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

Welcome back, everybody. I'm Joe Moore. Very happy to have with us today Giel Rutten, the CEO of Amkor. Giel, thanks for being here again. I think this is your second hear at least at this conference, and so I appreciate you coming. Can you just talk in general about the company's priorities? The last couple of years, you've seen a pretty big transformation around some of the new packaging technologies. Can you just talk about where your focus here on R&D, where you're focused strategically for Amkor?

Thanks, Joe. Well, let me start for the people in the room. I mean Amkor is a manufacturing service company. We offer assembly and test services for the semiconductor industry. We're one of the leaders in the industry. We have a strong global footprint. And our strategy is based on four pillars: basically, technology leadership; secondly, geographic diversities; and third is a strong focus on the semiconductor industry distinct growth drivers. So with respect to priorities for the company and R&D priorities, the way that we work is we engage with lead customers in the industry, and we work on innovative new products in the individual market segment. So in communication, Joe, our priorities are centered around introducing new products when it comes to the next generation of AI-enabled smartphones, that means a higher memory density, higher interconnect density at the apps processors, and that means next-generation package technology. Introducing these technologies take quite a bit of effort and time, but we expect that the next wave of smartphones will be AI-enabled smartphones. Now that's one specific area. And of course, the lead customers in that area are not specifically the lead customers in other areas. If we take the compute segment, Joe, we work closely with the leaders in the compute area. AI is important there. We hold a strong position with leaders in GPUs for AI -- enabling AI in the data centers. So that's a strong business for us. We invest in that. And we expect that innovation is driven very much around that team. 2.5D -- or TSMC introducers CoWoS. 2.5D for Amkor is a strong innovation area. We are ramping up. I think we tripled our capacity last year. We're introducing the next-generation interposer technologies into the market this year. And so we expect that, that road map will drive significant innovation, but also significant potential business for a company like Amkor. So that's a second important area, compute. The third one is automotive. The automotive industry specifically here in Europe is a little bit going through a difficult time at the moment. We also see that from the business volume. But we believe that innovation in the automotive industry will continue to be executed. For Amkor, there are 2 areas that are important in automotive. One is power management and specifically power modules. We had an announcement on the cooperation in our Portugal factory with the leader in power. That's a semiconductor company here in Europe. And we are also investing significantly in next-generation ADAS processors in the car, and we see more and more advanced technology entering into the car at this point in time. And last but not least, Amkor is focusing very much on wearable electronics. The drivers there are very much miniaturization, and we have a whole portfolio of system and package where more and more electronics are going into a smaller footprint and wearable electronics. That can be audio devices, it can be watches, it can be medical devices. But we expect that to continue and to drive further growth. So these are the four areas. In each of these areas, the company sets priorities for innovation. The critical thing for Amkor is, we work with the leaders in a co-development stage, bringing technology to the market.

Great. Well, maybe we could walk through each of those kind of priorities. Maybe starting with the overview comment that you made about geographic diversification. You talked about, I guess, a little more than a year ago, your facility in Arizona. And you usually think of the U.S. as being a place where there's a lot of semiconductor packaging. But can you talk about the Arizona initiative, what that means in the broader context of geopolitics, the relationship with TSMC, things like that?

Yes. I mean for Amkor, venturing out with the manufacturing facility in the U.S. is, of course, a big commitment. On the other hand, we get significant support in the U.S. from the U.S. government. And the way that we strategize that facility is that we work closely with critical partners in that supply chain. I mean you mentioned TSMC is one, and we recently announced cooperation with TSMC on technology to assure that we have a seamless manufacturing footprint in Asia as well as in the U.S., and that makes it easier for key customers being served by TSMC and by Amkor to manufacture in the U.S. That's important. We also work with lead customers in the U.S. to define the road map, what makes sense to do locally and what makes sense to keep in Asia because we have to keep in mind, I mean, not the whole supply chain will move into the U.S. It will be very distinct areas in data centers, in specific devices, in military applications that will move into the U.S. So that's the way we configure it. We get good support. We work with other people in the supply chain and there are lead customers that are committed to manufacture in the U.S., Joe.

And you mentioned the CHIPS Act and the U.S. commitment to more domestic manufacturing. It seems like there is an awareness of the back end of the packaging and test business from the government. How much of a priority do you think that is? And do you think that persists into the new administration?

That's a good point. I mean, back end or packaging technologies became a higher priority in the U.S. over time because there are two main factors here. One is advanced packaging is becoming a more integral part of semiconductor development because with a transition into chiplets, the transition into advanced packaging like CoWoS like technologies we see that advanced packaging is very important. So manufacturing wafers in the U.S. and then send these wafers to Asia for packaging is not a feasible supply chain model. And the government realized that in the course of this process also based on feedback from customers, and that initiated significant interest in order to support an advanced packaging facility in U.S. and that led to our engagement with the CHIPS Act and also successfully getting engaged there.

Great. And then you mentioned the compute market and the sort of CoWoS like technology. It's really interesting, I mean the degree to which AI is pushing the envelope of advanced packaging and maybe pushing you guys into areas that -- where the process looks a little bit more like a wafer process to stack these chips. Can you talk about your 2.5D process, where you are with that and the scale of your ambitions in that area?

Yes. I mean currently, our 2.5D process, we manufacture that in Korea in our K5 factory, that's our most advanced facility. It will also go into our U.S. facility going forward. We significantly invest and invested in capacity in that technology. We tripled the capacity '23 going into '24, and we continue to expand that going into 2025. So the scale is significant, I would say, even compared to TSMC, and we'll further expand there because we believe that this technology domain besides GPUs also now starts covering other areas in the data center where, for example, switches are also moving from more traditional package technology into 2.5D technologies. So we believe that this diversification will drive further growth. It's not driving the most advanced 2.5D or CoWoS technologies, but definitely, there are multiple followup applications that will continue to drive volume in 2.5D. And that makes this investment feasible and it's a very solid investment for Amkor. There's another thing. I think you mentioned wafer level processing. Within Amkor, we have a longstanding history in wafer level processing. It started with bumping multiple years ago. But here in Europe, we acquired NANIUM, which is now part of our manufacturing footprint. They were experts in wafer-level fan-out technology. And we integrated that in our manufacturing footprint. So we have a longstanding history in wafer processing.

And as you think about that 2.5D packaging, people ask me your market share or try to get the scope of your -- of how big you are. Just maybe you could talk a little bit generally about how market share is driven there? And how is your business in partnership with certain customers? How should we think about the growth for Amkor in this market?

We believe that the market on AI devices and consequentially, the further proliferation and data center devices is only in the early stage. We now see that the lead customer there is moving up and driving the most advanced technologies. And we believe that, that technology stays exclusively at TSMC, where, let's say, the previous generations will proliferate into the OSAT domain. That's what we see, while other technology, other, let's say, devices are entering the same domain and drive volume. So that will be the model, and it has been the model working with TSMC in the past, where TSMC introduces first, a certain technology in order to enable customers with their silicon to go to market and whether that technology matures, it proliferates into the OSAT industry. And we expect that to happen also this time. For Amkor, from a market share perspective, we believe that ultimately, there are only 2 OSATs, 3 companies globally will drive this, and that will also give a reference of which market share we're aiming for.

Okay. Great. And then maybe if we could step back and just talk about advanced packaging as a concept. Can you give us a sense for how are you defining advanced packaging and give us a sense of the margin differential between the older legacy technologies and what you're doing?

Well, for Amkor, advanced packaging is defined basically as all the package technologies that are not wire bond or leadframe, so this is flip chip technologies. It's flip-chip BGA, large body size, flip chip, it's multichip modules, it is system in package and it's technologies like 2.5D and 3D technology. That's what we label under advanced packaging. It's slightly different than, for example, TSMC labels in a very narrow context. They look only at mostly the compute segment, but we take a little bit of a broader definition there. For Amkor advanced packaging in the definition that we use is currently around 75% of our total business. So we have a small footprint on the wire bond, and we maintain that mostly in automotive to be complementary to our customers. So advanced packaging is definitely the growth area going forward. Going back to your margin profile, I mean, margins profiles for our customers largely depends on elements like utilization of our lines in certain areas and material content of certain package technologies. So take SIPs, where we have a high material content. And in that sense, for that part of our product portfolio lower, let's say, margin percentage, that we would consider on a lower margin perspective. On the other hand, 2.5D, large body size flip chip, definitely has a higher margin, and that would go up to margins that you are seeing in the foundry domain, for example. So within Amkor, we have a broad range of products. But you have to keep in mind in a manufacturing service company, the most decisive element that defines our margin is factory utilization, how do we utilize our manufacturing. And unfortunately, the last 2 years, the average utilization is relatively low. We run our factories with, let's say, an average utilization of about 60%, 65%. If you go back to 2022, it was 85%. So with a higher utilization, of course, your margins significantly improve. That means that currently, our flow-through -- if we have incremental revenue, the flow-through from margin is around 40%. So there's a significant flow through if you have the capacity installed.

Great. And I know you mentioned lower utilization currently, but I was just in Taiwan last week and it's also clear that it's very dynamic. There's a lot of stuff that's being expedited, things that are accelerating markets that are slowing down, like how difficult is it to adjust to all of that? And do you have any bottlenecks right now even at that low level of utilization that you think about?

Yes. I mean in general, we can take upsides, currently, very, very rapidly. I think we have capacity available. As I said, our average utilization currently hovers around 65%, maybe 70%. However, there are a few distinct manufacturing technologies and factories that are fully utilized. We talked about 2.5D. Currently, for that level of packaging, we are currently sold out. We're ramping up a second customer now in volume, and we have multiple customers where we actually are qualifying or being qualified and ramping up in 2025. So that's definitely an area where we'll continue to invest and we run at a very high utilization. Other areas that relate to automotive, for example, let's say, automotive ADAS devices, with the companies that are leading there, we see that utilization is less than 60%, and the automotive market still struggles to get out of the current, I would say, inventory blip in the industry. Although as Amkor, we are the lead OSAT in automotive, and we believe that going forward, automotive will continue to be a driver for semiconductor industry. More content per car, we believe that, that will continue. Also the transition into EVs will continue, and then power supply is an important driver for semiconductors.

Great. So I want to look at the end markets a little bit in more detail. But first, maybe you could also talk about the importance of test, and test is an important part of your business. Can you talk about where -- how that fits into the overall picture?

Indeed, I mean, within test, in a company like Amkor, we offer a complete portfolio from test, and that means that if wafers enter into our factories, you do probe, you do final test, you do burn-in, you do system-level test, and that whole portfolio is important because the bigger customers are looking for a turnkey manufacturing service. So basically taking wafers from the foundry or their internal factories, and we're shipping end products in drop shipping the FPTA side. That's our business model. Test is becoming more important. I think we have significant test sites in Taiwan, in Korea. Currently, it's roughly around 15% of our business. The investment level is higher, but the margin profile is also significantly higher. And currently, we're looking on a global basis what's the best way to invest in test. We want to grow test either organically or potentially inorganically where possible, but we believe that test is becoming more important based on the demand from customers to have a full turnkey service. So basically, you need to manage end-to-end your services.

Okay. Great. So coming back to the end markets, you mentioned auto a little slow in the short term, but also long-term growth drivers. Is that just slowness now, it seems to me is just a sequencing. We saw weakness in consumer smartphone, it kind of went down, got better. Industrial come down, hopefully bottoming out now and automotive, the last market to really enter into a downturn. Is it just as simple as that? Is it that people had a different sequence of where they reduce inventory? Or is there something more significant than that, do you think?

Well, I mean, there are a couple of shifts in the automotive industry as we see it. And of course, we're not an automotive company with a manufacturing service company for semiconductors. But one thing is the demand crisis or the supply crisis after COVID where the traditional model in the automotive industry of just in time, the delivery didn't work out. There were shortages after COVID and that initiated a change in the supply chain where the automotive companies as well as the semiconductor companies were increasing their inventory. Currently, we see that they're switching back to the old inventory model, and that means that they are adjusting that inventory closer to a just-in-time inventory level. That's a significant change going forward. So although the end market, the number of cars being built on a global basis is not fundamentally reduced, the inventory is being corrected. Then there's, of course, the dynamic of slowdown in electrical vehicles. On average, electrical vehicles have more semiconductor content than combustion engine vehicles, and that is gradually or at least less ambitious growth, and that means that there is a slower down. And then there is the China automotive industry, which is growing rapidly, also exporting cars and they are more and more using local content. So these three elements with inventories, with electrical vehicles and the China supply chain is sort of slowing down the automotive industry at this point in time.

Can you talk about the China dynamic because it does seem like that's where a lot of the EV innovation is coming from and that's where a lot of the growth is currently. We saw TI talk about the China market growing, all the other markets are weak, things like that. How does that affect you guys? How are you able to tap into the China market? And does that displace traditional Tier 1s and customers you may have?

Well, I mean, for a U.S. company, operating in China is getting more difficult. I mean -- and that not only holds for a U.S. company, generally for international companies. We see that, by policy, the China industry is looking for local supply chain, local for local. And that is also the preference. That's preference on the manufacturing side, and that's the preference also on the component supply chain. So that definitely doesn't make it easier to do business in China. On the other hand, if you bring in differentiation on technology or manufacturing performance, then you can build a footprint in China. I think we have a significant manufacturing site in Shanghai. And they are operating at a very high level of operational excellence. We have a very broad technology portfolio there. And you see that, that attracts local customers to do manufacturing in that facility, but to ship products out of China, manufacture them Korea, then bring them back is more difficult. But since we have a local footprint, that makes it somewhat easier.

Okay. Great. And then shifting gears, maybe we can talk a little bit about smartphone. Feels like there's a lot of enthusiasm for what AI in the phone may bring from a volume perspective, accelerating upgrade. It also seems like that's maybe next year or the year beyond in terms of we're not seeing the strongest market currently. Can you just talk about the prospects for you guys in that market, both with your biggest customer and then with the breadth of customers?

Well, we believe that AI entry in the phone, so the AI functionality will drive the possibility of new applications in the phone. It will also drive an upgrade cycle on semiconductors, so it will drive innovation. It's important to see what the killer applications are that will drive that upgrade cycle. I mean, currently, there are a few applications that enter into the phone that will drive some upgrades, but it's not really a killer application that will come. So after talking to many people also in the industry, we believe that there will be an upgrade cycle coming. There will be applications in the phone that will utilize AI functionality. Will it be next year? We'll see. I think that the first phones are being launched now, and I think the first functionalities and the first applications are being launched also. It may go into the second half of next year or even 2026. Now for Amkor, we have a strong position. We work both on the Android ecosystem as well as on the iOS ecosystem, and we manufacture mostly components for premium tier phones. So we believe that this functionality on AI will enter not from the low-end phones and then up, but it will enter the industry through premium tier phones, not only on the phone side, but also on the PC side, but definitely premium tier phones. With our position in premium tier phones, we see that we believe that we can benefit there. We were able to grow significant market share over the last 4 years in premium tier smartphones, certainly, on the iOS side, but also on the Android side, we see good traction there. So this is an opportunity for us. It's a bit of a slower start than many of us had hoped, but it will come in the course of '25 or maybe early '26.

Great. And then I was going to ask about wearables, but let me just see first, if there's any questions from the audience. No, maybe you could talk a little bit. You've discussed the wearables market. It's pretty clear the value of advanced packaging when you get into the small form factors. Can you talk about your prospects in those markets?

The wearable market, I mean, for us, it's -- this year, we launched a significant new audio device with a high level of integration. And we see more and more functionality enter into a smaller form factor. This platform of heterogeneous integration where you combine multiple components and multiple functionality in a very small form factor, I think, is increasingly getting attention. We built over the last 5, 6 years, a very strong foothold there. And we started with the first wearable audio devices with advanced functionality, but now in the third or fourth generation, and we will -- we believe that, that will continue. It's a sizable business. It's going up to 15%, 20% of our total revenue.

Great. And then just if we could wrap up with just profitability. Your gross margins had been under some pressure, as you mentioned, around utilization. What's the path to get gross margins back to 20%? What has to happen, if you could tell?

I mean, for us, the main element to get our gross margins back to, let's say, the level that were in 2025 is to drive factory utilization. Of course, currently, we're working very much on managing our cost, and that brings you to some extent, an improvement. But ultimately, as a manufacturing company, we're sitting on a large fixed cost base. And if you utilize your fixed cost within a utilization rate of 60%, your margins are under pressure. So with the flow-through of 30% to 40%, if utilization goes back to the 80% level, we believe that our gross margins go back to the 2022 level, so that means 20% plus. But that's an important driver. The second driver is our SIP portfolio. It's growing faster than average and the profitability because of the material content is lower. So automotive -- generally our automotive business has a much better margin than our SIP business, with automotive being currently at the low side. We believe that with that -- if that comes back and once that comes back, that will help to drive margins up.

Great. Well, I think we're about at time. So we'll wrap it up there, Giel, thank you very much. Appreciate it.

Thank you.