this. They imposed a total ban on sales of oil into China. I completely agree with your view.

[Mayuki Hashimoto, executive]

I don't think the U.S. would go to that kind of extreme stance. We are in the 21st century after all. Having said that, I couldn't have imagined Russia invading Ukraine either. Who knows what might happen. As management, the only thing we can do is to ensure that we are resilient and agile in adapting to change. It is probably necessary to have multiple scenarios and to develop skills to be highly adaptable to a changing environment.

[Mikiya Yamada, analyst]

Understood.

[Mayuki Hashimoto, executive]

Thank you, although I don't know if I really answered your question.

[Hiroshi Shibaya, executive]

Thank you. Thank you to everyone. We are grateful for your participation today. We will end the meeting here. Thank you.

[Statements in English on this transcript were spoken by an interpreter present on the live call.]