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

Okay. So morning, everyone. Thanks for coming to the conference. My name is Rod Hall. I'm the Hardware and Infrastructure Technology Analyst, Consumer Systems 2, we call it, and also Optical. So I'm here with Alan Lowe, who's the CEO of Lumentum, a long-standing industry participant. Kind of one of the old guard in optical side, [ and let me ] use the word old, but -- you know what I mean.

I'm feeling old now, Rod.

So welcome, Alan. Thanks for coming.

Thanks for having us.

So I wanted to kick off on 3D. And as we were talking before, this is a question that gets asked all the time. And so I thought maybe I'd get you to kick us off with some thoughts on 3D sensing, particularly some of your recent comments on market share? And then maybe I'll follow up on that.

Sure. So I mean, I think it's been more than recent. I mean we've been talking about share and normalization for years, and I think it's finally coming to fruition as the chips that are at our lead customer this year are the same as they were last year. And so our competitor has caught up, and that's why we guided to have 3D sensing down this fiscal year, 40% to 50%. So part of that is share shift and part of that, as you would expect, the average selling price reduction that our customer expects year-over-year. And so I think not a surprise to us. I think what's more exciting is what we've got going on in new applications in auto and XR, and all those things that are actually starting to ship today and should lead us to a more diverse customer and an application base in the coming years.

So we know there are a lot of different parts in 3D sensing. There's the flood illuminator, the dot matrix projector and then there's the world-facing side of things. Could you maybe talk about your competitive position on each of those things? We know the flood illuminator has been -- the share has been shared there for quite a while. Now, I guess we're talking about the dot projector. But what about the world-facing side? Do you still feel like you have a lead there, or is the competitor sort of catching up there as well?

Yes. I mean, I think there's a viable competitor for all of the VCSELs in our lead customers' products today, whether those are the same chips next year or not, or if one of the new chips that we're working on gets incorporated into the product, I think that would then change the competitive landscape and could potentially lead to a differentiation in share. Right now, we're trying to keep everybody down on -- with respect to growing our share, and we expect that we will keep that 50%-ish kind of share going forward, unless there's a new technology or a new chip or something that changes that game, or a new competitor coming in, for that matter.

All right. And a lot of times before the phones shipped, there's one view of share, as you know. And then we'd get the teardowns, and we have a whole different view. Do you think there's still -- is that still a situation that we should kind of keep an open mind about? Or do you think it's more certain this year, kind of how things are going to shake out?

I think you should keep tearing down phones so we can sell more. But I think, given that the phone was announced, and the phones are shipping to the stores and things like that, we're pretty confident that what we said on the call a few weeks ago is reality. And who knows? We've thought this before, but this one feels like they actually have the ability to produce what our lead customer needs and the quantities they need.

Okay. All right. Enough on that. I'm going to ask you the $1 million question of the conference, which I'm not sure anybody knows the answer to. But what do you think about demand visibility, consumer demand? We're going to be asking everybody this. As you look out the next 12 months or so, 6 or 12 months, what's the Alan Lowe view of the world, Lumentum view of the world?

You mean the economy?

More consumer -- consumer demand for, at least, the electronics, smartphones and so on that you're exposed to. I think it'd be better to stick with that. The economy, who knows?

Yes. I mean, I think maybe I can reflect on the supply situation that we're seeing, and some of our suppliers saying that they're starting to see free up of capacity as a result of consumer slowdown. That said, I don't know about what share shifts and all that stuff between our lead customer and other smartphone devices. But I'd say that overall, we're seeing more chips available, which led us to grow in our June quarter and give guidance to grow dramatically in our Telecom side of the business in the September quarter. So I think, who knows what happens? I don't see the consumer sentiment changing much. And I think, I'm not an economist, so hard for me to really comment on is it a gentle slowdown or a recession, who knows? I think on that point, though, the demand for bandwidth is going to continue to grow dramatically. And so from that perspective, I don't see the Telecom, datacom side of the business being impacted by what happens in the economy. Now, our laser business would be, because it really ties to industrial production, and the consumer side of our 3D sensing would also. But I think all of the stuff we've got going on in new applications, that's not going to be impacted at all.

Right. Okay. 3D sensing, you touched on other places the technology will be used, other applications. Could you maybe dig into where you think the biggest opportunities are there? I know we've talked about automotive sensing in the past, but what's happening? Maybe give us an update on some of those other opportunities.

Sure. I mean, we're working with many, many LiDAR companies as well as we're talking to the OEMs themselves. And in the June quarter, we shipped a $2 million dollars of VCSELs to one customer for LiDAR, and we expect that to grow through the year. So that's just one, but we are working with many, many others and expect that to really become more meaningful. And fiscal '24 and '25, which, again, will give us a diversification of effort. So we have VCSELs for solid-state LiDAR, and we also have coherent lasers for FMCW that we're working on. So we have all the different light sources needed for the various different types of LiDAR. So when a customer comes to us, we don't care what they pick, and we show them what we have. But there's so many [ LiDAR ] companies here, we have to bet on the -- try to bet on the winners, and I think we've done a good job so far.

When you talk about the $2 million LiDAR customer, can you give us any idea of the application there? Is that an automotive application?

Yes, automotive application. Chinese customer, we had an announcement about it a few quarters ago that we were working with Hesai. Hesai is the lead LiDAR customer of ours, and I think they're winning a lot of OEMs. And so from that perspective, they expect high growth, and we're going to grow with them on it.

I know one of the things that's so exciting about automotive and LiDAR is you get multiple sensing chips around the vehicle. Can you talk a little bit about how many they're utilizing? I don't know if I've seen anything on the technical design there.

Yes. I mean, I think they have various different applications, from full robo taxi to Level 3 type applications, and there'll be different numbers of LiDAR based on what the use case is. But I'd say that we have many chips in each LiDAR module because it is solid state. So they're smaller chips, but because there's not a beam steerer, we use the VCSELs to -- they can individually light them up, so there's many, many chips in any given LiDAR module.

And do you -- when would you predict that a major Western brand, either European or U.S. brand, starts to implement use these chips in their devices, their cars?

Well, they're certainly trying them now, right? It's -- they're developing products in cars using our VCSELs, whether that comes from Hesai or elsewhere. I'd say it's -- for the Western, it's probably a couple of years out at least. But I think the China adoption is going to be much more rapid and much more broad-based.

Okay. Let's move on to datacom, and talk about that a little bit. EMLs have been doing really well for you from a demand point of view. As you said, a lot of the networking applications have been doing really, really well. Can you talk a little bit about capacity expansion there in the near term, and then what your plans for '23 are?

Sure. I think we grew some 60% year-on-year in the June quarter in EMLs, and had record EML shipments. And continue to see strong growth, especially as we get into the next generation of leading-edge chip where the hyperscalers will be very motivated and compelled to make that shift to the next generation of 200 gig per lane laser. So we added capacity. We talked about it over the last 1.5 years. That came online, that's why we saw the big uptick in capacity and output in the June quarter. We also pulled the trigger on additional capacity several quarters ago that are coming online early next year in anticipation of strong growth in that next generation of products. So that really hits the mainstream [ ramp ] middle of next calendar year. So we're expecting to continue to grow that EML business for the long run.

When do you expect supply to meet demand there?

I mean, I think we're getting close just because of our added capacity. I'd say we still have DML capacity that is available, and so some of the things we're doing on high-speed DMLs which would be driving down the cost of modules should increase our utilization of DMLs that will free up some EML capacity. So I'd say that long-term, we'll have ups and downs, but we're still constrained in the given quarter, and we'll see what happens in the next few quarters.

Maybe for people that aren't as familiar with this, what layer of the data center are you seeing the 200 per lane applications, mostly aggregation layer stuff? Or are you...

I think it's going to be broad-based just because the economics are so compelling when you double the speed of a transceiver and use the same amount of fiber and -- basically similar to electronics. So the incremental cost of the data center is small compared to the doubling of speed. So wherever they need more capacity, I think it's going to continue to be used across the board.

I mean, Google published a white paper not too long ago on using optical switching in the aggregation layer. I don't know if you guys have seen that side of things yet or if it's sort of too far out to talk about, but we thought it was interesting because it could replace some switching with optical components. I don't know if you have any particular thoughts on that or...

No, I think it will. And I think that's why -- the -- if you look at the optical percentage of spend in a data center compared to where it was, it's going to continue to grow. And if you look at optics in routers, optics in switches, all of those optics are growing as a percentage of total overall spend. So I think -- I haven't seen the Google article, but I can imagine what it says, and that's why we're so optimistic about the future.

Let's move to ROADMs, we'll talk about that a little bit. So you've been doing better at supplying, even increasing supply. But then we talk to your customers and they say, we still need more?

That's a good thing, right?

Yes. That's good for you. So can you talk a little bit about what's going on with supply there? What are the dynamics? What are you seeing in terms of meeting demand that's out there?

Yes. I mean, we're still gating our ability to supply what our customers [ need ], especially in ROADMs because of the large semiconductor componentry of a ROADM or ROADM line card. I'd say that compared to 6 or 9 months ago, it's a whole lot [indiscernible] suppliers, but we still have some [indiscernible] suppliers that we need to continue to work with. And they're trying to help us out, and I think we're starting to see some freeing up of capacity, as we talked about earlier, because of some consumer -- potential consumer slowdown. And so from that perspective, we were able to grow in the June quarter, we're going to grow ROADMs in this quarter. And then I think we're going to continue to grow through the year because demand is not our problem in ROADMs. It's our ability to get the chips, and our team -- our supply chain team is doing a great job in getting what -- not what we need, but getting more of what we need to supply to our customers.

Maybe a little bit bigger picture there, especially on complex ROADMs, NxN type ROADMs. You guys have been the market leader for a long time. Do you see any competitors coming up to challenge that at any point in the near future? I mean, how -- just talk to us a little bit about that competitive environment for ROADMs?

Yes. I mean, we took 4-plus years to develop that NxN ROADM, and I think our competitors didn't think it was possible. So we're not seeing anyone today shipping production product. I think there's discussion about competitors developing it, but I'm super paranoid all the time about what our competitors are doing. So what we're trying to do is the next generation of NxNs, and continuing that to distance ourselves from our competitors. But I think there's certainly a leadership today that we have that's multiple years.

In the network topologies, we've seen China, I think, for reasons to supply more than anything else, go to some lower complexity ROADMs out closer to the customer. But then Western network topology seem to be driving more and more complex ROADMs closer and closer. Is that kind of -- is that the way you see things? I mean, do you think the...?

Well, I'd say that actually, China, we had a partnership with a large OEM in China to develop the NxN, and in fact, got some funding from them. So they actually deployed the complex NxN before the West. I'd say today, we're seeing more of the Western customers deploying them and growing much more rapidly than the China deployments. And I'd say that that's where we're having the most inability to meet the demand as the demand for that very high end NxN or even very, very high port count ROADM because our customers are highly dependent upon us and because we're sole source in many of those cases.

In China, they -- when all the trade issues popped up, I know that they at least out near the access layer, they were planning to do more complex, did lower complexity ROADMs. Are you saying they've come back and replaced some of that with NxN, or the NxN up further in the aggregation layer...?

Yes. I think the NxN is more in the core network. And to your point, the lower port count ROADMs are being deployed near the edge. That said, I think last quarter, between high port count and NxN, 70% of our ROADM revenue came from those products. Now, there's a huge ASP difference between NxN and 1x9 single ROADM, so that doesn't represent the units, that's just the revenue.

Does it -- do you think the Chinese will ever come back? And if they could, get access to more supply, rip and replace some of that -- like you say, 1x9, that kind of stuff? Moving more to an NxN type of architecture?

They certainly could. I don't know that in China, they're going to rip and replace stuff. I think they'll add new greenfields to add that capacity to their network needs as demand continues to grow.

Okay. Talk a little bit about just shifting gears again to Neo, and the acquisition there. Now that it's complete, how that fits into your strategy, just to kind of explain it to investors? And how are customers reacting to that closing?

Yes. I mean, very favorably. In fact, if you go back in time, our customers were the ones that were asking us to consolidate the industry there and have a more financially stable supply of critically important components and products. So the Neo acquisition really broadened our product portfolio. It gave us a 400-gig ZR. It gave us the ultra-narrow line with tunable laser, which is critically important to most of the world in networks today. So that's been very well received from our Western customers as well as our Asia-based customers. So across the board, very, very favorable. And the team is on board and being integrated well, and very happy to be part of Lumentum.

Do you think there's any -- like, the deal closure point, is it a breakpoint in time at all for people to place orders? So do you think order activity just continuing normally, or maybe some people holding up waiting for the deal to close? Any thoughts on that, or...

Yes, I don't think so. Well, I mean, certainly from a hyperscaler maybe for a ZR module, but for the tunable lasers and other coherent components, I think customer order flow is normal throughout that process. And they're so critically important to everyone else's ZR product or everyone else's 800-gig products that that business is continuing to do quite well.

I mean, while we're on this subject, I guess, broader -- you've been a leading consolidator in optical, it's needed consolidating for a long time. Broader thoughts on industry consolidation, further consolidation? How do you see the industry now structurally?

Well, I mean, it's who do you have in mind? I mean, I think there's not a lot left between us and [indiscernible] in the optical telecom side of the business. Our acquisition of IPG's telecom business is something that was opportunistic, came to us and we said, "wow, this is great." I don't know if there's a lot of those out there, but I think we're going to continue to be acquisitive, and we're going to continue to drive the value to our shareholders as a result of capturing synergies, like what we did on Oclaro, and we're starting to do now on NeoPhotonics. So we're going to continue in that game. And I think we're also going to get -- work to get into new markets. Our lasers business, for instance, we're getting into solar cell processing and EV battery processing, which will really grow our available market. And we think we can do the similar kinds of things with that technology base that we have across the company.

Okay. On Neo -- just to continue the Neo a little bit, the synergies. You've talked about $50 million, $30 million in COGS, $20 million in OpEx. Could you talk a little bit about actions that you need to take to realize those synergies splitting maybe between the 2 things?

Sure. I'd say that the day 1 public company costs are gone, and that's part of that $20 million of OpEx. We still have a little bit of work to do as we consolidate facilities, and that's happening. So I have all the confidence in the world that the team will come through with at least $20 million of OpEx synergies. On the COGS synergies, they take a little bit longer in that -- these are -- well, the initial ability to drive cost reduction because one of us are paying more for the [indiscernible], and we're in that process of talking to our suppliers about charging the lowest price that we get between the 2 companies. And so that will give us some initial savings. But then consolidating factories and fabs, they have a fab in Japan that's 2 train stops from our fab in Japan. We're going to put those together. But those kinds of moves take 18 to 24 months or more, and so those COGS synergies will take a couple of years. We'll get some now, but I think over the next couple of years, we'll get a whole lot more.

Do the supply chain difficulties affect your ability to execute there? Or how [indiscernible] that?

Yes and no. I mean, now we're a bigger purchaser of a lot of the supply chain stuff. So I think from that perspective, I have lots of conversations with executives of semiconductor companies, and they see the opportunity for us to consume more of their product, and so they try to help out. Given the tough supply situation, it's difficult to negotiate price reductions with semiconductor companies today, much like it's hard for our customers to negotiate prices down as we are in a constrained environment. So I think the team is on track. We review our synergy attainment and plans, and everything seems to be on track to exceed that $50 million.

Do you think -- and it's -- I don't -- I'm not particularly worried about this, but I'll ask the question anyway. Do you think that -- we've seen on the consumer electronics side of things, people are ordering [ or give me all ] I can take to get too much inventory now, and now they've realized, oh, demand is dropping, and I've got this inventory snaking toward me. That's a problem. And then they started to, as you know, get rid of inventory. Do you ever see a situation in any of your networking business, ROADMs, that kind of thing where people have been over ordering, ordering as much as they can get, but are maybe worried about the economy next year and they slow down, or anything like that?

I'm not -- I don't think so. And that the inventories that our customers is growing, and I think it's a result of, as they call, it the golden screw of getting that last component to be able to ship a system. They don't want to tell us not to ship a high-end ROADM because we know -- they know that we'll ship it somewhere else. And so I think that as the semiconductors become more available, their inventories will deplenish but their backlogs across the board are very, very strong. So that encourages me to continue to invest, and we're going to spend more in CapEx this year than we spent in the prior years, and that's a result of my confidence in the long-term growth and our need for fab expansion as our Telecom and Datacom business should continue to grow very strongly.

Based on what you see now, when [indiscernible] there's kind of a consensus on this mid next year. But then do you think it's getting better, or...?

Well, no, it's certainly getting better from 2 aspects. One is [indiscernible] problems. And second is the broker market, we're not having to go out so much to get the components. And when we do, it's not 100x the dollar part, it's a much more reasonable amount. So I think from that perspective, it's getting better. I think we'll still have challenges through -- to your point, through this fiscal year and into June. I don't think we'll be able to satisfy our customer demand through that period of time.

If we talked 2 months ago to now, if you look at the 2 time frames, has your view changed of the supply situation? Are you thinking it's different now than it would have been 2 months ago, if we...?

No. I mean, I think if you look at what we said in our May earnings call, we were confident that we would grow in June and grow again in September, and we are seeing the freeing up of capacity and [indiscernible] problems. So I don't know that the outlook has changed much other than demand continues to be very strong. And so it's going to take us a little bit longer to satisfy all that demand.

Great. Okay. Gross margins, let's talk about that a little bit. Just -- we've got our views on margin mix and so on. But I'm wondering if you could kind of give us a better understanding, at least from your point of view, of how the 3D sensing declines, and then the Neo acquisition impact gross margins in aggregate?

Yes. I mean, we gave a fiscal '23 outlook that had gross margins -- well, had operating margin being down from our model, and that's for 2 reasons. One of which, if you look at our organic pre-Neo acquisition, our gross margin forecast for fiscal '23 was 50%, even with the lower 3D sensing as we driven cost reductions and the price reductions are not as they've been in the past. So organic business, 50%, that's at our model. NeoPhotonics comes in at the low 30s. And so until we get those synergies, that's going to be a drag on gross margins. And I think that as we look at fiscal '24, we should be able, throughout some point in time in fiscal '24, we should be able to get those Neo margins back up through the synergies and drive more leverage of the top line as we grow quite rapidly to get to north of 50 points of margin, even with the product mix that's less favorable in 3D sensing.

On the leverage side of the equation, particularly in the OpEx lines, I wanted to ask you about R&D. R&D continue to move up. But rightly so, because you've got so many different opportunities that you're addressing. But do you ever see R&D stabilizing at either a percent of sales or an absolute level? Or how do you think about R&D? How should investors think about it in the next 2 or 3 years?

So I'd say that since the top line didn't grow in fiscal '22 because of inability to get supply, we grew R&D as a percentage of revenue. And with the acquisition of the IPG product in the DSP, that's going to have another tick up in R&D spend in the magnitude of about 1 point. Now, I think that's probably the right level and it's going to continue there, and we'll get that benefit once that DSP is finished and in our module from a gross margin standpoint. So we'll see a pickup in gross margin, increase in R&D and operating margin should still be at model, at 30 points.

So you're saying once you add the 1 point, when you say at that level, absolute R&D at that level, or R&D as a percentage of sales?

Percentage of sales. We're going to continue to grow. We have a lot of great opportunities to expand into new markets, and so we're spending more on R&D as investigations into new markets to grow that available market.

The way we think about OpEx leverage then is R&D as a percentage of sales [ says ] the same, and the leverage comes through SG&A? Is that the way...?

Yes, SG&A and our variable cost in a gross margin standpoint. So as we sell more and we have the same fixed cost, our gross margins should go up more rapidly than otherwise.

Great. Okay. So we're winding down to 10 minutes. I want to make sure the audience has a chance to ask questions if they have them. I've got more questions to go, but anybody in the audience have any questions for Alan? You don't have them right now, you can think about them and ask him in a second. Yes, maybe wait for the mic.

The question is on 3D sensing. A year ago, you gave us a framework or maybe it was earlier this year, about plus or minus 5% over the next 3 years. Obviously, year 1 is going to be a step down. Is that still a reasonable framework to think about over a 3-year period, or is that something that you have to revisit?

Yes. I think 2 things. One of which is we actually grew our 3D sensing more in fiscal '22 than we expected. So the step down in '23 is more dramatic as a result of that. And then I think as we talked about earlier, how fast the adoption of extended reality and automotive, as well as I'm still hopeful that an Android ecosystem will incorporate 3D sensing in some shape or form. So I think it's going to be a U-shape where as these new applications come on board, one could see probably on the lower end of the plus or minus 5%, 5% compounded reduction per year, I think, is reasonable.

Got it. And the other question I had was you talked about your share for some time, and anticipated that, that high share wasn't sustainable. This year, obviously, we'll see it come down. Do you think it sort of stabilizes at these levels, or that step-function reduction remains a 2-player sort of market?

I think it just depends on -- as they introduce new chips and whether or not they're more complex, and I think we've got a lot of stuff on the drawing board that is more complex. And then we'll see if -- who gets to market fast first. And typically, they've launched with one supplier. I don't think they want to do that. But I think because of the importance to get one out, they've typically gone with us on new chips. And so we'll see. I think having a customer between 10% and 15% like that is a little bit more comfortable, frankly, than having a 28% or whatever it was last fiscal year. That's a big chunk for one customer. So as we grow our other parts of our business, I think that's a more diverse customer base and a more diverse set of applications that I think makes us a healthier company.

All right. Other questions in the audience? All right. While we're on 3D again. XR. Can you talked a little bit about complexity of the products and the chips going into them? I mean we have some ideas of what's going in there. It seems like a lot more, but maybe just give us a little bit more color on that since that seems like something that's going to start developing this fall?

Yes. I think -- and again, they're very, very secretive of -- all of those customers are very, very secretive of what they're doing. But one can imagine a use case where there's eye tracking, gesture recognition and other kinds of -- room surveillance and things like that that you would imagine the content per glasses or headset is significantly more than a phone. Until they introduce the products, it's hard to really tell. And some of their forecasts are unbelievable. Not sure I totally believe the adoption rates of these things, although watching my son move around with those things in his hands is pretty fun to watch. So we'll see. We all had them last Christmas on the team, but yes, we got tired of them pretty quick after we injured ourselves a couple of times.

So the things that I -- chip wise, though, eye tracking, yes. Room scanning, facial expression scanning, as well as what we see out there being developed. So it feels like 3 or 4 parts per device, or how?

Yes, it could be more. More than that. It could be more depending upon what level of use they use the chips for.

Okay. Back to more mundane things, backlog. Can you talk a little bit about your backlog at this point? And how the dating of the backlog looks?

Yes. I mean, I guess from our perspective, backlog is less important than what our customers are telling us. And why I say that is that still, we have several large customers that are on vendor-managed inventory. So they give us a forecast and we stick it into their hub, and they have a certain amount of time that they have to pull it. Those hubs are dry, so as we put them in, they come out very rapidly. That's usually a good indication of their forecast is good. But as I said earlier, I think no customer is in the mindset of saying, "Hey, I don't want your ROADM" because they know we'll ship them somewhere else, and it may be a while before they get them back. So I think from that perspective, on telecom in particular, coherent components, modules and ROADMs and amplifiers and pumps, all of those things are growing rapidly. And organic growth this year, putting aside Neo on the telecom and datacom, more than 25%, and then you layer on NeoPhotonics, more than 50%. So the outlook is very solid, and we're very excited about our broad product offerings and our customers' acceptance of those products.

Do you -- we've seen this on the consumer side, which is why I ask, some of the -- so cost inflation is starting to roll over on some of these parts as the demand weakens, and you wonder -- and particularly on finished goods, so it's a little bit of a mixed metaphor here. But what we worry about, and we've already actually seen to start happening is the finished goods because the cost is coming down or coming out at one end of the pipe cheaper than the stuff that's in inventory, or in your case, in order backlog. So do you worry that maybe the inflation of some of these parts, I don't know, maybe it's 3 months from now, maybe it's 6 months from now when the supply chain gets better, you still got backlog of orders, you worry people come back and say, "Hey, I want to cancel that higher-priced order and put in a lower-priced order"?

Yes. We don't have -- I mean, if you look at our pricing to our customers, the price increases or reductions -- I mean, overall, it's pretty flat. And so there's not a lot of, hey, that order from 6 months ago is at higher price or lower price than the order I gave you here, not a lot of that. And I don't know what you mean by -- what did you say, the inflation reduction?

Yes. Yes. I mean the logistics, all these costs are going in. No, this is in finished goods metaphor, which I want to say it's mixed, but you see all that cost going in. Some of it is coming back out. Chip prices in isolation also coming down in some cases, so...

I haven't seen that.

Okay. Well, a lot of this is like consumer stuff. PC guys dumping [indiscernible] brokers and stuff like that.

Yes. I mean, we are seeing a reduction in what we have to pay the brokers because I think there is more availability. But we haven't seen any of our semiconductor suppliers coming and knocking on our door saying, "Hey, do you want some cheaper chip"?

Not yet.

But I will say that we've heard that their foundry partners have said, "Hey, we have this more wafers if you want to take them" and they're not asking for a premium, which is very different than it was 6 months ago.

Right. That's a good sign. Little bit more on laser visibility, given what's going on with the economy and so on. How do you feel about that? What does the visibility look like on demand?

Again, I don't know what the economy is going to do a year from now, but I'd say that the demand is very strong. And as we get into these new applications that I talked about in last quarter's revenue, it was very small in those new applications. And expect it to grow, especially in our ultrafast lasers that we've introduced a whole new set of products over the last couple of quarters, that -- when you get into a new market like that, the growth of our business should be much more than the overall market itself. And that's what we're seeing today. And then our large Amada fiber laser customer is doing quite well. I don't know, they may be taking some share from IPG, we don't have visibility to that. But if a customer was an IPG customer, now they're an Amada customer, that's market share gain for them and for us.

So just maybe finishing up, we've just got about a minute left. On capital allocation, it sounds like your M&A, not a lot of targets out there, maybe some alternative things to do, et cetera. But maybe just talk to us about how you intend to allocate capital here looking forward?

Yes. I mean, over the last couple of years, we've bought back $1 billion of our shares and over 9 million shares. We have $400 million left in our buyback, so that's -- we continue to look at doing that. I would say we're going to be very acquisitive. I think it's been -- we have that engine now that's running a little bit more smoothly and getting the synergies and being able to integrate companies. And so we're going to continue down that path as well, especially if we generate $400 million, $500 million a year of cash.

Maybe a word on CapEx the next couple 3 years, how are you thinking about that?

Up this year, for sure, as we need to grow our fabs and consolidate the Neo fabs, so -- especially transmission fab growth. And regardless, frankly, of what happens to the economy, we're going to invest that because it's an investment that we're making today for 2 and 3 years from now.

Great. Okay. We're out of time. So we'll leave it there.

Great. Thanks [indiscernible]. Appreciate it.