I'm Mihashi. I will present the financial results. Please turn the page four , on the third quarter result for FY 2023. The third quarter sales and profits exceeded those of the second quarter. Unfortunately, as you can see, they are still trending in negative territory year-on-year, but signs of recovery are starting to emerge. In the area of generative AI, sales grew by responding to robust demand for memory testers. In addition, sales of testers for some high-performance SoC semiconductors and system-level test increased quarter-on-quarter. Details will be explained sequentially in the following pages. Let me move on to the next page. On FY 2023 third quarter sales by segment. Regarding the third quarter for fiscal year 2023, all segments recorded an increase in sales quarter-on-quarter.
Regarding semiconductor and component test systems, SoC tester sales were JPY 63.3 billion, almost flat quarter-over-quarter. Sales for mature processes, such as automotive and industrial applications, decreased quarter-over-quarter, as there was a lull in customer investment after an elevated level of delivery up until the previous quarter. On the other hand, for advanced processes, an increase in sales for applications processors, in addition to high performance computing, offset a decrease in sales for mature processes, although such increase in sales does not represent full-fledged recovery. Memory tester sales were JPY 25 billion, an increase of JPY 6.2 billion quarter-over-quarter. Sales for high-performance DRAM for generative AI increased. For mechatronics systems, sales of device interface product increased in tandem with an increase in tester sales. Regarding service, support, and others in SLT business, sales increased for high-end SoC.
Please turn to next page. On FY 2023 third quarter sales by region. As you can see, for Taiwan, in addition to a large increase in sales of SoC tester sales for high-end SoC, sales of SLT increased. Regarding South Korea, while SoC tester sales declined, memory tester sales increased mainly for DRAM. As a result, total sales remained elevated as in the previous quarter. Regarding China, despite a decline in sales for high-end SoC, the overall level of sales remained at high level. Please turn to next page. On sales, gross profit, and operating income. While sales increased quarter-on-quarter, gross profit margin increased only slightly quarter-on-quarter. SG&A increased JPY 3.5 billion quarter-on-quarter. However, second quarter includes approximately JPY 3.2 billion of other incomes, which includes settlement income. Therefore, SG&A was effectively on par with the second quarter.
Now, if you turn to the next page, investments and cash flow. R&D, CapEx, and D&A remained on par with the previous quarter. On cash flow, free cash flow was an outflow of JPY 7.1 billion due to payment of bonuses and income taxes based on FY 2022 income. Now, if you turn to the consolidated financial position, inventories remained at an elevated level, and supply capacity expansion is being implemented for businesses with robust demand, such as high-performance memory, while responding to requests for pushout from some customers in other businesses. On the liabilities side, taking into account the working capital needs, we have taken on additional debt funding. In order to monetize inventories as quickly as possible, we will make effort to refine our supply chain management. That's all. Next, I would like to talk about the FY 2023 outlook.
Now please look at page 11. Business environment, semiconductor tester market trends. I would like to share with you current outlook.
Now, calendar year 2023, semiconductor market was more sluggish than expected. Due to delayed recovery of a major consumer electronics appliances, such as smartphones and PC, ramp up of tester demand, which was expected to take place in the second half of calendar year 2023. In our initial forecast, as of the beginning of calendar year 2023, has also been delayed. Therefore, we estimate calendar year 2023 SoC tester market size to have been between $3.3 billion-$3.4 billion. In calendar year 2024, we expect demand ramp up to take place in the second half, led by high performance semiconductors, such as HPC and AI. We estimate calendar year 2024 SoC tester market size to be between $3.3 billion-$3.6 billion.
We estimate, calendar year 2023 memory tester market size to have been approximately $1.1 billion. In calendar year 2024, we estimate memory tester market to be between $1.3 billion-$1.6 billion. While non-volatile memory is expected to gradually show tester demand recovery from the second half of the calendar year 2024. High performance DRAM tester demand is likely to ramp up rapidly and stay active throughout the year, in tandem with customers' increased production plan and rising quality assurance requirements. We estimate that our market share for both SoC testers and memory testers have reached the upper 50% range in calendar year 2023. Now please turn to page 12. FY 2023 forecast.
Based on the performance in the first nine-month period and an outlook for fourth quarter, we are revising up our full year consolidated forecast. Sales forecast is revised up by JPY 10 billion to JPY 480 billion, due to factors including increased demand for memory testers. Gross profit margin forecast remains unchanged from the previous forecast of approximately 50%. Operating income forecast is revised up by JPY 5 billion, as there is no change in the forecast for SG&A. Exchange rate assumptions for fourth quarter are JPY 145 for the U.S. dollar and JPY 155 for the Euro.
Our latest forecast for the impact of exchange rate fluctuations on FY 2023 operating income is JPY +0.8 billion per JPY 1 of yen depreciation versus U.S. dollar and JPY -0.3 billion per yen depreciation versus the euro. Due to the expected decline in sales from the U.S. dollar-based transactions, we have lowered our sensitivity estimate for the U.S. dollar by JPY 0.1 billion f rom October.
Please turn to page 13. This is the outlook by segment. Let me explain the outlook for the semiconductor and component test system for FY 2023. FY 2023 SoC tester sales forecast is JPY 244 billion, which is JPY 4 billion lower compared to the October forecast. For advanced processes, demand recovery is gradually observed. On the other hand, for mature processes, such as automotive and industrial applications, the level of tester delivery, which had stayed elevated, appears to be taking a pause toward the second half. For advanced processes, despite customers' production for increased demand in areas such as generative AI, which are gradually prompting the digestion of excess tester capacity, full-fledged recovery of tester demand is likely to take a while longer. FY 2023 memory tester sales forecast is JPY 81 billion, which is JPY 5 billion higher compared to the October forecast.
In high performance DRAM applications such as HBM, which is related to generative AI, robust demand is expected against the backdrop of medium-term demand growth and high reliability requirement. In order to respond to such demand, we are working on expanding our supply capacity and have raised our sales forecast. Please turn to page 14. This is the outlook for mechatronics system for FY 2023. The sales forecast is revised up by JPY 5 billion compared to the October forecast. Owing to an upward sales revision of memory testers, sales forecast for related device interfaces is revised up. FY 2023 services, support, and other sales forecast is revised up by JPY 4 billion compared to the October forecast. For maintenance services, solid demand is expected due to the steady growth of our installed base.
System-level test business is expected to be flat year-on-year, due to the impact of specific customers' investment trends affected by the sluggish semiconductor market conditions. Going forward, we will pursue a medium- to long-term sales growth, as well as profitability improvement, as we move forward to provide solutions and expand our customer base into areas such as for HPC or AI, where we can expect adoption of the system-level test. Please turn to page 15. This is the summary. We have been expanding our supply capacity structure in response to tester demand growth for high-performance memory. We have revised up our sales forecast by JPY 10 billion from the October forecast of JPY 470 billion. On the other hand, recovery of SoC tester demand is slower than expected.
While tester demand for high performance semiconductors, such as such as HPC and AI, is turning upward, digestion of excess tester capacity is taking longer than originally expected at the beginning of the fiscal year, although signs of recovery are emerging. In 2024, supply/demand balance improvement of semiconductors is expected. In addition to demand recovery of semiconductors for consumer applications, increasing demand of semiconductors and capacity increases on the supply side for generative AI, is expected to lead to related CapEx and drive tester demand in the second half. In the mid to long term, as generative AI technology advances, structure of industries as well as society at large are likely to change. We aim to expand the scale of our customer base by grasping changes in the semiconductor value chain and providing new solutions in order to pursue further growth.
Finally, I would like to touch on profitability improvement. Changes in the past year were significant. However, semiconductor demand is likely to continue to fluctuate going forward. We will work to refine our supply chain management to improve our responsiveness to demand fluctuations. The basic approach to improving profitability will be through the provision of higher value-added solutions in response to technological evolution of new high performance devices. In parallel, we will review the selling prices in light of rising materials procurement costs and strengthen our cost reduction activities. This concludes my explanation. Thank you for your attention.
Well then, Yoshida-san from CLSA.
Can you hear me?
Yes, we can hear you.
Thank you. Slide 11, tester market, I have a question. FY 2024 outlook has been presented. Thank you very much. If possible, SoC tester, memory tester, for each, if you have the breakdown for FY 2023 and FY 2024, could you explain your view, particularly regarding SoC, HPC customers, for computer customers, mobile customers? What's your prospect for the market and for memory, DRAM, particularly HBM composition? Will that change from FY 2023 to FY 2024? Based on that, what is your prospect of the market share for FY 2023? You said that it's around 50% for both. Toward FY 2024, what is the forecast, and as well as what's your view on the gross profit? Thank you.
I'm Nakahara, in charge of sales. Let me respond. Page 11, FY 2024. The calendar year 2024 estimate are provided. SoC first, JPY 3.3 billion-JPY 3.4 billion for... is the projections for this term. We expect a slight growth, as was mentioned during the presentation. Generative AI-related logic device tester demand is the area where we believe the growth is coming. Regarding SoC tester, at test house and/or at OSAT, tester utilization, if it's on the full utilization of the new device, are to be a purchase. That is the business model that we have. Therefore, with the, the utilization has heightened, however, it is not at the full capacity, full utilization yet. And from the second half of 2024 for large logic device tester, we believe that the growth can start to emerge. So SoC tester for 2024, AI HPC related logic device is the area that will become a big business for the company. At the same time, for calendar year 2024, in the latter half, smartphone related, our customers, applications processors customers, they've been shifting to the new process node. I believe that it is highly likely, therefore, the smartphone related demand, it was slowing down. However, starting from the second half of calendar year 2024, we may see some movement. And another point regarding memory, high performance, high-end DRAM, the LPDDR5 or HBM for calendar year 2024 will become a very big market.
At the same time, toward the second half of the year, RAM related business, to a certain extent, can recover, and it can come closer to the upper end of this JPY 1.6 billion. So HBM may be the main business. Now, talking about the market share, for calendar year 2023, both for memory and SoC, 50%, high 50%, is what we have stated. Now, for calendar year 2024, same level of market share, can be secured and maintained.
Another point about the gross profit margin, Yoshida-san, this is Mihashi speaking. Good afternoon. Regarding gross and profit, what are your view is another question that you have raised. Yoshida or Nakahara has already explained about this point as well.
In the first half of next year, the memory will continue to be strong, so recovery of the SoC may be still a bit slow. From that perspective, the current situation that we're in, as well as for the first half of next year, we don't expect to see a large change. Therefore, with regards to gross profit, it will start with the same level, and in the second half, with the recovery of the SoC, there will be the improvement of the product mix. And what shall be the improvement of the graphic process, gross profit is something that we will place a close eye on.
That is it. Thank you. Thank you for your response.
Thank you very much. Next, from Goldman Sachs, Mr. Nakamura, please.
Hello? HBM tester is my question. So sales, memory testers, sales, was upgraded. That is because of the increase in for HBM tester. Three months ago, or in November, there was IR day, and in comparison to that timing, demand outlook for HBM testers, how has it changed as to the sales increase revision? Is it because of the demand increase, or is it because of the manufacturing volume increase of yours? Can you tell us that as well?
Yoshida is going to answer.
As of around November, already for HBM, demand has started to increase. Toward the end of the year, the plan, looking at the following year, was fixed among our customers, and the outlook for the next full year forecast was given to us by our customers. According to that, in comparison to the outlook in September and October, there is, there was further elevation from that time. As for our supply ability, memory testers are manufactured in Japan, and how we want to improve the supply capacity of that is one of our challenges. In order to fulfill all the demand increase in the market, we need to make further efforts. Having said that, what I would like you to be cautious about is, demand change could always occur.
So in our manufacturing, we want to levelize things as much as possible, and at the same time, we want to fulfill the customer's manufacturing plan. So to the next full year, we want to improve our manufacturing capacity. At this moment, we are not at the stage that we are able to provide sufficient supply. So though we see an increase in sales, but because of the constraint of the manufacturing or supply, the situation is as such. That's the reflection of the reality.
Thank you very much. As a follow-up question, what you said is you have visibility up until next year, means up to 2025, you have some business inquiries for the manufacturing. Is that the correct understanding? And at the time of IR Day in November for HBM tester sales toward the next fiscal year versus previous year. So I understand it's increased by 50% or double. Hasn't that been changed?
So when you say double, it may be too much of an exaggeration, but the September, October, November, in comparison to our estimate back then, it could be possible, the amount could be double. And, looking at the year 2025, we don't have a clear visibility, but, until the end of 2024, we have visibility that the level will be staying at high, which will roll over to 2025. But, until what time in 2025 will that continue? We don't have a clear visibility, meaning that at our customers' level, we don't have that, they don't have that much plan.
I understand it really well. Thank you very much.
So supplementary comment, at the time of memory explanation, meeting, what we said is the following: HBM market, in terms of CAGR, 50% per year, meaning that as for next year, about 1.5 x, the following year, about 2 x or double. That's what we communicated. This is a supplementary comment.
Thank you very much. Thank you very much.
Hirakawa-san, of BofA Securities, please go ahead.
Thank you. My name is Hirakawa from BofA Securities. I have a question about SoC testers. In the fourth quarter, SoC tester sales, I believe if you do the math, it is JPY 57.1 billion, and this is a plan dropping from the second and the third quarter. And in the fourth quarter, I believe it is supposed to go up, but it is actually coming down. I think it is related to the mature node. So, can you please explain the reason? Also, I would like to confirm one thing. As I was listening to you, recovery expected at the time of previous briefing is going to be delayed by six to nine months, and I believe the situation is still the same from three months ago.
Is there anything that you can elaborate on, on this matter? Thank you.
Hirakawa-san, this is Mihashi speaking. Thank you very much for your question. Toward the fourth quarter, the number for SLC testers seems slow. I believe that is your question, and I believe you have done the math. And as you saw, in the fourth quarter, SLC is weaker compared to the third quarter. Normally, in the second half of the calendar year, our SLC customers drive smartphone-related demands. That is normally observed. However, as you all know, smartphone business in the recent situation is becoming weaker. Therefore, normally, strong numbers come out for the fourth quarter on QoQ basis, but that is not the case this year. And related to smartphones...
Well, there were opportunities, businesses that were pulled up from the fourth quarter. However, the continuation of those opportunities are not guaranteed at this moment.
Thank you. SoC tester market recovery is expected to be delayed by six to nine months. Hasn't that changed since three months ago?
I believe that is being pushed out to a certain extent.
Can you please elaborate on that?
As we explained during the briefing. Well, this is related to what Nakahara-san said earlier, but computer communications and automotive breakdown, if you compare those two, automotive and consumer products were stronger in the third quarter, but they are becoming weaker. And smartphone-related business has been pushed out or has been becoming slower.
That is pushing out the expected recovery of the SLC tester business.
Thank you.
Thank you. Next, Jefferies, Nakanomyo-san.
Hello? This is Nakanomyo from Jefferies. Can you hear me? Yes, we can hear you clearly. Thank you. System-level testing, third quarter, the sales increased considerably, but on a full year basis, you have revised downwards the forecast. What are the situations as well as the forecast for next, next year?
Go ahead.
Yeah, I can, I can take this question. Thank you. As far as SLT business goes, we've experienced, because of the downturn, particularly in mobility, a little bit of softness. But we've seen, you know, some level of recovery in the recent quarter, both from some capacity buys as well as new customers. And then also, we've seen an emergence of some compute-level requirements for system-level tests. So, still, year-over-year, we expect the situation to be relatively flat, but we're encouraged by some of the things we've seen in the past quarter. Yeah, the next fiscal year, it's still uncertain, but as we see mobility come back, that should drive capacity purchase at our key mobility customers.
And then, as in the remarks, we've seen some HPC opportunities come into SLT insertions for extended coverage, and for yield improvements. So difficult to say, as it is a very dynamic situation, but long term, we still feel that the SLT insertion will continue to grow.
Thank you very much. Now, based on what you have mentioned, for 2024 overall, how do you view the next year? Previously, initially, 2024 will recover to the 2022 level, and the previous briefing, you mentioned that it may be slightly delayed. What is your perspective for next fiscal year? Any general image, general view toward the next fiscal year?
2023, recovery started, and in 2024, should recover to the 2022. That is the scenario that we talked about in the past one year. However, recovery in 2023 is not as high as we expected. There is a push out. That is the actual situation we are in. However, what will happen in 2024, centering on AI? In other words, supply capability of AI would go up considerably. Demand is there, however, the supply is not catching up.
With the improvement in that front, SoC, memory, I believe that the demand itself would increase. Therefore, for 2024, if we reach the higher range of outlook, JPY 5.2 billion, or maybe the level, which is the level of the 2024, combining the memory and the SoC tester. But the SoC tester was still weak, and the memory is showing more robust recovery. So for the entire tester market, there is the opportunity to recover to 2022 and 2024. However, the memory tester, even if it becomes stronger for the profitability, we may not be able to catch up, and that is the challenge of our company.
Understood. Thank you.
Thank you very much. Moving on to Nomura, Yoshioka-san, please.
Can you hear me?
Yes. Hello.
Thank you. Yoshioka from Nomura. Away from the earnings result, I would like to ask questions of my concerns. North American biggest competitor of tester with a large probe card manufacturer, they formed a partnership. That was what happened in November last year. To your business, moving forward, what are the possible impact in the mid to long term in the short term and mid-term? And what kind of actions or policies are you going to have on this topic? Can you share with us your view?
Yeah. Yeah, thank you for the question. I think it's a very good question. As far as the impact that we'll see, the discussions we've had have been confidential, but the policy we have is that there's an open ecosystem for testers and for probe cards. That's the understanding that we have with the probe card suppliers. It's the understanding that our customers have with us, as well as the probe card suppliers. So we're confident that even though our competitor has made an investment in one of our partners, that there will be no effect. So we have confidence in that point.
Thank you very much. I understood. So if I may do supplement, the probe card is used together with tester. So install the unit of tester. When there is a more installed tester, then the amount used for probe card also increases. So the investment by competitor, won't they do business with us?
No, that's not the case. So our customer, when they use our tester, and then it's natural that the transaction will be made with them. And when competitor having interest in consumable business. As mentioned in our grand design, that is, looked from a total solution, it seems our competitor is thinking the same thing as us. So by observing the competitor's movement, we also have to think about the way of our business.
Thank you very much for your comment. I understood. Thank you very much.
Thank you, Yoshioka-san. Shimamoto-san of Okasan Securities, please go ahead.
This is Shimamoto speaking from Okasan Securities. About SoC testers, in the second half of 2024, recovery is anticipated, so it has been pushed out. But regarding the timing of recovery, to what extent do you have a visibility? For example, from receiving orders to delivery, I believe the lead time, if it is 6 months, then recovery in the second half will have a fair probability, I would think. So the timing of recovery, is there any risk of being pushed out even further? How are you looking at the situation?
Then I would like to answer your question. The lead time of testers has come back to a normal level. It depends on the model and configuration, of course, but it is between two to six month timeframe.
So market recovery is not anticipated in the second half just because of the lead time. But what we are monitoring is the utilization rate of our testers installed in the market. According to the recovery of the utilization rate, well, if we look at the installed testers, configuration is changed from smartphone to HPC AI, for instance. Such trend is occurring, and that is leading to improvement of utilization rate. When the utilization rate peaks out and larger configuration SoC testers' demand will be recovering in the second half of the year. That is our expectation.
Since the previous financial results announcement, maybe you are feeling that it's slightly pushed out further, but as Mihashi-san said, automotive, industrial, related SoC tester demand is slightly weaker, and that may lead to potential adjustment of the timing. But when it comes to the probability, well, I have to say that the market changes constantly, so we don't have a full visibility. But depending on my experience and depending on customers' plan for next devices, we are trying to develop our plan.
Thank you. I have a follow-up question. If possible, the situation of utilization rate, three months ago, six months ago, and today, how has it been trending? If it is possible, can you please share it with us?
Well, it is difficult to give you concrete numbers, but in around spring, I would say in April, the utilization rate was quite low, and it was the bottom. But compared to six months ago, the situation is improving. The peak is at around 90%. However, it's going to take a while longer to reach that level. That is the current situation.
Thank you. That is clear. Thank you. That's it from me.
Thank you. Next, Daiwa Securities, Sugiura-san.
I'm Sugiura. Thank you. Thank you. Regarding the SoC's tester, how particularly the market outlook for high-end SoC tester in the second half of the year, high-end HPC around smartphone, you expect to see the recovery. That's what you have mentioned. And if this has been expedited and front-loaded, what may be the trigger for that? Take, for example, PC and smartphone. Without the recovery in the volume, and if there will be the expedited schedule, what may be the trigger? Do you have any ideas on that?
Yoshida will respond.
The possibility of front-loading of the schedule, what we are closely monitoring is the supply side. The demand is there, AI chip itself, but we are not manufacturing the amount that is required. Foundry, the largest foundry-
is increased the productions from January, and it is constantly going up. And I don't think it is highly likely that the schedule will be front-loading. However, if the capacity increase will proceed in line with the plan, then sometimes in June and in July, they can reach to a ramp up stage to increase the production. And the testers, there may be some shortfall in the testers that is up and running, but for mobile, can it be further reactivated? We cannot say that for certain. Take, for example, what was announced by Samsung AI chip to be incorporated into the mobile phone, and if that will further spread. So such a new product will be launched in a larger volume, then there is a possibility that the things will be front-loaded. However, that we have not seen the sign of that. AI for the mobile device is what you have mentioned.
Now, even without the increase of the numbers of the unit for mobile, with the enhanced feature of the AI, then, there will be the increase, on the bottom demand. Is, is it correct to understand that?
Well, yes, that is what we expect, inclusive of the possibility of the new entrant.
Thank you.
Thank you, Sugiura-san. Next. So we had a connection issue. Once again, Morgan Stanley MUFG Securities, Wadaki-san?
Hello? So there is noise.
What is the benefit for you?
So I was not able to hear you clearly at the beginning. What were you asking? About what did you ask? Did you say AI? AI smartphone, did you say? Wadaki-san, can you repeat? Maybe you are on the move. Wadaki-san, can you hear us?
I can hear you. It's about AI smartphone. Can you elaborate on AI smartphones?
So AI smartphones. Doug-san, can you respond?
I didn't quite catch the question-
Oh, okay.
but I think it's referencing AI smartphones.
AI, yes.
Well, you know what I can say, everybody is hopeful that there's gonna be new applications in handsets, using AI that will drive a new replacement cycle for phones. Already, most of the APU manufacturers have built in AI cores into the AP. So the capabilities are in some of the phones already, and the use cases are what's coming. So it's been very public from many CEOs in the industry about the need to do more edge compute, using AI because of a lot of the power consumption, privacy, workloads. And so we see that coming, and when it does come, that's gonna drive a lot of volumes.
So in the time right now, we've seen a pivot from mobility into HPC that is digesting a lot of capacity in the cycle, and then on top of that, there's gonna be the addition of mobility and consumer using AI at the edge, which should drive some very, very large growth for the industry.
Did I answer your question, Mr. Wadaki? Most probably, he was able to hear us, so sorry. So we would like to conclude our answer. Thank you.
Now, we are almost at 5:00 P.M., so I would like to take the last question from the ones raising their hand. Yoshida-san of CLSA, please go ahead.
Thank you. This is my second time. Thank you for this opportunity. I have a question about the market share outlook for 2024. In 2024, in SoC, AI, HPC, and application processor and in memory and tester, DDR5, HBM demand are expected to grow, especially in the second half, according to your explanation. These applications have a relatively high market share of ADVANTEST. So I think we can anticipate a further growth and not flat year-on-year. But do you think you can realize a growth higher than the market growth in 2024?
Thank you, Yoshida-san, for your question. Regarding the market share, it is not easy to give you a direct response. The two companies are dominating the market. However, there are other competitors too. We have high 50% market share, and it is not easy for us to say we will aim at even higher market share. But according to the market situation, there's HPC and HBM, high-performance semiconductor testing demand, which are our strength, and there is high expectation for our company. That is true. However, setting a concrete target for market share or saying that explicitly is something we choose not to do in order not to offend other competitor.
Then I believe that is still a possibility, right?
Well, looking at the customer base and our product portfolio, we believe there is a fair probability. That is my personal view.
Thank you. That's it for me.