All right. Thank you, and good afternoon, everybody. Welcome to the end of day one here for our TMT Conference from Morgan Stanley. Great to have you all here, and I'm Mark Edelstone from Morgan Stanley. Really happy to have Hong Hou, the CEO of Semtech, and Mark Lin, the CFO of the company, and we'll definitely have time for Q&A from the audience, but want to just spend the time trying to help everybody kind of understand the key issues facing the company, the drivers, and all those things, so maybe, Hong, you could just sort of start out. The company's been through a lot in the last few years. You've been the CEO for nine months. All the things that would have transpired up before that point, you had no involvement with.
So maybe to start at a high level, kind of give your perspective of how you see the company positioned today, where you see the key challenges, opportunities, and so on, just to kind of level set everybody.
Okay. Thank you, Mark. Yeah. No, the company has gone through a lot. I'm the third CEO in two years, and I'm nine months on the job. But I do feel that one side, you can say the board is not afraid of making changes if it's not on the right track. I do feel we are on the right track right now. And we have a very clear strategy: how do we go forward and capture the emerging opportunities and accelerate the growth? And there's a lot of opportunities ahead of us. That's why we are trying to do some portfolio optimization. In the last few months, I have really tried to instill a strong culture to change the way of customer engagement. So I do feel like we're on the right track, like any CEO will say, right?
But I do think our board is very involved in the governance of the company and to do right things in strategy and capture the opportunities.
Yeah. Maybe let's focus on culture for a moment. I've had the good fortune of being here for a long time, and I'm here at Morgan Stanley because our culture is amazing. So your perception as you came into Semtech, what would you define as the culture before, and what do you going forward?
So I benefited from being a board member for a year prior as an outsider looking inside and found that the company had this very strong technology base, very well respected by the industry. You got a very talented and dedicated employee base, but we're just somehow kind of not engaging with the customers, with the market, front footed. For example, the commercial side, we would wait for the industry to define a standard. Then we do launch the R&D effort. Then we go after that. In these days, in a very fast-moving world, that's just not enough. So that's the one thing: to change the customer engagement, the way to engage with the customers, and to really identify opportunities and provide solutions and not be afraid of doing custom design and purpose-built. And that's tremendously enhanced the customer intimacy.
We were able to get our roadmap aligned with the customer's need from multiple generations. That way, we can really invest our R&D dollars more efficiently.
Yeah. Great. So maybe then just talk a little about strategy, high level. What do you see as kind of the key strategic objectives for you and your team?
Yeah. So as for the strategy side, Semtech has been traditionally a pure-play semiconductor company. With the Sierra Wireless acquisition, we get a broader portfolio, but we also get a mixed stack. So in some areas, we continue to be in the leading position in the industry. And for example, in the data center, the AI connectivity, we can leverage our analog and mixed-signal design expertise and our established leading position in the industry to further enhance our portfolio and really provide low power, low latency, and cost-effective solutions to the AI connectivity. But there's the other end of the spectrum. For example, some modules, we probably are competing with ODMs. So you will never get the gross margin and profit margin profile in the IC areas.
That is a strategy we'll try to transform into a pure-play semiconductor company and specializing in analog and mixed-signal serving the AI connectivities, IoT, and high-end consumer electronics applications.
Okay. Great. And we'll dig into all those things. But let's just go back. I thought you did something that was really quite important. You obviously came in at a really difficult time for the company. And I personally think with your background, you're kind of the perfect person to lead Semtech out of the issues that we have. But I think your first 100 days on the job or something like that, you set up your three near-term priorities that you had. Can you maybe just give an update to those, how you think they've played out so far? Do you need to make adjustments at all to what those near-term priorities are?
Yeah. Thank you. So yeah, I did set these three near-term priorities when I started with the company as a CEO. The first is to strengthen the balance sheet, rationalize our product portfolio. And the second is to discipline the innovation. As a technology company, new product and new technology development is a lifeline for the future business growth. Then the third area is enhancing the engagement of the employee and engagement with the customers. We were somewhat kind of like a not very clear in the future direction, partly because of the challenges in the balance sheet. So I will say now, nine months later, the product rationalization is very clear, what's core and non-core. And the balance sheet has improved tremendously. We reduced the leverage ratio from 11.2x to now 2.2x.
And that allows us to focus on accelerating the growth through some disciplined investment in the second area of the priority of the R&D. Then the third area, I couldn't be happier that the engagement score with the employees based on the recent survey has demonstrated tremendous improvement. So we are reestablishing the trust with the leadership team, with the management. And then the employees are leaning forward, engaging. We launched a Semtech Rising initiative. So the employees really feel that they are getting the true information. They are inspired to really take the empowerment and accountability and do whatever they do their best to build up this winning culture. So I'm really happy about the progress over the last few months.
Yeah. Is it time to start adding some additional near-term priorities?
Yes. So yeah, I think that, for example, the balance sheet is improved. That's no longer the priority. The rationalization, the portfolio is rationalized, but we need to focus on the transformation. Some non-core asset, we are in the process of running the divestiture. Some of the core asset, we're going to be adding more R&Ds. We may identify some gaps. We may do some small tuck-ins to make the portfolio align with the market needs. And as for the culture, really, right now, I think this year, if it's a first year, the focus on Semtech Rising this year will focus on Semtech Transforming so that we'll get a good portfolio, good product lineup to, in the future years, Semtech Accelerating.
Yeah. You've got a broad portfolio, for sure. So maybe you can just help everybody understand what you see as the most strategic parts of the portfolio that you want to continue to basically foster and invest behind.
Sure. So in general, the IC focus, we are a high-performance analog and mixed-signal company. We have been in the space of connectivity for high speed. For the AI connectivity, the low latency, high speed, lower power are very, very critical. So we will continue to focus in that area, and we do see some gaps on our portfolio, and we can spend some R&D to fill the gap, and the gaps were identified through the alignment with the customers, through frequent meetings at a technical level, and really, it's great. I feel like the investment on those products will be low risk. Basically, we build, they will come. LoRa, long-range IoT, that technology was created by Semtech, and the industry was created by Semtech. Over the last 15 years, we really have worked with our ecosystem, demonstrated a lot of successful use cases.
Now is the time for us to provide more enablement, provide broadening of the portfolio to accelerate the growth. That's core. Even high-end consumer electronics, we're in a great position. We can leverage the same core competency in analog and mixed-signal design to get a stronger portfolio in there and proliferating in major electronics companies. And for some, one of the rationales at that time of the acquisition for Sierra Wireless is to use a service platform and offer LoRa services. And recently, I've done some deep dives with my management team. We see that that synergy can play out. So those are the core areas. What can be non-core is, for example, it's a cellular module, taking the chipset from Qualcomm and making modules to be integrated in the different routers and gateways and some other devices for different cellular operator applications.
That may not be core for us because we don't have a low-cost manufacturer base. But it can be core to some company. They have that infrastructure. The great thing is the business has inflected to growth. We have demonstrated that quarter-over-quarter growth over the last three quarters in a row. And there are also tremendous tailwinds that play into the strength. So this will be the good time to find a great buyer to get that portfolio divested.
Yeah. Fantastic. And if something could do that, maybe Mark, give your sort of priorities. If you're able to divest non-core assets and monetize that accordingly, what would you use the cash for?
Yeah. Currently, we're still looking at debt repayment as a number one area. We still have a credit facility outstanding that has a fairly high interest rate. So that would be the number one use there, but continue to strengthen the balance sheet. I think that strategy has served us well, and we'll continue with that path.
Yeah. You've obviously done a good job at delevering here pretty meaningfully in the last nine months. So do you have a target of where you'd actually like to see sort of long-term leverage for the company?
Yeah. So we've given a long-term leverage target of less than two and a half times that. We're below that now, right? So there is some optionality. But in cases where we want to deliver a little bit more, to have some dry powder maybe for a potential tuck-in that Hong mentioned. But really, it's focused on reducing our leverage.
Yeah. Maybe go back, Hong, to your comment on just customer engagement and really making sure that you're delivering what customers want. I would argue, and I've basically followed Semtech for decades, I would argue that under Mohan's leadership, he did that with LoRa, it seemed to me. LoRa was clearly something he was significantly passionate about and probably over-indexed on it, actually, quite frankly. But when you look at things in signal integrity and the broader sort of port protection and the broader portfolio of core Semtech, is this sort of customer alignment, is that a new thing, or you just enhanced what was being done before? Just give us a sense for how much was really deep in the weeds with your customers to make sure that you're designing products for what they really want.
Yeah. So Mark, in that area, I will say even up to today, we have made tremendous progress. There's still more to go. I think company probably, I don't know exactly the reason. I feel like a very passive used to be. I mean, maybe because about more than 70% of our revenue were through channels. So the commercial team is very used to just push the channels, have them to drive the design wins, drive the engagement with the customers. I was overstating it. I said we were sitting behind and pushing them. I said, "I want everyone to jump to the front to pull, to generate the demand, and not only engaging with the customer, but also engaging with the customer's customer." I know now ACC is kind of like a generated a little bit the pocket, but that was a result of engaging with the customer's customer.
We see that is very effective. At this point, right now, the innovation cycle is so short, so accelerated. You just don't have the luxury to wait for the smart guys in the room to debate and formulate the standards, then you launch your R&D to do it, and another thing is really benefited, especially in the AI and high-end consumer electronics area. You have the leading customer. They are hyperscale. Even you may come up with a solution that may not be universal, may not be ubiquitous to address the entire industry's need. As long as you lined up with one or two hyperscalers, you're going to be having a very sizable demand, so we're changing the way and really getting people to go out and talk to the customers so frequently.
I would say the increase of the frequency of touchpoint with the customers may be by two orders of magnitude.
Yeah.
Because I was the customer of Semtech at the Intel days. I know how frequent they can bother us, but now I want them to engage with our customers almost like several times a week.
Right. Yeah. As I said earlier, your background is so perfect for Semtech, laid out against the core competencies of the company. So maybe in that, I think one of the more exciting areas is signal integrity. So maybe talk about that. You talked a little bit about the kind of core analog, mixed signal, the ability to basically drive performance at low power. But just contextualize that business within infrastructure. We can start with data center to begin with. Maybe size the opportunity and then give a sense for where you think the real opportunities would exist in data center for you going forward.
Great. Yeah. Sure. The data center area right now, the demand is increasing driven by the CapEx spending of all the CSPs and service providers and then the enterprise and then the sovereign, the building the data center everywhere. We provide the key electronics IC components to the optical module suppliers. Semtech has been recognized as a leader in the TIA or transimpedance amplifier industry for a couple of decades. Arguably, we provide the best components in the industry. That buys us a ticket into the game. Right now, the strategy is how do we increase the share of the wallet. How do we bring in the TIAs and bring in the drivers. Then bring the CDRs. Certainly, the DSP has been in very high demand, but DSP is also a very high barrier. For us, it's too little, too late.
We had several unsuccessful attempts over the last 20 years at Semtech alone. I will leave that business to the capable industry participants for them to continue the innovation. But we can have the analog version of CDR, Clock and Data Recovery, to cater for the specific needs of low power, low latency, and lower cost. So that's on the electronics part. Of course, if I have some compelling optical components I can bring along, then I'm changing my position from just being a commodity component supplier to possibly a solution provider.
Yeah.
So, this is really think about, this is very exciting. The ACC is part of that solution. What’s ACC? Behind ACC is a linear equalizer. That is basically provide a very low-cost, low-power, low-latency solution for signal integrity and high-speed. LPO can drive that transformation for the share of wallet from a few dollars to tens of dollars to several tens of dollars, that level of the growth. So, enrich the portfolio, transforming the business strategy. And that’s basically what I’m trying to do to grow the signal integrity product business.
Right. Okay. So a lot of drivers, which I think is good. And it seems to me that in this data-centric era of AI, it just increases the opportunities that you should have. The rate of change should be high. Obviously, there should be pretty big TAM for you to go after. You had to put out an 8-K on the CopperEdge with ACC. Maybe just give a little bit of details on that just so people can fully understand it. And then also kind of how we see that basically sort of going forward. Because I would argue it's a relatively small part of your business, and it seemed to have a pretty adverse reaction in people's minds.
Yeah. No, I know that's a sore subject. You're absolutely right. It's a small incremental business. So the ACC is, again, the linear equalizer we developed, co-developed with a leading GPU manufacturer to solve the interconnect between racks for the GPU compute cluster, so we guided a number right before I joined the company. I feel like we feel the management feel like that's a material information out there. A couple of days before the 8-K event, we were given the new forecast and also informed on the platform change, which will create air pocket for demand by two to four quarters, depend on their next generation platform. We were told basically they wanted to ensure the generation after the two generations after the current one, it will be on time, so that will clearly reduce our ramp of the linear equalizer for this specific customer below $50 million.
But we have been using the same engagement methodology with a broader customer base from CSPs in the U.S., CSPs in China, switch vendor in U.S., switch vendor in China, ODMs in Taiwan, cable companies around the world. So we probably engage with 20, 25 possible customers using ACC. So the response has been super positive. And they are at different stages of the verification and qualification. I will not be surprised. One of the major ones can start buying in volume toward the end of this year. But it does have this temporary air pocket. It definitely was not handled as well, kind of like an 8-K release Friday afternoon without there being some big news we buried something.
But it is a small impact to the overall business in the long run. But we just wanted to get in front of the investors and analysts to explain the situation. Other part of the business fundamentals has not changed. We will have our earnings call next Thursday. Certainly, you'll hear what the results for the Q4 and FY25 prior to this 8-K came about.
Okay. Great. Maybe just you've obviously have other things that you do in infrastructure. Maybe talk about that part of the business as well.
The other part of the infrastructure we have in there. So Mark, maybe I can hand it over to you.
Yeah. So in other parts of the infrastructure business, we have a telecommunications piece of the business. That area, there's one section that's passive optical network or PON. And then we have what we call wireless backhaul, so supporting cellular infrastructure. Those areas, especially with cellular infrastructure, I'd say it's been fairly muted for quite some time. So in that area, we're seeing in the industry, there's some tailwinds. But in that area, we'll probably count it as a growth area when we see some carrier CapEx come through, right? And then in PON, passive optical network, by the way, this is the same design team that works on the rest of our infrastructure business. So there's a lot of synergies in terms of R&D and capability there. With PON, it's largely aligned to a tender process.
So a tender process is a little bit cyclical, but I think that PON has recovered definitely off of the lows from last year.
Yeah, so maybe just if you kind of wrap that all together and looked at what you think maybe like a five-year growth rate of data center and maybe overall infrastructure could be given the portfolio that you have and the dynamics in the end markets, what does that look like?
Yeah. So Mark, what we were trying to do, I wanted to transform to a compelling portfolio that can achieve faster than the market growth rate. So we set a target of 50% higher than the market. For example, if the market grew 10%, wanted to be in a position to grow 15%. What gave us the reason for that? Basically, we wanted to really anticipate what the market needs. And we have done four-month-long strategic planning process. I think we map out the industry. We really had a lot of debate. And the plan we put in place, I do believe, will give us a comprehensive and compelling portfolio to enable us to accelerate the growth than the industry.
Yeah. And is that just within data center and infrastructure or is that like the whole company?
In every one of our core areas, but of course, I have a couple of moves I need to do in order to get the portfolio optimized.
Right. Okay. Very good. And then maybe on that point, when you think about, and then we'll take questions in the audience in one second. Maybe Mark for you, you have the range of products today, post-Sierra Wireless. And we know the module business is not a high-margin business as an example. But if you envision a world where you're back to what core Semtech would have looked like before, plus maybe the connected services business, which is high margin, what does the kind of end state look like for gross and operating margins when you have the portfolio more aligned again?
Yeah. Definitely. I think you can get there pretty quickly, right? When you look at the portfolio and just maybe looking at some of our disclosures, I think in the 60% gross margin is a very kind of achievable area. That's the place that we want to be. And then the little bit of a drive would be to move operating margins in starting with a three, so in the 30s, right? There's a little bit more work to do on operating margin, a little bit more of scale. But definitely on the gross margin side, that's a more near-term achievable target.
Right. Okay. So very much top tier. Okay. We have a question in the audience. Just wait for the mic. Thanks.
Yeah. Two questions. The first one is kind of on your CDR business. I think that's a business that should do very well, both the ClearEdge and the TriEdge for NRZ and PAM4. How much is that? What can you share in terms of how much is that in your business and what the growth rate is for that? Second question, kind of longer term is kind of co-packaged copper. Right now, that backplane on the NVL72 is big, ugly, and hard to do anything with if something goes wrong. But co-packaged copper looks like a great way to scale up and wanted to understand how do you guys play co-packaged copper. So really CDR question and co-packaged copper. Thanks.
Right. The CDR, so the TriEdge product we have right now is a bit dated. It was developed four or five years ago with PAM4 50 gig, and it's used in active optical cables, so 200 gig aggregated bandwidth and 400 gig aggregated bandwidth. The good thing is it has demonstrated really lower power than a DSP-enabled counterpart, and so I know it was not easy at the time. That's why we did not continue for 100 gig and 200 gig, but we had some architecture breakthrough. We also see the industry processing capability improvement enables a new performance of fT max of the transistors, so we have now adding 100 gigabit CDR and 200 gigabit CDR into our product portfolio, so when we communicated with our customers, this makes sense. Do you think you may have a use? Are you kidding me? That's great.
So it's very well anticipated. It can be used immediately in AOCs, AEC where they need more knobs than ACC, but they don't need the power and complexity of DSP. So I think that's going to be a good product. That is on our roadmap. We have a good schedule to sample to the customers on different generations of products. As for the copper backplane, the CPC co-packaged copper looks so slick. It's just beautiful. So TE Connectivity and our partner and then others, Luxshare, they had similar thing as well. And they got the interconnect management done so well and so seamlessly. That's going to improve the airflow because of better layout and improve the signal integrity as well because it's not buckled and bundled, all of this running into each other thing. Our play there is a linear equalizer in the connectors. So there's a lot of innovation.
Now we see that when they see the linear equalizer is capable and provides the performance they anticipated.
Thanks.
Yes, over here.
Hi, Dr. Hou. I think in last November, we had reported last quarter results. You mentioned you see a lot of CSP and other companies use linear equalizer, no matter it's on the board, on the cable, on the connector. So basically, you just mentioned about what's your progress in engaging with the cable customers. So do you have any progress how to engage using linear equalizer on board? Do you see any update, design win for this? Thanks.
Good question. So we engage with probably a total of 20-25 potential, well, customers. They are using in different configurations the paddle board and then a little daughter board and some built into the PCBs because they sometimes ask us to do some little bit modification from pitches of the bonding, the wire bond, or the surface mount pitches. The primary applications are several. One of them is the backplane. And another one is the inter-rack connectivity. Another one is the NIC port to top of the rack. And there are CPC co-packaged copper cable applications. There are five, six different applications. As I said, one of them they have gone through pretty much the validation of the different corner cases. They pretty much finalized and ready to go, but they're waiting for their deployment volume before the end of the year.
But then there are additional applications that is under development.
Do you think on board linear equalizer can replace a retimer? Thanks.
So, on board retimer, so the DSP-based retimer, if you need a DSP-based retimer, it'll probably be easier to be in the pluggable modules. I haven't seen many use cases that they got a DSP integrated on board. And plus the DSP consumes a lot more power. Then it really has a lot more involved in you do DSP, but you have to analog, digital, digital analog conversion. I don't know the use cases of DSP on board, and then that need to be replaced by the linear equalizer. And maybe closest one is LPO. That is a little bit motherboard, DSP-based board being replaced by linear equalizer-based board.
Other questions? Yes, over here. Just wait for the mic, please.
Hi. My name is Victor. I'm not too familiar with this space. Can you educate me a little bit about ACC versus AEC, and to me, it feels like a competing product, but there's pros and cons in both ways. You guys are lower power, low latency, lower cost, etc., but based on whether it be Credo or Astera Labs, you guys are saying AEC is a superior approach that the CSPs prefer. I'm just wondering what your thought is?
Sure. Yeah. AEC, just for those of you not familiar with the terminology, stands for active electrical cable. This product was designed and introduced about four or five years ago, was primarily to address the extension and reach the typical DAC cable or direct attach cable. It goes about a meter or slightly over, depending on the host signal-to-noise ratio or host quality, and in many use cases in a data center, that's just not long enough for 100 gigabit trees. That's why the AEC was developed. It was very quickly adopted by many CSPs, well, several CSPs. Today, AEC, as you named, there's a couple of leading suppliers. They have a pretty good volume and business shipping to three out of four DSPs. What the ACC can do, the same thing, basically improve the signal integrity.
For the 200 gigabit, we right now can transmit a signal up to three meters. For 100 gigabit, in some cases, we have demonstrated it can transmit up to five meters. So AEC, the standard specification, I think it's about seven meters, right? Seven meters at 100 gigabit. So we are getting there. But AEC still can transmit for longer distance. As for power, ACC is consuming a lot less power than AEC, maybe about 20% of total power consumption of AEC. And then that's not introduced any signal delay latency because there are no DSP involved.
So you can say pros and cons. Pros, certainly low cost, low latency, low power for ACC.
Con is not as far reached marginally. So I do believe the ACC gives them time, give a concerted effort to market this technology, make aware of the advantage of this technology to the marketplace. In a couple of years, it can grow into the level of business of AEC.
Okay. Great. I think we will stop there. So thank you all for your questions. Thank you for being with us here at the end of day one of our conference. Look forward to seeing you tomorrow. Thanks to Hong and Mark.
Thank you.
Thank you, Mark.