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

Great. Thank you so much. My name is Alex Henderson. I'm the Needham networking and security analyst. It's a pleasure to have Chris Coldren and Kathy Ta from Lumentum here. We're going to do a fireside chat for the next 35 minutes. If you have a question, there's a dialogue box. I've got that open and can see the question. I will relay it as soon as I can feather it into the conversation. And if you prefer, you can e-mail me at [email protected]. And with that, welcome, guys.

Yes. Thanks, Alex. Thanks for hosting us here at the fireside as well as at the overall event. It's been great.

So we've got a lot of stuff to talk about, but let me start off with kind of an observation and an overview, which is you guys are more like a semiconductor than anything else in the sense that you're a component company. And so I ask investors all the time, when is the right time to buy a semiconductor company? And it's usually -- the answer to that is usually at the trough of the cycle. And it feels like we're kind of there in the sense that the inventory correction has been run its course and some of pieces of your business in the process of finishing, bottoming out and others. The 3D sensing has come down, that all of which has pushed your multiple up even as the stock has been weak because your earnings have come in with it. But as is the case with semiconductor companies, you normally want to buy them when their multiple is through the roof because that's when they turn around. I keep on reminding myself that somebody bought the semiconductor industry the day Lehman went belly up. And by the time the market troughed in March, it was up 25%. So somebody's got a lot of nerve to do that. And I think that's the right way to approach you guys here. So let's step through some pieces of the business. The first one that we really need to get out of the way is the 3D sensing piece. This thing peaked at about $500 million in revenues. In our model, we have it down to around $100 million in revenues and with 50% operating margins, that's a huge hit to the business. The good news is that it seems like it's gotten small enough so it ceases to be a hit to your business and there are other 3D sensing opportunities other than the one with your prime customer. Is it reasonable to think that that's going to start to level off and even maybe a year or 2 out, start to grow again, not so much in the smartphones, but rather from the other stuff offsetting the decline in the smartphone business.

Sure. I mean I think we continue to invest in product development in this business line. And despite the, as you highlighted, very significant sort of share normalization that's happened over the last several years, it is a profitable business line. We architected the manufacturing in such a way that is more of a variable cost model so we can make money even in a much lower revenue level. And as you allude to, there are applications, whether it be in other consumer customers or as we look to the automotive space and industrial space that are adopting laser-based imaging and sensing. I mean the most obvious that people look at is automotive LiDAR and so supplying laser chips into multiple LiDAR module manufacturers, if you will, that go into automotive space, both in North America, Europe, but as well as in Asia is an area that we're involved in as well as emerging industrial applications where maybe the volumes are a little less, but you're able to leverage more of Lumentum's product portfolio to pull through more of an imaging and sensing module, for lack of a better word, for manufacturing applications and customers that are not too different than who we supply industrial lasers to.

So the answer to that question sounds like, yes, there are some other opportunities that can then start to flatten that out or maybe even turn it up at some point.

For sure. Absolutely.

Now within the industrial tech business, the other product line is the commercial laser business, that also has been under duress along with a lot of other production technologies. There seems to be a couple of dynamics there. One, your primary customer is Amada and Amada looks like they've changed their focus from high-end, highest quality products to more of a value-based laser because of macro pressures and that caused a little bit of a mismatch in the inventory that they had on their books and a shift to lower-margin product from you. Is that played its course yet? Or is there a couple of more quarters of pressure from that trend? And when do we start to see that flatten out and start to turn back up?

Yes. Well, thanks, Alex. That's a really astute observation about the product mix shift and what's happening there. And so I think maybe rewind slightly and say, as the macro market was up into the right, people were selling and per -- model were selling, customers were purchasing very high-performance, high-end systems. And as markets slowed down a little bit, there was an earlier shift than we had anticipated, and we were developing a product very apt for this more cost-sensitive application, but we weren't quite there in terms of the cost structure, but our customer is our partner, and we said, okay, let's move forward with this and switch to be able to keep sales going, if you will, with regards to the high-end kind of turning off and this mid-range solution turning on. So over the next 12 months, we're very focused on driving cost out of that solution, if you will, to be able to improve margins and follow our commercial laser or fiber laser business over many years. We know what to do there. It's just -- these are very hard products to initially make and then you spend a period of time ramping them in production and driving cost out of it. We just got a little ahead of where we wanted to be because the market shifted a little faster than we expected. But yes, I think over the next 12 months, you'll start to see an improvement in the gross margins that are coming from that business that kind of compress due to this effect. I would say more on the macro recovery, a little hard to say whether that's a calendar '24 event. I think it probably stabilizes during calendar '24 and then maybe starts to grow as you exit the calendar year. It's kind of I think of our cloud versus networking versus industrial, they tend to have slightly different time constants with the industrial piece being kind of the slowest to decline and then the slowest to recovery, whereas the cloud was the earliest to go through an inventory correction and then the most rapid to recover and telecom is kind of the goldilocks in the middle.

Yes. Just to remind people on the 3D sensing, which I forgot to mention, there is a seasonality to that business. I think you did, what $44 million or something like that in the September quarter. It's expected to be down, what, $10 million-or-so in the December quarter. Then the March and June quarter is a seasonally weak window, [indiscernible] in the low 20s and then it pops back up in the September, December quarter, again, that's the normal seasonal pattern to that. Am I in the ballpark on those numbers?

Yes. I think you're describing the seasonal pattern in the rough order of magnitude probably relatively well and nothing to suggest that as we sit here today, obviously, we don't guide that product line nor that time period, but nothing to suggest in the market that there's anything unusual going on that would cause an atypical seasonality, if you will.

Perfect. Well, we've now talked about 20% of the business for the most of this call. So that's -- we got to move to the other 80% that's much more interesting. So let's talk about the telco piece first, and then we'll go to the Datacom and the Cloud Light acquisition. Telecom has been a challenging end market. First, we had a supply chain problem, we couldn't get enough parts, that built up a backlog that, in turn, resulted in a big backlogs at a lot of the systems vendors that in turn resulted them over-ordering and that, in turn, resulted in their balance sheets getting very heavy 300% to 400% above normal inventories at a lot of these OEMs, whether that's Cisco or Arista or Juniper or Nokia, Ericsson, pick your name. So when I look at the telecom piece, not only do we have a correction in inventory at the telcos, but we also have a correction in inventory at the OEMs. How do we think about how that plays through relative to your telco piece? And can you split between transmission and the more non-speed related products?

Sure, sure. So I would say that the -- what we are seeing now, which is over the past 9 months-or-so now since our -- probably December quarter of last year is where we had our highest telecom and datacom revenue, both as supply -- our supply to us became stronger and we were able to ship more, but we had also recently closed the NeoPhotonics acquisition. But since that period of time, as you alluded to, our customers have built a lot of inventory over the prior -- probably several years, frankly, out of security supply concerns through the pandemic, et cetera, and in anticipation of, as you said, the big backlog they built. So they've been reducing that inventory down given our lead times and ability to suppliers are quite healthy at this point and so desiring to reduce their balance sheet is only very natural. So that's why when we continue to reiterate, we've come down much more in revenue while our customers are frankly probably, if anything, up in revenue over the same time period in many cases. And that's just clearly them burning off inventory of our product. We're under shipping end market demand. And when you look at it from a product-specific standpoint, I think 2 things come to bear. As you said, transmission and transport. I would say we were more supply constrained on transport historically than transmission. So that would suggest maybe there's more historical excess inventory in transmission than transport. But then secondly, our understanding of what is going out into the field is very transport heavy, if you will, because that's what goes out first. And then the transmission year tends to lag. So as we've seen in our own revenue, maybe transmission has declined a bunch more than transport revenues have declined. They've kind of not held in there. They're certainly down off of where they were historically, but they're a little healthier than transmission revenues, and we expect that the same thing will happen that they will recover a little earlier and then transmission in bulk will recover slightly later, but to a much stronger degree with the one asteric that some of our telco transmission products do land in cloud end customers who are growing. So there is some undercurrent of transmission inventory also being consumed because the cloud end market is increasing in health and consuming inventory, right? We commented on the call that our own -- we supply components to folks building ZR modules. But even our own ZR modules are starting to tick upwards after having a softer several quarters.

So just to be clear, transport includes ROADMs, amps, [indiscernible], optical line systems?

Correct. Correct. And then transmission would include coherent transceivers, the components to build those transceivers or at the edge of the network, we call direct detect transceivers, the 10, 25 gig tunable transceivers that kind of covers our transmission portfolio.

Okay. So my understanding is the backlog of the systems companies is very heavily skewed to the transport piece and optical line systems could be as much as 40% of the backlog of orders that they are waiting on. Now the interesting part of that is that optical line systems are the first things you put in before you put in all the other stuff. And they are a precursor of demand for transceivers, which goes into the transmission side. If I put in an optical -- a lot of optical line systems over the next 6 months, 6, 9, 12 months later, doesn't that drive an uptick in demand for transceivers?

Yes. I mean, absolutely -- or maybe said differently, you have plans to put that much network capacity in. And the first thing you need to do is install the sort of common infrastructure, the line systems, the amplifiers, ROADMs, et cetera, and then the other will follow. So one is a leading indicator generally, not always, but generally future demand for transceivers or transmission gear in whatever format it goes into the network. And that's why we're very excited that this kind of inventory correction kind of comes off of a sort of a new network architectures that our customers have developed using our next generation, whether it be ROADMs or transmission gear has kind of just been delayed by -- between the pandemic and now this sort of hangover, if you will, from the pandemic with regard to inventory. But the real part of that deployment is really ahead of us as opposed to behind us. It's just maybe perhaps history suggests that we shipped a little ahead of it and now are suffering through that phenomenon. But as things normalize, then we should not only recover to current end market shipments, but then obviously, as you're alluding to, end market shipments should accelerate over the next 12, 24 months, and then that will cause further growth from the replenished sort of inventory level that our customers will need to build back of our products.

So if I think about telecom as a group, around $200 million in trailing sales in the September quarter. Fairly flat into the December quarter, maybe up a tick. Normally, it's down in the March quarter. But given what you've been talking about, I think you're thinking it might be up a hair in the March quarter? And then, the June quarter is probably the last quarter of declines on a year-over-year basis before you start to grow again in FY '25?

Yes. I mean I think being able to handicap exactly the sort of the shape of the recovery is someway a fool's errand, right, that these things can do. I think, as you've alluded to, that not only do investors buy the stock at that point in time, but also these markets can tend to flip overnight as we've seen what happened in our datacom business where you're sold out, then you can't get anybody to buy your product for a period of time. And then overnight, you can't make enough. We don't have perfect visibility to the exact timing of the recovery, but I think during calendar '23 is certainly our expectation, and the conditions are set up, that it should happen sooner than later. But again...

You've mean calendar '24, right?

Calendar '24 -- apologies, yes, calendar '24.

Yes. So by the end of the fiscal year ending June, that should be turning up at a minimum, though...

Yes. That's generally our expectation. But again, it could happen sooner or it could push out a little, but I think the view that our fiscal '25, i.e., starting mid-calendar year and into the next calendar year, we believe there's a reasonable probability that we're kind of through the inventory correction and getting back to much more normal run rates in line with our customer shipments.

Let's shift over to the datacom piece. So datacom, a pretty small piece of the big puzzle in the September quarter. I'm guessing somewhere around $30 million in datacom revenues. Kind of at a trough CPU-based datacom, not doing a whole lot, but starting to recover, aided by availability of the substrates, the platters. And then you did this acquisition that really changes things quite dramatically. So do you want to give us a quick thumbnail around the Cloud Light acquisition?

Sure. I mean, and maybe the segue into it was great talking about our chip business that over the last 4 years-or-so, we put a lot of effort into developing very performance, technology, quality, reliability, best chips in the industry that power transceivers, but that's only a portion of what goes into a transceiver. There's still a lot more value that gets wrapped around at those laser chips, if you will. So we've been trying to find the right asset at the right time to more fully benefit from the growth that's going on in the cloud space. And I think it was a great confluence of factors with regards to us having built up not only the laser chip and other component capabilities, but also really over the last N number of years expanding our manufacturing footprint and having a great set of operations in Southeast Asia and particularly in Thailand, where we have our own in-house vertical integration in manufacturing of products that are analogous to datacom transceivers. So with AI creating a large surge in spending, particularly at the highest end of the market, we say, okay, well, wow, our chip business now is very, very healthy, can't make enough for our customers, maybe this is the right time to acquire a business that's very focused on 400 and 800 gig and what's coming next after it and find another outlet for our component capabilities and technologies but also bring us closer to the table directly with the cloud customers who frankly are also becoming large and influential for all the other telco gear we supply, eventually optical switching will become a lot more mainstream within data centers, and that's a big opportunity.

Optical switching being ROADMs?

Or ROADM-like, in this case, it may not be wavelength selective but being able to reconfigure infrastructure within a data center optically is a very powerful concept that certain hyperscalers are using today and others will adopt, we believe, over time. So being able to have a stronger sort of first-person direct relationship with cloud operators was really important, not just because of the data center transceiver opportunity, but where everything else in our business is increasingly landing ultimately is increasingly going into the cloud space. So we found this company, Cloud Light, who has a long history in the datacom data center space that used to be part of TDK, SAE years ago then they spun-off to be more agile and focused on the cloud space alone. And over the last 4, 5 years of the company really knocked it out of the park in terms of creating best-in-class manufacturing technology and capability, optoelectronics packaging, which then enables great transceivers and not just what you're hearing about today is 800 gig silicon photonic-based transceivers but historically, they've been very strong as well in VCSEL-based transceivers as well. It just happens that silicon photonics is more apt for the applications that are growing today, but eventually VCSEL-based transceivers and cables will follow afterwards in much higher volumes than they are today. So we think we were skating to where the puck will be with regards to focusing on the highest speed portion of the market and an asset that punches above its weight certainly with regards to not dragging along a large lower-speed business that will be growth headwind, but maximally leveraging ourselves to the highest growth portion of the market. Then on top of that, obviously, the component and fee capabilities we have between lasers, RFICs and other photonic components is very high. And so therefore, we're able to make Cloud Light once integrated fully into Lumentum much more profitable, but also from a customer value proposition, they see a supplier that has a more secure supply chain, given it's in-house and we manufacture more broadly globally, so I think it checks a lot of boxes of taking a private company based in Hong Kong to now being part of what we hope or aspire to be a trusted partner of our customers here based in the United States. It really transforms who Cloud Light is in many ways as they appear to and are able to service a broader customer base over time.

So you've said that Cloud Light growing at about a 30% clip trailing revenues around $200 million. But that really kind of -- it's a little misleading in the sense that there's a big difference between the first quarter and fourth quarter on those numbers. My understanding is that, that was a $70 million to $80 million in the fourth quarter of their fiscal year or trailing year. And the growth rate is actually accelerating as the mix shifts more and more to 800 gig. Am I in the ballpark of thinking that, that's the right kind of scaling.

Yes. Well, I would say we haven't disclosed the most recent quarters. We will certainly, when we announce our December results, obviously, you'll see it in the numbers. But I think your -- it's a fair statement to say that on a $200 million trailing 12 months that -- the first quarter of that year was a lot less than the last quarter of that year. So the run rates are higher than the implied average of $50 million a quarter. And certainly, we have great aspirations to grow. And we said 30% CAGR is over a multiyear period, but obviously, with 800 gig really sort of just launching in the last few quarters and really not hitting its drive perhaps for another few quarters from a growth rate standpoint. There's lots of opportunity to grow with the current customer base that the current design sockets that Cloud Light has. And then in parallel, we've got a lot of effort underway and a lot of customer receptivity around more design-ins on other sockets as well as design-ins that were underway that maybe -- the transaction helps push a little more share our way and provide a little more certainty of outcome with those customers. So it's been very well received by all of the hyperscale customers. And they both provide that feedback to us and to the Cloud Light team who -- they're very proud they've built this great business, and they want to see it also grow significantly, and so they're very excited about what they're hearing from customers as well.

Well, so from a $30 million quarter in '23 and $150 million worth of datacom last year for the full year, it seems pretty reasonable that by the end of fiscal or calendar '24 and into '25, this could easily be $150 million to $200 million a quarter business. Is that a reasonable objective?

I would say that the market opportunity is there and we have the capabilities. I think it's just now we got to go win more business, and then we'll see.

And the margins here are pretty low to start with. Can you talk about the margin improvement opportunities? I think they're, what, 10% to 15% range currently. Could you get them up into the 20s?

Yes. So what you're referencing are the operating margins for the business. And I believe absolutely that: one, they are smaller in scale. So as volume goes up, there's significant operating leverage; but then secondly, the product in-feed synergies that bringing more and more momentum photonic content going into their transceivers is a significant gross margin, and therefore, operating margin lift. So we believe being able to increase operating margins north of 20% is certainly a reasonable objective given the nature of the products and the nature of the components that we will be able to supply into over time. Now not all of those components are available right now, some we have to develop or customize for their needs. And then secondly, often customers push off qualifications until there's a like a product transition, right? You're going from Gen 1 to a Gen 2 of a system or a platform or in this case, a certain network architecture in a data center. So our ability to drive in margin improvements will be kind of probably sequenced a little bit off of new product introductions than sort of requalifying old designs, but there could be some of that as well. So that's why we highlighted that these synergies, these product in-feed synergies, probably take 18 months to 2 years to say with confidence that we can get them all under the belt because that's, in some ways, the product life cycle rolling off and a next-gen turning up, but there's some opportunity to do a little bit earlier. As I said, some customers, they're not worried about us cost reducing, that's on us, right? But what they are worried about is security of supply and as we bring more component -- Lumentum component capabilities into the transceivers, that's perceived as improving the security supply, which it really is, and that is a justification for them to maybe requalify products and get us ramped up with our own content in a little sooner. But those -- we're only sort of a week into this deal being closed, so I don't want to get too ahead of. That's something that we'll be able to talk a lot more about as we make more progress.

Just mechanically, there's a couple of adjustments to the income statement that come out of this. The first one is, you've got the OpEx kicking in for them for 2/3 of the quarter and the December quarter and another 1/3 in the March quarter. So that takes you from about $100 million of OpEx up to around, what, $107-ish million in the December quarter and then with 1 more month's worth of add-ish in the March quarter, it gets up to, what, $112-ish million something like that and then fairly steady at that level?

Yes. So we don't necessarily guide operating expenses, but I would say that those are not unreasonable assumptions in the near term. Obviously, as you've seen historically, the March quarter tends to be a little higher on operating expenses, given there are some new calendar year payroll tax and fringe benefit resetting, if you will, that occurs at that point in time. But yes, there's only 2 months of Cloud Light operating expenses in the December quarter guidance, and so there will be a little bit of incremental OpEx due to Cloud Light for the full quarter. Then the seasonal impact of the March quarter and then conceptually OpEx could come down a little bit in subsequent quarters. But depending on where the top line is. If the top line is growing, then obviously, OpEx could stay more flat as you were alluding to.

The other offset would be on the interest line, which were $16.8 million in the September quarter. I think it drops down to around $7 million-ish in the December quarter and then a little bit more hit in the March quarter, again, from a full quarter impact.

Yes. I mean we paid consideration, wrote the check, if you will, that came off of our balance sheet and it was earning a lot of interest. So one has to keep track of. There's now foregone or lost interest coming off the income statement. But more than compensating for that is the operating income that's flowing through the income statement. So now I think it's in some ways a healthier position to be in that the earnings per share is more driven by operating income than interest income on cash sitting on the balance sheet.

Not that I care to lose any interest income, I like interest income. But that point aside, I think it was clearly a good acquisition and accretive day 1. So we've only got 2 minutes left here. Maybe we'll just summarize it. So it sounds like we're kind of at the trough of the cycle. Datacom has already turned up, the Cloud Light acquisition to drive nice improvement. The drag from 3D sensing should be increasingly negligible if not even turn positive. The telco piece still has a couple of quarters' worth of correction in it, but it's very rapidly getting cleaned up and by midyear next year should be growing again and the margin should have a fair amount of leverage in it. So again, thinking like a semi analyst, you want to buy these things when they're at the trough, and I think that's where we're at here. So with that, I'd like to call it a wrap. Thanks so much, Chris and Kathy, for joining us and all of those who zoomed in. We appreciate you listening in. And again, I'm around if you would -- if you have questions on the company. Thanks.

Great. Thank you, Alex. Thanks for hosting us. It's a great set of meetings, and thanks, everybody, for attending.