Good morning. No, we're not starting just yet, but this is a one-minute warning. Thank you. Good morning, everybody. For those of you I know, great to see you again. For those of you I don't, thank you. Thank you all for joining us, those in the audience physically and those online for the webcast. My name is Paul Silverstein. I'm Head of Investor Relations for Coherent. This is the 2025 Coherent Investor and Analyst Event, and we greatly appreciate all of you taking time out to join us here today. We have a great program. Before I get into the agenda, however, I've got to refer you to our forward-looking statements. All the disclosures apply to all the statements made throughout the day. Please take a look at it at your convenience.
As far as the agenda, we're going to start off with our CEO, Jim Anderson, who's going to provide an overview of our strategy and our future direction. Jim's going to be followed by Dr. Julie Eng, who's our CTO, who's going to discuss our data center and communications business. After that, we're going to have a short 10-minute break. Please come back. We're going to start promptly after that. We're going to have Dr. Chris Dorman, who's our EVP of Lasers, who's going to discuss our industrial lasers and materials business. To wrap up the day, we're going to have our CFO, Sherri Luther, who's going to talk about our financials, including our long-term model. I know none of you are interested in that. Finally, we're going to have a Q&A session.
For those of you who are here physically, we will have an informal lunch, at which you'll have another opportunity to speak with the executives. Before Jim comes up on the stage, we're going to have a short video. Before the short video, let me just mention, we do have products here. At the end of the day, after the presentations, please feel free. We've got the products. We've got some folks who could talk to you about the products. One other thing, should you need to leave the room outside of the breaks, please use the back exits to minimize the disruption to your colleagues as well as the speakers. Once again, thank you all for being here. We're now going to have a video, and then Jim will come up. Thank you.
At the heart of Coherent is a single particle, a quantum of light, the photon. Photons are powerful. We generate, modulate, amplify, direct, and detect them, turning them into tools that transmit, transform, and measure. Powered by light, we enable the technologies behind the world's most transformative innovations, from data centers to life-saving surgeries, from next-gen mobility to immersive communication. In AI-powered data centers, Coherent leads with the broadest and deepest portfolio of solutions, from materials, components, and transceivers to fully integrated systems that form the backbone of modern networking. Our technologies power optical networks at every level, from data centers to long-haul, submarine, and satellite infrastructure. Coherent delivers the precision tools that drive modern manufacturing across semiconductors, EVs, display fabrication, research, and biophotonics. We don't just build devices. We produce the foundational materials, components, lasers, and systems to keep our customers competitive. Supply chain resilience is our hallmark.
Our global teams continue to innovate at the physical level, pushing what's possible with light every single day. Photons are everywhere, and they are the foundation of Coherent.
All right. Thank you, everybody. Thanks for being here today. We've been really looking forward to this event, so thanks for spending time with us today. I joined Coherent a little bit less than a year ago, so let's see. Next month will be my one-year anniversary. I can tell you, after spending a year with this team, this is an incredible team, I have even more confidence and excitement about where this company is headed and the opportunity in front of us than I did when I joined a year ago. I am really happy to share with you today some of our long-term plans. Thanks again. Where I thought I'd start is with the question that I've gotten most asked over the last year.
Actually, a lot of you here in the room have asked me this question, which is, "Hey, why did you join Coherent, and what's the opportunity that you see for the company?" That's pretty simple, actually. Two reasons, right? Number one is when we look at photonics, the photon, that is becoming more and more critical to so many different applications, so many different industries, and you're definitely going to hear about that today. The second reason is there's really no other company in the world that's better positioned in photonics than Coherent, better positioned in terms of the opportunity with photonics, but also better positioned in terms of helping our customers innovate in photonics. Before I joined Coherent, I spent almost my entire career in the semiconductor industry, which is really in the world of the electrons, right?
We made some pretty amazing advancements in the world of electrons. Even from the vantage point of the world of electrons, you could see that photons and photonics are becoming more and more important in lots of critical applications like data center, industrial, et cetera. Photons have some pretty incredible properties relative to electrons, some kind of magical properties. What I thought I'd do is start with a couple of examples of that first question of why is photonics becoming so much more important? Let's start with that. Let's start with data centers. This data centers, this is near and dear to my heart because I actually started my career in data centers. I started as a microprocessor architect for servers that go into data centers. I spent most of my career around data centers.
In AI data centers, what you see is definitely optical networking and photonics technology becoming more and more important, becoming important to how we architect data centers, how we design data centers. That is certainly becoming more important. If we look at why is that, so why are photonics becoming more important? It really comes down to just basic physics. As you increase the data rate of the data transmission within a data center, or if you increase the distance, what you do is you basically hit the limits of electrical connections. Physics forces you to transmit that data optically with photons instead. Obviously, data rates within the data center just continue to go up. This has driven already a significant conversion of the data center from electrical to optical.
When we first started building data centers, the communication links between the computing nodes were 100% electrical. Now, if you look at a data center today, the entire scale-out part of the network is optical. We have already converted the scale-out portion. That scale-out network continues to grow and expand at actually a faster rate than the underlying computing nodes. Now, through the rest of this decade, what we are going to start to see is the scale-up part of the network start to convert from electrical to optical. The scale-up is the connections within the rack. This is going to further expand the optical networking TAM within the data center. When you look at Coherent and the technology that we have, we are incredibly well positioned for growth in this segment.
That is why we expect our market opportunity in data center to roughly double over the course of the rest of this decade. That is one example. Another example, though, let's look at industrial. This is our other big market. I love our industrial business. If you look at photonics, photonics is so important to so many different industrial, scientific, medical applications and becoming increasingly important. Let's take just one—there are lots of different applications for photonics, but let's take just one example. Let's take advanced manufacturing. This is manufacturing a really complicated device that requires high levels of accuracy, precision, and you are trying to manufacture more and more per year. As you drive higher levels of accuracy or precision in your manufacturing, invariably you are moving away from mechanical manufacturing systems to photonic, to laser-based systems.
We see this all the time throughout industrial, where more and more of the industrial process steps are requiring lasers and photonics. A great example of this is the device that all of us carry every single day for most of the day, and that is your smartphone. If you look at your smartphone, this thing is a marvel of modern manufacturing. If you look at the process steps to build that smartphone, there are over 40 of those process steps that require a laser to complete that process step. It is not just one laser. It is five different types of lasers, and Coherent being the only company that offers all five of those lasers. Here again, great opportunity for us in terms of the continued growth and expansion of photonics and laser content in all sorts of industrial applications.
If we go back to Coherent now, we've got more and more photonics being used in lots of different applications. What you'll see today as we walk through the products, the technology is there's really no other company on the planet that's better positioned to benefit from that opportunity than Coherent. Not just benefit, but also help our customers innovate. As photonics is becoming more important to their applications, we're really well positioned to help them innovate. That's why we think our fundamental mission is around harnessing the power of photonics to help our customers innovate. As I've spent about a year with the company now, there's a number of other things that I think give us a great foundation to build on from here. Number one is I love the innovative culture of this company.
This company is amazingly innovative, the most innovative company I've ever worked for. We innovate not just at the system level and the software level. We definitely innovate there, and you'll hear about that. We innovate all the way down through the technology stack, at the device level, down at the material science level, the very physical science level. That deep innovation, that's helped us drive just an amazingly broad and deep technology portfolio around photonics. That innovation, that is something that our customers depend on. We have over 10,000 customers depending on that innovation that we drive in photonics. Those 10,000 customers are giving us a really diversified set of markets and revenue streams. Beyond just innovating and designing the devices, the other important aspect is, hey, can you actually deliver it? Can you actually manufacture a complex photonic system?
We not just design them, but we manufacture complex photonic systems, devices, materials for those customers. I think something that's underappreciated about this company, at least I've certainly come to appreciate this over the last year, is if you look over the last two decades, this company has grown at about 20% per year over the last two decades, a combination of organic and inorganic growth. A great track record of growth over the past years. Now, where are we headed from here? I think this is a really great foundation that we've built, but where are we headed from here? Number one, definitely focused on continuing to drive robust top-line growth, so strong double-digit growth on the top line. Sherri will definitely talk about focused on improving our financial metrics as well.
Beyond that, a few of the themes you'll hear from us today is, number one, about concentrating and focusing our investment on the areas of greatest growth and profit for us over the long term, but how that increased concentration and focus of the investment, how that's accelerating our innovation engine, how we're getting more products out to the market quicker, especially in our key growth areas. We'll certainly talk about how we're raising the bar on operational discipline, but also financial metrics. We'll talk about the higher targets we're setting for ourselves on financial metrics. Let me start with focus. I think that focus always starts with what are the markets that you're focused on. If we look at the markets we're focused on, it's really straightforward: two markets, data center and communications and industrial.
Data center and communications are our largest, fastest-growing market. We think that market opportunity there for us grows to over $40 billion by the end of this decade. The second market, industrial, a little bit smaller market, but still a good-sized market. For us, really importantly, a high-margin market. These two markets combine to give us not just a significant market opportunity, but a really complementary, diversified end market exposure. If you look at all the different innovation that we drive across photonics, especially at the foundational level, at the physical science level, that innovation, that R&D investment, that's leveraged across both of those markets, is shared across both of those markets. Okay? That's our market focus. Let's talk next about products and technology.
This is where the company, I think, has done just an outstanding job building an incredible portfolio of photonic technology, not just through organic investment, but through some really great acquisitions as well, some really key acquisitions that have turned into some of our largest and fastest-growing product lines. A great portfolio of technology, but definitely also some room for improvement on the portfolio, some room to better optimize, better focus that portfolio. That is why you heard me after I joined, I talked about we were doing a portfolio assessment. We looked at the entire portfolio of Coherent, looked at all the different products and technologies that we're investing in, and looked at which of those products and technologies are key for the next not just one to three years, but the next five to ten years.
What we found is really two things when we did that analysis. Number one is that we were investing too much of our investment in non-core product areas. Product lines that really were not going to move the needle for us long-term. Sometimes I referred to them as science projects. When we looked at our fiscal 2024 R&D spend, about 20% of our R&D dollars were flowing into these non-core areas. That was takeaway number one. Takeaway number two was that when we looked at our growth and profit engines, we were actually underfunding our growth and profit engines. There was the opportunity to dial up the investment to get better growth and profit over the long term. That is what we have been focused on doing over the past year.
This is still a work in progress, but shifting that investment away from our non-core areas into our areas of greatest growth and profit. In the non-core area, what that means is for any product lines or technologies that fall in that part of our portfolio, that's either shutting down or divesting those product lines. This is still a work in progress. There are still actions that we have in flight here, but certainly focused on moving that non-core investment towards our growth and profit engines. Also, facilities and assets, looking at are there underutilized assets that we can divest and drive better asset and utilization efficiency across the company. All of that being refocused on the areas of greatest growth and profit. What I want to talk to you next about is, okay, where are those big growth areas?
Where are we investing moving forward? Let's start with data center and communications. Julie will come up in a few minutes, and she'll give you the whole detailed overview of data center and communications. I want to flag the top three places that we're investing, but also where we see the greatest areas of growth over the long term. First of all, and this is probably our biggest opportunity, this is optical data transmission in AI data centers. This is growing very fast. The industry expects, if you look at total data center bandwidth, the industry expects the bandwidth in the data centers to increase by about 10x over the coming years. I mean, that's a huge amount of increase. That's all going to be driven or primarily driven by optical communication. A great opportunity for us.
Whether it is a pluggable or a CPO transceiver, we are investing to accelerate our roadmap, make sure that we have got the right products and technology for our customers. Julie will talk more about that. Beyond that part of our growth engine, we are also expanding our addressable market, bringing on new products and new revenue streams. A great example of this is optical switching. This will add about $2 billion of incremental SAM for us over the coming years. Optical switching we see as a key growth area with our customers. If you look at our technology, it is very, very differentiated versus what else is out there, based on a non-mechanical digital liquid crystal technology that is very differentiated, great customer traction on this across multiple customers.
I'm actually really happy to announce today that just as of the last week or so, we achieved first revenue on this product line. This product line is now generating revenue for the company. That's actually ahead of schedule, ahead of what I thought the team was going to achieve. They've done an outstanding job of driving this into production. I'm really pleased that this is starting to generate revenue for us. The third area that I would flag is data center interconnect. We see big AI workloads that are spanning now multiple different data centers, and that's driving a lot of increase in traffic between data centers. This part of our market is expected to grow to about $4 billion over the coming years. We're really well positioned in the DCI interconnect with our ZR, ZR Plus products.
Again, Julie will touch more on this in a few minutes. Those are our key areas in data center and communications. Now, the other part of our business, we get a lot of focus on data center and communications, but our industrial business, I think, is underappreciated. I love this business, very sticky, long revenue streams and a high-margin part of our business. We are investing here as well in key growth and profit engines. The common theme you will hear from Chris will come up and talk about this segment, but the common theme you will hear from him is it is more than just about the underlying market growth. What we are seeing here is just this continuous expansion in the amount of laser photonic content that is used in many different industrial applications. The first one here where we are investing, this is a great example, semicap equipment.
We have seen over the years just a steady increase in the amount of laser content used in semicap equipment. When I started my career, we were kind of at the quarter-micron technology node in semiconductors. If you looked at the processing in quarter-micron, there were maybe two or three optical inspection steps back then. Now, you fast forward to today in the modern technology nodes, 3 nanometer, 2 nanometer, we have now about 100 different optical inspection steps. That is all driven by lasers and photonics. We think that continues moving forward. We are definitely investing in semicap equipment and additional expansion here. Another area of growth is display manufacturing. This is where the industry expects on OLED displays for the amount of OLED surface area to roughly double over the coming years as we bring OLED to new devices as an industry.
This means that the amount of manufacturing capacity has to double. There is only really one way to manufacture a high-quality OLED display, and that is to use a laser from Coherent. The third one, precision manufacturing. I have already kind of talked about this. A great example here is smartphones, but there are lots of other devices where more and more lasers and photonic equipment are getting used for advanced manufacturing. These are the three big growth areas in industrial. Chris will talk more about this in a few minutes as well. I have talked about the markets that we are focused in, some of the key growth areas that we are driving and investing in. I want to take a minute to talk about customers as well. What do customers care about? What is important to them?
Whether it's a big customer or a small customer, an industrial customer or a data center customer, every customer discussion I have revolves around two things. Number one, it starts with the technology. What's the technology you can bring? How can you help us innovate on a multi-generational basis? That's where it always starts, right? The other part of the discussion, and this has certainly become more important over the past years, is around manufacturing. Okay, you got the technology, but can you actually deliver it, and can you deliver it at scale? When you look at those two factors for Coherent, there's no other company out there that's got the combination of technology, breadth, and depth in photonics, and the ability to deliver that, to manufacture that at scale, at high quality. I want to click down on both of those.
Let's start with the technology. You will definitely hear more about this from both Julie and Chris. It's important with Coherent to remember that we don't just build the end product, right? We certainly do that, and we build a lot of software that goes on those lasers, etc. We innovate and build all of the key photonic devices and materials that go into those products. In data center, what that means is we don't just assemble and test the transceiver. That's probably the easiest part of what we do. We build all of the key photonic ingredients that go into that transceiver. That's really important to our customer. That's one of our key advantages. Our customers tell me that all the time. The same is true in industrial as well. We're building all the key photonic ingredients in our industrial lasers.
That helps us partner with our customers on a multi-generational basis to drive deep levels of innovation. The second part of the discussion is always about, okay, great, we got the technology, but can you deliver this at scale? This is the other big strong capability of Coherent, the ability to not just design complex photonic systems, but to deliver that at scale. If you look at assembly and test capacity, certainly a lot of assembly and test capacity, we've expanded that. We've basically doubled that over the past couple of years. Manufacturing deeper in that stack as well, manufacturing the key photonic devices that go into those systems. A great example of this is indium phosphide capability. Indium phosphide is one of the key technologies used in lots of different types of transceivers in the data center.
Indium phosphide capability, something that we've had in-house for over 20 years. We've tripled our indium phosphide capacity over the past 12 months. Next quarter, we'll start production on the world's first 6-inch indium phosphide line. Over half of our indium phosphide transceivers use our own internally manufactured EML. This is just one example of the manufacturing depth that we have. It's not just the capacity and the manufacturing of the devices. It's where we're doing that as well. If you look at where all of our production facilities are, first of all, we have over 60 different production facilities spread across 14 different countries, with about half those production facilities in the U.S. That gives us an incredible level of flexibility and adaptability for our customers to move things around as necessary for our customers.
For our big customers, they care about, hey, can you ramp quickly? Here again, great capability. If you take just one example, which is our 800G transceivers, we ramped our 800G transceivers at twice the rate to 400G. On 1.6T, we have plans in place to ramp 1.6T even faster than 800G. The ability to ramp quickly as well. For our customers, deep technology stack for innovation, but also the ability to deliver that at scale. What does that all mean in terms of financials? This is probably what you guys are mostly interested in, the financial piece of it. Let's start with, and Sherri will cover the financials in detail, but let's start with just how we've done over the past year. If you look at FY2025, we'll wrap up FY2025 end of June.
Where we expect to land for this year, for this fiscal year, is revenue growth over 20%, G ross margin expansion of 360 basis points, gross Operating margin expansion of 470 basis points, and EPS just about tripling. I would say, hey, good progress. I'm pleased with the team's progress this year. Good initial results. Hey, also a lot more work to do. We have a lot more work from here and a lot more opportunity. Where are we headed from here? First, let's start with revenue growth. Certainly focused on continuing to drive robust double-digit revenue growth. At a company level, targeting 10% to 15% revenue growth over the next three to four years, with 15% to 20% or maybe a little bit better in the data center and communications business, and about 5-10% in the industrial segment.
If you look at those two segments, where is that growth coming from? If we look at data center and communications, I talked about the key investment areas earlier, but the other way to look at it is from an end market perspective, we're assuming data center continues to grow at a reasonable clip. There may be some peaks and valleys along the way, but we're still very early in the AI adoption cycle. We believe AI data centers continue to grow through the rest of this decade. Beyond that, continuing to expand our SAM, create new revenue streams, bring new products to market. A great example of this is optical switching. I believe we have significant share gain opportunity.
If you take that combination of technology and innovation at the photonics with the ability to manufacture that, we have, I think, significant opportunity to gain share. In our industrial business, in terms of the end market, a little bit of near-term softness in the industrial business. Over the long term, we believe that industrial continues to grow as a segment. The bigger factor in our revenue growth is the expansion in the amount of content, in photonics content that's used in all sorts of industrial applications. Chris, in his section, he'll spend a lot of time talking about where we see that content expansion opportunity. The third piece of growth in industrial, and I think this is very underappreciated, is that in our industrial business, about a quarter of our revenue is recurring revenue.
This is revenue from services or from replacement parts for the huge installed base of industrial lasers that we have. Every single year, every time we ship a new system, we add to that installed base of lasers that require service and replacement parts over time. Chris will definitely talk about this as well. That portion of our revenue, which is today 25%, we expect that to continue to grow as a percentage of revenue. That also drives growth for us. That is where we are headed in terms of top-line growth. Let's talk about operating metrics as well. Cheri will get into the details of the operating metrics. I want to talk about just some of the key goals that we are focused on. First one, gross margin. You have definitely heard me talk a lot about gross margin since I joined.
I joined at the end of fiscal 2024. Fiscal 2024, our gross margin was about 34%. In my first earnings call, I said, "This company needs to do better than that." Our Gross margin, given the technology that we provide, should be much higher than 34%. In my first earnings call, I said, "Hey, we should be at least at 40% or better." We made some good progress over the past year. Now, having spent a full year with the team, getting a better sense of the product, the technology, the customers, very comfortable raising that target. We are raising our long-term Gross margin target to 42%. Sherri will walk you through more details around that. Beyond that, Operating expenses targeting 18% with the bias, the little bit more of that 18% bias towards R&D versus SG&A.
Operating margin of above 24%. I think if you take that combination of double-digit growth with kind of mid-20s Operating margin, I think what you'll see is that's an incredible opportunity to drive shareholder value creation over the coming years. I want to go back to what I said at the very beginning, which is number one, photonics continues to become more and more important. You're definitely going to hear that from Julie and Chris. Number two, this company is better positioned than any other company in the world to take advantage of that opportunity. With that, thank you one more time for being with us here today. I'm going to pass it off to our CTO, Julie, to talk about data center and communications. Great. Thank you. Julie?
Thank you. As Jim said, I'm Julie Eng . I'm Chief Technology Officer of the company. It's great to be here with you in New York this morning. It's nice to see so many familiar faces. For those of you who I don't already know, I've been with the company for 22 years. I ran our data comm transceiver product development for 15 years when we grew that product line from $100 million to $1 billion. I also ran our 3D sensing business unit, which included the bring-up of the Sherman Fab from its inception as an empty shell of a building to qualified product shipping in volume to customers. Before all of that, I was at the former AT&T Bell Labs, which was like the epicenter of optics and photonics in the U.S. at that time. I've been in optics and photonics my entire career.
This morning, I get the opportunity to share with you our data center and communications business. This is a growth engine for the company. I feel like we are in a very strong position. I am glad to have the opportunity to share that with you today. As Jim mentioned, all of our modern communication systems, say like after smoke signals, became electrical wires. The first transatlantic calls that were sent from Europe to the U.S., electrical wires. For that distance, as the data rate goes up and up and up, after you are sending a few phone calls and you want to send a lot of phone calls, or you want to send video, or you want to have two-directional streaming video, that same distance, higher data rate, we transition from wires to optics.
Optics becomes the only way, or the most cost-efficient, or the most energy-efficient way to send that data. The way that we do that is we take the electrical signal, we change it into the optical domain using a laser, we shine that signal down an optical fiber, which is about the size of a human hair. At the other end, we use a photodetector to change back from the optical domain into the electrical domain. That basically is a transceiver that does that transmission back and forth. That happened on the undersea; same thing happened on terrestrial. Started out electrical, changed to optical. Same thing happened in the data center. Started out all the connections were electrical; eventually, they all became optical. As Jim mentioned, the last kind of electrical connections in the data center are the scale-up.
Those will too become optical. It's just physics. There's actually no question about if it will happen; it's just when. In that way, photonics has really enabled all of our modern communications. What I see, having been in this industry for a very long time, is that in the past, photonics used to be thought of as like a conduit, kind of an interconnect. Now it's become like a fundamental part of how our customers architect the data center. As Jim showed, this is our SAM that we see for the data center and communications. For data center, we mean everything inside the data center. Communications is everything outside. That includes DCI, the telecom, transport. We see it as a $44 billion SAM by the end of this decade.
We see about $32 billion for data center and about $12 billion for communications. I'm going to start first and talk about the data center portion. I'm going to highlight three things, which are pluggable transceivers, co-packaged optics, and the optical circuit switch. Let's start first just talking about the data center and the opportunities in the data center and sort of some of the characteristics of the AI portion of the data center. You've probably heard this terminology, scale-up and scale-out, as Jim has mentioned. What really is that? What scale-up is, is actually a combining together of multiple processing units like GPUs or TPUs to act together like a single compute node.
The scale-out network is the networking of all of those nodes together to form one huge supercomputer that basically is the AI portion of the data center or the AI factory. Those two portions of the data center have very different attributes that lead them to kind of have very different product needs. Let's talk about those attributes. The scale-out network, as Jim said, is all optical today, primarily served by pluggable transceivers. One of the things about the scale-out network is that it's a very heterogeneous environment. What I mean by that, for instance, is the distances can be very different. For instance, if you're just going from one rack two racks over, that might be a 10-meter link. You'd use a certain type of connection for that, a certain type of optics.
There might be another connection that has to go up through the roof of this building across like a Costco-style size warehouse down to the other side. That could be a 10 km link. Those two things have very different solutions. In that space, in the scale-out domain, you might have network interface cards from one vendor, routers from another vendor, switches from another vendor. For all these reasons, flexibility is actually very, very important to our customers in this portion of the network. That is why you see a lot of pluggable transceivers in this portion of the network because they offer flexibility. I'll give you a little bit of a deep dive on why in a few slides. The scale-up network, in contrast, is very different. Today, it's 100% copper, always less than 10 m
The person who owns one end owns the other end. It is what we call a closed system. When you think about it, you need to fit everything inside the rack. Size is very important. You're going to replicate it a lot of times. Power consumption is very important. Cost is very important. Density, cost, and power are very, very important and arguably more important than flexibility. What you see is that co-packaged optics, in particular, I think is a great fit for the scale-up portion of the network as that portion of the network transitions from electrical to optical, which is a SAM expansion for us. Here is our market model for pluggable transceivers and co-packaged optics. This is showing our view of the SAM through 2030. First and foremost, you see everything grows. That is good news.
The pluggable transceivers grow to about $25 billion by the end of the decade. The CPO combination in scale-out and scale-up is about $5 billion by the end of the decade, with a little bit more in the scale-up domain. Here is another just comments about the data center market. Again, I've been in this field for a long time. I had this feeling that the data center, that the data rates were just coming faster and faster and faster. I went back and looked at all the historical data and actually confirmed that, that the time we had between data rates, say when we moved from 10 gig to 100 gig, was much longer than we have now when we move from 800 gig to 1.6T. That is actually great for us because faster adoption cycles favor a technology leader.
We're a technology leader. That's a great market dynamic for us. In addition, we're seeing these overlapping cycles of growth. What I mean by this is that as we see 1.6T increasing, we don't see 800G falling off a cliff. If you look at the bottom, this is our view of the SAM as a function of data rate through the end of the decade. What you see is we're showing 800 gig growing for two more years, actually. This is very favorable to us because we're a supplier with a very broad portfolio. Faster adoption rates, overlapping cycles of growth, very good for us as a technology leader and a broad portfolio supplier. Let's dig down one step deeper and start talking about the products. As I mentioned, pluggable transceivers offer flexibility. Now, why really is that?
Part of the reason is it's a standardized multi-vendor ecosystem. What is this? It offers our customers security of supply, which is actually very important. They're easily serviceable, which means if something goes wrong, you pull it out, you put another one in, and it's very easily serviceable. What I think is almost the most important is what I call it delays your architectural commitment. What does that mean? Our customers design a switch, transceivers plug in through the front faceplate. They can decide after that switch is on the data center floor, they might decide a third of those ports, they want to only go 10 m and use a VCSEL-based solution. Maybe another third, they want to go 500 m and use a silicon photonics-based solution.
Maybe the other third, they're going to go 10 km and use an indium phosphide-based solution. They can make all of those decisions at the end. They can even change those decisions later on if they want to because that's the flexibility offered by the pluggable transceivers. In addition, it retains what we call the pay-as-you-grow model, which is that not all of our customers actually require the whole switch to this bandwidth on day one. They can put in only as much bandwidth as they need until they can generate revenue with more bandwidth and then add more bandwidth. In this way, pluggable transceivers really offer a lot of flexibility. That is why you see a lot of our customers speaking very strongly about pluggable transceivers. As I mentioned, co-packaged optics does have a place also.
When you look at co-packaged optics, what really is that? When you think about it, it's really just a miniaturization of the transceiver and moving that transceiver inside the box next to the switch or GPU chip. Now, why would you do that? Because now you just give up all those great flexibility attributes that I was mentioning about pluggable transceivers. The main reason you would do that is otherwise you have to drive that electrical trace to the front faceplate. That takes equalization. Equalization is power consumption and some silicon. It is power consumption and cost. By moving that inside, you can save some power consumption and cost. Also, when you think about it, you do not have to put the entire transceiver in through the front faceplate. The majority of the transceiver is now inside the box.
You only have to bring that optical connection out. You can fit more of them on the front faceplate. That is what we mean when we say it increases faceplate density. That solution to me is actually a perfect fit for the scale-up domain. Again, which is a SAM expansion for us as an optical vendor. We, as Coherent, have the deepest and broadest technology stack in the entire industry. I think this is an undisputed fact, actually. I will try to explain undisputable, I mean, fact. I will try to explain to you why that is. In the upper left, I am showing the picture of the inside of a transceiver. In the bottom left, I am showing you an architectural implementation of co-packaged optics.
First, and then on the right-hand side, I'm listing all the attributes that you need inside transceivers or co-packaged optics. First and foremost, as Jim mentioned, for both those things, you need assembly and test. That actually does have IP in it. It actually is hard to deliver at scale with very good quality. That's something we're very, very good at. As Jim mentioned, a lot of the deep IP in a transceiver is in the ingredients that go in the transceiver. That includes lasers, detectors, passive optics, integrated circuits. Almost all of those things, actually all of those things, we have the capability to design ourselves and, in most cases, manufacture ourselves. That's an internal capability that we have. Now, why would you want that internal capability if you can just buy it from other people?
Part of the reason is it can impact our time to market. I'm going to give you some examples of cases where we have a time-to-market lead because we chose to invest in a particular product or technology ahead of the rest of the industry because we had the capability. Another reason is it can offer differentiated features. I'm going to show you some example of that. Finally, it can help you also with cost to have better cost basis. Also, security of supply because in the end, we control some amount of the capacity. The investment in that capacity is at our discretion, at our timeline. It gives us freedom to make those decisions. All that being said, we use a lot of external partners, ecosystem partners.
Now, why would we do that if we have all these great capabilities in-house? The main reason to do that, it gives us security of supply, which actually offers our customers security of supply. It also allows us to focus our R&D dollars on what we think offers the best technological or cost advantage. Then we can leverage the rest of the ecosystem for all the other components. We have a balanced approach. We have this capability to design and manufacture internally to use it whenever we see fit. The last point I kind of wanted to make on this slide was just to look at what are the ingredients in a pluggable transceiver compared to the ingredients in co-packaged optics. What you see is the ingredients are basically the same. You need a laser. You need some way to modulate it.
You need detectors. You need passive optics. You need thermal control in some cases. You need drivers and TIAs. In the co-packaged optics, in some of the implementations, you might have a separate external laser source. You might have polarization maintaining fiber and this product called a fiber attach unit. The ingredients are basically the same. This broad and deep technology stack that we have for pluggable transceivers is directly translatable to co-packaged optics. Okay, now I want to double-click a little bit on give you a little more technology on our sources, lasers, detectors, and passive optics. I'll just zoom in on those three topics. Let's start off first with lasers. First, I think we are the undisputed leader in semiconductor lasers for data com. I don't think there's any nobody can question or challenge it. I'll try to explain to you why.
As you all know, we've made our own gallium arsenide vertical cavity surface emitting lasers or VCSELs for over 20 years. We ship those in our own transceivers, but we also sell them to our competitors who then ship them to our customers in their transceivers. That's another way we deliver content to the data center. At the Optical Fiber Conference this year in March in San Francisco, we showed a 200 gig per lane VCSEL. We showed a 1.6T transceiver based on eight lanes of that 200 gig per lane VCSEL. I was very proud of our team for that demo. In addition to our strength in gallium arsenide VCSELs, we are also a strong and powerhouse provider of indium phosphide. As Jim mentioned, the majority of our pluggable transceiver revenue is actually indium phosphide-based.
The majority of our EML transceivers ship with our own EML inside of them. We demonstrated a 400 gig, what's called differential EML at Optical Fiber Conference. We were one of the few people in the world to demonstrate 400G. We were the only one who demonstrated what's called differential. The differential EML helps with power consumption. It is better power consumption and better lower crosstalk. It is a good fit for the PAM4 system. I'm super proud of our team for that demo also. That 400G element is very important because as we move from 1.6T to 3.2T, we are going to need eight lanes of 400G. We are laying the foundation here both for our own future revenue, but also for the future of the industry to move the datacom pluggable transceivers to have a strong roadmap from 1.6T, 3.2T, even 6.4T.
In addition to the gallium arsenide and the indium phosphide EMLs, we also have indium phosphide CW lasers and silicon photonics. We have indium phosphide high-power CW lasers in production. We have a robust roadmap to increase the optical power of those as needed by either pluggable transceivers or CPO. We have an in-house silicon photonics design team. There we design to external silicon fabs, as is common in the silicon industry. We demonstrated a 1.6T DR8 transceiver at the Optical Fiber Conference last year with our own indium phosphide laser and our own 200G per lane silicon photonics. When you look across this chart, there is no one else in the whole industry that has demonstrated a 200 gig VCSEL, a 400 gig EML, a CW laser, and 200 gig silicon photonics. We are the only one.
That is why I say you can see we are a technology powerhouse. As Jim said, it is not just about technology. We have also delivered all of these devices at scale. We have delivered more than 1 billion VCSELs for datacom transceivers, not including 3D sensing, just datacom transceivers, 1 billion VCSELs. We have also delivered over 250 million indium phosphide lasers for datacom transceivers. Technology, expertise, broad and deep portfolio, manufacturing scale. Now I want to talk a little bit about photodetectors, which is maybe something we have not talked about as much in the past, but I had the opportunity to share with you this morning. This is the device that changes from the optical domain back to the electrical domain. It is a very important part of a transceiver.
This is one example that we use to get a time-to-market lead above other people. Many of you who have covered us for a long time may remember the 100 gig CFP. It was a big form factor. The first time the industry delivered 100 gig Ethernet. We were actually the first person to release an IEEE compliant 100 gig transceiver. We had almost an 18-month lead over the whole rest of the industry. One of the elements that we designed, not the only element, but one of the elements that enabled us to do that was our decision to invest in a 4x25 gig photodiode. An array of 4x25 gig photodiodes, which today ship in volume by many vendors, but back then nobody in the world had ever done before.
We made a decision to do that along with a couple of other key decisions that allowed us to have that sole source position for 18 months. In addition, we're working on photodetectors for 200 gig per lane, 400 gig per lane. This is another case where we have technology and scale. We've actually shipped over a billion gallium arsenide and indium phosphide photodetectors for data com. Another of our deep expertise that I believe gives us a competitive advantage that maybe we haven't talked as much about externally, but I have an opportunity to share with you today, is passive optics. What is passive optics? Basically, passive optics is anything inside the transceiver or the CPO that doesn't generate or detect light. It moves the light around, helps guide the light, things like that.
One great example of our expertise in passive optics is the optical isolator. What's an isolator? If you have the laser here, the light goes out down the fiber. Eventually, you have a patch panel or something. You might get a back reflection into the laser. That has a bad deleterious effect on the laser's performance. What we put in place is called an optical isolator, which prevents the back reflection from getting back into the laser. We make those optical isolators. We use them ourselves in our transceivers. We also sell them to other transceiver makers. As Jim said, we do not just make the free space isolator. The free space isolator is actually based on a magneto optic material called garnet. We actually grow the garnet in New Jersey, actually, not very far from here.
We actually grow the garnet that goes in the isolators that then we sell in our own transceivers and we sell to others in the industry. We also make what are called optical multiplexers and demultiplexers. We both design and manufacture these. These combine many wavelengths together to go down one fiber. For those of you who know the fiber or the standards, that would be like a CWDM4 and LR4. Our customers use that when they want to go very far distances so you do not have to have many fibers. It is more cost-efficient to put all the wavelengths down one fiber. Making and designing those is very complex. We do that ourselves. We also make these fiber assemblies, which can be used in transceivers, but also are very important in co-packaged optics.
This expertise in passive optics, I believe, adds to our competitive advantage. If you put all of that together, I'm very pleased with our strong roadmap in pluggable transceivers. We have 800G in production. You might have seen at the Optical Fiber Conference that we introduced a new product line of transceivers that were specially optimized to work with an optical circuit switch. We expect to have 1.6T in production this year. We actually had three different 1.6T demos at OFC. We have 3.2T in development. Specifically, as I mentioned, we built this 400G starting the optical elements that you need as the whole industry to raise up the data rate to 3.2T and 6.4T. We're also deeply working in co-packaged optics. We're engaged with a very broad set of customers. The implementations are very wide-ranging.
I hope I showed you on the previous slides that we have a broad portfolio of CPO-enabling technologies. I'm sure most people probably saw that at GTC, we were called out by NVIDIA as an ecosystem innovation partner for co-packaged optics. We're very proud of that. We're very proud to partner with NVIDIA and all our other customers working on co-packaged optics. Now I'm going to change topics a little bit. We're still inside the data center, but I'm moving away from pluggable transceivers and CPO now to the optical circuit switch. First of all, what is an optical circuit switch? When you want to transfer optical signals or packets from one place to another in the data center, today that generally happens by transitioning from the optical domain back to the electrical domain. Switching is done electronically.
You transmit back to the optical domain. What the optical circuit switch is, is it allows you to make those transitions while staying in the optical domain the entire time. Now, why would you want to do this? One of our customers posted a paper in a published paper in late 2022 that showed that they actually saved over 30% on cost and 40% on power consumption by implementing an optical circuit switch. Everybody loves to save on cost. There's no question about that. I'm sure you all know that there's a lot of reports of data centers being power limited. If something gives you the ability to save 40% on power, what that means for our customers is they can use that power now to run more GPUs, which is something that they can actually generate revenue off of.
This got a lot of broad attention in the customer base. In addition, another of our customers actually published a paper saying that they use it to improve their network uptime, which again is very, very important in the AI data center. They actually have a rack of equipment. They keep a spare rack of equipment. If they have any failure in one rack, they switch over to the other rack to keep the uptime for the AI data center. They use the optical switch to do that very quickly. It improves their network reliability. We see this as a $2 billion SAM by the end of the decade. As Jim mentioned, this is a SAM expansion for us. It is a whole product line that we were not selling before. We believe we have deeply differentiated optical circuit switch technology.
I'll try to walk you through that today. The conventional way to implement an optical circuit switch is based on MEMS, which stands for microelectromechanical systems. These are fingers that basically move up and down. The light deflects based on the up and down moving fingers. You all know that anything that moves up and down, sometimes it could stick down or stick up. MEMS have an innate reliability weakness. We are using liquid crystal technology. With liquid crystal technology, you put a voltage, it just changes the orientation of the molecules in the material. There are no moving parts. We have used this liquid crystal technology in our undersea ROTAMs and wavelength selective switches. The undersea is the most high reliability requirement in the optics industry. We have shipped that for many, many, many years into that application.
In addition, for MEMS to move these fingers up and down, it takes high voltage components. For the digital liquid crystal to just change the orientation of the molecules, it takes low voltage. High voltage components fail more than low voltage components. If you put that all together, we believe we have a differentiated, reliable—our core competency, core differentiation is reliability. Reliability is very, very, very important to our customers. In addition, we actually were the first tier one optical vendor to demonstrate an optical circuit switch. We did an OFC of 2024 a whole year ago. Part of the way in which we were able to pull this together so quickly is because of that broad and deep technology portfolio. We had this liquid crystal technology that we were using in our telecom products.
We actually had shipped those passive optics I told you about. We had shipped those into optical circuit switches for over a decade. We also have a great software capability that I think maybe we have not talked about externally that much, but I have an opportunity to explain it to you today. We have a really deep software capability. We were able to pull together that liquid crystal, passive optics, and software very quickly to have a demo by OFC 2024. Not just a demo, but we already had announced that we had POs. As Jim mentioned, now we have already shipped first revenue on that product. Very excited about the SAM expansion with the optical circuit switch and very pleased with our position in the optical circuit switch. Also, optical circuit switch, it is not just a single product.
It's also a whole roadmap. We have a strong roadmap of products. We expect to production release our 64x64 this year. We are working on higher capacities also for this calendar year. We see interest from customers in using the optical circuit switch for DCI. We need to change to the C-band from the O-band, change different wavelengths for the optical circuit switch for DCI. We see interest in scale up where people want a more modular optical circuit switch, which we're working on. Eventually, people always want higher and higher, what we call RADIX, meaning you can connect more and more and more GPUs together. Long, strong roadmap of optical circuit switch. Okay, now that's all we're going to talk about the data center today. Now I'm moving to the other $12 billion of our $44 billion SAM by 2030.
This is our communications business. Just as a reminder for us, now communications is everything outside the data center. I'm going to focus on two primary topics. One, which is DCI or data center interconnect. The second one is our transport. We'll talk about those real briefly. As Jim mentioned, and as I said, as we start to see data centers become actually power limited, you see people having the need to spread AI workloads over more than one data center. This drives up the need for very high bandwidth connections between data centers. We call that the DCI. These are served by transceivers. Again, you may know them by their standard name, which is ZR and ZR Plus. We see the SAM for that market growing to $4 billion by 2030.
This is another place where we have a very broad and deep technology stack. In this case, it has helped us in time to market. I'll try to give you a few examples. Probably some of you may have followed that we introduced a product called a 100G ZR QSFP28. What this product was, was we have customers who have 10 gig, 100 kilometer links. By deeply engaging with those customers, our engineers and product managers understood that our customers needed a simple way to upgrade that portion of the network to 100 gig without ripping out and replacing a whole bunch of different technology. One way that we could enable them to do that is to cram all of these things that you put into 100G ZR into the small QSFP28 footprint, which was very, very challenging.
Because we own all these elements, we could make that decision to make this product before roughly anybody else in the industry even knew you needed this product. We have a multi-year lead on that product. One of the ways in which we got a multi-year lead on that product was our indium phosphide tunable lasers, which we also design and manufacture in-house. Our team, which has 20 years of deep experience, worked on a very low power novel indium phosphide tunable laser. That is one of the elements that enabled that product. We also were first to demo and are first to market on 800G ZR. That also is enabled by indium phosphide. For that product, we use our capability that is called indium phosphide photonic integrated circuits.
If you look at the picture on the bottom left there, that picture is actually a complex integrated circuit. It has four modulators, has semiconductor optical amplifiers, it has photodetectors, all in one piece of indium phosphide. Having that capability allowed us to be first to demo and first to market at 800G ZR. In addition, some of our customers are interested in L-band, meaning adding new wavelengths down the same fiber so they can expand their capacity. They can send from one place to another without digging new fiber trenches. Indium phosphide, because of these built-in semiconductor optical amplifiers, is a great match for that. Again, we were among the first, or I think we were the first to show L-band capability on these 800G ZR products. We also have a strong roadmap to support the DCI transceivers.
We have 100G and 400G in production. We're ramping our 800G now. We have our plans in development for 1.6T and 3.2T. Very robust roadmap pipeline to address that SAM. Okay, final topic for you today is transport equipment. First and foremost, what is transport equipment? If you think about it, if you want to send a signal, say, across the entire United States, you can't do that without amplification. The signal will decay as it goes partway, and you need to boost it back up. We do that with our fiber amplifiers. In addition, for long distance, we use wavelength division multiplexing as an industry. You might want to peel off, say, one wavelength for New York, a different wavelength for Boston. You need the ability to do wavelength switching and routing. You need to monitor all of this.
All those pieces together are what we're calling transport equipment. They amplify, switch, route, and monitor what's going on in the longer distance telecom portion of the network. We see this as the $5 billion SAM by 2030. In this space also, we have a very wide and very deep technology portfolio. Some of the products we sell into this space are erbium-doped fiber amplifiers. We sell wavelength selective switches, reconfigurable optical add/drop multiplexers, optical channel monitors. One of the things I'm personally very, very proud of is our gallium arsenide pump lasers. We've made those for over 20 years. They're a key part of the amplifier, of the amplification system. We have excellent output power out of those pump lasers and excellent efficiency. We also package those pump lasers ourselves.
The combination of those two things put together, we believe, gives us a multi-generational lead over other people in pump lasers. Finally, having all of these capabilities in-house, we were able to deliver another industry first. This was a problem that our customers have brought to us that now they want to send more and more data across, say, the United States, for example, goes through these huts, which provide the amplification, can do switching. Now they want to send four times the amount of information through that hut. The hut already is there. It is not going to grow by 4x. You certainly are not going to be able to provide 4x more power to it. They want a solution that allows them to put through that hut 4x more without growing 4x more power, 4x more size.
We were able to do that. It's called this multi-rail technology. We were the only person to demonstrate that at the Optical Fiber Conference this year. Again, broad and deep portfolio. Bringing it all back together now for our data center and communications business. Again, showing on the left what you can see, hopefully I've been able to convey, that if you think about communications from undersea across the United States in the data center, in the AI data center, you could make a network like that and optically connect it only with Coherent products. We make every product that you need end to end. This is a growth engine for the company. I believe we have a very strong position.
We have, hopefully, I showed you our broad and deep product portfolio and how we've used it to do industry first and to innovate. We're deeply engaged with all of the top customers in this space. With that, just thank you for your time and interest. I will pass it back to Paul.
Thank you, Julie. Thank you, Jim. We've got a great program after the break. I know I recognize and appreciate all the interest in the data center and communications business. But we've also have a great industrial business. My colleague, Chris Dorman, is going to tell you about it after the break, followed by Sherri Luther, our CFO. Please be back in your seats by 10:15, 10-minute break. Thank you, everybody. Hello? Am I on? Can you all hear me out there? Greetings, greetings. If I could ask everybody to start getting back into their seats, we want to keep the show on time. Thank you. Thank you again for joining us. I think we've got an equally good second half of the show. We're going to start off with my colleague, Christopher Dorman, who's our EVP of Lasers. He's going to talk about our underappreciated, and I do mean that, industrial and lasers business. Chris?
Thanks, Paul. I'm here. Hi, good morning. I'm Christopher Dorman. First, I'll give you a little bit of my background. I started out in London, late 1990s, doing a PhD in quantum physics. The good news is I've almost entirely recovered now and have successfully reintegrated with society. Why is that relevant? That's relevant because it turns out to do quantum physics experiments, you need to use a lot of different lasers technology. After a couple of years in the semi-industry, I made my way back to photons, which is significantly more exciting than electrons. I started working for Coherent 22 years ago and started with managing a product and then gradually more and more products until I'm now responsible for the lasers segment. I'm going to be talking about the industrial segment, as Jim covered before.
I'm going to be, first of all, I said I'd completely recovered. I haven't completely recovered. I'm going to start out by doing a few science slides, talking about key parameters that you need to know about in the laser landscape that will give more color to how we interact with the various end markets. I'm going to talk about our portfolio, the five different kinds of lasers that Jim referenced before and how important it is to have a broad portfolio and a deep portfolio. I'm going to finish by digging into these end markets and seeing the dynamics, the long-term dynamics for each of these markets in terms of the increasing amount of photons that you need, the increasing amount of laser processes you need, and the increasing complexity of the lasers which are used for those processes.
Let's go back to the science. Just three slides. Be patient. The three critical parameters when you're talking about laser technology are power, wavelength, pulse length. Power, wavelength, pulse length. I'll go through each of these in terms of why they're important. Now, first, power. Power is the one you're probably quite familiar with. If you're a fan of James Bond, it's the one that you might be recognizing from Goldfinger with James Bond tied to a table and saying, "Do you expect me to talk?" "No, I expect you to die, Mr. Bond." This is a kind of initial conception most folk have about lasers. It's the amount of power. Certainly, it's important. It's about how many photons you can deliver. It's about delivering heat in a very precise way. This is melting metal. This is scribing, welding.
It's the first thing you think about when you think about laser technology. We've seen quite a lot of evolution here and incredible advancements over the last 20 years. Twenty years ago, a CO2 laser cutting through metal, 10-millimeter metal, half-inch metal at a snail's pace. Today, slicing through quickly with 20 kW of fiber laser power. Let's look at what a fiber laser is in terms of what it's achieving. Today's modern fiber laser is taking the energy, the power consumption of a neighborhood. It's delivering it down a fiber the width of a human hair. The intensity of light that you're getting at the end of that fiber is around about the same intensity of light you'd experience if you were happily sitting on the surface of a star. Incredible advancements, incredible technology. This is a super exciting advancement.
But it's important to understand wavelength, and it's important to understand pulse length in the context of the evolution of the laser industry. The second important parameter is wavelength. Twenty years ago, a typical laser would be 532, a nice green color. Today, it's 266. Halving the wavelength, doubling the frequency, much, forgive me any physicists listening, but much smaller photons looking for smaller things. In terms of modifying material, this means you can focus the light smaller with a shorter wavelength, but also it's absorbed in a shorter distance. X, Y, and Z, much higher levels of accuracy of machining, of scribing, of interacting with materials. The other important element of wavelength evolution to shorter and shorter wavelengths is when you're detecting. A semiconductor wafer inspection is a good illustration of this.
You're looking for defects on a semi-wafer, either a bare wafer or a patterned wafer. If you were to take a semiconductor wafer and scale it up to the size of Manhattan, looking for defects is the equivalent of scanning for an ant, then identifying the species of the ant all within 20 seconds. Scanning Manhattan for an ant, finding it, identifying it, and finding out what species it is within 20 seconds. That's how accurate modern semiconductor wafer inspection is. Wavelength, another critical parameter. The third one, and this is probably the most underappreciated parameter of all, is pulse length. Early 2000s, maybe a nanosecond. Late 2000s, a picosecond was a primary laser. Today, femtosecond. Let me just explain what the scale is for nano, pico, femto.
If I was to get a torch or a flashlight and fire it into the sky for one second and then turn it off, that would be a pulse 187,000 km long. That would go off into space. If I was to turn on a flashlight for 10 nanoseconds, let's say 8 nanoseconds, it would be a pulse about this long. If I was to turn on a torch for 300 femtoseconds, which is modern laser technology, that would be a pulse length about the width of a human hair. You have gone from the scale of 187,000 km in a second to the width of a human hair in terms of how long that pulse is. Incredible advancements in pulse length. What does that mean for interacting with material?
Let's just take the average power from the power from an incandescent bulb, and let's do laser things with it. First of all, we'll point it in one direction, and then we'll collect it into a 300 femtosecond pulse. You take that 40 watts, direct it, compress it into 300 femtoseconds. The peak energy in that 300 femtosecond pulse will be about 100 million watts, which is about the power consumption level of a city. From 40 W, you've created a power level of the consumption levels of a city. Why is that important? That's important because in a modern manufacturing, like a smartphone, you're trying to cut without depositing heat. If you took a smartphone and put it on the surface of a star, it wouldn't really operate for very long.
If you're cutting the delicate materials inside a smartphone, the SIP packages, the screen, the glass on the screen, the OLED, the sapphire, you need very short pulses to cut without heat. Three hundred femtoseconds is much faster than the speed heat travels. You want to get in, pull molecules apart, be gone before the molecule next to the molecule you've impacted even realized anything was happening. This is cutting without heat. Power, wavelength, pulse length, all key parameters in laser technology. I wanted now to talk about our portfolio and how it fits into those three parameters. What you see is we have a portfolio of every kind of laser that's needed for modern industrial manufacturing, arranged here by wavelength, but also calling out the pulse length and the power levels that you need.
You need that kind of full portfolio to cover the breadth of modern manufacturing. A level deeper here, here is how those parameters, how those lasers interact with each of our end markets, semi-precision manufacturing, display, instrumentation. What you can see is that you need a full suite of the technologies, a full suite of these five lasers in order to cover those end markets. If you're missing one, you've got a problem. I mean, the competitive landscape in which we live tends to be confined to companies with either a specific technology, in which case they're confined to a particular column on here, or they've grown from a specific market, in which case they are confined to a specific row here. For modern manufacturing, you can't afford to be confined to a column.
You can't afford to be confined to a row on this table. As Jim explained before, I'll just take one example, one of hundreds of examples of modern manufacturing, a marvel of modern manufacturing, the smartphone. I'll take those five laser categories, and you can see how you need each of those different kinds of laser technologies in order to manufacture a modern smartphone: the screen, the camera, the electronics inside it. You need a complete array of technologies in order to manufacture a modern device, whether it's a phone, a car, or many other kinds of modern technology. What's also important, and Jim touched this as well, is not only the breadth of the portfolio, the five kinds of lasers, but also the depth of the portfolio, the deep tech stack in which we have.
It's not easy to develop a modern femtosecond laser without having access to the crystal technology, the fiber technology, the pump technology. You need all of these in your toolkits because the engineering teams work with each other to push the boundaries of modern technology. Each of the callouts on this slide are internal technologies, engineering teams working together to push the boundaries of modern industrial laser technology. I should also point out that another critical component of laser technology is the software. There's an incredible amount of software inside our laser systems. There's analog electronics, digital electronics, control software, the software to communicate with our customers, pulling together common architectures to leverage our expertise on software. The same applies in our materials capabilities. What's interesting here with our materials capabilities, several of our photonics material capabilities here, is the overlap.
Whether it's a transceiver, a laser, or the materials we sell to our customers, for example, in semi, we innovate at a material level, a fundamental material level, which then makes its way through into the laser systems. There's a significant overlap between those capabilities. An illustration here would be the modern semiconductor manufacturing processes and how that overlaps with our material capabilities. Depending on how you define it, 10 key steps in semiconductor manufacturing. Eight of those steps are enabled by Coherent's unique and incredibly deep materials expertise. I'm going to talk about industrial end markets, a focus on lasers, go through each one in turn. You see the critical parameters in those markets. In each one, I'll explain the number of photonic steps. The number of laser processes is increasing with time.
The use of those lasers is increasing with time and each generation of that technology. Also, the photonics technology you're using, whether it's a wavelength, a pulse length, all these other parameters, means that the lasers are increasing with complexity. A combination of more and more laser processes in modern manufacturing, together with the fact that those lasers are getting more and more complex on a generational basis, underlines the growth of the Coherent industrial growth is greater than the growth of the end markets. More photon processes, more complex photons. First of all, a good illustration is the semiconductor market. It's one which we tend to be more familiar with, and we're all very familiar with Moore's Law. Jim, you said quarter micron. I'm going to use a physicist term of 250 nanometers if that's okay. You're an engineer. I'm a physicist.
Definitely 250 nanometers, 20, 25 years ago, evolving in node size from 250 to around 3 today and moving on even smaller. What does that look like for semiconductor wafer inspection technology? As I mentioned, more and more photonic steps, a single defect of a certain size is going to give you much more of a problem as your node size gets smaller. In order to have appropriate yields, in order to not have a wafer full of defect-ridden dies, you need to run that inspection process many more times. From 2, 20 years ago through maybe 80 now, pushing up towards 100 in the future. More and more photonics processes. The photons there are getting more complex as well. Twenty years ago, 25 years ago, you were using a green laser. Today, you're using a 266 laser.
The complexity of the laser is increasing as well as the use of the laser for wafer inspection. I want to take a look inside the one in the middle there. It's called an Azure laser. You can actually have a look at one after the presentations. We'll open it up and look inside. I want to give you a visceral example of why this laser is more complex than the previous generation. If you look at the green breakout box, that represents the entire laser from 20 years ago. Generation to generation, what was last time around laser is now just a subcomponent. It's a green CW laser, constant flow of green photons. We have to take this to 266 to look for smaller defects. What does that involve? That involves something called resonant enhancement. We have to take this green light.
We have to make it go round and round a cavity hundreds of times through a crystal, which doubles the frequency, halves the wavelength. It has to go round and round that. Every time that light goes around, it has to completely align with the waves from the previous trip around the cavity, which is quite difficult. It means controlling the mirror position. This is kind of difficult to get your head around. You've got to control the mirror position in that cavity to less than the width of an atom. In order to have the light travel around hundreds of times in order to generate the UV. That's, as I said, quite difficult, a bit of British understatement. It's actually not the most difficult part.
What's actually really, really difficult is ensuring that this laser will run for 20,000 hours, having that close link with the crystal technology groups, having the software to allow this laser to run for many thousands of hours in an industrial environment. If this laser stops, the fab stops. So reliability and productization is actually even more complicated than controlling a mirror to less than the width of an atom. Let's visit another couple of markets we've partially touched on. I'll divide this in the precision manufacturing. I'll divide into power, lots and as many photons as possible for cutting and welding and melting, and ultra-fast, which is about cutting delicate materials without transmitting heat. The car is a very good example of precision manufacturing with precisely delivered heat. Twenty years ago, relatively few processes. Today, 40.
In the future, that continues to grow with the adoption of electric vehicles. Each generation of automotive manufacturing technology has an increasing complexity of those lasers. In this case, more and more photons, higher and higher power, and now the interaction of pulse. On the smartphone, a similar dynamic, an increased number of laser processes. Really, every single part of a modern phone is touched by a laser or manufactured by one of our five key types of lasers. An increasing complexity as we move from nanosecond to femtosecond to UV femtosecond in the future. A femtosecond, by the way, is a millionth of a billionth of a second. Evolution there, as with semi, more photonic steps, increasingly complex photonic technology. This is what I'd like to touch on, an enabling factor, the depth of our portfolio.
There's a slightly different dynamic, but relatively similar in each of these markets. Really, in kilowatt fiber laser manufacturing, the only way to be successful is having a deep tech stack, which includes all of the critical components. In order to be competitive there, you have to have access to the diodes, the fiber, to the fundamental EPI that goes into the laser diodes. On the ultra-fast side, where you're driving pulses shorter, it's critical to have alignment between the different teams for those various technologies to be able to innovate quickly, to be able to stay ahead, to drive pulses shorter, to drive wavelengths shorter in the ultra-fast regime. We're busy launching. We have a show called Laser Munich coming up very soon, in which we'll see many more launches. I wanted to pull out a few of the most recent launches.
We launched in March, the latest in our kilowatt fiber laser series, the AMFL, very precise automotive, medical device manufacturing, consumer goods manufacturing. If you follow, we're really building out a pretty exciting portfolio of kilowatt fiber lasers. Now, display, I wanted to do similar echoes through each of these. I wanted to explore the dynamics of what's happening in the OLED market, where we produce very powerful UV photons using an excimer laser, so referring back to those parameter space, to manufacture to anneal OLED as it goes through its manufacturing process. The dynamics there are about half of phones today are using OLED screens. That continues to rise. In the future, that will push closer to 100%. The big change in the OLED market is the adoption of OLED into laptops and into tablets.
If you do the mathematics of the number made, the surface area, multiply it all through, work out what's happening in the next few years, you can see that the requirement for square footage of OLED is approximately doubling in the next few years. That's a big dynamic in the OLED market and gives us a great tailwind in the OLED business. Super exciting for me is where TVs are going. This is towards a technology called microLED. For Coherent, with the technology I'll be discussing in a couple of slides' time, we see a $1 billion opportunity per 10% of microLED adoption in TV-sized panels. OLED first. I'll do OLED first, then I'll cover microLED. OLED, I think many of you will be familiar with this.
It's about the generation of the panel sizes from which you cut the phones or the tablets. OLED, really not a feature in the 2010s. Today, 10 million square meters per year of OLED are produced. In future, 20 million. That factor of two I was talking about. The lasers are growing in complexity with each generation. That's because we not only make the laser, but we also take the laser and spread its beam into a long, thin line for passing the OLED past and annealing the silicon backplane. The lasers are getting more powerful, and the line beam delivery systems are getting more complex. You can see that when I say complex, I mean complex.
I don't know if you can see the person in the picture there, but this is a significant scale of laser technology that we are shipping to our key customers. We are launching now, around about now, at the moment, a tool called the UV Transfer. Very proud of the team in Gottingen in the excimer laser business unit, using excimer lasers to take the microLEDs that are produced on a wafer and transfer them to the TV-sized panel. We are launching the first tool, which is compatible with manufacturing levels of transfer rates for microLED production, transfer more than 50,000 microLEDs per second, which takes the transfer time for a panel from hours to minutes. It is super exciting that yet again, excimer laser technology is at the heart of evolution in display technology.
Finally, bioinstrumentation, health sciences, one of the main applications in health sciences is something called flow cytometry, where you're analyzing cells per second, analyzing cells for cancer, for disease. You're analyzing them to see their response to a vaccine. You have the count of number of cells per second that you can analyze here. Today, 50,000 cells can pass by a bunch of lasers. You can analyze them for a range of parameters. You can detect leukemia in one blood cell in 50,000 in one second that's passing by. Super important technology. Again, showing that similar dynamic of increased laser usage and increased complexity. In this case, flow cytometry in the 2000s would use two different color lasers. Today, six different color lasers. In future, nine plus, which we're now packaging into a common platform to make it a little easier for our customers.
With the backdrop of an aging population, with the backdrop of a move towards personalized medicine, you can see the importance and the evolution of this particular technology. A similar evolution to semi, a similar evolution to display, a similar evolution to precision manufacturing also applies in health sciences. We are launching two worlds colliding here, ultra-fast and bioinstrumentation. Launched this quarter, femtosecond laser, which is used for looking at neurons, very accurately mapping neurons in the brain, three-dimensionally imaging not only neurons but neuronal connections on a mobile subject. Super exciting to be at the forefront of the fight against neurodegenerative diseases. I think you can see one of these behind on the display later on. This is the underappreciated aspect of the way we interact with our customers.
I get to tell a personal story, and it's great to be able to talk about myself a little bit more. No, I'm just kidding. This is a story of a laser called the Azure Laser. It really, really illustrates the relationship we have with our customers and how service revenue fits in. House service is not optionally phoning up and wondering if there's a part in stock. The Azure Laser is a laser for semiconductor wafer inspection. It's the first market I was talking about. It's looking for those tiny defects on a wafer. This laser launched, this particular one, it's actually a couple of generations later. We don't make this one anymore, but it was the one I first sold when I was a product line manager 22 years ago. It's a UV laser. UV photons are not kind to optics.
They're not kind to crystal. So this laser has various crystal spots. It burns through one in 250 hours, automatically moves to the next crystal spot until it's used all of those spots up, in which case it needs to be refurbed. Refurbing means sending it back to the factory where it gets opened up in a clean room. It can't be opened up anywhere else. All the UV bits replaced. We would have sold this laser when I was just starting out for around about $200,000. In the subsequent 20 years, it's come back through the factory for refurbishment 20 times, 20 times. Each of those refurbishment activities is approximately a $60,000 activity.
It is relatively simple mathematics to work out that if the original sale was a 1X, that actually most of the revenue is coming from the service activity of that laser in order to keep it running. The great news is that fabs do not run for five years, then shut. Fabs are kept open for 20 years plus. The lasers I was negotiating and talking to my customer about 20 years ago are still going back round through the factory, getting refurbished. Most of the revenue for those lasers has come from the service activity. I think that is not so well appreciated in terms of how service works for us. We are at 25% now in our industrial segment of service and consumable revenue. That is growing.
If you do the mathematics of our increasing install base for all of these lasers over time, you can see that this is a fantastic foundation for the business. In summary, more lasers for modern manufacturing, all of the processes, more complex lasers, so faster than market growth. You need the five key pillars, the five key laser types in order to interact sensibly with any modern manufacturing capability. The portfolio is important. You need the deep technology stack because you're not going to be able to innovate quickly. You're not going to have the right cost structure unless you have access to a range of internal all of the optics technology inside. On the service side, you can see long life, high margin revenues. Thank you very much. I'd like to hand over to Sherri Luther, our CFO.
Thank you, Chris. It's great to see everyone here at the New York Stock Exchange. Very exciting to share with you our financial model. It's also great for me to be back at Coherent. Very excited. As many of you may know, I spent about 17 years at Coherent, or Coherent Classic, as we internally refer to the company. Some of the key things that I really enjoyed in my earlier tenure at Coherent, of course, the people, the collaborative environment, but the innovative technology was very exciting. Technology that, in many cases, Coherent was the only company who could develop and manufacture these products. When I did my due diligence before rejoining Coherent, I saw that this new, larger company had an even greater depth and breadth of technology than what Coherent had. That was really exciting to me.
Some of the other things that I saw when I did my due diligence prior to joining the company, in particular from a financial perspective, included certainly opportunity to drive higher revenue growth, double-digit revenue growth. After all, we have this deep and innovative amount of this powerhouse of innovation at the company. I saw the opportunity to drive higher gross margins. After all, Coherent has higher gross margins than what the company has today. I like higher gross margins. I also saw the opportunity to increase the efficiency in the way that we make our R&D investment decisions to make sure that we're investing for the long-term ROI of the company. I saw opportunity to increase and improve the level of efficiency in our SG&A, in particular the G&A portion of the business.
I also saw opportunity to be more disciplined about the way that we allocate our capital to reduce the debt on the balance sheet and reduce the interest expense that hits our P&L. All of these areas really represent key financial priorities for the company. I'm excited today to take you through our strategy as to how we plan to drive increased profitability and enhanced shareholder value. First, I'd like to take you through a little bit of where we've been and the progress that we've made to date. We've had strong revenue growth, 24% from our most recent quarter, year over year, on AI data center growth. I'm pleased with the progress that we've made on gross margin, 490 basis points year over year. We've been disciplined about the way we've been making our OpEx decisions.
have reduced our OpEx as a percentage of revenue to 19.9% year- over- year. The combination of margin expansion and OpEx discipline has driven operating income improvement of over 600 basis points year over year. With our debt paydown and debt repricing, we have reduced the amount of interest expense in our P&L by 22%. All of that has driven 2.4 times increase in our EPS. While this is good progress, it is not enough. We can do so much more. Let me start with revenue.
Jim talked about driving double-digit revenue growth with our company target, the organic company target growth rate of 10% to 15% plus on a three to four-year CAGR, with our industrial market segment growing 5% to 10% on the greater increased use of lasers, as Chris talked about, and our data center and communications segment growing 15% to 20% plus. We're in early adoption of AI, as Jim mentioned. We're well positioned for share gain. With that double-digit revenue growth target of 10% to 15%, we're also focused on expanding our gross margin. We've been talking to you now over the course of about a year of driving a gross margin to a goal of greater than 40%. Today, we are raising our Gross margin target to greater than 42%. I'm very excited about raising our target and driving our gross margin expansion strategy.
Let me tell you a little bit about the elements of that strategy. Certainly, double-digit revenue growth is going to provide a benefit to Gross margin. Higher revenue is a tailwind to gross margin. The flip side of that higher revenue is the mix of that revenue. The data center and communications part of our business, that Gross margin is lower than the industrial part of our business's Gross margin. That increased mix of data center can provide a headwind to mix or related to mix. The net of the higher revenue offset by mix impacts is a net positive to our gross margin. There are two other areas that we're focused on for Gross margin expansion. The first is pricing optimization. The second is cost reductions. Let's talk about pricing optimization and what that means.
Pricing optimization is pricing our products for the differentiated value that they provide to our customers. We've developed tools and analytics to help us price our products for the value they provide. Also, the depth and the breadth of our product offerings are beneficial as well to our gross margin. In fact, Chris talked just now about how excited he was on the service revenue stream. That is also exciting to me because that's a credo to our gross margin. We have long-standing relationships with our customers where they are vested in our long-term product roadmap and the development of products that help them differentiate their products. That allows us to price our products for the value that we provide to our customers. Cost reductions, the second area. This is a very broad area. It's everything in cost of sales. It's direct material cost reductions.
It's reducing our overhead cost, our fixed cost components. It's yield improvements. All of these areas are encompassed in cost reductions. I'll give you some examples. Direct material cost reductions. What we've done is we've looked at the products that we sell the most of, the bills of material for the products we sell the most of. We've developed plans as to how we can cost reduce those bills of materials to reduce cost. Other examples that we've talked about as well are yield improvements. These initiatives are not one-time projects that once you finish one, you're done. These are initiatives that are ongoing. This is the way we do business. This is in our DNA in terms of how we do business and trying to drive cost reductions and pricing optimization.
Now, you may ask, what parts of the business should you expect to see pricing optimization benefits versus cost reductions? The way to think about that is that pricing optimization, we will expect to see more of the benefit of that element of our strategy in the industrial part of our business. For the data center and communications part of our business, we are focused more on share gain. Cost reductions, that is everywhere. That is in industrial. That is in the data center and communications part of our business. That will be everywhere in terms of those initiatives that we expect to see. In terms of the relative order of magnitude of each of these initiatives, the net positive impact that I talked to you about in terms of volume offset somewhat by mix, that is a net positive to our Gross margin.
When you look at pricing optimization, that is incrementally higher than the net positive from volume. Cost reductions is incrementally higher than pricing optimization. As you can see depicted on this chart, cost reductions, that's the biggest area where we expect to see the benefit, followed by pricing, followed by that net positive on volume. All of these initiatives combined together are what are going to drive us to our Gross margin of over 42%. You may ask, why not higher? Can you do more? I'm sure you will ask. Of course, we will try to do more. As far as our target of over 42%, let's get there first, and then we'll be happy to raise our target at that point in time. Let's get there first.
I'm very confident with the initiatives that we put in place that we can get there and very excited to drive these initiatives. Now, let's talk a little bit about the way that we are investing for our long-term growth of our business, in particular in OpEx. Our target for OpEx is 18%. And within that, R&D, 10% of revenue, and SG&A at 8% of revenue. One of the things I saw when I did my due diligence on the company before rejoining was that the company spends more in SG&A than it does on R&D. We are a tech company. We should not be spending more on SG&A than we are on R&D. So what are we going to do about that?
With our target of 8% for SG&A spend, we're going to drive greater efficiency in the way that we manage our SG&A, in particular the G&A component. I'll give you some examples. We're going to establish shared services in low-cost regions for some of our G&A functions, low-cost regions in Asia, low-cost regions in Europe. That will not only allow us to reduce cost, but it allows us to put like work streams together that make us more efficient as well as more cost-effective. Another example is Jim talked about the inorganic investments that the company has made over the years, acquisitions that have been really good for the company. The other thing that that has given us is a lot more ERP systems than we really need. We have kicked off an ERP implementation consolidation plan to get the company on one ERP system.
Of course, that drives greater efficiency as well, like work streams together, more efficiency, reducing costs. Jim also talked about facility consolidations. Those are other examples of areas where we can reduce our SG&A costs. We are very focused on driving efficiencies in the SG&A function to get us down to that target of 8%. Now, of course, leverage higher revenue will benefit from an SG&A percentage of revenue. The math will just tell you that. We are not relying on higher revenue to get our SG&A target down to 8%. We are focused on driving these greater efficiency initiatives because that allows us to do better business, right? That is just good business. Now, from an R&D perspective, our target is 10% of revenue. We believe that is the right amount of spend for the company to drive the innovations that we want to drive for long-term shareholder value.
You've heard from Chris and Julie today what a great innovative powerhouse the company is. We want to continue those innovations. Our customers rely on us for those innovations. Certainly with higher revenue, 10% of revenue, again, the math means we can spend more on R&D. The other thing that you are already seeing in our numbers from an R&D perspective is that shift, that shift of our R&D spend away from the non-strategic businesses that Jim talked about to the profit and growth engines. That shift is well underway, allowing us, again, to focus our R&D dollars on the parts of the business that is really driving the growth and profit for long-term value. That's what's built into the 10% for R&D as a percentage of revenue. Again, target of 18%, really excited with the initiatives that we have to drive that.
What does that all equate to? Higher Gross margin, more disciplined SG&A spend. That drives greater profitability. Our operating margin target is greater than 24%, again, the result of executing in all those areas and certainly double-digit revenue growth as well. Now, higher profitability means greater cash generation. I like cash and cash generation. In our most recent quarter, our debt leverage ratio is down to 2.1 as defined in the credit agreement. We have a target here for debt leverage now as part of this long-term model of less than two times. I'm really pleased with the progress that we've made to date. In fact, in FY2025 to date, we've paid down 70% more than what we paid down in debt in all of FY2024. We are really focused on reducing debt.
The debt leverage, the ratio in reducing it is all good and certainly really pleased with our target here. Lower debt means less interest expense. I shared with you earlier the 22% reduction that we've already achieved in interest expense, that burn on the P&L. We want to continue reducing that impact to our P&L. We've also implemented several repricings in the course of the past fiscal year that also reduces our interest rate. We're really focused on debt deleveraging, but also reducing that interest expense on the P&L. From a total capital allocation perspective, what all of this means is that number one priority is the organic investment in our business, driving those R&D investments that drive long-term ROI for the company, CapEx investments as well that drive the long-term ROI for the company.
A very close number two priority is making sure that we are deleveraging our balance sheet, making sure that we're reducing that interest expense and really strengthening the Balance sheet even further. From an M&A perspective, we're focusing on looking at those opportunities that are complementary and adjacent to our core business. All of this focused on increasing profitability but driving enhanced shareholder value. From a total financial target model perspective, we spelled it all out here for you on one page. Very excited to drive the opportunities that I saw before rejoining the company. I even have greater conviction after joining the company that we can drive these initiatives. I'm very excited to drive that forward. Again, revenue growth, double-digit revenue growth, growing faster than the market, 10% to 15%. This is organic revenue growth.
Gross margin greater than 42%, really executing on that gross margin expansion strategy, certainly getting better leverage from higher revenue volume, but really focusing on pricing and cost reductions. Our OpEx target at 18%, being very disciplined about the way that we invest our dollars and reducing that SG&A as a percent of revenue in particular, but really making sure we're spending, investing for the long term in R&D. Our operating margin target of greater than 24%, really expanding our margins and our profitability. All of this focused on driving long-term shareholder value. A couple of other points I'll make before I close is just to reiterate what we said in our last earnings call regarding tariffs and the impact of tariffs to the business in our current quarter, where we have said that we do not expect it to be significant.
The other thing I'll just reemphasize here is what Jim talked about is the flexible and adaptable supply chain that we have that our customers value very highly. It is certainly a benefit to us. What you can expect to see from a reporting perspective, starting our FY 2026, which is the beginning of July, we are moving from a three-segment reporting structure to a two-segment reporting structure. On the left-hand side, you see the three segments that we have been reporting on and will continue through the end of FY 2025: lasers, materials, and networking. That is today and through FY 2025. Starting FY 2026, industrial, data center, and communications, those will be the two market segments that we will report on from a financial perspective. Why are we changing that?
Because it aligns with the way we look at our business and the markets that we look at our business from. Pretty simple reason why. I just wanted to give you guys the heads up to expect that starting FY 2026. The other thing I'll just remind you, industrial business, high gross margin business, happy to have that as a separate segment. The data center and communications, that's our fast growth, fast growing business. The key takeaways for you today, we are our revenue target of double-digit revenue growth, 10% to 15% plus. That is our revenue target. Margin and profit expansion, gross margin greater than 42%, really focused on our gross margin expansion strategy. Pricing optimization, cost reductions, that is in our DNA.
That is the way that we are managing our business to really drive enhanced profitability as well as shareholder value. Certainly being disciplined about the way we invest our OpEx, very important. We need to invest in the R&D spend that drives the highest ROI for the company, but be more disciplined about how we're managing our G&A in particular. Disciplined capital allocation, very important. We want to make sure that we are investing in the organic part of our business, number one priority. Number two, very close number two is debt paydown and reducing the interest expense that's hitting our P&L as well as our debt leverage. Thank you for your time today. Enjoyed seeing all of you here today. Thank you. I'll turn it over to Paul.
Thank you, Sherri. Thank you, Chris. We're going to have a five-minute break.
We're going to start up again at 10 after with Q&A. Jim, Sherri, Chris, Julie. Thanks, guys. What we're going to do, you're fine, coach. This one we're going to leave on in the control room. We're going to power it off here. John, wait a second. Oh, sorry. And then turn it back off. Wait a second, and then boom, it's on. And then turn it back off. Hello? We're going to start up again in a minute. If everyone can take their seats, please. Appreciate it. All right, we're going to do a 45-minute Q&A session. Karl Ackerman.
Hi, thanks everyone. Really appreciate the opportunity today. Jim or Sherri, I was hoping you'd discuss whether your revenue growth and margin targets exclude any cost actions in flight, such as the portfolio realignment or asset sales and/or manufacturing consolidation.
If yes, should we expect that most of the cost optimization plans are in COGS, and maybe those cost savings will be used to reduce debt? I ask because your OpEx target assumes that half of the non-core product R&D resources appear to be reinvested back in growth areas. Thank you.
Yeah, I think, and Sherri should definitely comment on it as well. I think the way you can think about it is that the targets that Sherri went over today, we think we can achieve that without any significant portfolio optimization, right? To the extent that we are able to optimize the portfolio further, it kind of accelerates it, right? Let's take for instance, if we were to divest a product line, we would expect that to help accelerate our progress towards those targets, right?
We believe we can achieve the targets without, but it will make it faster. We will get there faster if we are able to divest non-core product lines. The reason for that is because if you look at that non-core product line category, those product lines are running generally lower gross margin than the company average, much lower operating margin percentages. Just taking those out drives the higher operating metrics. Sherri should comment, but any proceeds that we derived, any cash proceeds from those divestitures, we would certainly use that to pay down debt, right? That would be the most obvious usage of it.
Yeah, definitely. I mean, the Newton-Aycliffe sale was an example in prior quarters that we talked about, and we did just that, which I find very exciting because any proceeds we get from those sales definitely would want to earmark that to pay down debt, and that would obviously reduce interest expense on the P&L, et cetera. I am pleased about those initiatives as they happen.
Next question, Simon.
Thank you very much. Simon Leopold with Raymond James. Just quick clarification first. On those CAGRs you gave us, base year is fiscal 2025 for that?
Yeah, we use 2025 as the base year, I think, in those graphs, yeah.
Great. Got that question a couple of times already, so I want to make sure I understood that. In terms of some of the trends in datacom, there is this argument that VCSELs go into decline and get displaced by CPO. I appreciate the commentary you gave us in the last earnings call about the mix within datacom. I'd like to get a sense of how you see the mix evolving over time and what your view is on this idea that VCSELs are going away and that's bad for your business. Thank you.
Yeah, I would say that's entirely incorrect, right? I think Julie should comment a little bit on some of the future usages of VCSELs that we see in CPO applications. We'll come back to that. I think the way to think about it in terms of laser technology is, look, we invest in all the key laser technologies that are used in data center and telecom applications. That's VCSEL, that's EML, that's CW lasers as well.
All of those different lasers have different pros and cons, and they have different places where they're better to be used or not, right? Depending on the application, what we do is we work with the customer, we're agnostic to it, to just work with the customer to decide, hey, which is the best laser technology to use for that particular application. In CPO, I certainly want Julie to comment on definitely CW lasers we expect to be used in CPO, but we're seeing some interesting work that we're doing with customers on the application of VCSEL to CPO applications. I think Julie will comment on that. Also, the other thing that I would add is remember what we talked about with Indium Phosphide capability.
If you look at our transceiver sales today, over half of our transceiver sales today are based on Indium Phosphide, and over half of those transceivers use our own internally based EMLs. CW lasers we've made for many, many years, and that Indium Phosphide capacity that we're ramping up, right? We've tripled it over the last year. We're bringing on six-inch Indium Phosphide capacity next quarter. That's not just for EML, right? That's the near-term use is EML for transceivers, but we're also building that capacity for future CW ramps as well. Why don't you talk about VCSEL?
Sure, sure, yeah. As I showed, and as you know, and as Jim said, we have the VCSEL, EML, and Indium Phosphide lasers with silicon photonics. You have those three technologies for different reasons because they each address a different part of the application space very well.
If you could only have one technology, of course our customers would only want to qualify one technology. But in generally speaking, the VCSEL is the lowest cost and lowest power. If you actually look over history, which I have done, I actually showed it at OFC, the ASPs from third-party market reports of transceivers, the VCSEL is generally the lowest cost or has been the lowest cost for two decades and also the lowest power. The Indium Phosphide has the best performance. As we start going to 400G, actually, people are kind of coming to the conclusion it's pretty hard to do 400G with silicon modulators, actually, right? Indium Phosphide gets a lot more attention, both for the longer reaches, higher performance, 400G. Silicon photonics can provide a cost benefit if it can meet the exact spec.
It kind of falls in between those two. Those fundamentally do not change just because you move the transceiver inside the box, right? To Jim's point, I actually had a talk at OFC this year where I built up a model of a VCSEL-based CPO and a silicon photonics-based CPO to this, well, our team built the model, I presented it, but was comparing them to the same set of specs. What comes out very clearly is that the VCSEL is lower power. Power is very important in the data center. I think just like we see all three of those technologies having their applications in pluggable transceivers, you are going to see all three of those technologies have their applications in CPO. That is my view. Yep. Super.
Samik. For the purpose of the transcript, for those of you asking questions, if you could please announce your name along with your institution. Thanks very much.
Hi, Samik, JPMorgan. Maybe I'll ask you a two-parter, but Jim, just to follow up on the last comment there about Indium Phosphide capacity. You highlighted the vertical integration today, but can you just talk about if any of that capacity in the future, other than being used in your own transceivers, will be for the merchant market as a component itself? A broader question just in terms of the data center growth, data center and communication growth that you're highlighting today of 15% to 20% plus. Just want to understand how much of that is the underlying growth you see in the existing products you're in versus the additional SAM that you're outlining in terms of OCS and DCI.
Just can you talk about what that plus really entails in terms of the opportunity? Okay, got it. On the first, yeah, going back to the first part of your question, the short answer would be absolutely that's a strategic option for the company, right?
We've tripled the capacity. We're going to, with six-inch Indium Phosphide production starting next quarter, that's another nice step function in increased capacity. Actually, what six-inch Indium Phosphide gives us also is a nice cost advantage. The cost improvement is a step function improvement. It is a big improvement for us. Initially, we're using that capacity for our own internal use, but that capacity could certainly be redirected for outside sales as well, right? That could be EML sales to somebody else, or that could be high-power CW lasers used in CPO applications, but that's definitely a strategic option for the company.
On the second part of your question, yeah, we're looking at data center and comms at the 15%-20% plus range is what we said. The plus was kind of to acknowledge the fact that, hey, data center demand, it might be faster than what we're baking into the base model. It's possible. We've certainly seen over the last year or two very strong data center demand, and it's possible that it exceeds what we're factoring into our model.
In terms of how much of that growth is due to SAM expansion, I think that's what versus the general market growth, I would say it's hard to separate that out, but I would say we believe, I would believe that we could achieve that growth level just through the expected market expansion and share gain, that the SAM expansion isn't actually required to get within that range, but it certainly helps us or would help us be at the higher end of the range. Okay. Tom?
Hey guys, thanks for hosting today. Appreciate it. I have one on the data com side and then one on the industrial side. In the data com side, you showed kind of the pluggable roadmap through kind of 2030. You showed CPO both from scale up and scale out, and then you kind of showed a slide where you check the boxes for what you do in a transceiver today, and then you check the boxes for what you see in CPO. From like a dollars perspective, I just want to understand. Today you integrate a DSP into a module, and then you mark that up and you get more dollars for that. Are you trying to say that in the world in which CPO occurs and you're selling all of these products, you can more than offset what you lose with the DSP moving out of your purview? What's the message there? Is there more volume there that helps offset that? Are there more products there that you didn't have today? Can you help me understand that first?
Yeah, when we look at CPO, first of all, the CPO design can be different by customer, right? We're working with multiple customers on CPO designs, and the amount of content can be different depending on how that customer is implementing it. Julie may want to comment on this or give some examples. In some cases, CPO, the total amount of content is about the same as a pluggable transceiver. Sometimes it'll be less. Sometimes there's actually examples where it will be more. One of the things that's very clear to us is CPO will likely be higher gross margin, right? Because take for instance the thing that you talked about with the DSP. We really don't get a lot of margin markup on the DSP, right?
That's just something we're integrating into the final product, and there's not a lot of margin on top of that DSP, right? But there is, most of the margin dollars I consider actually come from all of the optical ingredients for us, right? If you take that DSP out and we're really selling the optical ingredients or the final assembly and test of that photonic, of that CPO module or that external laser source, that's actually better gross margin for us. I think gross margin dollars or percentage-wise, it's actually higher. Did anything I missed, Julie, that you don't?
No, no, not at all. Just to add that what Jim was referring to about extra content is you think about it, you package that CW laser in a separate package. That doesn't happen inside of a transceiver, so that's number one. Second thing is you actually connect that package to, if it's a silicon photonics implementation, to the CPO with a polarization maintaining fiber. We actually make that too. Actually, we did not talk about it today, but we do make it too. You have to get the light from that fiber into, say, the silicon photonics. That is another one of our great things in passive optics, as we make these little micro lens arrays and prism micro lens arrays. You integrate that into a fiber attach unit. The components are partitioned differently. Until you look at the exact implementation of the customers, it is a little sometimes hard to make an apples-to-apples comparison, but there is content there for sure.
If you look closely at that chart that Julie showed of transceiver CPO or pluggable CPO right next to each other, and you look at the bottom, that's what she's referring to is there's some ingredients that are required for CPO that aren't actually required for a transceiver. There's some additional ingredients. Got it, super helpful. Paul? By the way, in our product section behind us, you can see one of the external laser source modules. You can see some of this in the product section behind us.
If I could just add also, as we showed, a lot part of what we see of CPO is actually a SAM expansion, actually, right? The CPO for scale up. Oh, for scale up. SAM expansion, yeah.
Super helpful. Paul, sorry, Tom O'Malley with Barclays. The second one is on the industrial side, so I thought it was really useful understanding kind of the three pillars there. As you move through the portfolio, can you help us give a, just give us a feel for what is at the lower end of that portfolio today and what is at the higher end? You obviously showed growth vectors in both the FPD area and then on the semicap area, but there is also a portion of that portfolio that is more of the cutting and welding, which is kind of perceived on the lower end. Something that I think people ask about a lot and I think would be interesting is just from like a percentage basis, if you look at that, what is that mix look like today in the industrial side?
What would you deem as like high-quality stuff and low-quality stuff if you could lay that out?
Yeah, so Chris, maybe talk about the five different laser types, right? Maybe there's a relative gross margin differentiation. I think, for instance, the products that go into excimer laser or go into display manufacturing, relatively high margin. So there's a range of margins within that business. I know you're very focused on driving gross margin up across that entire business, but do you want to talk about just kind of relative profitability?
Yeah, so if I go through the, I'll go through the five from, it was ordered by wavelength. So excimer laser on the far end, the shortest wavelength. UV photons are pretty horrible to glasses and optics. So those lasers are the high-power UV lasers. They have very good service revenue used for display.
That's a gross margin engine for us. The iron laser that was next is kind of a historical UV laser. It was the first way of getting to very actually low power, but CW UV, and it was the first generation of laser technology to be used in semiconductor wafer inspection. Again, that's a nice market for gross margin, as is the UV section of the diode pump solid state market. More recently, we've introduced a range of kilowatt fiber lasers. This is SAM expansion for us. This is a market we didn't really play in so much before. Bringing the various components together through acquisition over the years has created a situation where we're now very competitively positioned to attack that market, and we're in a great position. We're launching a suite of kilowatt fiber lasers.
That's a competitive market, but there's a real growth in that area. The final one is CO2 lasers by wavelength. There's no other way of producing this very long wavelength. It's 10.2 microns, 10,000 nanometers, very far infrared. This is a power play. There's no other way in solid state of delivering this many powerful photons. That's a kind of market that was, the CO2 laser was used for cutting metals. That's kind of been years ago now, ancient history being replaced by fiber lasers. CO2 lasers now are used for a range of precision manufacturing markets. Exciting for me is there are multiple semiconductor market applications emerging for CO2 lasers. That's kind of in between somewhere in terms of those various markets.
Yeah, and one of the things I absolutely love about this business, and Chris talked about it at the end, is look, as we build the installed base of these industrial lasers, we continue to expand the service revenue, right? The service and replacement parts revenue, that's a really nice revenue stream, very stable, but also a nice margin revenue stream. As we continue to build out that installed base, that just grows as a percentage of the business. It makes that business very, very sticky and very, very stable and a great business over the long term.
Reuben? I'll just stop.
I have a lot of questions. Reuben Rogers, Steve Holtman. Jim, I noticed that you put software on a lot of the slides, and you mentioned software was underappreciated. Maybe, I don't think this is new, right, on OCS systems or some of the industrial laser systems, but how are you thinking about software longer term?
Yeah, that is a great question, and I'm glad you asked. The short answer is we are going to be investing much more in software moving forward. In Chris's business in the industrial side, it's hard to underestimate the actual stickiness and the power of that software because, okay, we deliver the laser system, we're servicing it for years and years, but that software layer that sits on top of our laser, right, that gets integrated into the whole control system in the fab, right? That software layer is incredibly sticky because once you design that software into your overall software control system for your manufacturing site, your fab, whatever facility, that makes our solution very, very sticky.
We'll absolutely be investing more there. I think this is an area where we historically may have underinvested a bit here. I think there's a big opportunity to dial up the investment here and to dial up the differentiation and the stickiness related to software. By the way, same thing on the OCS. Julie touched on this, but I just want to reiterate, we're delivering the whole software layer that sits on that optical switch, right? That gets integrated into the overall data center management software. The feedback that we've gotten from customers on that software layer has been super positive, right? I don't know if you have any?
Yeah, I've been in multiple customer meetings where they give us super strong positive feedback on our software capability, yeah.
Yeah, so I think the company maybe hasn't thought of itself typically as a software company. That's something that we're changing moving forward is really thinking about how we could invest in software to drive stickiness of the products.
Thank you. If I could just follow up quickly on the OCS, in terms of the $2 billion TAM, obviously a lot going on with hyperscalers and it's an area that you're really excited about. In the $2 billion TAM, is that mostly hyperscaler or some of the other areas that you talked about, Jim, part of that?
Yeah, I would say mostly hyperscaler. That's who we would expect to certainly be the early adopters and to drive that certainly that first few years of adoption, right?
I think eventually we would expect to see greater adoption, wider adoption to kind of the next, the second tier, et cetera, right? Initial adoption, hyperscalers, yep.
Vivek? Great, thank you. Vivek Arya?
Vivek Arya from Bank of America Securities. Thanks for a very informative analyst day. I had two questions. First one on gross margins. If we kind of double-click on the two segments, could you give us a sense for where the gross margins are relative to the corporate average today? What I'm really going with that is what's going to be the bigger driver of getting to your 42% plus target, right? Is it the comm segment where I think, Sherri, you mentioned you're optimizing for market share, or is it the improvement in the industrial segment? How revenue-dependent is the achievement of those gross margins?
Yeah, so the industrial part of our segment, I mean, that of our company, that gross margin is higher. That is higher than the corporate average. The data center and communications gross margin is lower than the industrial gross margin. That is the way to think about the gross margins of both of those segments. In terms of what is going to be the biggest impact that drives us to help to get us to the greater than 42%, the biggest box on that chart was cost. Even though data center and communications we talked about, we are trying to share gain is very important to us, but we will get benefits on cost, right? I mean, there is lots of opportunity in data center as that revenue grows to get even more efficiencies and more cost reductions there. Yield improvements are one. I am certainly direct material cost.
The greater volume we ship of products, if we reduce the cost of those build materials, that is an impact, right? That really adds up. The cost element is the biggest because that's the entire company. It's not just data center, it's also industrial. It's really the cost piece of it. Pricing is important. Industrial, even though they have a higher gross margin, we can get more benefit for the value we provide to our customers on the industrial side of our business. I love that service revenue stream. That's accretive to our gross margin. Hopefully that's helpful in understanding some of the elements there.
For my follow-up, Jim, on the restructuring, what is the right way to understand where Coherent is in the restructuring journey? Is everything that you could restructure has already been restructured? Is it that things that, because I think you kind of—I wish I could do it that way. How much of that is still on the come, and is that contemplated in your 10%-15% sales kegger?
Yeah, think about, as I said earlier, think about all of the targets as really organic targets, setting aside any inorganic either divestitures or acquisitions. Think of those as organic targets without M&A, right? What I would say is M&A will, we believe like divestitures would certainly accelerate our progress towards our operating metrics, like better gross margin, better operating margin, right? What I would say on the revenue side is what we would be looking at potentially divesting would be less than 10% of the company's total amount of revenue.
I do not see these as big revenue reductions, but significant improvements in the operating metrics, like I said, gross margin and operating income. Those metrics would benefit, right? The first part of your question was just kind of where are we at on that? Yeah, I would say I am pleased with the progress we have made to date. There are a lot of other actions running in parallel. I would say stay tuned and expect to hear more about some of those restructuring actions and reducing that non-core investment over the coming quarters, et cetera. We still have a good chunk of work to do. As an example of what we have done so far, we have shut down five product lines, most recent of which is we shut down our investment on silicon carbide devices and modules and focused that investment on wafers and EPI.
We have more work to do. I would say stay tuned, you'll hear more on that.
Thank you.
Yep. Blaine?
Thank you, Ryan, comes with Needham. With regards to the comms and data center product lines—
Ryan, that was a nice switchboarding. I said Blaine, but you can go ahead. That's all right. That's all right. Oh, sorry about that.
With regards to the comms and data center, can you talk about the customer segments you sell to there between operators and OEMs and then how you go to market for those? Is it all direct sales or just a little comment would be helpful?
Yeah, I mean, we sell to everybody. I was trying to think of somebody we do not sell to. I think we sell to everybody. It's almost all direct. Yeah, I'm trying to think of, yeah, I'm trying to think of an example of where we don't sell direct. I believe it's all direct.
I think one case that someone might not consider it direct is if we sell, say, a tunable laser subcomponent that goes into, say, a telecom OEM and then sells to—oh, and we do sell that as well. That's a good clarification,
yeah. So we do sell components as well. So we'll build a transceiver for data center or a transceiver for DCI, but then we'll also sell components separately. That would be, I guess, indirect to the end customer.
The majority, the vast majority is direct, as Jim said.
Yeah. All right. Thanks, Blaine, Chris, and Jefferies. I want to ask about on the data center side, you said you want to maximize for share. Is there any particular targets there?
I'm kind of curious, is it more content per device or is it more absolute share of what you're already selling? I'm just kind of curious, the temperature of the world, you being considered a US company, obviously manufacturing in China, can you maybe address what the customers are asking of you, where they want it manufactured, kind of the tariff angle, as well as are you seeing any shifts from pure Chinese vendors to the US ones?
Yeah, look, in the data center business, let's take pluggable transceivers. I think we have a lot of share gain opportunity. I think, and by the way, I think we've gained share over the last two to three quarters, but I think we have more opportunity ahead of us. I think it's a combination of the technology. We spent a lot of time talking about the technology depth.
When the, as the big, especially the big strategic customers are partnering with us, that depth of technology and the ability to innovate across pluggable, CPO, VCSEL, EML, et cetera, that's very valuable to them that they can go to us and know that they get the full suite of photonics technology. The other big part of the discussion is what we talked about manufacturing and not just the scale of what we can manufacture, but the geographic diversity, right? From a customer perspective, they would look at us and say, "Yeah, we're a US-based company." Check. That's definitely a positive, right? We're also a US-based company that has 60 different production facilities across 14 different countries. Half of those production facilities are in the US. We are clearly investing in more US-based capacity.
That six-inch indium phosphide line, which you better believe our customers care deeply about, right? That is going into production in Sherman, Texas, right? We are certainly investing in more capacity in the U.S. as well. I think that whole supply chain resiliency, adaptability, flexibility, where can we make things, where are we making things, or where do we make things today? That is definitely top of mind with customers. They would see us as having probably the most resilient, most adaptable supply chain in the industry. I think we have a question right down here in the front row.
Let me give Meetam, and then we will come to—I know Chris has been chomping at the bit. Okay. We will give Meetam, then we will go to Chris. Great.
Thanks. Meta Marshall, Morgan Stanley. First question, just on OCS, is the opportunity, are there opportunities to gain share with kind of current customers who are implementing OCS, or does most of that growth come from new customers? Maybe I'll just get my follow-up. For Sherri, just as you realize you have a lot of flexibility with the Bain investment, but just anything that we should be thinking of in terms of it has now been kind of four or five years that they've been involved. Thanks.
Answer to the first question is easy. It's both, right? I think there's opportunity with both customers that have already deployed OCS historically, but also with a good set of customers that we're engaged with beyond that as well.
I think one of the things is now that we believe we've kind of solved the—there was customer concern around reliability of OCS because it was a mechanical—the prior versions of OCS were mechanical-based systems, MEMS-based systems. There were some customers that really viewed that mechanical-based technology as a reliability issue and did not want to put it in the data centers. I think that we've solved that problem for them, right? With digital liquid crystal technology, non-mechanical technology, that concern, that hesitation has gone away, right? That is why we're seeing a broader engagement or broader excitement around that technology, I think, with other customers beyond ones that have already deployed it.
Yeah. I think just to make sure I'm answering the right question, did you say the Bain investment? Yeah. Okay. I just want to make sure. Yeah. Sure. We love Bain. We love them.
They're certainly an important investor in the company. I think you're probably referring to the preferred shares. There's all kinds of detailed criteria that's in our 10-K that you probably already have read, but if you haven't, I would encourage you to read it. We're focused on what will maximize value for the company at such time as any of those initiatives or as any of those criteria are met. We obviously keep watching it and see where we are on it, but that's how we think about it.
Chris?
Chris Rolland, Susquehanna. Thanks for the day. I can understand the portfolio optimization or review is still underway and stay tuned on to perhaps some non-core assets still yet to wind down.
Perhaps you can talk about maybe some businesses that we were not necessarily obviously sure on that are going to end up confidently in the growth bucket. Thanks, Jim.
Yeah, I think the way to think about it is the assessment that we did is I just want to clarify something that you said at the beginning. The assessment is done. What we are doing is we are driving all the actions that came out of that assessment. There are some actions that we could execute very quickly. There are other actions that just take a little bit longer and are still in flight, right? When I say stay tuned, it is not so much that we have not finished the assessment. It is more that there are some actions that just take longer that are not done yet, right?
That is what I would say is to expect over the coming quarters, right? The second part of your question, though, is around businesses that need—yeah, there's a profitable bucket and then there's a growth bucket. I was wondering any businesses that might have been on the bubble that we were considering that you've confidently designated as growth. As growth. And we'll keep. There's definitely businesses that we would see as key growth drivers. I mean, certainly, I talked about six of them today, right? In data center and communications, our three big growth areas that I highlighted being transceivers, which I include both pluggable and CPO data comm transceivers. That is absolutely a growth area for us. Optical switching. The third area that we talked about was DCI transceivers. Those three in data center and comms definitely growth areas.
I gave you three, in my prepared remarks, three industrial examples as well. The lasers going into semicap equipment and all of the optical and materials that go into semicap equipment. The OLED, the screen, the display manufacturing equipment like excimer lasers that do OLED manufacturing. The third was precision manufacturing. There's a number of lasers that go into our precision manufacturing segment, which we view as key growth areas. Those would be the six—I would say those are kind of the six core, three in each one of our segments. Those are the six core growth engines that we're investing in.
Yeah. Papa?
Thank you. Papa Sylla from Citi. Just wanted to clarify a little bit in terms of the data center segment between lasers and products or transceivers in general, what's the mix at this point and what's the rationale between kind of selling the lasers independently or selling the transceivers? Is the rationale you always prefer to sell the transceivers when you can and the lasers when there is enough capacity to do so?
Yeah, I would say that we look at it as there's not a—because we sell lots of different components, there's not a simplistic answer to that question. Certainly, if we can sell the transceiver, the complete end product, we're certainly happy to do that, right? It has more to do with what the customer wants from us, right?
If the customer wants us to build the full transceiver and for us to deliver that transceiver to them fully tested, ready to go, we'll do that, absolutely, right? If the customer says, "Hey, you know, I would actually prefer it if I just want to buy your lasers," right? Or, "I just want to buy your isolators," for our big strategic customers, we're okay with that as well. In fact, we make higher gross margin dollars on the components anyway. It really has to do with whatever's best for that customer in that engagement. That's kind of the way we approach it.
Yeah, no, that's helpful. Just a quick follow-up on the CPO. It's pretty clear that obviously everything is—we are selling a bunch of your components. Content could be higher for CPO versus transceiver. I'm curious from the project that we already know were announced, is that the case? That's the first question. In terms of the CPO ecosystem, what kind of behaviors are you seeing from your customers? Is it more they double source component? Is it winner stakeholder component? What kind of behaviors are you seeing so far?
Yeah, I would say on the first part of your question, I can't comment on a specific customer program, right? We try to respect our customers' confidentiality on that, right? On the second part of your question, it depends a lot on the particular application or the customer. You'll see some customers that for that application, they'll sole source it and they'll rely on only one vendor. You'll see others that for a particular application, they'll have a second source or maybe a third source, right?
It really depends on their particular strategy for that particular component. We see both ends of the spectrum, both sole sourced and multi-sourced.
Papa, you're good. Ananda?
Thanks, guys. As you can see me back here.
I can, but I think you're in the last row there. Go ahead.
Appreciate it, Jim. Yeah, thanks. Ananda Barua, Loop Capital. Thanks for all the information today. I guess just two, if I could. Any—as bandwidth gets built out from all the application propagation you guys talked about today, would love to get any thoughts you have on 5G, 6G potential build in the coming years. If that were to amplify, does that fit into your model envelope or could that be something incremental? I have a quick follow-up as well.
We've tried to capture that, right? We've tried to use what we think are reasonable kind of assumptions around the adoption of 6G in our base model and tried to use what are kind of middle of the road third-party estimates to try to factor that into—that's the communications part of our data center and communications. We have tried to factor that in to the extent that it would be adopted faster than what the industry would expect. That would certainly be upside for us.
That's really helpful. The follow-up is just going back to the transceiver business in your conversation around market share opportunity. The conversation that's taken place, not today, but just generally speaking around share shift back to domestic transceiver providers, should any thoughts—should it occur, any thoughts on if it would occur gradually or could there be some sort of tipping point dynamic in adoption where we see a more amplified shift towards U.S. producers? Any thoughts there? And that's it for me, thanks
. Yeah, I definitely think that dynamic is a tailwind for us for sure, right? It's definitely a tailwind. I think that tailwind we've already benefited to some extent in some of the share gain that we've gained over the last couple of quarters. I don't think that was the only reason, but I think that was a factor in it. I think it's certainly a tailwind moving forward. It sometimes can take a while because these products get qualified, right?
They get qualified in a particular usage case and it takes a while before that starts to ramp. There can be some latency. If the application is big enough and highest volume enough, there can be step function increases in volume if it is a big enough application. Yeah.
Carl?
Hi, thanks for the follow-up. Carl Ackerman from BNP Paribas. You highlighted how the data comm growth will be 15%-20% over this three- to four-year period. That, of course, is quite compelling. I was hoping to hear your confidence level in the sustainability or the ramp of 800-gig data comm transceiver cycle in 2025 and 2026 ahead of the 1.6T ramp, which seems to be broadening at this point in time.
As you address that, should we expect that your internal EML capacity with the addition of your 6-inch EML fab today should allow you to increase your in-house EML capacity to say 75% of your total EML procurement versus roughly 50% today? Any way to frame that, the internal mix of that would be great. Thank you.
Yeah, in the first part of your question, in Julie's section, she broke down, she showed the market, how our expectations on market growth of CPO and pluggable transceiver. She also broke it down by data rate. If you look at that slide, I know we do not have it in front of us, but it actually breaks down what we are expecting in terms of 800-gig growth at an industry level. You could assume that our revenue would roughly match the industry breakdown, right?
As she showed, yeah, we are expecting 800 gig to continue to grow in this calendar year to be more, and it'll grow again in next calendar year, right? I would refer you back to that chart of just kind of what's the distribution of demand across 800 gig, 1.6, etc. As she talked about, we're seeing these overlapping cycles. Even though we expect 1.6 to start to ramp this year, 800 gig will still grow underneath that. We still have a significant number of customers that are still transitioning from 400 gig to 800 gig, right? As I said earlier, we service almost every customer, right? There is a good chunk of customers that still need to make the 800 gig transition or are still early in their transition. We think there will be overlapping cycles.
On the second part of your question, which was on Indium, say it one more time. The mix of your EML capacity. Oh, mix. Okay, yeah. What I would say is that we may choose to kind of shift more of our internal capacity towards our own transceivers or shift it more towards internal. I would anticipate us continuing to use external suppliers for Indium Phosphide for EML in particular. We have some really good external partners that do a great job of supporting us. I think there is value in us with our customers being able to offer the internally sourced manufacturing, but also external EMLs as well.
I think that way we're able to provide them not just geographic diversity in where we make things, but even greater supply chain resiliency by saying, "Hey, we source not just from ourselves, but external." That's a value to our customers. They view that as a benefit. As long as our customers view that as a value, I would say we continue to do that.
Tom? We're going to make this the last question. Thanks. Oh, boy.
That's good. It's got to be good then. All right.
OCS. Google's been the big deployer to date, obviously kind of pushing the boundaries forward in the data center. You haven't seen much revenue yet. You just talked about your first shipment. Just started revenue, yeah. Is that one, I guess, why were you guys and others not involved in Google supply chain? Essentially, were they doing that themselves? Is this moving to other hyperscalers? What are they doing differently that's bringing you into the fold?
Yeah, first I want to correct that perception. Julie talked about it. We've actually been supplying components into OCS for 10 years. We've been supplying components into that OCS implementation for many, many years. Think about it as we've already been participating in this market as a components supplier for a long time, right? What's bringing us in now at the system level is we've introduced a new technology, which is the digital liquid crystal technology, which has a significant benefit over all the other solutions out there, which are based on MEMS technology. With this digital liquid crystal technology, it's basically enabling us to offer a system-level solution that's better than what's out there.
That is compelling to both existing customers, but new customers as well. Remember what I talked about earlier is there were customers that were hesitating to adopt OCS because of the reliability issues around a mechanical-based system. Now that we have solved that problem, that hesitation from some of those customers is gone. Now they are much more open to adopting it. I would also mention that the software that we are providing on top, that has also been beneficial to getting customers to adopt it as well. That has helped ease their sort of adoption of that technology.
All right, before we cut over, for those of you in the audience with us live, before we cut over to the informal lunch and also give you the opportunity to see the demos, I want to let Jim give you some closing remarks. I want to thank all of you, both those of you in attendance physically and those virtually. Thank you for joining us today. We greatly appreciate it. Jim?
I just mostly want to just say thank you. Thanks for everyone's time today. Hopefully, as you heard from Chris and Julie, you've got a sense for photonics is becoming so much more important to so many different applications. We talked about.