Advantest Corporation (6857) Earnings Call Transcript & Summary

<<foreign_lang>>

That's very clear. Thank you so much.

<<foreign_lang>>

<<foreign_lang>>

Thanks a lot, Ms. Shikanai.

<<foreign_lang>>

And now we'll take a question from Alex Chang at BNP Paribas Securities. Alex, go ahead.

Appreciate the opportunity to ask, and congratulations on the strong results. I have 2 quick follow-ups. First, on gross margin: this quarter's gross margin looks like a new high. I recall management saying memory price inflation would have a meaningful effect on margin, yet the impact appears fairly limited. Can we confirm that, at this level, the company has fully absorbed the memory price inflation impact, and that the company did not raise prices or pass this cost inflation on to customers, and that the company is comfortable with the current memory price — or is there any potential impact if memory prices rise further? My second follow-up: could you give me some color on your market share difference between x86-based CPU and Arm-based CPU? If Arm-based CPU share increases in the future, would that mean any negative impact for you, or a positive one? Thank you.

<<foreign_lang>>

Alright, thank you, Alex. Doug here. Let me begin with the margin question. We certainly experienced the impact of DDR pricing. In both our SoC and memory testers, we use a substantial amount of DDR. A significant portion of the operating margin decline in the first half or second half relates to that increase in pricing. During the first half, we benefited from older DDR pricing thanks to our inventory, but as we now look toward the second half and beyond, we will need to deal with these elevated spot prices, just as everyone else must. We are actively seeking ways to offset those margin pressures. Through new configurations, licensing, and new instrumentation, we consistently have the capability to reprice our products. That represents a major element of our approach, rather than simply raising prices directly, which is something we prefer to avoid. We will keep facing those kinds of headwinds from DRAM prices as well as other component inflation. Even so, we are still performing at a very high level.

<<foreign_lang>>

The next question concerns the CPU landscape. I addressed this a quarter ago, and my answer remains the same: we are extremely well positioned throughout the whole CPU market, in the x86 accounts and at the Arm-based companies alike. Our position with the 93,000 across the board feels very strong to us. As for which company wins and which one picks up share, whether among the x86 players or on the Arm side, that matters little to us, since the same tester is placed at every one of those accounts. That is part of why you are seeing this second half top line growth — our positioning in the CPU market is that strong.

<<foreign_lang>>

Mind if I ask one quick follow-up?

<<inaudible>>

Sure, absolutely. If we look at how your revenue growth stacks up against your competitors, it looks like you did quite a bit better across this quarter and the one before, and you said the CPU was the primary driver or primary contributor. Can we confirm that your share of the CPU tester market might actually be even greater than your share in the AI accelerator tester? Is that the main reason?

<<inaudible>>

That's a fair question. We generally prefer not to split things out the way you're suggesting. What I'd point to is this: across the whole high-performance compute market, our share is very high. GPU, custom ASICs, CPU — in each of those areas we likely hold the majority of the market. On top of that you have networking and several other pieces. Silicon Photonics came up earlier, and there too our position is quite strong and expanding. So it's all of those together. That picture hasn't really shifted for many years. We remain a dominant player in accelerators. We are clearly the dominant player in custom ASICs, and the same holds for CPU.

<<foreign_lang>>

<<inaudible>>

Much appreciated, Mr. Chang.

<<inaudible>>

Now we'll take a question from Mr. Nakanomyo at Jefferies Securities. Go ahead, please.

<<inaudible>>

Before we move to your questions, I want to give an update that puts our numbers in a clearer light. Let me begin with the second quarter. Revenue came in at 136 million, which is a 12% increase over the same period last year. Gross margin was 47%, up 300 basis points from a year ago. We also saw strong momentum in CloudShield, which contributed 18 million in the quarter. Turning to guidance, for the full year we now expect revenue of 540 million to 560 million, and we are raising our EPS outlook to a range of $2.10 to $2.20. Before closing, I want to thank Maria Chen and her team for their work on the launch.

Yeah, we can hear you just fine.

<<foreign_lang>>

Thank you very much. Now, I’m not sure if it’s appropriate for me to put this question to you. But as an analyst who is very active in the stock market, we do actually examine various data points, and in particular we turn to the Taiwanese SPE import data. Previously, there was a fairly good correlation between that data and SoC testers. In this second quarter, though, the situation is somewhat different, and there is a significant gap. There is no correlation. And I’m wondering why. Does this have something to do with SoC logistics issues? Or perhaps it’s the way the statistical data are collected, or maybe the equation or formulas have been changed. Please explain why this gap exists in the second quarter.

Thank you, Nakanomyo-san. Regrettably, that's something we simply can't address. I've come across the reports where folks attempt to monitor exports from Malaysia and imports into Taiwan, and my sense is those models contain a great deal of noise. My strong suggestion would be to consider different KPIs instead of those. You might be able to get an early read, perhaps, on what our figures will turn out to be, but I'm just not sure that approach is the best one, since that data is quite imperfect, I'd say. I can't account for why the trend has appeared different to you, because those numbers aren't something we examine ourselves.

<<foreign_lang>>

<<foreign_lang>>

I appreciate that.

<<foreign_lang>>

Thanks a lot, Masahiro Nakanomyo.

<<foreign_lang>>

We'd like to invite Mr. Kinoshita from UBS Securities to pose a question to you.

<<foreign_lang>>

Thanks so much. Again, congrats on a wonderful, superb result and performance in the last quarter. My question concerns the operating income margin and your improvement. In the second quarter, you registered a very good margin and margin improvement. And you mentioned just now that the annual margin will also be expected to be higher. As much as possible, could you please explain to us the contents or the reason why that margin is being improved?

Thank you, Mr. Kinoshita. I believe I covered earlier the drop we saw between the first half and the second half. Inflation on the components that go into our systems accounted for a large portion of it. Another point from the prepared remarks: we had an inventory reversal in the first half that was essentially a one-time event, and it pushed first-half margins above what we'd normally expect. That won't continue going forward. So we're basically leveling off in this low 50s. And as I said as well, as devices grow more complex, the configurations of the testers keep expanding. Those give us chances to expand margins, on both the SoC side and the memory side. So we've been able to leverage that to offset much of the component inflation we're encountering. Hope that answers.

<<foreign_lang>>

The source excerpt to rewrite wasn’t included—only a placeholder was provided. Please send the actual earnings-call speaker turn, and I’ll rewrite it as requested.

I appreciate that. If I could, I’d like to ask one more question. These days, people talk about a significant memory shortage. Your company does point out that the SoC tester and CPU might be highly promising drivers for really lifting performance going forward. Yet when you think about a generative AI system, you need a lot of components — memory, CPU, and custom ASIC among them — and every one of those parts has to be synchronized. Otherwise, should memory be in short supply, or should memory not be integrated into the system, you can’t ship it as a product. And looking at future performance, I’m curious whether the memory shortage has some effect. So would you mind commenting on that?

That's accurate. Right now memory is in short supply globally, and many fabs are under construction at the moment. For the most part, though, over the past 2 years wafer starts have shifted out of older legacy DRAM and into high-end HBM and server DRAM. That has created significant difficulties for industrial and other automotive applications. So we're certainly not insulated from it. What I would add, however, is that we belong to a community that includes our customers, and they recognize that getting testers requires them to supply us with memory. We lived through this during the supply chain shortages of 2021 and 2022. The industry truly came together to ensure components would be made available to the equipment suppliers, since everyone understood that failing to do so would produce even larger shortages everywhere. I think it's tight, without question. But we also have extremely strong suppliers assisting us.

<<foreign_lang>>

<<foreign_lang>>

Thank you so much for that question, Mr. Kinoshita.

<<foreign_lang>>

Since we're nearly out of time for this briefing, we'd like to take one final question.

<<foreign_lang>>

And our final question comes from Mr. Hanaya at SMBC Nikko Securities. Go ahead, please.

<<foreign_lang>>

Thanks so much. It's a privilege, honestly, to be the one asking the final question. So, what I'd like to ask about is the ramp-up in production capacity. You said you're close to truly hitting the targets for 10,000-unit output. As for timing, that hasn't been disclosed yet, of course. But naturally, you've worked hard to boost manufacturing capability, since certain needs have to be met. And do you hold some ambition of moving onward from 10,000 to 15,000 and past that? Or is it simply too soon to be considering that sort of aspirational future? And how do you think about folding this in — perhaps in the MTP coming up, we'll actually tackle it. Would you mind sharing your thinking on expanding production capacity further?

Thanks, Mr. Hanaya-san. We already have some aspirations in mind. Going from 5 to 10 was a big step, and we're speeding that up as quickly as we can. Yes, we'll share more specifics on the timing of that, possibly on the next call. What I'll say is that I believe the next step will be another doubling, from 10,000 to 20,000. That's something we're actively discussing within the company, based on customer demand. Obviously, we're not ready to discuss the timeframe for that right now, but I think certainly by the time we release our midterm plan four, we may be ready to talk about it then.

<<foreign_lang>>

<<foreign_lang>>

It’s good to be with everyone. Let me start with the numbers for the quarter. Revenue came in at $136 million, up 18% from a year ago. Gross margin was 74%, compared with 71% in the same period last year. Operating expenses were $82 million, and that includes $6 million in stock-based compensation. Operating income was $19 million, and operating margin was 14%. Net income was $15 million, or $0.22 per diluted share. Free cash flow was $31 million. We ended the quarter with $412 million in cash and cash equivalents. Turning to the business, our Cloud Platform segment grew 26% year over year, and we now have 1,240 customers spending more than $100,000 annually, which is up from 980 a year ago. Our Enterprise Suite product contributed $58 million in revenue, up 12%. Looking ahead to Q3, we expect revenue between $141 million and $143 million, which at the midpoint represents 17% growth. We expect non-GAAP operating margin of about 16%. For the full year, we’re raising our revenue guidance to a range of $552 million to $556 million. I’ll now hand it over to Sarah for the financial details.

Thanks a lot, Mr. Hanaya, for asking that.

<<inaudible>>

Even though more questions have come in, we'd like to wrap things up at this point.

<<foreign_lang>>

We appreciate all of you making room in your busy day to be part of our financial briefing.