Operator: Quarter earnings call. We thank you very much indeed. For today, simultaneous interpretation is made available. Please click the interpretation icon at the bottom of the screen and select the language of your choice. Mr. Shinkai, please turn your video on. Today's session is attended by our CFO, Mr. Shuhei Shinkai, as well as other staff members of the company.
Operator: After a few words from Mr. Shinkai, we would like to provide you with explanation of the Q3 results, which will be followed by a Q&A session. We expect to finish the entire session in about 60 minutes. The materials to be used for today's presentation is already posted on the IR site of our corporate homepage. Mr. Shinkai, please turn your microphone on and begin your presentation.
Shuhei Shinkai: Thank you. I am Shinkai, CFO of the company. Good morning to you. From this time around, the earnings call update method was reviewed in June of this year at the Capital Market Day. Our mid-to-long-term business direction and strategies, as well as the aspirational financial targets were presented to you in a comprehensive manner. The progress of those mid-to-long-term things will be explained during the full year earnings call, also at the Capital Market Days from Mr. Shibata. The recent or the quarterly performance of the company will be explained mainly by myself. Please be aware of that. Now, if I could begin the presentation using the presentation slides. But before that, in the Q3 of this year, I would like to just briefly talk about the highlights of the Q3 results, as well as the outlook for the Q4.
Shuhei Shinkai: The Q3 actuals, with the Industrial and Infrastructure sales came in higher than the projected numbers. On the actual apple-to-apple basis, without the impact of forex, it was a 1.18% increase compared to the plan. As for the Q3, on 1 July, the Timing Business was transferred. We completed that transfer, so the GAAP-based operating profit, we recognized the proceeds from the sales of JPY 450.3 billion. For the Q3, on a QoQ basis, excluding the foreign exchange impact on apple-to-apple basis, 4.2% QoQ increases was recorded. The details was driven by the growth of automotive, SoCs, and also the Gen 4 took off, started off, and also the MCU, 28nm adoption spread. So that was also a factor. And also Industrial Infrastructure continued to be strong.
Shuhei Shinkai: The Q4 outlook, if I try to summarize for the full year of 2026 on a year-on-year basis, 26.4% revenue increase is expected for the landing. There's also foreign exchange favorability, but on a year-on-year basis, a 20% increase is expected for the full year. Now, for the mid-to-long-term aspirational targets, towards these goals, we will continue to grow, and we are also tackling to increase the production capability. The in-house production will be expanded through CAPEX. In the Q3, Q4, we will continue to make those capital investments. Now I would like to go into the explanation using the presentation slides. Now, if you can go to page 3, this is a disclaimer here.
Shuhei Shinkai: As I said at the outset, on 1 July, the Timing Business was transferred, so those were excluded from the non-GAAP numbers. The GAAP-based Q3 numbers recognize the proceeds from the transfer of the Timing Business, the gains from the sale. On the fourth page, the Q3 actuals are presented here. If you look at the shaded part on the left, those are the Q3. If you can look at that, for the revenue, JPY 435.2 billion. Gross margin, 59.4%. Operating profit, JPY 150.5 billion. Operating margin, 34.6%. EBITDA, JPY 173.4 billion. Profit attributable to the owners or parent, JPY 125 billion. Foreign exchange, JPY 161 to the dollar, JPY 185 to the euro. As always, the Timing Business, excluding all that, the year-on-year comparison excluding Timing Business is presented on the right-hand side of the screen.
Shuhei Shinkai: For Q3, there was one-off accounting-related factor. Excluding that, we wanted to show this apple-to-apple as much as possible so that we can explain the real strength. Excluding the foreign exchange and the one-off factor, if you look at the second column there, that is already presented. It is the one that says that excluding the foreign exchange and also the foreign exchange impact and also the one-off factor. The ship and debit, that transaction-related adjustments are this one-off factor, the ship and debit. We have made a quotation for the provisions, and then based on the actual sales of the act to the customers, we have settled that. We have changed the methodology for the quotation, so that is the reason why we have decided to make this adjustment one-off here.
Shuhei Shinkai: The number was quite significant, so that is the reason why we have decided to exclude this one as a one-off factor. It is just a technical adjustment. If you can look at the next page, this is the comparison vis-à-vis the plan. The Q3 revenue growth margin and also the OP margin, I would like to explain them one by one versus the plan. If you look at the column versus the plan, revenue was up +1.2%. However, on a like-for-like basis, it was higher. It came in higher on apple-to-apple basis, higher than the plan. But because of the foreign exchange impact and also the one-off factor, those factors canceled out each other. The actual demand, 1.8% incremental revenue. If you look at the actual demand, automotive compared to plan was slightly higher. Infrastructure, IoT, industrial, and IoT was higher.
Shuhei Shinkai: Therefore, the vast majority came from the contribution of IIoT. There are three major factors here. Data center actual demand increased, consumers, and the back-end process supply improved. Therefore, we were able to have an increase in collection. As for the mass market in the industrial sector, AI and semiconductor-related demand increased. Those were the three major factors behind this upside compared to the plan. As for the gross margin, compared to the plan, 1.9 percentage point increase compared to the plan. Weaker yen mix improvement and the cost reduction for production were the major factors. Cheaper weaker yen, 20%, mix improvement, 50%, and the cost reduction, 30%. Those are the contribution towards this improvement versus the plan.
Shuhei Shinkai: As far as the mix improvement is concerned, the data center and also the supply improvement, that was the factor behind the revenue improvement. Likewise here, the digital power DPU improvement was a contributor for the mix improvement. Also, when it comes to the production cost reduction, the write-down and write-off of inventory did not occur as much as what we had planned, so that resulted in lower-than-expected expenses. Also for the OP, or operating margin, compared to the plan, 2.1% higher compared to the plan. The gross margin, the upside factors for the gross margin was the same here for the OP margin. There were increase in expenses, but the increase in the revenue, so those canceled out each other more or less.
Shuhei Shinkai: The decline was the decline of expenses due to the freezing or the review of the hiring plan, and also AI infrastructure investment inside the company. Also, the increase of the SDI-related provisions offset those favorability. So, the actual landing numbers came in more or less in line with the plan. If you look at the right-hand side, on a QoQ basis, gross margin was more or less in line with the projection. Also for the operating margin, on a QoQ basis, we anticipated expense increase, and we landed according to the plan. By segment, IIoT, on a QoQ basis, there has been a significant increase, as you can see here. But when it comes to the gross margin, foreign exchange and mix was also a factor. Also for the operating profit, the scale increase was the major driver behind the good performance.
Shuhei Shinkai: If you look at the next page, this is the quarterly performance of the revenue. Overall, year-on-year, 30.2% increase on a QoQ, 7.4% increase. The numbers on the brackets are the actuals, so apples to apples, excluding the one-off factor and the foreign exchange, 22.3% increase year-on-year, and 8.6% increase QoQ. The next page. Inventory. Left-hand side is the DOI, and QoQ Q3. Working progress and completed is increasing, and die bank increase, and production increase too, and working progress also increase. These are the major factors. Q4, QoQ, there is increase to be expected. Die bank and production capacity fulfillment by which working progress will be increased, and Q4 has inventory increases to be expected. In addition, on the right-hand side, sales channel inventory shows Q3 QoQ for inventory increases, and auto and industry has been increasing in inventory.
Shuhei Shinkai: On the other hand, for automobile, more than the assumption
Shuhei Shinkai: The sell-through is increasing, so for fulfillment, it is not so large, but it is increasing as a trend. Q4 channel inventory fulfillment is to be continued. For auto, as mentioned at the beginning, 28nm and R-Car mainly will be fulfilled, so that demand fulfillment preparation will be made. For industry infrastructure, data center is still at low level in inventory levels, but in actual amount, it will be increased, and we will work on that. Next page, please. CAPEX and utilization. Front processes wafer. Q3 is as assumed. Kawashiri plant earthquake impact wafer input had been delayed, but impact is very small, and it is 59% as a result. Q4 expects some increases. On the right-hand side, there is a CAPEX situation shown.
Shuhei Shinkai: Q3 has MCU and digital power to be increased in production and in-house factory utilization increases and R&D investments that have been in-house, which is JPY 111.2 billion to be fixed. Increase in production is 80% of overall. For MCU, domestic Naka plant, and for power, mainly for the back-end process for Malaysia factories to be handled for increase in production, and domestic Saijo factory where production will be increased, and the post-process and interim processes is where it will handle those processes. Q3 continues for AI infrastructure post-processes mostly, and in-house factory capacity increases will be worked on. Next page. Q4 and full year forecast. For Q4, on the left-hand side, there is a shaded portion to be referred to.
Shuhei Shinkai: Revenue and profit, JPY 454 billion for revenue, and ratio, 58.0%, 32.0% for OP, and Forex USD 157, and EUR 178 expected. Respectively, the contents will be explained. Revenue, JPY 454 billion is shown at the center. Year-on-year on the right-hand side is 29.4%. QoQ is 4.3% of increases expected. QoQ 4.3% will be explained where yen appreciation factors and Q3 one-time factor may be included. Excluding that, it will be around 4.2% of increase to be expected in revenue, as written at the revenue, excluding Forex and one-time impact. This 4.2% QoQ increase for the revenue for data center and industrial increases are mostly the portions for automobile, 28nm MCU deployments, and R-Car Gen 4 deployments are the main drivers for the growth. Gross margin, 58%, is QoQ 1.4 points decrease to be expected. Yen appreciation and mix and production expense increases are included.
Shuhei Shinkai: More than 50% is mix factor, and the remainder is Forex and production expense increases. For the production expense in CAPEX, as has been explained, gradually there is more investment into production increases, so investment-related expenses in Q4 is increasing. Year-end maintenance is also necessary, so repair works expenses will concentrate at the end of the year. 32.0% of operating margin, 2.6% reduction occurs. In Q4, OPEX increase is expected, and there are about three factors to this. One is the office upgrading and infrastructure, internal IT and AI infrastructure to be prepared for, and investments for that are necessary, and R&D investments. There are seasonal factors where it concentrates towards the end of the year, and other portions are also included. QoQ operating expenses is increasing thereby. At the bottom, there is the sensitivity of Forex.
Shuhei Shinkai: Q4 foreign exchange sensitivity has it where JPY 1, USD 1 to yen fluctuation and impact to operating profit is shown. In dollar case, it is JPY 2.2 billion, and operating profit JPY 1.1 billion. In euro, in sales, EUR 0.2 billion, and operating profit impact is EUR 0.1 billion. So USD 100 and EUR 120 means the performance forecast 32.0% is 26.2%. In addition, for the appendix, please jump over to page 17. Operating profit, the third from the top, JPY 563 billion is shown. As mentioned at the outset, Timing Business transfer proceeds is booked here, equal to JPY 430 billion. So that is the conclusion of the slide explanation.
Operator: Thank you. Now we would like to move on to the Q&A session. I would like to first explain how to give a question. If you have a question, please press the raise hand button on the screen. From those who have raised their hands, we would like to call the name of the person and the company. If you are nominated, you are able to make a statement, so please unmute the microphone yourself and begin your questions. In the interest of time, we would like to limit the number of questions to 2 questions per one questioner. All right, first of all, Goldman Sachs, Takayama-san, please begin. Please unmute yourself and begin your question.
[Analyst] (Goldman Sachs): Thank you. My first question is the revenue for data centers and also digital power that is a portion of that in Q3 and also for Q4. Can you give us the outlook for Q4? Also, as you ramp up your production inside the company, what is the output that can be expected for Q1 of next fiscal year? How further can you increase your production? I would like to confirm these points. When it comes to the profit margin in Q3, because IIoT profit margin has been increasing, so I am sure that data center performance is improving, but other than memory interface, the digital power is also, I believe, is improving. Can you explain those trends if you could? Also, related to data center, agentic AI impact, if you see any impact from that, if you can comment on that point as well. That is my first question.
Shuhei Shinkai: All right. Agentic AI impact. It is not possible for us to quantify that yet. However, the industrial and also the mass market business of ours is indirectly affected by that in the form of revenue increase. That is how we see it. We need to further analyze and study this impact. Also, digital power and all memory interface trends in Q3 to Q4, we are expecting the growth QoQ. Data center overall at the whole company level in Q3, 21% was the contribution to our total sales. In Q4, its contribution will rise further from the 21% level to a higher level. Therefore, this will be one driver of our growth. As far as CAPEX is concerned, digital power front end and also digital power middle end and also the back-end process, for those needs, we are going to make investments.
Shuhei Shinkai: Also, the memory interface front end currently from foundry, we are securing these from the foundries, but we are going to increase the capacity for test at the back end, and that is our plan. Therefore, CAPEX impact on revenue growth in that regard, digital power contribution will likely be higher in terms, and also its scale will be larger. Its impact scale will be larger. In a phased manner, the supply will increase gradually. One of the major steps will be next year, Q1. At Saijo, those capacity increase investment output will come in in full strength from Q1 next year. Therefore, we expect that will have a boost to our revenues.
Shuhei Shinkai: Also, digital power, if you single out that portion only, production increase investment, and also the foundry capacity increase investment, including all these things in 2026, compared to fiscal 2026, double to 2.5 times at the run rate, production increase is expected from next fiscal year Q1. That is the contribution that we foresee right now. Also, as far as the mix is concerned, memory interface, digital power, the customer architecture platform, those are quite well diverse. Therefore, the cost configuration and also gross margin breakdown is quite different from customer to customer, from platform to platform. Just like market share, this changes quite dynamically. That was the major factor. That will be the major factor behind the change of the mix.
Shuhei Shinkai: As a general trend, gradually, as the product matures, the yield will also improve, and also, depending on the maturity of the production, the gross margin will also improve. But in the near term, it will be affected by the mix, and that may seem to be significant. Thank you.
[Analyst] (Goldman Sachs): Thank you very much for that. A supplementary question regarding the data center breakdown. You mentioned the breakdown. On a QoQ growth, Q3 versus the Q4, what is your projection for that? Also, that is going to result in increasing in the Q1 on a run rate basis. Is that what you see? Is that correct?
Shuhei Shinkai: Q3, Q4 comparison. If I give you some numbers, we are expecting an increase. Approximately 10 some percent, mid-teen percent is the growth that we are expecting. We can fairly say that if you consider the comparison between Q3 to Q4, mid-teen level of increase is what we project. Also from 2027 onwards, fiscal 2027 onwards, especially when it comes because we are going to increase the capacity mainly for digital power with increased production capacity. So we are expecting a significant increase, growth from 2027 onwards. So Q3, Q4, Q1 to increase of mid-teen level, and the Q1, you will be able to expect a higher growth rate because of the increased capacity. The capacity of the overall data centers, digital power is the area where we are going to increase the capacity, and that will account for, let's say, less than 50% of the total. So, digital power will increase due to the capacity increase, but for others, it will grow at a normalized level. If you can understand it that way, that would be good.
[Analyst] (Goldman Sachs): Thank you very much. The second large question is about, with the SiTime, JPY 450 billion of proceeds is expected, and for its use, 3 months or 6 months ago, treasury stocks or some kinds of investments were mentioned. We would like to know what your thoughts are about this. From a wider perspective, I suppose performance expense and valuation and stock prices are rather difficult to link together. What are the factors by which improvements and evaluation increases can be gained, including the treasury stocks to be acquired? What is the intent of the company?
Shuhei Shinkai: SiTime proceeds was booked, but half is cash and half is in stocks that we acquired, and the stock portion is not cashed yet. On the other hand, the cash portion has been acquired. For its use, capital allocation framework basically accordingly is applied in order so that it can be used. CAPEX and deleverage and also the investments towards growth areas would be considered. Presently, as mentioned within this slide, our in-house capacity will be increased so that CAPEX will be reinforced for that sake, is our thought. Other than that, growth area investment possibilities will be searched for continuously, and we will prioritize.
Shuhei Shinkai: And determine what is most important, and accordingly, whether we will acquire treasury stocks or not will be determined. At this moment, CAPEX is what we prioritize more. And for valuation, always, Capital Market Day timing is where we talk about our matrix, and is this as in line with that matrix. Organic growth through which sales growth may be gained and scale expansion to be materialized is one point, and valuation gap will be filled in and further reinforced, and then there will be payout out of that to be worked on, and business itself will have the mixes to be changed through digitalization. Composition of the businesses will be changed. Those are the plans we have. That is all for the answer.
[Analyst] (Goldman Sachs): Is there some kind of delay in the payout for them? Because for the cash, there are ample amounts, and there is some kind of additional portion to be considered. Investments, is it so heavy next year, or after the year has ended, maybe you can think in more neutral terms to use them in other ways?
Shuhei Shinkai: It is the latter policy. What kind of items or investments we have, as in balance with the annual budget, we will consider. In addition to that, if it is M&A, there are opportunities for various projects that may be arising, and we will look at them carefully at appropriate timings. We will decide what is best. One timing is at the year-end, and that may be one of the milestones.
[Analyst] (Goldman Sachs): Thank you very much.
Operator: Thank you. Now we are going to move on to BofA Securities. Hirakawa-san, please unmute yourself and begin your question.
[Analyst] (BofA Securities): Thank you. BofA, my name is Hirakawa. I have a question relating to automotive semiconductors. On a year-on-year basis, 19% increase, so you are seeing signs of recovery. I have two questions here. One is that in this, what is the actual demand recovery? If you divide that with the inventory elimination, what is the contribution of these two factors for the increased performance? As far as the demand recovery is concerned, you said R-Car Gen 4 and 28nm. I think you talked about the possibility of share gain, share improvement. So, the market improvement and also the share gains, if you look at your own opportunities for business expansion, can you just sort out what is the contribution of these two factors for your performance improvement? If you can also talk about the sustainability of this performance, that would be appreciated.
Shuhei Shinkai: The automotive year-on-year increase of 19%, how we analyze that I think was the question, right? Before the exchange, the weaker yen favorability is also included in this number. If you exclude that, on a year-on-year basis, 13.3% increase in revenue.
[Analyst] (BofA Securities): What is the result for Q3?
Shuhei Shinkai: The inventory in the channel, I said that we have been building up the channel inventory, but when you look at the sell-through, only half of that level is the sell-through. The actual demand has been improving, and that accounted for the half of the 13.3% increase, and the remaining half is due to the increase of inventory.
Shuhei Shinkai: The sell-through improvement, which translates into the actual demand improvement, the driver behind this was the, as I said earlier, the 28nm MCU, especially in China. China's OEM is adopting this more broadly. That is one factor. Also SoCs, the new Gen 4. Recently, I think this is primarily expanding among the Japanese customers. From Q3 to Q4, those are all increasing. 28nm MCU, mainly among the Chinese customers, we are expanding our sales, and that is also projected going forward. Also the R-Car Gen 4, beyond Japanese customers, the adoption by other regions' customers are expected going forward. Overall, the legacy decline is muting the growth to some extent. To some extent, there has been offset factor from that. But still, our unique growth drivers are the MCU of new 28nm and Gen 4 R-Car. That is the two major distinctive driver for us. Follow up on that. The 28nm MCU expanded adoption by the Chinese OEMs.
[Analyst] (BofA Securities): Is this only one company or multiple companies?
Shuhei Shinkai: Multiple companies.
[Analyst] (BofA Securities): Okay. Got it. Thank you very much. My second question, this relates to all segments of your business. Currently, backlog and the product lead time for each sub-segment, can you talk about the current status of your lead time, if there is any difference between different segments? Also, in 2027 or towards early 2027, given the market situation, how is the order faring, if you can give us some guidance towards 2027 up to, let us say, H1 into 2027?
Shuhei Shinkai: The backlog for each of the segments basically is performing very favorably. Internally, the data center related and also others, we classify into these two different categories and look at the general trends. Also for these other segments, we continue to see strong backlog demand, the backlogs coming in quite strongly, and also for the infrastructure, I am sorry, the data center backlog. I am sorry, the lead time for data center, sorry, is also getting longer, and therefore the backlog cover until H1 of next fiscal year is already there. That is the current status. On an ongoing basis, until H1 of next year, we have a coverage with the backlog already.
[Analyst] (BofA Securities): Okay, then. That applies to non-data center related, for automotive and IIoT, you believe the current market conditions will continue into H1 of next fiscal year?
Shuhei Shinkai: To be more precise, the data center, the lead time is getting longer, so we have a H1 coverage already. Other than that, the situation is mixed. But backlog, generally speaking, is building up favorably, even for other segments. That is what I wanted to say. Thank you.
[Analyst] (BofA Securities): Thank you very much.
Operator: Next, UBS Securities, Yasui-san, please. Please release the mute and post your question.
[Analyst] (UBS Securities): Thank you very much. UBS. Yasui is my name. There are two for myself. The unit price increases is first question. Around from this month, each competitor has started to increase prices, as you said. When would the effects arise? I think it should be around July or thereabouts. Is there effect of the unit price and also the gross margin increases should be expected? Is there such an effect for unit prices?
Shuhei Shinkai: We have been looking at them very carefully, and basically, the cost increases in supply chain and others have been considered in order to work for that.
[Analyst] (UBS Securities): What about the Q3 was the question.
Shuhei Shinkai: To that, it is not so much, but slightly unit prices have been revisited, and that in the Q3 has been impacting.
Shuhei Shinkai: On the other hand, gross profit margin to be improved significantly is not the level of impact that has been reached. It depends on the levels.
[Analyst] (UBS Securities): 0.5%, for example, of contribution, is that the level of contribution that was achieved?
Shuhei Shinkai: Not to that level. Please understand that it is not to that level.
[Analyst] (UBS Securities): Thank you very much. There are various products. It is not for all. On the other hand, what is lacking should be rising in unit prices. What is the product that is easier to raise the unit prices?
Shuhei Shinkai: Generally speaking, mass market products have unit prices that are quite easy to increase. On the other hand, there are cost increases to be added on from that perspective, as according to product. Exposure or cost structure may change as according to how easy it is to see that the costs are increased. All in all, on average basis, mass market products are easier to revise the prices.
[Analyst] (UBS Securities): Thank you very much. The second question is about data center. Digital power, considering the capacity is 2 to 2.5x to be expected for next year. Memory interface or CPU having a major strength is what I have heard about. Capacity considered, what is the sense of level? And in data center, mass markets are growing, and I am not sure of its definition. What is the sense of scale of the data center's mass market, is it? And what kind of products are included in that? Can you give some insight? Digital power- 2 to 2.5x the run rate production capacity. And it is not that it will contribute cleanly next year. It will start up in order. So that is one point I needed to revise and comment on.
Shuhei Shinkai: For data center or memory interface, it is where the front-end process can be considered, and capacity for that to be expanded and acquired is where presently we are working on. What if we have the capacity? It is reverse to that. The actuality is about the capacity to be more considered and the mass market. For example, for data center, air conditioning, direct application is not graphed in many areas which are in the mass market, and that in AI, data center, semiconductor production equipments is where they will be applied into. That is what we assume. This has to be reviewed more deeply, and I would like to explain more details when things are more clear.
[Analyst] (UBS Securities): Intuitively, industrial is where mass markets are included in. Is that the case?
Shuhei Shinkai: That is so. Thank you very much.
[Analyst] (UBS Securities): That is all. Thank you very much.
Operator: Thank you. Moving on to the next question. Daiwa Securities, Oka-san, please begin. Please unmute yourself and begin your question.
[Analyst] (Daiwa Securities): Thank you very much for today. My name is Oka from Daiwa. I also have two questions. Some is overlapping. Regarding the gross margin, I would like to reconfirm. The Q3 gross margin has improved significantly. Memory interface was good, but digital power was not favorable in the past. This is affected by the mix already, but is it also improving? In Q4, if you listen to your story, the industrial and data center is also likely to increase, and the mix may improve, I think, according to what you said, and also because of the unit price improvement. The profit margin for Q4, what is your outlook with the additional contribution may coming in, but is that really going to decline this much? If you can give us your view for Q3, Q4, that would be appreciated.
Shuhei Shinkai: Well, Q3, digital power, first of all, if it increases, the gross margin coming down, not really. What we call digital power includes many diverse sub-product lines, where the different types of products are included under there. Therefore, depending on how these products, the results are different because digital power comprises of new products as well as old products and things of many different platforms. It is a mixture of them. Depending on the change of the mix, the numbers could change dynamically. Digital power overall is not going to push down the gross margin necessarily.
Shuhei Shinkai: That is one comment that I wanted to add in the beginning. Now, Q3 compared to the forecast, gross margin improved compared to the plan. This was due mainly to the digital power and memory interface. Those were the drivers behind this performance. Amid digital power, there are many different elements there. Some with the higher gross margin was the driver increased, and that was a driver. Digital and memory interface also had a contribution because its gross margin is also better. Towards Q4, on a QoQ basis, gross margin 1.4 percentage point decrease is expected. The mix accounts for, let us say, 50% or more of that decline. Because within digital power, the many mix will change in itself. This time around, this will have a negative impact on the gross margin.
Shuhei Shinkai: So that is the difference between the Q3 and the Q4. More than half of the element was the mix factor, but the remaining half will be, as I mentioned earlier, this will be coming from the investment for production capacity. We are going to increase our cost, and also the maintenance towards the year-end would also be done, so that expenses are also included there. So although these are one-off things, that will have an impact to deteriorate the gross margin. That is all. Thank you.
[Analyst] (Daiwa Securities): Thank you. So within digital power, the gross margin, some have a lower gross margin. Is there any room for improvement for these? Are you contemplating any measures to improve the gross margin of those products?
Shuhei Shinkai: As I said a little bit earlier, these are new products, so therefore the yield is generally lower. Because of these reasons, gross margin are low with some of the products. So for those, as the product continues to mature, we can expect an improvement. But on the other hand, the production mix, depending on the site of production, maybe we are using some old sites or foundries, there is a difference there. So, as those cost increases by the site of production, those will have an impact to both deteriorate or improve the margin. So that is how we see it. Thank you.
[Analyst] (Daiwa Securities): My second question is about the memory interface and the digital power. Compared to 3 months ago, can you give us an update of the demand situation? You talked about the run rate 2 or 2.5x. That is quite a large difference. Also, memory interface, the CAGR 2% increase was your original projection, but I think you are already achieving some upside to that. Can you give us a mid- to long-term direction on the Q4? I heard that you are going to elaborate on this in the Q4 or the Capital Market Day, but can you give us an answer for that?
Shuhei Shinkai: Compared to 3 months ago, I would say that on a continual basis, we continue to hold a very bullish outlook. That is the reason why we are going to increase the capacity for production so that we can respond to this increased demand. That is our priority that we are tackling right now. Most recently, the backlog, this is not a midterm thing, but the backlog is increasing in a very steadfast fashion. So, that is another reason we are taking this view. So, compared to 3 months ago, we continue to see a very strong outlook.
[Analyst] (Daiwa Securities): The growth rate, is it rising because of the prices or the banks?
Shuhei Shinkai: I would say, rather, the visibility into the next platform and how much we can contribute to them. So, the visibility vis-à-vis the program is getting higher. That is the situation.
[Analyst] (Daiwa Securities): Thank you.
Operator: Next, Nikkei BP Nakanishi-san, please unmute and pose your question.
[Analyst] (Nikkei BP): Thank you very much. Can you hear me? Yes. Thank you. Nikkei BP Nakanishi is my name. The market perspective and data center especially is the area of question I have. Next year, there will be direct current 800V in the market to be launched. Towards that, the development situation, and when will it be where it will contribute towards the performance?
Shuhei Shinkai: 800V is the version we are very much proactive in working on. Firstly, in our device, it is GaN, gallium nitride, to be applied for the 800V solution. Major customers' architecture candidate structure is including that. In order to secure the projects, we are proceeding with our activities. For the technology lineup or capability, 800, 42, or 12, 800 to 6, respectively. As of this October trade show, the minimum size version has been exhibited. There are various lineups to be supplied so that it can be applied within the architecture. It is not that we have had them applied and that sales is set up yet, but it is a very important area, so we would like to focus on it so that we can secure the business cases. That is the situation for the 800V version. Thank you very much.
[Analyst] (Nikkei BP): With the customers at this moment, is there some kind of dialogue and some kind of considerations?
Shuhei Shinkai: I suppose the application is not yet established, but in the market, when it is launched and if it is applied, then some contribution or in timeframe-wise perspective, what do you expect is the progress? In our mid- to long-term plan, 2030 or 2035, being this and that level for sales plan or target are set. This 800V DC solution in 2030 and 2035 is in our minds, and I am sure it will contribute greatly within that timeframe.
[Analyst] (Nikkei BP): Thank you very much. To add on to this, with this 800V and others next year, the perspective of the market is what I would like to know about. For data center, the capacity level has been mentioned. When will this continue? In 2027, the demand may fluctuate to a certain level, or EV compared, a more stronger situation may be continuing. Can you comment on this?
Shuhei Shinkai: For data center, you mean? For the data center, demand will continue to be strong, and sales channel and inventory would become continuously thin. As previously, there is a lot of accumulation of the background, and in 2027, there would be expected major increases in H2 2027. Is that trend to continue towards H2 2027, or rather, it will continue to be strong.
[Analyst] (Nikkei BP): Thank you very much.
Shuhei Shinkai: Thank you.
Operator: Since we are running out of time, the next question will be the last question for today. MUFG Morgan Stanley, Yoshikawa-san, please begin your question after unmuting yourself.
[Analyst] (Morgan Stanley): Morgan Stanley, my name is Yoshikawa. I have two points. It was already asked earlier regarding your capital allocation plan. For CapEx, you are going to increase the allocation to CapEx. I think that was the nuance that you have given to us. The quarterly disclosure numbers are based on decisions, so it is very difficult for us to foresee. But next year or the 2 years from now, on a payment basis versus a sales number, 55%, I think, was the projection on your capital allocation plan. What kind of upside should we expect?
Shuhei Shinkai: The CapEx amongst capital allocation, on average, it is going to be 5% to the revenue as the midterm average. In reality, it changes year by year. Right now, what is already decided at this point of time, and altogether for this fiscal year, more than JPY 200 billion of CapEx was planned for this fiscal year. This is mainly linked with the investment when we make for the equipment or facilities. Cash out is generally done at that point of time. There is a 1-year or 2-year lapse vis-à-vis the plan. The cash-out timing, from that perspective, the cash-out is expected to be done from next fiscal year onwards. That is our current projection.
Shuhei Shinkai: For the near term, 5%, there are some years where we will spend more than 5% of revenue, but if you even them out or take an average, 5% of revenue is the medium-term model, and I think the actual numbers will converge it with that level.
[Analyst] (Morgan Stanley): All right, thank you. Understood. If that is the case, the revenue is growing in itself, so the investment is going to be a downside pressure in the free cash flow. For your leverage, already at this quarter, your target level of net debt to EBITDA 1x or lower is already achieved at this quarter. I am not just saying that the return is the absolute good, but your target of shareholder return, phased increase, and more than 30% of free cash flow, I think you are well-positioned to achieve that. It will balance well with other growth investments. What is your view on that?
Shuhei Shinkai: Yes. As you rightly pointed out, at the right timing, at the adequate timing, we would like to discern the best priority and make a decision. I think that is what we should do. As it was asked earlier, one good milestone is the end of the fiscal year.
[Analyst] (Morgan Stanley): Okay, thank you. My second question, sorry, is about your SG&A. Up until the H1 of this year, when you average out, JPY 100 billion per quarter was the total amount for the combined amount of our R&D and SG&A. Now, at the H2 of the year, if I also take into consideration your guidance, JPY 110 billion per quarter was the run rate, I think, is currently the run rate, I think. You mentioned that you have to make office investments and also some items, you raised some items for expenditures. Is this level going to be your new normal, or is this just a one-time thing? Is this inflated due to one-off costs?
Shuhei Shinkai: As we previously mentioned, the SG&A OPEX run rate would be JPY 100 billion plus per quarter. This year, if you take an average for this year, that is going to be the level for per quarter. As Mr. Yoshikawa, as you mentioned, for the H2, JPY 110 billion per quarter is the average amount. The financial model, when we talked about the financial model, as far as OPEX is concerned, especially primarily R&D expenses, for the growth of the company for the future, we have to make proper investments. That is the policy that we are adhering to.
Shuhei Shinkai: Of course, this is about a balance with the revenue growth rate, but the OPEX investment size is not going to be considered a one-off thing. We would like to make continuous investments. That is our view.
[Analyst] (Morgan Stanley): Okay, understood. Thank you.
Operator: Q&A session will be concluded here. Lastly, there will be remarks from Mr Shuhei.
Shuhei Shinkai: Ladies and gentlemen, thank you very much for today. The Q3 and Q4 mainly have been explained. Recently, there is a very healthy situation that is continuing. For automobile, production volumes are very strong, but we have our independent strength. 28nm MCU and Gen 4 are starting up, and the strong situation will be continued accordingly. Infrastructure and industrial is also expected. Going forward, growth will be continued, and for that sake, R&D and CAPEX and other factors will be continued as we mentioned today. We would like to conclude at this point. Thank you very much for your participation.
Operator: Thank you and goodbye.
