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

Good day, everyone, and welcome to the Lumentum Holdings Special Conference Call for investors. [Operator Instructions]. Please also note today's event is being recorded for replay purposes. Following the presentation, we will conduct a question-and-answer session. [Operator Instructions]. At this time, I would like to turn the conference call over to Kathy Ta, Vice President of Investor Relations. Ms. Ta, please go ahead.

Thank you, and welcome to Lumentum's Special Conference Call to discuss its announcement of entering into an agreement to acquire Cloud Light. This is Kathy Ta, Lumentum's Vice President of Investor Relations. Joining me today are Alan Lowe, President and Chief Executive Officer; Wajid Ali, Chief Financial Officer; and Chris Coldren, Senior Vice President and Chief Strategy and Corporate Development Officer. Today's call will include forward-looking statements, including statements regarding our expectations regarding the acquisition, including market opportunity, macroeconomic trends, trends and expectations for our products and technology, our end markets, market opportunities and customers and our expected financial performance, including statements regarding our future revenues, financial model and margin targets and statements regarding our expected first quarter operating results. These statements are subject to risks and uncertainties that could cause actual results to differ materially from those projected. We encourage you to review the risk factors described in our most recent filings with the SEC, including Lumentum's 10-K for the fiscal year ended July 1, 2023. The forward-looking statements provided during this call are based on Lumentum's current expectations. Lumentum undertakes no obligation to update these statements or any other forward-looking statements. Please also note that unless otherwise stated, all financial projections discussed in this call are non-GAAP. We have not provided a reconciliation of these projections as one is not available without unreasonable efforts due to the variability and unpredictability of excluded items. Non-GAAP financials are not to be considered as a substitute for or superior to financials prepared in accordance with GAAP. On today's call, we will reference a presentation on today's acquisition announcement, which you will find posted on Lumentum's Investor Relations website. Please turn to Page 4 of this presentation. With that, I'll turn the call over to Alan for his comments.

Thank you, Kathy, and good morning, everyone. Today is a very exciting day, and we have a lot to cover. Earlier this morning, we announced that Lumentum will acquire Cloud Light, a leading provider of high-speed fiber optic transceivers and active optical cable products to hyperscale cloud customers. This acquisition is expected to accelerate Lumentum's growth in the coming years and will be immediately accretive to our EPS. It squarely positions Lumentum as a leader in providing photonics to cloud operators at a time when artificial intelligence is strongly accelerating. The addition of Cloud Light is expected to result in a more than fivefold expansion in our served opportunity inside of cloud data centers. This is also expected to more than double our revenue from cloud operators over the next 12 months after close. Cloud Light provides the highest speed optical transceivers in the market, which is where photonics to cloud operators is growing rapidly. Nearly all of Cloud Light's more than $200 million of revenue in the last 12 months was derived from 400G or higher speed transceiver sales. In the most recent quarter, over half of Cloud Light's optical transceiver revenue was derived from 800G modules. The combination of our companies enhances security of supply for their customers with a broader combined global manufacturing footprint and enhanced opportunity for component level vertical integration. In addition to immediately and significantly expanding our participation in the fastest-growing portion of the data center opportunity, this transaction positions us to address future customer challenges around network bottlenecks that impact AI data center workload scaling. It combines Lumentum's advanced vertically integrated chip-scale photonic integration, RF electronics and data transmission technologies with Cloud Light's in-house highly automated, high-volume optical assembly production platform, along with an excellent transceiver design team. Along with Lumentum's optical switching capabilities, these are all essential ingredients to create the innovative optical interconnect and network solutions essential for data center workload growth. We believe we can rapidly scale to even higher volumes of these advanced transceivers needed by hyperscale cloud and AI infrastructure customers. The combination is also expected to have broader benefits for Lumentum. It increases our direct relationship with and relevance to cloud customers who are increasingly driving product road maps, volumes and new consumption models for our telecom products. In addition, at the network edge, our tunable DWDM transceivers are increasingly becoming cost-effective solutions for next-generation wireless and distributed access architecture networks. With Cloud Light's low-cost, high-volume transceiver manufacturing platform, we expect we can further expand the application of our DWDM transceiver products. Please turn to Page 5. To better illustrate our strategic rationale for this transaction, let me first provide some context. AI models are driving an exponential increase in compute requirements, where performance is now doubling every 3 to 4 months compared to the historical doubling every 2 years according to Moore's Law. The number of parameters in a typical AI model has grown over 100,000 times in the past decade. This rise in complexity requires more and more compute power to train and deploy these AI models. Additionally, the sheer number of AI models being trained and deployed is expanding dramatically. For example, according to Google's public comment, the number of AI models trained on Google Cloud TPU v4 Pods has increased by over 100 fold in just the past year alone. Finally, the speed at which these complex AI models can be trained and deployed has become a critical issue as more companies and user activity embrace AI to make real-time decisions. The relentless demand for compute power to train and deploy AI models is training traditional compute resources. As a result, many companies are turning to cloud computing platforms to meet their AI compute needs. Cloud platforms offer scalability, elasticity and on-demand access to compute resources, making them ideal for AI workloads. Now please turn to Page 6. AI workloads are both enormously data and compute intensive. This drives new AI supercomputer architectures with computation distributed across thousands of parallel network processors, which are specialized GPUs, TPUs or CPUs. These processors exchange data with each other and high-bandwidth memory for a large number of high-speed data links across numerous compute cycles. The time communicating across these links and networks is frequently a significant fraction of the overall job completion time. A public study from Meta estimates that 5 to 10x more network bandwidth is needed in data centers that are architected for AI workloads. All of this drives the need for very large numbers of the highest data rate -- data links within data centers and AI clusters. Please turn to Page 7. This growth of AI computing and data centers is, in turn, driving an explosion in demand for leading-edge photonics products, in particular, the strong demand for 800G optical transceivers. 800G transceivers offer today's highest data linked capacity, making them essential for supporting growing heavy AI workloads. Recent third-party research indicates that unit shipments of 800G transceivers will soon surpass those of 400G transceivers. By 2028, industry forecasts predict that cloud data centers will require 26 million 800G units with another 5 million 1.6 terabit units on top of that. This surge in high-speed transceiver demand is almost entirely due to the proliferation of AI computing in hyperscale cloud data centers. Now please turn to Page 8. Third-party research also estimates that total revenue from sales of high-speed transceivers utilized within data centers will grow from $2.4 billion in 2023 to more than $10 billion in 2028, a 5-year compound annual growth rate of more than 30%. This rapidly growing demand is focused on the highest speed portions of the cloud intra-data center transceiver opportunity, including 400G, 800G and 1.6T transceivers as well as active optical cable or AOCs, for short-reach intra-data center applications. Cloud Light's product portfolio aligns very well with these trends. With this transaction, our addressable opportunity within cloud data centers also expands significantly. Historically, we have focused on providing industry-leading laser transmitter chips such as EMLs, which comprise approximately 10% to 20% of the total transceiver revenue opportunity. By acquiring Cloud Light, we expect a fivefold expansion in Lumentum's Cloud Intra-Data Center Served Opportunity. Now please turn to Page 9. Cloud Light has a strong track record of manufacturing leadership and delivering innovative products quickly as evidenced by a strong position in 800G transceivers. Nearly all of Cloud Light's more than $200 million of revenue in the last 12 months was derived from 400G or higher-speed transceiver sales. In the most recent quarter, over half of Cloud Light's optical transceiver revenue was derived from 800G modules. Cloud Light has achieved and consistently maintained its leading time to market and volume scale and capability through high levels of automation in its optical module and transceiver manufacturing processes. Their products have been qualified by a number of top hyperscale cloud customers. Please turn to Page 10. With this transaction, we've become a clear supplier of choice for intra-data center applications with a broad and expanding product portfolio. For intra-data center applications, we will supply the full suite of required laser solutions and high-speed transceivers and AOCs. The breadth of our combined product portfolio also creates vertical integration opportunities that will provide customers with higher performance and lower cost solutions. As we mentioned in our March Investor event at OFC, we anticipate a new and rapidly emerging opportunity for optical circuit switches or OCSs within data centers. Lumentum is well positioned to provide OCS solutions to leading hyperscale customers, given our core capabilities in optical switching, amplification and monitoring developed for longer reach networks. Serving the complex switching fabric within hyperscale data centers is another exciting opportunity beyond what we have already covered today. Outside of the data center, we serve cloud customers with our full set of ROADM-related products for metro and long-haul efforts. And following our FY '23 acquisition of NeoPhotonics, we now have a leadership position in coherent components for pluggable ZR and ZR+ modules that are used in data center interconnect applications and a growing ZR module business. All our cloud data center products and product road maps reflect Lumentum's continued commitment to investing in fundamental technologies that anticipate customer needs and deliver the highest performing, highest quality and highest reliability products. Now please turn to Page 11. As you have seen, we believe we're extremely well positioned to address the growing needs of our cloud customers as they scale to meet the demands of AI computing. This transaction combines Cloud Light's strong track record of manufacturing and first-to-market product leadership with Lumentum's commitment to lead in R&D innovation over the long term. We will lead the industry's transition to 800G and 1.6 terabit transceivers, including high-speed, short-reach active optical cables and we will provide solutions for both Ethernet and InfiniBand applications. Our cloud customers are responding very positively to this transaction, and they appreciate the rationale for the combination. We're looking forward to benefiting from having a broader, more strategic supplier and product development partner once the transaction closes. Please turn to Page 12. To summarize, acquiring Cloud Light expands our cloud intra-data center served opportunity over fivefold as we will serve not only our existing laser transmitter customers, but now the optical transceiver opportunity as well. This large served opportunity is expected to grow at a CAGR of over 30%, driven by AI workloads requiring deployment of an escalating number of the highest speed transceivers, such as 800G and 1.6 terabit products. We expect the opportunity for these high-speed modules to exceed $10 billion in 2028. This transaction will be immediately accretive to our earnings per share and accelerate our revenue growth. We also expect our cloud intra-data center revenue to more than double in the 12 months following the deal close. With that, I will now hand it over to Wajid to provide more details on the financial terms and benefits of this transaction. Wajid?

Thank you, Alan. Good morning, everyone. I would like to echo Alan's sentiment about the significant benefits of this transaction. This combination accelerates Lumentum's top line growth, creates opportunities to benefit from economies of scale and to become a broad solutions provider for our data center customers. Please turn to Slide 13. Under the terms of the transaction, which has been unanimously approved by Lumentum's Board of Directors, the total transaction value is approximately $750 million, subject to certain adjustments and will be paid in cash and the assumption of substitution of Cloud Light outstanding unvesting options at the time of closing. This combination is expected to be immediately accretive to Lumentum's non-GAAP earnings per share and is expected to more than double our cloud intra-data center revenue in the 12-month period following the close. During our due diligence, Cloud Light showed us their impressive growth where they delivered more than $200 million in revenue in the last 12 months at low teens non-GAAP operating margin and high teens EBITDA margin. We expect Cloud Light revenue to grow at a greater than a 30% CAGR over the next 3 to 5 years and even faster in the near term. We also expect that Cloud Light's non-GAAP operating margin percent will improve to the high teens within 24 months of the close through synergies derived from increased manufacturing utilization and efficiency, increased leverage of Lumentum components in cloud-like transceivers and the leverage of our combined global manufacturing and engineering teams. Because of the anticipated synergy scale with the high expected revenue growth, we're providing margin expansion expectations versus a specific dollar figure for these synergies. We currently expect the transaction to close by the end of calendar 2023, subject to required regulatory approvals and other customary closing conditions. I would also like to provide a brief comment regarding Lumentum soon to be announced Q1 financial results. We expect our Q1 results for revenue to be approximately $317 million, which is above the midpoint of our guidance range and EPS to be above the midpoint of our guidance range as well, which we provided during our August earnings call. As Kathy mentioned at the start of this call, we will defer further comments about Q1 results and our Q2 guidance until our upcoming earnings call, which is scheduled to take place on November 8. With that, I'll turn the call back to Kathy to start the Q&A session. Kathy?

Thank you, Wajid. Before we start the Q&A session, I would like to remind everyone that during today's call, we will be addressing questions that relate to this acquisition announcement, and we will not be making further comments on our Q1 results or Q2 outlook. I would also like to ask everyone to keep to 1 question and 1 follow-up. This should help us get to as many participants as possible before the end of our allotted time. Now let's begin the Q&A session.

[Operator Instructions]. Your first question comes from the line of Simon Leopold from Raymond James.

Great. Could you give us a little bit of background in terms of Cloud Light's essentially business model. I'm presuming that they are basically a manufacturer and not vertically integrated and wondering, have they been a historic Lumentum customer? Just a little bit background, maybe including sort of where are their factories, how -- what's sort of the square footage? I'm just not that familiar with them.

Yes. Thanks, Simon. Let me try to tackle as much as I can. Cloud Light has been a small customer of Lumentum, and that's one of the avenues by which we started learning about how impressive they were, especially at the very highest speed transceiver design and manufacturing capabilities. The primary manufacturing site is in Dongguan, China and have some operations in the Philippines. They design -- design teams are in Hong Kong and in Taiwan. I don't have the square footage, but I just would say, as we said in our prepared remarks, the team is extremely impressive. Their manufacturing automation is very impressive. And I think the combination between Lumentum and Cloud Light really provide an opportunity for further component vertical integration as well as driving to the best possible cost for our customers.

Great. And then just maybe a quick follow-up. You did mention that they've got sales into multiple hyperscalers. Maybe if you could give us a little bit of insight in terms of any degree of customer concentration, have they had greater than 10% customers? And do they have any customers that are meaningful, but not hyperscalers? So somebody perhaps manufacturing an element of an AI cluster or something like that.

I don't want to really comment on their customers per se, but given that there are only a handful of hyperscalers and to your point that the customer partner that does do a lot of the GPU cluster work, I think we'll save that for when we become a little bit more familiar with how they perform. But I would say that the combination does enable us to broaden the sales to further customer opportunities. And I think, that's why I'm very excited about the combination. Given that there are only a handful of customers, of course, there's 10% customers at their revenue. But we'll see as we combine the company -- will it be a 10% overall customer, and that's -- we'll let you know as we further integrate the 2 companies post close.

Your next question comes from the line of Samik Chatterjee from JPMorgan.

Maybe if I can just ask you to clarify something first, and I'm sure you went through this and I didn't understand this completely as -- does Cloud Light do the full transceiver module or the transmitter IC, which you're sort of highlighting as 10% to 20% of the revenue opportunity, given transceiver? And the reason I ask is I believe you were in the transceiver module business and decided to pay back that business over time. And so how is this sort of opportunity different from the datacom module business that you were in earlier? And I have a quick follow-up.

Yes, sure. Samik. This is Chris. So yes, Cloud Light does produce or design, develop, produce full optical transceivers as well as the optical subassemblies that go into those transceivers. And then kind of turning back to our history, yes, we previously had a datacom module business, but it was much more oriented towards, let's say, the enterprise world of the past. So therefore, less cost competitive, fast moving relative to the needs of today's hyperscalers. Cloud Light is very different. They're at the leading edge of cost and speed and agility, very well regarded by the hyperscalers. As well, Lumentum has changed quite a bit since we were last in the datacom module space with regard to having a much larger and innovative transmission business from which lots of technology is being developed. We've revamped our datacom chip capabilities, and we've brought online a large in-house factory in Southeast Asia from being essentially a start-up at the time we got out of the datacom module business to being a preeminent photonics factory across the industry. So a lot of things have changed in the industry world, and ourselves since we exited the datacom module business.

Okay. And -- sorry, for my follow-up, if I can just ask when I think about the competition for this, are the typical competitors, the ones we think of like the Innolights and Coherents of the world? And if that's the case, how you're thinking about sort of differentiation in that market as the lowest cost producer? Is that the differentiation and to enable that is it really more automation and capital investments? Or is it more going down the path of contract manufacturing? What's the sort of way you sustain that differentiation?

Yes, Samik, this is Alan. I'd say twofold, one of which is first to market design and technology that Cloud Light has shown, really does enable them to capture the most leading-edge requirements from customers. And so we were super impressed with their capabilities there. And as we said in the script, over half of their transceiver revenue in the last quarter was 800 gig. And I would hesitate to think anybody has that type of mix from our competitors. I'd say we're still very committed to supporting our other transceiver customers with datacom chips, and that's going to continue to be a major growth engine for us as well. But that's no different than how we treat our telecom customers as well. We'll support them with low-level components, subassemblies as well as entire products. So I'd say the competitive advantage is really first-to-market technology in manufacturing that is super impressive that will drive low cost. And then you combine the 2 companies where we will have a lot more product in feed to drive vertical integration and cost reduction, I think that. And then what Chris talked about having our own manufacturing expertise in Thailand for potential growth areas as they're outgrowing their existing footprint and we'll need to expand dramatically post close.

Your next question comes from the line of David Vogt from UBS.

Great. I just want to follow up on the comment that Alan and Chris made regarding your datacom and laser business, your chip business. So post transaction, are we to assume that there's no change in your relationship with, say, like and unlike, going forward? And that you have enough capacity to meet the demands of obviously, your internal growth projections, as well as your customers? And then I'll just give you a follow-up while I have you. You mentioned, obviously, you're excited about having manufacturing capacity in Thailand and other parts of Southeast Asia. Is there an opportunity to rationalize the footprint to maybe extract more profitability and/or synergies out of the disparate footprints that you currently have from the manufacturing capacity?

Sure. Let me address the first question. The vast majority of the laser chips we provide to Innolight and Coherent and others is really based on our EML technology and the capabilities that we have in providing them with the leading-edge EMLs for 400 and 800 gig today as well as VCSELs and VCSEL arrays for those shorter-reach applications. And we're going to continue to support those customers with those products and technology. The majority of the -- well, I'd say all of the transceivers and active optical cables that Cloud Light manufactures today are based on silicon photonics as well as VCSELs and VCSEL arrays for those shorter-reach applications. So it's not really a competitive offering per se, it's more of a complementary ability that we can provide VCSELs and high-power CW lasers for the Cloud Light transceivers and that we can continue to provide EMLs and DMLs for the rest of the market. Chris, do you have anything to add on that?

No, I think the rest of the question around manufacturing footprint, I would just highlight that our focus is on growing the business right now and post closing of the transaction. We will support the growth in existing footprint. And as Alan highlighted, then expand perhaps in other existing Lumentum footprint to handle the rapidly growing demand.

Your next question comes from the line of Meta Marshall from Morgan Stanley.

Great. Maybe first question. Just is there any increase in capital requirements, just as you guys look to kind of expand our capacity over the next year that we should think of as being a part of the model? And then second, I noted that you guys were hoping to close this by the end of the calendar year. Can you just kind of give us a sense of what China approvals would be necessary as part of this?

Sure. Let me address the capital. As we've talked about the rapid growth, Cloud Light has made significant capital investments over the past 12 months to position for that rapid growth that we see as Wajid talked about 30% CAGR, but in the short term, even faster than that. So capital has been deployed. I'd say that as we look forward and as we get more close to closing and the integration planning, we will have a better picture of capital requirements and how we might be able to continue to drive more efficiencies there. So I wouldn't say that it's more capital intensive per revenue dollar than typical, but we'll learn more as we get a little closer. As for the regulatory requirements, the vast majority of the revenue for Cloud Light is outside of China. So the main regulatory requirement is with U.S. HSR, and that's why we think we'll be able to close before the end of the calendar year.

Your next question comes from the line of Alex Henderson from Needham.

So I was hoping you could talk a little bit about the percentage of their revenues coming from active optical cables versus the percentage coming from longer reach understanding that in AI environments, there's a higher percentage of AOCs in that environment? And within that context, can you talk a little bit about what subassembly components they manufacture that are -- how to say it, specific to them, i.e., internal content? Or is it just simply a matter of the vertical integration of these components that they're supplying from other people?

Alex, this is Chris. So today, I would say the majority of their revenue is coming from silicon photonic based transceivers. So they give us sort of the intermediate or longer reach distances at least relative to inside data center distances. I think the shorter reach transceivers and AOCs are more of a longer-term growth, as you said, as AI architectures starts to increase the utilization of optical cables versus electrical cables. So Cloud Light is well positioned in both from a product standpoint and a capability standpoint, but the mix today is leaning more towards the intermediate to longer reach silicon photonic modules. From a standpoint of, I think, as you mentioned, subassembly components versus vertical integration, I would say the primary driver of the cost synergies between the 2 of us is driven by the vertical in-feed opportunity we have, given the nature or sort of the BOM cost, how it breaks down. So the optical components that we will be able to provide into their transceivers.

I'm a little confused by the answer on the first part. I thought you were saying that virtually everything they're selling is inside the data center, datacom product. And your answers seem to imply that some of that was for longer reach outside of the data center. Am I -- can you clarify that, please?

Yes. Sorry, Alex. What I meant by longer reach was not the short reach, tens of meters type applications. So longer reach inside the data center is hundreds of meters to a kilometer or so. That was the implication.

Okay. I got it. I understand. And then 1 last question, if I could. One of our companies we follow has gone from essentially 0 to $140 million in quarterly reported AI-related revenues in the span of about a year. And that trajectory is expected to roughly double over the next year. And clearly, that trajectory is tied to the spectacular growth rates in AI that has been reported by people like NVIDIA. Is that the type of trajectory that we're talking about here? Are they tied into those order flows? And therefore, it's really not a function of demand, but rather a function of capacity availability. And to that extent, if that's the case, has the orders been put in place to build out the capacity to produce these products, 3, 4, 6, 8 quarters in advance, which is the lead times for the systems that required to manufacture them?

Yes, Alex, I hesitate to think that we're going to go from 0 to $140 million in a period of 4 quarters. But I would say that as Wajid indicated, the short-term CAGR is very steep. So you're spot on with respect to capacity being the gating item, not demand. And so, Cloud Light has been running the business very effectively being able to ramp significantly over the last few quarters and do have orders for future capital needs to satisfy the order intake that they have and expected forecast from new customer qualifications that are in the works. So I wouldn't say that there's any concern on our part with respect to them having the capacity to meet the needs of -- or try to meet the needs of the very, very strong demand that they're seeing in their business. I would say that they're going to run out of factory footprint before they run out of demand. So that's why I think the combination really does lead to a 1 plus 1 equaling more than 2.

So the orders for the equipment has been put in place, this is what you're saying?

Absolutely.

Your next question comes from the line of Ruben Roy from Stifel.