Lumentum Holdings Inc. (LITE) Earnings Call Transcript & Summary

Welcome back, everyone. I'm Tom O'Malley, SMID cap U.S. semiconductors and semiconductor capital equipment analyst here at Barclays, very pleased to have Wajid Ali and Chris Coldren here from Lumentum. Thank you for joining us, guys.

Thank you for having us.

Thank you for coming across the country, I know it's a busy time of the year. Why don't we start with another time that we were all out here on the West Coast at the beginning of 2022? You guys laid out some growth profiles of your different businesses. Clearly, this year has seen a bunch of pushes, a bunch of takes from these different end markets. Can you just talk about the 3 big businesses, what you think in terms of growth rates for those and kind of compare and contrast what you were thinking about them then and maybe what you're thinking about them now?

Yes, sure, Tom, thanks, and thanks for having us. So the purpose of what we talked about, as you said almost a year ago, was to really highlight that we have been foreshadowing that we would lose or reduce from a very high market share position in our 3D sensing business to a more normalized market share because we had gained so much market share in that business in the prior year. But that still, the company at the aggregate level could grow double digit growth rates over the next several years. And the way that would be accomplished as the 3D sensing business normalized, that was almost 10% negative headwind to top line growth. But given we're at the front end of a large technology transition in telecommunications and data communications markets, moving to 400, 600 and 800 gig and coming off the pandemic, a lot of pent-up demand that those businesses would grow above their normal growth rates in the range of 20%, 30%. And lasers, same thing and more than offset that. Now what's happened in the September quarter, telecom and datacom, now you have to go pro forma for Neo plus Lumentum, the acquisition we closed in August, we were up 20% year-over-year. So not quite that 30%, but we highlighted that we had some supply constraints that did not improve per our expectations to the tune of $70 million, $80 million. That's one piece of the puzzle. So we still see very strong, healthy telecommunications and datacom demand outside of one sort of customer inventory challenge that we're working through. So we're still very optimistic about that 3-year time frame outlook. But as you alluded to, some setbacks in supply in general and some inventory built up at a customer that probably grows a little more conservative in their spending as they -- as everybody looks out to the macro environment and grows a little concerned uncertainty. So I think things are doing okay on that on a couple of year perspective, but obviously, some setbacks on the supply side.

Sure. So we talked -- well, you just talked about 3 separate issues there. Let's start with the supply side. So I think quarter-over-quarter, it's flattish or maybe even a little bit worse, I think it's 75% to 80% or in that range both quarters. Can you just talk about what you're seeing in the supply chain right now? I'd like to get to what we saw with Ciena this morning in a bit. But from your perspective, where are the gaining issues in terms of supply? I think historically, there had been some FPGA issues. Has that changed? And then is supply getting better or getting worse as this quarter is kind of going on here?

Yes. So I mean, like Chris alluded to earlier, we had supply constraints on FPGA devices and analog components coming into our fiscal year. As we move through the first quarter, we saw some improvement in FPGA supply, but the analog supply, even going into our fiscal Q2, wasn't where we had expected when we had originally guided for the year. We were thinking we would be able to get some supply of those specific IC components by our fiscal Q2, and then we could work it through our supply chain and be able to really -- I mean, it's hard to get it down to 0, but get it close enough to 0 by the end of the fiscal year from a supply-demand standpoint. As we work through the quarter and the months, we got not great news from our analog IC suppliers, and that caused us to reforecast when we actually thought we'd able to meet that demand from our customers that largely impacts our telecom business. And so some of what you're hearing outside of Lumentum in terms of the strength of telecom, we're not able to fulfill all of that demand primarily because of that. And our thinking coming into the quarter is that we'll probably see that impact us throughout the entire fiscal year, and if not, even a little bit into the first part of fiscal year '24. So it's unfortunate because we really wanted that to be a tailwind for us as we got through the year, but it's still persistent. And as we get feedback from our suppliers, it continues to be that way from what we had expected coming into October, November time frame.

So I think you've had a ton of conversations say, as I'm sure you have on people trying to square the circle. So Ciena says on their call this morning, "It's PMICs that are the issue that have been holding us back." You guys have said, "Hey, it's power chips as well." You've mentioned PMICs, I think, a couple of times as well. So there's obviously a month offset in your 2 quarters. Is today something that's incremental in terms of the supply chain that you guys have seen? Or do you think that this is more of them getting PMICs in areas outside of your product, it's just difficult to kind of parse is this in your lane or maybe another lane?

I think maybe one thing to add to what Wajid said that maybe help that is -- I mean, our telecom business did grow quite a bit quarter-on-quarter, and things like ROADMs grew 20% going into the Q4 to Q1, so from -- into the September quarter. So there is an increase, which is allowing our customers to grow. It's just now, we've reached a second bottleneck, meaning, as you talked about, the FPGAs, that's no longer a problem, allowed a big jump up in telecom growth. But then we sort of clamped at what the next bottleneck is, which is, I think, the kind of chips you're referencing.

And something that I struggle with is that if you look at a guy like a TI, they're building inventory on their balance sheet, there's product that they're producing that they can't get rid of at this point. Is it a certain type of power chip that you guys specifically need? Is it a higher SKU? Just help me understand like why the general purpose PMICs aren't the ones that you guys need and maybe something more specific.

Yes, that we are using something that's maybe a combination of either unique or older, let's say, that's not used as -- by as many folks. So it's -- this isn't a situation where if demand dropped off somewhere else that they could redirect those chips to us. It is generally targeted towards us. And the challenge then being is conceptually, we could have redesigned the product to use something else. But it takes time to redesign a product, it takes time to get your customer to recall it, and we had full expectation from a very credible, major supplier that something would materialize on a certain time frame that made redesigning the product not needed or not logical. Now in hindsight, had we done that, we would have been better off, but we didn't go that route.

Got it. One more on supply and then I promise I'll move on as this is a tiresome topic, I know, for you guys. So when I look at your financials, you non-GAAP out some charges related to procuring components. You had a competitor who did it for a quarter and then stopped doing it. Can you just talk about what, in your eyes, mandates something that is a non-onetime item from a procurement perspective?

Yes, I mean, from time to time, our procurement team has done spot buys where we're paying $1 for something, we'll pay $1.25 or $1.30 or we'll have some type of expedite fees, but they're not exuberant in nature. And really, when it came to the broker buys that we had to do to support our customer demand, and that customer demand is now flowing through as you kind of alluded to with Ciena, we felt it was much more important to satisfy that customer demand than to not purchase the components just because of a pricing issue that's related to those components. The charges are exuberant when you take a look at it on a per-part basis. I mean it can be multiples, many, many multiples of what we would normally pay for that product. Our view was, is that -- or is, is that, that is something that is going to take a few quarters to kind of put through. We are seeing that we're having to buy less from a broker standpoint. And so it's playing out as we had expected when we had first put that charge into our financial statements and had disclosed that, we had talked about the fact that we felt it was something that would last for a few quarters. And that's where we think it's going to end up. I think exiting our fiscal year this year, we should be able to see a lot of that cleared out from our inventory. And because we're not buying as many anymore, we shouldn't be able to see that anymore. But it was really so that people had some focus around how much that was really impacting us, especially with the NeoPhotonics acquisition and the change in our business model, we actually felt that provided more transparency to us.

Got it. Okay. So supply shortage is one. Second one is the datacom business. So you're ramping EML lasers. You have a headline customer, which has been done as Amazon, obviously, 400 gig. You were capacity constrained for multiple quarters. And it seems now that, that customer is working through some inventory issues. Can you just talk about the communication that you have with them? When did this spicket get turned off? Was it instantaneous? And is there any other dynamic there where there was maybe more competition? Just walk me through that process and how it went from being capacity constrained to being something that isn't growing for a couple of quarters there.

Yes. I mean I would say that ahead of the earnings call, obviously, we got a message that, hey, we don't need as much for a little while and what the explanation as to why it's not -- customers don't give us a whole laundry list of why they're not doing something, but I think it's pretty easy to surmise that we were aware of some other areas. The same customer is trimming inventory and preserving cash, if you will. So it seems very logical. And perhaps we overdeliver, meaning that we were supply constrained and then delivered on our commitments to customers and took the parts. So I think you put those 2 together that maybe they were buying an insurance policy, if you will, from a supply standpoint and then we delivered. And at the same time, grew more conservative as they look outward to the world and say, "Wow, do I want to tie up cash and inventory? Or would I rather put that cash to a better use?" So we view this at this point in time as more of a sort of an inventory correction, a customer-specific situation, probably not too unusual. I think when you have a situation where customers have been not getting what they want, they perhaps lean in on ordering. I know certainly with our suppliers, we drive them pretty hard. And you don't under-order and then hope that they can deliver more, you order everything you think you could ever use and then maybe use a little less once the situation at hand, by the time you get the supply, have not seen that proliferate to other customers and believe that, that will impact the December quarter. Maybe the March quarter kind of be the trough and then start growing from that point forward.

Yes, that's exactly what I was going to ask you next is I think a theme of some of the conversations I've been having over the last 2 days is an incremental weakening on the data center side versus expectations from even the earnings point, where -- whether it be additional inventory or some just weakening orders? We definitely sensed that. So I guess my question for you is you're more exposed to a few customers, but you definitely have a very good view of the broad hyperscale market. When you look out into '23, what's your feeling in terms of what you think spend will be for those guys? Do you think it will be a growth year for cloud? We've heard numbers down all the way up to plus mid-teens though.

Well, I think maybe from a Lumentum standpoint, we're -- there's a limited set of cloud guys. In general, it's not -- there's not 1,000 of them in the world. There's a handful that really matter. But our revenue is really focused in -- last quarter, it's 82%, 83%, 200-gig and 400-gig EMLs. So that's even sort of the bleeding edge, which means perhaps what happens is if somebody pushes out the next big build a little bit, then you bear the brunt of it, whereas we're not backfilling capacity, let's say, in an older data center where they're just adding incremental capacity. But as I look ahead, I think beyond a correction, I think it's a good spot to be because we're still in the very early days of these higher speed. So that market's going to continue to grow and at lower speeds are eventually going to be cannibalized. So I think for Lumentum, that it's a good spot to be in, to be in the next gen of what's coming because that's something that's going to be, probably, more resilient, if you will, even if there's a little bit of turbulence along the way to get there. But also, one thing that we're focused on is the reason why our revenue is so concentrated in e-mails. If you go back in time, right, we used to be a transceiver vendor, got rid of that business because it was a challenging business to make money in. What we were left with were the chips we had at the time. So we've taken that from millions of dollars to a quarter to, call it, $50 million a quarter kind of run rate. But in parallel, we've also invested a lot of money in R&D to develop 100-gig VCSELs. So as copper goes to fiber, that's something where we have 0 revenue today. It's a big market. There's very few competitors. Same thing about taking cost reducing 400 gig by making directly modulated lasers so that customers are more willing to adopt 400 gig. And then third, a portfolio of high-power lasers for silicon photonic-based transceivers or silicon photonic-based optical interconnects for new AI-oriented data center architectures. So the message really is yes, our EML business will recover. We're also broadening out our portfolio to have a wide range of products that play into new trends, not older trends so that will go from essentially 0 revenue on those kinds of products or very low revenue these days to adding things that can be as meaningful as our existing EML business, but more diverse.

Got it. It's amazing we're getting halfway through this year, and we haven't started on the 3D sensing side. I feel like it's been years since we've made it through the first 5 minutes. But turning to the consumer side. I think you guys have been very upfront and have done a good job of messaging the pattern of orders and your content for the last, really, 5 years since this, really, story started. This year, clearly, you're seeing some share normalization, as you called it, and there's also really not a ton of new content in these modules. Can you just talk about -- one, you've clearly given the CAGR long term of a flat business. So I think that tells us something about what you're thinking. But is there an opportunity at the largest customer for content gains in the future where you have a shot at seeing some increased content in future phones? Or is that view of a flat business overall kind of your way of saying this story is kind of done, and we're taking this off the table and derisking? I -- just walk me through potential opportunities and what you guys expect there.

Yes, I would say it's more of the latter. I mean we have new programs for new types of lasers and other content, if you will, with the lead customer. What we always do, right? And you don't know which are ultimately going to make it out into production. I think we look at it and say if there's a new feature or a new application that can be enabled that they can charge more money for, then certainly, there's a logic to a content gain. But in absence of that, why would somebody want to pay more to get a similar functionality? So we don't really see that on the time horizon. So what we're really focused on is obviously continuing to defend that business, but really on new applications, whether they be in consumer electronics or in auto or industrial for the same underlying core technology so we can leverage that capability and frankly, the volume and cost structure into other markets in that automobiles for LiDAR and for in-cabin sensing, extended reality, right, those are things that today, very little revenue is sourced from, but the opportunities could be very significant. And that view, as you said, a flattish, but that was flattish from a preshare normalization time, so it's adding another $100 million or $200 million over the next 3 or 4 years from new applications is something that we're targeting, if you will. It's not without risk because these are brand-new applications that are individually capable of driving those kinds of revenue, all about the timing of when do they take off. Obviously, automotive's a long, long cycle-type business, but once it happens, it's also a long life cycle business and therefore, it can be quite sticky and long-term, good for us.

So yes, I remember the chart you guys put up at OFC, and you gave some CAGR on some of the subsegments within that business as well. Would you say that auto is the area within 3D sensing that is the most exciting outside of the consumer-facing smartphones? Or is there another area that...

I think it's the time scale, right? Auto and industrial are the biggest and probably the best industry structure, if you will, but they're 3 years to inflect and really 5- and 10-year type opportunities. Between then and now, I think extended reality is the opportunity that has the most ability to ramp quickly because it is a consumer-type product, right? And I don't -- not something that fully resonates with me, but boy, if my 9-year-old son begging dad for a VR headset, and he's claiming all of his friends are getting them for Christmas, so maybe I'm wrong, maybe it will happen a little sooner, but it is something that has that potential as a consumer product.

Okay. Let's pivot to a bright spot here on the telecom side. So clearly, upstream, some of the larger guys are doing quite well. They're talking about getting product, and there's clearly a backlog and a demand profile that they say that they're shipping to. We've seen in other end markets that, that backlog always doesn't hold up, but it seems as of now, it's pretty steady. Can you guys talk about -- obviously, I like to think of the business in 3 ways, ROADM, pumps and then on the transmission side, just transceivers. So can you just walk through -- you obviously have a double-digit growth profile out there for the business. I think most people are pretty excited about this year. Where is that growth coming from? And can you just talk to what you're most excited about into this next year?

Yes, I mean, the big driver is a little bit of chicken and egg, but data center -- overall data center and Internet traffic has been growing very, very rapidly and relentlessly. And the technology to allow that growth needs to evolve. And for several years, we've been preparing for this 400 gig in the data center and 400, 600 and 800 gig in the telecom side of things. Pandemic kind of delayed that adoption, if you will. And now we're finally starting to see that occur. And kind of some of the key data points that I like to point to is -- we've talked about this on earnings call, the things like pump lasers you mentioned, those go into optical amplifiers that boost the optical signal as light goes through the fiber, those were up in our fiscal '22 for the whole year versus '21 50%, 60%, 70%, depending on which component -- pumps, passives or overall aggregate amplifiers. But I mean that's a big number. And those are some of the first things that inflect when you build a new network. What's been lagging behind that is ROADMs, where they should be shipping at similar rates but they're actually down, not from an order standpoint, but more from an ability to supply, as we've talked about. So I view that as, as supply comes in, they should jump up and be shipping at those same rates. At the same time, 400-, 600- and 800-gig components, they doubled in '22 over '21. Now off of obviously a very low base, but it's a signal that this transition is just starting to happen. Then obviously, our largest telecom customer, this morning, highlighting that they're now starting to accelerate. We tend to be a little bit before them just because of the timing of supply chains. So I think we're excited that there's a lot of opportunity out there, it's just this lingering supply problem continues to hold us back.

You've seen revenue exposure to that customer increase pretty drastically throughout the year. If you were to go to your customer today, do you think that they'd point to you guys as a chokehold for them? Or is it still other parts? I'm just trying to figure out, it seems like you're doing a better job of getting product to them, and they're clearly opening the floodgates on the revenue side. Could you just walk me through those conversations? Is it still very challenging to get them product? Is that a large portion of what you're calling out as shortages right now?

Yes. But I don't know if we're ultimately -- where we are [ in the limits, they're ] building a system. It's the old if you're missing one golden screw, right, and I don't know if we're necessarily that. But I would say we've been able to grow within our telecom customers because -- a couple of factors, right? We've really focused on next-gen technology. And in fact, you turn the clock back post the Oclaro transaction, we actually got rid of some of the older technologies that would have been cannibalized by divesting them. So it kind of -- in every technology transition, the old declines and the new grows and they get this net growth rate, we already kind of took out some of the old and vice versa, the pandemic suppressed some of the old. So the ability to grow quickly in the new products, which is where we have disproportionate exposure at our customers is probably why we're growing a little bit faster, a little earlier than they've been.

Got it. Another area, and I can direct more towards a CFO question here. Obviously, the Neo acquisition, really exciting, cements your position in the ZR space, but there is a business, obviously, the different gross margin profile than you guys have historically had. So obviously, in the near term dilutive, but can you talk what you guys are doing to bring that gross margin profile up? You've seen that business flash moments of higher gross margin. Walk me through what you guys are doing to bring it up to what you'd say is your corporate long-term model.

Yes, I mean, one of the things with our corporate long-term model is that we've got a portfolio of margins with -- generally, our chip business is being above our corporate gross margins and our module type or transceiver-type businesses being below our corporate gross margins. NeoPhotonics business, as you probably saw it when they were public, kind of in the $80 million a quarter range, they weren't doing so well. But as soon as they got up to $90 million or $95 million a quarter, the business was doing quite well from -- at least from a telecom benchmark standpoint. They were still -- they are still very far below kind of our telecom business. When we acquired NeoPhotonics, we pointed out that we felt like we had $50 million worth of synergy opportunities with some of the OpEx opportunities coming in within the first fiscal year. And then many of the cost of sales synergy opportunities coming in probably in the latter half of the first fiscal year and then into this second fiscal year especially as it relates to renegotiating pricing on components and also making a lot of the product transitions so that we have less platforms that are more impactful and then we can consolidate our factories and our operations. They've got a location in Japan, we do as well. They've got locations in China, we do as well. So there's a lot of opportunity there from our perspective. We're quite confident as we've gone through the first 3 or 4 months with NeoPhotonics, and we've actually started pushing through the execution plans on those synergies that we will be able to achieve, if not exceed, those type of synergies. That would add, at a company level, 200 to 300 basis of operating margin to the company. So we'll provide more updates on that at our next earnings call, but things are going quite well as it relates to that. So...

So you also did another acquisition. And I would say at first, when I saw, so it threw me for a loop a little bit. I remember back in the heyday of the last optical cycle, everyone felt as though they needed to own their own componentry and particularly owning a leading-edge digital asset is traditionally very difficult. If you're a transceiver vendor, if your DSP doesn't work, then you can't sell the whole transceiver, right? So what made you guys feel like you needed to go out and get a DSP? And intuitively, I'm thinking to myself, it doesn't seem like you want to compete with the leading bleeding edge. If it's in the telecom business, the Acacias and Marvells of the world, why did you feel like you needed that? Is that for other applications like in the backhaul, 5G area? Or is that just trying to be another provider of like the N-1 or the N-2 telecom company?

I mean, Chris can probably follow up with a better answer. But from our perspective, it's really about the cost of goods sold, right? I mean we're paying -- I don't want to say what price we're paying for a DSP publicly, but it's quite exuberant. And when you take a look at the opportunity we've got to be a follower, and we're not trying to be a leader from a DSP standpoint, but even being a follower and the type of cost of sales improvements we can see, especially given the demand that we are expecting on our 400 ZR+ products, it's in the tens of millions of dollars of opportunity as we're able to integrate our own DSP into our products. And so us being able to do that, especially given some of the long-term demand trends in transmission, I think, will be quite beneficial for us. We've seen some very good milestone achievements from the team in Brazil. And so we're quite excited about it. But from our perspective, it really allows us to control the supply chain and really improve our operating margins within our telecom business as we get that going and integrate it into our products. Chris, do you want to add?

Yes. I mean, I think the only thing I'd add is over the years, through acquisitions and organically, we're extremely vertically integrated now on these modules and at the same time, something we haven't talked about, but we're selling not coherent modules but 10 gig and now 25 gig, we call direct detect, so not coherent. It used to be the mainstay of telecom 10 years ago. It's pushing out to the edge and access-type applications, where the volumes are very large They're a little more cost-sensitive applications, but they're very large. Our -- that business grew 30% year-over-year, and it's -- quarter -- Q1, it's up doubled year-over-year. And on a meaningful number, that will eventually -- not next year or the next year, eventually be replaced by 100-gig coherent. So we felt both a cost reduction for DCI and applications like that, but also a necessary ingredient for these edge access applications, where, as you said, we're not competing with our customers in these. These are applications they don't even touch, but they're very high volume and suited for us when you're fully vertically integrated.

Great. Well, with that, we're running out of time here. I really appreciate you both being at the conference. And good luck for the remainder of the year and into next.

Thanks, Tom.

Thanks, Tom.

Take care.