Coherent Corp. (COHR) Earnings Call Transcript & Summary

Thank you, everybody, for joining us. Thank you very much, Mary Jane. Thank you, Sanjai. And we welcome everyone to this fireside with Coherent. I got to keep remembering it's not II-VI anymore. And I think we're going to have a really great next 40, 45 minutes here to really dig in to the company. So thank you both for joining us. Just a quick housekeeping here. If anybody has any questions for any of us, please you can either type in via the text box in Open-Xchange, and we can ask or just hit me direct, sometimes that's just the easiest at [email protected]. And with that, we can begin. I guess, Mary Jane, probably most people know you, but Sanjai, maybe if you could introduce yourself, that would be great.

Yes, sure. I'm Sanjai, I'm the Chief Marketing Officer for Coherent. I've been in this role since July of 2019. And we look at markets by market vertical. So our -- my team is organized by market vertical. And within the 4 broad industrial -- the 4 broad market groups that we play in today, industrial communications, electronics and instrumentation.

Great. and we will probably get into all of that. I guess with that, I'll throw it out to both of you. But I get a lot of questions around the optical cycle. Obviously, we have an economic cycle, we have a semiconductor cycle. And of course, we have an optical cycle. And they move in different kind of modulations. I guess I'll open it up to either one of you. I get maybe a lot of pushback on -- it's been a while before we -- since we've seen the last top of the last optical cycle. It's been a long time since the bottom, does that worry you guys? Where do you think we are in it? How do you kind of mix and match all these different cycles and describe where we might be today?

You can start if you'd like to, Sanjai?

Yes. Yes, sure. I can -- okay, thank you. I can start with datacom and telecom to optical on the communications front. So it is -- the demand is growing, things like ChartGPT results in more AI/ML, more AI/ML requires more intra data center processing, which all drives transceivers. So we don't -- we really -- I wouldn't call it a cycle, it's still up, up in a way, but maybe the slope of the curve changes a little bit, but it's still going up. And the same with telecom with the increase in 5G and broadband access, all the government incentives that are kicking in world over, all of that ultimately drives demand for trunk bandwidth, metro access, all of the optical networks. So it's still growth. So we -- at least we don't anticipate seeing the -- we don't see the top yet. We are still on the growth part.

Anything to add, Mary Jane, or that was pretty well said?

Well. No, he did a great job. I would say that those for at least 2 years, if not 3, we've said that we expected 5G to be a multiyear cycle. So while everyone is correct, but it seems like a long time since the last trough. We expected that actually. And I think because 5G enables so many things, not just whether or not it's making your wireless phone work, it by definition, was a longer build out. I don't know that the continuation of our growth here, as Sanjai was saying necessarily is perfectly smooth, but I do think it is continuing to go up, yes.

Great. And maybe that's a good lead in to Finisar. So Finisar before you bought them was kind of notorious for lower margin, high -- kind of higher volatility type business. And under your leadership, I think you've really smoothed out a lot of that. And it's really looked like kind of up and to the right. All very, very positive. Some wonder if there are some risks underneath the surface or whether this is going to continue. Do you think we're through that period of great volatility with -- under your management now? Or do you think we'll see periods of great volatility again?

Sanjai, what's your view?

So we've been -- I mean we do announce publicly our networking segment revenues. And you can see that it has grown significantly faster than the market. So that -- our datacom business, when you say Finisar, it's really -- it's mostly datacom, but there is also a telecom piece and there is some -- there was some 3D sensing piece as well. So -- but if you're looking at the datacom piece, I think our -- we've come a long way in the sense that the focus now over the last 3, 4 years has really been growing share in the highest end part of the datacom market, which was then 100G, now it's 200G, 400, 800 and 1.6 terabits. So that has been our focus. That's where those are the upgrade cycles that our largest customers, hyperscalers really want. And so we have retooled all of that activity to focus on that area. And I think that's what you're seeing. We are actually -- our high-speed business was, which we define as 200G and above, was a small fraction when we acquired Finisar. But now it's over 50% of all our datacom revenues comes from high speed. And we continue to innovate. We are doing -- you'll see a whole bunch of announcements and demos at OFC. It is -- all of it is really focused on the datacom side, it's the high speed. So -- and then on the telecom side, it's the Coherent transceiver business where legacy Finisar was a latecomer, but now we really -- we are making our own DSPs. We announced our own DSP for Coherent, our first products. So those DSPs will be demonstrated at OFC. And so the Coherent transceiver is a big growth vector for us. And so one of the fastest growing. I mean that market is growing at 35% or 37%. So overall, there are certain segments of the market that are growing very rapidly, and that's really where we are focusing on time and efforts on.

And how deep do we go with that DSP? Do we get to PAM-4? Do we get to -- where do you go with it?

So legacy Finisar always had IC capability. We have a whole bunch of ICs that we do. This DSP, the [indiscernible] product, which we announced some time ago, 6 months ago was our first coherent DSP. And this completely -- it's a purpose-built DSP. It enables a new -- it's a market creation effort. It enables a new segment of the market that did not exist before. So that's our first foray into DSP. We've been investing in DSP for over 3 years. And while we don't talk about product road maps, you can imagine that we are not just going to stop with one -- with sort of a one-stop, we are going to expand our DSP portfolio.

And I guess since we're on the topic, our last fireside was actually with Marvell, and they have Inphi. And they said something interesting. What they said was, so for example, your competitor Lumentum, on their last call, what they said, well, 2 calls ago, they said that they were having a 400G inventory issue. It was primarily around 1 large hyperscale customer and they were digesting that inventory. And then this quarter, they expanded that from 2 quarters of digestion to a full year of digestion. Marvell, I think also they kind of partner in some ways. There are some parts that all go into the same transceivers that go to the same customer. And so I guess my question for you guys. So Marvell said -- so Lumentum kind of said maybe this is 400G inventory digestion, moving to 800G later on this year or next. And then Marvell today said, well, we actually expect 800G to be pretty limited, and we're moving right to 1.6 T. Would you agree with any of that? Do you think that 800 could be abbreviated and there's a stronger move to 1.6?

Yes. I mean, Mary Jane, I can start it off and then if you can maybe support me a little later. So the cycles are not like a 2-year event. Everybody does 800G and then it goes on. These cycles are very long. We've been -- we are still growing our -- the markets still growing 100G, it's not over. It's still -- these cycles -- these are overlapping cycles. So we do believe 800G will be significant, but it will be a significant opportunity for us. And we -- it's not -- customers are not -- yes, they are interested in thing with the data. Datacom market is you've got different hyperscalers and they are in different segments of their upgrade and their approach is also different. So it's not a given that 1 company is doing very badly the rest of the other. So it isn't -- it depends on which hyperscalers you're exposed to and what product ramp-ups you're exposed to. So I think that's really where we are today. We -- and almost by definition, if we are growing share -- because we are completely vertically integrated. If we are growing share at the higher speed, it's coming at the expense of somebody, who's losing share. And if those suppliers are not vertically integrated, then the downstream -- the upstream suppliers into that value chain are going to be affected, right? I mean that's just natural. Mary Jane, I don't know whether you'd like to add more to it.

Well, I think also -- I mean, first of all, on the inventory part, I had a click on that you moved from inventory to the actual sale of products. I mean, Sanjai has that obviously completely correct. And also, irrespective of how people think they might move, sometimes when they think they're trying to do that, it doesn't actually work, right? So I don't know that someone's saying we think it will go this way makes it definitive. Sometimes it just doesn't go at that speed. But as Sanjai said, it really depends on which aspects you're associated with. I think there's no question though that we think it's going to be a continuous progression. By that, I don't mean to suggest it must go through 800 for a certain period of time, just that there is a natural progression to the higher data rates for sure.

Okay. Great. Maybe just finishing up here, like who are you guys taking share from? And then secondly, how do you feel about Broadcom buying back FIT, and we had Broadcom last night, talked to Hawk. Hawk suggested he's the one area of investment that he sees for 2023 for him is in the optics. So I guess, how do you feel about competition? Who are you taking share from? And ultimately, do you worry about Broadcom really investing in this area?

First of all, you're talking about them having bought it back in 2019?

Yes, I believe that, that was the time frame.

So our view at the time was -- I mean, if you remember, it wasn't exactly, I think, their press release. So it wasn't particularly clear. But from what we could surmise and Sanjai, I see what you think of this, they didn't actually buy back all the actual manufacturing and assembly. They bought back all the engineers. Because I think what Jamani has said for some time now on higher data rate transceivers, the integration of the integrated circuit, the laser and the electronics is actually very complicated. And that's a very difficult design to toss over the wall. This is what we're thinking, see if you can make it. You actually have to pretty much shepherd that through. And so when we bought Finisar, for example, they, at the time did not make integrated circuits, but they have an integrated circuit design team, because they were very, very involved in exactly how does this need to work. So it would be probably an analogy across to Broadcom from that perspective. And if they wanted to keep making them at all or whatever components they have going into them, it isn't probably in their best interest to make sure that they're having them go correctly. Sanjai, what would you say?

Yes, absolutely, Mary Jane. I mean this is something -- so I mean, you just said that they divested it and then brought it back. So there was a reason to bring it back. And the reason is really, it's almost imperative at the very high speeds at 800G and 1.6 tera that you're completely vertically integrated and you're able to talk with the different groups. Because if you're -- if I am buying a laser and I need to make a little tweak. I'm talking about laser manufacturer. They're going to take a few weeks to come back. And it's just a much different than when we do transceiver design, it's 1 team. It's 1 seamless team that works on it. And that becomes critical at higher speeds. And you could argue that as the higher speeds, what is today's higher speed is not so high next year, okay, there will be a learning curve. But the ones that are out there first would really have like a 2- to 3-year advantages and that would be my view.

I think Hawk also wants to eventually move to onboard optics in package and then eventually, silicon photonics. And I think they even had a demo or a product release around silicon photonics. I wondered if you had any thoughts on that?

Yes. I mean, silicon photonics is one of our tools in our toolkit. I mean we've got products that are shipping today in silicon photonics, both Coherent products as well as datacom products. I think if you're looking at specifically socketed CPOs or co-package optics, et cetera, again, these are places that we look and we invest as well. But we think the best application of co-packaging is actually in a pluggable transceiver. The idea of -- because ultimately, the market went disaggregated for a reason, right? The hyperscalers, they just will -- I just don't think they're going to be a hostage by 1 company that does everything there. There are a whole bunch of reasons why that's just going to be very hard. But I do anticipate some sort of a hybrid environment where there is the availability of interoperability and multi-vendor sourcing, but yet, you have the advantages of co-packaging, I think that's -- but for the next 4, 5, 6, 7 years until 1.6 tera is going to be a transceiver. It's going to be a pluggable transceiver. So that kind of tells you that, okay, you've got -- and we may have co-packaging built into the pluggable transceiver. We've got wafer scale integration platforms, chip-on-glass technology. There's a lot of stuff that we do, but we will -- they'll all feature within a pluggable transceiver.

You have chip on glass, and it works...

Wafer scale optics. Yes.

Is it shipping today?

No, it is -- we've got them in an NPI right now. So it is slated for some programs or is not shipping as of today.

So that's great. That vertical integration just really shows one of your advantages. Telecom, maybe talk about what's going on there. I know there is this fiber investment in the U.S., that could be some of it. But what are the big things to think about in terms of drivers over the next couple of years?

Yes. Disaggregation is a really big deal right now. I think there are the coherent pluggable market, there are more ports that go on coherent pluggables now than on the traditional fully integrated Black Box kind -- type systems. So coherent pluggables is a big focus for the market. So it's a very -- it grows very fast, and we expect to participate in that growth cycle. And disaggregation goes beyond coherent pluggables, [indiscernible] systems are getting disaggregated, amplifiers, a combination of amplifiers, monitoring, channel monitoring, optical time domain, reflectometry for monitoring the health of the fiber plant ROADM. The entire system is getting disaggregated. So when the system gets disaggregated, it just opens up opportunities for a larger group of component suppliers such as us. And those of us who are vertically integrated can offer similar solutions. So that's a big opportunity. We are also seeing SATCOM. We are selling into the Spacecom, SATCOM market. That is -- we are able to leverage our capabilities and in -- from our Aerospace and Defense Group, marrying that with our communications products. So we are able to -- that's -- everything from fiber, I mean with the Coherent acquisition brought us fiber -- advanced fiber capability. So these are the specialty fiber, erbium fiber, terbium fiber to make amplifiers for these high-power lasers for communication. So that -- so those are sort of the big drivers for our telecom business.

Great. We have a question here Telecom is growing very strong, just 30% growth in telecom versus NEM growth of high teens mean will be below market growth or can it keep growing faster than systems?

So is the question whether or not the growth in '23 is because we pulled it forward?

I think what they're saying is maybe yes. Yes, perhaps.

I'll let Sanjai answer most of it. But basically, what's causing telecom at least so far this year to grow faster is the fact that, it was on pause for such a long time waiting for integrated circuits. And as that supply chain began to ease, we were able to ship more of it. But Sanjai, do you have an overall view on how we expect to see growth just in a general matter going forward?

Yes. I mean we still -- our SAM is increasing as we do more disaggregation. That's sort of -- and so we do expect to grow. Submarine is coming up again in a way. The submarine market was kind of depressed for some time, now it's on an up cycle. So there are a few things that are going on. Our access networks our products that go into the access networks have also grown nicely. So I think it's a combination of what Mary Jane just said, we were -- we had constraints and prevented us from shipping, but there are also new growth areas that are coming up.

Great. I get a ton of questions here, and that's around silicon carbide. I guess, maybe let's just start with your plans there? I think you have some big numbers that you've thrown out in terms of some CapEx. Maybe talk about where you are today. I think you're on 150-millimeter or 100-millimeter going to 150. This is an area that truthfully, I probably not done enough research on. So maybe you can tell us where we are today? What you're investing in? And where are we going?

Right. So our company has doubled the capacity of silicon carbide every 18 months for probably the better part of the last 7 or 8 years. We can grow all diameters. We were the first ones to introduce the 200-millimeter substrate back in 2014 for power. The world, not us, the world is actually largely standardized on 150, which means all of the machines that by a silicon carbide substrate are fixtured to take 150. So if you think about putting an epitaxial layer on a substrate, that fixturing has got to be super precise or your epitaxi recipe is kind of messed up in the attempt to make a device. So we would expect that the world will, over time, standardize on 200, and we're ready to sell that today, sell it to some customers, but largely, the vast majority of customers are buying 150. We still have some customers on 100, because that's what they need. They don't need it at 150. And I'm sure that everyone that makes silicon carbide would probably say the same thing. It's impossible, but I don't think anyone in the world is buying only 200. But 200 or diameter expansion as a general matter, is absolutely one of the greatest ways to leverage capacity expansion, if not yield improvement, the -- what we would call the increase of the surface area, because not every material we grow is in a bull. Some of them grow up the side of a furnace, so we do cubic inches. But nonetheless, that absolutely is the way you would increase the capacity, and that's very important as the world broadly adopts the silicon carbide without a vision on how the world would have more supply, nobody would change their car lines. So we're very, very bullish on this market and are excited about how we're seeing it develop.

Yes. Like Wolfspeed is probably the leader in materials here. They've had a pretty aggressive few years. Your CapEx plans are nowhere near there. But I guess if you've been doing this forever, why not be more aggressive here overall?

Well, I think we feel that we're being as aggressive as needed for what the timing is that our customers are demanding. Keep in mind, we're probably some equivalent size to what we're selling into the merchant market. But will speak for many years have made devices and modules. And so their expansion plans are for all 3. We are just beginning to get into devices and in these early years of CapEx, most of ours is for the engineering. But nonetheless, as the market ratchets up, and we are already seeing that even in this fiscal year, that may cause us to change our plans. But for right now, we think we're being as aggressive as needed.

You've said, I think, 35% to 50% of your fiscal year '23 CapEx is going to go towards SiC capacity, I think I have that right. And then is that furnaces? Or what are we talking about?

It's generally growth stations, but also the back-end silicon carbide processing that's needed to finally make a finished substrate.

Okay. And then how hard is this? Wolf says they want to tell people that this is impossible. And it's extremely difficult. And some people will have silicon carbide substrates out there, but they won't be of the highest quality. And how do we balance all of that? And do you feel like you're going to have the quality that they say is required out there considering they have kind of a head start at volume capacity.

Well, actually, I think for purposes of the volume in the merchant market, we've been making and selling silicon carbide for 20 years. I think we have the market required extremely high level of quality, and it is very difficult to make. It's very difficult to make. Sometimes on some materials, excellent epi could make up for a kind of so-so substrate. That's really not the case with silicon carbide. The quality of the substrate is really important, and it's very difficult to make. It's also not a technology. Grown materials as a general matter, are not technologies that lend themselves to typical leapfrog of technology. So you can't grow 1 inch and then all of a sudden grow 8-inch the next day. If you can't grow 2 or 3 or 4 or 5, you have no ability to grow 8 inches. So -- and you have to be sure that you have perfect uniformity across that entire bowl. And if you don't have that experience, I mean, this is a classic, you start your process it's very, very carefully characterized, et cetera, but you close the door and you wait until it's finished. You can't go in and tinker with it. It's very, very difficult to make. We have seen customers in the past try to backward integrate into materials. And it isn't so much that they couldn't be very successful and sometimes they actually are to some extent. But over time, what actually happens is they determine that it isn't core. Because backward integrating all the way to the material is really far. I mean, think about us, we licensed GE's very excellent technology for power devices that they were probably not going to use much. I mean this is the world, if not the largest -- one of the world's largest maker of large electric motors and trying to backward integrate that far proved to be maybe something in the whole scheme of that entire massive company wasn't going to get funded, whereas it's all we do all day long. So I -- we agreed with Wolfspeed that it's very, very difficult to make.

Okay. Great. And then I guess, lastly, around this, how do you interpret Tesla's comments about using 75% less silicon carbide?

Well, as a general matter, I mean, it's not really news. I mean, we expected that. I mean I'm not sure that anybody built a model on, Tesla uses X,Y, Z so much silicon carbide and they multiplied it by the volume of the vehicles and got a market TAM. That is in how the market TAM was developed, and I'll let Sanjai talk about that. Anyone who was a first-generation designer is a brilliant company does marvelous things really can correct through the market in a very, very interesting and powerful way. But I would say also 9x out of 10, and this probably applies to us, too. Oftentimes first-generation products are a little bit overengineered, because you don't really know what you need yet. And so then you naturally true it up, plus the fact it would depend what power you want, lots of different things that will drive exactly how you design the car. So I know everyone seems quite up in the trees about this, but generally, I mean, as far as we're concerned, this was not only -- not unexpected, I mean, we expected to see things like this, but we almost weren't sure why we were seeing it as news. But Sanjai, what would you say? He's the TAM man.

No. I mean I think the simplest way to put it is our TAM that we had on silicon carbide 3 days ago has not changed today. It's still the same. This is -- to reduce cost, you have to go better to wait for 200 millimeters that's there. You need to reduce the size of the devices, that's there. You need to -- the problem of overspecification they have had many, many, many millions of miles and over the 6 years of experience. So they're probably overspecified, overengineered all that extra is coming out. But this is just natural. I mean, it happens in all industries when ship size has come down. So...

Okay. Good. Good. I think that we'll do it for silicon carbide. To another hot topic on 3D sensing. So there were some tweets out by a well-known analyst saying, it was a little confusing the tweet, because he was basically saying Lumentum was out, but they were a little unclear on II-VI or Coherent. I guess, if you want to just squash that rumor, that would probably be great. But maybe how do you think about this market? How do you think about your opportunity there? You've had some great share gains over the past couple of years. Do you hold on to those? And how do you feel about a third entrant potentially in the market here?

I can start, and then Sanjai can follow up with me. I mean, we have new opportunities in the sensing market being qualified all the time. So actually, I think from the very beginning in 2017, when we first started talking about 3D sensing and sensing overall, we always expected there to be more than 2 players in this market. If you think about it, it's an interesting parallel to silicon carbide. The market can't really develop if there's too few players. There's just not enough supply and enough ability to diversify your designs among different types of players. But Sanjai, what would you add?

Yes. I mean, I think you said it, Mary Jane, this isn't a 1 product, 1 cycle, right? There's -- there are overlapping products, there are new applications that are coming in. And I think the -- and we are in all of those applications and a large fraction of us. So we started with Face ID and face ID well facing, then Face ID well phrasing, something else and something else, and something else. So we do -- so there would be new players. I mean it's just natural as a market matures. But the fact is the market, the overall market, our TAM for the consumer electronics market is growing at a 20% CAGR over the next 5 years. So it is -- as long as it's growing, there is great opportunities for all of us. So it was pretty good.

You guys have done extremely well here, taking a lot of share over the past few years. It looks like versus your competitor, at least it's permanent, but did you have any thoughts there? And then as we look forward in terms of road map, do you think rearchitectures benefit you guys?

I can just tell you, maybe it's the breadth of the platforms that we have, gallium arsenide and indium phosphide, we have gallium antimonide, which is mid-infrared 2-nanometer, 2-micron laser. And for sensing, there are different things that are sensitive, different wavelengths are sensitive to different materials and different species. So if you're trying to sense glucose, you need wavelengths at those ranges. So it is -- and the transition from gallium arsenide to indium phosphide. I mean, those things happen. Indium phosphide is a better wavelength, because you can put more power out of it without burning people's eyes, like 30x more power. So it's fun to have such a broad portfolio as we can expand on new applications. I mean that's what I would do.

Excellent. Let's see. Supply. So you guys talked about a $67 million shortfall from constraints, I think, that was higher than September, but better than June, but can you talk about supply? One of the key themes from our conference is that supply is absolutely loosening up. And so are you guys seeing something similar? And when does this normalize? And does that then benefit you guys by that amount in a quarter?

I'd say that we've said since the 6/30 reporting at the end of August that we were beginning to see the supply chain easing. And at the time, I described it as all the lanes on the turnpike are blocked and they just opened the shoulder. So because people immediately said, "Oh, great, it's over. Having said that, move fast forward to where we are today, our comments in the 12/31 reporting were that we expected that we might see similar constraints here in the 3/31 quarter. But that by June, we thought they would be considerably down. And then for fiscal year '24 commencing in July, we thought that it would largely be entirely behind us.

Okay. Great. My next question was on backlog and then I think there's a follow-up to the question earlier. So on optical, Coherent has highlighted $2.9 billion backlog, down $150 million Q-on-Q, and the forward outlook indicated continued consumption of backlog to the mid-$2 billion exiting the year, so down $150 million per quarter. Would this imply that your overshipping current demand by that $150 million rather, if not, why is backlog coming down?

Well, a couple of things. The downside is speaking to a lot of excellent financial people is that they start doing that. We were asked the question of if the backlog comes down, should people be worried. And our answer to that was what we're seeing from customers. And we said this since the September quarter, we are seeing a return to more normal ordering. So during the supply chain shortage times, we had much longer-dated orders from customers especially in the networking group where they might have been as long as 12 months, which is very, very unusual. In order to try and raise the priority with integrated circuit suppliers by demonstrating what the demand was. As integrated circuits become more available, it is not necessary to give such longer-dated orders and they'll probably come down closer to maybe 6 months, but the typical networking time line is actually 13 weeks. And so when we said that if the resting backlog came to be in the middle 2s, like $2.4 billion, $2.5 billion, people should not the least, but be worried and that, that might be more like the resting backlog. Under no circumstances, did we say it was coming down $150 million a quarter, that probably is math if people wanted to do it. It was more trying to signal that as we return to more normal levels of ordering, the backlog will naturally come down. And that has nothing to do with shipping ahead. It just has to do with which orders are rolling off and what are they being replaced with, it's as simple as that.

Okay. Yes. That totally makes sense. This is an out left field question, but do you guys happen to do a material called gallium oxide? And is it a replacement potentially for silicon carbide? Or are you guys unaware of that?

We are there. I don't know whether Mary Jane can comment on what we are actually -- the programs in our...

Well, possibly not. I don't know that we have a lot out there. But as you can imagine, there's probably a number of materials that we work in that some of which come to fruition, some don't.

Yes. I mean you guys -- that is your heritage really or some of these being on the leading edge of new materials. So I thought I would ask. Great. And then maybe finishing off here, what -- I don't know if you want to give us some sort of a preview of what to expect for OFC if not from you guys, from the industry? And what -- Sanjai, for example, what are the things that are most interesting for you that you're going to be looking at more broadly for the industry?

Yes. Maybe as a marketing guy, I'll start with what we are showing off at OFC. So we are showing off with telecom, and we announced new products for both about telecom and datacom. Datacom in strictly about 800G, 1.6 tera, 200G lasers. On the coherent side, we are showcasing a demonstration of our 100ZR, which is the transceiver for access networks. When access networks go from 10G direct detect today to Coherent, we believe that they're going to use this, this is going to be the platform of choice. And that was built with our internally designed purpose-built DSP. So that's -- so we are showing that. We've got some systems demos. We also have some equipment demos. We have a small equipment business, telecom instrumentation business. So we're showing some of those. And those are -- and then the next-generation VCSEL. So we've got a pretty -- all focusing on high speed, whether it's Coherent or datacom. In terms of what I guess I'll be looking at is, that is an area we -- the market is also changing a little bit. The OEMs are starting to offer components. And that is a new dynamic. They're integrating upstream, so to speak, which is interesting, because we would -- we are, again, companies such as who are in our layer are going up, got going downstream. So you've got all that playing out, but -- so I'll be looking for all of that. But this is our first OFC after the Finisar acquisition. So it's a big show for us. So we've got lots of speakers, lots of panelists and Johnny will be the executive forum on Monday. So it's...

And next component you mentioned was that like the onboard optics and package...

Yes. So the onboard optics, we announced some really tiny components, passive components to make tiny amplifiers that go into these coherent transceivers. So with silicon photonics, what ends up happening is you just don't have enough power for these transumers. So some of our customers are building tiny amplifiers that actually go inside the transceiver and we make all of those components. We introduced -- we just announced a high-power pump, dual-chip pump. So we've got a whole range of different products that we are excited about, both demos as well as product announcements.

I'm excited to see them all.

Yes, are looking forward to that.

Thank you so much, Sanjai. Thank you so much, Mary Jane. Really appreciate it.

You too.

And I hope to see you guys in San Diego in a few days.

Okay. Great. Thank you.

Bye-bye. Thank you, Sanjai.

Yes. Thanks. Bye.

Bye, guys. Cheers.