Amkor Technology, Inc. (AMKR) Earnings Call Transcript & Summary

All right. Great. Thanks, everyone, for joining this session. I'm Randy Abrams. I cover Asian semiconductors out of Taiwan for Crédit Suisse. For the next session, we're pleased to have Amkor, one of the leaders in back-end test and packaging. With us, we have Giel Rutten, who's CEO, was -- has been at the company for a long time and became CEO this past summer, also Megan Faust as CFO. So format for this, we'll do a fireside chat. Feel free to e-mail me if you have any questions, and we'll try to fit it into the discussion. So yes, I think for Giel, we can start just to get everyone up to speed. If you could give an overview now for Amkor, just within back-end test and assembly space, how is the market positioned and a bit of profile just the applications. There's a lot of drivers across semiconductor space, which has helped the company positioned for those drivers within the back-end space.

Yes, Randy. I think -- now let me give a short introduction on Amkor. Let's start with this year and where we see the company being positioned this year. Amkor is a company, indeed, in the assembly and test space in the semiconductor industry, we are a manufacturing service company. We offer end-to-end services from let's say, wafer bumping, wafer probing, assembly and final testing. We have a strong position in the communication markets, where we do -- about 45% of our revenue is generated in the communication business, mainly with our advanced product portfolio from wafer level to flip chip to System-in-Package portfolio that we supply in that market. But we're also present in all the other key markets that are there in the semiconductor industry from consumers to computing and the, let's say, the automotive markets, where in the automotive market, we definitely hold a leading position and that's a strong market for us. Of course, this year, the automotive market is under pressure. But we still see, as a company, we're performing fairly well this year. We see growth in all the other segments that we're serving, and we expect to see this year growth year-on-year of over 20%. So we're performing well. The products segments that are doing well for us is System-in-Package across multiple applications, not only the communication market, but also the consumer market. And that's -- as a high-level introduction, Randy. Of course, I can go on and on.

Okay. We'll keep you going on and on just with a few -- a bunch of follow-ups to that. For the -- I guess, as we close out the year, I mean, one thing we've seen, even after you reported, it seemed like a lot of the companies we cover in Asia, they're still at the conference, pretty upbeat as we go into the end of the year. I'm curious, for fourth quarter, where you're guiding maybe a bit of a sequential decline. It's coming off a couple of strong quarters. But if we just factor in like the later launch of flagship phone, auto seems like it's turning the corner. So is there kind of maybe a factor you're still seeing that seasonality kick in for -- as we close out the year in fourth quarter, that might be driving a bit of softness?

Yes, I think across the segments, I mean, we see auto recovering indeed, I should say, although the recovery if you look to the overall supply chain, where we serve the supply chain is slightly slower than our customers because the inventory is still being corrected going forward. So a lot of the end market recovery is supplied out of inventories. We also see some impact on some of our SiP business, impacted by, let's say, component and supply shortages where end customers make choices across applications. That will impact our consumer business in the fourth quarter. Also with the shift with some of the products moving from one generation to another. So we have some puts and takes in the quarter. So we expect a few percentages down quarter-on-quarter. Of course, we came back -- we came from a very strong third quarter.

Yes. That's right, especially for the year, still growing over 20%. I think one thing that's happening, we're seeing -- and it seems like the Asian suppliers are starting to mention the wire bond, like traditional packaging that hasn't been tight for a long time. Getting very tight, raising prices. I think 2 parts to that. But one, if -- given their capacity tightness, if any potential incremental benefit if you have capacity available. And also, to the extent you're seeing that mainstream part of the business also starting to tighten up?

Yes. Before I answer that, I mean, the profile of our wire bond business is slightly different than, let's say, the general market. I think we are very much tuned towards supplying our automotive customers with wire bond solutions, more the conventional wire bond. So I think roughly 40% to 50% of our wire bond business is going into automotive. So we have fairly little exposure to, let's say, the communication market or the compute market with our wire bond businesses, mainly let's say, automotive and industrial and some specific segments even in that market like microcontroller. So we see that recovery coming in the fourth quarter, but also next year when the automotive market recovers, we see also our wire bond market recovery. If we look to the interaction with our customer base, and it's generally covered by longer-term contractual agreement. It's beyond a tactical engagement where you can -- where prices go up and down. So we see some opportunities to take some of that upside, but that's relatively small. So the dynamics in what we consider our wire bond business is slightly different maybe than in some other parts of this market. So we see yes, some corrections because some of the component prices is going up like, for example, gold, as a material it goes up. So that means that the prices in the end market go up, but that's more like a pass-through pricing structure here. So I don't think that we see a significant price increase on the wire bond side.

Okay. That's fair. You mentioned earlier tightness, it's limiting a few things, like some of the System-in-Package or some of the wearable-type products. Maybe the extent because we do see there's been a couple even accidents at companies like AKM and Unimicron, and there was already some tightness on things like substrates. To what extent are you seeing in your supply chain tightness as a bottleneck or limitations? And how you're managing through the supply side at this stage?

Yes. I think on the supply side, of course, the substrate supply is very tight, certainly with the recent events with one of the bigger Taiwan suppliers. We have a specific team in place to look for alternatives, to work with customers, to qualify alternative suppliers in Korea, but also some other, let's say, jurisdictions. We have a strong procurement team in Japan to look for alternative supply there. Overall, we are able to minimize the impact of that tight supply with -- to a very low number in the fourth quarter, slightly bigger number in the first quarter, but still overall to our business very small. There are other titles of supply. Some of the components are very tight, and they are managed by our customers and our customers make choices where to allocate these components. So overall, yes, there is tightness. Some capacity buckets on the OSAT space is also tight on our rate level packaging. So -- but impact is still manageable, I would say, from our perspective.

Okay. Great. So in the -- I think one other thing that could trigger tightness, but I'm curious if you're seeing any of that. One of the China suppliers JCET was put on the military end user list. Have you seen any types of diversification or maybe we should say like reverse, China. China has been localizing, but maybe the reverse trend of some business maybe shifting from international customers, it could be to your benefit? Or do you see even recently with the JCET restrictions, if it may have any impact or given maybe less U.S. content in the back end, it may not have that much impact. I'm curious what you're seeing from that front.

Well, I cannot comment whether this is specifically related to one of our competitors. But I see a general trend that some of our customers in specific areas are derisking their supply chain because of their perceived risk of having either single-source supply out of China, or some industries that want to derisk that China supply chain, for example, in the automotive domain. So we see some of that. Yes, where there is derisking of supply chain risks. And it can either be done by either us looking for second source. And then, of course, for Amkor, that could be potential opportunities or shifting more dramatically a full supply chain out of China. But I think -- if anything, I think it would be some tailwind for us, and we see that.

Okay. That's good. And the other tailwind, it's -- I mean, typically, it's maybe the delayed reaction, automotive, where you mentioned earlier, there's still the inventory. Where do you think we're at in terms of that inventory getting back to normal if -- sorry, as far as timing, when you may start to see the general market automotive turn the corner? And I guess, the customer feedback at this stage, how they're viewing next year? At this stage if I get through that inventory correction?

Yes. That's a good question. I mean, if we look to, let's say, the data that our customers are publicizing them. We see that a lot of the end market upside in the third quarter was supplied out of inventory. Of course, companies hold different levels of inventory. But we're seeing that at the end of the third quarter and also this quarter, it's much more in balance than at the end of the second quarter. So I personally see that at the end of this year, the automotive supply chain is very much in balance. And maybe even some replenishment is needed in the early part of next year. Then the production numbers of cars next year, what I see, it still will not -- is not expected to be back at 2019 numbers, still be a little bit depressed but a much more normalized supply chain. It's not a V-shaped recovery. It's a gradual recovery. But definitely, I think longer term, for Amkor, we strongly believe that the automotive market is a strong growth engine. And even companies that follow this market carefully, they expect that the semiconductor growth in the automotive market is higher than the average semiconductor market growth. So for us, it's still an important market. We continue to invest in it, not only in capacity but also in building up capabilities.

Okay. One market on the other side, away from auto, but just into high-performance computing. Could you discuss, in 2 areas, but just your exposure overall into these type of applications and if you see this segment maybe more of a foundry segment, dealing with these large die, very advanced chips or maybe fab-lite process, but how you see the opportunity for high performance computing? And also as you can say both out of TSMC and Intel, these devices get split up into chiplets, the opportunity for back-end relative to what -- what kind of sits at a foundry or IDM for -- in this market segment?

Yes. I think in general, let's step back and in general look at that market segment and the opportunities for OSAT. I mean, if we take let's say, several years ago, it was pretty much a vertically integrated market where a few companies dominated that market. They did internal manufacturing both on the wafer side as well as on the assembly side. We now see new entrants coming in, and that gives us more opportunities than they were before in this market. If you look specifically to the technology evolution towards, let's say, chiplet technology where you break up, let's say, large SoCs in multiple dies and you bring them together, either through silicon bridges or on a substrate by reintegration. If we bring them together with technology that is advertised by Intel or by TSMC, I would consider that clearly as a post wafer fab technology, which is not typically an OSAT domain to access. However, the moment that it's either reintegrated on, let's say, high-density board or through an interposer or you take these reintegrated parts with chiplets and you put them on, let's say, a large body flip chip BGA, then you come into the OSAT domain. I mean, we hold long-term, let's say, relationship, both with the companies that you mentioned. And we are not looking for a competition here but we're clearly looking for a partnership model to grow our business in that domain. Overall, the high-performance computing market, as I said, it's deverticalizing and that offers opportunities, and I strongly believe that through partnerships with these companies, we can grow our business.

Okay. Yes. Does the Intel opportunity change if -- is it contingent to on the decision they make as far as foundry? Does that maybe swing the decision on OSAT if they move in the direction towards more outsourcing to foundry?

It's difficult to say. I cannot comment on behalf of Intel. Of course, Intel is a significant customer for Amkor. They already outsource significant volume in the industry, both the foundry side as well as the OSAT side. Specifically for the high-performance computing and their strategies for, let's say, their specific silicon bridge technology or chiplet technology, I cannot comment to that. Will the dynamics change? I mean, they've built up so much expertise internally. I don't know. Difficult to say Randy.

Okay. Fair enough. Makes sense. I want to turn to System-in-Package. Last year, last 18 months has seen very strong growth for Amkor. Could you talk how you've been able to come into that market? Like what your approach has been versus some of the Asian OSATs that have been in there for a few years, where you've been picking up new opportunity in SiP? And how do you see kind of next couple of years playing out for your SiP business? Where you're seeing additional pipeline of opportunities?

Yes. Good question. I mean, we started entering the SiP business in 2014. Basically entering that in the RF domain, where in the mobile market, specifically, there was integration of functionality, and we are as part of the phone into SiP going from 3G into 4G and now into 5G. So the component count in this RF front end increased and the integration platform became very much a System-in-Package platform. So that's where we built our initial share. And actually, we hold a very strong market share still, now also moving into 5G, where that integration level goes up. I mean to give you an example, in 3G, we had about 15 to 20 components per SiP, currently goes up well above 150 components per SiP. So that's an initial market. But now it proliferates in other parts of both the mobile phone market, the smartphone market as well as other domains like the wearable markets, in the consumer space, consumer wearables, but also in the automotive domain, where we see a higher integration level and SiP platform happening. So the good thing is that we use the same manufacturing platform in order to serve also these markets, so our goal was very much to diversify from our initial footprint into multiple markets and into a broader customer base. And I have to say, and I have to admit that we are fairly successful in doing that. But also in our industrial strategy to assure that we use one uniform industrial platform where we can basically remind the modifications served for these individual product portfolios. And that was our strategy started in 2014. And I think we still believe that going forward, SiP, as a product format as an integration platform for functionality will have a lot of future.

Okay. And maybe from now to look at the scale of the business, like after being in at 6 years, how large is that SiP business within the total? What type of growth are you seeing in it? And maybe also help go through the margins because I think initially, it showed up as gross margin coming off a bit when it came up, but it's also a different return profile. So if you could talk maybe a little more about the revenue and the margin profile and if it's growing, how it may affect the overall corporate margin?

Why don't I leave this over to Megan to comment. Megan?

Sure. Happy to do that. So just to kind of level set, we were about $1 billion business in advanced SiP in 2019. Significant growth, we're looking to probably double that in 2020. And that, of course, is for some primary programs that we won and ramped in 2020. From a growth profile future, we still see that as a very strong growth with the way that technology is proliferating into the other markets. From a gross margin perspective, we have educated the advanced SiP with how many components, as Giel mentioned, does have high material content. And so when you have that high material content, it does put a strain on the gross margin percentage. But that being said, with the capital efficiency and the ability to use that CapEx, we are seeing really good drop-through to operating income as well as the bottom line represented by our last 3 quarters, which were peak advanced SiP quarters. So from a margin perspective, we just have to be mindful that it's not just a gross margin percentage story. There's good cash flow and good drop-through to the bottom line.

Okay. And Megan, I'll maybe follow-up, a couple of questions for you. But a bigger picture, gross margin. If your view now, I mean, a couple of years ago, I think peak leverage would get to a 20% gross margin. Factoring the size and how much this has grown, is it maybe a different -- how should we think of kind of the peak season potential for gross margin operating margin? If you have a kind of a target next couple of years or framework on that?

Sure. So let's look at what happened in Q3 as a good starting point in our guide for Q4, right? So our guidance for Q4 and the midpoint is about 18.5%. And there's still some dynamics there with respect to utilization. Our mainstream business and our automotive business is still much lower than it was, say, more than a year ago. So when you think about our incremental margin model, we expect 40% to 50% flow through. So with the return and the increase in utilization, that would definitely have some uplift to gross margin percentage. The other factor is we're really focused on our footprint in Japan. So we're reducing our footprint there. That's going to have a good impact to reducing fixed costs, and that will also have an uplift to gross margin going forward.

Okay. That's good. The other piece, I mean, it's tied to that. OpEx, I feel you could get some leverage back. In the fourth quarter, it's back up to $120 million with some of the Japan restructuring. But maybe at what level? Like how much could come back out of OpEx? And if I carry it forward, the other place, I guess I see as potentially the interest expense. Like -- and maybe other parts of the P&L, you could actually get some additional leverage?

Yes, absolutely. It's a great observation, Randy. So in 2020, we'll have recognized around $20 million in restructuring charges in OpEx. Going forward, that will be much smaller. We'll have a little bit in 2021. So from an OpEx perspective, I would assume that would go back down to on average of nearly $10 million a quarter. And that's fairly constant. So we do get really good leverage and drop-through on that. From an interest expense, we've done a lot of really good work, lowering our cost of debt over the years. We're going to have about a $7 million decrease in interest expense this year. And given the actions we've taken near the end of this year, we expect to see continued decline in interest expense going forward.

Okay. The one other piece, you've been generating a lot of -- a fair amount more free cash flow. CapEx, though, came back up. And on the last conference, you talked about even a swing that a bit more CapEx could even pull in. As we get to like a tighter -- I know mainstream, you have capacity, but say, for advanced packaging, wafer-level where it's a bit tighter. Do you see kind of a ramp back up in the CapEx to address some of the opportunities on more of the advanced side?

Sure. So from a CapEx perspective, we're operating in the low teens capital intensity. And we did indicate on our last earnings call, there could be an increase in our target CapEx for 2020. Given where we are now, we're not going to have to increase CapEx. We're going to be able to meet the time line of all those programs that are occurring in the first half of 2021. So we're maintaining our $550 million CapEx target. And even going forward, looking at the new programs and the growth that we're expecting, we still see the low mid-teens capital intensity being sufficient to allow that growth. I mean, looking at our over 20% growth in 2020 on that capital intensity, we expect we can maintain growth at that same level going forward.

Okay. I mean in factoring that it was, I think, a huge milestone, your first dividend payout. So that was a change. But you are still carrying and think about $400 million net debt which is down -- definitely down from where it was. Are you now at the point you'll keep relatively similar? I think you're near your debt-to-equity targets. So now the approach to start moving toward shareholder returns or still some balance that you'll do both. Maybe just how we should think about it if you continue to generate a couple of hundred million free cash flow each year?

Sure. Yes. So we're really excited we were able to initiate that dividend. We're really confident in our long-term outlook to do that. So from a cash flow and an allocation of capital, obviously, investing in the business is the primary -- the priority for that. We did hit our leverage target here in Q3. That being said, we're carrying quite a bit of cash and liquidity. So we'll continue to optimize what we may have to see if there's ways to continue to lower that cost of debt. And then we think the dividend is a great addition that provides shareholder return.

Okay. Good. Yes. I'll turn it back to Giel. Just -- we just have a couple of minutes left, but it would be helpful kind of to wrap it up. We started with a summary of -- that your end market positioning. But if we look forward into 2021, just what you see as some of the key growth drivers for the business? And then maybe just broader sector, but then within Amkor, where you might see some opportunities, whether it's project or market share. But how you think 2021 looks like it's going to be playing out at this stage?

Yes. Let's come to the markets, Randy. I mean, on the automotive market, we already cover that to some extent. We expect the continual recovery in the automotive markets. We saw already a good recovery, and we mentioned that in our last earnings call for our advanced products in automotive, and that will continue. Definitely, that will be a positive for the year. If we take the communication markets, we expect further growth in 5G. This year, about 220 million phones next year, probably closer to 500 million. So semiconductor content will be more significant. We're well positioned. And there are a couple of tailwinds also. One is the Huawei HiSilicon story where we traditionally didn't really participate to a large extent. So assuming that that's a zero-based gain that if these phones would be picked up by alternative supplies that would give us some tailwinds. And then the overall smartphone market is expected this year to decline, let's say, low teens, that would come back to some extent next year. So the overall volume would also help us a little bit. So we expect a strong communication market next year also. High-performance computing and consumer, it is also expected to see continued strength. So overall, we're looking at a solid year next year. The exact -- the question is, how would the second half look versus the first half, that's difficult to predict.

Yes. No, that makes sense. Okay. I think with that, we're about running out of time. So I want to thank Giel, also Megan for joining and everyone in the audience also for joining us today. And thanks a lot. I think at this time, we can end this session. Thank you.

Thanks.

Thanks, Randy.