Well, let's get started. Good morning. I'm Tristan Gerra, senior semiconductor analyst at Baird. I would like to introduce Diodes, a leading supplier of analog and power platform solutions serving a variety of end markets, including AI data center. We're pleased to have with us today Emily Yang, Senior Vice President, Worldwide Sales and Marketing. With that, let's get started. It's going to be a fire chat with questions only today. We have about 29 minutes. I'm going to start with my first question, which is about the trends in power. Clearly, things are changing fairly quickly these days. You've obviously had a significant share gain over the past several years. You're benefiting from an upcycle. It looks as if there are secular trends that are layering on top of that normal increase. I know you're going to provide more data around data center and the AI part of this, but there are some emerging exciting themes coming, with 800 volt native. Maybe at a high level before you get a chance to, later on this year, provide more metrics. Maybe at a high level, how do you see those trends in data center driving your business and any opportunities that you want to highlight? Oh, perfect. Well, first of all, good morning, everyone. Good to be here with you guys. AI, data center, power, conversion's definitely a very hot topic at this moment. I think I mentioned AI is a whole ecosystem. We did share some of the AI solutions, especially focused on the server applications. We've been growing really fast in this area and benefit a lot for the last few quarters. This is actually specific focusing on the PCI Express support, sending out additional ports on the timing solutions and stuff like that. On top of it, I think Tristan just actually mentioned 800 volt. This is a transition from 48 volt to a 800 volt application. A lot of, I would say, devices focus on, I call, the power supply area. This is actually the power supply supporting the AI applications, whether it's a data center, whether it's a server, whether different application. What we see, the power supply actually consists of different units. There's a power supply unit, we call the PSU. There is battery backup unit, BBU. There's actually power delivery unit, PDU, and there's also isolated, DC-DC conversion portion. I would say all in all, I don't think you guys need to remember all these acronyms. I think with this voltage translation create a lot of opportunity overall for Diodes. If we look at the PSU, which Tristan specifically referred to earlier as a good example. In the PSU area, we actually seeing a lot of silicon carbide requirement. This is actually not only just on the MOSFET area, we actually also have the diodes over this area as well. That's actually a great opportunity for us moving forward. On top of it, we actually seeing a lot of isolation solutions require. This actually covers our gate drivers, isolation gate drivers. We're seeing a lot of current sensing requirement. We also have something called digital isolation, overall, in the ORing. This is more on the redundancy of the power supply, also create a lot of opportunity. I would say all in all, this is quite significant, and like Tristan mentioned earlier, we are putting a lot of materials together that we actually going to share in our investor relationship deck. Just stay tuned, but we are extremely excited. On top of it, we have the BBU opportunities, we have the PDUs. We also have different areas of expansion. Couple areas. One is actually with the current device, current requirement. We also working new products supporting the future requirements. Great. You mentioned silicon carbide. How central is that as part of your product offering? Also, if you could talk to the extent that you can about plan for doing some of the epitaxy internally. Are you fabless right now? What is the product roadmap on silicon carbide and how big do you think it becomes ultimately as part of your portfolio? Is it just a small addition that's nice to have on top of some of the products that you have in PSUs or is it going to become more central? Yeah. silicon carbide is a big, I would say, important portion of our product engagement, product development. This is actually fell under our discrete power management area because a lot of the silicon carbide, whether it's the diodes or MOSFETs, is all belongs in the same category. This is only a part of a product offering, definitely is not everything that we offers. I think currently what we working with, especially on the epi area, we work with all the tier one suppliers like the Episil, EpiWorld stuff. all the actual manufacturing happen in our Taiwan facility, our own fab. We definitely want to continue to engage and continue to expand in this area. Like I mentioned earlier, all this product is actually linked together. What we really want to do is actually total solutions. If you think about Diodes, whether it's discrete, whether it's analog, whether it's mixed signal, whether it's timing, we want to actually loop all this things together provide a total solution to the customers. A lot of time, even with the PSU we mentioned, it's not just about silicon carbide. There's analog, there's discrete on the power protection, whether it's power or data line protection. With some of the other AIs, high speed protocol, PCI Express, whether it's USB Type-C, USB 3, right? All this is actually linked together. If you think about Diodes, other than we don't have CPU, GPU, we don't have the ASICs or FPGA, we actually have a lot of complementary solutions that actually offer to the customers together. I would say silicon carbide is one of the area, but definitely is not the whole thing about Diodes. I know you haven't provided a breakdown of product, but I think historically, it's probably fair to assume that a little bit more than 50% of your product mix has been power, discrete, and the rest being analog. At a high level, how have trends changed in favor potentially of power products? The discrete side, where I think in the past, people, your peers, were really focusing just on analog. I've heard Keh-Shew Lu saying in the past that, clearly there's very good gross margin opportunities in power as well. I just wanted to understand better, with data center, with this whole revamp of power supply, is discrete, is power getting an acceleration of the growth? Is it becoming more important? Is this a portfolio, a mix of product that you're happy with given what we're seeing in terms of trends unfolding in the market? Yeah. What we're seeing, the growth is actually coming both from the discrete as well as analog. Since we cover a wide range of analog power management, whether it's DC-DC, whether it's LDO area together on the discrete side. Right? We talk specific about 800 volt, but 48 volt is not getting away. Right? We still have 12 volt applications as well. I would say for us, we treat everything equally and some specific end application can drive some of the momentum, doesn't mean that others will not. Right? I would say, it's really a balance overall for us. We want to be a well-balanced overall supplier. Even with the end market focus, we talk a lot about AI, but auto, industrial. If you really think about automotive, there's also a power conversion moving towards from 48 volt- 800 volt or 400 volt. There's a lot of similarity going on. I would say all in all, it's actually very good potential for Diodes overall in the longer term. Great. A number of companies have talked about supply constraints over the last earning season, given the demand notably coming from AI that seems to be creating some scarcity. in other end market. How do you mitigate that? How big or small is it of an issue? Do you see the potential at some point for the supply constraint to really have an impact on demand where some analog companies could guide on supply as opposed to demand? Again, what steps are you taking to mitigate that? Okay. Demand supply is always there, right? We're definitely seeing very strong demand across the board. I think I talk about it with last earnings call. We're seeing a really good momentum picking up, even in the industrial market segment, automotive market segment, which is great. We focus on the hybrid manufacturing model. Basically what that means is actually we leverage our internal manufacturing capability together with partner outside as well. We believe with our hybrid manufacturing model, give us a lot of flexibility to dynamically adjusting things to better support the customer. What we're seeing is definitely there's constraint both in the backend, I will say more in the backend, and majority of the backend assembly and tests, we actually have in-house capability. We are extending the capacity aggressively, but more focused on the strategic level. We're not just close our eyes, expand everything. We want to make sure the area we expand is actually the future trending area. For example, smaller package, more density of the package, that's the area we expanding. On the front end, we actually own a lot of our, I would say fab in-house together with external partners. With the in-house manufacturing, we are aggressively qualifying, porting our own internal fab. We bought a fab from TI about five, six years ago in Scotland. We bought onsemi fab in South Portland, not too far from here, about two years ago. We are still going through the transition of qualifying the technology process as well as devices. We believe with the front and backend combination actually help Diodes overall better support the customer, working with their real demand, true demand. There are situations are very hand to mouth. It's not like we just have endless capacity, but we try to manage it very closely, stay very close, especially our tier 1, tier 2 customers, and work with them on their real demand. On the same topic as we see demand exceeding supply, and we've seen lead time expand a lot. Are we getting back to an environment that perhaps feels like 2020 where pricing is stabilizing? What do you see in terms of opportunities for peers to raise pricing? I mention peers because you tend to have a fairly price stable philosophy where you emphasize share gains. How do we look at the pricing environment, specifically in discrete, maybe in analog? Also embedded in that question, given those supply constraints, how do you see customers' steps to secure potentially 2027 capacity? On the pricing side, demand, supply always drive different pricing behavior overall in the market. We definitely see that happen already, some of my peers actually announced the second round of price increases. On the Diodes side, I think Tristan mentioned, he's right. I was talk about long-term benefits more important than the short-term gain. There are situations we have to adjust our prices as well, I don't want to say we're not doing it, we want to do it more strategically. We want to continue to gain the market share, design in some of the newer, better product with better margin. If you really think about at the end, we're talking about product margin or gross margin improvement. If I can secure some of the newer designs, with the production cycle, it will actually give me more stability. I call it sustainability for a longer term. I would say that's really more our key focus overall. Talking about long-term supply guarantee, amazingly during the COVID time, there's a lot of LTA, long-term agreements people put in place and then got burned and stuff like that, and customers kind of shy away from it. Currently, there's more customer want to talk about LTAs guarantee for a longer term, one to two years minimum supply, overall. I think that is a good indicator of the market overall change, and people definitely are concerned more on the supply than the specific pricing at this moment. I think overall we also seeing the lead time extended, the cycle time, because there's a queue time. It's definitely happening in the market at this moment. How do you feel your position relative to your peers, in terms of front-end capacity? We've seen, obviously, a big build of capacity in analog, including TI Analog Devices, all at the tail end of the last upcycle in 2023. In retrospect, does it look like all that aggressive capacity build was actually a good thing? Because demand seems to be stronger for the next probably three, five years than it's been. the past 10 years, if you have any comment on that in terms of was that big ramp in capacity industry-wide really ill-fated, or is that instead really going to help for the next few years, and how are you positioned within that? I think from the fab technology point of view, we probably need to separate advanced technology versus I call the more mature technology. All the Diodes products more sitting on the mature technology. We haven't really seen a massive expansion in this area. I mean, it's always difficult for us to time the market, right? If we go back to year 2000, 2020 to COVID, right? At that time, if you ask everybody, say capacity is not enough. We have the post-COVID that everybody say we have too many of the capacity. I think for Diodes, what we try to focus, we didn't really scale back our capacity, whether it's up or down that much. We really focus on the longer term, where we want to be as a company. I think last quarter, we talk about it. We finish our last year, $1.48 billion. We want to grow to $2 billion within the next three years. What we're looking at is actually our current landscape or manufacturing capability, whether it's enough to support our $2 billion goal, right? You can have a little bit up and down. Who knows what's going to happen tomorrow in the market? We want to make sure we look at our long-term goal to match the overall. If you really think about, with the $2 billion, we actually estimated 10.5% CAGR growth for the next three years. If I just look at the first half of this year versus last year first half, I believe it represent about 19% first half versus last year first half growth already. That's really where we want to focus, right? We believe the current manufacturing footprint will enable Diodes to continue to grow towards that direction. I think the South Portland fab in U.S., Maine, would be very focused on the analog, standard linear logic type of product focus. The Scotland fab in Greenock will be more focused on the discrete MOSFET type of area. With both of this structure or the footprint will actually enable us continue to be competitive in the market and continue to support from the supply-demand point of view in a longer term. Given the context of lead time expansion, does that facilitate the push industry-wide for customer re-qualification? In your case, that's for outsourced capacity? Of analog product back in-house, for companies like TI and ST that's trying to migrate people to 300 mm. What's the appetite for customers to re-qualify, I guess, in a way to secure capacity? Maybe that was not the case the past two years when we were in a down cycle. Right. Customers are always more flexible and willing to qualify a new product, whether it's pricing, whether it's capacity. We are leverage every single angle we can and work with customers to better support them in a longer term. Right? I would say all in all, this is all positive for us. Okay. In terms of timing, is this something that started earlier this year or is the bulk of those re-qualification fitting the Greenock fab and the South Portland, Maine fab happening next year? Are you already seeing customer re-qualification this year at those fabs, or is it more about new product ramp? Yeah. I think with the qualification, we have to separate into two areas. One is process qualification. We have to make sure we build a new process with the qualification, and that usually take a little bit longer. Once the process is qualified, then we qualify devices sit on the specific process technology node. What we're seeing is it's progressing. It's not going to be a cliff up or a cliff down. It's going to be every quarter gradually improvement throughout, I would say, next year and the year after. We will start seeing improvement gradually. I think we openly talk about it. There's a lot of changes with the manufacturing service agreement and stuff like that. We believe the worst is pretty much over, so everything is tailwind from now on. Great. Are you able to share maybe some of your target in terms of front-end in-house versus outsourced and how it is looking now in terms of the split? I think right now probably on the fab side, on the front end, I would say probably a 55-45% split. Eventually we want to be maybe more 60-70% in-house, then the outsource portion may be about a percentage somewhere between, I would say, 30-40%. On the back end, we always have a pretty high percentage that's in-house, so we expect to maintain similar percentage. I would say that's somewhere between 85-90% in-house at this moment. We want to continue to engage with our partners outside, especially on the technology on the new product area. We want to, whether it is Pericom type of product, we actually use a lot of third-party partners to enable the technology also enable the new product introduction. That is not going to change. We definitely would continue to remain the hybrid manufacturing model for the years to come. Back a few weeks before last earnings season, there was already feedback from investors that the recovery in analog was counterintuitive because of incremental weakness in PCs and smartphone because of DRAM pricing. Obviously those concerns were dismissed on the last earning season. Could you talk about how do you see those trends unfolding? You've mentioned on the last earnings call that you expect those trends to continue in the second half. You must have very good visibility. Maybe if you could give us a sense of the visibility that you have, how does that compare versus last year? And maybe your quick review by end-market in terms of- Sure demand trends. Yeah. Because we specific talk about computing, consumer, maybe we start with three Cs and then I'll talk about auto industrial. I think the biggest challenge for this year is actually memory shortage. The visibility overall in this area probably still kind of limited because most of the customer maybe have three-month visibility versus a longer term. I think that remains a concern. All in all, if you look at the size of the customers, I think the bigger company probably tend to have more of the capacity guaranteed versus smaller companies. I think there's still a lot of dynamic in this area. Going to the PC area, we've definitely seen very strong momentum continues in the AI-related area, especially AI server. We definitely see a good forecast and good momentum all in all. We're also expanding our content in this area. We believe that's going to continue to help drive for the 2026 as well as 2027. On the PC and motherboard, even with the last quarter, we definitely see some adjustment down, but it kind of help with the AI portion. All in all, that probably still remain a little bit uncertainty. Usually Q3, Q2 are more of the peak, and then usually Q4 is a little bit slowing down. That's based on our usual seasonality trend. Like I said, we do expect that we're going to be a double-digit growth this year. I think even with your estimate, definitely reflecting the expectation of the street, so we want to continue to execute. Consumer, it's even a little bit more uncertainty, very customer-dependent. Some of the customer, they can get very good support. Some definitely challenge a little. I would say, from the historic point of view, this hasn't been our number one focus overall. Right? Communication, we're seeing a lot of momentum also driven by the AI. There's a lot of switches and routers in the networking, very focused on the AI support and application overall. We're seeing good momentum. Auto continue to be our key focus and we want to continue to grow automotive. Based on last quarter, we finished about 20% and we want to grow much bigger. We have a very good momentum so far. If I just compare for the last 13 years, we actually have a compounded annual growth rate more than 20%, and we expect that momentum continues. This is more focused on the content expansion, not specific to the unit output. Industrial, we're actually seeing very good momentum. I think last quarter, I talked about it in the earnings call. Inventory shadow is over, right? Now we're definitely seeing a healthy momentum of picking up of the real demand, we expect that will continue to grow for this year as well. Great. Any signs yet that customers are starting to rebuild inventories? We're looking at ODM inventory days maybe picking up a bit the past quarter, even though at OEMs, inventory levels were still down year-over-year, quarter-over-quarter. In automotive and OEMs, our data is that we're probably back to pre-2021 levels, which is amazing. Are you getting feedback from customers that they wish they had a bit more buffer inventory, or do you expect that to happen into next year? A lot of customer definitely want to build up. If you really think about when we have an inventory correction, sometimes people overreact, correct it to a really low number, which may not be healthy to support their overall production. There's a lot of customer want to rebuild, but I want to be honest with you. On Diodes' situation, we are very hand to mouth at this moment. I don't really think we have a lot of luxury to rebuild the inventory. If you look at end of Q1 resale, I think I talk about it, both the channel in terms of inventory days, in terms of inventory dollar, decreased quarter-over-quarter. I would say, I believe if this rebuild would happen, probably will be sometime next year. I just don't see that as a possibility for this year at this moment. Because of the fact that you're not giving the customers all of what they need, and that's despite the fabs, the Greenock fab or the South Portland, Maine fab. Is it just a different mix where you see very strong demand in areas that are not served by those fabs or how should we look at this? Yeah. I think it's a combination of both. Number one, we also don't want to repeat the same situation that we went through in 2023, right? I think there's a lot of lesson learned. We want to make sure we are building the inventory at the right place for the right customers, for the right parts. That's number one. Number two, there's also some product mix change. I think last quarter, I talk about one of my peers, they have some supply interruption happen, right? That also kind of skewed the mix a little. I would say combination of both, but I think the key message is we want to work with the customer for their real demand, because remember 2023, we have the golden screw situation, so we don't want to repeat that. We want to make sure our shipping and ship-out at a very healthy level. We monitor the POS, point of sales, with distributions very closely. That's really something we try to do better learning from before. Any quick question from the audience? We have 30 seconds left.