[Foreign language]
Thank you very much for joining Advantest Corporation's financial briefing for the first quarter of fiscal 2024, despite a busy schedule.
[Foreign language]
I'd like to introduce the attendees on our side today. Mr. Douglas Lefever, Representative Director, Senior Executive Officer, and Group CEO.
[Foreign language]
Next, Mr. Tsukui, Representative Director, Senior Executive Officer, President, and Group COO.
[Foreign language]
Mr. Mihashi, Senior Executive Officer, CFO, CSO, Executive Vice President of Corporate Strategy Group. Mr. Nakahara, Senior Executive Officer, CCRO, and Executive Vice President of Sales Group. Serving as a moderator of today's session, I am Oike of IR Department of Corporate Strategy Group. Today, Mr. Mihashi, Mr. Douglas will talk about the summary of this financial briefing, and after that, Mr. Mihashi will touch upon the financial results of the first quarter of fiscal 2024. After that, Mr. Lefever will present fiscal 2024 outlook before entertaining questions from the audience. We are going to finish up today's session at 7:00 P.M. Japan time. In today's financial briefing, we will use English-Japanese simultaneous interpretation.
If you prefer to use a Japanese language channel, you are kindly requested to click the globe icon on the lower left of the Webex screen and select Japanese in the menu of My Preferred Interpretation Language.
[Foreign language]
If you slide the bar of the balance to the right end, Interpreter, you will hear interpretation into Japanese when the original language is English.
[Foreign language]
If you prefer to hear original audio containing Japanese and English, you don't have to change the setting. Please join us with the default setting. Today's presentation materials are posted on TDnet and our website. The audience joining us through a telephone line is kindly requested to download the materials. Now, during today's briefing, we will project the Japanese version of the materials. Before we begin, we would like to remind you that today's briefing contains forward-looking statements, all of which are subject to risks and uncertainty that may cause our actual results to be different from those in such forward-looking statements. We appreciate your understanding. Firstly, Mr. Lefever will present the brief summary. Please refer to page 4 of the materials.
Yes. Thank you, Junko-san. Hello, everyone, and thank you for joining. First, I'll provide an overview of the FY 2024 before delving into the results of the first quarter. Over the past year, we have outlined the business opportunities arising from generative AI, and in the first quarter, we saw some of this manifest. While we had anticipated the semiconductor test-related market to revert to a growth cycle in FY 2024, increasing complexity is raising tester demand more rapidly than originally expected, especially for SoC testers. In the first quarter, sales, operating income, and net income increased quarter-over-quarter and year-over-year. Against a backdrop of continued demand growth driven by semiconductor performance and rising complexity, our product deliveries to customers exceeded our original plan, and our sales mix improved. Therefore, we are revising up our full-year guidance for FY 2024.
As such, we are off to a very good start for our third Midterm Management Plan. Now, Mihashi-san will explain our first quarter results.
[Foreign language]
[Foreign language]
Now, I'd like to talk about the first quarter summary of the results. Could you refer to page 5, please? In the first quarter, we posted an increase in sales and profit quarter-over-quarter, partly boosted by the yen depreciation against the U.S. dollar.
[Foreign language] .
There was robust CapEx spending by customers for high-performance semiconductors in both SoC and memory, mainly related to generative AI. On the other hand, demand for mature process application had remained soft since the third quarter of the previous fiscal year, resulting in quarter-over-quarter decline in sales. Details of these results will be explained in the following pages. Could you refer to page 6, please? SoC tester sales were JPY 69.1 billion, an increase of JPY 10.4 billion quarter-over-quarter. Sales increased due to a significant increase in demand for the advanced process applications, such as HPC and AI. On the other hand, for mature process application, in addition to the soft demand of automotive and industrial just earlier, sales for display driver IC also decreased. Display driver IC is DDIC. Memory tester sales were flat quarter-over-quarter at JPY 32.0 billion.
Demand remained elevated for high-performance DRAM, particularly HBM. Next, mechatronics systems and service support others. Sales of device interface increased quarter-over-quarter in tandem with increased sales of testers. However, segment sales decreased quarter-to-quarter due to the absence of bulk sales of nanotechnology products booked in the previous quarter. As for service support and others, in the system-level test business, which has high sales exposure to a limited number of customers, sales decreased as the demand remained soft in the consumer sector. Page 7. Now we can see the sales by region. In Taiwan, first of all, increasing complexity of HPC AI-related semiconductors driven mainly by U.S. public companies resulted in the longer-than-expected test time, leading to an increase in tester volume demand from the related country and OSAT. As a result, capacity of testers has been largely digested, and sales of SoC testers increased furthermore.
Next, in Korea, memory tester sales increased as customers' investment appetite remained robust, particularly in the DRAM applications. In China, while sales of both SoC testers and memory testers decreased quarter-over-quarter, the sales level remained high. Regarding the tightening of restrictions on the export of semiconductor production equipment to China by the U.S. and its allies, the direct impact on our fiscal 2024 earnings is expected to be limited under the current rules and regulations, but we will continue to closely monitor the situation.
[Foreign language]
Next page, please. Slide 8. So the sales, gross profit, operating income in first quarter 2024. Gross margin increased quarter-over-quarter, mainly due to the improved mix, reflecting an increase in the sales composition for high-end SoC. SG&A, including all the other incomes and expenses, decreased by JPY 4.5 billion quarter-over-quarter. However, core SG&A increased quarter-over-quarter due to the booking of impairment loss of a portion of goodwill of about JPY 9 billion as other expenses in the previous quarter, as well as the impact of yen depreciation and increase in provisional performance-linked bonuses. Next, slide 9. This is investment and cash flow. R&D expenses, CapEx, and D&A are illustrated on this slide. R&D expenses remained high following the previous quarter. As for cash flow, the operating cash flow decreased quarter-over-quarter due to the bonus and other payment in the first quarter.
With regard to investment cash flow, we completed the acquisition of Salland Engineering, a Dutch test service engineering company, on April 2, 2024. This resulted in an increase in cash outflow from investing activities in the first quarter. The acquisition is intended to strengthen the midterm and long-term growth foundation, as not expected to have a significant impact on the company's financial performance in fiscal 2024. Next, page 10. This is balance sheet. Inventories increased slightly quarter-over-quarter. While our sales have increased, the status of demand recovery varies from one application to another. For high-performance semiconductor, where test demand is growing, we are strengthening procurement to increase production. On the other hand, for product where demand recovery is likely to take time, we will continue to work on inventory management. This concludes my presentation.
Now, I'd like to hand over to Doug, who will go over fiscal 2024 outlook.
Thank you, Yas. Okay. The recovery of the semiconductor market in 2024 presents a mixed picture. While semiconductors for generative AI are likely to grow faster than expected, semiconductor demand for automotive and industrial equipment is expected to remain soft. The SoC tester market size in calendar 2024 is estimated to be in the range of $3.2 billion-$3.5 billion, an upward revision of $300 million from our estimate three months ago. While recovery of tester demand for automotive and industrial applications is likely to take time, test demand is likely to exceed our original expectation due to the increasing complexity of SoC semiconductors, particularly in the HPC segment for generative AI.
On the memory side, the calendar 2024 memory tester market size estimate has also been revised upward to a range of $1.6 billion-$1.8 billion. Prompt supply capacity expansion of testers in response to robust demand for high-performance DRAM has also contributed to the upward revision of the memory tester market size estimate. Taking all these factors into account, the semiconductor tester market size is estimated to increase by roughly 15% year-over-year at the midpoint in 2024, and tester demand in 2024 is expected to grow more than we assumed at the beginning of the period. In light of the first quarter result and the outlook going forward, we are raising our full-year forecast as follows: sales up to JPY 600 billion, operating income up to JPY 138 billion, income before tax of JPY 138.5 billion, and net income up to JPY 105 billion.
Of the sales upward revision totaling JPY 75 billion, SoC testers account for the biggest portion, with high demand for SoC testers expected in both the first and second half of the fiscal year. Memory testers are expected to see a further increase in sales in the second half of the fiscal year. With sales expected to exceed the original estimate, we continue to work on our supply chain management in order to keep up with demand. The gross margin for the full years is now estimated to be 52%, up from the previous forecast of 48%, and this is due to higher sales and a better product mix. The exchange rate assumptions from the second quarter onward are JPY 140 to the USD and JPY 155 for the EUR.
The latest estimate for the impact of exchange rate fluctuation of FY 2024 operating income is positive JPY 900 million per yen of Japanese yen depreciation versus the U.S. dollar and negative JPY 300 million of every Japanese yen depreciation versus the euro. Due to the expected increase in sales from U.S. dollar-based transactions, we have raised our sensitivity estimate for the U.S. dollar by JPY 200 million from April. Next, I'll explain the details of the sales forecast. First, let's look at our semiconductor and component test system segment. The FY 2024 SoC sales forecast is revised upward by JPY 47 billion from the April forecast. At the financial briefing in April, we explained that the lead time normalization for SoC testers has caused the visibility of the business pipeline to be shorter.
Now that three months have passed in such an environment, we are revising up our sales forecast as the increase in test demand has become clearer, driven by increasing semiconductor complexity for HPC and AI applications. On the other hand, for application processors, demand recovery remains uncertain, although demand from factors such as the shift to 3-nanometer process can be expected. Demand recovery for mature processes applications such as automotive, industrial, and display driver ICs is likely to take time. For memory, the FY 2024 sales forecast is revised upward by JPY 17 billion from the April forecast. Although the strength and customers' appetite for investment has been steady from the previous year, we are raising the sales forecast this time due to prospects for a prompt increase in our production capacity.
Now, let's look at our mechatronics segment, where our forecast is revised upward by JPY 8 billion from the April forecast. With tester sales growing, sales of related device interfaces have been raised from the previous forecast as well. For service support and others, our sales forecast is revised upward also by JPY 3 billion from the April forecast. For support services, we expect solid demand due to the steady growth of our installed base. For the system-level test business, sales forecast is raised slightly from our April forecast. Despite continued uncertainty in consumer applications, sales growth is expected in computing and automotive applications on the back of increasing complexity and high reliability requirements for semiconductors. Finally, as I explained, FY 2024 is off to a great start. In the meantime, there are main challenges.
In the recently released Midterm Plan 3, we disclosed a 3-year average of our cost-profit target model in which the operating margin is projected to be 25% with an average sales level of JPY 630 billion. In contrast, our current FY 2024 forecast assumes an operating margin of 23% on a sales level of JPY 600 billion. So, in order to reach an operating margin of 25%, as shown in the cost-profit target model, further profitability improvements are needed in addition to our sales growth. In order to achieve a profit structure that is less affected by product mix, we will push forward with measures such as improving profitability for our HBM business, which is one of the challenges we are facing currently, and we will strive to achieve our MTP3 goals. So, this concludes my presentation. Thank you for your attention.
[Foreign language]
Now we'll go into the Q&A session. We'll take questions in the following fashion. On your web screen, if you press an attendee icon, you'll see a raise your hand button. If the moderator nominates you, you'll see a button saying unmute popping up. Therefore, please speak after unmuting yourself. Please state your company name as well as your name. In order to take as many questions as possible from many participants, we would like to limit to one question per person if there is any excess time. We'll take questions until the very end of the session. We ask you to please speak slowly and concisely since we are distributing this using Japanese-English simultaneous translation. Also, we kindly ask you to refrain from asking questions about specific individual companies. We will take questions in both Japanese as well as English. Now, we'll start the Q&A session.
[Foreign language]
Hello, my name is Yoshida from CLSA. Can you hear me? Yes, we can.
[Foreign language]
Yes, please go ahead with your question.
In the last month, you announced ATE market to be $5.9 billion as an average in the next three years. But this time, you've just revised up your ATE market outlook for 2024. So, could you please share with us your current ATE market outlook for 2025 and 2026, reflecting this in a strong near-term tester demand driven by AI? Could you also please comment about the margin assumptions in the midterm as well if you have any updates reflecting this better product mix? Thank you.
Thank you, Yoshida, for the question. As far as the market shares outlook beyond our 2024, it's very difficult for us to estimate that far out. I can only give you kind of qualitative answers. I think the important thing is that we are now clearly in an upcycle, and we don't see that upcycle stopping within the year. If we had to give an opinion about the 2025 market size, it would definitely be at a growth level beyond this year. We don't know. There's too many factors that go into that calculation. As to, I'm sorry, the second part of the question was, hold on, the margins. Maybe I'll let, you know, Mihashi, do you want to answer on the start of the questions? Thank you.
[Foreign language].
Yoshida-san, with regards to your second question about the future margin outlook. So, as Doug just explained now, with regards to fiscal 2025, we think that growth will continue. But then, this environment does still present uncertainty. Having said that, with regards to high-performance computing, there will be applications related to that that will drive tester demand, which we believe will continue. So, we expect continued growth into next year. So, we have a 93,000 model for high-end market, and we enjoy a dominant position. So, we believe that we can enjoy decent profitability. And with regards to a concerning area, which is HBM, alongside evolution of HBM, performance spec features will be added onto HBM. And with this backdrop, we have opportunities like license upgrades, which we believe can progress forward.
In addition, with generation upgrades, we can try, we can expect new solutions to be adopted, which should drive profitability improvement. Thank you very much. Thank you.
Thank you very much. Next question is from Mr. Yamamoto of Mizuho Securities. Mr. Yamamoto, please.
[Foreign language] あ、水戸商機山本です。
Hi, I'm Yamamoto from Mizuho Securities. Can you hear me?
Yes, I can hear you. Thank you very much for your question. Thank you very much. So, regarding tester outlook, I think this is so difficult to make a prediction for tester. I know that. But having said that, I have a question. In your company, the fabless companies' production plan, for example, or foundry production plan for the future, so that sort of likelihood of the future potential production outlook of fabless or foundry, I think those things will be changing. But against those potential demand, you may have some plan for the tester demand in the future. So now we can see some upward revision this time. So, HPC and AI GPU, compared with those numbers, from your viewpoint, you think your capacity is good enough, or are you producing?
Do you think the customers are buying more testers than necessary, or do you think there are going to be higher demand for the testers in the future? I'm sorry, you have just revised your projection, but now I'd like to know the time for testing. Test time is changing a lot. I know that, but number of chips to be produced in the future. And how do you think about the inquiries of the testers? Do you think this is still strong, or do you think more testers will be required in the future?
Let me answer to your question. I'm Nakahara of Sales Group. Let me answer to your question. So this time, so ever since April financial briefing, SOC area, so JPY 47 billion upward revision was made this time.
[ Foreign language ]
As you said in your question, we try to understand collecting information regarding customers' production plan for the future. We did have some plan as of April. However, things have been changed compared with the information back in April. For the past 1 month, there's a major change, particularly in the area of, so actually, our customer supply chain, advanced packaging capacity is expanding significantly. That's one thing. That means, as you said, number of chips, I think. In addition, actual high-end generative AI logic devices. So I think the very sharp startup, the production, our company as well. So we didn't have such information 1 month ago, but test time changes. That's a new, quite news. Actually, the test time is much higher than our expectation. That's what we learned over the past 1 month.
On top of that, over the past few financial briefings in the past, SoC business, as you know, mainly in Taiwan and also customers, their utilization ratio for smartphone dropped, but it's so difficult for them to recover their utilization rate so far. However, generative AI, the latest platform for V93000, I think their utilization rate is close to 100%. So because of those several factors, now we have revised our forecast by JPY 47 billion in the future. I don't know what happens in the future. So within one month, such a big change takes place in the market. So it's so difficult for us to think about the future. But when you look at the background, number of chips, test time, and utilization ratio, so various factors are considered when you come up with new revision. I have one follow-up question for test time.
So last year, I think it's difficult for you to say that for AI processor, maybe for one year, I don't think test time gets shorter. Is that correct understanding? Test timing for AI will not decrease. It's a very difficult question to answer. Our company, within one month, information drastically changes. So try to find out what happens in the future. But it's so difficult for us to make comment about the future change of the test time. But in principle, customers try their best to reduce their test time. I think that will have what's happening in the future, about one year. But having said that, in parallel, as you know, the latest GPU, very, very fast to revise the roadmap. I think next year, they introduce new devices of GPU. So the development cycle is very short. Under such circumstances, you may think another factor which impacts test time. It's so difficult for us to say what happens in the future. Thank you very much for your answer.
May I add one more thing? May I add one thing? Exactly HPC and HBM, the semiconductor industry, customers, and we, all of us are making many challenges. It's so difficult to imagine. But as for test time in the future as well, it's so difficult to predict what happens to test time. Might increase drastically, or we are making every effort to reduce test time. One thing which is sure is now complexity creates test time difficulty for prediction. This is a kind of huge challenge for us to overcome. Therefore, in the future, the fabless customers, all OSAT customers, foundry customers, we need to have a good networking with those different customers.
We must work hard to enhance the networking. That's how we can strengthen our capability. At the same time, our production capacity, supply capacity should be flexible enough. We are making every effort to enhance our capability to make our supply capacity flexible. Thank you.
[ Foreign language ]
Yamamoto-san, thank you very much. Next, Morgan Stanley MUFG Securities, Wadaki-san, please.
[ Foreign language]
Hello, this is Wadaki speaking from Morgan Stanley.
[ Foreign language ]
Well, what an impressive upward revision. Tsukui-san, Mihashi-san, or Yasuo Mihashi been fairly busy.
[ Foreign language ]
Tsukui-san, just explain. From Taiwan, I'm hearing information about capacity constraints. Second quarter sales is JPY 135 billion. For the second half, forecast is JPY 145 billion on average, looking at your forecast. To what extent will it be able to handle sales delivery? What kind of measures do you have in place?
Wadaki-san, thank you. Well, with regards to HPC, we have been receiving demands from customers for quite some time. We have been preparing for a while. We are working not only by ourselves, but also along with suppliers. For HPC, with regards to our SoC platform, the test time extension has been an urgent matter. This is information that we have been presented with.
To respond to this, if you look at our balance sheet, you'll see we have been strategically trying to increase our inventory to meet such customer demands. But in terms of to what extent we can increase our capacity, if you look over our past history, you'll probably be able to guess that we should be able to handle capacity. And we are speaking with customers. And as long as they provide information in advance, while it's difficult to say, to mention time horizon, we do can say that we are preparing ourselves. Was this answer sufficient?
[ Foreign language ]
Yamamoto-san, thank you for your question. Next question is from Mr. Nakamura of Goldman Sachs Japan. Mr. Nakamura, please.
[ Foreign language ]
Can you hear me? Yes. Thank you very much. Regarding profitability, I have a question. First quarter, the profitability is 5.4% for gross profit margin. So great improvement has been achieved. So I want to know what happens in the background. The SoC tester goes up, the mix is affected, or the memory as well. Can you see the improvement in the profitability more than expected? In addition, so Douglas said earlier, so the annual GPM is 52%, gross profit margin is 52%. That means the gross profit margin may decline in the second half of this year. So could you just let me know what happens?
Thank you very much, Mr. Nakamura. Let me answer to your question. This is Mihashi. So for the first quarter,
[ Foreign language ]
So gross profit margin goes up. There are two major reasons. So number one, the mix has been improved drastically. So high-end HPC, AI server applications, the GPU acceleration processor units, they are booming and picking up very rapidly. Because of that background, our SoC testers are sold a lot because product performance is rather high. So memory and SoC, so the proportion of the sales between the two improved a lot. That is the reason why we have improved our JNP in the first quarter. For HBM, the profitability, we have some concern for profitability. As I said before, for HBM, the necessary upgrade should be made little by little. And that is happening little by little. So toward the second half of this year, we are going to improve the performance furthermore. And we can see some tangible effects of that.
Because of that, the first quarter gross profit margin was improved. When you look at the full year, what will happen for the GPM? As a whole, SoC, as I said earlier, as Mr. Nakahara said, JPY 47 billion additional sales is expected. I think SoC will continuously sold well. On the other hand, when you look at memory tester, for HBM tester, production capacity should be improved. Along with this improvement in capacity, the sales will be increasing toward the second half of this year. So product mix, because of product mix, the second half of this year, the gross profit margin slightly declined.
[ Foreign language ]
Thank you very much.
[ Foreign language ]
Thank you. Next, we'll take a question from Hirakawa-san. BofA, please go ahead.
[ Foreign language ]
Hello, this is Hirakawa from BofA speaking. Hi, my question is about SoC testers. Q1 revenue was fairly high at JPY 69 billion, and if you multiply by 4, but then the full year forecast is JPY 192 billion, which is less than 69 multiplied by 4 times. What are the upsides and downsides? In Doug's opening remark, you mentioned that APs for smartphones are still uncertain. I'd like to also get an update for that as well.
[ Foreign language]
Hirakawa-san, so let me just confirm your question. So for Q1, SoC, if you multiply SoC revenue in Q1, and if you compare that with the full year forecast, is that multiplied number higher than the full year number? Was that your first question?
Yeah, more or less.
[ Foreign language ]
I'll take the question, your question. This is Nakahara speaking from sales. So as of now, as a result of the upward revision of JPY 47 billion, we have come up with the full year forecast as just disclosed. Currently, we are seeing strong demand driven by test time, as well as the increasing capacity on assembly side, up assembly capacity. And utilization is good. And so all three factors are tailwinds for us. So we feel a very good, strong business pipeline. So the numbers that we presented, I feel we feel fairly confident. With regards to downside, well, if there are events or matters that are beyond our imagination, which we never know, we don't really think that unless there's something very negative, our confidence will waver.
[ Foreign language]
Hirakawa-san? Oh, with regards to, can you also comment on the upside from smartphones? Well, smartphone-wise, in comparison with demand generated by GenAI, demand is still smaller, but then overall utilization rate has been heading higher. So for application processor units, there are APUs with AI features that are coming out now. And so there is new SoC tester demand that's being driven more visibly from such AI featuring APUs. Thank you very much. Hirakawa-san, so was there a question about, I think you said something about asking a question about Doug. I think we missed that. Can you repeat yourself? Oh, that's all right. Doug mentioned about APUs for smartphones. Like he said, it's still uncertain, which is why I wanted to ask about utilization rate for smartphones. Oh, thank you for making yourself clear. It's all clear. It's all good. Thank you.
The next question is from Damian Thong from Macquarie Capital Securities.
[ Foreign language ]
Thank you very much. Yes.
All right. Thank you. This is a question about Doug. Congrats on a great result and strong items increase. Just two questions. So the first one is, could you give an update on the timeline for you to develop the next generation of HBM test solutions? Obviously, there's a lot of talk of HBM3E and HBM4 coming around the quarter. What kind of timeline do you think you will have an updated test solution to meet this opportunity?
Yeah, thanks for the question. Good afternoon, Damian. We're already actively in development for those next generation HBM, HBM4. So we have a very strong position with the memory customers. And so we work lockstep, hand in hand with those customers to meet the timelines for their device release.
Got it. So sometime in the next two years, perhaps? I mean, this okay, got it.
Yes.
On the second one, I guess this links a little bit to Teradyne's comments, I guess. I would say I think everyone recognizes that Advantest has a strong position, especially at the GPU major. And going forward, this is the second opportunity, I guess, in ASICs with what they call vertically integrated producers, basically large tech companies in their own in-house silicon. Can you ever talk about the progress you've had in securing those customers? Is your market share increasing? In other words, are you seeing a broadening range of customers coming through for your product? Thank you. That's it.
Yeah, I really like that question. Thank you for asking. I think it's important to differentiate the system and hyperscale companies that have been our traditional customers or their new customers as they develop their own silicon. But as you point out, the ASIC companies that are also working with those hyperscale companies are becoming increasingly important to our business. And I'm happy to report that we have a very, very strong position with those. There's not too many of them. There's just only a small handful of those customers. And so we've been positioned with those customers for many, many years. And without talking about details, I can report that we're very comfortable with our market share at those ASIC companies as well as those end system hyperscale companies.
Got it. Thank you. Just one last follow-up, and then I'll wrap up.
Sure.
If you had to look at it on a 2-year timeframe, would this business be material, I guess, large enough? Because clearly, things are test time, complexity increase. It's universal, I guess. But do you think that this could be an additional growth driver, perhaps as early as next year?
Well, yeah, absolutely. It already is. So we've been active in those accounts for quite some time, as I mentioned. But it's absolutely going to be a meaningful part of the SoC market.
That's great. Thank you very much.
Thank you.
[ Foreign language ]
Thank you, Damian. Next, Okasan Securities, Shimamoto-san, please go ahead.
[ Foreign language]
[ Foreign language ]
With regards to your upward revision for SoC testers, with regards to the change that took place in the last three months, test time was longer than expected. Is this the biggest factor causing upward revision? I'd like to know the details about the change in the last three months. You mentioned about also utilization rate improvement. So what is the thing that you got wrong in terms of your forecast deviating from your original forecast?
I'll take that first. Nak can address some of the utilization details. But yeah, certainly we were caught off guard in this first quarter with some of the unpredictabilities on the test times. And to be clear, our customers were also somewhat surprised as they were releasing new devices. Just the sheer complexity of these devices creates a lot of uncertainties when they model their test times. And only when they go out and start ramping and actually the yields are becoming apparent, do they then have to add tests into the test program or add additional test insertions in order to achieve the desired yield. So we were fortunate that we were able to react quickly because these are some of the biggest companies now in the world that had these demands.
So yeah, we were caught off guard, and that's why it obviously did not show up in the full year disclosure that we had given back in April. As to the utilization rates and the OSAT, I'll let Nakahara-san maybe add some color if that would help.
Sure. Sure. Let me give you a...
Oh, okay.
Oh.
Well, as Doug just explained, the biggest moving needle, I would say, maybe not for the last three months, but more of the last one month. So rapid production ramp-up for high-end SOC really cost longer than expected test time. And this really is the gist or the reason for the upward revision of JPY 47 billion. Aside from this, packaging capacity was expanded. And also utilization in the testing market, these were factors that were somewhat predictable. As I said, for utilization rate in our latest EXA Scale SoC platform, it was full. It reached full rate, which actually reached full rate earlier than expected. But still, I think that the biggest surprise for us was longer than expected test time. Thank you. I just want to check one thing quickly. About one month ago, you held an MTP3 briefing.
Did you take the upward revision into account at that point?
[ Foreign language ]
Thank you. So the MTP3 range, with regards to MTP3, it's unlikely that the numbers will be close to the low end of the range. Well, I wish I could say that. But as I said for FY2024, I think there was a question in the beginning about the FY2025 outlook. We did mention, we did say some make some positive remarks. But then in terms of how the business will progress toward FY2025 or 2026, we don't really yet have full confidence as of now. We need to work with customers to gain input from them to figure out the trajectories for 2025 and 2026. So as of now, I think that we are off to a fairly good start. But in terms of the future trajectory, we are not quite sure how things will land.
We need to sort out our information and report to you as appropriate in due course. Thank you very much.
Mr. Shimamoto, thank you very much for your question. Let me just repeat once again. If you have any question, please select the following method. First of all, on the web screen, please click the participant icon, and please click the raise hand button, which is popped up on the screen. We do have some time. I'm very happy to take the second question as well. Next, Mr. Yoshioka of Nomura Securities. Mr. Yoshioka, please.
[ Foreign language ]
Thank you very much. I am Yoshioka from Nomura Securities. One question I want to ask you. This fiscal year SoC testers, the sales concept is the question I want to ask. As we discussed earlier, for this fiscal year, maybe second quarter and full month, the sales will be rather flat. Second half of this year, the memory tester for HBM is expected to grow.
So when you think about SoC tester, second half sales will be slightly declining starting from the third quarter. That's my prediction. SoC tester sales slightly decline in the second half of this year just because of the short lead time. So there are some uncertainties. So you are very conservative for the prediction of the sales for the SoC. Do you think some upside change in the area? What is your thought about the second half performance of your SoC testers? Thank you very much for your questions. Earlier, Hirakawa-san asked a similar question, I think. At present, as Yoshida-san said now, so when you think of first and second half, SoC sales remain almost flat. Second half, towards the second half, there are slight declines towards second half this year.
However, for the second half of this year itself, actually, we don't know what happens in the entire period of second half. There are some factors which are uncertain. So we are rather conservative at this moment. So now we are seeing some decline in the second half, which might be some conservative consolation. So in the second half, you don't think any factors which drive down your sales. No, no. That's correct. So for profitability, it's improving, but declining. And we are trying to ready for this capacity so that we can increase the number of memory testers. I think that will have a positive impact on our sales.
Thank you very much. I think that is very clear now.
Thank you very much.
Next, SMBC Nikko Securities, Hanaya-san, please go ahead.
[Foreign Language]
Hello. Thank you for your presentation today. I just have a quick question on the gross margin. Margin deterioration has been taking place, and your explanation is mixed deterioration. In the upward revised guidance, looking at the numbers and based on my math, in terms of SoC memory mix, it doesn't seem to be a huge fluctuation. Memory mix doesn't seem to be all that much. But then margin has gotten a lot better from 48% to 52%. Could it be that within SoC, there is a change in mix maybe between SoC as in HPC AI versus auto industrial? Do you think that this is more of a factor compared to memory versus SoC mix?
Doug, I'll take this question or try to do the second. Yeah, you're correct in that the mix is very favorable for us with the SoC, but actually, we were able to execute on some of the things that we talked about in our midterm plan in June as well as in April in terms of improving our margin on the memory side. There were two big areas that we outlined for margin expansion for memory. One was to take advantage of some value-added features for some of the technical inflection points that were happening as the speeds increased. So we were executing on that. Then we were also taking advantage of some economies of scale in our production and our procurement area. We've executed on both of that.
The other thing I might add is we talk about product mix between SoC and memory very often. But then even within memory, there's a mix within the mix. So there's HBM versus DRAM versus flash. And those mixes within the mix can also have a favorable effect on the overall gross margin for the business. And we saw that occur in the first quarter. So hopefully, that helps to answer that question.
[Foreign language]
Hanaya-san, let me add a point or two. So if you look at the screen, if you look at the slide, there's a computing and compute, and there's an auto industrial consumer DDIC mix. And so FY2024, the mix is 80-20. And within compute and communication, HBM revenue is significant this fiscal year. And HBM-related business presents us with richer configurations for our test systems. And so that presents higher average selling price, which comes with better margin for us. So the suggestion that or the implication Hanaya-san mentioned is correct. Thank you very much.
Finally. So last question from Mr. Yoshida from CLSA Securities Japan, please. So regarding Hanaya-san's question, I have an additional question to Mr. Hanaya's question. In the past, computing and communication SOC testers, the sales, so HPC, about half of them, and mobile AP, maybe 20%-30%, and remainder is for RF. So I think that's the kind of composition or breakdown of the demands. But this fiscal year, you said 80% is the proportion of SOC computing and communication, 80% within SOC. What sort of breakdown can we see, especially for AI and HPC? I would like to know the breakdown within the 80%. And you said it's so difficult to see what happens in the second half of this fiscal year. But if there are some upswing somewhat, what sort of application will give you some upward revision of the second half of your sales?
[Foreign language]
Yoshida-san, thank you very much for your question. In the past, we don't announce very precise breakdown of our sales. I think you know our presumption when you ask this question. For this fiscal year, as we said, answering Hanaya-san's question, high-end performance computing for AI accelerator, high-performance computing accounts for the huge portion. That has started up within this fiscal year. That is a major factor. Therefore, relatively speaking, that drives the revenue, first of all. What is your HPC proportion and mobile proportion? As for mobile, as Nakahara said earlier, at present, we don't see any strong market recovery yet. However, in the future, HPC goes on and business may recover to some extent. In addition, HPC improvement, innovation is from PC and other applications or other players will join.
I think that will give us the additional opportunities, the new players. We don't know what players were coming in, but we are going to talk with our customers to find out what sort of new opportunities we can grasp. Based on that information, I'd like to have some opportunity to explain our plan for the future. Once again, in that sense, for upside, so potential for upside is AI is one thing. And it comes to the edge, so smartphone or PC, and there might be some other factors, but HBM should be the upside potential factors. Is that a correct understanding?
I'll try to address that question. Right now, it's being driven so much by the data center business, and that's going to continue. But we are seeing the first signs of some of the edge AI applications, the devices that are going into PCs as well as phones. I read a report where for 2024, they estimate that 20% of the smartphones will be equipped with AI APUs. And the growth rate, the CAGR for that is expected to be 80%. And so we're seeing those things happen, and that presents really, really nice upside because of the volumes associated with the handsets and the consumer electronics are so large as compared to the data center. So as we see those businesses get layered onto what is now foundational with the data center, it's going to even further ignite the industry.
Thank you. Thank you so much.
[Foreign language]
Thank you very much, Mr. Yoshida, for your question. We have received many other questions, but now we'd like to close today's session. Thank you very much for joining us in Advantest Corporation, the first quarter 2024 financial briefing, despite a very busy schedule.