Lumentum Holdings Inc.

Good day everyone and welcome to Lamentum Holdings first quarter fiscal year 2025 earnings call. All participants will be in listen only mode. Please also note today's event is being recorded for replay purposes. At this time I would like to turn the conference over to Kathy Ta, Vice President of Investing Relations. Ms. Ta, please go ahead.

Thank you and welcome to Lamentum's fiscal first quarter 2025 earnings call. This is Kathy Ta, Lamentum's Vice President of Investing Relations. Joining me today are Alan Moke, President and Chief Executive Officer, Wajid Ali, Executive Vice President and Chief Financial Officer, and Chris Cauldron, Senior Vice President and Chief Strategy and Corporate Development Officer. Today's call will include forward-looking statements, including statements regarding our strategies, trends, and expectations for our products and technologies, including demand, our customers, our end markets and market opportunities, our expectations and beliefs regarding recent acquisitions, including CloudLight, macroeconomic trends, and our expected financial and operating performance, including our guidance, as well as statements regarding our future revenues, financial model, and margin targets. These statements are subject to risks and uncertainties that could cause actual results to differ materially from our current expectations, particularly the risk factors described in our SEC filings. We encourage you to review our most recent filings with the SEC, particularly the risk factors described in our most recent 10-K and in our 10-Q that will be filed soon. Forward-looking statements provided during this call are based on Lamentum's reasonable beliefs and expectations as of today. Lamentum undertakes no obligation to update these statements except as required by applicable law. Please also note that unless otherwise stated, all financial results and projections discussed in this call are non-GAP. Non-GAP financials are not to be considered as the substitute for or superior to financials prepared in accordance with GAP. Lamentum's press release with the fiscal first quarter results and accompanying supplemental slides are available on our website at .lamentum.com under the investors section. With that, I'll turn the call over to Alan.

Thank you, Kathy. Good afternoon, everyone. In the first quarter, we exceeded the high end of our guidance for both revenue and earnings per share. We set a new record for Datacom laser chip orders, including 200-gig EML chips, reflecting strong demand from multiple customers, including an AI infrastructure customer. Based on expanding cloud demand and improving trends in the broader networking market, we expect double-digit sequential revenue growth in the second quarter. During the first quarter, we advanced our strategy to expand and diversify our cloud and AI opportunities. We secured an additional hyperscale transceiver customer with a new qualification and initial volume orders. This is beyond the new award we highlighted last quarter. We expect to start shipping volume production against these new customer awards in the first half of calendar 2025, and they will ramp through the year, consistent with the revenue targets we set out previously. Over the past few quarters, I outlined a three-prong strategy to grow our cloud data center business. As a reminder, the first prong is to expand our cloud and AI customer opportunities at the component and transceiver levels as customers migrate to higher speeds. The second is to scale significant production capacity for components and transceivers outside of China. And the third is to advance our differentiated technology roadmaps, enabling customers to future generations of cloud and AI data center architectures to hire compute capacities in a more cost-effective and power-efficient manner. I would like to share more details on recent progress in each of these areas. Regarding the expansion of our customer opportunities, as the industry transitions to higher speeds, our differentiated technology becomes increasingly valuable, especially to market and technology-leading customers. For example, in the coming year, the transition to 200G lanespeeds will drive growth and increase the importance of single-mode optics and indium phosphide laser transmitters. This shift aligns well with our market and technology leadership. Our indium phosphide EML transmitters have established a strong reputation for high performance, high quality, and reliability. Underscoring this, our 100G EMLs are currently shipping in high volumes to a wide range of optical transceiver suppliers for use in leading edge, single-mode 400G, and more importantly, 800G optical transceivers. These customers are now designing our 200G EMLs into their next generation of transceivers, positioning us well for the upcoming transition to 200G per lane. We have also made significant progress in expanding our opportunities on the high-speed datacom transceiver side of our business. As we previously forecasted, our datacom transceiver shipments are expected to increase sequentially this December quarter, and we expect our transceiver production to continue growing throughout calendar year 2025, driven by demand from multiple hyperscale cloud and AI customers. As I mentioned earlier, we received an initial set of orders from new hyperscale customers and expect to start shipping production transceiver volumes for these new customers in the first half of calendar 2025. In addition, we are actively working to finalize additional awards for multiple other opportunities. The second prong of our cloud and AI datacenter strategy is to expand our manufacturing capacity at established Lumentum facilities outside of China. This expansion is essential to ensuring a secure and reliable supply chain for our cloud and AI customers. Our Indium-phosphide laser chips are critical to datacenter infrastructure. As discussed in our last earnings call, to meet the growing demand, we are on track to increase EML production capacity by 40% in Q4 of fiscal 2025 compared to our capacity in Q4 of fiscal 2024. This expansion will help alleviate the industry-wide shortage of Indium-phosphide capacity. However, we still expect to be on allocation throughout calendar year 2025. Our datacom transceiver capacity expansion at our Thailand campus is progressing as planned. The first production line is now operational and we anticipate completing additional expansion phases over the next 18 months to meet the high volume of demand from our customers. This includes completing construction of our new three-story cleanroom facility on the same campus, which remains on schedule. The third prong of our cloud and AI strategy focuses on delivering innovative technologies to address the escalating challenges of scaling datacenter compute capacity. We are collaborating with leading edge customers to develop breakthrough solutions to enable higher data link capacities with enhanced energy efficiency that will support their multi-year cloud and AI infrastructure roadmaps. Inside the datacenter, optical switching is expected to be critical for future generation cloud and AI network architectures. Lamentum's power-efficient, high bandwidth, and low latency optical switches are well positioned to meet the demands of these evolving networks. We have already shipped evaluation units to customers who have provided overwhelmingly positive feedback on our performance. While we are beginning to ramp our 200G laser chips, we are already working in close alignment with customers on technologies for future generations of yet higher speed optical links, such as 400G per lane, and new architectures including co-packaged optics, which will require unique ultra-high power lasers. Our advanced in the embossed light and platonic integrated circuit capabilities are essential for meeting these upcoming demands. While we don't expect the deployment of these technologies to start until our fiscal 2026, we are actively collaborating with customers to shape the future of optical technology. Executing our three-pronged strategy will drive significant revenue growth fueled by new opportunities in AI and cloud datacenters. This growth, combined with a recovering network business, positions us to return to a double digit operating margin as our capacity utilization improves and we maintain strict cost discipline. As we outlined at the OFC conference last spring, we are targeting an operating margin of 17 to 20 percent once our quarterly revenue surpasses $600 million per quarter. Now let me move to additional fiscal first quarter revenue and product highlights. Our first quarter cloud networking segment revenue grew 11 percent sequentially and 23 percent year over year, driven in part by strong end market demand from cloud hyperscale customers as they invest within and outside of AI datacenters. We are also encouraged by continued growth in shipments of our newest networking products. This combined with improving inventory levels of our products at our networking customers is encouraging as we look to the future. We had a sequential increase in demand for our narrow line with tunable lasers used in high speed long haul applications as well as 400 ZR and higher speed solutions in DCI applications. With our existing design wins in this area, we anticipate maintaining a leading market share in laser components for ZR and ZR plus applications this fiscal year and beyond. Customer interest in our advanced coherent transmission and next generation transport solutions is on the rise. In leading edge coherent transmission, we are experiencing robust demand for 130 and 200 gigabaud coherent products. This demand is driven by the need for greater capacity and spectral efficiency amid continued bandwidth growth. Our differentiated technology, manufacturing capabilities, and strong design win momentum position us to maintain leadership in these critical products as well. Complementing growing demand for our coherent transmission components, we are seeing increased demand for integrated C plus L band transport solutions, high port count rotums, and the emergence of demand for multi rail rotums and amplifiers. Multi rail is the next step in efficiently scaling the optical transport layer and offering increased flexibility in high capacity line systems. We anticipate strong sequential growth in our cloud and networking revenue in fiscal Q2 due to rapid growth in our products addressing cloud and AI applications and broad based improvement in demand across our product portfolio. Now let me move to our industrial tech segment. As expected, our industrial tech segment revenue grew 2% sequentially while being down 38% from the same quarter last year. Like others in this space, demand continues to be challenged due to the weak industrial and market. In industrial tech, we are focused on developing innovative industrial laser products to meet customers' needs for higher precision with minimal heat damage. This creates opportunities for our ultra fast laser products in heat sensitive applications like semiconductors, displays, and advanced packaging. We are working closely with leading semiconductor equipment manufacturers in interposer and advanced semiconductor packaging applications. We recently delivered high and low power femtoblade demo units for advanced display applications and we have received positive customer feedback. Looking to fiscal Q2, we expect industrial tech to be approximately flat sequentially due to an uptick in industrial lasers led by our ultra fast lasers offset by a sequential decline in 3D sensing revenue. To summarize, we have made significant progress in executing our strategy to grow our cloud business. We set a new record for datacom chip orders and are working diligently to fulfill this demand over the next several quarters. We also made excellent strides with multiple high speed optical transceiver customer engagements, including securing meaningful transceiver orders from a third hyperscale customer. Additionally, we are actively working to secure more awards from more new customers. Our robust pipeline of cloud customer engagements combined with improving trends in the traditional networking market reinforces our confidence in achieving our previously stated target of growing quarterly revenue to $500 million per quarter by the end of calendar year 2025. We anticipate continued significant growth into 2026 and 2027 as we capitalize on new cloud and AI opportunities, positioning our cloud business for a multi-billion dollar annual run rate in the coming years. Before I turn the call over to watch it, I want to express my sincere gratitude to our employees for their unwavering focus and dedication and to our customers worldwide for their partnership and collaboration. With that, budget.

Thank you, Alan. Before I outline our Q1 financial results, please note that consistent with our discussion last quarter, we have changed our presentation to include certain charges that were previously excluded, primarily related to abnormal access capacity in our non-GAAP results starting in the first quarter of fiscal 2025. Comparable historical periods have been recast to align with this presentation for consistency. First quarter revenue of $336.9 million and non-GAAP EPS of 18 cents were above the high end of our ranges. Gap gross margin for the first quarter was 23.1%. Gap operating loss was 24.5%. And gap net loss per share was $1.21. Turning to our non-GAAP results, first quarter non-GAAP gross margin was 32.8%, which was up sequentially and flat year on year on higher revenue. In future quarters, we anticipate company gross margins will sequentially increase as manufacturing utilization improves due to an improving telecom outlook, as well as an increase in data com laser shipments. First quarter non-GAAP operating margin was 3%, which was up sequentially and year on year. First quarter non-GAAP operating profit was $10 million and adjusted EBITDA was $37 million. First quarter non-GAAP operating expenses totaled $100.4 million or .8% of revenue, a decrease of $1 million from the fourth quarter and down $1.9 million from the year ago quarter. Despite increased investment in our expanding cloud opportunities, Q1 operating expenses were due to restructuring actions that were taken during fiscal 2024, as well as overall stringent cost controls across the company. Q1 non-GAAP SG&A expense was $37.7 million. Non-GAAP R&D expense was $62.7 million. Interest and other income was $4.6 million on a non-GAAP basis, driven by interest earned on our cash and investments. First quarter non-GAAP net income was $12.2 million and non-GAAP diluted net income per share was $0.18. Our fully diluted share count for the first quarter was 69.1 million shares on a non-GAAP basis. During the first quarter, our cash and short-term investments increased by $29 million to $916 million. Our working capital came in as expected, with inventory levels increasing sequentially to support the expected growth and revenue across our cloud modules business. In Q1, we invested $74 million in CAPEX, primarily driven by investments in high-speed transceiver capacity additions at our Thailand manufacturing site, as well as indium phosphide wafer production capacity. Turning to segment details, first quarter cloud and networking segment revenue at $282.3 million increased 11% sequentially and 23% -on-year. Cloud and networking segment profit at 13% increased sequentially and -on-year. Our first quarter industrial tech segment revenue at $54.6 million was up 2% sequentially and down 38% -on-year. First quarter industrial tech segment profit of 4% increased sequentially and decreased -on-year. Now let me move to our Q2. We expect net revenue for the second quarter of fiscal 25, which is on a non-GAAP basis and is based on our assumptions as of today. We expect net revenue for the second quarter of fiscal 25 to be in the range of $380 million to $400 million. This Q2 revenue forecast includes the following assumptions. Cloud and networking to be up sequentially with strong growth in products addressing cloud applications and improving networking customer demand. And industrial tech to be approximately flat sequentially with increased industrial laser shipments offset by a decline in 3D sensing. Based on this, we project second quarter non-GAAP operating margin to be in range of .5% to .5% and diluted net income per share to be in the range of 30 cents to 40 cents. Our non-GAAP EPS guidance for the second quarter is based on a non-GAAP annual effective tax rate of 16.5%. These projections also assume an approximate share count of 69.6 million shares. With that, I'll turn the call back to Kathy to start the Q&A session. Kathy?

Thank you, Wajid. To allow everyone an opportunity to ask questions, please keep to one question and one follow-up. Now let's begin the Q&A session.

If you would like to ask a question, please press star followed by one on your telephone keypad. To remove your question, press star followed by two. And if you are using a speakerphone, please pick up your handset before asking your question. Our first question today comes from Sameek Chatterjee with JP Morgan. Your line is now open.

Hi. Thanks for taking my questions and congrats on the strong guide here. If I can just ask on the guide first, I think the implied increase -over-quarter in cloud and networking is about 50-60 million -over-quarter. Can you sort of parse that for us in relation to how much of that is telecom improving versus maybe data-com chips versus data-com modules if you can actually point us to where most of that improvement is coming from and have a follow-up. Thank you. Sure. Thanks, Sameek.

This is Alan. I'd say it's a mix of all of the above. I'd say chip growth is not a lot until the next couple of quarters because the capacity comes in increments and chunks. And so, as I said in the script, the Q4 number will have a step up, but not a lot of chips increased this quarter. The telecom is modest and I don't expect that to be strong growth other than things that are going into DCI applications, but there's a large pick-up on the data-com modules and we expect that to continue, as we said, in the script through calendar 2025.

Got it. Got it. And for my follow-up, you announced another sort of new customer win. Now you have, I mean, if I take and aggregate the two new customers you've announced in recent quarters, how should we think about the magnitude of the opportunity maybe after a full ramp with both of them compared to what was the original customer for CloudLight? How should we think about the magnitude relative to the first original customer there when we take the two new customers together? Thank you.

Yeah, I think that's hard to answer in that it's going to come down to our ability to execute and provide the kind of quality and delivery and capacity that the customers need. So the new customers, as I said before, will start ramping in the first half of next year and it'll be ours to earn their growth of their business in the second half. And I think we're positioning ourselves very well. We're having the capacity in place, as I said, with clean rooms as well as equipment. So we're planning for success, but we still have to earn that. But that said, these are all very, very large customers that consume a lot of transceivers. And so getting in the door is step one and earning bigger share is really what we're striving to do now. Okay, great. Thank you. I'm putting

my question.

Our next question comes from Simon Levin-Pose with Jeremy James. Your line has got open.

Great. Thank you very much. I wanted to dig a bit more into what's going on with the CloudLight business. Reflecting back on when you reported your third fiscal quarter last year, you had talked about sort of the cadence of that business where we'd see a bit of a pause in June and September, which I think we've seen and sort of the snapback in December. And so what I'm trying to get an assessment of is sort of what's changed in that December outlook versus what you thought several quarters ago? Is it above, below, and what's different in terms of informing that expectation on CloudLight? Thank you.

Thanks, Simon. Yeah, nothing's really changed. I think it's coming in as expected as we've talked about the transition of new products. We're ramping the new products now. And so it's kind of come together as we had expected and as we had talked about on our previous two earnings calls. So nothing unexpected from my perspective.

And then just as a follow-up, I'd like to get your perspective on how you see the market for 1.6 terabit devices evolving and what you would anticipate your role would be in that market, really sort of timing and magnitude perspective. Thank you.

Yeah, you know, that one is a tougher one in that lots of interest from lots of customers for 1.6t and in fact, you know, it's kind of 1.6t is one version of the next generation of products. There's also 800 gig using 200 gig lanes. And so we're going to participate at the component level, as I talked about in the script with our 200 gig EMLs. And that ramps really more towards the summer of next year. But we're in the qualification stages today and have received volume orders today as our capacity is quite constrained. So we'll play at the component level with EMLs and mostly and then we'll play at the transceiver level. We've sampled transceivers to customers and feedback very, very positive as we talked about on the last call. So, you know, we're in the hunt, we're in the game and we just need to continue to work on the qualifications. And then the timing is really up to the customers as well. So, you know, some are ready sooner than others. But I'd say that by the end of next calendar year, 1.6t should be, you know, ramping in a significant way.

Thank you. Thanks,

Simon. Our next question comes from Metta Marshall with Morgan Stanley. Your line is now open.

Thank you. This is Conjiv Akar on for META. Just quickly and congrats on the quarter. I guess, are there any changes to your timeline for vertical integration with CloudLight? I think you mentioned sort of no shipments until the end of fiscal 25, but just sort of updated timeline on your vertical integration there?

No real change. I mean, we're building products with our lasers and EMLs and VIXELs to be qualified. But I'd say, you know, the timing is really unchanged in that products that we were working on before the acquisition are coming to market now. And those mostly didn't change. But I'd say that, you know, we have certainly opportunities in calendar 2025 to have more product in feeds. And that should help margins, you know, late in the calendar year and into calendar 2026.

Okay, that's helpful. And then a quick follow up on the telecom side. Are you sort of expecting sequential growth primarily from new products or products that didn't previously had inventory build up? I think you mentioned that last quarter, I guess, how do you expect traditional telecom products to track and sequentially across the year? Thank you.

Sure. I mean, I'd say that, you know, we're seeing dramatic strength in anything ZR, anything to connect data centers as data centers are being built out. And that can take the form of ZR modules. But given our share of tunable lasers that go into ZRs, that's where we're going to see a dramatic pickup in the telecom side. So, you know, I'd say that the hyperscale growth is very solid outside of the data center. And that includes transport or open line systems. I'd say that carrier spending is still muted and inventory still needs to get burnt off there. But, you know, there are signs of light. So new products are flowing out the door, as we talked about. And I think we're a quarter or two away from having returned to more normal inventories. But we're seeing certainly good signs of inventory reductions of our product at our customers.

Thank you. Our next question comes from Monado Barua with Loop Capital. Your line is now open.

Okay, yeah, guys. Good afternoon. Thanks for taking the question. We'd just love to get your thoughts, Alan, on how to think about, you know, tariff positioning, tariff risk, hot topics, you know, kind of these days. So get your thoughts there in the context of diversifying out of China. And then I have a quick follow-up as well,

thanks. Yeah, I don't think our strategy is changing based on who is going to be in the White House. We've been on a strategy to, you know, to be a U.S. headquartered company with manufacturing outside of China. And that is not changing as a result of, you know, potentially increased tariffs. So I would say that it positions us very well, as we've had an established manufacturing facility that we're growing dramatically in Thailand. And our laser chips come out of the U.S. as well as the U.K. and Japan. So from that perspective, I think we're as tariff-free as you can get with respect to our future. So I don't think any strategy change is happening. But I'd say that that resonates with a lot of our customers who want to have supply chain resilience. And we provide that for them, both with vertical integration, as well as, you know, a limited footprint in China and a growing footprint outside of China.

That super helps the context. Thanks. And then just a quick follow-up is just on EML's lasers, given this constraint position, is there any pricing power that you guys have or that you're using to control it? Thanks. And that's it for

me. You know, I think we look at pricing on a portfolio basis. And I think that as things are constrained, they go down slower or go up faster. And so I think in the constrained environment, as we look forward, and as we sit here today, I think, you know, we have to pay for a lot of capital is going into add capacity. So I think that we're justified in looking at price optimization on EML's as one area that we are looking at and have implemented some strategic pricing for those products.

Thanks, Anand. Thanks, guys. Appreciate it,

Alan. Operator?

Yes, Sierra. We're ready for our next question.

I apologize. My line was muted. Mike Genevieve from Rosenblatt. Your line is not open.

Okay. Thanks very much. I guess that's for more color on the telecom market. You know, and when you make comments about 26 and 27 and that kind of visibility, is that all the data, data, data, com side or how far does the help the telecom visibility extend?

I mean, I think the telecom, traditional telecom or carrier telecom, the visibility is not great. Although I would say that the deployments have been below the speed of bandwidth demand growth. And so there is going to have to be some additional spending and deployments on new networks. And we're seeing that today in China as they're rolling out new networks. Visibility of telecom carrier network spending in 26 and 27. I'd be kidding you if I said I had good visibility to that. I think more so the data center and the data center interconnect markets, we believe is going to be very strong in the future. And frankly, we have a fairly small footprint and small share of data com transceivers. And with our US headquarter and manufacturing outside of China, I think there's a compelling reason for customers to come our way. So we expect to not only grow our data com module business and email chip business, but gain significant share through the next coming years.

Great. That's very helpful. And then finally, I guess, can you give me more color on the optical switch product in the data center? Like what's going to do in the data center? And is this kind of a completely new function? Or is it, you know, kind of like the next? Will it replace anything or replace part of anything? And I'm specifically thinking about internet switches, whether there's anything there.

Yeah, Mike, this is Chris. Let me take a stab at that. It's performing obviously the name optical switching. And there are, as you know, in a data center, there's a tiering multiple levels of optical switches to get the large scale that's needed to connectivity across a very large data center. And so the ability to produce high radix, so 300 by 300 kind of level switches, I guess you could argue replacing some level of electronic switches, I think that a little bit more of an and in that it enables the data center to scale more efficiently so that the power consumption that's, you know, in today's switches or the switches that you would use had you scaled today's architectures, then can be used for more GPUs and more networking, optics, etc. to interconnect AI accelerators. So that's really the idea here is that to bring a lot more energy efficiency when you're talking about big pipes going across a very large data center. The other advantage also to be clear is that as you add layers of switches, do have added latency and optical switching has very little latency. So that's also a big advantage as you increase the size of data centers. Sorry, Mike, you were saying? Okay, thank you.

Oh, well, I was going to say, so basically sounds like larger cluster sizes of GPU, as they get larger and larger, you need larger optical switches, basically, just to sort of

Yeah, other uses and maybe a couple things to also point out that, you know, for us, this is a natural extension that's leveraging the technologies that we've pioneered in the Rotem space over past several decades. And so it's really an adapt, adapting of technologies that we've had to a new data center opportunity.

Perfect. Thanks so much.

Thanks, Mike. Our next question comes from Tom O'Malley with Sparklays. Your line is now open.

Hey, guys, nice results. And thanks for taking the question here. So I wanted to understand the qualification process on the transceiver side. So you've now won another award. And you talked last quarter about kind of qualification in the first quarter time period with that first customer. Can you maybe refresh us here? So two new customers, are both of them qualified yet? And when they are qualified, when do you normally get a feel for the volume? Obviously, being qualified and the amount that you actually end up shipping is a different thing, but you're very confident to next year. So just can you remind us when are these customers going to be qualified? And then when will you have a feel for those volume forecasts for those customers?

Yeah, I'd say we're still on track with what we said last time with the first new customer that we talked about on the last earnings call with qualification happening in the first calendar quarter. I would say the new customer that we added this quarter is qualified now. And we received orders now that will be beginning to ramp in the first half of next year. So a little bit different timing, but we're happy with both. And as I said before, we've got a lot of other qualifications going on that we'll talk about on the next earnings call, I hope.

Super helpful. Just looking at the moving parts to kind of get to that $500 million a quarter, you guys have kind of expressed an increased confidence there, perhaps a little bit of bounce back. So you're kind of in the low $300 million a quarter right now. You talked about EML setting the record, I think previously you were at $80 million or so a quarter. So let's just assume a little bit above that. But you kind of talked about capacity being outstripped 2020, so you're basically sold out there. So when I look at where you go from here, the low $300 million to kind of that $500 million range is more of that incremental, the EML lasers or more of that incremental, the qualifications of new receiver customers and their ramping. Any hint there would be really helpful. Thank you.

Sure, sure, Tom. I think starting with our midpoint of our guidance for our fiscal Q2, which is the quarter we just guided, we're guiding to 390s, the midpoint. And I'd say that there's a combination of what we expect to win and grow between now and the end of next calendar year. As I talked about in the script, growing EML capacity by 40% from the June quarter of last year to the June quarter of this year, this coming fiscal year is part of that growth. But I'd say transceivers are growing. Anything regarding data center interconnect is going to grow. And we're not counting tremendously on a bounce back in the traditional telecom business with carrier spending. And if that happens, then we'll get there sooner. But I'd say that's kind of the three big prongs, which is growth in the chips, starting really in the summer time, growth in transceivers that we've talked about, really starting the December quarter, and then data center interconnect are really the three big drivers for getting us from 390s, the midpoint to the 500 million.

Thanks,

Tom. Our next question comes from Carl Ackerman with BMP. Your line is about open.

Yes, thank you. I wanted to double click on the hyperscale win as well as the design when you spoke about last quarter. Specifically, are these transceiver opportunities being driven by hyperscalers or large customers connecting their own AI processors? Or is it being driven by you being included on reference designs of maybe perhaps new GPU platforms? And as you address that question, are these initial transceivers now integrating your EMLs? If not, does that push out the ramp of these customers or hyperscalers ramping your optical products in 2025? Thank you.

Yeah, thanks, Carl. I think we don't want to necessarily get into customer architectures and what they're deploying. But we can certainly comment on our transceivers. And so these in general are transceivers that won't have our EMLs at initial launch, if you will, because these have been in development and designed over the past year or so. Obviously, these accounts have other opportunities, to Alan's point, as we succeed and execute with them. Not only will there be more share, but there will be more skews and other types of transceivers where we can introduce more of our own content. Okay, thank

you. Hey, Carl, did you have a final comment?

Yeah, I guess perhaps along the same lines then, could you discuss whether the new wins that you're receiving are driven by data center customers reallocating procurement of these optical transceivers to U.S. domiciled suppliers like yourself? I'm going to just comment on that. And then perhaps in addition to that, are you able to quantify maybe the expanding CAM that you see as these AI clusters that now have the option of disaggregating some of the components of scale-up architectures? Are you also seeing that as a new opportunity? Thank you.

So maybe add that these are new opportunities, so it's not a reallocation. I would say that customers are certainly like the fact that we are U.S. headquartered and non-China manufacturing and a long roadmap of that. So I think that makes us more competitive in winning in these new opportunities. These are definitely scale-out for the time being. The introduction of optics into more of the scale-up is probably a year or year and a half out kind of time frame. There's certainly people making progress towards that, but these are traditional pluggable transceivers for scale-out, hyperscale opportunities as people deploy a lot more AI gear, if you will. If you recall that, you know, if you turn the clock back a year ago, there were really only two folks building AI gear. And I think you'll have more, I think you brought up the word, to an extent some level of disaggregated architectures as we roll forward into calendar 25 and 26.

Thanks, Carl. Our next question comes from David Vought with UBS. Your line is now open.

Hey guys, thanks for taking the question. This is Brian on for David. Can you clarify if these transceiver customers were contemplated in your 4Q25 guide and then I have a follow-up?

4Q25 guide. Oh, yeah, we talked about calendar 2025 and getting to 500 million. You know, I think we have plenty of opportunities, whether this one or that one was, you know, in their number. We don't need to win all the ones we're going after. So it's hard for me to say that, yeah, we're right on track with these two that we had planned on. So I'd say that's a hard one to answer. You know, they certainly play a part in getting us there. And we're really, really happy with the progress we've made with these customers as well as other customers that we expect to, you know, be talking about on future calls.

Got it. That's helpful. And then for my follow-up, in the qualification process, you mentioned that you have to earn more volume. Does that mean you're a secondary provider in these deals? And then if so, what type of share would be reachable in your view?

Well, I'd say that, you know, in any of our businesses, we have to earn trust and earn share and that's done through quality execution and resilient supply chain and doing what we say we're going to do. And, you know, share does shift in all of our business as a result of performing well or not performing well. And so that's what I intended to mean by that discussion. I'd say that there are, you know, hyperscale demand is very big and no one relies on one supplier. So as they ramp up two or three suppliers, the ones that perform better are going to get more share in the long run. So that's what we're really working on to make sure that we can count on us to provide more of their product that they need and that's what we're working on.

Thanks, Brian.

Thanks for the call. Our

next question comes from Ryan Koontz with Needham & Co. Eli is not open.

Great. That's the question. I wanted to circle back to your comments around DCI and understand your exposure there on tunables, but can you maybe talk about what your opportunities in the module space there are? Are you present giving design wins in ZR today for ZR modules? And how should investors think about that opportunity for the company?

Yeah, so DCI exposure, I would say the biggest revenue exposure for us is what I said before, which is the tunable lasers and to a lesser degree other components that go into other people's ZR and ZR Plus modules. We do have exposure to our own ZR module and we're making very good progress on the 800 gig ZR and ZR Plus both domestically as well as outside of the US. So we're favorably encouraged by the progress there. I'd say the 400 gig module, frankly, we were late to market. We have small share and we're doing our best to try to frankly, the market is very, very strong and we're very happy in growing our tunable laser business there and supporting our customers who are winning significant share in a market that's growing very rapidly as data centers are moving further apart.

Yeah, that makes sense. It's a more mature market there. And then a quick question for Rajit around gross margins and kind of in lieu kind of a lot of puts and takes going on here. How should we think about gross margins as the fiscal year unfolds here?

Yeah, thanks, Ryan. So like we talked about in the script, we're expecting gross margins to improve sequentially. We're even seeing it in Q2. You can see that our operating margin profile is significantly different than what we how we performed in Q1. And so we're expecting to see that gross margin improvement move into Q3 and into Q4 as well as more of our supply comes online for EMLs. Having said that, we are expecting to increase our R&D spend in Q2 and into Q3 and Q4. You can probably tell from Alan's comments that the customer pull is strong and the opportunities are quite strong as well. And so because of that, we're probably going to be spending ahead of revenue from an R&D standpoint. And we've already talked about some of the capacity investments we're making. So yes, we'll see gross margins tick up sequentially through the fiscal year, but it'll be muted a little bit from an operating margin standpoint because of the increased R&D investment we're making just given the amount of customer pull we have.

That's great. Does that capacity you just talked about adding, does that capacity talk about, does that have an overhead impact also a little headwind there until it's fully? Yeah, it does. Yeah, thank you. Yeah,

thank you. No, it does have a little bit of overhead impact because we're building out in our Thailand facility as we move more of our production of transceivers to Thailand. And so as we move that up and ramp that up, there will be a couple quarters of overhead expenses associated with that. And so that's already contemplated in the sequential increases and margins. But then as that volume ramps up in the middle part of next calendar year, we'll be able to see the benefit of that moving through the quarter. So we'll explain more about that as the quarters happen. But thank you for asking about that. Sure, that's all. Thanks a lot, guys.

Thanks,

Ryan. Our next question comes from Christopher Rowland with SIG. Your line is not open.

Hey, guys, thanks for the question. So, Allen, I wanted to double click on your prior answer around vertical integration into lasers in particular, into cloud light. You said it mostly didn't change. Wanted to know what actually did change there. And then you suggested the second half of 25 for those margin benefits. So I would assume that's a timetable for integration. And perhaps I may have been wrong here in my assumption, but my assumption is that you would be more meaningfully integrating in the first half. So I'd love to know, is this a qualification issue? Is this not a delay? Was this just my understanding that's wrong? Is this a delay in the integration and what's causing it?

Yeah, so good question, Christopher. I'd say a couple things, one of which is customers don't want us to change product that's already in flight. That's number one. Number two, our laser capacity is constrained. So we have to make versus buy decisions around what mostly the product going through the transceiver business today is silicon photonics based. So we use a lot of CW lasers, not EML lasers yet in the products that we're shipping and released today. So we can buy those CW lasers externally, or we can use our very critical EML capacity to add those CW lasers into our products. We've done the math. There's a lot of good CW lasers suppliers. It makes more sense for us to buy those CW lasers and free up that EML capacity to ship to our customers than it would be to convert that EML capacity to CW lasers, for example. So that's one of the things that's pushing off that integration of CW lasers into our products until the second half. I'd say that we are working on EML based designs, and those will come to market in the second half of the calendar year. We have to get qualified and go through that. But today, most of the products that we're producing are silicon-photonics based using CW lasers from our strategic supplier partners to keep that EML capacity for our customers.

Yeah, and I'd add that the real performance advantage of EML starts to come in as we talk about 1.6T and future generations of higher performance 1.6T. So our natural roadmap also aligns with using more vertical integration where the technologies are much more differentiated at those speeds.

Thank you very much, Guy. That actually cleans up a lot for me, clears up a lot at least. It also probably brings me to my second question, which is trying to understand more about EMLs. I mean, it just seems like the whole market is moving here. AI demands inflecting units. There's one other big EML player as far as I've recognized at least. It seems like if you have this EML capacity available, it actually has pull into your transceiver units, pipeline, market share, customers, new customers, etc. So why are we talking about 40%? Why aren't we talking about 100% or even more here? Why is 40% the right number here? How capital intensive is this? If you wanted to, how fast could you get new lines up? Would you be leaning on overbuilding versus underbuilding, just given what seems like the importance of this part into this new cycle? Thanks.

Christopher, you sound a lot like my board of directors, so that's a very good question. It does take time to add capacity. One of the things that we did as we were working on integrating and getting synergies, we had two wafer fabs in Japan. We were going to close one of them and consolidate it. We stopped that several quarters ago and said we need that extra wafer fab capacity. We're ramping that up now. If you go back a year and a half ago, we had idle capacity in our wafer fab. It takes a year and a half to add capacity. That's what's going on. We are adding more. We're converting to larger wafer sizes. That takes time. We're adding epi actors that takes time as well. Very good question. If I had the ability to add 100% between now and June, I would do it. But that's the lead time of capital equipment and bringing up and qualifying new tools is just painful. The good news is there's barriers to entry in that market. We're going to grow it as fast as we possibly can. At the same time, we're going to introduce new products. The 200 gig per lane is a new product. We're getting qualified. That takes time and resources and starts typically at a lower yield. We have to get through that learning curve and that yield. I think we're going to do well in the second half of the calendar year on that. We are adding capacity beyond the 40% for sure after the June quarter. We're not sitting idle for sure.

Perhaps just one more follow-up. Why don't you use your limited EML capacity to bring new cloud-like customers online? It seems like that would be two birds, one stone.

Oh, yeah. That's a good question. I think we have silicon photonics, we have multimode Bixel-based transceivers, and we have EML-based transceivers. We are designing each of those for a given application. Some applications work better with single-mode silicon photonics. Some have a better cost structure with EMLs. As Chris said, a lot of the new next generation of 1.6C likes EMLs better, especially as you get into multi-wavelengths where EMLs can really play a key role there. We're absolutely going to do that. It probably makes more sense in the second half of the calendar year as we get into these more advanced 1.6T products. Then we have next generation 200 gig EMLs, which are really going to differentiate us from our competitors. I think that really gives us the ability to drive incremental differentiation at the transceiver level and drive higher gross margins.

Thanks, guys. Sorry for the third question.

Thanks, Chris. Sierra, I think we'll squeeze in one more question.

Understood. Our final question today comes from Richard Shannon with Craig Hallam. Your line is not open.

Glad to get in under the wire here. Thanks for taking my questions, guys. If you'll follow up on the topic of the optical circuit switching, here's a couple earlier questions and comments. Maybe you can just kind of delineate the time frame for qualification and volume ramps here just to get a sense there. Then also, your close peer who reported recently has obviously talked about a product necessary. Maybe you can discuss how you were differentiated or different or complementary to what they're doing.

Yeah, I can tell you about what we're doing. I can't really comment on what they're doing. I'm sure their technology is unique and interesting. I think our technology and our approach is different. Ours is MEMS-based. We have, as Chris mentioned earlier, years of experience with MEMS designs and MEMS in our WSSs. In fact, our most advanced WSSs have multiple technologies like LCOS and MEMS. We are getting positive feedback, as we said, from our customers. There are customers that are more advanced and more ready to put optical circuit switches into their data centers. Then there's customers that need to re-architect or write software to enable this optical circuit switching. It's really dependent upon which customers are ready when. I would say that a deployment in calendar 2025 is realistic. Really, the more meaningful growth will probably be in calendar 2026.

Okay, that's great detail. Thanks for that, my second question.

I was just going to add one thing, Richard. We've already transferred the product to our Thailand operation. We're making qualification units and beta units in Thailand today. We'll be ready when the customers say go.

Okay, great. Thanks for that, Alan. My second question is on a topic that has been asked in a while, and that's 3D sensing. Is there anything going on here, either in your past markets or new markets here that's going to give us any signs of growth and improvement anytime soon?

Hey, Richard, this is Chris. I would say a little bit of yes and a little bit of no. I would say that obviously the high-volume applications that we have been playing into for several years is mature, and there's not an obvious catalyst in that specific application other than if the Android world begins to adopt in mass. No suggestion that they will in the short term or midterm, but that's always a possibility. I would say as we look to other opportunities in metaverse, augmented virtual reality, and applications like that are helping a bit as well as the automotive space, but both of have been a little slower to adopt than we've all hoped. But there are folks continue to invest strongly in that, and as a leader in the technology, we continue to partner with those leaders, but we'll keep things muted in terms of thoughts on that until that new application really comes along.

Thanks, Richard. Great.

Thank you,

guys. I think Sienna, yeah, thank you. Sienna, I think we'll pass the call back over to Alan for some concluding remarks.

Great. Thank you. I'd like to leave everyone with a few thoughts as we wrap up the call. We are excited about our growth prospects, our comprehensive photonics portfolio built on differentiated in-house technology and proven manufacturing capabilities addresses critical challenges like connectivity bottlenecks and power consumption. With strong progress on our three-pronged strategy, we are highly confident that we will reach $500 million in quarterly revenue by the end of calendar 2025 and in the long term expand our cloud business to a multi-billion dollar annual run rate. Thank you for joining our call today. We look forward to seeing you again at investor conferences and upcoming meetings this quarter. Have a great day.

Thank you. That will conclude our conference call today. Thank you all for your participation. You may now disconnect your line.