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

Okay. Great. Welcome to our investor briefing at OFC. I'm Kathy Ta, VP of Investor Relations. And very pleased to have you all here. You'll see that we have been infinitely more civilized than some of our peer group because we actually have seats for everyone but I do expect the meeting to fill in as we go along. If you have an issue seeing the screen, there are a few seats up in front that you can grab now or you can follow along with our webcast. So our webcast should be live and we'll follow along with the screens. If you are following along with the webcast, please turn off your speakers so that the time delay of the audio won't mess us up. So with that, I'll introduce the topic for today is Innovating for the Era of AI-Optimized Networking. And I will carefully read every word on this slide. So we will be making some forward-looking statements and our actual results may differ materially from the statements that are made today. You can read the full language to your heart's content on our website. Also, I want to mention that our slides will be available for download at 11 a.m. Pacific Time. So right when we're done with the presentation today and open up for Q&A. And the pleasure to announce today's speakers, Michael Hurlston, our new President and CEO; Wupen Yuen, who is the President of our Cloud and Networking business; and Wajid Ali, our CFO. And then Chris and I will be available for Q&A along with the speakers at the end. So we'll start the Q&A session around 11:00 a.m. Pacific. And with that, I'd like to introduce Michael to the stage.

Thank you, Kathy. Good to see many people. The first question I get asked is, why did you come back to optics, okay? One answer is I'm absolutely insane. This is a tough business. Alan set us up for incredible success but this is a tough business. But the second answer and that's really the answer, is the opportunity. I mean there's no question this market is on fire. And as we'll talk about the switch now from electronics to the optical domain is really what's driving our business and what we see as an unfolding immense, immense opportunity. We really have an opportunity to grow this company both on the top line and in terms of margins and operating margins. So we'll go through all of that as I click through, hopefully, I got this clicker to work. As we think about the optical domain, this is all about power, efficiency and scale. And so as we start thinking about now moving these data centers more and more to the optical domain, which Wupen will take us through, you see incredible power reductions, which ultimately allows these data centers and the AI data centers to have more scale. You can run more workloads in the data center. You can run more servers in the data center at the same power level, at the same efficiency because we're moving things to a much lower power consumptive environment and that's really what optics brings. And so as we think about now moving the switches that are all electronic today and the spine and the leaf down to the optical domain, the power savings that, that enables is incredible. So we really believe that this world is switching quickly to the optical domain. And if we can execute properly and capture that opportunity, we will do phenomenally well. You can see, as we think about the company, one of the core competencies is clearly our components. And we have components really for every make and model. You think about now and Wupen will talk about this, we're -- we believe we're first to market with a 400-gig EML. We already have 200 gig and we have an opportunity to move that to a differential from a single entity into a differential, which again, saves power. So we are not only engaged in the transceiver market, which is an unbelievable opportunity but we are supplying our components to almost every transceiver going into all end customers. We really believe that this is a key differentiator for us, the ability not only to have components but then to bring those components together in solutions and bring those solutions to our customers. So we win kind of no matter what. As our end systems go into customers, we obviously do very well. But as our competitors, in some cases, they bring transceivers and they bring solutions to the data center, we still win because our components, more often than not, are inside those subsystems, so this is really a key for us. As we think about that, again, where do we do really well? It's the marriage of material science and engineering design. That's where we do really well. So we have, as many of you know, wafer fabs, those wafer fabs really bring together indium phosphide, a material, a key material, gallium arsenide, again key material and our laser designers. So the combination of design and material science is really our underlying differentiator. And why we do well over and over again, not only in our own transceivers and our own system products but as we sell our components out to other competitors and customers, that's where we win. So this underlying technology in the company, the marriage of material science and design is a fundamental, fundamental differentiator. As we think about the market, again, this is an explosive opportunity. We see a 25% -- greater than 25% compound annual growth for the part of the optics market that we play in over the next handful of years. And you can see where do we focus CPO, co-packaged optics, LPO, optical circuit switches, these are the areas in which we play that we see the greatest market growth. It's an unbounded opportunity almost, right? As we think about keeping our solutions in the optical domain, our ability to grow and scale the top line while focusing on profitability is where we think we can operate the company. So our scale-up photonics, we saw that last week at GTC. We really think we have a differentiator in our CW laser, high-power CW laser that would go into co-packaged optics. But we think we can play even more broadly in that opportunity and in optical circuit switches, which Wupen will talk about here in a second. What again is a differentiator with Lumentum is our ability to play in all parts of the optical market. Everybody now is focused on what's happening inside the data center because, of course, that's the biggest part of the pie. But our long-haul optics, which has been a fundamental underpinning of the company for years and years, this is starting to drive the biggest transport people now, the biggest customers of long-distance, long-haul networks are the hyperscalers. So as we think about our business and scaling it from end to end from subsea to long haul and metro data center interconnect inside the data center, we have solutions for all of it. We're very unique in the optical world in that we have bits, we have solutions for undersea. We have solutions for metro. We have solutions for data center interconnect. Of course, inside the data center is where we are -- a lot of the focus is. But we do have a true end-to-end portfolio. And this market is explosive not only inside the data center but also as we're thinking about moving bits under the ocean, moving bits from data center to data center, this is the kind of the underappreciated part of the market, a market where we're equally strong and have a very big footprint. So as we kind of conclude here my piece before I hand off to Wupen, a couple of key comments. One, optics brings the lowest power per bit. It really has the underpinnings to allow these data centers to scale up, to allow data centers to carry more complex and difficult workloads. Second, we see this market growing very, very rapidly. 25% compound annual growth over the next 5 years. This is a very sustainable growth rate that we see, particularly in the segment of the optics business that we participate in. And then finally -- and Wajid will talk about this as we think about our business, we see a near-term track to revenues of greater than $3 billion annually and a 20% operating margin. So as we're growing the company, which I don't think anybody questions, we think we can grow it profitably, both from a gross margin perspective and from an operating margin perspective. So we really, really feel good about where the company is positioned and the opportunity ahead of us. Okay. With that, I'm going to turn it over to Wupen, who's going to go through the technical details of all of this good stuff. Mr. Yuen? Okay. Show is yours.

Thank you, Michael. Just that you will stare at my picture for a little longer. So before I get started, I want to share with you, first of all, a few kind of top-level views that will be good for you to think about and kind of put everything we're going to discuss, details into kind of a framework, right? So #1, what has changed? The optical industry has been a boring industry for the last 20 years, right? So what has changed between the old optics and the new optics now? I think the way I would summarize it really is that optics used to be used for communications. Now it's power compute. That sounds a very simple, very plain sentence but actually, it captures the true essence of why optics are now becoming so big and so important, right? Today, as you know, you can hear last week in GTC, without optics, AI does not scale. right? It's now in the middle of the compute revolution and that's why the opportunity is really big, right? So that giving you kind of the overall framework. Then let's talk about the intra data center growth. I want to give you some ideas why it become so big again. When we were so all big in the cloud networking from let's say, mid-20, the first 10 years of year 2000 to maybe 2022, 2022, time frame, All we hear about is cloud computing. And cloud computing only use -- today, we call it the front-end network, right? And that was already pretty exciting. Then when the AI shows up, the GPU connectivity basically required a brand-new network that's called a back-end network, it's 5x the capacity of the front-end network. So all of a sudden, you have a instantaneous big jump in the capacity required for optics. And third, which is yet to come, and you also kind of got a hint of this last week at GTC is that when the GPUs are really connected with the scale-up optics, the capacity requirement is 10x of the back-end capacity. So think about one being the front end, 5 being the back end, the scale up is 50x, right? That, coupled with all the computation requirement needed for AI, whether it's training or inferencing is what's actually really causing the optics to be now thoroughly in this whole revolution of compute. The third point I want to just echo what Michael just talked about, right? And what we're seeing now is that it's not only data center self is growing like crazy but also that traffic that's diffused out of data centers. Why? What's causing all this? First of all, the GPUs are hungry monsters. You got to feed them with data, right? So people today are building data centers, they will actually match the fiber capacity into data centers with the kind of power wattage they're going to actually consume. Why? Because that proportional to compute, #1. #2, today, a lot of data centers are -- because they're so power hungry, they have to build near power sources. And guess what, those areas don't have fibers. And therefore, laying down new fibers, build new data centers also requires more transport capacity. #3, there are some, in the DCI level, cross data center training work that's going on, right? So all this, all combined together is also driving, as Michael talked about, the transport, transmission in a traditional sense but again, is to serve compute, is not to serve communications, right? So these are the kind of the driving factors, right? It's really driving the optical business. And then come back to this and then say, okay, what does it take to serve that market? Right? This is where things get really interesting because the customers, now hyperscalers are looking for performance, speed, scale, quality and security, all 5 of them, right? And that's not your grandmother's optics anymore. It requires not only this one dimension performance or innovation, it requires scale, quality, security, everything, the whole 9 yards. So optics has to go into a very different game to really make itself a growth, a platform that our customers finally can actually count on, right? One of the things we hear about all the day today is actually you're not scalable, are you big enough and so on and so forth, right? So that's where Lumentum actually comes in. I want to come back to Lumentum story is that Lumentum has 25 years' experience in the optics domain, scaled up all these technologies and all the investments and all the fabs and manufacturing capacity is what actually positioned Lumentum squarely in the middle of all these optics revolution, because of these requirements coming from the application itself. So with that, let me jump into the details of my presentation. Okay. So first of all, the optical bandwidth growth, we talked about it already, right? The front end and the back end now then including the scale-up growth is really driving a huge amount of, let me call the optical lanes, right? And to actually meet that optical lanes, the traffic growth, the number of optical lanes now they have to increase by a lot, they also have to increase in speed. Right? And frankly speaking, without that optical lanes been ramping at the same time with power, with the efficiency and speed increasing, the left-hand chart of the power consumption growth is going to be even more. Right? As you can see, more and more, the optical -- the communications and then the computation that's caused that optics become a very dominant part of the whole equation in figuring out how to scale the AI infrastructures. So in order to address the power consumption challenges that data center will be facing, there are a few very important points we will actually go through. #1, you use higher and higher-speed optics, right? Because by and large, when you go to the higher data rate, the power consumption does not really increase very much. So by scaling to a higher bandwidth, higher debt rate, you scale down the power consumption per bit, on a per bit basis. #2, another way to address it is like what discussed last week at GTC, so you call it co-packaged optics. What it does solve the problem for you is that you remove the copper loss between the switch to the front panel, right? That loss is huge at 200 gig per lambda or per lane. Right? So that saving that actually is a very major saving in terms of getting the power consumption down. And third, another way you can do this actually is to change how you do switching. Today, all the switching is done at the electrical domain. At some point of the network, it might be advantageous to actually do it in optical way so that you remove all this very power-hungry silicon switching, and also, frankly, remove some of the pluggable power consumption too. So I'll give you some more details in that regard. And finally, one last thing we'll talk about today is, also talk about how data center interconnectivity can be served with the technologies to at least make the power be more available. You may not lower the power consumption but make it more available to data center locations. So and then as the data center comes to scale in capacity and also in the -- in scale, just in size, there is a very clear trend now going forward is that the traditional gallium arsenide VCSEL-based solutions is not going to be sufficient. It doesn't have enough distance. You can go probably 30 meters at 100 gig per lane, maybe less than that for 200 gig per lane if it's ever doable. Right? And therefore, when you scale up to 100,000, 200,000 median GPUs, the only way to connect at such a vast scale and also at such a high speed is through indium phosphide. Right? Whether you do it with the traditional indium phosphide, EMLs or lasers like that or with the CW laser signal photonics, overall, the amount of indium phosphide usage is only going to increase. Therefore -- and once the data center folks have installed single-mode fiber as their structural connectivity plan, there's no reason to go back anymore. Right? So definitely a single-mode fiber plus high-speed indium-phosphide is the direction that industry is going. And Lumentum having invested in indium phosphate for the last 25 years, build the fabs, build technology basis, foundations and the reputation for quality, we're well poised to benefit from this indium phosphide, taking the leadership in the market share. And then over the last, I guess, 25 years or 20 years or so, right, these are just kind of some of the publications we have done in the industry, all the way from 40 gig, 25 gig, all the way to actually 400 gig. Just now, we actually announced a 400-gig device and I am going to show you some of the eye diagrams here, it's actually pretty amazing. But one thing I want to share with you, too, is that the following, right? People talk about leapfrogging, leapfrogging, leapfrogging. I think in the semiconductor world, frankly, there's no leapfrogging because every single generation of technologies considering performance, scale and quality and speed, you really have to build on a very solid foundation of the previous generation, right? And so one of the core strengths of Lumentum really is that 25 years nonstop continued innovation and continued scale-up of our operations. And that's what's causing us to be able to really implement the next speed node much, much faster than anybody who doesn't have the same kind of history and scale as we have today. So we talked about -- one of the first thing you do is to increase the data rate per lane, right? We're happy to report that we're going to have 2 demonstrations at this OFC. First one is a 400-gig EML, right? Actually, when I first saw this eye diagram, I was really impressed. This is really a very, very clean eye diagrams, very centered, no [ skew ], looks beautiful, right? So this one, we're going to have a demonstration together with [ Keysight ] at their booth, right? This -- I would think probably one of the world's first 400-gig EML. On the right-hand side, actually is a even more interesting, this is a [indiscernible] integrated chip, right, that you can control the chip, you can go to longer and longer distances even over campus. And it also -- this actually driven from a DSP, right? So now it demonstrates that a DSP and 400 gig combination is now possible, right? This will be demoed at our booth, right? So I would say this definitely put us as the front runner, as a leader for the 400G per lane generation and then that will poise -- put us in a really strong spot to be the first to address the market with 3.2T optics. So second thing what you do, a little bit of physical description here is that how do you get the power down at a given data rate. So one of the things that if you understand the semiconductors is evolving, right? While digital circuits are scaling as a smaller and smaller node of semiconductor, the analog portion doesn't scale very much, right? And therefore, like the drivers, TIAs, things like that, the way to get the power down is by lowering the swing voltage, right? The lower the swing that is required, the lower the power. So what we have developed so-called differential EML, basically, you drive it in a differential way, the natural way you're coming out of a DSP. And then by being able to do this, you basically have the voltage swing that's needed, which then results in a power consumption saving, right? And in addition, because you drive it in a differential way, you also give you a better signal integrity, improve the performance of the optical modules. Right? So this actually is a new technology we're bringing to market and this can also be extending into 400 gig per wavelength going forward, right? This will be on top of the another 2 technology we just showed on the previous chart, right, to again extend the use cases, the performance and the scale of our indium phosphide leading-edge high-speed devices. And the third way we talked about already here is to how to reduce the power consumption is to minimize the loss itself, right? Because when you drive a optical module from a switch that long trace of 20 inches or so of copper traces is very, very power hungry. Right? And therefore, the way to actually remove all that power is by integrating the optics right there, right next to a switch and really making distance so short that it almost doesn't cause any power consumption, right? If you do that, you probably can save more than 50% of the power consumption, which actually makes a big deal, you also remove the DSP pretty much completely from the optical chain, right? Now our innovation actually, which is like we talked about just now for Michael, is that our Raman pump technologies, again, developed over the last 20 years has given us a foundational technology starting point to develop a ultra-high performance, high signal integrity and high-power laser that's ideally suited for this application. And this was the announcement at GTC last week, right? And this basically highlights the Lumentum's laser is being used together in this -- the CPO platform, both for the quantum switch and the spectrum switch, right? If you look at this chart, something, a little detail there that you will see that the density of the CPOs on the switch is very high, right? And actually, NVIDIA talked about the use of microring modulator, right? That is the cause of such high density or so-called shortline density of the CPOs, right? And our laser is specifically designed to match the need of performance for the microring modulator. And we believe microring modulator is now being proven feasible, will become the norm going forward. And therefore, we're highly confident with our lasers position and performance. We're definitely increasing the scale of these lasers to meet the future demand of the CPO going forward. Now it's going to take some time, right? It's not going to be overnight. It's a very complex technology over time but we do believe that the future is going to be pretty bright for this technology. And we talked about it already, right? Because of CPO, there's a word called blast radius, right? You have this big switch. If the laser fails, it's going to be a big problem, right? And therefore, our Raman pump technology and proven reliability is the core, right? That got our customers -- the attention, get the confidence and everything is going pretty well. Of course, we'll be ramping Raman pump lasers as we speak as well because the transport capacity is also ramping up very quickly. So if you put all this together, right, actually Lumentum has a full road map covering the next 2 generations, 3 generations of optical technology. We're just taking this really from the page of GTC from last week, right? But right now, we're in 800G generation. Next year, it will be 1.6T generation. 2 years later will be 3.2T generations. We have a whole slew of devices, whether it's a single -- is a transmitter, EML transmitter or high-power lasers, right, whether it's a pluggable module or the CPO form factors, we have technology to actually serve it all. Right? So these are all on our road map, right? And we do intend to not only have the technology but also have the scale to serve at the time lines that the customers need. Finally, we're not talking about today is the scale-up optics on the last row, right? Today, it's still very much a copper cable going to a copper backplane going forward, right? But at some point, we go scale beyond a few hundred GPUs, it's going to be very, very challenging to scale at that data rate over that number of GPUs. At some point, optics has to be used. And we think that in the next few years, there's another opportunities that we at Lumentum intend to address in a big way. And hopefully, that will again add to the future opportunities for Lumentum going forward. Okay. So that was about indium phosphide. If I have not convinced you with indium phosphide, our strength and position, let's talk more. But we want to also talk about this optical circuit switch, right? This is a brand-new innovation that have been talked about and Lumentum again, is in a very strong position. Why? Because Lumentum has been doing MEMS-based free space optics design for the last 25 years. Our first WSS were actually built on MEMS. Every single MEMS problem that everybody has seen, we've seen it all. And with reliability to demonstrate for it, billions -- more than 1 trillion actually MEMS hours operating in the field, right? And this actually allows to put together a very simple, very elegant MEMS-based OCS design, right, that got our customers' attention. Not only the performance is fantastic in terms of [indiscernible] loss and also return loss and things like that. But reliability was really, really, really important for the customer because this again has a huge blast radius in the MEMS sales, right? So we're very proud of this because it's built on, again, a strong foundation of 25 years of work on MEMS-based optical products. And this just shows you the history. I'm not going to go through the details. But since 25 years ago, our first WSS was based on MEMS. And we have continued that MEMS evolution development over the last 25 years. right? To this day, it's used in our -- some of our other WSS products and that's why we gain our customers' confidence to use our OCS, which is almost like a transparent no penalty kind of optical box that the best that you can think about and it's very simple. So one of the big things, if you think about the numbers, right, it's actually kind of mind boggling. You think about a regular data center switch, call it [ 1T or 2T ] switch or 51.2T switch, the switch plus the optics, which can cost you 1,500 watts to 2000 watts. So that switch, think about that switch, maybe 2 RU boxes, 4 RU boxes, the switch chip inside with all the front panel modules, about 2,000 watts. Our OCS consumes less than 150 watts, right? So the power savings, if you use this architecture, is meaningful. Now you may not use it at every single layer of the switching but certainly at a spine level, using optical switching actually makes a lot of sense, right? And therefore, we are really confident about the opportunities of the optical switch. Certainly, analysts has told us that and told the world that it's probably $1 billion opportunity or more in the next few years. And certainly, in some of the other hyperscalers people are using OCS for their own architecture choices, right? So we see OCS as being not only a customer specific but maybe become an industry-wide opportunity for the OCS. And finally, right, so we talk about power consumption reduction based on the speed, based on indium phosphide innovation, based on the use of optical switch. Finally, it's about power availability, right? So now we talk about data center have to be built next to power sources or frankly, when the AI cluster costs so much power, you need more data centers, kind of close by. The connectivity, the data center interconnect connectivity is now requiring more and more, I'll call it, pluggable optics, right? This again changes our indium phosphide innovation in the coherent optics domain and this is actually a huge opportunity. This year, combined between the pluggable coherent optics plus the coherent components roughly the size of the market about $2.5 billion will go up to about $7.5 billion, right? It's also coupled by the fact that so-called channel limits has been reached, right? So you're not getting more efficient as you move along technology is not really making transmission more efficient. So the more bandwidth you need, the more channel you need, the more channel you need, the more pluggable you need, right? And therefore, the ramp of the volume and also the market size become proportional to really how much bandwidth you need, right? So all these benefits coming from the scaling of the channel limit now has been reached. Right? So overall, we are very excited about the opportunity as well. Again, it's based on our long-term investment, fab scale and quality and reputation of the indium phosphide technology that we've been building at Lumentum. So how do we build this, right? Again, we talked about scale really is really, really super critical. I think when the back-end network showed up, everybody was caught by surprise, right? But we at Lumentum was able to really ramp our indium phosphide capacity really quickly. In the last 18 months, we basically doubled our capacity in indium phosphide. We have a long-term plan to also increase the indium phosphide to even further beyond this to serve the EML, to serve the CW laser, to serve the ultra high-power lasers going forward, right? We know how to run indium phosphide, including changing the wafer size, including more machines. And all these are fully automated that guarantee the quality and also scalability. We're really proud of the work doing here. On top of that, considering all the geopolitical challenges and we actually have a very large operation for assembly and test in Nava, in Thailand. That's where we put all our tests, the back-end test and assembly processes over there, right? So we have our wafer fabs all around the world, in Japan, in Europe, in the U.S. We have our assembly and test operations in Thailand. So we're also very well set up as a global company to serve the needs of our hyperscaler end customers, right, and meet their needs on all different dimensions that we talked about in the very beginning. So just to summarize, right? So Lumentum is a company, I want to give you the view, right, Lumentum is a company fundamentally built on its chip fab device strength, right? We also, coupled with that, the free space optics, the pump technologies that we've done over the 25 years and the indium phosphide portion of it is very well suited from speed, power consumption, scale, point of view to serve the future. We put it together into some of the module products to accelerate or to get us closer to our customers. And coupled with the optical switching, that's also super important for our customers to scale their network. Finally, the ZR coherent optics is able to scale the distances, right, allow us to be in that part of the market as well, right? So overall, as Lumentum is very well positioned in entire AI-driven optical evolution. And hopefully, then we're able to see our business growth accordingly as what we will talk about next. So if you have time, come to our booth, we're going to show you the whole 9 yards of this. You have the optical switching in terms of per fiber switching, right, the OCS. You see the optical switching by wave lengths, the WSS, you can see ZR optics, you can see the data center optics, all there. You also see the 400 gig demonstration in our booth as well. So you'll not be disappointed if you come to our booth and take a look at what we have to offer to our customers and to the industry. Thank you very much.

All right. Good morning, everyone. So thanks very much for that Wupen. So what I'll do is I'll talk a little bit about our model and how we really think we're going to be able to execute to the model that we're going to put in front of you today. The punchline, Michael started with right away and that was our ability to execute to a $3 billion top line plan, which should give us an operating margin profile that's greater than 20%. And we've -- in previous investor briefings, we've spoken many times about the natural leverage in the business. And whether that natural leverage comes from the capital investments that we've already made, that we're able to scale or it's through some of our foundational R&D investments that we've had that allow us to execute quickly on new opportunities as they arise. Wupen talked about 2 critical foundation technologies that I think deserves to be rementioned. One is really around the foundational technology that we've got in our WSS products that's really allowed us to scale quickly, that's going to allow us to scale very quickly with OCS. We talked about OCS at our last year's investor briefing. And this year, we're now in a position to be able to communicate that we're going to be shipping OCS products in the second half of calendar year '25. We're seeing extremely strong demand from our customers and pull from our customers. And the operating margin profile of our OCS products are able to benefit from the foundation technology that we've got, both with MEMS and with WSS products. That's the first one. The second one that Wupen mentioned is really around the years of pump technology that we've got. And we talked a little bit about ultra-high-powered lasers in our previous investor briefings. But really to see it flush through and for us to be able to have a large AI infrastructure customer, providing us with a lot of demand for that product has come through because of the foundational technology that we've had in the past and that we've been able to leverage from and to really support the financial model that we're going to talk about today. And so effectively, our capital deployment strategy as we move forward has not changed, is to have those core foundational technologies that we can leverage as we see new opportunities in the market. So a couple of things that many of you asked whether you're speaking to me or to Chris or to Kathy is what does the next $500 million quarter look like? And we've publicly communicated that we expect to be able to achieve that $500 million quarter and the associated -- operating leverage associated with that by the end of this calendar year. And you can see the last time we were at $500 million, it was largely because of strength that we saw in our telecom business, although we didn't see as much DCI demand at that time. It's mostly just kind of natural telecom business as well as a lot of the 3DS demand that we had. And so that was kind of the old $500 million. The new $500 million is much more split evenly between our datacom chip demand. We'll start to see some OCS shipments as part of that $500 million but also our cloud module customers coming online. And then the other 45% is really our telecom business. You can see that we've kept it a little bit muted, not because we don't expect to see those historical type of growth numbers in telecom but just so that we can use that as a point of upside and focus in on our DCI products, specifically our tunables products as being the main reason behind us getting to our next $500 million. And you can see that the industrial tech portion of our business as part of that future quarter is substantially less than what we've had before. Now that's obviously changed a little bit of the dynamics in terms of what type of operating margin profile that has. But having a good sense of what that looks like from a product mix standpoint, really kind of drives the thinking around not only our R&D but some of the CapEx investments that we're going to make in order to support numbers beyond what we see by the exiting of this calendar year. Okay. And there will be time for Q&A at the end. So you can ask questions then. Okay. So Michael talked about the $3 billion. And so again, I'd encourage everybody to think about that on a quarterly run rate basis, right? And so in Q2 of fiscal '25, we achieved $400 million. The way to think about $400 million to $500 million a quarter is really just our EML capacity coming online. And so our EML capacity will double between last -- the Q2 that just ended and the Q2 '26 that we've talked about. And so that will be a big source of the delta and then us really being able to ramp up our Thailand facility to support the tunables demand that we're seeing because of data center interconnect that -- those 2 things should really drive the majority of the delta in terms of getting from $400 million to $500 million. And so $500 million to $750 million a quarter, the reason we're able to talk about it this year is because of OCS and ultra high-powered laser. And so the bridge to think about is going from the $500 million to the $750 million are really those 2 major product lines driving a substantial portion of the growth from where we expect to be at the end of this year to when we get to the $750 million a quarter. Now the other positive thing that you'll see on the next 2 slides, is that our gross margin profile for both of those product lines that are going to drive that growth are above the corporate gross margins of the company. And so the ASP profile are substantially different for both of those products because the ultra high-power products are chip products, chip-based products and OCS is more of a module-based product. So the ASP delta is substantial but both of them carry a combined gross margin profile that is above the company gross margin average. And so what you'll see is you'll see some nice gross margin leverage for both of those and you'll see some nice operating margin leverage. A lot of it has to do with the fact that ultra high-power laser will be produced by us and are -- in 2 of our fabrication facilities and OCS, a lot of the work will be done at our Thailand facility that we'll be investing in, in order to take advantage of the leverage there. And then when you flow that through, when you flow that revenue growth through to operating margins because the gross margin profile is better than the corporate company average and the R&D investments, much of those R&D investments have already been made and we've got that as a baseline. And those same R&D resources can then move on to the next generation of OCS, which Wupen hasn't talked about yet, we really don't see a major increase in operating expenses as that flows through. So the target financial model. So last year, when we got up and talked about our target financial model, the first column had a $1.6 billion to $1.7 billion number in it. And so that effectively equated to $425 million a quarter at about 10% to 10.5% operating margin profile. You can see from the progression of results that we've had for the last number of quarters as well as the guidance that we provided for our fiscal Q3, really, we've fallen in line with the gross margin profile and the operating margin profile that we communicated to all of you last year. The $2.4 billion model we also presented to you last year and there haven't been really any changes to that with us being able to start to get towards 40% gross margins and operating margins that start approaching 20%. And then as we get into volume production for both OCS products and our ultra high-powered products, we will then see the bridge to a $750 million quarter run rate and we should be able to see operating margins that are greater than 20%. And you can see that we see a path to start nudging up against 42% gross margins with operating expenses being 18%. If you start to quickly do the math on 18% on $3 billion and you compare it to really what we're spending right now, you can see that actually there's a lot of investment that's built into that number as part of the $3 billion model. So there's certainly opportunity there as we continue to use the same foundation of resources to drive that level of revenue. So that's really kind of how we're thinking about the model moving forward. And you can see that really not a lot of changes in the first 2 columns from what we have presented last year. All right. So I think I've said everything here already, probably shouldn't be any takeaways that you haven't already kind of thought through. So with that, I'm going to pass it to Kathy to start the Q&A session.

Okay. Thanks, Wajid. So we'll be taking questions from the people in the room. [Operator Instructions].

Vivek Arya from Bank of America Securities. So I had a few questions. First is the timing, Wajid, for the $600 million and then the $750 million. And then secondly, you mentioned that the delta was really OCS and high-powered lasers. So I'm surprised not to hear transceivers right in that. What's your -- what's the market for OCS today? What's your share? And how much do you think -- where is that market going? How many customers because it seems like a very niche market, so for it to generate a large TAM and for you to have a high market share at accretive margins.

Okay. So Wupen and I will talk through this one together in terms of market share. I do think that -- my view is, I do think the market is evolving. And so we are starting to see bottoms-up demand from multiple customers in multiple geographies. And so I think OCS right now, I mean, we can look at TAM information and kind of conclude based on how we think the overall TAM is working. But based on what we're seeing from a bottoms-up standpoint, our internal forecasts are very well supported. And I would say, are below 50% market share in terms of the opportunities that we're seeing over the midterm. And just to be -- not to be cute about midterm, so we've openly talked about $500 million being by the end of this calendar year. $600 million, we will probably see 6 to 9 months after that, just in terms of run rate and based on what we understand about the timing of our 200G EMLs ramping up. I mean there's a lot of products I haven't mentioned just to keep the storyline simple. But the reality is, is that 200G EMLs are also ramping up in the time frame of the beginning of calendar year '26. We're going to start to see OCS ramp up probably around the middle of calendar year '26 and high-volume production. We're going to start shipping now but we'll start to see high-volume production. And then ultra high-powered laser is probably the middle to back half of calendar year 2026. And so I don't want to put a time frame on the $750 million, but that's really the progression of the timing of volume shipments and so as we get closer to that, we'll be able to put a stake in the ground and say, okay, within this 6-month period, this is when we see it hitting. I think the other thing is, is that we have undercalled in the very first slide, I was open about it. We have undercalled our telecom demand. And so we're seeing WSS demand. We're seeing tunables demand. We're seeing CDM demand. And a lot of it until our customers report. We can't actually tell how much of it is natural telecom and how much of it is actually related to data center growth. But across all of those product lines, we're seeing an uptick in demand. And so even though I'm calling it DCI, I think there is some -- definitely some DCI. But I do think that there is also some underlying telecom demand that is coming back that could kind of push and pull on our revenue profile, not mostly because of how we can supply to that demand profile rather than the demand from our customers itself. So I hope that makes sense. I don't know, Wupen, you want to...

So on OCS, right? I think you are right, OCS is still a nascent market, right? Again, one customer has a very specific use case, the volume is pretty high. And there are other customers who are actually looking at it, right, as a way to the spine switch replacement or in some cases, actually scaled up, right? So there are multiple use cases, right? Today, our share is low, which is getting started. But like Wajid said, right, we do have visibility into ramping up to some very meaningful volume second half of 2026.

One quick follow-up. So in your EML lasers, how do you decide whether to use it in your transceivers versus selling it to others? What are the difference in margin and your opportunity?

Yes. So I mean, just kind of backing up, as we've been taking a look at our capacity, one of the conscious decisions we've made is actually the difference between our EML lasers and our CW lasers just given the margin profile per wafer. And so the margin profile per wafer is better on EMLs, and it's less so on CW lasers just given that market dynamic. And so as our capacity is ramping up, we now have more room to ship our own CW lasers, whether externally or internally for our own demand. And so our thinking is that if we have a customer that wants our EMLs as part of the transceiver solution, we will go satisfy that customer first and then take a look at the margin profile of the second, and then we look at capacity and then we decide on CW laser. So it's kind of 1, 2, 3. That's how we're thinking about it.

Chris Rolland from Susquehanna. Just a clarification on OCS and then I wanted to hit EMLs. So I guess, first of all, on OCS, the $1.25 billion by '29 that market. Just to clarify here, this is for the full box as opposed to WSS components and the margin profile of the full box.

Is greater than our company average.

Is greater than corporate average...

Yes, it's greater than our company average.

Great. Yes. Thank you for that. Yes. And sorry, one follow-up there would be if you had a market share expectation or something like that around this. I know it's fairly limited market...

I mean if you just do the math, right, it clearly says we're in the kind of 30% to 40% market share number based on the high-level math, but our bottoms-up level is kind of coming to the same result. It actually feels like customers are coming to us first, just based on the intensity of how quickly they'd like to see our products and when they'd like to come in, but it just might be a bump in the road and it will normalize there. So that's why we're very comfortable with the numbers we're presenting on OCS too...

Chris, we have a big advantage on high port count. So as you get up to 300 by 300, we think we have a big advantage. On lower port counts, more competitive. Still, we're a major player. But you can see our share getting into the 40s very quickly. I think the market is moving to the higher port count. And as that happens, it plays to our strength.

Fantastic. Thank you, Michael. On EMLs, and this is really a debate, I think, here at OFC EMLs versus SiPho. It seems like EMLs at 800 and then 1.6, have a speed advantage, speed to market advantage. How are you guys thinking about some experts claim that the gap for SiPho is narrowing? So how are you guys thinking about your position in EMLs? How are you thinking about this going forward, your capacity ramp? And is there a situation in which SiPho ramps quickly and you have additional extra EML capacity, is that fungible for something else?

So the capacity is fungible. So I'll answer that part of the question because we're constantly making allocation trade-offs between EML versus CW laser. And then within EML, internal, external and then within external, which customers. But let's hand it over to Wupen.

Yes. The EML for CW laser or silicon photonics is not just technology per se, right? Actually, last year, until earlier this year, the EML supply -- I mean we talked about it, we're just inundated by EML, right? And people say, oh, EML is not quite available. Let's go to CW lasers, go to silicon photonics. And guess what, it's also constrained, right? So people have -- I think they have maybe an intuitive feeling about, oh, just choosing one versus the other. The fact is indium phosphide is the constraint. Right? And therefore, we're like, okay, let's focus on EML, making sure we make our gross margin dollars there. Now capacity ramps up, we also serve the CW laser, right? So that's one very important factor that kind of cost that shift a little bit. And single photonics, of course, it's actually coming, it's actually getting better. Frankly speaking, though, almost everybody uses the same silicon photonics foundry today. Right? So that could also become a constraint, right? So I think as technology starts to kind of play itself out, it's going to be performance and also the availability of capacity. Right? So again, like we said, we're going to optimize our gross margin going forward, and we have both technology we can serve the entire market.

Tom O'Malley with Barclays. Just really quickly, the theme of the show here is both the optical switching and then also co-packaged optics. So to the extent that you can on a like-for-like basis, I don't know if we're talking about the aggregation layer or you could talk about the scale up. Can you talk about what you would get from an ASP perspective from a module today versus all of the cool like solutions that you showed? And then kind of the second part of that question, if you look at your financial forecast, is there an internal assumption about how quickly that takes off and where you see that taking off first inside of the rack or at the aggregation player?

So let me answer that question. So the way we look at it following, right? I probably won't answer the ASP question. But the way we look at the math is following. So we have a -- we believe we'll have a meaningful share in the CPO market, especially for the laser part because of the performance advantages, right? So we think we're going to have a pretty high market share in the laser side, right? And then on the pluggable side, we're going to have a lower market share on the pluggable side. Therefore, you actually calculate the opportunity, the revenue opportunity per switch for us is almost the same, right? And then the CPO approach will have a higher gross margin, right? So therefore, for us, while some of our peer companies have a very high share in the pluggable space, this will be a headwind. CPO is going to be a headwind for that. But for us, actually, it's not so much of a headwind because of the share being different and the strength being different, right? So that's kind of how we are looking at the headwind of CPO in our business. Wajid, you can answer the second part of the question.

Yes. So in terms of material revenue, material levels of revenue, not $10 million or $20 million here or there. If you just take a look at our CapEx profile, so obviously, EMLs will benefit the financial profile first just because we've been working on that capacity for almost a year now, right, in terms of putting that in. And so we're starting to see the benefits of that. The second part will really be OCS because OCS, there is some back-end investments that we need to make in our Thailand facility, but those are quicker to get to and quicker to kind of shift over from a volume standpoint and actually achieve that revenue profile. And then ultra high-powered lasers, we do have some capacity at our facility in San Jose, but we're going to be making incremental investments there. And so we'll start to see those -- that CapEx profile come in, in the middle of the year next year and then start to be able to ramp volume levels of production on ultra-high powered. So it's -- some of it is our customers' road map. Some of it is just constrained by the timing of how the CapEx will all work and whether it's a back-end CapEx or a front-end CapEx.

And then just one more, if I may. On the telecom side, some good growth assumptions and where you guys see yourself at the $500 million. When you split that out between transport and transmission, can you talk about what's generating that growth? And then just remind me, just that I'm straight. In terms of the ZR revenue that you guys get or the ZR+, is that all still going to be in the datacom side or the telecom side? Tell me how you're going to define that in that growth.

Yes. So on the ZR side, a lot of the growth that we're going to see is because of our tunable lasers, right? And so our tunable laser demand is just -- it's really high. All the way kind of backlog that we've got right now that we have unfulfilled all the way out to the end of the calendar year that we see demand from customers. We actually have 2 different manufacturing locations trying to meet that tunable laser demand, which really is going into ZRs of our customers. And so that's kind of how we're thinking about the ZR demand. On the rest of kind of telecom, if you put the DCI aside, we're seeing a lot of growth on the transport side of the business with our WSS products and our CDM products. And so yes, that's where we're seeing a lot of demand. Now could some of it be telecom, could some of it be DCI? That I'm a little bit unsure of.

On pump lasers as well, actually.

On pump lasers, sorry. Pump lasers aren't really high too right now.

Simon Leopold from Raymond James. I want to get a sense of how you see the mix of the flavors of lasers going forward in that there's sort of conventional [ winsomeness ], CPO starts displacing more and more VCSELs and then CW maybe get squeezed from the top, from EMLs to ZML gets squeezed by coherent devices coming in the data center. So I'm trying to think about not this year, but years out, what is the mix of laser types look like? And then ultimately, what does that mean for Lumentum?

Well, that's a complicated question.

That's loaded.

Yes, that's loaded question. So I think a couple of things. Number one, 100G per lane is going to be there for a long time, right? So the way we look at it is following the 100G is going to be continuing to ramp like this, right? They're going to add 200G on top of it, right? So within each bucket, there's going to be a share between EMLs and CW laser. Today, it's probably still mostly EMLs that could change, right? And then 200G per lane probably started with the EMLs and maybe gradually shifting over to CW lasers, right? So overall, I would say, if you look at this, that's why I think we look at it, we're going to be there for all different categories, right, 200G or 100G per lane and CW and EMLs, right? So that's -- it's going to evolve, right? So today, 100G, 200G like this and you're going to see the share split between the 2, right? I think the CW, the ultra high-power CPO is going to come in probably in the next 2, 3, 4 years. It's going to gradually eat into, probably not scale up, more scale out, right? It's going to start to replace some of the EML or CW laser market share, right? But again, once we are actually well positioned in there, it doesn't matter to us, right? I think for the scale up portion of it, it's probably going to come in 2028 time frame-ish, probably still a few years away.

And what about the idea that coherent light devices from the likes of Ciena or Cisco get inside the data center?

Yes, that's actually something we didn't talk about. So we talked about may come in the future. I think coherent light is becoming increasingly interesting because the challenges at 400 gig -- 400 gig per lane going forward for any distance at all becomes really high. Right? As we're now getting into the closer and closer, have a close look on technology itself, I think the challenge is going to be -- it might make sense now to think about paying a little penalty on the power consumption, but really get much more linked budget. Right? So I think it's becoming really interesting. And we're definitely not absent in that investment either.

Karan Juvekar, Morgan Stanley, on behalf of Meta Marshall. So 2 quick questions. One, kind of if you're expecting a 25% growth CAGR and doubling EML capacity this year, why not commit to further capacity expansions going into...

Sorry. Why not...

Commit to further capacity...

Commit to further capacity expansions beyond calendar '25?

I mean we are committing to capacity expansions that we've already made, and that CapEx is going to start coming in, in calendar year '26. And so the lead times are 6 months, 9 months, 12 months for even simple things like reactors. And so many of those commitments are currently being made and some of them have been made 6 months ago that we're not going to see until calendar '26. So I think -- we always want to be thoughtful about how quickly we pull the trigger on CapEx, and we are certainly leaning in, but we're not using some of the market growth numbers to determine how we allocate or deploy capital. I'll be honest with you about that. So it is a very balanced and thoughtful approach. But it is taking time just because of the lead time offsets on CapEx.

Makes sense. And then another quick question, just time line to Cloud Light vertical integration. Any updates there and how you're thinking about that?

Yes. So we're typically doing it, right? So as you know, Cloud Light has a silicon photonics-based platform. So we're definitely vertically integrating our CW lasers into our silicon photonics platform. As we talked about just now, we have been favoring EMLs just because of the gross margin. Now we're also using our own laser for our internal use. Again, we are not going to conflict with our external laser customers. So we're going to be [ host loop ], putting our own silicon photonics, our own CW lasers and our own detectors actually in our Cloud Light kind of the module products going forward.

Ananda Baruah, Loop Capital. I guess a clarification quickly and then 2 questions. Wajid, on the EML business, did you say that your first order of business will be to serve your merchant transceiver customers?

Our customers, yes. Yes. I mean if we've got a customer that wants our EMLs and our transceivers, job #1 is to get our EMLs into our transceiver customers because even though the margin dollar profile isn't as good as selling to somebody else, it certainly creates a competitive advantage for us that Wupen convinces me is better than today's margin dollars.

And then after that, did you say you take a look at the margin profile...

Of the EMLs that are left over. We allocate that capacity. And I think that sometime in early calendar '26, maybe even exiting calendar year '25, we'll start to have some capacity for more CW lasers I talked to our head of our fab operations almost every other day to kind of understand that. So it's certainly at top of mind for us because Wupen is bringing in CW laser demand, and we want to make sure that we're able to commit to those customers based on the capacity we see as well as the risk management of our EML demand.

And that's helpful. And then -- so is that to say the decision is -- was it gross profit dollars or...

Margin dollars per wafer. Because the die size is different between EMLs and CW lasers.

So you got -- so this is your heat map going forward?

That's your heat map. Yes, thank you.

Got it. And that dovetails into my second question, which is I believe in the slide deck 2029, like momentum-focused TAM, is it $30 billion?

Yes, that's correct. Yes.

And so $3 billion is this target out there that sort of suggests -- it's not a 2029 target, but it sort of suggests 10%. So 10% share. So can you foot the $3 billion to the $30 billion and kind of share wise, how do you guys think...

It's not tops down. So Ananda, it's bottoms up. And so I think as we work through OCS with various hyperscale customers, we'll be able to have a more thoughtful answer to that. And so the 25% is more about kind of the long-term consistency of the CapEx we're creating and not trying to say 10%, but I'll give it to Chris to give a better answer.

Let's not get apples and oranges. The $30 billion is 5 years out, Wajid highlighted a nearer-term $3 billion objective. I would also say that clearly, where we've highlighted market leadership positions on the components, lower ASPs, better gross profit dollars. So I think you'll see across our portfolio fairly high relative market share across components and then transceivers, share gain or OCS, we should come out of the gate very strong.

Joe Cardoso from JPMorgan. I guess just one bigger picture question for me and particularly on Michael's comments yesterday on the CEO Forum -- our CEO panel. I think you made a comment around not getting paid for the value that you guys are driving and maybe from an industry perspective. And it kind of echoes a similar comment from your peers, but where they saw an opportunity is more on the industrial portfolio. So as you think about Lumentum's opportunity, particularly that you're more weighted towards communication, how are you thinking about driving that value for yourselves? And is there any particular areas of the portfolio where you see that as more low-hanging fruit, et cetera?

Yes. Look, I think there's a significant opportunity for us to expand gross margin, right? I think, as I said yesterday, I don't think that we get paid for what we do. And I think as we go through the portfolio in the next year, there's going to be areas where we're going to focus, the high-margin parts of the business. Wupen spent the time talking about EML lasers. We think that we do very, very well there. There are parts of the portfolio that, frankly, don't deliver the margin and there's no growth attached to it. So if you think about the older telecom side of the business, we're engaged in that today. Is there an opportunity for us to do some cleanup in the portfolio? So that's what we're going to look at, I think, over the short term, how do we get the gross margin up overall into the 40s.

Got it. And maybe a clarification for Wajid is just as we think about the financial model you laid out today, is that kind of baked into that? Or is that...

It is not baked into that. The model does not take into account the daily e-mails I get from Michael Hurlston on the orders that have come in and the standard margin per order. It does not contemplate that, no.

Ryan Koontz with Needham. Just follow-up on telecom commentary there. It sounds like you're being conservative. What do you assume for a TAM growth rate there in the telecom sector talking mid-single digits kind of thing or...

Yes, low to mid-single digits for classic telecom outside of DCI.

Yes. Right. And then with respect to DCI and ZR, like what is your attach rate look like there? Is it higher when you go to a DCI model than traditional systems? Maybe you could talk about the mix and how that changes your attach rate margins.

I think from a component point of view, in laser, for example, I think the attach rate is as high or maybe even higher for the traditional telecom products, right? But look at our kind of high-speed product for telecom, for coherent transmission, detachment is also very high, very high. So I think overall, I would say maybe close, very close to each other because, again, they build on the component strength, right, DCI or transmission, it's the same. They're like Datacom. Yes.

Yes. That's great. And then on ZR specifically, is that a market you're going to try to compete in competitively as a module supplier?

Yes, we think so. We continue to have that opportunities going forward, right? 400ZR is kind of a current tense. The future is going to be 800ZR and 1.6T. We definitely will retain those [ modules ].

Tim Savageaux from Northland. I've got a couple of questions. First, in terms of differentiators, you focused a lot on design, material science as a chip fab device, didn't really hear manufacturing in there anywhere. And does that potentially portend a change in strategy in terms of manufacturing internally that has arguably been a source of execution struggles for the company? Question one. Question two, what's going on with Amazon? Everyone is issuing warrants, you aren't. What does that say about your potential market share across the board? That's it for me.

Yes. Tim, look, I think that the manufacturing is sort of a means to an end, right? So we think we have a flexible manufacturing footprint. We've got Chinese manufacturing. We've got a footprint in Thailand, as you know, and that actually has allowed us some measure of scaling into these various opportunities. So having that has really enabled us on transceivers, for example, to capture market share, probably quicker than we would have had we gone outside. As we look out 5, 6, 7 years, how are we thinking about it, that's still a question, I think, for me, and I think the rest of the management team very, very important, but is it really going to be a gross margin growth driver? We're going to take a look at that. What was the second question?

It's a little company in Seattle....

Oh, yes, yes, for that question -- look, we supply -- yes, we supply our lasers to people that are going to supply into Amazon. That's really our means right now. So we don't have a direct relationship per se, something that I think we can grow into quickly as we think about the transceiver opportunity, OCS, it's a move that we can get into Amazon. But today, our path in is through other transceiver suppliers that are taking our lasers in.

And just to point out, I mean, as you all know, the warrants are dilutive to gross margins. So it's effectively a rebate that you're providing. So irrespective of kind of that you're using equity for it, it is a gross margin.

I mean if it's $4 billion, then it's super dilutive, right?

Yes, exactly. So just so that we're aware of that.

This is Yang Pu from BNP Paribas for Karl Ackerman. So I have a question about VCSEL. You had a slide showing that VCSEL share is going to decrease from like 40% to 20%. Is that fair, we are moving from VCSEL to EML, but some companies are showing 200-gig VCSEL in the show flow. So I wonder what's really -- if we just look at data center, what's the real application for 200-gig VCSEL? Do we still need that? And -- or do you think that the 200-gig VCSEL already missed its adoption window?

That's a great question. I think the -- 200-gig VCSEL, I think you will find applications in AI clusters. Right? Even if it works, it means a heavy DSP compensation. The VCSEL technology is just very challenged at this data rate. So with a heavy DSP compensation, you probably can still reach like 20, 30 meters, something like that, in that range. So it would be very much limited to AI cluster kind of thing. It cannot be across data center kind of scaling, right? It doesn't do really kind of real scale-out implementation, right? So I think it will remain more of a data -- AI clusters getting kind of application.

So there is still some space for 200-gig VCSEL for server to switch.

Yes, something like that. Something like that or GPU to switch or switch to server, that kind of thing, yes.

Okay. And another question is about CW laser -- your indium phosphide capacities. So I know it's fungible between EML and CW. Just how fungible it is? Like what's the real lead time to switch between each application? Also, how fungible it is for ultra-high-power CW lasers?

Yes. Actually, the fungibility or the time to switch is just really just the lead time for the fab. Yes, we started wafer, starting EML, starting CW laser just when to start. It's just the lead time. It's the same tool to implement everything. Okay. So there's no real change time.

What about ultra-high power?

Same. Same kind of thing. Yes. So it's just the lead time, application lead time.

Can you give us a little bit some sense about like the margin for ultra-high power laser related comparing to EML and the...

Ultra-high power is above our company gross margin average. And so I think that's what I'll say about that.

This is [ Emily from Generation ]. So I have 2 questions. First one is a clarification question. So we talk about from $500 million per quarter run rate to $600 million run rate about 6 to 9 months. I want to ask if we achieve actually $5 million per quarter earlier, than the end of calendar year '25, are we still looking for 6 to 9 months to achieve $600 million? Or we imply it's a...

I'm not going to put a stake in the ground right now. And so it was more directional. It was a directional statement about kind of the time line between when the products hit. So it's really about products and CapEx coming in to drive that level of growth. The other thing is we see customers pulling in and pushing out, making road map changes even across an 800G line depending on different components. And so I think just more as a general rule, it's 6 to 9 months rather than a stake in the ground.

Got it. So second question is about OCS. So are we looking for like a particular customer in list? Or are we looking at multiple customer will adopt this?

Multiple customers.

Okay. I think that is all of the questions that we have. I would like to invite you to visit us at the booth. So our CTO, Dr. Matt Sysak, will be hosting office hours today. There will maybe homework involved, but there will be office hours at our booth at 1:30 to 3 p.m. today or tomorrow from 10 until noon. So yes, if you could come during those time windows that would be great, then we can be sure to have Matt on hand to answer your questions about our demos. Thank you very much for attending. Really appreciate your interest.