Good evening, everybody. It's so great to see all of you here, so many familiar faces. For those of you who don't know me, I'm Sanjai Parthasarathi. I'm the Chief Marketing Officer of Coherent. We are so excited that all of you could join us here in person, and those of you who are joining us by webcast for our launch of our PhotonLink platform. PhotonLink is our integrated optics platform, purpose-built for AI data center connectivity. Before we get started, I'll have to refer you to slide two of our presentation, which contains our forward-looking statements and disclosures. The presentation in its entirety will be available in the Investor Relations section of our website after the event. Moving on to an agenda. We've got a great program for you today. Jim Anderson, our CEO, will start us off by introducing PhotonLink. Jim will be followed by Beck Mason, EVP of Semiconductor Devices, who will present the generate section of PhotonLink. Beck will be followed by Julie Eng, our EVP of Optical Components and our CTO, who will present the shape, guide, and receive sections of the PhotonLink. Julie will also present our integrated solutions. Without further ado, please join me in welcoming Jim Anderson, our CEO, to the stage. All right. Thank you, Sanjai, and thank you everybody for being with us here today. We've been really excited about the launch of PhotonLink. We've been looking forward to this for a long time. Thanks. We're here in Spain at ECOC. Thanks to everybody that's joining us on the webcast, but thank you especially for everybody that's joining us here in this room because there is a beautiful sunny beach about 200 m away from this location. The fact that you chose to sit in this conference room instead of on a sunny beach, I appreciate it. There are a few people that walked in with flip-flops and shorts. Again, thanks for being with us here today. We're really excited to share this with you. Just to kick it off, what is the motivation? Why are we launching PhotonLink? The motivation or the reason for us to launch PhotonLink is very simple. It's to try to make it as easy as possible for our customers to move from electrical connections to photonic connections in the data center, to make that transition seamless and easy to bring the complete solution to our customers. If you look at any particular distance of links in data centers, if you look at long distances connecting two data centers, if you look at the very short distances connecting maybe two chips that are sitting right next to each other on a board, regardless of the distance, as you increase the data rate, as you increase the amount of bandwidth you're trying to send through that connection, if it's an electrical connection, you'll eventually hit the limit to that connection. You will want to do is you will have to convert to a photonic link in order to achieve the data rate and the data transmission goals that you are trying to achieve. That is what PhotonLink is all about, is making that transition from electrical to photonic link as easy as possible. If you look at the data center today, actually a lot of the data center is already completely optical photonic. The telecom network, the scale-across, the scale-out networks, all of those originally started out as electrical networks but were long ago converted to photonic optical networks. Over the coming years, we will start to convert the only remaining electrical portions of the network. The scale-up network and even the chip-to-chip links will become photonic over the coming years. That is what we are really excited about, helping our customers make that transition. Now, for Coherent, that is also a big opportunity for us. When we look out at the end of this decade, at 2030, if we look at our addressable market, it is about $60 billion for the existing portfolio products that we have. What PhotonLink does is it allows us to access another $30 billion of integrated optics market opportunity by the end of this decade. So another reason we are really excited about this technology. Now, as we have engaged with customers across integrated optics over the past months, actually, the level of engagement and the intensity of engagement, the number of engagements with customers over just the last six to 12 months has ramped up really significantly. We have got a number of different engagements across many different architectures. It is very clear to us that there is not just a one size fits all, one particular architecture that will be implemented. We are going to see a wide range of different architectures that our customers are going to implement as they implement integrated optics. Some customers will adopt co-packaged optics right away. Some customers will first adopt near-packaged optics and then over time, move to CPO later on. We are already engaged with a number of customers talking about photonic links between the individual chips themselves. So a lot of different architectural options that we are working on with our customers. Regardless of those architectural options, one of the things that our customers are struggling with is if you look at the amount of optical technology that needs to be pulled together, there is a significant amount of complexity there. That is a combination of laser sources, the fiber optic cable that has to be pulled together, the complex optical assemblies, passive optics. A lot of technology has to be stitched together. A number of our customers came to us and asked us, "Hey, rather than just providing components or a subset of components, can you bring the complete optical solution, the complete assembly, to make it easier for us to transition to optical?" That is exactly the purpose of PhotonLink, is to bring a single integrated platform that is architecture agnostic, that works across CPO, NPO, chip-to-chip as well, but brings the whole solution to our customers. When we look at PhotonLink, it spans all the way from the beginning, where we generate the photon, through the shaping and guiding of that signal, to the detection of the signal on the other end. So that is the complete PhotonLink platform. Now, we believe that Coherent, more than any other company in the world, is best positioned to help our customers make this transition for a couple key reasons. First of all, because there's no other company in the world that has that complete set of photonic optical technology. If you look across the generate, shape, guide, detect, and look at all the different optical connections, all the different optical technology that's required to make that happen, it's really only Coherent that's got that full portfolio that we can bring together, stitch together in a single solution. There's other vendors that provide the laser source. There's other vendors that provide maybe the fiber optic cable, but only Coherent has all that technology in-house, and that helps our customers innovate faster and get those optical connections to market much quicker. That's number one is the complete technology portfolio, but the second reason is the ability to manufacture that at tremendous scale, the scale that's required for data center. If you look at complex optical components, complex optical assemblies, over the past years, we've delivered hundreds of millions of components and optical assemblies. Not just the full portfolio of technology, but the ability to ramp that quickly and ramp that at scale for our customers. An example of the complete solution that we can bring, not just the component ingredient technologies, is here I'm showing just an example of, this could be a CPO or an NPO example for a switch chip or a compute chip, using an external laser source. In this example, we'd be bringing the complete external laser source, the complete optical assembly that goes from that ELS to the switch chip or the processor chip that it's supporting, and then the complete optical assembly that goes back out to the faceplate to provide the I/O, the photonic I/O, to the rest of the system. That complete set of solutions and all the underlying components that go into that, and we can manufacture that at scale. We have anchor customers already secured for both CPO as well as NPO architectures. We'll talk a little bit more about that. To give you an idea of the value of the content that we can bring, if we're using the full PhotonLink solution, if we take, for instance, a switch chip or a processor chip that, let's say, needs 100 Tb of I/O capacity, and let's say we're using 200 Gbps optical lanes, the content for Coherent would be up to about $15,000 per system or per 100 Tbps chip, in the case of where we're bringing all the content for the customer. A significant amount of content that we're delivering for our customers. We'll talk about today, of course, all the different ingredient technologies. I've got our two experts, Julie and Beck, to talk through all the ingredient technologies that are underneath PhotonLink. I think more importantly, what we'll also do is we'll talk about a number of specific examples that we're working on with customers, different architectures. We'll show a CPO example, a couple different NPO examples, and a chip-to-chip example, too, of the solutions that we're bringing to our customers. Across these different architectural choices, a number of different types of light sources, whether it's CW lasers or VCSEL arrays, et cetera. As I mentioned earlier, the customer engagement across these different architectures has ramped up very significantly, especially over the last six months. We now have over 10 unique customer engagements on CPO, over 10 on NPO, and over five different engagements on chip-to-chip connections. A lot of activity across the customer base. I talked about how much content we're delivering in the case of when we bring the full solution. To give you an idea of the production timing, for CPO, our first production will start to ramp in our December quarter, this Q4 of this calendar year. That initial CPO will ramp for scale-out applications. About a year later, in the second half of 2027, we'll start ramping CPO for scale-up applications. NPO similar, we expect that to start to ramp in the second half of next calendar year, again, primarily for scale-up applications. Chip-to-chip is a little further out there. We expect that to be further out in the 2029, 2030 timeline, but definitely a lot of intensity with our customers on designs around chip-to-chip photonic connections as well. A lot of customer activity going on. I'm going to pass it next to our two experts, Beck and Julie. What they're going to talk about is, think about it as two parts of their presentation. Number one is they're going to walk you through some of the specific technologies underneath PhotonLink that are really some of the key technologies that we stitch together into the full solution. Julie, at the end, will talk about three different architectural examples, CPO, NPO, and chip-to-chip, and give you specific applications and what this full solution is that we're bringing to our customers. Thanks again for being with us here today, and I'm going to pass it off to expert number one, Dr. Beck Mason. Thank you, Jim. I appreciate it. I'm going to present to you today on our generate technology. It generates maybe the core or the heart of the PhotonLink from a photon generation standpoint. What I'm going to cover are the four main technology platforms within PhotonLink that generate the photons that serve all of that link capability. These are our ultra-high-power lasers, our high-power CW lasers, our VCSEL arrays, and our high-density VCSEL arrays. First, I'm going to start with the ultra-high-power CW laser array. This is one of the areas where we have really strong traction from a number of customers. It's an exciting platform. In this platform, the lasers are actually remote from the transmission, and they're put inside an external laser source. That external laser source contains up to eight ultra-high-power CW lasers. Each ultra-high-power CW laser can typically feed multiple links, usually around four. This provides the photons that go into the silicon photonic modulators that are co-packaged with the XPU or switch chip, or near package with that switch chip or XPU in order to provide the source for the link. The lasers are designed specifically to produce very low noise, high output power, and very high efficiency to enable the link. We are very excited about this platform because we have two long-term agreements signed with leading hyperscale AI data center customers for both CPO and NPO solutions. Our ultra-high-power laser is designed to deliver industry-leading performance. It is a very unique design. It enables very high output power at very high efficiency and very low noise performance. Why that matters, high output power lets a single laser drive multiple links with that solution, so you can split it and share it. That helps you drive cost and efficiency. High power conversion efficiency is important. These are producing very high optical powers. High power efficiency lowers the overall power dissipation in the system. That is critical in AI data centers where power is key. Finally, very low phase noise and very low amplitude noise from these lasers delivers superior link performance with better bit error rate performance. All of these lasers today exceed all of our customer requirements. We are starting to ramp these in the fourth quarter of this calendar year, and we are seeing significant volume throughout next year. These are ramping on our 6 in indium phosphide platform that we have in our Sherman, Texas fab, where we are seeing very good yield and very great performance from the lasers. The next laser I am going to talk about is our high-power CW laser for NPO applications. In some architectures, the NPO solution, or near- package optical solution, has the laser embedded within it, co-packaged with the silicon photonic modulator. That gives you greater flexibility and a different approach to the architecture that can be valuable in certain situations. The smaller size and lower power dissipation and high efficiency of our high power buried heterostructure CW lasers is ideal for this application. This is a laser that we are already shipping millions of units a month in because it is used widely in our pluggable data center transceiver solutions as well. It is a laser with proven field reliability and proven link performance, and it is optimized for both low power dissipation and very high efficiency. It is production-ready now. As I said, we are already shipping it. We have got a long history in indium phosphide, and we have shipped over 300 million indium phosphide lasers to the field so far, and the pace of that is only ramping up and continuing to increase. The next type of solution for Generate is our VCSEL arrays. VCSEL arrays for NPO come in two different flavors. The reason why VCSELs are very attractive for this application is they enable the lowest energy per bit for NPO applications. They also allow us to get to very high channel counts in a very dense solution, giving us the lowest cost and the best bandwidth density of any application that we can use for integrated photonic links. The massively parallel optical interconnectivity that we can achieve here is really ideal for NPO applications where we need high density. I am going to talk about two types, our Type 1 and our Type 2, the Type 1 being conventional and Type 2 being high density. Type 1 is designed for use with ribbon fiber assemblies, and it is similar to what we have been selling for many, many years. We have been supplying VCSEL arrays and PD arrays for more than 10 years. In fact, we are the world's largest supplier of datacom VCSELs today. Those PD arrays and VCSEL arrays are used in pluggable transceivers. These new arrays are high-density arrays. That means they are two-dimensional. They have more VCSEL channel count. They are also available in flip-chip versions that allows us to mount them directly on top of driver ICs. The photodiodes are also available in flip-chip versions, so they can be mounted on TIAs. Because they are flip-chip, we can put lenses directly on the backside of the devices. Those integrated lenses allow much more relaxed coupling, higher coupling efficiency, and better overall link performance. These solutions will be production-ready in the first half of next calendar year. The other area is our high-density 2D VCSELs, and these are actually some of the most exciting and interesting solutions, and we are actually demonstrating this product today in our booth on the ECOC floor. I almost said OFC. These are super compact solutions for NPO, and they have very, very high density arrays, and the VCSELs are so tightly spaced here that we are able to couple all of these VCSELs into a single fiber, which is a multi-core fiber. We make fibers with up to 37 separate cores, and we have a VCSEL array and a PD array with 37 elements. We can flip-chip those. We have the backside lenses, and with that solution, we can get very high capacity in a single fiber. It is the most dense solution we have out there. This is suitable for both PAM-4 and NRZ applications. It does slow and wide, or I like to say medium frequency and wide as well as fairly fast and narrow. These will be in production in the second half of calendar 2027. Just to complement this, I am going to speak very briefly about our photodetectors. It is important because to date, we have shipped over 1 billion photodetectors to the field, both short-wave gallium arsenide and long-wave indium phosphide. Our latest generation of photodiodes goes all the way up to 200 Gbps today, and we have 400 Gbps devices in development. These devices are capable of being flip chip directly onto TIAs, for example, with backside integrated lenses to allow better coupling efficiency, and they are an exciting solution for us for the other part of the CPO link. All the devices that we make today come from our internal wafer fabs, and we have four major wafer fabs that are spread out across the globe. Our largest, and arguably the world's largest and most advanced indium phosphide and gallium arsenide fab is in Sherman, Texas. Here we have 6 in indium phosphide lines and 6 in gallium arsenide lines for multiple different devices running. The second fab we have is in Järfälla, Sweden, also running 6 in indium phosphide. Our Järfälla, Sweden is the place where we've developed some of our most advanced devices. We've been in production there for over two decades, and we have a broad mix of solutions, again, there. The third fab is in Zurich, Switzerland. Zurich, Switzerland was historically our gallium arsenide fab, where we run 6 in VCSELs in very high volume for datacom applications. We are now ramping up 6 in indium phosphide in that fab as well to enable the immense ramp we're seeing in demand on high-power lasers for PhotonLink and CPO applications. Finally, we have a fab in Fremont, California. This fab is running our 3 in production line and is where a lot of our most advanced devices are developed as well. That rounds out our complete suite of wafer fabs. We're expanding capacity in all of these fabs today. Our laser capacity expansion has really been driven by our 6 in transition. About six years ago, we transitioned all of our gallium arsenide production over to 6 in. We've been running 6 in gallium arsenide in both Sherman, Texas and Zurich, Switzerland since that time. We finally completed around last year the final transition, which was our pump laser platforms over to 6 in. Through that period of time, we've increased capacity on our gallium arsenide lines by about a factor of four. Today, we have capacity for over 1 billion VCSEL arrays per year. I can't even count the number of VCSELs that is, but it's a huge volume capability. Indium phosphide, we've learned from the success of that 6 in transition in gallium arsenide. We're now transitioning all of our indium phosphide lines to 6 in, and that's been a tremendous success for us so far. We have the bulk of our high-power CW lasers, our EMLs, and our photodiodes all qualified now on our 6 in platform and ramping. Last year, we doubled our capacity of indium phosphide globally. We did that a quarter ahead of schedule, and next year we're planning to double that again, or more than double that again. We're laser focused on our ramp, no pun intended, and we're doing it with our largest form factor wafer size in the world. We're the first to ramp on 6 in indium phosphide of anyone globally. That wraps up my section. I've gone through the four critical types of lasers that we use for the generate function for our integrated optics PhotonLink platform. Our ultra-high power CW lasers that are good for remote applications with the highest performance, both CPO and NPO. Our high-power CW lasers that are flexible and replaceable and can be easily used inside NPO solutions. Our conventional VCSEL arrays, which provide low power, high efficiency, and high density. Then our ultra-dense VCSEL arrays, which provide the smallest and most compact solution for getting high-capacity data across these short links. I'm going to pass it to my colleague, Julie, who will give us an overview of the shape, guide, and receive portion of the PhotonLink architecture. Julie, thank you. Hi, good evening, everybody. It's great to see so many familiar faces. For any of you I haven't met, I'm Julie Eng. I'm EVP of our Optical Components group and also the CTO. As Beck said, and Jim, I'm going to share with you the shape, guide, and receive portions of the PhotonLink portfolio. I'm also going to give you some examples of how we're bringing these integrated solutions to our customers. As Jim said, what we're going to do here, we're going to follow the path of the photon from where Beck generated them, which he called it the heart. I think the rest of it is like the cardiovascular system, shape, guide, and receive. Actually, if you double-click on it's much more complex than this. This is a picture of what it actually looks like. There's a tremendous amount of optical components and technology that are between the laser and the receiver. I hope I'll show you today that we have the broadest and deepest portfolio of products and of technologies. Also, as Jim mentioned and Beck mentioned, we're manufacturing these at scale. In these optical components that I'll introduce to you today, we shipped over 650 million optical components. We're in high volume in all of these components. If you start first, Beck showed you about the laser. We have to get the laser light into an isolator, into a fiber. The first way we do that is with glass molded lenses. The glass molded lenses give us efficient light coupling. We have excellent coupling performance. These are in production by the end of this calendar year. If you look at our entire lens portfolio, we've actually shipped 250 million of lenses. These lenses would go inside of the ELS and focus the light into the isolator and into the fiber. Once the light gets focused, the next step is the isolator. The isolator protects the laser from unwanted back reflection that otherwise harms the performance of the link. Isolators are based on a magneto-optic material called garnet. Coherent, we're vertically integrated in garnet. We literally grow the garnet boule in the United States, actually. Then we slice that garnet boule up into substrates and polish them. We grow the garnet epitaxy on top of those substrates. We dice them, and then we attach polarizers, which is how you make an isolator. We actually supply a significant fraction of the industry's garnet and isolators today. Isolators are becoming a key component in the CPO and NPO ramp. In fact, we have customers securing capacity for garnet and isolators with long-term agreements. The ELS, you can really think of the external laser source as an integration platform. It provides a platform to integrate that laser with the coupling lenses, with the isolator. There's other components in there, like an ELS fiber assembly, a thermoelectric cooler, which keeps the laser at the constant temperature. All of those internal components we design and manufacture, and then we integrate them into the ELS. Because we have all those vertically integrated components, we can give faster responses to customer specifications. The assembly and test of an ELS is built on our transceiver line, which we have demonstrated manufacturing at scale. We are sampling ELSs to customers today. Okay, we have to get the light from the ELS to the receive side. The way we do that, of course, is by fiber. It's maybe not so well-known, but Coherent, we have an internal fiber design and manufacturing capability. We've actually had that for 25 years. We've shipped over 300 million meters of fiber. But where we focus is on the specialty fiber, like really hard-to-make fibers. We're not competing on standard commodity SMF or MMF, just really hard stuff to make, and we're really, really good at making the hard stuff. One thing that's hard is polarization-maintaining fiber. Polarization-maintaining fiber is key for the silicon photonics with the remote laser CPO architecture, and we're production-ready on that fiber now. Here at the ECOC 2026, we're introducing this Mode-Matched Fiber, and what this is a fiber that improves coupling efficiency to edge-coupled silicon photonics. Finally, on the right, we have Multi-Core Fiber, and as Beck mentioned, that's very valuable, especially for these VCSEL, highly dense VCSEL array-based solutions, from NPO, CPO, and chip-to-chip. Right now we're undergoing executing a very significant fiber manufacturing expansion for both CPO and NPO. Okay, once we get the light through the fiber, now you have to get the light into the receiver. In the silicon photonics implementations, we're making micro lens arrays. Micro lens arrays are lithographically formed in silicon lenses. They have to be very precisely uniform one to the next and all have good focusing capability. If you're using grading coupled silicon photonics, they turn the light also. That's called Prism Micro Lens Arrays. We're making those. Fiber guides. This is a precisely V-groove etch that then holds the fiber. The two of those things together help you get good, precise coupling efficiency in your integrated solution. Then we can integrate those components with our polarization-maintaining fiber into a fiber attach unit. That provides an integration step that we can deliver to the customers. Here at ECOC, we're also introducing our detachable FAU connector, which helps with serviceability and installation. We can integrate all those FAUs also into a precision fiber assembly. This helps route and protect the fibers for ease of deployment and also for reduction of fiber damage. All of those things put together are either in production now, by the end of this year or early next year. This is one way in which we can help solve the integration problem for our customers. Once you start having fiber inside, say, the rack at very high amounts, it becomes very important to our customers to try to reduce the fiber count. One way you can do that is with a component called a circulator. Circulator takes a beam that might be going this direction with a signal that might be going this direction and combines them out on the same fiber. It reduces your fiber count by a factor of two. Circulators are also based on garnet, and they also require a crystal called yttrium vanadate, which we also manufacture and design internally. Then we have WDM multiplexers. These take many wavelengths and combine them onto one wavelength. Those rely on super precision optical coatings, which we are also an expert at, and we deliver in scale. Finally, fiber shuffles organize that high density fiber routing and can help reduce fiber breakage. We can put the circulators and the WDM multiplexers in that fiber shuffle in the ELS or in connectors. All of those products are in production. As we get to the receive side, Beck told you about indium phosphide and gallium arsenide receivers. Some of the CPO and NPO implementations have silicon photonics receivers. We have an in-house silicon photonics design team. We use multiple foundries, as is common in the silicon industry, and then we do in-house assembly and test. We have demonstrated 200 Gbps per lane silicon photonics, and we are shipping that in production in our transceivers today. At OFC earlier this year, the Optical Fiber Conference in March of this year, we demonstrated 400 Gb link based on silicon photonics. We are the only person in the world, to my knowledge, that has shown that and demonstrated. We demonstrated on the show floor and also got a post-deadline paper at OFC on that topic. Just like the ELS can integrate a lot of components, little NPO modules can also integrate a lot of the components, do the integration work for our customers. Here at ECOC and also at OFC, we showed our 6.4 Tbps Silicon Photonics NPO module. This integrates the silicon photonics with the drivers. The silicon photonics has modulators and photodetectors with drivers, TIAs, and all the optical coupling, and this is 32 lanes of 200 Gbps. In addition to that then, if you look at NPO, you can have a common form factor, common socket that can accommodate either silicon photonics or VCSELs. We also showed our VCSEL NPO at OFC, and that integrated the VCSEL, the photodiodes, the drivers, the TIAs, and the coupling optics into a single deployable NPO. We are a founder of the Open CPX MSA, so we are helping to develop an open ecosystem for a common connector among vendors, which gives our customers actually architectural flexibility. Hopefully now I have shown you the products and the technology portfolio, both the depth, the breadth, the ability of us to manufacture it at scale. Now I am going to give you some PhotonLink examples of how these integrated optics come together to solve solutions for our customers. I am going to give you four examples. I will give you a CPO example, two NPO examples, and then a chip-to-chip example. Starting first with CPO. This is a CPO solution with an external laser source. The CPO architecture is putting the optical engine actually literally on the package of the switch or XPU chip. It offers the lowest power consumption and in many cases, the highest density. We have an end-to-end solution for CPO with ELS and a remote laser. You can see we can provide the ELS with our own laser, lens, isolators, TECs, and fiber assembly. You can couple from that to our FAU, which has our PMLA, our fiber guide, and our PM fiber. We can integrate all of the FAUs into a precision fiber subassembly, so handling all that fiber routing and fiber protection for our customers. Then we can provide the modulation and the detection with our 6.4 Tbps silicon photonics. What you can see here is we can provide an integrated end-to-end solution for our customers. As Jim mentioned, we have over 10 active customer engagements in this area. He estimated our content at $15,000 per 100 Tbps I/O, and we do have an anchor customer, an LTA secured. This is our NVIDIA agreement, which is a public agreement, which we can mention our customer for that agreement. We expect to see scale out in Q4 of 2026 and scale up in second half of next year. Okay, now I'm going to take an example NPO solution. The NPO solution takes that little 6.4 Tbps silicon photonics, socketed NPO I mentioned, just for those of you who think it this way. This is actually 15 Gb p s per millimeter. You think about that like the end of a pencil is like a millimeter, and we're cranking out 15 Gb p s in every millimeter. You could start from that optical engine, go into our FAU, the same precision fiber assembly, and if you're doing it remotely, you could have an ELS. Of course, you could integrate the laser into the silicon photonics NPO also. In these type of applications, we have greater than 10 customer engagements also, similar Coherent content. In this application, we also have an anchor customer and an LTA secured, and we see this ramp at the second half of next year. Then, as I mentioned, because NPO gives you a common form factor, I forgot to say, but the NPO is moving off of the package onto the board right next to the package, and there are trade-offs with that. An advantage of the NPO is it's easier to deploy and it's more serviceable. But it comes at a price, and that price is you move it a little bit away from the chip, that requires power consumption, and that socket impacts the signal integrity. There are trade-offs with all of these architectural decisions. But one great thing about the NPO is it's an architectural platform. You can use silicon photonics or you can use VCSELs. This VCSEL one I'm showing you on the left is our NPO solution with the 2D VCSEL array that Beck mentioned. In this case, we integrate the substrates, the drivers and TIAs, our 2D VCSELs, 2D photodiode arrays, a heat spreader, a lens array. We've actually demonstrated this at 1.2 pJ / bit, which as far as I know, is the lowest CPO or NPO power consumption, including the laser. Then we also have an NPO solution based on our high-density VCSEL that Beck mentioned, and we have the high-density VCSEL and photodiode there. We integrate it with the TIAs and the drivers. Then some of our customers want to take the light out the top. Some of our customers want to take the light out the side. You take the light out the side, we incorporate also a turning mirror. We have lenses. We have a lensed fiber array. We can bring that whole end-to-end solution for our customers. Then the last example I'll tell you about is called chip-to-chip interconnect. We're seeing a lot more interest in this in the last, I would say, three to nine months, so this calendar year. And I think the way you can think about it is that as the bandwidth between compute and memory scales, the optical link moves from the edge of the package to the interior of the package. Or another way to think about it, as Jim mentioned, with photonics transition from electronics to photonics, always happens when the data rate goes up and the distance is fixed, say. It happened first on the intercontinental telecom network, then the terrestrial network, which is a shorter than intercontinental, then the next thing was inside the data center. Now we're talking about inside the rack. This is the next frontier, inside the chip package. We're seeing people look at multiple architectures, including ELS with ultra-high power lasers or integrated lasers, but also high-density VCSEL solutions for this. One thing you see is with all the lanes here, there's a lot of discussion about using WDM. This is a case where circulators and WDM multiplexers come in as a very important architectural tool. Here we have over five customer engagements, and this type of deployment is more like an end-of-the-decade deployment. Hopefully I've been able to demonstrate through the CPO, NPO, and chip-to-chip how we bring this PhotonLink platform an integrated end-to-end solution for our customers. As I mentioned, we have a ton of customer engagement on all three of these areas, and we've discussed the timing of end of this year, end of next year, and then 2029, 2030. But what I also wanted to point out was one of the ways in which we can be the trusted advisor to the customer is because we have all these technologies. If you, as our customer, come to us and say, "I'm trying to solve this architectural problem," we can openly tell you, "If you do it with VCSELs, here's your power consumption, here's your cost. If you have integrated laser, here's your complexity, here's your impact to your signal integrity." We can give you an honest assessment of all the choices if you tell us your architectural requirements. I think that our customers find that very powerful, that they're not just being sold a solution, because we have all the solutions, we think about your problem and try to give you the best tools in our toolkit to provide you an integrated solution with PhotonLink. With that'll conclude the formal part of our presentation. Hopefully, between myself and Jim and Beck, we've been able to show you that we do have, I believe we have the broadest portfolio and the deepest portfolio and technologies in optics for the integrated photonic solution. With our PhotonLink platform, we're pulling this together as a complete integrated end-to-end solution for our customers, and we're highly capable. We've demonstrated in the past that we're able to manufacture at scale, and we're definitely scaling all of these technologies and products today. With that, I'd like to thank you all again for coming. Thank you for your attention, and I will hand it off back to Sanjai. Thanks