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

Good morning. I'm Samik Chatterjee. I cover the hardware and networking companies at JPMorgan. And for the next session, I have the pleasure of hosting Lumentum and Michael Hurlston from Lumentum, CEO of Lumentum. So Michael, thanks for taking time to come to the conference, and thank you to the audience as well for attending.

Michael, I'll start you off with a few questions we're getting mostly on the macro from investors. And I think this is based on sort of overall hesitation or concerns we see from investors more around the macro than anything else at this point. You obviously get a lot of feedback from your customers in terms of how they're thinking. So how would you characterize sort of the likelihood that we see a significant slowdown in demand later in the year or see potentially a recession? How are you thinking about those sort of risks to the overall business?

No. Thanks for the question, Samik. So first, thanks for having me. Good event. We haven't seen any slowdown at all in ordering patterns. Book-to-bill remains positive. Generally speaking, I just finished around with all the customers. As you know, I'm new to the seat. So I've been spending quite a bit of time with the customers and at least what they're reflecting to us is no slowdown. So we'll see how it plays out. And of course, we're continuing to monitor and modulate our spending. But at least over the course of the next 4 or 5 quarters, we don't see a slowdown in sight.

And I'm going to sort of hit some of these macro topics first. There's also -- and I'm sure you get questions on this all the time, is a lot of concern whether the cloud companies continue to invest in AI to the same extent that they have in 2024, for example -- or in the first half of 2025. How do you sort of -- when you look at your pipeline of engagement with the hyperscalers, how do you get comfort around that magnitude or even sort of a robust spend continuing, which entails then you to sort of be focusing more resources on that front as well?

Yes. Again, as you know, having spent time with Kathy and with the company over a long period of time, we're somewhat indifferent to AI and just data center in general. We don't see any slowdown in CapEx. I think everybody has come out with their CapEx spend for 2026. And it same or better than we saw in 2025. We have opportunity in our landscape, as you know, to pick up share in a couple of areas. Our transceiver share is relatively modest. And we have opportunities out in front of us on OCS and some other things that we can see that give us maybe a unique set of tailwinds. But generally speaking, similar to the first question, we're not seeing any appreciable slowdown in spending.

Okay. And then talking about tariffs specifically and things keep changing every day, but maybe talk about, one, sort of how you're navigating the unpredictability here of tariffs? And then overall, how are you planning to mitigate the impact?

Yes. I mean we characterized on our call, our call was last week, and you and I had a really good discussion following the call, the impact that we characterized was 100 basis points to margin. And that 100 basis points to margin was largely coming from the reverse tariff where we are importing goods to make our different optical goodies in China. So as things come into China from the United States, that reverse tariff was hitting us to the tune of 100 basis points. Interestingly enough, as you know, we have manufacturing footprints, both in China and Thailand, and we didn't see appreciable impact from the tariffs on our ship out. So most of our impact was actually on ship in. Obviously, the situation has gotten slightly positive even since last week, and it's something we continue to monitor. We have a fairly unique situation as a company in that we do have 2 manufacturing locations. We think we can steer around tariff impacts by using that supply chain flexibly. And in events where tariff might be imposed on goods coming out of China, we think we can maneuver into Thailand. And even beyond that, we have good relationships with a set of contract manufacturers where if need be, we could skate our manufacturing and our supply chain to that, where we could actually utilize the contract manufacturers more than we currently do. So we have a lot at our disposal on the manufacturing line that helps us mitigate outgoing impact on tariff.

Okay. Some investors asked us after the investor -- after your earnings call itself about how to think about plan B for Lumentum. I know you're talking about moving capacity to Thailand more and more. And you also said in most of those cases, you're not the importer of record. So you're not the one paying the tariff as such. But if there is a tariff on Thailand as well that's in play eventually after some of this 90-day sort of pause goes away. Is there a plan B in terms of capacity that you need to then address? Or would you just more than focus on contract manufacturers that are not impacted by the same?

Yes. I mean, look, we're -- our first plan is to continue to scale up in Thailand. As you and others that follow the company and know very well, we've enlarged our manufacturing footprint there. We're still not at capacity. We still have room to grow in Thailand, and we would intend to do that as sort of plan A. And even for some of our flows, which you and others have covered well, where we manufacture in China and actually goods flow out of China, we have the opportunity to steer that such that we do finishing work and have the country of origin be Thailand rather than China. Beyond all of that, and that's a lot of moving pieces, we probably uniquely in the optics industry have a strong relationship with contract manufacturers. It wouldn't be overnight. But over the course of a few quarters, we could move more of our manufacturing footprint to the contract manufacturers and have them pick up some of the work in countries that were more tariff favorable, in the event that a significant tariff, for example, was applied to Thailand.

Got it. Okay. Okay. Great. So now going back to more sort of company-specific drivers, but maybe we'll start with your sort of experience at the new seat. You were in the optical industry before going to Synaptics and you've come back to the optical industry. Most investors I talk to are always concerned about the historical trends of the optical industry where if you take the industry in aggregate, like finding a period of sustained revenue growth and revenue growth at healthy margins is pretty few. You don't really find that over multiple time periods that the optical industry has delivered that. So one, do you see the industry as something that's changed from the -- what the historical trends were in the prior years? And what drove -- what was the motivation for you to then come back to the optical industry and try to change that overall that do you see a period of healthy growth and healthy margins at the same time?

Good complex question there. You've done your work. Look, maybe I'm crazy to come back to the optics industry. I think that that's the headline. But I think 3 things have changed. I mean we sold Finisar to II-VI, which has become Coherent. And now, of course, the majority of the revenue coming from Coherent actually is from Finisar. So I understand the business over there pretty well. I think 3 things are different. First, Lumentum has a much higher component mix. The thing that I like about Lumentum is the components. And Kathy and others have characterized our laser-footprint as being second to none. As we look out at co-packaged optics, opportunities like scale up, it's going to be very component-centric. So having that strong foundation, I think, is a differentiator and something very attractive. But to your question, I think 2 things are now different about the industry that were drivers when I ended up selling Finisar and are now sort of attractive forces. One is globalization, right? So that was not a thing in 2018 and 2019. There was an equal playing field in Broadcom with Hock, Avago originally, Lumentum and Coherent were all major transceiver suppliers. Lumentum actually dropped out as did Broadcom. So Coherent and Finisar were the last people standing, but we could see a lot of pressure coming from the Chinese. And Alan Lowe is a smart guy and Hock is obviously a smart guy, and I felt probably they have -- they're reading the signals right, time to withdraw. That's changed. I mean I think there's a big push from the hyperscalers to source from U.S. names, even though the Chinese names now have moved their manufacturing footprint to other jurisdictions for security concerns, both security of supply and actual security, there is a big push for the U.S. hyperscalers to source from the U.S. And then the second dynamic that's changed, and you've covered it well, is just the rate of change in the market. There's -- the market is moving much more quickly, which speaks to innovation, which is typically a strong suit of U.S. customers and U.S. companies. So as the hyperscalers have become an increasing part of the overall market in optics, the rate of change that's being driven is very, very advantageous to a company that's willing to invest to innovate, which is typically the hallmark of a U.S. company. So dynamics in the 5 years since Finisar are now coming back into the market are appreciably different.

Okay. Fair. What's your early assessment of the Datacom assets that Lumentum has? Even Cloud Light is a relatively newer acquisition. If you were to sort of think about allocating capital today, is the focus going to be more on the transceiver side with Cloud Light? Or is the focus going to be more on the chip business or the component business as you call it?

Near term, obviously, a lot of our growth is going to come from transceivers. We're a small player in the transceiver market. We have low single-digit share in the market. We have a high degree of exposure to one customer, as you know. So our plan certainly in the next 4, 6, 8 quarters is to grow in the transceiver market. I see transceivers as a means to an end. We want to grow our component business, and we see such an opportunity with co-packaged optics coming online sometime in the next year or so, scale up being a big opportunity that's in front of us from a component standpoint. So in a sort of 5-year, 6-year asymptote, I'd like to skew the company much more to components. I think that, that's higher gross margin, higher shareholder returns if we can do that. But in the interim, I'm going to use the transceiver business to build cash flow to allow me to invest to sort of fix the -- clean up the balance sheet and get to a place where we can have an even higher majority of our shipments be components than we see today.

Okay. And on the transceiver side of the business, how do you think about how critical vertical integration is? I mean, particularly given that you have capabilities on both ends in terms of the transceiver assembly as well as components of your own. Do you see that as a prerequisite to sort of drive more vertical integration and deliver shareholder return? Or do you think shareholder returns are independent of driving vertical integration on the transceiver business?

Certainly -- it helps. I mean, I think you and I had a discussion last night over dinner. The transceiver business for us is definitely operating at margin levels right now that we don't like, right? It has a lot of room for growth. But even in the asymptote, I don't see the transceiver business getting anywhere above mid-30s gross margin. And that includes vertical integration. So we would -- today, as you know, we don't use our own components in the transceivers that we ship out. We want to fix that. We want to change that. And so we have a strategy to now put our own components in our own transceivers. But I don't know that, that's a differentiator. InnoLight, Eopto, Hisense, right, all Chinese competitors, nobody -- none of those guys have vertical integration. And frankly, they're doing better than us in the market, both in terms of market share and margin. So we have room to run for sure. I think the vertical integration strategy gives us a level of differentiation and a level of help on the gross margin line that will be appreciated in the asymptote, but I don't think it's a prerequisite to your question. I mean the other guys are doing a fantastic job without having that vertical integration. In fact, as you know, we supply some of the critical components to most of the players.

On the transceiver side, you've been pretty busy or Lumentum has been pretty busy even before you joined in terms of announcing new wins with hyperscalers. Should we largely take that as a reflection of already some of the move from the Chinese suppliers starting to materialize in your wins? Or how would you encourage us to interpret the amount of wins that you've announced recently in relation to Cloud Light with hyperscalers?

Yes. I think it's definitely driven by a desire to have a domestic name. I think between us and Coherent, if everything is running seamlessly, we're still able to supply less than half the market. So although there's a big push to have U.S. suppliers, and I think we're taking share from overseas suppliers, it still is not a situation where we will be able to fulfill 100% of the demand. The demand is still very hot to the first set of questions you asked. There's a lot of runway ahead of us in terms of building share and momentum in the market. And look, we're a very, very small minority player today. But where we're moving the share needle, we're certainly moving it from China, less so from Coherent.

Okay. Got it. Let's talk CPO. Everyone is focused on that transition. How should we think about sort of the drivers in terms of how to think about revenue versus margin? What does that impact on Lumentum look like if the industry were to hypothetically go to CPO sort of very quickly?

Yes. Let me -- maybe the more difficult question first. Certainly, from a revenue perspective for us, and this is probably unique to us among all the suppliers, we would see it being neutral, right? Because the number of lasers that we supply as a transceiver, assuming in a transceiver architecture and then our share of the transceiver market is relative -- is unequal to what we would get at CPO. And CPO, I think we're going to get a very, very high share, certainly from the leading deployer of CPO. And so if you look at it on a revenue line, and Kathy has done the calculations, it's roughly the same, right? But the difference comes in the fact that our share is so high relative to transceivers and relative to lasers that we might supply into third-party transceivers. The margin dollars are considerably higher. So it's a real benefit for us for the world to go CPO, revenue neutral, margin significantly accretive. And we have learned a lot from NVIDIA in terms of how they're thinking about architecture, and we hope to extend that relationship into transceivers into other things that we might do with them from this point forward.

Okay. I know we had this discussion last night, but wanted to get this -- have this discussion in this forum as well. Primarily when you discuss share in transceivers being low and then the offset being the high share you have on the laser side. Maybe just flesh that out a bit more. Where is your market share today? Or what are your aspirations in terms of market share when it comes to the lasers that support the CPO solution?

Yes. I mean, out of the chute, we think we have 100% share. I mean we've already started shipments into the single announced CPO opportunity. We think we can maintain that given the differentiation we have. So I think Kathy pointed this out last night, we are coming from a place of real history in CPO. We are leveraging a laser that we've typically used for submarine products, high power to go transmit along long distances on undersea cable and an unbelievably high reliability. And those 2 things are key factors because as you think about the argument against CPO has often been, look, I have to go in and sort of detach a bunch of things from the motherboard where before I had a pluggable, right? This is -- if we're able to generate a truly highly reliable laser, which we have, we have a long history of that with the underlying technology for this UHP, the laser we're deploying in CPO, and we're able to generate the power levels that are required when you but the optics with the switch, we think we have a very unique differentiator. So I do think it's going to take a long period of time for somebody else to catch up. It's not insurmountable. I mean, if people put their mind to it, but it's not an overnight thing. It's really a long, long investment cycle that we've had in this product. And we think that's going to be a real competitive moat as we look at CPO.

Got it. You outlined your long-term road map is to move more towards components. I mean if technologies like CPO become more relevant, that's moving more revenue dollars to components. I mean, in 10 years' time or even 5 years' time, how does Lumentum look relative to a semiconductor company that's vertically integrated from the design to the fab level?

Yes. I mean, look, I think we've certainly outlined a near-term path to get our margins above 40%. And in that structure, $500 million, going to $600 million, going to $750 million, there's a big element of transceivers that fold in there. So in our calculus, certainly in the very near term and intermediate term, transceivers are going to play a significant role. But I think as you go at for $750 million and beyond, and we'd like to get this thing to a $5 billion-plus revenue company, we are going to tip a lot more to components. So we're going to use cash that's generated to start pursuing opportunities and scale up in different parts of the optics market. We don't have a DSP today, for example. So there's a lot of things we think we can do to inflect our component content such that when you start talking about revenue levels of $1 billion a quarter, we would be much more biased as a company toward components.

Okay. Got it. Going back to sort of the technology level on the component side, you have either design capabilities or even fab capacity when it comes to VCSELs, EMLs, CW lasers. I know you'll be sort of -- it doesn't matter to you where the industry goes because you have capabilities across all 3. But do you have a view on where the overall sort of optical landscape when it comes to inside the data center, which technology sort of gets more adopted relative to the other? What are you hearing from your customers on that front?

I think if you look at the mix now of transceivers, it's definitely biased to CW. So silicon photonics is a decent share, a significant share of the overall transceiver that are deployed. That doesn't mean that our EML capacity, which is completely sold out, is going to change anytime soon. The world supply of EMLs can't keep up with EML demand. But if you look at the landscape right now, there's no doubt that the majority of deployments at 800 and 1.6T are CW. That's why we said, hey, we're going to put our toe in the water with CW. We've not deployed -- we've deployed, but not in this construct, a Datacom transceiver, we've not deployed a CW laser. We would certainly -- look we are planning to do that in the fourth quarter. I think you know that of the calendar year, start shipping production CW lasers. And that's mostly strategic because we would be in-sourcing those CW lasers into our silicon photonics-based transceivers for shipment to our largest customers and such that we can start on this journey. But again, the world supply of EML, we're completely sold out for the foreseeable future. We will bias our manufacturing to the extent we can toward EML as that's a better margin opportunity, a better ASP opportunity for us overall.

Okay. Got it. Let me take a pause there and see if any questions in the room.

There are some contract manufacturers are making -- directly for cloud. That could be a good opportunity for you, but also competition if you're making your own. How do you think that's going to evolve over time? Do you think the clouds will eventually design most of their own and not use third parties and do it with a contract manufacturer and you could sell components? Or do you think that will be a small part because there's a lot of innovation that you guys can bring to the table as we move on these speeds moving up?

It's always going to be a mix. I mean, if you actually look at it, how did we get into the transceivers? Well, we bought a Cloud Light, which was arguably a contract manufacturer for a large customer, right? So this has been a -- this is nothing new. This has been around for a long time, where there's a mix of contract manufacturing opportunities. And as you correctly point out, that's a great opportunity for us to sell components. And then there's a mix of, hey, we're going to go to the open market and source it for somebody who can give -- bring us a more turnkey solution. We have a mix today. We have some things we're doing on spec. We have some things that are -- where we would look somewhat like a contract manufacturer. We have some things that are done as turnkey designs. So it really depends. And to us, that landscape has been there. It will continue to be there. And we feel like we have opportunities in both.

Any other questions in the audience?

You mentioned on the gross margins, I think did you say mid-30s for transceivers is probably the right place to be thinking about because that's where the guys like InnoLight are and then you're going to bring in some of your own vertical integration. But even with that, that's probably the right level to think about.

Yes. I think -- I mean, today, we're behind that, right? We're not at -- we're unfortunately nowhere near that. So we have a lot of room to run. I mean if you look at Coherent, I think they're sort of mid-30s, perhaps high 30s, 35%, 36%. InnoLight actually reportedly is in the 40s, but they think they have different ways of accounting for depreciation and some other things that give them an artificial tailwind. We would expect to be low to mid-30s. I think somewhere 33%, 35% is going to be where we end up. And so our job is to manage the rest of the business to gross margin levels significantly higher than that. We think we can run the rest of the business even as this is growing at margins much higher than 40% to bring our overall average to 40% and beyond. So we have a lot of levers, we think, in the business to continue to improve and work on gross margin in the rest of the business. Components are obviously at semiconductor type margins, but there are other parts of our business that are operating at levels we think we can improve to get our -- the gross margin line moving above 40% in a pretty rapid fashion.

Okay. So if I continue, we haven't discussed OCS yet. So maybe help us think about the revenue opportunity that you're sort of envisioning around OCS. And you mentioned product shipping later this year, I think. But when do we start to see that become material where it sort of is more noticeable to the P&L?

Yes. Look, this is a great opportunity for us. I mean, we think greenfield data centers are transitioning very clearly to optical at the spine layer and in some cases, in the leaf layer. So we won't supplant all the electrical switches. I think we've given out numbers like 10% to 15% of the opportunity that could go to optical. It makes sense because the more transitions you have from electrical to optical, the more power you end up consuming. And every one of those transitions power is an issue. OCS has advantages in cost. It has advantages in power. It has advantages in latency. And so I think if you start looking at, again, greenfield deployments, and there are many now coming online, those are strongly considering OCS as an alternative to electrical for a number of the layers inside the data center. So the way we see it playing out is we think we'll kind of get early, early revenue toward the end of the calendar year and then really material revenue starting to build probably in the second quarter of fiscal -- of 2026, calendar 2026 and then really starting to impact our fiscal 2027. So we really like this. I mean you think about the ASPs, you're talking somewhere between $75,000 and $90,000. So you don't have to sell too many of these things to really impact the bottom line, the top and bottom line. And they are -- for us, they're accretive even to these high-margin targets we've set out. So if we can make a dent in this market with the OCS, we will really be able to move the needle.

Got it. Okay. Margins on OCS?

Yes, it's above even the targets we've set out. So very accretive on the margin line on OCS.

Telecom, seeing quite a bit of revival recently. How much of that is traditional telecom equipment versus what is more purpose for ZR or DCI applications?

Yes. I mean it's fascinating. I mean, really, I would argue that the build-out of -- from the hyperscalers is actually trailing the in data center. So the majority, Samik, of our revenue is actually -- although we would sell it to traditional telecom people like Nokia, Cisco, Ciena, are traditionally strong customers, they, in turn, are reselling to the hyperscalers. And if you look at the dollars that are being deployed, there's been so much to build out the data center, but very little to connect, to build backbone to really -- to facilitate DCI -- and that spend is starting to pick up. We certainly see it from the traditional telecom guys. But it's not the AT&T's, the Verizon's, the traditional telcos that are driving that. It's much more the hyperscalers that are causing our revenue line to move.

Okay. Any views on what the sort of the underlying telco spend is on their own network because that obviously used to be a big driver of revenue for Lumentum in the past. Like is that starting to continue to moderate? Or is that now just stable at this point?

I would say, generally speaking, as you know, having been around the company, there's a big inventory problem, and that was mostly coming from the traditional telco and MSO type of customers. I think we're largely through the inventory. There's still pockets, but that business has yet to fully revive. So it's flattish, right? And there's pockets where we're up, pockets where we're down. But if you look at telecom in general, it's becoming more and more driven by the hyperscalers rather than the traditional carriers.

Okay. Got it. Maybe lastly, just on Industrial Tech. outline sort of what -- where do you see opportunities in both industrial applications, but also vis-a-vis that like the traditional face ID business that you had. Do you see opportunities on either of those that would be material in terms of an incremental opportunity?

Yes. I mean face ID has been -- I played it on both sides, right? As you know, I was built the face ID factory for Finisar that now Coherent has taken on. I think that story has mostly played out. I think we're -- we went through a boom. And now I think the market is relatively flat line. I think you've got rational behavior in the market, but I don't expect to move the needle all that much. I think there's some opportunity for us as a company to improve profitability. We have probably too many engineers deployed on it given the size of the opportunity. So I think we'll look at how we can rationalize that and make it a bit more profitable. The other component that you correctly talked to is industrial lasers. And again, I think there's opportunity there to grow the business. We really are driven largely by one customer, particularly on our fiber laser side. We think that as the semiconductor and semiconductor packaging industry grows and is looking more and more laser deployment, there's opportunity to do better there. And there's definitely opportunity to improve the margins, right? As that business has sort of traded sideways over the last few quarters, we have not done a lot either on our cost or on our prices, and I think there's some opportunity to improve the business by looking at those kind of levers.

Okay. I will wrap it up there before we run out of time. So thank you, everyone, for coming to the conference. Thank you, Michael.

Yes. Thank you, Samik. Good seeing you.