I'm Meta Marshall. I head up networking here at Morgan Stanley. We're delighted to have Lumentum here with Chris Coldren, SVP, Chief Strategy and Corporate Development Officer. And we kind of always are adding new titles to you, Chris. So pleasure to have you here. Lumentum has undergone a number of changes over the past few years. How does the company look at the market opportunity and ways in which it can participate in the multiple avenues of growth after Cloud Light, NeoPhotonics, and kind of the AI explosion?
Yeah. Well, thanks, Meta, and thanks for everybody attending. That's been a great conference so far. A lot of people eager and interested in talking about what's going on. In terms of what we've tried to do at Lumentum over the last, frankly, 10 years plus, is really make sure that we have foundational photonic technologies that are well aligned with where we think the overall world is going over the long run. And so that started with internet infrastructure, as well as some industrial and consumer applications. And now, as you've highlighted, with the AI explosion, is driving tremendous demand for the kinds of technologies that we provide, and in particular, many of our products.
So as we look ahead, I think what's really exciting is that what we're seeing now is really only the tip of the iceberg because there are multiple industry technology transitions ahead, particularly in the cloud and data center space, where photonics and our type of photonics will be used much more so. For example, in the coming year, our indium phosphide lasers that we make, market leader in those, they played only in a certain set of the data center applications. Now they're going to play in a much larger set in the 1.6T generation. As you build out all these data centers and put in a massive amount of equipment inside of them, there's a lot of infrastructure that needs to go in connecting the data centers together. So that's driving the outlook in the next 10 years of our telecom product business.
Eventually, photonics is going to replace copper and copper interconnects in data centers. The volume opportunity there is an order of magnitude larger than the things that we're doing today. So all of these things provide significant growth opportunities for our technologies, not only in the next year, but in the years to come.
Got it. You've mentioned a lot of opportunities, a lot of different ways in which you're participating in those opportunities. You've just had a CEO transition. Just kind of why was this the time and how does Mike best position you for success as you try to ramp kind of to all of these opportunities?
Sure. Sure. So I think Michael Hurlston, new CEO, came into the company a couple of weeks ago. And his experience is very deep in the semiconductor space. And the semiconductor space is becoming increasingly intertwined with the photonic space, where I guess we're becoming more of the compute and on-the-computer roadmap. And photonics are used to interconnect semiconductor chips. So his experience is something that we have not had historically within the company. So he brings a very interesting and new perspective and experience that we can benefit from. And at the same time, he has spent time in the photonics industry, so we can straddle both worlds. So given the opportunities ahead, it's an opportune time given as we become much more part of the electrical and compute roadmap going forward.
OK. Got it. So we've talked about some of these opportunities ahead of you. You're ramping Cloud Light. You're ramping EMLs. You're ramping ZRs. You're vertically integrating Cloud Light. Just in addition to just kind of ramping the traditional industrial lasers business, just how are you, as Chief Strategy Officer, kind of figuring out what to focus on?
Yeah. I think that the most important guiding principle is listening to the customer, listening to the market, that our technologies are so integral to our customer roadmaps that we really need to make sure that we are very well aligned with what they need, not just today from a product and manufacturing capacity standpoint, but at those forward-looking product roadmaps and working very closely together with them to ensure that we are both aligned on what will work or what will make sense for us in the industry. That's the overarching principle.
OK. And so maybe starting with EMLs, you noted on the fiscal Q2 call that despite kind of these additions to capacity you're going to have throughout the year, you expected it to be in a state of capacity constraint through the end of calendar year 2025. How has the customer base expanded here over the past year? And just kind of how much visibility do you have into that demand as you invest in kind of expanding capacity?
Yeah. Yeah, sure. We supply, I would say, at this point, to just about anybody who's building an optical transceiver, and particularly the guys making EML-based transceivers. And that's really a recognition of we've historically always been focused on the high end of the market, sort of highest performance, almost a premium product, if you will. And AI has really pushed the volume and demand at the high end of the market, which was a little bit different than the traditional data centers, where maybe the high end was initially where things were, but the bulk of the market was at something that was a few years old. So I think that having that indispensability to our customers and ultimately the end customers gives us a level of visibility where they need to work with us to ensure we're planning adequate manufacturing capacity over the next few years.
Even at some point, the end customers are also involved either purchasing product directly or helping us decide where product should ultimately land in which supplier so that they can ensure they get what they need versus having our direct customers maybe lead us in one direction or another on demand. Visibility, I think, is as good as it's ever been in that we are talking to the end customers and have an understanding of what they need. At the same time, now being also in the transceiver business, that gives us a different view or a different seat at the table to hear what the actual transceiver opportunities are and what those volumes are and can correlate and triangulate with what we're hearing about the chip needs for other people that are serving those applications.
OK. Would you expect kind of EML pricing to remain attractive throughout it being capacity constrained? Or are you seeing kind of normal contracts being put into place where, yes, we'd like you to continue to invest in capacity, but we'd like to kind of enter into these multi-year contracts?
Yeah. I mean, I think obviously when demand exceeds supply, that's favorable in general to pricing. I would say that given our product differentiation and performance, that also allows us to be in a portion of the market that's far more worried about ultimately performance and delivery and quality. So I do think that the favorable price environment is durable. And I think customers are in some ways happy to receive supply, even if it comes at little changes in price over time. But I also think the thing that's really important to keep track of is kind of what I said earlier, that we're introducing, for example, now going from 100 gig per lane EMLs to 200 gig per lane EMLs. There's obviously more value in a 200 gig per lane EML.
The product cycles, the product cadences are shortening from what used to be four or so years, now down to two years. So as you have that relentless innovation cycle, there's not a lot of time save for slower-moving competitors to ultimately catch up and come into the market and changing the pricing dynamics, and that's true at the chip level, but it's also the same thing true at the transceiver level. So then that puts less pressure from the transceiver manufacturers down to us because they're operating at that leading edge as well.
OK. Got it. Maybe turning to that transceiver business. On Cloud Light, you've had one lead customer when you acquired the company, but you've added kind of two additional customers who are ramping. Just what is the ramp path for these three customers? And kind of what's the line of sight to kind of qualifying additional customers?
Sure. Sure. So with our initial largest customer, we've gone through, we're now on the other side of it, a product transition where that resulted in a reduction in demand as the older products rolled off and the newer products are just starting to ramp up. With the two other customers, we have one that's qualified and we're just beginning initial volume production. We expect that to accelerate as we go through the year. And with the third transceiver customer that we've talked about, that's still in qualifications. We expect that it will complete. I mean, some of that's not fully in our control as to do with the customer's timeline and their resources towards the middle of the year, kind of in the June quarter time frame. And then we expect it would start to ramp and grow from that point forward.
And with customers four and five, we certainly are engaged with a range of other customers and delivering samples. The real challenge here is aligning. You generally are inserting on a new generation or a next generation product. And so for us, a lot of that orients around things like 1.6T. And 1.6T is in its very early days. And that'll launch probably with a single customer who's really driving the industry. And then there'll be other customers after that. So I think we've got a shot, certainly not only at landing customers four and five, but something that we talked a little bit about on our prior call, that also delivering products two, three, and four with customer number two and customer number three so that we both grow within the customers that we have, as well as adding additional customers.
Got it. I guess given geopolitical concerns, it would be natural for customers to try to line up multiple sources. Just how do you assure that kind of qualifications turn into meaningful volumes that justify your investment?
Sure. So certainly that's why when we've talked about, use the words awards on earnings calls, meaning that we have some kind of commercial understanding around share or volume or something that, again, provides some of the assurance that you're talking about. But I'd also add that while transceiver customers are trying to obviously land or line up multiple suppliers, I think we're in an interesting situation given geopolitical factors that drives many end customers towards somebody like Lumentum. Given we are a U.S.-headquartered manufacturer, we have substantial transceiver capacity coming online in countries like Thailand and the Philippines that aren't China and are vertically integrated, so control more of our own supply chain on the optics. That geopolitical instability or concerns is certainly driving a lot of the traction and discussion we have with customers.
So I do think, if anything, the playing field is maybe slightly tilted in our favor and helps ensure that the capacity we add will be filled over time.
I mean, in most cases, are they minimums or are they true volumes that you?
Yeah. I don't want to get too much into the details of the commercials other than to say that as I go back to that formula of U.S.-headquartered company, non-China manufacturing, there's not a lot of other folks that fit in that category. And our combined capacity is nowhere near the needs of the industry. So I do think we've certainly got an opportunity to sell what we can make ultimately.
Got it. I mean, today, kind of the transceivers you sell are not vertically integrated from Cloud Light. Just when do you think or what is that timeline for when the transceivers could have more Lumentum content?
Yeah. There's really two reasons for that. First is the transceivers we have today, were developed pre-acquisition time frames. And we weren't designed into Cloud Light as a supplier. But since then, our own demand for our own internal laser supply has gone crazy outside the company. And we can't make enough. So our strategy here is to make sure that we're satisfying all of our external customers so we don't let them down. And then we're adding significant indium phosphide capacity to be able to ensure that we can meet the growing demand from external customers as well as for our internal needs. We believe that intersection point is likely in the next calendar year. But demand is also continuing to go up. So it is a bit of a moving target, but we do expect it in the next calendar year.
OK. I mean, you spoke about kind of supply chain limitations around areas like CW lasers on the fiscal Q2 call. Just how does that impair kind of your ability to gain share currently?
Yeah. I think that the good or bad news on that question is that the overall industry is constrained. So I don't think we're any worse off necessarily from a supply base than anybody else is. So we continue to engage on existing customers and new sockets and new customers. So it's not impeding any traction with customers. Certainly in the quarter, it impeded some of our revenue growth. But certainly our transceiver customers are telling us that we're holding back their ability to grow via lack of laser supply. And I don't think it's just us. I think it's tight indium phosphide capacity across the industry.
And just when do you think kind of that CW laser constraint should clear? Is it similar timing as EMLs where maybe we are constrained through most of the year or?
Yeah. I think we're going to see improvement in supply sequentially. So I think we're going to see through the year improving CW laser supply. To say that supply and demand are going to be in balance, I think is a bit aggressive. I think that takes into the next calendar year.
OK. All right. We got 16 minutes in before we mentioned CPO. So just how do you think about CPO as a threat versus an opportunity for Lumentum?
Yeah. I think of CPO as clearly a long-term opportunity for Lumentum and the industry and certainly a midterm opportunity for us. CPO is not rolling out in the next quarter in large volumes. And maybe for folks that are listening that don't know what CPO is, CPO is an approach where you take some of the optics that are inside of today's transceivers and move them in close proximity to whether it's a switch silicon or a compute silicon. And the reason for doing that is that as we scale up to higher speeds, the power consumption and maybe even the signal integrity get very challenging. And so by doing this, you're able to reduce power consumption and operate at higher speeds. And I think the benefit of that not only solves the customer's problems, it's bringing down power consumption.
I mean, power is what's ultimately limited how much gear is going into a data center. So if power comes down, then more gear goes in. So ultimately, what's being eliminated are power-consuming electronic devices, not photonic devices. So I think ultimately it drives more photonic devices. And at the same time, there are opportunities where we don't play in today at all, which is certain copper applications called scale-up interconnects in data centers. And I think techniques like CPO will find a home in that as well. And that's a much larger market than or a much higher volume market than the current scale-out market. So that's why I think it's a tremendous opportunity over the long run for the industry. For us in particular, we're a high-performance laser vendor first. That's really the core of our data center and data com business today.
And as such, in CPO, there is a need for very high-performance laser. In this case, you're using one laser to maybe power multiple optical channels. But the amount of laser dollar content per channel kind of stays the same or perhaps goes up given the uniqueness of the technology that we have. So I think our ability to partner with, as you asked, how do we prioritize, partnering with lead customers to ensure that we enable them to be successful on that application. And then that will proliferate within their platforms and to other customers. So it will be good for us and then overall good for the industry over the long run.
So in terms of kind of content that you would put into a CPO system, should we think of it as higher dollar EMLs, but you have the loss of potential transceiver revenue? Just like, how do you guys think about kind of that, what your content was before versus what your content becomes?
Sure. So yeah, maybe I'll start just at the laser level because we play at both sides. So if you think about today and/or soon-to-be a 1.6T transceiver, that's eight 200 gig lanes. So that might be eight EMLs. Or it might be photonic-based. And maybe we have four lasers and split the lasers in half. And they power two optical lanes. But the laser's bigger and more expensive that we use to do that. So you carry that to an extreme where you say in a co-packaged optics approach, there's not, you don't need to gang them together in eight lanes because you're looking at all the lanes. So maybe having a laser that serves more than two or four channels is a better way to go because it's more efficient. It takes up less space.
Ultimately, it might be more reliable because there's few lasers involved. But that laser is clearly more valuable. And so as I said earlier, I think the dollar content per lane stays the same and maybe even is a bit higher over time. We'll see. In terms of comparing that to a transceiver, certainly we would not be selling everything that's in a transceiver that's in the CPO. So you're certainly looking at some level of content reduction. But that's because in a transceiver, we're buying an electronic DSP that's not and not marking that up by much, frankly, in a transceiver. And then putting that, the CPO eliminates that element from. So I don't think it's really a profit dollar loss in any way. There is some reduction in revenue.
But I would also highlight that because of the comment that I said earlier, that I look at it saying if it's replacing a transceiver, it's replacing a transceiver that power wouldn't have allowed to generally be there anyway. And so I view it still as an overall favorable to the industry. Second, I don't think you're going to see a mass conversion to everything being one technology base. I think you're going to see CPO emerge in the coming couple of years. And you're still going to have pluggable transceivers. You might have linear pluggable transceivers all existing at the same time. And then kind of this comment about geopolitics, certainly for Lumentum, I think we're going to continue to sell a lot of transceivers.
And I don't think our overall opportunity is reduced even if part of what theoretically might have been a transceiver if you had infinite power is now CPO. But it isn't theoretical. The reality is power is what's limiting it. And I think ultimately we'll still sell a lot of transceivers. And we're going to sell even more lasers based on the CPO development.
Got it. I mean, you talked about kind of that incremental opportunity more in scale-up networks because of replacing copper. Just kind of can you just give us a sense of both been in this industry a long time, both been talking about CPO for a long time, just kind of how you see that timeline of maybe a difference between scale-up and scale-out CPO?
Yeah. I think it's going to take several years. I mean, you're going to obviously hear of probably over the next year of initial products coming to market, but really more ramping in the next calendar year. And then probably that's starting more on the switching side or someplace like that where there is sort of more of a direct analog to what is done today. And then ultimately penetrating over subsequent years into much larger scale-up opportunities. But again, I think a lot of this are still being debated and decided in the industry and by customers. It's again not going to be something that happens in the next quarters. It's really going to be from a volume production standpoint, as Alan alluded to on the earnings call of, hey, we'll go through this calendar year providing some products into it.
But really in calendar 2026 is maybe where an opportunity at a customer starts to ramp up a little more aggressively. And then as we go into the subsequent years, it's more with the same customer and more with more customers.
Got it. So you've talked about how we're kind of at initial days of 1.6T. Just kind of how do you see that adoption path being different for 1.6T than maybe what we saw with 800 gig?
I think it'll actually have a lot of analogy with 800 gig in that we expect it to start with a small or concentrated set of customers who will be the first adopters. They tend to control maybe some of their own silicon. And therefore, that's why they're able to incorporate 1.6T initially. But then in a couple of quarters later, there'll be more, let's call it sort of merchant silicon available for switches. And then you'll see more customers adopting 1.6T.
Got it. And then just how do innovations like DeepSeek change how you think about either the amount of speed transitions or just overall demand?
Yeah. I think it goes back to something that we've hit on a couple of times of anything that can improve efficiency, whether it's efficiency in algorithm or software or if it's in hardware, that enables overall the ability to deploy more gear or to lower the cost of the operation so more users will use it. I think that's all good for trying to drive the overall industry forward. I think that the real headline coming out of that, and certainly you've heard that come from hyperscalers and AI infrastructure guys in the last number of weeks, is that, hey, what really is the innovation is now reasoning models are driving massive compute needs, which obviously is a good thing for us and driving further needs for photonics. So we don't see that, again, efficiency improvements, whether it's hardware or software, as a negative.
Because at the end of the day, AI is not where any of us want it to be. I mean, it's great for the application and use cases we have today. But the promise of it requires vastly more compute power. And I don't see anybody slowing down towards those goals over the mid to long term. And therefore, these improvements just move us ever closer to getting to the market ballooning.
OK. I have a ton of other questions on the rest of the business. But I wanted to maybe open it up to questions to see if anybody had any kind of transceiver data com questions. Maybe one in the back. You can just yell it and I'll repeat it. What's causing the shortage of supply for CPO?
Yeah. I think it just boils down to that the AI surge that started, at least for indium phosphide lasers back a year and a half ago, we only had so much existing wafer fab capacity. And the market is now multiples of what it was back pre-AI. And so the time constant to add wafer fab capacity is fairly long. It's not something you can do in a matter of months. It does take quarters or in some cases years to add capacity. So we've been adding capacity and expanding output from our wafer fab that serves EMLs over the last year. It's been an amazing accomplishment how much output the team has been able to deliver. And now we're underway of adding additional capacity to be able to continue with the steep inflection in volumes that happened with AI.
Do tariffs or additional actions relative to China-based, excuse me?
No problem.
Suppliers impact your forecast? And I would presume it's a tailwind relative to the competitors, but a headwind relative to your Cloud Light opportunity. But I'm just curious broadly how you think that affects customers in the U.S. and what they're choosing.
Yeah. Unless I misheard it, I would say that the tariff situation in general is a cost to the industry. So somebody's going to pay something. But from a relative standpoint, I think it's more of a tailwind than a headwind for Lumentum in that being a company that has significant footprint outside of China from a manufacturing standpoint, we are supplying today outside of China and have plans to vastly expand production outside of China. So as Meta asked about, we have customers two and three and maybe more to come. And first customer all eyeing capacity, in some cases requiring it to come from non-China manufacturing. And so I think it actually is more of an opportunity for us than any kind of headwind. Does that answer your question?
OK. I can take off my tariff question then. All right. So one of the things that caused kind of a transition during the quarter or fiscal Q2 was that you had shut down kind of the NeoPhotonics fab. Just how has that influenced kind of the timing of how you think about the ZR ramp or kind of some of the telco business being able to capture that telco business recovery?
Yeah. So as we had on the call, we've had a very significant growth, not just in actual telecom revenue, but in our demand outlook is really turning up quite strongly due to data center interconnect. All this capacity going inside of the data centers, we need to now build out outside the data center. And the traditional service provider telco market, we believe, is starting to improve. Our customers are telling us they're seeing trends of spending. And we're now shipping more of those products that you know kind of land in a typical service provider. But in terms of holding back our ability to grow and meet customer demand, I wouldn't say it's anything having to do with integrating NeoPhotonics or prior acquisitions at this point. It has much more to do with the fact that two years ago, the business came down rather dramatically.
And therefore, we lived off of inventory and turned off our supply base that supplies into us. And given that, it's taken time for our suppliers to ramp back up. And the demand for telecom products is kind of coming on pretty strongly. So the ability to, like the earlier question of, it's not really capacity, but it's more material flowing through our suppliers' factories. They just take time to ramp up. And that's really what we're seeing is more of a delay in being able to meet customer demand in the telecom space.
Got it. So you've had this kind of $500 million quarterly exit run rate exiting calendar year 2025 kind of target. As you look forward, just kind of what are the biggest factors towards meeting or exceeding that, as well as kind of this 40% gross margin target longer term?
Sure. I mean, I think when we set that out a ways back, we probably had more demand concerns than we do today. I would say certainly between the AI market, the telecom market, virtually every part of the cloud and networking markets we participate in are on growth trajectories now. So really, the ability to deliver on the revenue goes to the few things we've talked about in the data com space. It's getting those wafer fab capacity or output from the wafer fabs. It's closing on these qualifications and ramping up transceivers. So these things are in our or our customer's control, let's say, less than per se an overall demand environment to reach the revenue levels. Then from the margin standpoint, I think the most important factor is utilizing the capacity we have.
In the case of the telecom side of things, we have a fair bit of capacity already. So the variable margin on the ramp back up is pretty good. We just got to get ramped back up and get the supply to be able to do that. Similarly, as we add transceiver capacity for these new design wins and new awards or same thing on the chip side, we just got to utilize the capacity. The margin should flow.
Got it. OK, and then maybe just last question on the industrial tech side. You're seeing kind of demand remain challenged just given kind of weaker industrial end markets. Just how are you thinking about kind of recovery or opportunities within this market?
Yeah. I think there's a couple of things on that. First is historically, we have seen slowdowns in the industrial tech business. And it tends to have a bit longer time to recover than the traditional communications and networking markets. That's not a surprise. We don't have perfect visibility to how it's going to be this quarter off into the future. But I would say we feel pretty good that the market has bottomed. And we're starting to see some green shoots in traditional industrial markets. But that said, we've also put a lot of effort into introducing new products where we weren't selling them historically. So that's all new revenue that's sort of offsetting some of the down pressure. So we feel pretty good about ultimately the historical product lines are going to recover probably in the next year or so.
And the new product lines, particularly on these ultrafast lasers that are used in semiconductor manufacturing processes and display and solar cell manufacturing, brand new products for us that we're just really ramping up the production of. So that should also help provide some market-independent tailwinds for the industrial tech business.
Great. Well, Chris, thanks so much for being here today.
Thank you for having us.