Okay, thank you everyone for coming out. So I'm Randy Abrams, based out in Taiwan, and we've actually covered Amkor for a number of years. They're a leading company in package and test, and now on the UBS platform, we also recently picked up coverage. They're a company at the central place of a lot of the key end markets in tech, have always had a good exposure to automotive, smartphone, and this year has also been involved quite a bit in some of the high-end AI systems as well. With us, we're pleased to have the Amkor management team, Giel Rutten, CEO for Amkor. He'll start with a quick presentation, about 5 to 10 minutes, and then we'll go to Fireside.
If you have questions, send it here, and then I'll try to ask your questions as well. With that, I'll turn it to Giel.
Okay. Randy, thanks for the introduction. My name is Giel Rutten. I'm the CEO of Amkor. I will do only a few slides for the introduction of the company, so I will leave some space for questions later on. Now Amkor is part of the semiconductor supply chain already for a long time. I think Amkor was founded 55 years ago, so we are actually were pioneering the business model of outsourced semiconductor manufacturing. Now, over that period, we built strong relationship with all key players in the industry, but on top of that, we also built a very strong manufacturing base globally, where, you know, we were able to work with lead customers to bring new technology to the market, but also to ramp it to high volume manufacturing.
We are the only at-scale OSAT headquartered in the U.S., and given the current geopolitical tensions, I think that gives us some advantage. Now, we invest about, you know, in 2020 close to $1 billion, $900 million, in both capability as well as in manufacturing scale. We have, you know, we are considered the technology leader in our industry. We have a team of about 700 people in Korea, mostly in Korea, for R&D on new package developments. And overall, our manufacturing staff is about 30,000 people around the world. We have a couple of leadership positions, certainly in automotive, automotive semiconductors, where we have a presence in the key automotive areas, both in Japan as well as in Europe, as well as critical customers in the U.S. we serve with our services.
With automotive electronics being expanding, that gives us a good leverage there. But besides the automotive position, we also hold strong positions in the smartphone business, as well as in high-performance computing and consumer segments. We are centered around advanced packaging. That's our strategy going forward, and I will show you a few slides later on. Advanced packaging is the core of our business. 75% in 2022 of our business was in advanced packaging. We're part of an integral semiconductor supply chain, where we are closely operating with semiconductor foundries, but also we work with internal factories of our customers, where we receive these wafers, and we assemble and test them.
You know, the next step in that supply chain is we send package products out to contract manufacturers, and there are a couple of important trends in this supply chain. I already mentioned the geopolitical shifts in these supply chain, where we see that core, locations of, manufacturing are moving away from the traditional locations. So that's the reason that, for example, we built the last manufacturing location in Vietnam, that we opened recently. But also we have a good presence in Europe, in our Portugal facility that caters for the European automotive industry. Another trend is that, a larger percentage of, the manufacturing is being outsourced. So, you know, if we would go back, let's say, 10, 15 years ago, there was a significant percentage that was internally manufactured by what we call IDMs.
But if things now move into more advanced technology, we see that the outsourced percentage is increasing. Now, in that environment, Amkor's strategy is built on three pillars. Though the first one is our technology leadership, where and I already mentioned that 75% of our revenue is in advanced packaging. Customers knock on Amkor's door generally because they are looking for a technology partner, a company that has the capability to work in an early stage of product development with customers to bring new technology to the market, but also to ramp it at scale very rapidly. So the technology availability in combination with operational excellence is one of the core strength of Amkor and is also a critical differentiator. Geographic footprint is the second one.
I think we are present in all the core supply chains for semiconductors, and with the shifts that are occurring there, that work towards our advantage. I think we are present in Japan. We're working with all the critical automotive suppliers there. We're present in Europe, and we're currently, and that's announced in our latest earnings call, we applied for. Let's say, we did a pre-application for CHIPS funding in the US, and we believe that that offers a good opportunity for Amkor going forward. From a market perspective, you know, we believe that the semiconductor industry is built and is growing along, let's say, four core secular growth drivers. In these growth drivers, we are focused on the leaders that drive innovation.
We engage with them, we bring new technology to the market, and while we have that technology available, we ramp with subsequent parties in that market. We do that for the communication market in 5G or in RF application, in automotive, with functionality like ADAS and infotainment, but also in computing, 2.5D, high-performance computing, but also in IoT devices. Now, our performance over the last year, we had strong performance on the financial side. We grew significantly at an average CAGR over the last four years of 13%. This is 2022. This year, of course, we're in a cycle year. We're a little bit slower, but on average, we are growing faster than the market. Our profitability and our cash flow is strong.
I think we hold a very strong balance sheet, and our earnings, per share, actually grew sixfold over the last couple of years, so our investors, like that. Now, overall, in summary, and then we can go to questions, I think our investors- investment appeal is, first of all, Advanced Packaging. That's the foundation of all the core drivers in the semiconductor industry. Geographically diverse position, I mentioned it before, and then we are very much leveraged, to industry mega trends and have strong working relationship with the leading customers in these individual industries. And that's the introduction-
Okay.
Randy.
Yeah, let's. We'll start actually on near term. Actually, start with third quarter, because there were a lot of companies actually already seeing a bit of slowdown in areas like auto industrial. Your business still grew 25% sequential. So, could you recap that, the strength you were seeing for third quarter, just by the key when you went through applications like communications, auto industrial, high-performance computing?
Yeah. Yeah, I think we had a very strong third quarter. We grew sequentially, indeed, well over 20%. The drivers there are related to our position in premium-tier smartphones. I think we hold a very good position in the premium-tier market, and the launch of new phone models drove our growth in the third quarter. You know, we work with the leading players there, and we bring, I think, innovative solutions, both on the RF side, for example, but also in display, in cameras, et cetera, and that drives our growth. Now, if you take a year-on-year view, we even grew year on year in a declining market.
If you take the smartphone market, volume-wise, it declined over the last two years, about 15%, but we were still able to grow on a year-on-year basis with 3%. So that's underlining our strong position, but also our market share growth in some of these segments.
Okay.
Now, automotive, it was sort of flat, and we see a slight decline towards the end of the year, and that's inventory correction. I wouldn't see that. I see this more as, you know, cleaning up excess inventory coming out of the COVID years, where the automotive industry built up some inventory, and we now see with a more balanced supply chain that they build that down. I wouldn't. I don't see, we don't see any structural corrections. I think we still believe the automotive industry, midterm and long term, is a very strong contender. High-performance computing, a little bit mixed. I think we're very strong on the high-performance computing side with our 2.5D solutions.
You know, we are the only OSAT currently that drives a full turnkey 2.5D solutions with leaders in the industry, and that gives us upside. I think we see the PC side in that market still correcting, not fully in balance yet, but overall, longer term, also in this market, we believe that continues to grow.
Yeah, I mean, if we take the second part, how do you see as we go to year-end and initial cut on first quarter? Like, you normally have seasonality, but how do you see the correction playing out? Could we have to go through a couple quarters below seasonal, and then we clear the decks, and we're actually into a good recovery as we move through 2024?
Yeah. I mean, if we talk to our customers, we see that customers are still very conservative.
Mm
... to guide into next year, also to put forecasts out or to have orders in the books for next year. I would say there's a strong-
Mm
psychological effect, because, you know, generally, whether it's, it's the fourth quarter or early next year, I think we expect that we reach the trough, and there should be a year-on-year recovery. As I said, you know, there, there is conservatism, on the, on the customer side. On the other hand, on the wafer foundry side, we see good optimism going into next year.
Mm.
So in that, in that overall view, my personal view is that there will be a recovery next year.
Okay.
It will be, although small, but a year-on-year growth.
Okay. I'm curious, the disconnect. What do you think is driving foundry, like, a bit more optimism? When you speak to the foundry partners, what's driving their optimism? And then from the customer conservatism, is it worry about inventory level, demand level, macro? Like, what are kind of their concerns to keep conservative?
... Yeah, I think on the customer side, the conservative position is still related to inventory. If you talk-
Yeah
For example, the smartphone markets, they are conservative in putting new orders in, in the industry. If you take the foundry optimism, the mix of markets in the foundry is different than, for example, in our business. They are very much exposed to the compute market, and specifically the data center and the AI market, and that gives them an uplift, which is giving us an uplift also, but they are exponentially bigger-
Yeah
When it comes to that extent.
On that, you mentioned earlier, okay, when AI took off, I'd say late spring, it felt like you were seeing that opportunity. Some of the OSATs we cover in Taiwan have been a little more conservative going after that. What's been Amkor's approach? Like, why have you kind of been aggressive on the front foot, seeing big opportunity to expand? And could you talk about maybe your technology there, capability, and what you're seeing for growth opportunities from, like, AI, high-performance computing?
Okay. Yeah, good point. I think we were indeed early in building up that capability, but also expanding capacity. We started our first engagements with lead customers there in 2017. So currently, we're in the third generation of technology development. And besides the technologies, you also need to have the manufacturing capability to do that. So we went through three cycles of products to build up that capability, and we now see the market taking off, and that gives us that head start. And we believe that mid-term, as well as longer term, this will be a significant growth opportunity for us. I mean, with AI functionality specifically entering into the data centers, with a huge install base on data centers, I think we believe that over the next 3-5 years, this is an upside.
You know, barrier of entry is very high. On the foundry side, this business is supported, but it's not part of their core business. So we believe that this is giving us a good opportunity, Randy.
Yeah. And can you recap, just to explain for a lot of investors, some are probably newer, just exactly the role. TSMC has CoWoS. They usually talked about vertical integration, but where do you come in as a... Like, in package and test, what are the opportunities you can come in on?
Yeah. So I think, you know, let's take the example of high-end GPUs, graphic processors, that are being deployed for AI functionality in data centers. Generally, these GPUs are packaged together with high bandwidth memory. And the package technology that you use in order to provide an interconnect structure between the memory and the SoC, we call that 2.5D. Of course, you know, you package them on a certain substrate-
Yeah
That's called an interposer, and then that will go into an overall package. I think that technology initially was launched by TSMC. I would say, you know, in their standard, you know, you could call it the waterfall model, where they enable customers to utilize their advanced silicon, and later on, when the technology matures, it's being proliferated further into the market. So TSMC has that capability. We started first in 2017 to further develop that with the lead customer there, and I think we now have both that capability. Of course, we also cater for a supply chain where customers can get and procure their interposers from third parties, while TSMC only caters for a vertically integrated supply chain.
We offer more flexibility, and I believe that when this market is, let's say, deverticalizing, that positions us well to capture some of that growth.
Yeah. Two parts to that. I mean, one is, do you have any concern TSMC... They were a bit on the back foot, or everyone was, like NVIDIA, about how fast this took off. Is there any concern they finally put in a lot of capacity in place? Like, how's the comfort as far as your orders, for the business that's not going through TSMC to continue to grow? And second part, I think the investor concern is, it's so strong in 2024, there's a fear it's a mini bubble that will come off a high base and have a correction. Do you have any, excuse me, initial indications, sustainability, if you see-
Yeah
further growth in the 25 continuing?
Well, well, long term, we believe that this is a strong market and so strong technology domain. Mid-term, you know, yes, I think there can be some cyclicality to that. You know, two things, as a comment for our investment and return on these investments. For Amkor, the investments are highly fungible. That means we buy equipment for this technology domain, but we can also use that same equipment for other areas, like wafer-level fan-out, like bumping, et cetera. So I think underutilization is less of a risk for us. For TSMC, it's more of a risk, but okay, let's put it aside. You know, if we look to how the supply chain is developing, then we see multiple alternatives being developed-
Right
for specifically the interposer supply chain. And ultimately, the end customer decides where to get their packaging technology or their packaging manufactured. There are advantages, there are advantages to do it vertically integrated with TSMC, but there are also advantages to do it in a different supply chain model. In that sense, we see TSMC more as a complementary partner for us than a strong head-on competitor. We have advantage ultimately, we know how to cost down, and bring this to a cost level that it can be broadly proliferated in the market.
Actually, just one final question here. It's mostly a, excuse me, it's mostly a silicon interposer now, but the market's been talking about a shift. It's an expensive piece of silicon, and they want to move to redistribution interposer, where you don't actually need the piece of silicon.
Yep.
Like, when do you see that trend? And if that takes off, would it be less captive to foundry? It would be more opportunity for OSAT to play in that space.
Yep. Yeah, we believe that that's the next step in the technology roadmap, that the interconnect structure changes from an interposer to an RDL-
Yeah
- structure. That capability is in place. The equipment is similar, to a large extent, than the equipment that we use for the 2.5D. You know, actually, all customers are evaluating that at this point in time, and we believe that it will be ready for ramp up probably in 1-2 years from now.
1-2 years. Okay.
Yeah.
Okay. One other area getting more attention here, so a lot's been about the cloud AI, which you're participating now with CoWoS. Edge AI, how do you see a opportunity on edge devices, smartphone, PC? And there's a related area to ARM. CPU was always an x86 market, but now you have Qualcomm with an ARM CPU, you've got others trying to come in. How do you see the both edge AI, but also these alternative compute, alternative processors driving the opportunity for your business?
Yeah. Well, for Amkor, we see the computing domain as a significant growth area, also beyond the AI functionality. I think there are a couple of important trends. One, you already mentioned, the trend from, you know, of a shift of x86 architecture into ARM architecture, and not only for the consumer devices, but also for the computing devices and data centers. You know, drivers are there, for example, power consumption, and we see that being deployed already quite broadly in the data center domain. In the PC domain and in the handheld devices domain, you know, the smartphones are completely ARM-based, I think, to a large extent. In the PC domain, you know, there were the first pioneers that moved into an ARM-based architecture. They now, you know, have a full range of products.
We see some of the China players also moving into ARM-based architectures, and we see that, you know, Qualcomm publicly announced that also with the Nuvia core. Now, I believe that there is a structural advantage in that architecture that relates to power consumption, but also computing performance. You know, they will coexist next to each other, but I believe there is a significant opportunity for that architecture to gain share, and that's good for Amkor.
Yeah.
Basically, it's good for Amkor because, you know, we go from a vertically integrated supply chain to a horizontal supply chain, where there's an outsourced manufacturing base to companies like TSMC and to companies like Amkor.
Mm-hmm. If we take the smartphone market, there's maybe two questions. One, the short term. Qualcomm, last half hour, talked about stabilizing market, inventory finally getting back to normal. How are you seeing it? Are you seeing signs of life? I think you talked about there's a little bit of seasonality coming off a high third quarter. But do you see more optimism that there's a bit of a turn? That's kind of a short term, and then midterm, because it's a key part of the business, do you still see growth drivers for that, or it's going to come more from these other applications, auto, PC-
Yeah
... compute?
Yeah, let me try to answer your first question. There is indeed optimism on the smartphone market. The optimism is very much, let's say, triggered by a shift to high-end devices and smartphone market. So currently, we see that, you know, and that is triggered by, for example, the launch of the Huawei P30 phone. That's clearly positioned as a high-end phone. The new smartphones launched by Apple and the iOS ecosystem. You know, there is a high interest in high-end, high-functionality smartphones. Customers seemingly, and that's the feedback that we get from customers, are willing to spend a significant amount of money on high-end phones, and that favors that segment-
Right
... of the market. That's also the segment where the innovation takes place, and that's also the segment where we hold a good position. We are quite optimistic, and we also see that now back in, you know, not only in the iOS side but also in the Android side-
Yeah
with an increasing order portfolio for these high-end devices.
Okay, that's good. The adjacent market, you've had a lot of SiP business. Some of that goes into smartphone, also goes into consumer products. Could you remind everyone what the driver has been, like, SiP, where you see opportunities both in that smartphone platform and the consumer platform, where you've had good position? Where do you see the opportunities for that business over there?
Let's start with the consumer platform. We indeed have a proven platform where we integrate, let's say, the majority of the electronic functionality in a very small form factor... and that can be customized to the form factor, to the electronic device. So that we deploy mostly in wearable devices, whether it's audio, whether it's, you know, smartwatches, et cetera. You know, there's a weakness in that market, you know, based on an economic downturn, but we expect that market will pick up. We have a fairly strong pipeline there. So going into next year, that market, I believe, will recover. In the smartphone part, you know, that platform of System-i n- package integration is used in multiple parts of the smartphone.
I mean, whether it's display and display components, whether it's connectivity and connectivity components, the RF part of the phone, and more and more also in the optical part of the phone, where, you know, camera modules and all the integrated mechanical things that relates to autofocus are actually packaged into a system-i n- package. So these three domain, whether it's RF, whether it's the camera, whether it's the display, whether it's the connectivity in the phone, are all moving from single package components-
Yeah.
into System-i n-P ackage, and that's a big, that's a big opportunity for us.
Yeah. It's a segue just for your factory network. When you gave the introductory talk, it was more geographic. I mean, there's kind of the focus now on geographic diversification. Talk at a high level, how you see yourself positioned, any gaps in the portfolio, like, where Greater China, like, how you're positioned in that area, but then also the opportunities, with your footprint elsewhere and also with the new Vietnam facility.
Yeah. I think we see a couple of trends in that supply chain. I think even independent of the geopolitical situation, we see that the experience during the COVID, let's say, period woke up companies that they had to revisit their supply chain strategies-
Right.
-specific in industries like the automotive industry. To give you a few examples there, I think there's, I would say, a revitalization in Japan-
Yeah
... to bring back semiconductor manufacturing into Japan. We hold a good position in Japan. We have five factories in Japan, and we see also there that that will continue. I think their front-end manufacturing is being built, TSMC is investing there, and I think with our presence there-
Mm
... and a strong team in Japan, we see that as an advantage. Similar trend in Europe, I think, initiated by the European automotive industry-
Mm
... supported by actually governments in Europe and the EU, we see a reshoring of semiconductor manufacturing there.
Yeah.
Again, TSMC is also reaching-
Yeah
... in that domain, but. And this was a public announcement: We're teaming up with GlobalFoundries in Dresden and our Portugal factory to offer a seamless supply chain in Europe. We see huge interest there. I think we're ramping up our factory in Portugal for multiple applications. So we see that as a positive trend.
Yeah.
Vietnam is a bit separate. I think-
Mm
... Vietnam, you need to position in the grander, let's say, Asia supply chain shifts, where the bigger companies that have strong presence in semiconductor manufacturing in China are looking for balancing their risk. And that doesn't mean that everything is massively out, moving out-
Right
... but it's more, balancing between different countries. You see a trend into India, and you see a trend into Vietnam. And, you know, our Vietnam investment was also triggered by some of the customers that actually motivated us to give them an alternative. And so we believe that, that although Vietnam will be the smaller player, it will still be a very strong presence for Amkor.
The two last geographies. So China, you have quite a good facility in Shanghai. How is that factory positioned, both international business in China and then also for domestic customers? So what's the main focus? Also, with a lot of back end, a lot of capacity investment in China-
Yeah
... how do you see that operation positioned?
Yeah. Our China factory was, you know, an excellent performing factory and is an excellent performing factory for Amkor. Of course, also they saw an underutilization during this-
Yeah
... this year downturn. You know, in that factory, we roughly balance our loading, 65%, you know, non-China customers, 35% China customers. And we now are working to shift more to a local for local, manufacturing support out of that facility in China, and that means that, you know, we already had strong relations with local customers, but we're going to intensify that. We still believe that since the, let's say, some of the customers like the performance of the factory, for non-critical components, China is not out of the question, even for non-China customers.
Yeah, makes sense. Actually, I'll ask one more. We're running close to time. But U.S., could you talk a bit more about the opportunity there? Because you mentioned about the CHIPS funding. What kind of scale do you see, or what type of operation are you planning for U.S.? And how's the economics running a package and test from U.S.? I think the concern has been cost, geography, labor force. So how do you see that opportunity playing out?
Yeah. In putting together a plan for a U.S. facility-
Mm
... there are a couple of elements. One is the choice of technology. I think we're not going to put a wire bond factory in the U.S.
Yeah.
I think that's clear. It would be advanced packaging, supporting certain applications in the U.S. I think that's related to advanced computing, high-performance computing-
Mm-hmm
... automotive, and some of the mobile devices. I think that would be the technology scope. If you would say, what, what makes sense in the US? Well, it would be a highly automated factory.
Yeah.
That would make sense in the US, where we would basically use our learnings in our Korea factories. You know, then when it comes to business models and cooperation models, we need a stronger engagement with lead customers in the US because, you know, you need to basically-
Right
... see a path to a fully utilized manufacturing facility. That's one part. That's on the customer side, and then on the governance side, I think you need to have support, financial support to prime the pump-
Mm
... in the US. I would say on all these ingredients, we're working. You know, we have not finalized our plans.
Yeah.
But, you know, if you would compare the situation now as it was, compared with last year ago, a year ago, I would say we're much more optimistic now that this is feasible-
Mm
... that it can be executed, and that customers are willing to see the value of U.S. manufacturing. If there's value, then we can also build an economic model around that.
Yeah. It'll be here, in this area?
I can't, I cannot comment on that.
Okay. Yeah. Okay. Okay, good. But with that, we'll have to wrap up. I want to thank everyone for coming out to the conference and also to Amkor. Also, thank you and the management team from Amkor as well.
Okay.
Okay. Yeah, thank you.
Thanks.