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

Welcome, everybody. I'll read the disclosures first. For any disclosures, please see Morgan Stanley website at morganstanley.com/researchdisclosures or speak to your sales rep. That's it. I'm Meta Marshall. I cover the networking space here at Morgan Stanley. We're delighted to have Lumentum here today, Wajid Ali, CFO. The optical transceiver space is all of a sudden -- or not all of a sudden, it's come back in vogue for a little bit right now.

We hope for a while.

Exactly, for a while. And so excited to have you here today. Lumentum's story has undergone a number of transformations over the past couple of years from China build-out to 3D sensing, to consolidation synergies and now kind of, as we just alluded to, to the Datacom AI opportunity with the acquisition of Cloud Light.

What are you -- would you say are the keys to understanding the Lumentum investment opportunity today?

Yes. I mean when you take a look at all the different products we're investing in, and not just the products but also a lot of where our capital investments are going, we're really trying to be a great partner to hyperscale customers as well as to NEMs to help them solve their toughest problems that are within the data center and outside of the data center. And for the last number of years, what we've been doing is building out the infrastructure and also gaining a lot of the capabilities, whether they're just very basic capabilities that we refer to now with our indium phosphide capability or our capabilities at our Sagamihara facility for Datacom EMLs. But really, what we've been doing all along is building out that capability so that we can be a very key partner to many of the customers that are working with us now to -- for us to be relevant, both inside the data center and outside the data center. And so what you mentioned around some of the -- I would almost call them external events that have kind of happened while we've been flushing out our strategy and executing to it, whether it's the synergies that are related to NeoPhotonics and us really gaining value out of that asset that we had purchased or whether it's some of the geopolitical issues that we, as an industry, have been dealing with as it relates to shipping into China and things of that nature. So I almost view those as part of the process as we've been going out to execute on our business strategy. And as we sit here today from an investment standpoint and you take a look at some of the things that we've talked to our shareholders about and all of our stakeholders about is how we think we're going to win with 1.6T from a Datacom transceiver standpoint, how we're sampling our 800G ZR or ZR+ products to be ahead of the curve either with a laser or with a module in order to take advantage of that data center interconnect opportunity, our 200G EMLs, which we believe we're going to be first to market on and be shipping at the end of this calendar year, and what that will really do for our average selling prices with EMLs and how that will enable -- we believe it will enable the transition to 1.6T and also to take advantage of any AOC opportunities that come about with our 100G VCSELs that are also expected to intercept that opportunity later on in this calendar year. And so as we talk about that and you kind of step back and you say, hey, here's a company that's been building on all these market opportunities and has been lining up the product road maps to be able to succeed from a data center standpoint. That's really the investment thesis for Lumentum.

Okay. So -- over the past few years, you've kind of narrowed your Datacom focus to Datacom chips, where you saw kind of the most lucrative opportunity, and you have actually exited the full transceiver market. You recently reentered that with Cloud Light. So what was it about the asset or just kind of the size of this opportunity that made you kind of decide to jump back into the full transceiver market?

Yes. I mean I think there's -- we saw a real opportunity from an overall market standpoint and where the market was going with enterprise-level customers. And I don't think that, that opportunity with enterprise-level customers was there many years ago. There's also the fact that Cloud Light brings with it a unique capability from a design and R&D standpoint that I actually don't think that we had many years ago. And then you combine that with the acquisitions we've done in NeoPhotonics and gaining some CW laser technology and the opportunity for vertical integration with EMLs, it started to look like a whole unique value proposition. And I think that one thing that's come about over the last 3 or 4 years ever since we exited that part of the business is really the importance of being a U.S.-headquartered partner to Datacom transceivers. And there's very few of us. And so that's really the reason we went back in and said, hey, we think that this is the right time to reaccelerate in this part of the market.

Okay. And not to get too nuanced for a general audience here, but there have been kind of competing laser technologies or kind of ways to view the Datacom market and where the growth would be. In general, I think maybe 6, 9 months ago, people viewed EMLs as maybe for too long of reach opportunities, they weren't going to kind of see the early growth in AI and data center use cases. So what is it that's changed about the market for most estimates to kind of have seemed like EMLs are going to see much more growth with AI transceivers?

Yes. I think the biggest thing is, is that 200G EMLs create a pathway to 1.6T. And 800G products, you could do with 8x100G EMLs. But we're seeing a transition even with an 800G to move to 4 by 200s. And that's really kind of -- I don't want to call it a pipe cleaner, but it's almost a path to prove that 200G can actually be successful. Now it could be successful using silicon photonics, and it can be successful using EMLs. Obviously, we -- our thesis is that there's a lot of power savings that go along with moving to 200G, maybe not half of what 8 100Gs, but certainly a material amount. And when you take a look at a lot of the technology innovations that are occurring, it's not just about speed, but it's about speed at the lowest cost per bit, at the lowest power per bit. And so how do you achieve that perfect triangle almost or trifecta? So that's really what -- that's really why we see 200G and EMLs doing well as we move forward.

And so can you maybe just give kind of a brief -- and not to bring you into a CTO versus CFO position. But...

I would not do -- well, I have Kathy here [ for that ].

Yes, exactly. We might have to bring Kathy onstage. But can you just give a sense for where you would be using EMLs versus VCSELs versus silicon photonics and how people should think of those different use cases?

Yes. No, I think that -- I think there's a market for each. And historically, Cloud Light did have a very good portion of its business that was AOCs. Now they were using a different VCSEL than ours. And so there's an opportunity, like I mentioned earlier, to intercept that for really, I would say, a very basic type of data center interconnect opportunities. And that's really kind of where you'll see a market opportunity for that. And then for silicon photonics versus EML, I think that there is still opportunity for both types of design architectures. It's very customer dependent. Every customer wants to design their own data center architecture and almost think through which solution is better for them. And the customer interactions that we've had, as the Cloud Light opportunity has become ours or has become part of our business, has revolved equally around both silicon photonics and its power saving abilities versus really 200G providing the highest level of performance and providing a lot of wavelength stability. And so as I look at it, I said, okay, okay, if we want to be a part of this business, then -- and we want to succeed, then our ability to offer solutions across all 3 of those platforms and be able to customize around the design of our customers, that's an important capability to have. And so that's really what we're focusing on.

Okay. So maybe VCSEL is kind of the shorter reach, most basic and then kind of depending on customers, silicon photonics versus EML for anything kind of a little bit more...

Yes. We're -- I would almost say we're agnostic about it. And although our preference -- because of our ability to ship 200G EMLs, those EMLs, and that's the greatest opportunity, but we're agnostic at some level.

Okay. And maybe that leads into kind of the conversation of obviously, Cloud Light uses -- you're not vertically integrated with Cloud Light today. You just closed that acquisition. What are the opportunities to kind of vertically integrate that business?

I think the biggest opportunity that's happened with Lumentum joining forces with Cloud Light is our ability to gain the trust of some very large customers and for them to have the confidence that there is a Silicon Valley-based partner that has expansive manufacturing capabilities in Thailand. And ones that we've proven, we can build out and execute too for NEMs customers, and we can apply that same type of model for their demands and for their business. And so there's a lot of cost vertical integration opportunities. And usually, we'll talk about, hey, we can take our EMLs and put that into the design of a transceiver. We can take our CW laser and put that into the design of a transceiver or our VCSELs into an AOC. But I think that, that's actually missing the big opportunity. The big opportunity is Lumentum being a company with a global footprint for manufacturing and being able to produce outside of China in a meaningful way through its access to Thailand manufacturing.

Okay. And then maybe you can just speak to when you bought Cloud Light, it had one major customer. They're obviously kind of going -- not obviously, but they're going to be going through kind of a design transition that you guys will have to work through at the beginning of this year. But can you just talk about, one, that transition and just the impact to your business? And two, as you just mentioned, you're kind of broadening out that customer base. And so kind of what are the offsets in terms of that digestion with other opportunities to expand the customer base?

Yes. And I mean if you step back and you say, okay, well, why is that transition happening? Well, the transition is happening because that large customer wants to move to designs that have lower power. And so it has nothing to do with us or with Cloud Light, and it's really the desire to move to lower power. And so we're going to be moving from 1 800G flavor to actually 2 800G flavors with that customer. And I think that, that digestion will then be followed on with sampling of 1.6T product by the end of the summer and then hopefully, some production volumes of 1.6T kind of as soon as you get through the whole sampling process and things like that. But I think what you're seeing is you're seeing that transition happen for good reason, really for power reasons and then how you kind of have the offset with that particular customer is with 1.6T and then after that, diversifying to other customers that are also evaluating our ability to meet their demands with 1.6T as well.

Okay. And then just broadening that customer base from one customer, just how long does it take to kind of qualify other customers? Just when do you think you guys will kind of have more customers?

I think from a production standpoint on 1.6T with our current customer, it's probably a end of calendar '24 event. And even if we had a design in with those other customers, we probably wouldn't be able to announce it. But a reasonable assumption would be the first part of calendar '25, if we were able to kind of get over all the technical and qualification hurdles in a timely way.

Okay. And I mean there's a lot of enthusiasm about 800 gig and 1.6 terabit, and you guys are clearly leading on the 200-gig EMLs. Does -- what are other bottlenecks that we should be thinking of, of what is kind of the inhibitors of some of the transition?

Yes. No, that's a great question because it's not all about us, right? It's -- the supply chain ecosystem has to work alongside us in order for us to be able to execute on that ramp. So whether it's the DSP or whether it's the switching technology, and I'm not going to mention vendor names, but I think people know who they are. And those partners within that supply chain ecosystem also have to kind of come along. They have every economic incentive to do so. And I think that our end customers are also involved in ensuring that there is alignment to make sure that, that happens. But yes, there are things that are not within Lumentum's kind of execution abilities for that to happen.

Assuring alignment is a nice way of saying -- heavy arm to it, I think, is what you meant to say. Okay. And so just in terms of 1.6 terabit, just what is your thinking just in terms of when you're thinking that ecosystem kind of comes together?

Yes. I mean our internal plans are for the first part of calendar '25. And so we're marching along from a CapEx standpoint, from a planning standpoint. We've got our operations team. I think our operations leaders were there 2 weeks ago in Thailand, working through the space planning of our various buildings to say, okay, if we have one customer, what does it look like? If we have a second customer, what does that space planning look like? And what does that mean from a tools investment standpoint as well as from a layout standpoint? So there is a team that's currently working on that internally at Lumentum.

Okay. And then at the same time, as kind of all this was happening, you were also kind of going through a digestion period on your Datacom chip business that you had with kind of a different end customer or a different customer set. Just how have kind of that -- how has that business evolved post-Cloud Light or even post kind of the inventory digestion we're going through? It seems like we're back on an upswing.

Yes. I mean the end demand for those 100G EMLs has been driven by the 800G transceiver opportunities across the various hyperscalers, right? And so that's gone through its normal price curve, I'd say. And so that's why I think it's really critical that our ability to ramp, our customers' ability to call on 200G is very important because although that will give a better cost per bit for our customers, it will raise average selling prices for us. And then it can kind of reset the tone of what that Datacom business is. But really, that end customer demand is being driven by 800G transceivers, which has created pull for us. In the last couple of quarters, we've had -- we've been on allocation because as we went through a bit of a downturn at the beginning of calendar '23, we had reduced costs quite significantly there. And then when the demand came back, we had to go through the training and bringing back some direct labor for us to ship. But we've -- ever since June, we've been shipping more the sequential quarter than we did the quarter prior on the Datacom side.

Okay. As we move to faster speeds within the network, there's kind of been more of a discussion of either linear drive optics or co-packaged optics or basically kind of eliminating the traditional transceiver structure. Just how would you explain how you see participating in some of these opportunities? And what are some of the timing you see for some of these opportunities?

I mean -- so I mean, linear drive optics is doing very well right now. And so we know of the parties and the current ecosystem that have been benefiting from linear drive as it's happening right now. I'd say that, that is a lower value add from a partnership standpoint for a company like Lumentum because we really want to be part of the customization that's needed in order to deliver the final solution, and linear optics doesn't really give you that type of opportunity to do so. Co-packaging optics, I think you do need our CTO here to talk through that, but it's something that I've been hearing about for years on the business. I'm sure it will be a topic of discussion at OFC. We're almost agnostic. I think that if there's co-packaging opportunities, we're certainly there to help build customized solutions for our customers. Our thesis right now just based on what we've been investing in is that 1.6T with EMLs and 1.6T with silicon photonics are really where the market opportunity is, and it's where we can add value, and so therefore, we can make more margin than we would on a linear optics type of product. So that's kind of our thinking around those things.

Okay. The CTO will answer [indiscernible]...

Yes. Yes, we can...

The rest of that question.

The question, yes.

That's helpful. So with the market moving so fast from 800 gig to different generations of 800 gig to 1.6 terabit, how do you think the traditional cycle dynamics of Datacom either apply or don't apply to the cycle? Because I think for those of us who have followed the space for a while, there's kind of been this -- you get really great margins for a while and then everybody ramps and the margins come down and you just kind of rinse and repeat with these cycles. But are we getting to a point where these cycles are so fast and the opportunities are so big?

Yes. I will be the first to admit that the 800G to 1.6T transition seems to be happening a lot faster than even I thought. And again, I'm sure internally at Lumentum, we're aware of it, but I was a little bit surprised by how fast it was because the 400G to 800G transition took 4 to 5 years to happen. And 800G to 1.6T feels like it's under 24 months. And so being first to market or we believe we're going to be first to market with a 1.6T product, I think, really gives us a lead over the competition and makes us a really important partner as part of the overall ecosystem for delivering 1.6T. So I think this is one of those times that being fast to market and the market's changing so quickly is actually going to be to our benefit.

Okay. So we don't have enough time at 800 gig to see that margin digestion as we might have kind of in a traditional cycle. Okay. I'm going to jump into telco, but I just want to ask if there's any questions maybe on the Datacom side of the business before -- I think it's a good point to initially stop for questions.

Just on Datacom EMLs, I think that business maybe peaked at $200 million of annualized revenue at some point in the past before this kind of digestion. Can you just level set for us what applications in the data center were those chips really being used for? Because it would be helpful to understand how that continues to recover and then we kind of layer in this 200-gig EML ramp to 1.6T and beyond.

Yes. I mean the -- great question. So I think the primary application that was being used for was 800G transceivers. And so we were shipping in to all the known partners 100G EML chips for those 800G transceiver opportunities that customers were shipping to. And so that was really where it was. On 200G, the transition, some of it will be driven by 4 by 200G architectures or designs. And then most of it, we believe, will be coming about when we start shipping 8x200G. That's the way we're thinking about it.

Got it. Okay. We'll jump into the telecom side of the business. That business has undergone inventory digestion headwinds both on just telco and optical systems side. Just how do you think about recovery there? And will it lag systems vendors because we have to get through their inventory as well? Or just kind of how do you see that timing of when you could start to see demand come back there?

Yes. I mean it feels like the inventory digestion has gone on longer than we expected. And it's a combination of things. One is I think that the demand during the COVID-era period was exasperated by a shortage of supply just generally in the market. And I think that our partners also wanted to make sure that they had enough. And then I think that there's been some end customer demand drop-off that wasn't expected from our NEMs customers or their customers. And so the combination of those 2 has exasperated the amount of time that this inventory correction has taken place. Our thinking up until now has been that it would be a midyear -- mid-calendar year '24 time frame where we should start to see a bit of an uptick. And as we're getting closer to that time, I think we'll get more information around it. But it's still a little bit hazy, I'd say. And whether it's the middle part of this year or whether it's in the third quarter of this calendar year is still a little bit unknown. But we are -- we have been kind of floating at the bottom for a couple of quarters now as it relates to telecom inventory.

Okay. And just how does that telco timing impact or recovery timing impact attainment of NeoPhotonics synergies, which you were kind of in the midst of?

Yes. No. So I think the NeoPhotonics synergies have gone quite well. When we had originally acquired NeoPhotonics, we had said that we thought we'd find $60 million of synergies. A year later, last year, we increased that to $80 million. And at the most recent conference call -- earnings call we had, we raised that again to $100 million. And really, there's been kind of 3 components to it. One has been some of the public company costs and the operating expense savings from really getting onto one platform from a product standpoint. The second part that took a little bit longer is -- has been the manufacturing consolidation. And so just in December, we finished the consolidation of 2 of the manufacturing facilities in NeoPhotonics. And so we should start to see some synergy benefits from that. Then as that's occurred, we've seen some incremental operating expense synergies. But then the last piece is really vertical integration opportunities with things like vertically integrating our RFICs and our TIAs, and that's going to take a little bit longer. But I think the bulk of it, I think 2/3 of our total synergies we had achieved approximately plus or minus by the end of December. And we should be able to get through the majority of the remaining by the end of this calendar year. And then what will be really left is some of the vertical integration opportunities. So to your question, how does revenues being down impact us, I think that it impacts us in 2 ways. One is once the vertical integration opportunities happen, then we got to take a look at what volumes we're able to achieve on those. And then secondly, some of the manufacturing savings that we've been able to achieve have been masked by our overall underutilization at a company level. And so as revenue comes back, we should be able to see the benefit of that.

Okay. Maybe moving on to the industrial tech business. I would say in my conversations with investors, this is probably like the black hole of their understanding. You've seen inventory consumption at a large customer and kind of continued macro weakness. Just -- what gives you confidence on that industrial laser inventory should be corrected over the next 6 months? And just traction you're seeing with kind of newer products, newer customers?

Yes. I mean fortunately for us, our ultrafast products have done quite well. And so they've been offsetting some of the inventory corrections we've seen. We're in close contact with that large customer that's within our lasers business. They've been going through some inventory management themselves as there's year-end and things like that. And so the forecasted demand that they're seeing really gives us some ability to say, okay, later on in calendar '24, things should normalize. We saw this happen back in 2019 as well, where we had a couple of really good years, and then we had a correction for a 1-year period of time, and then things normalized again. And so that's our thinking around it. And hopefully, continued ultrafast laser success provides some buoyancy to that recovery.

And just in terms of kind of broadening out that customer base, I know you've kind of made recent traction, but just -- is there any -- like how should we think about the broadening out of that customer base?

Yes. I think the broadening out of that customer base has happened, but it's been slow. We've added new customers as our ultrafast lasers have come on board. But yes, it's been slower than we'd like, certainly.

Okay. Okay. I'm going to jump around here to a couple of different segments. 3D sensing, much smaller portion of revenue at this point, largely, I would think, derisked. But do you see any further downside risk or offsetting potential opportunities for a rebound in that business?

Yes. I mean all the readings through our R&D projects that we've been working on give us confidence that our end customer wants to have multiple suppliers in that part of the business. It brings about the most innovation and the most opportunity to make sure there's security of supply. So we certainly think that we'll continue to be a part of that business, but it will be at a much lower rate. I think the real opportunity for us is what comes out of some of the AR/VR investments that we've been making. But those are probably a couple of years out in terms of actual financial benefits for us.

And are there opportunities to kind of expand that customer base beyond kind of what has predominantly been kind of a singular customer?

I wouldn't put that into the base case of our investment thesis. But yes, certainly, there is.

Okay. And then you mentioned upfront kind of the ZR opportunity. Obviously, I think you guys made some acquisitions of DSP teams that haven't traditionally kind of been known as having that full stack. Can you just kind of say where you are there and where you see kind of the biggest opportunity there?

Yes. I mean we believe we're first to market with 800G ZR and ZR+ products. And with the 400G ZR+ product, we were only able to sell the laser. And a lot of that had to do with the fact that the ramp of that product happened when NeoPhotonics was stand-alone. But as it's become part of Lumentum, the confidence around being able to build full modules for 800G ZR has increased substantially. And so there's an opportunity for us to be first to market. We're already sampling 800G ZR products. And so there's a real opportunity for us to be first to market there, but also the opportunity to actually sell the full module rather than just the laser. And so I think that customers need to decide how quickly they want to move to that speed. But certainly, I think as soon as one customer does, others will follow along pretty quickly.

And largely cloud customers.

Yes. Thank you. Yes, largely cloud customers.

Okay. All right. Perfect. So let's get into your sweet spot of financial questions. One of your priorities has been to lower fixed cost base to help margins as the top line recovers. What have been some of these initiatives to lower those fixed costs? And just where are we thinking of where they primarily are focused?

Yes. No. So I think some of the manufacturing synergies that I spoke about earlier, that's been the first place that we've been focused because that will really -- not only just give us a lower fixed cost base, but also give us more nimbleness in our ability to service our customers. So I think that, that's really important for us to know where our products are going to be shipping from rather than have 3 or 4 manufacturing facilities. So that's an important part of our fixed cost base. The other is some of the advantages that we've had with the IPG acquisition. We've got a great team down in Brazil, and that's really allowed us to expand our capacity from an R&D standpoint to take on new R&D projects that might not make sense with a Silicon Valley-based R&D team, but more with a lower cost but very capable team in Brazil. And then the third thing has just generally been our operating expenses. We've been trying to bring that down on a quarterly basis to below $100 million a quarter. We're not -- we're still not there yet. We're still north of that, but we're working through how we can bring that down below $100 million a quarter.

Okay. I mean just given maybe the reduction of the 3D sensing business, some of the offsets that you've seen on demand, integration of NeoPhotonics, integration to Cloud Light, I think there's some misunderstanding of kind of what the margin structure really is of this business at this point. So you just mentioned kind of the $100 million OpEx -- quarterly OpEx from that standpoint. But just I know we'll get to some of those stuff at OFC or other places. But just how do we think about what the margin structure is of this business at this point?

Yes. So I think very much so, especially with Cloud Light coming in, we're very focused in on the operating margins of the business because the gross margins have gone through a little bit of a structural shift, and I agree with you. OFC is probably the right spot to talk about that. But I think kind of step one is for us to get to a double-digit non-GAAP operating profit. And the way to get there the fastest is really for our revenues on a quarterly basis to be north of $400 million a quarter. And so with some of the telecom recovery that we're expecting later this year, introduction of new products, Cloud Light revenues potentially picking up, we should be able to see a path to that. And so right now, our real focus is making sure we can quickly get back to double-digit non-GAAP operating profits for the company on a quarterly basis.

Okay. Any last questions from the audience?

Just on the data center interconnect piece, that 400 gig ZR business grew pretty strongly last quarter. Given the power draw from AI, it seems like there's a lot of demand for new sites and new builds on the data center side. How do you think about DCI intensity and spend over the next few years? Is this potentially a wave of very strong spend for multiple years moving forward?

That's a great question. I think that's really what we're hoping for with our new 800G ZR+ products is to be able to intercept and take advantage of that expected increased spending that's going to happen in the pluggables part of the market. And so yes -- no, I think we're very much aligned with that comment.

So maybe just a last question for me. A lot of initiatives there. How are you thinking from a capital intensity standpoint or just kind of a CapEx standpoint, kind of the spending that needs to take place?

Sure. I mean -- so last year, fiscal '23, we spent about $130 million of CapEx. The CapEx was much more focused in on ensuring we had sufficient capacity at our Sagamihara facility for EMLs and building out our indium phosphide capabilities in Caswell. And so I think that, that envelope of CapEx spend stays the same, at least for fiscal '24, but it will be much more focused on ensuring a revenue ramp for Cloud Light -- for the Cloud Light -- former Cloud Light products. And so the envelope will be similar, but it will be targeted to making sure that we can ramp up at the appropriate time for customers for 1.6T.

Okay. Perfect. Well, I think we did some good work, not being engineers, explaining this, and it's a pleasure having you here today. Thanks.

Thank you so much, Meta. Thank you.