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

My name is Simon Leopold, Raymond James data infrastructure analyst, and we are at the Raymond James Tech and Consumer Conference. We added consumer this year, by the way. For this session, we've got with us from Lumentum, Chris Coldren, who is the SVP of Strategy and Corporate Development. And in the audience, we have Kathy Ta, who runs the Investor Relations side of the business. So format is classic fireside chat. I've got an outline of questions, but if you've got questions, wave at me or I'll check with the audience towards the end to ask. Chris, I know the story, but maybe some of the folks are new.

How do you like to introduce Lumentum to folks and introduce yourself to folks?

Yes. I mean, Lumentum is a technology market leader making photonic devices. And so photonic devices historically have been deployed heavily in networking applications, but we also have other end markets we serve in the industrial space and still a little bit in the consumer space as well. And then my role within Lumentum is -- as title implies Strategy, Corporate Development, but over my career, I've had roles in product and technology development, marketing, finance. So seen the industry from a few different angles over many years.

So starting out with something kind of topical whole company is this topic of the post-pandemic world. We've had this big buildup of inventory, carriers of too much inventory, OEMs have too much inventory. Where are we in that cycle? And how is it affecting Lumentum's business overall?

Yes. Well, it's impacted just about every major business line within Lumentum. First, it started off in our datacom or business serving data centers a little over a year ago, where demand dropped quite significantly as customers and end customers started bringing down inventory levels. But over the past several months, demand has turned up quite significantly. And now revenue we're working to get up, but given the manufacturing time, in the case of selling laser chips into this end market is several quarters. We're through the trough from a demand standpoint, but looking to ramp up to prior levels or even exceed prior levels given the strength that we've seen in the end market. Telecom, on the other hand, started correcting a little bit later, earlier this year, maybe the March, April time frame, and has kind of unfolded since then. So hard to say if we're -- it feels like we're at trough-ish levels, plus or minus. And when we look at our telecom business being down almost 40% year-over-year, whereas our customers, in most cases, are actually growing, it's pretty clear we're undershipping end market demand. We don't have perfect visibility into customer inventory to know exactly when things will turn, but we certainly believe that telecom will start growing during calendar '24, hopefully sooner in calendar '24, but not perfect visibility as to that. Commercial lasers or industrial, the last portion of our business, it's currently not at trough levels. It's coming down a little bit. We think trough levels are probably early calendar '24. We have some new businesses selling into solar space, display processing, some other microelectronics applications based on some new ultrafast lasers that we've developed over the last several years that kind of ameliorates or partially offsets the broader industrial weakness. Smaller portion of the business, but for the sake of completeness, I think you can kind of see these 3 waves that have unfolded, unfortunately. And now certainly, the datacom and cloud portion improving, and then we expect sort of, in a similar sequence, telecom and then lasers to flow through. What that really impacts our business is it's kind of a bummer that kind of a lot of work over the last several years in terms of acquisitions, bringing them in, executing synergies that reduce sort of the fixed cost base of the business, improve margins and principal. But during the pandemic, we were supply constrained and then quickly went to demand constrained. So the lower revenue levels, the nice sort of variable margins we have now cut both ways, and on the down it's negatively impacted us, and margins due to under-absorption in our factories are a bit painful. But the flip side is we're very focused on doing everything we can do to ensure that, as the demand returns, that margins will grow much faster than revenue grows in that return to normal.

And trying to look past this sort of trough period, what does the business model look like in sort of normal world? So what's a normal growth rate, normal gross margin, normal operating margin?

Yes. I would say there's probably been a little bit of changes given the recent Cloud Light acquisition, and we've not updated any model or growth rates. But I think we highlighted earlier this calendar year, at OFC, we had the target models, both kind of a midterm model and a long-term model. And let's just focus maybe on the midterm model for the moment. It was about $200 million -- or sorry, $2 billion, so $500 million a quarter, or a little bit higher than that from a range, and getting into the low to mid-40% gross margin and low 20% range operating margins. I don't think those are readily achievable in our business. I mean, maybe the revenue needs to gets tweaked for the Cloud Light acquisition coming in. Our focus now in the near term is really let's do what we can to get revenue back to those levels as well as execute...

Just remind folks what sort of the last quarter revenue was.

Yes. Last quarter revenue was in the $300 million -- low $300 million range now. Obviously, with Cloud Light coming in, that increases where we're at a bit. But that's significantly below where we were previously and where we believe we should be running at. So now we've put a lot of our focus on investing in new products, so that as we come out of the trough, we're able to grow off of the secular trends and technology trends in the market, acquisition synergies and as well as for both our datacom chip business, expanding manufacturing capacity because there's a lot of demand out there driven by cloud trends, particularly those driven by AI, where our datacom chip excellence tends to be at the very highest speeds, and therefore, we maybe get more uplift than average given what's going on in AI. And then similarly adding manufacturing capacity for the Cloud Light acquisition as they're seeing very strong demand for 800-gig transceivers, again, most likely driven a lot by surge in AI build-outs.

So I can kind of guess the answer to this question, but stepping back and looking at the whole company, what do you see as the most promising opportunities for calendar '24?

Well, I think first is, and as we've alluded to here, is recovery. I mean, I think as we grow back to normal run rates through calendar '24, there'll be a significant margin uplift, and so -- and to my point about some of the acquisitions we've done and a lot of the hard work the team has done to optimize our factory footprint, et cetera, start to see the benefits of that in our numbers. But then I think as you were alluding to, we've, between NeoPhotonics and then Cloud Light, acquired -- and a lot of our own organic programs that have over the last several years focus on expanding our cloud business, which has been a smaller portion of the overall revenue of the company and believe that in calendar '24 we'll get that up to more than 30% of company revenues. And we all know that the cloud end markets are growing quite rapidly, so that should accelerate overall company revenue growth rates. So I would say that. And that just -- that not only impacts our datacom business but also our telecom business where cloud customers are also large purchasers of, whether they be coherent DWDM modules for data center interconnects or we supply components and modules through our network equipment manufacturer customers who ultimately land in the cloud space. So I think that's the largest 2 opportunities we have in front of us: general recovery and then seeing the benefit of all the hard work we've done over the past several years. And then two, accelerated cloud exposure and, therefore, the related growth.

Now how do you summarize this Cloud Light acquisition for somebody who maybe isn't as close to the story?

Yes. So what Cloud Light manufactures and -- develops and manufactures are optical transceivers to interconnect equipment within a data center, whether that's servers to switch, switch to switch, but ultimately, in an AI-driven world, there's lots of extra bandwidth needed to interconnect the distributed compute element. So there's a lot more network capacity needed and what Cloud Light delivers is high-performance optical transceivers to affect data links between these network elements. The rationale for us acquiring them was several fold. One is it increases our exposure to the cloud growth trends significantly, perhaps more than 2x our prior exposure, but also gives us a more direct relationship with the cloud operators or the cloud infrastructure providers. We've participated in the market in a more -- a smaller way supplying laser components, if you will, very good business, nonetheless. But you're one layer removed, and so less sort of connected to what's going on within data centers and the technology trends. So it gives us a seat at the table with the cloud operators who, as I alluded to earlier, also buy a lot of other telecom gear and a lot of other things that we as Lumentum could supply. So not only does it add a lot of revenue growth to the company, but it also increases our broader knowledge and relationship base with the customers. That also brings a lot of manufacturing, high-volume manufacturing technology that's applicable to any of our transceivers, frankly, even our telecom space can leverage these high-volume platforms, whereas in the telecom space, we've been much more focused on high performance, but at the same time, those high-performance volume or high-performance modules are now settling into applications where there is higher volumes. And so bringing in that manufacturing expertise is really critical.

So one of the things that I think some investors were critical of is you exited the datacom transceiver business and now you get back in after spending money to do so. But what I think that overlooks is how the business has changed from when you exited to today and maybe it's a combination of margin and size, but I have to imagine you've had this [ indiscernible ].

Sure. Yes. And I think it's -- I think one has to caveat that when we -- the market has changed, the world has changed, we've changed over the ensuing 5 or so years. And at the same time, what we had going back when we had a datacom module business was something that was really built around the enterprise world of old and was struggling to transition to the cloud world. And cloud was starting to take off and in fact, to take over the overall datacom world. Now you go 5 years later, I think that transition has fully happened in the market. As that transition was happening, there was a lot of uncertainty of who the competitors were, those emerging competitors, and we just made the decision. We weren't able -- we didn't think we could take what we had and transition it to a successful cloud business during a period where there was lots of sort of cross currents going on in the market, whereas we had a gem of laser technology that came from Oclaro that enabled us to sort of free ourselves of the module business, grow the...

And if I recall correctly, the modules had terrible, terrible, terrible...

Terrible margins because they were built for a different business that paid generally a higher price for a different set of features, if you will, than what the cloud customers were willing to pay for. Now you roll forward and say, wow, okay, the industry has stabilized a little bit more, who the key players are has stabilized. We've built up a significant component capability that can supply into cloud light or any datacom module. We've ramped up production in a large factory in Thailand, which 5 years ago was just starting to get off the ground, and we were at the time making transceivers in a contract manufacturing environment, so stacked margins not as cost effective. So now we have in-house manufacturing. So in a sense, we've grown to be a better partner to a -- or a better hand in glove fit with a transceiver business being within us. And I think Cloud Light brings a different set of capabilities that are a little more aligned with what a successful cloud vendor needs to be today and into the future.

Now how does this acquisition and your participation in the transceiver market affect your relationship with your customers that make transceivers with your lasers?

Yes. So I would say it certainly does change things a little bit. It adds some complexity to the dialogue. And when you're not a direct competitor, it's easier to share information. But on the other hand, we have been supplying laser chips to broader competitors in the market. I think the way to summarize it is, if you don't have product leadership, you're not going to sell product. And if you have product leadership, the end customers are going to recognize that, direct customers are going to recognize that, and you will continue to have a good business there. And so we view the 2 businesses as not being in a complete conflict with each other, but in some ways in -- and complementary. And at the same time, the Cloud Light acquisition becomes a customer for some of our in-house products. So we've been developing VCSEL laser chips to supply into the market, some of the largest customers for those are vertically integrated, so they might not be a great target customer, but Cloud Light makes VCSEL-based transceivers. So they become a customer of our own products. And Cloud light is also making silicon photonic-based transceivers, which are complementary to our largest product line on the chip side is EML, an indium phosphide-based laser. So we can play in -- end customers may have a preference for one or the other. In this way, we're able to play in both successfully.

Now there have been some other assets sold in the industry. So Intel sold its silicon photonics division and then AOI sold, but then canceled its Chinese data center business. Why were those assets -- why did you choose Cloud Light over those assets?

Yes. I mean we certainly are familiar with the range of potential combinations that are out there. Certainly, as Alan always says, you have to have a willing seller and buyer at that -- in order for a deal to happen. But I would also say, we genuinely feel that Cloud Light is the right fit for us in terms of we wanted to buy somebody that was disproportionately exposed to highest speeds, if you will. And as we've alluded to in the transaction announcement and on the subsequent call that Cloud Light has more than 50% revenue, from 800 gig speeds in recent time periods and over 90% of the revenue coming in from 400 gig and 800 gig speeds. So really focus on the highest speed portion of the market. We've talked about that sort of chip component complementarity that we can feed in. And then as we did our own due diligence in talking with customers a lot, pointed us towards Cloud Light is something they really liked and would love to see partnered with Lumentum. So you kind of put all those factors together, that was the right transaction for us.

Yes. Maybe just let's get into the weeds for a little bit because you make a number of different laser form factors for the data center. There's EMLs, DMLs and VCSELs. And VCSELs are sort of the shorter reach ones and EML are kind of the leading edge, right? What -- our sense is sort of the VCSELs are a hot topic now because of AI clusters. Where is your business in terms of the mix of these different kinds of components and where do you see that going?

So our chip business is highly biased towards EMLs. As we've alluded to, the VCSEL business is just starting to grow as there are new products that have been introduced, where we've had EML products for a long time. And I think what you're seeing in the market today is EMLs and then certain silicon photonics are leading the charge at 800 gig. Where VCSELs come into play is ultimately as copper gets replaced by -- or electrical interconnects get replaced by optical interconnects. That's what VCSELs will roll on whereas today at the shortest reach is, there's still a lot of electrical interconnects. So you can kind of look at it that in a cluster today, you probably have some EML-based or silicon photonic-based 800 gig. And then for the shorter distances, you have some probably even a little less optical and a lot of electrical cables. But as you roll forward, then at the shorter distance, we'll go to VCSEL-based and at the higher, longer distances they'll go continue to be EML and silicon photonic based.

And should we expect Lumentum will play in the different flavors?

We expect to play in all the different flavors. And that's one of the customer value propositions we have is we have capabilities to play with all the technologies from a laser standpoint, photonics standpoint, in-house and have leadership positions in them in-house. So we don't have to come and sort of, hey, we think you should buy this. We think it's the best because it's the only thing we have. We can come in and say, "Here's the pro and con of each technology," and choose the right technology together with the customer and then be able to deliver that. Each customer has their own perspective and each -- their data center architectures are a little different. And so we view having the whole portfolio as a strong asset with any individual customer, but as well to address the whole broader space.

So maybe drill down into sort of the artificial intelligence opportunity here in that -- to me, I sort of perceive optics as a way to ride the coattails of the GPUs because I don't have a debate over whether it's Ethernet or InfiniBand, I don't have a debate as to which GPU I'm using. I just need optics. How are you guys thinking about where is that market today? How does it develop through the year? And how do you sort of get your share of it?

Yes. I mean I think it's a fair point or a fair way of looking at it that -- what's happening is that in an AI data center, there's a lot more processor to processor and processor to memory communication given the nature of how AI workloads operate or are dispositioned in the data center. And so that's driving, as you said, a tremendous need for these data interconnects driven largely at the highest speeds and distances by photonics. And today, I would say the overall cloud market outside of perhaps the AI space is still kind of muted and maybe burning off, in some cases, inventory. So most of the growth you're seeing today in the market is being driven by AI. I think it's a smaller portion of the overall market, but will soon become a much more significant portion. So we -- on the announcement of the Cloud Light acquisition, we said that we felt the overall market is growing, let's call it, a 30% CAGR over the next 5 years, probably a little faster than that in the near term as these things typically go. So I think the ability -- I mean, AI is probably going to -- transceivers going into AI double year-over-year, so very, very high growth rate. So our strategy is certainly why we focused on the highest speed, so we have maximum leverage to that growth. And now it's really about both executing the product ramp, the volume ramp at Cloud light has been as well as on our own datacom chips, but as well as taking advantage of the combination. I think not all customers -- I think all customers liked Cloud Lights' products and capabilities that maybe didn't like that they were a private Hong Kong-based company. And the fact that they're now part of Lumentum, a U.S.-based company, we have more vertical integration, created more value proposition. I think there's lots of opportunity to pick up market share by either adding customers or growing within existing customers that we can grow at least as fast as the market, if not faster.

So I want to unpack some of what's going on in your telecom side of products. So I tend to think you've got a couple of different buckets. You've got your transmission products, you've got transceivers. You've even -- you've got a subset of the ZRs. You've got switching devices called ROADMs. Are they all sort of doing the same thing or are there nuanced cycles within there?

Yes, I would say that there's a little bit of nuance that today, current revenue shipments -- our products we call transmission products that sort of transmit and receive data. We call transport products, the switching or optical amplification products that really move data from point A to point B, but the transmission parts are on the opposite end, so like transmitting and receiving. We're probably a little more rich in the transport products today for 2 reasons than what I would call a normal run rate. Normally, I would say the transmission product should be 60-plus percent of our overall telecom business, whereas today, maybe they're 55% or something in that zip code. It's not a huge difference, but I think there's 2 things going on. One, we were a little more supply constrained on transport products than we were on transmission products. So perhaps customers built a little more inventory of transmission products. But secondly, we are at a point, and I think this -- you have spoken with and will speak more with some of our customers, there's a lot of new infrastructure going out in network deployments. So a lot more ROADMs, amplifiers, fiber being lit up, and then as sort of a leading indicator in our words of what's to come, of filling in more of the transmission products over time. So I think you've got both of those factors weighing into our current product mix. And then as we look out over the next 1, 2 years, you're going to obviously see nearer term recovery in the business and then a lot of transmission gear going out to complement or fill up the channels that have been shipped from the transport gear that went out.

And you guys have talked a little bit about doing more vertical integration, getting involved more silicon, things like DSPs. Where are you on that effort? And what's the vision?

So within our telecom business, our telecom products, I would really draw a distinction between -- there's sort of very high end, long-haul, submarine, subsea applications there. We tend to supply very high-performance components to network equipment manufacturer customers and partners where they build the latest and greatest very high-performance systems. There's another portion, which is more of the metro and maybe even, I'd argue a third bucket these days as you start pushing to the edge access side of things, where it's a little more cost-sensitive there and a little higher volume. So we've made a lot of investments organically and for M&A to bring in capabilities to serve those markets in a better way, one of which was with NeoPhotonics bringing in a silicon photonics platform for the ZR modules. With the ZR module in place, we already had tunable laser. NeoPhotonics brought more tunable laser. The other missing piece was bringing in electronics. NeoPhotonics had RF drivers and some TIA. So that's a great complement as well. And then we acquired some capabilities to produce coherent DSPs. We're in the process of developing that 400-gig DSP to serve a market that we think will be very long-lived and a high-volume market where 400 gig is a speed that will live on for a while. It's not to be first to market with this -- the next latest and greatest speed, but to really have a cost reduction pathway to be competitive over a long period of time, the metro. And then as particularly as you go to the access edge portion, that's an area where maybe this is a little overly simplistic, but the market has come to us, taking technology at 10 gig, now going to 25 gig that we've been selling for 10 years is now being adopted at the edge, WDM technology at high speeds. That eventually will go coherent. And so having more vertical integration is crucial to being able to serve those customer needs at that very high volume but cost-sensitive portion of the market.

This conversation were a few years ago, we would have started with 3D. Where is that 3D business today? And sort of what do you see your share? And is it sort of just slowly going to melt away? Or are there things out there we should be looking for?

Yes. So I think for the last several years, we had highlighted -- I mean that business is today and frankly, the market driven largely by one customer for one kind of application in the smartphone handset space. And we did -- let's say, had outsized position in that market for several years, but had long forewarned that ultimately, some competition will catch up and market share would normalize and that has happened over the last 1 or 2 years. And so I wouldn't characterize that specific customer opportunity as a growth business going forward. There's some upside we could pick up on a product transition. There's always competition. So there's maybe some downside that I would say has kind of normalized to a steady state, if you will. But coming off of that, there still are other applications, both automotive LiDAR, augmented and virtual reality. There are other consumer handset vendors that continue to flirt with using the LiDAR for world-facing and face ID. So we do have a slow or a modest growing business that still today is probably lost in the noise of what's going on with our largest customer. But we do expect, as we look to maybe a year out, 2 years out where LiDAR starts to inflect a little bit more. In augmented and virtual reality, hasn't adopted a lot of optical sensing to date. I mean there is units going out, but they tend to not have optical sensing. So that will cut in as well. So I think we will start to see that business grow as we go out another year or 2.

So we're just a bit over time. I want to ask you to close. What do you think is the least appreciated aspect of the Lumentum story?

I think it's really 2 things. First is how critical we are to -- whether it's cloud or networking infrastructure and how much the industry has changed with regards to -- there being both a just a finite set of competitors and how hard the technology really is. And we have upcoming technology inflections that are required to be able to continue that bandwidth growth. And I think that puts -- as you were talking about, hey, it's just to interconnect GPU to GPU, but photonics are really pushing out into lots of new applications and becoming much more relevant for those applications. So I think that provides a lot more long-term tailwind for our business than it is appreciated.

I agree. Well, great. Well, thanks, everybody, for joining. Chris, thank you.

Thank you.