Nokia Oyj (NOKIA) Earnings Call Transcript & Summary

All right. And now please give a big welcome to the stage, Vice President, Solutions Marketing, Tim Doiron.

Thank you all for joining us at our 2026 OFC Executive Briefing. We know that your time is maybe the most important thing you have at the show, and we're really thankful for you joining us and spending some of that time with us here this morning. We'll do our best to keep this train on time and running smoothly because we know you have commitments throughout the day, and we want to get you where we're going, but we also want to share some things with you here today. I have 2 pieces of good news for you, which I think you should all be very excited about. First is it's Wednesday morning. You have made it through 1/3 of your OFC journey. Congratulations. Way to go, you should feel good about that. The second is you have passed the L.A. Convention Center room numbering test, and you have come up 2 flights of stairs to what is now called the 400 level go figure, but you've made it here, and we couldn't be prouder of you for doing that as well. Settle in for the next hour here. We have cold AC or air conditioning and hot coffee, and we're ready to go. Okay. Before covering the agenda just briefly, just a reminder here that in today's presentation, we will be making some statements about future financial performance and business progress. Those statements are actually predictions and predictions carry uncertainty and risk and enough of that said. Let's move on next. Let's cover the agenda real quick. And so what's going to happen now is David Heard, our President of Network Infrastructure, will join us. He will hand the baton to Ron Johnson, Senior Vice President and General Manager of Optical Networks. He will then pass that baton to Rob Shore, who is Head of Optical Network Marketing. David will come back to close out the presentation portion and then a collection of folks will come back on stage and take your questions and do a Q&A session as we try to keep the train running on time and fit our schedule here. With that, welcome aboard, Mr. David Heard.

Good to see everybody. What a year it's been. Last year at this time, actually 1 year ago, Nokia completed the acquisition of Infinera. I was the CEO at Infinera. So I actually wasn't presenting last year. I was actually sitting in the audience as a potential synergy. And when I look at what has happened over the last year and why I'm still here because usually I am a synergy, I'm super excited about the future. When I think about the first few decades of my career and what's happened with the growth explosion in mobility, moving from packets or circuit switching to packet switching, all the generations of technology, the cloudification of the network, the coolest thing I see this week and that's happened over the last year is optical is in the absolute spot to find a way to ensure that we can keep up with the 60,000x improvement in compute over the last 20 years when networking has only improved by about 30x. So we've got some work to do. Power is at a premium. The volumes of what is required outside in the WAN inside to the data center have moved from things that are tens of thousands and 20s of thousands to hundreds of thousands to tens of millions to -- if you've been active in the show, you're seeing many of the inside the data center component technologies driving literally to hundreds of millions of units per year. You are going to hear from everybody that they're building out capacity. Three years ago at Infinera, we saw that be the case. One of the best-known secrets or unknown secrets of the industry is that we actually put the photonically integrated circuit out there, right, 25 years ago. We had a first fab in Sunnyvale, California. Three years ago, we decided to build a second fab because we saw the impact that volume, scale and power we're going to have out on the network. So we've seen just in the beginning with the integration with Nokia, an increase in the optical business, 14%. We'll talk more about what that means going forward for the year. Overall, a 36% increase in the cloud and hyperscale business. If you look back at Nokia in 2024, it's about -- when you look at kind of the dramatic impact of that, it's moved 10 to 12x in terms of that business, and we're just now beginning to see the true opportunity in front of us. And we achieved kind of #1 market share. So I was talking to Simon last night when I joined Infinera, we were #8 in a really crappy industry. We were all fighting for $9 billion of systems business in optical systems. You've seen again, that proliferation of optical in every nook and cranny of the network and now inside the data center, it is a grand business, and we are all scrambling for capacity. So we've launched our 800-gig pluggables. Ron is going to talk about some really exciting things that I'll flip to here in terms of our new product introductions. When I was at Infinera, we would spend $300 million a year on research and development. As the President of Network Infrastructure, we are investing about $2 billion in research and development and a ton of CapEx in building out demand in this very accretive business for us. So we have expanded those fab capacities in San Jose, 3 years ago, we kicked that off. We are qualifying that fab this year. Ron can talk more about that. We also have our own advanced packaging center in Allentown, Pennsylvania, which we also have extended. We had the support of the U.S. government via the CHIPS Act to be able to make that happen as Infinera. And now we have the scale of Nokia to really take things to the next level. So again, moving from us from #8 and Nokia from kind of a #4 to a #1 position, that's something that when I was at Infinera and said, ultimately, we're going to be #1. A lot of people looked at me and said, you're crazy. Well, hey, we were #1 in Q4. I don't measure just by that. I measure it the opportunity and growth opportunity that's in front of us and the execution that's in front of us. Just so you know, as we look at optical and where it's hitting the network, both in the WAN and inside the data center, there are some really important adjacencies that also attracted me to stay for this opportunity, which is as the President of Network Infrastructure, we have an optical networks business, as I mentioned, that's #1, scaling has a pluggables business, has a component play inside the data center. We are #1 in edge routing and IP networks and have entered and had our design wins ramp up this last year inside the data center with our switching portfolio. That has a very nice play with optics. Optical and switching together are really opening up huge opportunities for the hyperscalers and also in the wide area networks. And that scale allows us to leverage our technology, not just for inside the data center, but take those advantages, especially when you operate a fab and you're building DSPs, that scale effect gives our customers in the wide area network the lowest cost per bit and the lowest power per bit. And then in fixed networks, unknown fact, we're #1 in the world, including Huawei, including Huawei in broadband access and PON. So we're seeing technologies like when you're #1 in PON in the wide area network to be able to pull those technologies inside the data center, and we're seeing the technologies that we're using inside the data center that we can leverage outside in the wide area network. And so that, again, unleashes a huge opportunity for us. I'm not going to go through this slide. Nobody needs to tell you on Wednesday of OFC that the opportunity in front of us is absolutely enormous. Again, the scale effect of what we're seeing out there. We are at the right place at the right time with the right people, the right technology and again, the right vertical integration to be able to win. When we would come to OFC with the $300 million budget, we would announce a new DSP because we were doing systems and we were doing subsystem technology. We have to do DSP maybe every 3 years. We'd announced it 1 year, and then we talk about its progress and then we talk about early trials and then we talk about deployment. Cycles have moved from 4 years to when you're dealing with hyperscalers, neoclouds and even now in some of the wide area networks, the demand for lowest cost per bit, lowest power per bit, Moore's Law to be able to hit the optical network, it's requiring less than 2-year splits, 18-month splits on DSPs in many cases. So we're announcing here at the show, what we announced earlier this week, 4 new DSPs that together are the building blocks for 13 new coherent solutions. Ron is going to dig into that. Each of these, each of these is all indexed to lower CapEx, OpEx and particularly power, performance and reliability. We've been making lasers for 25 years and what we've done. We've never had a laser fail in the field, pretty important to the web scalers that we deal with. What's changed in the year? Super excited about what I just talked about. We are loaded for bear ready to go. But what you want to hear at OFC isn't me. You want to hear Ron and the team go into the implicit details of what's coming ahead and what profound effect does it have for our client base. So Ron, why don't you walk them through the specifics of what we're going through this week.

All right. I couldn't be prouder to be up here today representing optical for what was Infinera as well as Nokia Optical. It's really a great time to be able to bring these 2 resources together, an amazing talent on both sides of that equation. And the timeliness is really great because our customers, frankly, are fragmenting, where we're seeing all the different hyperscalers have different careabouts, different reasons for optimizing in different ways to even solve the same problem. And it's not that we want to build bespoke solutions. We're doing as much as we can to optimize what we do and how we do it. When you hear about us talking about 4 engines, but 13 applications, it's really because we're taking those same building blocks and repackaging them in different ways to drive value and efficiencies in our development. But it's not as simple as saying we are building one solution for the entire market. It's building what our customers want, when they want it, how they want to consume it and for different reasons. The market is shifting. We're certainly seeing a lot more focus on pluggables, a lot more focus on lower power. This is a representation of the market today in 2025 -- today, yesterday. This is in 2030. You see about a 65% growth expected in the business that we're talking about here today. It's another business that we're investing in. We're not talking about that a lot today. This is inside the data center. You can go to our booth and see a lot of the innovation that's going on for us inside the data center, but you'll see more. You'll hear more from us as we go throughout the year and approach ECOC in September. The big growth here is in pluggable-based solutions. It's because these solutions are cannibalizing more and more of the embedded base solutions. Embedded will become really the super high-end situations where fiber is a massive constraint or in subsea. But for everything else, pluggable-based solutions will pursue these opportunities. In fact, we'll see a blurring of even what pluggable-based solutions means. We'll get higher power devices and pluggables. We'll see higher powered pluggables address certain situations like XPO. Is XPO a pluggable? Is it a line card? It's kind of a similar power level to what we would have talked about in the past about an entire line card, but maybe have the opportunity to take some of the building blocks we're going to talk about in a minute and drive that into a 12-terabit pluggable. And then line systems is really exploding. You'll see in a minute why that's the case. Line systems is exploding because the capacity is not slowing down. Capacity is growing at a breakneck pace. The ability to pack more capacity onto fiber is not changing at that same pace. So our objective and what we sit down and talk to our customers about is really, how do we take and look at the paradigm today, look at 800 gig that can go 1,000 kilometers in most networks in a pluggable. And how do we focus on technologies that will drive 10x the capacity at the same CapEx, same power, same footprint. And luckily, because I think that's already happened with pluggables, luckily, the market is growing so fast. We're seeing opportunities to grow not at 10x, but 100x and even 1,000x the capacity that's in the network today. We will still see massive growth. We'll still see massive changes in the distribution of who wins in the network, but it will happen because of these innovations and because of delivering tremendous value in terms of cost per bit per kilometer. As we approach the Shannon limit and we get closer and closer, you've heard this for multiple years now. It's taken a lot of innovation to continue to get close to that. We'll continue to invest there. It still has tremendous value to do so. It's part of our plan. But there are situations where customers now are just saying, look, I want the most operationally simple solution that I can get. If I can get a simple pluggable that goes in every router port that I have in the network, and I can throw that bandwidth onto fiber very efficiently, very quickly. In a simple way, let's do that. In other situations where they're really fiber constrained or really have some other motivations to deliver more capacity per fiber, we're going to deliver those solutions as well. The power continues to scale up in the data center. We're doing that. We're seeing this kind of scale even though tremendous innovation is going into TPUs, GPUs the engines that run optical. And we're seeing this now just for -- I don't know how to think about terawatts, but playing with my friend, Gemini, I was able to look at it today. And if we look at all the data centers in the world, today, they add up to about the power required in the U.K. And if you add up all the data centers that were projected in 2030, it's about 4% of the total power that's required on earth. So it's a big number, but it's not out of the realm of possibilities to deliver this. So I try to simplify this and keep it simple. I think of it as freeways and cars or infrastructure line systems and engines. This is where our innovation is really focused around. On the line systems, we've created massive freeways with a huge number of lanes. We've extended the width of the freeway through investing in Super-C and Super-L, and that's taken off in some meaningful places in the network, but we also optimize around C+L. On the embedded side, we've done a tremendous amount of investment here that has given us in the [indiscernible] realm, a capability to drive 800 gig anywhere in the network. Now we've moved that -- those 2 capabilities forward, and we've delivered multi-fiber optimized line systems. Today, we can fit 20 ILAs in a single rack, which is a tremendous innovation that's occurred over the last few years to enable that. We've also taken this capability and squeeze it into pluggables, right? And we keep pushing the envelope further and further. We're working with our competitors, frankly, in this space to make interoperable solutions so that we can drive 800 gig 1,000 kilometers and more in the network, and it doesn't have to be us on both ends of the circuit. So it makes it easier for our customers to realize that opportunity. So if we really look at what we announced, and I'm just going to introduce it here on this slide, and I'm going to turn it over to Rob to really dig into those 13 applications. It's really twofold. One is we took those -- the 20 ILAs in a rack, which is state-of-the-art today, and we're going to deliver later this year, the ability to multiply that by 8. So we'll get 2 ILAs in every rack unit -- or 4 -- sorry, 4 ILAs in every rack unit, 40 rack units in a rack, 160 amplifiers in a single rack. And then on the DSP side, we have lots of different motivations, lots of different optimizations. If you go from left to right, it's really an optimization of applying a DSP per distance and trying to get the power out of it and the cost out of it. We've been investing in CL technology. Our variant of CL technology is mostly focused around our strengths, which is in coherent transmission. This is where our customers are trying to connect outside the data center length of 10 kilometers, 20-kilometer type distances. We've invested a lot here, but this is a changing space, OIF, if you follow it, has made some decisions recently that we have to incorporate back into the chip. So we can talk about that in detail, but we won't cover that here today, but it's a moving target, but we've invested in the technology to be ready as our customers are ready to leverage this. Huron, these are all lakes, by the way, or oceans, big bodies of water. Huron is a technology that evolves this 800-gig capability we have today with ZR/ZR+ into 1.6 terabit. This will tape out in a few months and give us the ability to deliver product next year, which has the capability to stretch 1.6 terabit to these types of 1,000, 1,500-kilometer type distances in the network. As we were building that and defining this, we found all kinds of things that we wanted to do, really good enhancements to our ability to squeeze more capacity into a DSP and drive it further. And that really became the definition of what we call Superior. Superior has the ability to scale up to 2.4 terabits and can work in some pluggable applications, but mostly can work better if we throw a little bit more power at it and provide some kind of hybrid between pluggable and embedded applications, and Rob will talk a little bit about that. And then we have Pacific, which has frankly been a placeholder for now for us for some time where we're putting all of the innovation that we can think of -- and as we've built out our road map, we keep pulling capabilities as it makes sense to the left. And so we've done that more and more, providing -- going from what was at 400 gig very short reach distances in the network to much longer, in many cases, long-haul distances at 800 gig, extending that a little bit further now with 1.6, and we'll continue to make innovations and drive that capability as we move forward. What that enables? DWDM plugables are going routers and switches, where we announced a double-sided device. So this -- you could think of as a transponder and a pluggable. It has the client side and the trunk side built into one device. So it sits in a transport platform that looks a lot like the density of a switch. We're enabling that in the GX products, which are the Infinera products as well as in the 1830 PSS products. So as our customers look to provide the capability to do MOFN applications for hyperscalers. They can leverage this under the same umbrella as the management that exists for both Infinera and for Nokia. Same with thin transponders. These devices may be hosted in this thin transponder. These devices may also be hosted in these thin transponders. This is the GX version of it at QSFP. There's an OSFP version of it that you could see on the floor. There's also a larger high-capacity shelf for the 1830 that we're building for the same capability. So this enables an easy path for our service provider customers to get investment protection. So where they've made massive investments in manageability in our 1830 platform or our GX platform, we want to allow them to continue that investment and win that new hyperscale opportunity. We've also been working very closely with our customers to get this concept of a full-band transponder. So if we really want to focus around how do we drop the density as much as possible, how do we eliminate cabling, how do we simplify taking a transport device, just having client optics and spit out a C+L solution that can go thousands of kilometers in the network. This enables that. And our customers are very excited about this, and we're very excited about this because it's a good blend of technologies. This is air cooled, the way it's shown here. This is 4RU to get the entire C+L+ line system, all in a very dense terminal. the likes of XPO that many of you heard of and seen this week will enable this with a liquid cooled device to actually dramatically even move that density much tighter. We could even get this into -- from going from 4RU to 2RU with some space left over. And then embedded transponders, we're going to continue to invest in embedded transponders to enable that upper right part of the curve where our customers, again, have fiber constraints and they need more capacity in the fiber and are willing to dedicate more space, more power to achieve that because the fiber is the constrained asset. All right. And with that, I'm going to turn it over to Mr. Shore to go into detail about each of these.

All right. So I get the fun job of going through all the detail of everything that we're doing, why we're doing it and the benefit it's going to bring our customers. So Ron kind of gave a pretty good high-level overview of all of these different things, but we're going to dig a bit more into it so we can explain the applications, the drivers and the benefits. And we are going to start here on the left-hand side of the screen for the line systems. So we talk about the scale of the network and how rapidly things are growing. It's really very few better examples of how rapidly things are growing than this, right? So this is a kind of a typical data center interconnect network. You got 2 data centers. If it's more than 80 kilometers apart, we have to put these things in between called in-line amplifiers to boost the optical signal to get it to where it needs to go. Historically, how much fiber did people put in between those locations? And we've seen a number of announcements from different network operators. Lumen and Zayo made public announcements about this. A number of other people talk about it. You've seen a lot of it on the show floor. Historically, people would put maybe 100 or 200 fibers between locations, and that was to cover the amount of capacity necessary. But as Ron pointed out, bandwidth is growing like crazy, but fiber capacity isn't, right, because of that little annoying thing called the Shannon limit. So the only real recourse our customers have and network operators have is to add more fibers. And this is the kind of increase in fiber capacity people are adding. We're going from hundreds of fibers to thousands of fibers. In fact, I just saw a presentation yesterday where Zayo was talking about 13,000 fibers between locations. And this is the kind of capacity that's required to meet these bandwidth demands. So last year, we talked about working on the endpoints. What do we do inside the data center to simplify the aggregation, deployment and demultiplexing of all those optical signals and we rolled that out last year, and we're now shipping that solution in volume at the endpoints. This year, we're focused on these intermediate locations, these intermediate amplifier locations. And again, you have these locations that look like this. That's basically a double-wide shipping container that they ship out to a field somewhere and run all the fibers between it and sometimes put this pretty rock face on it to make it blend seamlessly into its environment. But this is the idea. You got to get now from 200 fibers that you need to amplify to 7,000 fibers that need to be amplified. This is all about density. How can I create technical solutions that can enable us to enable our customers to amplify more fibers in this physical footprint, so they don't have to modify and upgrade their infrastructure. They can leverage what they already have. And this is what we've done. Today, this is the highest density solution in the industry right now, okay? This is 4 fiber pairs or 4 ILAs, 8 rack units. What we're doing is taking this and condensing it into a single rack unit for 4 in-line amplifiers, 8 bidirectional fibers in a single rack unit. This will be the most dense solution in the industry. You've probably heard some of our competitors out there talking about their version of this solution. They're getting 128 ILAs per rack. We're getting 160 ILAs per rack. So this is the current highest density solution. When this comes out at the end of the year, this will be the highest density solution, about 25% more dense than any other solution that will exist in that time frame. And again, density in this application is the name of the game. Whoever is going to enable a network operator to provide more services over that existing infrastructure is going to have a substantial advantage and enable network operators to do more with their existing infrastructure. One of the other real benefits of the way we've architected this is this fits seamlessly into our existing platform. okay? The GX is an extremely well-architected platform that really enables you to integrate virtually any technology into it. So you can do line systems, transponders, pluggables, thin transponders. And now this multi-rail line system, this multi-rail ILA just seamlessly fits into the existing platforms our customers are already deploying. So they don't need to onboard a new platform. They can just leverage what they've already done. And that's been a really big fundamental principle for Nokia for many years now is this idea of these modular-based platforms that is very easy to configure and upgrade, okay? Let's move on to optical engines because this is a really more interesting story, more involved story to tell. And we'll start with here. Ron kind of talked about, we've got a whole host of different distances that need to be addressed in the market, everything from these short-reach campus, metro regional long haul all the way to subsea for multiple thousands of kilometers, okay? That market right now, again, if you take out the line system and you just look at the engines, that's about a $9.3 billion market today, okay? And how are we addressing that market today? We're addressing it with these 2 optical engines, a pluggable and an embedded. And we use these 2 engines, everybody, this is an industry. We use these 2 engines to address every application across the entire spectrum. And even this is relatively new because historically, as Ron mentioned, we've been focused for 30-plus years in the industry on building engines specifically focused on maximizing capacity per fiber. That has been the focus of our innovation. This generation is the first generation where we've really shifted, right? Customers recognize they can't do everything they need to do on a single fiber pair. They have to go to multi-fiber networks. They're deploying 7,000 fibers. Now they can shift their focus and they want more cost base and power-optimized solutions. And that's what caused the industry to introduce this engine here, this 800 gig pluggable. And it was really the first foray for the industry into this shift of focus away from maximizing fiber capacity to cost base and power optimized solutions. So what's happening now in the industry? Well, a couple of key things are happening, right? Number one, that top line, you can see the market is growing really, really rapidly from $9.3 billion to $15.4 billion in just a few years. So the market is growing very, very rapidly. But at the same time, the applications are diversifying, right? We're going from just these current engines, right, these kind of space cost and power optimized engines that, by the way, are designed for a very specific power envelope. These engines, the pluggables we have today are specifically designed to go into routers. When you put a pluggable in a router, they got about a 30-watt power envelope. It has to be at 30 watts or less to operate inside a router. So we designed the engine specifically to do that, so you could plug it into a router, okay? One of the things we noticed as we go to the right here is those engines, even the 800 gig engine we have today, still has a lot of gas in the tank, you might say, right? We're only operating it at like maybe 50% or 60% of its actual capacity because we got to keep the engine cooler to fit into a router. If you actually push that all the way to the red line, these engines have a lot more capabilities than we are actually using them for today. So what we're seeing in this next generation is a whole new line of these same types of pluggable engines, these space, cost and power optimized engines, but operating at their peak capacity, maybe in more like the 60-ish watt range that maybe won't go into a router, but can be used in these thin transponder solutions. We'll talk a little bit more about that as we get into it. But now you can really get a lot more out of these same types of engines that are still, as I said, cost, space and power optimized, okay? On the other side, by the way, you have a bunch of these shorter distance solutions, these maybe campus or 80-kilometer type distances. A lot of these are addressed by direct detect optics today. So very cheap, very simple, very reasonably low performance optics. They were able to address these distances up to about 800 gigabits per second. As we go beyond that into 1.6T or 3.2T, they can't reach that distance anymore. Those direct detect optics are not capable of closing those spans. So this is opening up a whole new market for low-power optimized coherent engines that are optimized around that 80-ish kilometer range. A lot of people call that coherent light, okay? But that's a whole new market for coherent optical engines, right, and something that needs to be specifically addressed. And as Ron said, there will always be this right-hand -- left-hand side here, I guess, your right-hand side, my left-hand side. For people that are fiber constrained, if you only have a couple of fibers, then you will pay extra, you will give more power to solutions that maximize your capacity per fiber. So there will always be those, and you need to, of course, address those applications and have solutions optimized for those as well. But on top of these diversification of application, we're also seeing a much wider variety of network deployment models. All these different types of implementations that need to be addressed. As Ron said, how customers want to consume this is also increasing in variety. So whether it's these new kind of CPO switches or regular switches or routers, thin transponders, full fiber solutions, embedded transponders. And it's not just these solutions, but they want all kinds of different pairs, right? They want short-reach pluggables, short-reach thin transponders, short reach full fiber, long reach, high performance. So they want all these different combinations of solutions, both the type of performance from the engine and the implementation model. And so what we really realized is you cannot address this market effectively by just doing another version of the engines we have today. This is not going to get it done. Customers are increasingly wanting applications or solutions optimized for their specific applications. This might be a great solution like are the 800 gig plugs we have today are the absolute best solution you can find for roughly 1,000 kilometers. If your application is 1,000 kilometers, that is the best you can do. If you want to go 200 kilometers, well, now it's overkill, right? You could have a more optimized solution that would save you money. And this is exactly what we're talking about doing, right? Essentially, instead of just 2 solutions, 2 new engines, a new pluggable and a new embedded, it's really coming up with a different methodology for building these optical solutions based on a building block principle, essentially taking the technologies and building them in these very standardized implementation methodologies, standardized control languages, standardized data paths so that we can take those engines now, these building blocks and integrate them into a whole variety of network implementations. Now you can get a mix and match of combinations of solutions so that customers can have solutions that are optimized for their very specific applications, and they're not leaving a single penny or a single watt on the floor. And this is the concept. And by the way, this is not new. Infinera/Nokia, we did this about 10 years ago for the optical systems market. Historically, optical systems were these big giant monolithic custom design solutions. And about 10 years ago, we shifted to something called compact modular platforms. where we took those big monolithic solutions, and we broke them down into small discrete technology blocks called Sleds, and you could then configure those systems however you want to give you an optimized system for your application. and you could then configure those systems however you want to give you an optimized system for your application. What we're doing now is the same architects that came up with that concept for systems have now developed that same concept for optical engines, enabling us to have building blocks and a much wider variety of solution implementations that are customized for our customers' applications, okay? So let's talk about the building blocks. And Ron already talked a little bit about it, but we'll jump through it a little bit more. Four different building blocks, all designed for different distance applications. You've got Huron -- or Ontario, excuse me, which is short reach, low power, you're looking at 18-ish watts, 16 to 18 watts is what you're aiming for there. So when you have these short campus style applications, right, this is going to give you the optimal solution with the lowest power and the lowest cost. Moving one over Huron. This is really the spiritual successor to the current generation of 800-gig pluggables. It's the 1.6T variety. These are all based, by the way, on 2-nanometer-based technology. So we're going up one generation in node process, okay? But this is the spiritual successor to the 800 gig, just a generational upgrade. It's still an IP over DWDM optimized solution. So it's a device that can be plugged directly into a router and fit within that power envelope, but maximize the performance and give you that generational upgrade from the 800 gig today. Next, you have that Superior, and this is that engine that I was talking about. It takes the same technologies really that we're using in Huron, which really is running at about 50% maybe to fit into that power envelope and just stepping on the gas, throw in a couple of extra chromatic dispersion gates, maybe a few other things to enable you to go longer distances. And now you have the real true potential capability of the space cost and power optimized engines. Now you're talking about a device that can be put into a pluggable that can run at 2.4 terabits and take that over 1,000 kilometers. You have now a pluggable-based solution or a solution that can be integrated into a pluggable that can support virtually every terrestrial application and virtually in most subsea applications, okay? And we'll take a look at some of the specific details on the performance of that when it's integrated into a pluggable, okay? And last, of course, but not least, is Pacific, you will still have, right, these high-power solutions that really are optimized to maximize spectral efficiency, right, to put as much capacity onto a fiber even at the cost of cost, space and power, right? But for people that only have a limited number of fibers, they're willing to pay the extra for that, right? You can get 10%, 15% extra capacity per fiber. That's meaningful if you only have one fiber. Less meaningful if you have 7,000 fibers. But if you only have one fiber, that's an important value. Now these are the building blocks. The really interesting part of this implementation approach is our ability to integrate those into all of these different varieties of solution implementations. And we're going to go at each one of these, right? But you've got IP over DWDM, those double-sided plugables, which we'll take a look at, thin transponders, full band and embedded. Now these are the 3 we've been doing, right, pluggables that go into routers, pluggables that go into thin transponders and embedded. The 2 completely new solutions are right, these double-sided pluggables, never been done before in the industry. Full-band transponders, never been done before in the industry. And of course, by the way, each of these implementations can be a host for any of these building blocks. So even with the full band transponders, I can do a long reach, medium reach or short-reach version. So if you have a short-reach application and you want to use that solution, you're not wasting any space power or money with a more powerful engine than necessary, right? That's the idea, right? So digging into kind of the traditional ones, we're going to cover them both in one slide. These are basically just generational upgrades. This is what the industry has been doing for 30 years. If we didn't do this building block approach, this is all we would be announcing. It's probably what most of our competitors are going to announce. It's just these 2 solutions. And there's benefit to them, right? This is right, your traditional IP over DWDM, just the successor to 800 gig. This is your embedded, so your successor to i7 or PSE-6, right? And they provide value. They give you about 30%. This is what you normally get for a generational upgrade. You get about a 30% improvement in TCO and that space cost, power, fiber efficiency and operations, right? That is your normal benefit. And it's still good, and there are these applications, and we want to get those benefits. But if you have one of these other applications, you want something more optimized for that. And our customers are increasingly demanding that and selecting vendors based on who can provide them a solution most optimized for their application. So let's take a look at the double-sided pluggable. This is a fun one. One of the scenarios we're having to deal with in the industry now is these switches and routers that have integrated optics, whether it's CPO or LPO or LRO or XPO or NPO or any of the other POs out there, they have optics in them. Coming out of the fibers with optics already. I can't put a coherent pluggable into it because the optics are already integrated, and this is gray light. It's direct detect. I need to somehow convert that light. This is, by the way, great. If all I'm doing is 100 meters inside the data center, boom, I just connect those 2 things together, I'm good to go. But if I want to take this anywhere else outside the data center, 10, 20, 30 kilometers or even 7,000 kilometers, I need to convert this white light into coherent. How do I do that? The way I would do that today, this is the most efficient way to do that today. I buy a thin transponder. Of course, I got a chassis, a thin transponder. I have a client pluggable that connects to the switch, then it takes the signal across that thin transponder to the coherent transceiver. It converts to coherent light and off I go. That's how I would do it today. This is the most efficient way to make that happen today. What we're saying is we can optimize that solution. By the way, what are these optics right here? These are potentially 1.6 terabit direct detect optics. For us, that's ICE-D. We already have that solution. We make that solution. We can do now is take these 2 devices and integrate them into a single pluggable. Instead of an electrical interface on the back, it's now an optical interface. That's why we call it double-sided, optical in/optical out. Gray optics or white optics coming in, single DSP pointed this way and that way, coherent going out the other way. Now instead of a thin client optics, thin transponder, coherent optics, I have just a single pluggable. This will save you about 70% TCO. And again, space cost and power. instead of a big thin transponder, 2 pluggables, all the operational complexities, literally a pluggable, I plug it in, I'm done, right? And this is what I mean about application-optimized solutions. Could I solve this with our current generation of solutions? Yes. It's just unoptimized. So you're leaving space cost, power and money on the table if you use a solution that is not optimized for this application. And that's the first example. Second example, what we're doing with these thin transponders? Now of course, we do this today, thin transponders, and we use it essentially with the 800 gig, and we're certainly going to have a Huron implementation of that, which will give you pretty much the same kind of performance you're getting out of the 800 gig plugs today just at 1.6T. But I do not have the same power restrictions on this thin transponder that I do in a router. These can go up to 60, 65 watts. I can take the harness off that engine and push the gas all the way to the floor, and that's where you get this kind of performance, building a superior base solution into a pluggable that can now enable you to have this type of performance, right? If I compare it to what we're doing with the embedded engines today. So this is i7, PSE-6, WaveLogic 6, you name it. This is what you're getting today with an embedded engine, which is the big giant Sleds, okay? This is a pluggable, and it's giving you anywhere from 50% to 100% more capacity, so 2.4 terabits. It's taking the peak line rate almost 50% further, almost twice as far. So these take their peak rates about 600 kilometers. This will take 2.4T, 1,000 kilometers. Of course, it has PCS, so you can tune it down and take it really virtually any distance, 7,000, 8,000, 9,000 kilometers, right, at lower bit rates, just like this current generation, okay? But even at the peak rate, it has better performance, dramatically lower power per bit. This is a 73% savings in power per bit, better performance, better reach, and, oh by the way, better spectral efficiency. It's about a 13% improvement in spectral efficiency by tweaking the forward error correction and other signal transmission protocols, there we go. So you're getting higher rate, higher reach, higher performance and 73% less power in a pluggable form factor. And this is what I mean about when we shift our focus from just building bigger and fatter engines to building engines that are space cost and power optimized and optimized for these specific applications. And by the way, there's a lot of good reasons to use a thin transponder, right? Certainly, one of them is you're a wholesaler and you don't own the router. You can't put pluggables in somebody else's router and you're just trying to transmit that service. It provides domain separation, alarm correlation, lawful intercept if you need it. And also, of course, operational domain separation. You can do things like client splitting and client aggregation. So there's a lot of good benefits for thin transponders. Even if you can't -- even if you could put a pluggable in the router, there's still some benefits. So this is going to be a very popular solution. It's already a popular solution today, thin transponders. And now we're actually getting the full capabilities out of these thin transponders. And so that's the idea. And this, again, when I compare this 2.4 terabit pluggable to today's embedded engines, it's a 60% TCO savings. This has never happened in the industry before, where one generation to the next, you're getting a 60% reduction in TCO, right? This is really monumental. And there's actually -- it's an interesting thing we just looked up the other day called Jevons Paradox. Anybody heard Jevons Paradox? All right. Yes, thank you, Tim in the back, who looked it up with me the other day. There's this concept, right, that, oh my God, you're cutting cost and everything by 50%, isn't that going to cater the market like this thing you were selling for $100, you have to sell for $50. The answer is that's the paradox. Lowering the price actually increases the market because people find new applications for these solutions. This is pluggables, right? The 800 gig pluggable is about 50% lower cost per bit than the embedded today, okay? What happened? It created the entire scale across market that didn't exist and wouldn't exist if we weren't able to build pluggables in that form factor and cost profile. And this is the idea here. This is enabling customers to achieve their objectives and enabling new types of applications. Okay, Thin transponder. Last cool new solution, the full-band transponder, okay? This will come as no surprise to anybody, but we have customers today that aren't just lighting up full fiber day 1. They're lighting as many as 10 full fibers day 1, one purchase order, one installation effort, all at once, 10 full fibers worth of capacity. This is what you'd have to order today. This is just one fiber, 64 coherent pluggables, 64 client pluggables, 22 thin transponders, buy all of that inventory, receive it, install it, cable it all up. For 10 fibers, that's more than 1,000 components, more than 1,000 fibers that need to be connected, okay? Not only do customers saying, hey, this is way too much cost base in power, it's just way too operationally intensive. And of course, what's the #1 cause of failures in a network? Humans, right? So the less amount you can have people touching it, right? That's good for a little while until robots take all of our jobs. But until then, it's still humans. So they're saying, hey, I need a much more efficient way to do these types of solutions. When I have these kind of hyperscale demands, how do I more efficiently address it? And this is what we're doing. We're building this full band appliance. It takes not only the coherent optics and the client optics, but if you'll see in that animation, which we can all watch again because it's really fun. It integrates the DWDM line side integration as well, all into a single appliance, okay? So essentially now, instead of over 1,000 components to light up 10 fibers, I can do it with about a dozen components. And instead of 1,000 fibers, it's literally just one fiber because I've got no -- all of the connections between the coherent optics and the aggregation, all integrated onto the Sled. There are no fiber connections. It's all integrated into the Sled and chassis. And all I have coming out of this is a single fiber that's fully loaded with all my traffic. And by the way, yes, this is great for hyperscalers that have these massive bandwidth demands, but it could be good for service providers, too. We had a little thing at Infinera years ago, we called Instant Bandwidth, where we predeploy the bandwidth, but they only pay for what they actually use. That way, they can turn it up much faster, be much more responsive to their customers. They can use this solution for that same purpose. Turn up the whole fiber, pay for what you need as you need it or you got 10 fibers worth of capacity, you could turn it up in a fraction of the time, a fraction of the cost base in power and with many fewer errors due to humans misconnecting fibers or not cleaning the fiber cable properly. There's about a 70% TCO over that thin transponder implementation. And again, this is what I mean. I could use thin transponders to address that today, but it's not optimized for this application. You need an optimized solution. And with the building block approach, not only can I build these full band appliances for regular distances, call it, regular 1,000 kilometers, I can build it with Superior to enable you to do this over subsea applications or I can build it with an Ontario to enable you to do campus or scale across type applications. So again, this is the beauty of the building block. It's not only being able to introduce these new types of solutions, it's being able to integrate any of these building blocks into those different network implementations, right? And that's why really what we're focused on is not just a few new engines, but it's a new paradigm for building optical -- coherent optical solutions that enable us to address our customers' specific applications with the most optimized solutions. This is not only good for them because it will enable them to do things that are more ambitious than what current technology enables them to do, but it's good for us because if we're going into a competitive environment with somebody trying to use a 1 or 2-size-fits-all solution and we have an optimized solution, we will be 30% to 40% lower TCO for that customer for that application, okay? So this is the summary of the solutions. Next question, and Ron kind of already let the cat out of the bag on this one, but when are they available? We're going to be sampling these in the summer of next year. So when the first samples come out with the first of these solutions, which is going to be the Huron-based pluggable will be available at the end of next year. So now you might be asking, hey, Rob, why are you announcing this so far in advance? Well, there's actually a really good reason for that, which is we are building these solutions to help our customers plan their network strategies. Where are they going to put the next data center? How much power can they put in it? How much capacity can they do? What are the distances between them? They need to know that because they need to know that now for what's going to happen 2 or 3 years from now, so they can make those plans appropriately because it takes 2 or 3 years to build the data center. And when they turn it up, they want to make sure they're using the latest technology, and they've architected their infrastructure to best leverage what's available at the time. So -- and by the way, this probably goes without saying, but we don't do any of these things in a vacuum. This isn't just a bunch of people, engineers in the closet dreaming up a bunch of cool solutions. We work very closely with all of our customers to understand what their applications are. and develop solutions to specifically address them. And a lot of them are hyperscale, not exclusively, but a lot of them are hyperscale. So each of these applications are real. And the volumes are just massive, right? They are demanding customized solutions for those applications, okay? So this is the idea. We will be around after for questions, and I'm going to now hand it back to David to close this up, who's getting ready to run up on stage since I've got a few minutes over.

All good. By the way, Rob had coffee from the left carafe. So I suggest you stay away from that. I thought I was bad. Hey, Rob, thank you very much. Hey, when you take Bell Labs and you take scale and you take the largest coherent DSP team on the planet, when you take having 2 fabs in indium phosphide plus a team, if you remember, Nokia had bought Elenion and Silicon Photonics. And you put that all together and then you really listen to your clients, this environment is perfect for that. because especially hyperscalers and neocloud players and even now out in the wide area network with service providers and mission-critical enterprise, they are buying road map because they are making plans, they're buying data centers. They're buying facilities. They're planning HVAC right now. And the reason we are committed to this road map is they have helped us shape what is in each of these products. And you know what, that minimizes loss, less gates in a DSP. It says less lines of code in the software that we write to stitch this all together with, again, automation that goes across, moving into domain controllers and truly event-based automation. At the same time, we're able to plug this into the rest of our network infrastructure platform. So when we talk these scale across, we talked a little bit about this, Simon, last night, the scale across architectures, we're rolling out half a petabit switches in scale across with 576 suite 800 optical plugs in it. That is really rich for us in terms of business growth as we go forward. And so the opportunity has never been more important. It's about pulling together the right technology, the right people, the right interface with our client base and having the scale. So I couldn't be happier to be here a year from a year ago because we've got those things, and we are leaning forward in the industry. So we're going to kind of come up here. Sorry, we're running a little bit over, but we've got time for Q&A.

We have a couple of different folks joining us here. David Mulholland from our Investor Relations is here and on the far end. Julia Larikova from our Vice President, Product Line Management and a lot of the leadership behind the work that we're doing here. We have some amazing microphone sherpas that have mics, and we're ready to take your questions. Just raise your hand, and they will run by and hand you a microphone and give you an ability to ask really anything you want from this esteemed panel. Anybody, don't be shy. There he is. I knew Ryan would do it go right ahead.

Thanks, Ryan Koontz with Needham. What's the plan for the new fab? I mean, how are you using it now? What's the vision for the future? What kind of competitive advantage does that give you? Maybe walk through that a little bit.

Ron, do you want to start and then maybe, David?

Sure. Yes. I think I know the new fab is finally flowing gas, flowing chemicals. So it's -- all the equipment is being turned up. We're qualifying it this year. It will start production late this year, early next year. Our focus is to ramp that pluggable that is in such high demand, meet our customers' expectations to deliver these scale-out applications or scale across -- sorry, across scale across applications. And then there's a lot more capacity in that fab, which we're building multiple different things inside the data center. indium phosphide is in is a coveted thing right now to have capacity in indium phosphide. So we're being careful about how we're committing that, and we're working with our customers to do that in a meaningful way.

Is there anything we're doing with wafer size or anything with that?

Yes. No, it's 16 -- or sorry, 6-inch fab that...

That out there.

We're going to 16.

That's the latest -- you heard it here.

Like a laser disk.

Yes, 6-inch, which is 4x the capacity of a 3-inch wafer like we have today. That, along with the advancement in tools and the more tools gives us like 20x the capacity to build really, really complex things. As we build simpler things, like, for instance, lasers or inside the data center applications, it could drive many, many, many more wafers to address some critical needs in the network today.

In addition, we have the packaging in Pennsylvania, and we have expanded that capability as well, all driving both the embedded and pluggable growth that you see with these DSPs coming forward. And we're leveraging, again, the largest DSP team, coherent DSP team on the planet to make that happen. And that was one of the advantages of the acquisition is being able to split that team up to be able to drive that kind of building block innovation. Even within the DSPs, there's 9 to 10 building blocks that we're able to kind of reuse generation to generation.

Yes. Daryl Inniss from LightCounting. You basically made a promise today. You're talking about something that's going to happen in 2027 or later. It strikes me that this is new to the industry. I don't think that we have made these kinds of promises. I'm worried about your ability to deliver or examples of track record where you're able to hit this. And so anything that you can share with us would be helpful.