Advantest Corporation
Q1 2021 Earnings Call Transcript

Published: 31 Jul 2021

[Operator:]

Thank you for attending the earnings conference call of Advantest Corporation for Fiscal 2021 First Quarter Ended June 30, 2021. Participants from Advantest are Yoshiaki Yoshida, President and CEO; and Atsushi Fujita, CFO; Kimiya Sakamoto, Executive Vice President of Sales Group; and Yasuo Mihashi, Executive Vice President of Corporate Planning and Stakeholder Relations. First, CFO Fujita will go over the financial results for the first quarter, followed by a presentation on the fiscal 2021 outlook by CEO Yoshida. We will then take your questions. The entire conference call is scheduled for an hour. Materials are available from TDNet as well as the company website.

[Atsushi Fujita:]

Good afternoon, Fujita speaking. I will go over the first quarter results. Please turn to Slide 4. First, the business environment in the first quarter, the acceleration of digitalization as a result of the COVID-19 pandemic has boosted demand for semiconductors used in data centers, PCs and AI applications. In addition, competition among the 5G smartphone manufacturers has accelerated demand growth for related semiconductors and higher functionality requirements. Moreover, as final demand recovers amid the lifting of COVID-19 related restrictions, semiconductors shortages have become palpable in diverse sectors, including the automotive sector. In response, suppliers throughout the semiconductor market are now aggressively investing in production capacity enhancements and advanced technology. Amidst these circumstances, Advantest has endeavored to capture the expanding demand for semiconductor test from every angle by leveraging our broad product portfolio, which is one of our strengths as a solutions provider. We have put the greatest emphasis on stable procurement of parts, amidst tight semiconductor and electronic component supply conditions to ensure our ability to meet the ongoing robust demand, but procurement bottlenecks throughout our supply chain are increasing parts delivery time, which is also affecting the time on our products. Following the surge in the tester demand environment, our customers have been moving to secure products in advance in anticipation of longer product lead time. These trends help to boost orders received in the first quarter to a level far exceeding the record highs in the previous quarter. We also achieved record high quarterly results in terms of sales and operating income. Quarterly income decreased quarter-on-quarter due to the one-time impact of having posted deferred tax assets in the fourth quarter of the previous fiscal year. Next slide. As for the first quarter orders by segment, Semiconductor and Component Test Systems were up 58.7% quarter-on-quarter, at ¥116.2 billion. Orders for SoC testers increased by ¥49.4 billion quarter-on-quarter to ¥97.5 billion. Amidst overall strong growth, outstanding vigor was seen in orders for testers for high-end SoCs, especially application processors, or APUs, which are core smartphone components, and devices for high performance computing, or HPC. Memory tester orders were ¥18.7 billion, down ¥6.4 billion quarter-on-quarter while orders for DRAM testers increased, orders for flash memory tester declined in reaction to large orders in the fourth quarter.

[Yoshiaki Yoshida:]

This is Yoshida speaking to present the fiscal 2021 outlook. Please turn to Slide 12. Outlook of semiconductor tester market trends for 2021. As explained at the briefing session on our new, second mid-term management plan, held in May, Advantest is in the midst of a situation where we can benefit from the expected growth of the semiconductor and semiconductor test markets in the medium to long term, driven by the progress of the digital revolution and the push for energy efficiency improvement. In 2021, demand for semiconductors will grow along with increasing final demand and higher end-product functionality, driving aggressive investments in major regions targeting higher semiconductor performance and production capacity enhancements to solve the semiconductor supply shortage. In the SoC tester market, the production ramp-up of new APU and HPC products as well as increasing device complexity associated with miniaturization should support higher tester utilization at customer sites for the time being. We anticipate further increases in investments to expand test capacity, mainly for high-end SoCs. At this time, the calendar year 2021 SoC tester market size is expected to be US$3.8 billion. In the Memory tester market, device evaluation in terms of miniaturization, multilayering and higher speeds and bandwidth helped by digital revolution has been driving the tester demand continuously since last year. As a result of slightly adjusting our market size outlook, based on the latest customer trends, the calendar year 2021 Memory tester market sites is currently expected to be US$1.4 billion. As our customers have been moving to secure products early in anticipation of longer product lead times, it's now more difficult to estimate the market size than in the past. If the market trends as expected, it will grow more than 20% for the second year in a row. Page 13 shows fiscal year 2021 forecast. Based on our Q1 results as well as future outlook, we have raised our full year forecast to ¥400 billion in orders, ¥385 billion in sales, ¥100 billion in operating income, ¥100 billion in income before income taxes, and ¥75 billion in net income. In this forecast, if it is achieved, orders and sales will reach record highs. Operating income will reach the ¥100 billion level for the first time, setting a new record in 24 years. Regarding future orders, although we anticipate some falloffs in response to the sharp increase in orders in the first quarter, orders should continue to stay at a high level in the second quarter and beyond.

[Operator:]

The first questioner is Mr. Wadaki from Nomura Securities.

[Tetsuya Wadaki:]

Thank you. Wadaki from Nomura Securities. Thank you and congratulations on your presentation on spectacular financial results. I have a couple of questions. I believe that there are many business opportunities being presented, including system-level test, chiplets. China move towards lower power consumption of servers to name a few. I think the tester market itself is undergoing structural change. Can you comment on this structural change? Which driver do you think is most significant for Advantest?

[Yasuo Mihashi:]

Thank you, Mr. Wadaki. This is Mihashi speaking. Thank you for your continued support. You said structural change specifically what particular aspect do you have in mind?

[Tetsuya Wadaki:]

Well, there are several reports on this increase in system-level tests on chiplets, and also in relation to move towards lower power consumption of servers. Intel is losing its market share while accelerators are growing and there is a China factor. From these perspectives, I think there are several very strong tailwinds blowing in favor of Advantest. Are you finding anything that are different from the past? I would be interested to hear your view of Mr. Mihashi.

[Yasuo Mihashi:]

Thank you for that clarification. As you have correctly described, the advancement of various technologies continue to drive the semiconductor industry. For sure, as we've been saying for some time, there are node shrinkage, particularly double patterning in recent years and also going forward high NA – high resolution is to drive the advancement of scanners. Increased complexity of the semiconductor production techniques is good news for the testers, which are needed to ensure semiconductor quality. So going forward, we should continue to expect the technology advancement to be tailwinds in favor of our business. While we don't see a big change, we expect this pattern to be repeated in the future. Moreover, advances in chiplets and other devices with 3D structures that you mentioned will encourage manufacturers not only to thoroughly test individual chips on ATE, but additionally to perform SLTs to ensure the quality of the device once the multiple chips have been integrated. This will also serve as a tailwind for the testing growth. The advancement in semiconductors overall, including functions, low power consumption, these serve as a very strong tailwind.

[Tetsuya Wadaki:]

I see, thank you. My second question is on China. Shipment of semiconductor manufacturing equipment in China is extremely strong. At the same time, some say this is demand being boosted in anticipation of the sanctions being imposed on China. How much is the real demand versus artificial demand do you think? On the other hand, some suspect that there actually would be no sanctions. What is your view on the current state of the China market and what is your outlook?

[Yasuo Mihashi:]

Our sales in the Chinese market are growing, mainly in memory testers for the Chinese companies, of course, but when it comes to SoC testers, sales are not necessarily originating from Chinese companies alone, but rather original orders coming from U.S. fabless companies and elsewhere. So there is a mixture in terms of where the orders originate. What is currently active in our China business is not only one or two testers being supplied to fabless companies, but also larger orders consisting of products that cover the global supply chain, not just from within China. Therefore, we do not consider this to be a purely China risk. Going forward while there, of course, is concern over national economic security, it's unlikely that the U.S. could completely be independent of the Chinese supply chain. I don't think that will be easy. So in that sense, while there is a risk, I doubt either party would embark on the economic sanctions that will kill their respective business. That is our current view. But, of course, we remain vigilant because politics are unpredictable.

[Tetsuya Wadaki:]

I see, thank you. A follow-up question. You alluded to artificial demand and I was wondering if that was mostly attributable to China, but now I take it that that is not the case and that artificial demand is more global in nature. Am I correct?

[Yasuo Mihashi:]

I wouldn't call it artificial demand. It's real demand. It's just that customers have no choice, but to place orders way in advance since testers are not readily available, given that delivery time is getting longer. It used to be orders replaced three to four months in advance, but now they're getting longer. So it's not like double ordering is taking place as some people suspect.

[Tetsuya Wadaki:]

I see, very clear. Thank you very much.

[Operator:]

Thank you. Next questioner is Mr. Damian Thong of Macquarie Capital Securities.

[Damian Thong:]

Damian Thong of Macquarie Capital Securities. Thank you and congratulations on your very strong first quarter results. I have two questions. First is on test time. Clearly, centering on high-end SoC testing demand and production is growing, compared to the past, how much is test time increasing today? That is for this year and what about next year? I believe the test time is a factor that is driving demand for your business and that is the reason why I'm asking this question. My second question is on your forecast for the second quarter sales, which you expect to decline quarter-on-quarter. What is the reason for that? Orders for April to June were rather strong, some of which are to translate into sales in the second quarter? Or are you expecting sales to realize later in the October, December or January, March quarters? Are you expecting the lead time to be getting longer?

[Kimiya Sakamoto:]

This is Sakamoto speaking. Your question is on test time. As was mentioned earlier with further node shrinkage, devices become more complex and test intensity goes up and test time goes up. It's hard to say quantitatively as there are various different devices and we cannot say how much percent increase per different types of devices, but basically with increased complexity, test time tends to get longer, and we expect that trend to continue. Your second question on our second quarter sales forecast. True, slight decline expected quarter-on-quarter.

[Damian Thong:]

Yes. I'm asking for the reason for this slightly conservative forecast.

[Kimiya Sakamoto:]

As CEO, Yoshida, briefly explained earlier, we're not concerned about their production capacity at all. But procurement of semiconductors and electronic component is becoming more challenging and we are factoring in possible impact from that. Orders in the first quarter were strong. And conventionally with delivery time of three to four months, orders in the first quarter were to translate into sales in the second quarter that used to be the cycle. But with the delivery time becoming longer, with that in mind, we are projecting the second quarter sales as you see there.

[Yoshiaki Yoshida:]

This is Yoshida speaking. Should parts and materials procurement proceeds smoothly, it's highly likely that the actual sales could exceed the forecast, but we are factoring in some uncertainties and thus a conservative forecast.

[Damian Thong:]

I see. What is the delivery time right now?

[Yoshiaki Yoshida:]

Currently, around six months.

[Damian Thong:]

I see. So it's getting rather long. And I have a follow-up question. The new product that you announced this year, the EXA Scale test system, what's the impacts or how do orders look like?

[Yoshiaki Yoshida:]

Yes. Adoption by our key customers is proceeding very smoothly for engineering to mass production purposes. So going forward, we expect it to make contribution to SoC tester business.

[Damian Thong:]

I see. Thank you. That's all from me.

[Operator:]

Thank you. Next questioner is Mr. Yoshida from CLSA Securities.

[Yu Yoshida:]

Thank you. Yoshida from CLSA Securities. I hope you can hear me. My first question is on order forecast, which has been revised upward by ¥50 billion. By application, where is the increase coming from? SoC testers were very strong in the first quarter. So were those the source? Can you give us the breakdown by application as well?

[Yoshiaki Yoshida:]

We made upward revision to order forecast by ¥50 billion on a four year basis, mostly in relation to SoC testers, almost all in relation to SoC testers. By application, computing and communications account for nearly 70%, automotive, industrial equipment and consumer applications over 20% and DDICs a little less than 10%. That's the breakdown.

[Yu Yoshida:]

I see. In other words, basically in line with the current product mix, correct?

[Yoshiaki Yoshida:]

Yes. The breakdown is basically in line with the current makeup. We also made upward revision to the sales forecast for Mechatronics Systems, Services, Support and Others, but in terms of orders, they're not reflected right. For orders, SoC account for a large proportion. So in comparison, Mechatronics Systems and Services, Support and Others don't have much effect in terms of sales. Most of upward revision to sales is in relation to SoC testers.

[Yu Yoshida:]

I see. Thank you. My next question, tester market outlook is becoming strong for the remainder of the year. What about next year? What is your outlook for the market next year? And if you are projecting an increase or a decrease, what are the factors behind?

[Unidentified Company Representative:]

Indeed, we are having discussion to identify the outlook for 2022. For the second half of 2021, visibility is still very much limited, partly in relation to the delivery time becoming longer. We see signs of change in the pattern. So if I could talk about the second half of this year first, regarding the possible large investments by our major customers in fiscal year 2022, that decision has yet to be made by the customers. And as these decisions become clearer, we will have better accuracy in our guidance for the second half orders. Frankly, the accuracy of the current guidance is not that high. As CEO, Yoshida, briefly mentioned earlier, there is a possibility of upside to our guidance. So we need to take all that into consideration, but based on what is currently available, TAM for calendar year 2022 is expected to be similar to or greater than that of calendar year 2021. That is our expectation. So in that sense, fiscal year 2021 with orders of ¥400 billion and sales of ¥385 billion is going to be a very good year and we expect this to continue into fiscal 2022.

[Yu Yoshida:]

I see. Thank you. That's all for me.

[Operator:]

Thank you. The next questioner is Mr. Hirakawa from BofA securities.

[Mikio Hirakawa:]

Thank you. Hirakawa from BofA Securities. I have two questions. First is probably partly the same as Mr. Yoshida's question earlier. As for orders, you expect the ¥400 billion level to be sustainable on a constant basis. You raised your forecast for this year from ¥350 billion to ¥400 billion. Of this additional ¥50 billion, how much is coming from the semiconductors shortages and how much from technology advancement, which is more sustainable? I know it's very difficult to say exactly, but as a general feel, what is the ratio? If it's only semiconductor shortages, I'm afraid it is going to be temporary. That's my first question.

[Yoshiaki Yoshida:]

As you have correctly indicated it's really hard to say exactly, but my personal view is that the investments related to node shrinkage continues and there are foundries that have a capability for more advanced process node, although the number is limited. And the high-end devices and the mobile-related applications, we see migration to more advanced a process node, for example, 5-nanometer node this year, 4-nanometer next year, 3-nanometers beyond that. Of course, we don't know if that's exactly going to happen, but that is the current projection. And if that's the case and as indicated in relation to the very first question, test time is getting longer. And of course, before the next product is launched, there will be an attempt to shorten the time if that could be afforded. But actually with the adoption of more advanced node, the test time is increasing. So the more complicated the process, the more test items. So instead of time being shortened, the test time is getting longer. As long as this trend continues, the current situation that we find ourselves in, the very strong need for testers, we believe will continue. So we believe order growth to be maintained.

[Mikio Hirakawa:]

Thank you. My second question is about market share. The company has raised its market forecast by US$500 million of which US$400 million is for SoC and US$100 million is for Memory. On the other hand, the sales forecast has been raised by ¥35 billion, depending on ForEx. And I believe that the company expects to take 70% or so of the overall market growth. How do you expect your market share to develop in year 2021? Would you please share with us your thoughts on this point? What is the expected trend? That's all.

[Yoshiaki Yoshida:]

As you said, we have raised the total available market for SoC to US$3.8 billion from US$3.4 billion previously. That's an increase of US$400 million. The breakdown of the US$400 million is about $250 million for computing and communications and the $150 million for automotive and other. Based on that, our calculated market share for SoC is about 50%, that's the result of our simulation. Naturally, this will change as the environment changes, but this is our share expectation under the current environment. So that's 50%. For Memory, it is revised up by ¥100 million. Of that, DRAM accounts for 50% and flash accounts for 50%. Based on that, we expect our market share to be above 50%. That's what we are expecting.

[Mikio Hirakawa:]

Thank you very much.

[Yoshiaki Yoshida:]

Thank you for your question.

[Operator:]

The next question comes from Mr. Nakamura of Goldman Sachs.

[Shuhei Nakamura:]

Hello. Thank you very much for the opportunity. I have two questions. The first is about the sales target for fiscal year 2021, which was revised upward. According to the three-year plan that was released in May, the range was ¥350 billion to ¥380 billion. The current level seems to be already higher than that range. Taking in account the business environment for the next three years, is the company already in a situation where the target needs to be reversed upward again? Can you share your latest outlook, including market forecasts?

[Yoshiaki Yoshida:]

I do understand that it's difficult to foresee the future at this point. We do not believe that the cyclicality is gone. The business trends of going up and down repeatedly remains still unchanged. We formulated our midterm plan sometime in the latter half of last year towards the first half of this fiscal year. So there is no doubt that the recent trend is much more upbeat compared with what we had assumed at that time. If the business is strong next year, for example, for fiscal year 2022, it may decline in fiscal year 2023. If that will be the case, I believe that the average for the next few years could be ¥380 billion. So that's still possible. So that's my first point. Another point is about the upside potential. With the current shortage of semiconductors, there are moves to increase capacity around the world and plans to expand capacity, not only for 300-millimeter, but also for 200-millimeter. And they are announced one after another all over the world. It would not be surprising if a peak demand for testers come one more time when those new fabs are up and running. Assuming such an upside scenario, it's not surprising if the company outperforms its midterm plan targets. However, our midterm business plan has just begun, and we have not yet reached a point where we would revise it upwards. That's the current situation. I understand now. Thank you very much.

[Shuhei Nakamura:]

The second question is about the shortage of parts and materials mentioned earlier. If possible, please give us more detail on the shortage of those materials and how much it is actually affecting your production and sales? Also, please comment on the impact of the lockdown in Malaysia, which was started in June and is supposed to be lifted in August later? So could you give us your comment on this?

[Yoshiaki Yoshida:]

Let me answer the simple question first. The lockdown in Malaysia has not had a major impact on our production so far. In Penang Island, the situation is not that serious. We are told that the production itself is on schedule. So we do not think there will be any impact on the production capacity itself. However, as for the shortage, we are running out of inventory for a wide range of parts and materials. For example, in semiconductor manufacturing, there is a shortage of substrates. Some people are wondering how long this is going to last and say that it may last for next two to three years. That seems to be the situation. Today, it takes many more substrates to make a three-dimensional semiconductor than in the past. Even though it is suddenly needed many times over, it can easily take one or two years to build up the capacity. So the situation that we are facing at the moment has never been experienced before. It's not just the parts that are used to make our testers. But a lot of things are happening in upstream, and it is becoming a big problem for the entire semiconductor industry. As a result, it's difficult to procure some semiconductors to put on to our testers. This is a matter of whether the chicken comes first or the egg. If we can't make testers, they cannot make semiconductors, we cannot make testers. In order to avoid such a situation in the overall industry, CEOs and other top executives of companies are having discussions at the moment to address and continue procurement. Unfortunately, we cannot disclose specific company names though.

[Shuhei Nakamura:]

Understood. Thank you.

[Yoshiaki Yoshida:]

Thank you very much for your question.

[Operator:]

The next question is from Mr. Hasegawa of Mitsubishi UFJ Morgan Stanley Securities.

[Yoshihito Hasegawa:]

This is Hasegawa from Mitsubishi UFJ Securities. Let me ask you a quick question. What was the timing of the decision to revise the forecast this time? Earlier, you mentioned that the midterm plan was formulated in the second half of last year to first half of this year. Can you give us some details on when exactly you actually decided to announce the revision?

[Yoshiaki Yoshida:]

We discussed the figures for the midterm plan until – just before the announcement, and there were no major orders as of April that would require us to revise the figures we had already announced. The major orders started to come at the end of April and continued into May and June and also into July. In April, we could not foresee the current order trend. So that was what happened. Orders for the fourth quarter were quite significant. And therefore, we thought that there will be a fall in the first quarter. However, they trended in the opposite direction. This is what actually happened. So that's why we included the April trend when we formulated the midterm plan that we announced in May. And we thought that we could achieve the range of up to ¥380 billion on average. So we set that as our target. So in summary, the change in trend in the first quarter was beyond our expectation.

[Yoshihito Hasegawa:]

Thank you. The revised plan refers to a lead time of six months, but I understand that this lead time is the period from the time your company receives an order to the time of delivery. Can you comment on the assumptions for material procurement in this revised plan? Well, there are many materials and components with longer delivery times according to the revised results; the second half of the year will be generally flat in terms of the level of sales. I would like to know if you could share any assumptions about the procurement of parts and materials.

[Yoshiaki Yoshida:]

If we suddenly ask the suppliers to double their supply, they will not be able to respond to such a request. Naturally, we procure based on the supply and demand outlook, but we also procure parts with long lead time well in advance. However, orders are increasing rapidly now. So we are asking for delivery as soon as possible, but the suppliers respond to us only with delivery dates they are able to promise with confidence. However, over time, it is possible that the delivery dates will gradually shorten. So in that sense, we will be making efforts to bring forward sales as much as possible compared with what we currently expect based on our procurement plan. Well, that's a possibility with the current shortage of semiconductors all over the world. We do not think we can accelerate the delivery so dramatically, say, three months or two months. In that sense, you can understand that our current sales plan is within the range of levels that we can stably procure.

[Yoshihito Hasegawa:]

If that's the case, do you think that the level of ¥123.8 billion in the order backlog at the end of the fiscal year can be generally fulfilled, but not 100?

[Yoshiaki Yoshida:]

Even if there are ¥120 billion left at the end of the fiscal year, that doesn't mean that it has to be fulfilled in the next three months.

[Yoshihito Hasegawa:]

Understood. Thank you very much.

[Yoshiaki Yoshida:]

We don't know if it will end up at ¥120 billion or if it will increase any further. It also depends on when the delivery dates are. At that point in time, if customers still think that they can increase production, it's possible that the level of the red bar on the chart will be higher.

[Unidentified Corporate Representative:]

I would like to add that there is a considerable possibility that the order book may expand even more. So it seems that the current type situation will not be easily resolved anytime soon, but the company will do its best. At the moment, there are very few semiconductor suppliers. I believe we can confidently say that they will shift the required volume at any given time.

[Yoshihito Hasegawa:]

Understood. Thank you. The second question is completely different from the first. But I was wondering if you could explain the impact of the partnership with PDF Solutions. It's been about a year now since the investment was made last year.

[Yasuo Mihashi:]

This is Mihashi. I would like to answer your question. We started our partnership with PDF last year. Since then, we have been conducting activities to implement our cloud environment on their cloud platform. And in such an environment, we are currently rolling out various applications for customers to use. We have even started to operate some of these assets outside of the areas of mass production. However, having said that, in terms of the contribution to sales, we still have to deep dive into the desired customer value proposition. At this point, no significant impact has been generated yet. But first and foremost, we would like to utilize the data to improve the effectiveness of testing or the efficiency of the overall semiconductor manufacturing process, while continuing to conduct deep dive analysis.

[Yoshihito Hasegawa:]

Understood. Thank you. How do you evaluate then the results of year one in terms of ROIC, R-O-I-C, for the capital the company has invested.

[Yoshiaki Yoshida:]

Well, for year one, as mentioned earlier, we spent most of the time building an environment where we could do business using data before we could recover our investment in terms of ROIC. That's how we spent the first of the year. From the beginning, we did not plan to be able to recover our investment in terms of ROIC in a short period of time. That's not the nature of the investment. Regarding data analytics, it is a business that if we don't do, we could put our tester business in jeopardy. So there was such an implication that we had to protect the tester business. Based on that understanding, the investment was made from a medium-term perspective with the intention of building a software-centric business ourselves. So we are not planning to recover the investment in a certain number of years in terms of ROIC.

[Yoshihito Hasegawa:]

If so, is it correct to say that the investment was made in anticipation of the comprehensive impact on the company's business rather than to recover the investment in a certain period of time, such as three years?

[Yoshiaki Yoshida:]

Yes, you're correct. In that sense, we are already seeing an impact. Joint development with customers that directly leased to hardware sales is already underway. And if we consider the sales of testers that have come out of it, we can say that we are already recovering a large part of the investment. However, this impact is not counted as a return on investment in the software business per se.

[Yoshihito Hasegawa:]

Understood. Thank you.

[Yoshiaki Yoshida:]

Thank you for your question.

[Operator:]

With this, we would like to conclude the Q&A session. Thank you very much for your participation in our briefing today.