Ciena Corporation (CIEN)

Status update

← View all transcripts

Sep 18, 2026

Summary

AI-driven demand supports a projected $52B 2029 TAM, with the fastest expansion in AI WAN and interconnects. Financial targets include ~$14B revenue, ~50% gross margin, and 32%-35% operating margin by 2029.

Welcome. I'm Gregg Lampf, Ciena's VP of Investor Relations. Thank you for joining this webinar, where our team will discuss Ciena's strategy, innovation, leadership, and our financial goals during a pivotal period for our company and the industry. With decades of experience in innovation, Ciena has established itself as the clear leader in high-speed connectivity. Since last year, when we hosted a similar event, we've continued to advance our industry-leading optical technologies that underpin today's network infrastructure for AI workloads and that set a strong foundation for future demands. Backed by focused and leverageable R&D, deep expertise, and proven scale, we've extended our leadership in optical systems while increasingly applying our coherent optical capabilities to a growing portfolio of inside-the-data-center connectivity solutions, where rising bandwidth demand is pushing existing technologies to their limits. In today's constrained supply environment, we're working proactively to secure multi-year customer demand commitments and long-term supplier agreements, strengthening our abilities to support growing demand. These efforts are delivering results, driving operating leverage, and accelerating earnings growth. We encourage you to listen to the full webinar, including important financial updates on how we plan to continue converting our strong market position into long-term value for our customers and our shareholders. As per this slide indicates and directs accordingly, please note that we'll be discussing a number of topics that will be forward-looking and measures that will be non-GAAP. Thank you. Hi, this is David Rothenstein, Chief Strategy Officer with Ciena. In connection with the other presentations that we'll be giving at our Ottawa Investor Forum this year, and as a predicate for that content, I'd like to provide some detail and context around current and evolving industry dynamics, particularly those being driven by AI, our view about categorizing the key markets in which Ciena plays, and the architectural shifts going on within those markets, and how all of that translates into an expanding market opportunity for Ciena, particularly reflected in our view about the total addressable market for our business. At the highest level, AI is creating a new era in communications that is being built on top of previous eras. We can refer to the era from 1995 to 2010 as having been fundamentally enabled by the internet, which oversaw the growth of connectivity. The internet's global connectivity platform gave rise to the subsequent cloud era, with the introduction of the digital cloud economy, and that cloud era continues to grow. Now, in much the same way as cloud applications were built on internet connectivity, AI is building interactive intelligence on top of the cloud and is springboarding off of large data center infrastructure investment, largely driven by enterprise adoption of cloud service platforms. One of the key characteristics of these subsequent eras is the definition and redefinition of scale, with the scale of all aspects of AI dwarfing that of previous eras in communications. Another key characteristic is that optical technologies have enabled each successive era. In short, AI has created an inflection point for our industry. Whereas the internet era was driven by the telecom service providers, the drivers of industry change today are the cloud providers. Hyperscalers are funding this AI investment wave, and they've recently been joined by new entrants called the neoscalers or neoclouds, who are building new GPU-as-a-service or AI-as-a-service business models. Since ChatGPT was launched three and a half years ago, this customer segment has been investing and continues to invest unprecedented sums into the AI market. Analysts now expect their CapEx spend this year on everything relating to AI, including data centers, compute, power, and network infrastructure, to reach $1.2 trillion, up over $100 billion from the beginning of this year. That number is expected to grow to at least $1.6 trillion next year, and probably higher, and growing to some significantly larger number by 2030. Indeed, it's worth pointing out that this market growth has been very difficult to forecast. As we illustrate on the right-hand chart, actual spend has consistently been higher than the previous year's forecast of various industry analysts. While it's not shown on this slide, there's an important point to note. There's a lot of industry and media focus, understandably, on the cloud providers, but service providers are playing and will continue to play an important role around the world. They've largely underinvested their networks in recent years. Their access to fiber means that they will benefit from an expanding set of managed wholesale opportunities. Data center build-outs will increase to support data sovereignty rules and regulations, and traffic will ultimately need to leave the data center and traverse the service provider network over the WAN in order for AI investments to be monetized. The large-scale investment in AI infrastructure is showing up in data and bandwidth forecasts, and this slide highlights some examples. The amount of global data being generated continues to increase. IDC's latest model highlights the continued acceleration of data growth, with hundreds of zettabytes expected to be generated by 2029. Omdia has forecast the amount of data service providers are expected to move across their networks, which is a proxy for how much AI is influencing the bandwidth that end users consume. The data that's related to AI is growing at over 35%, indicating that AI will be a significant driver of new growth across the metro. In terms of bandwidth, we're running out of adjectives to describe the growth rates. Inside the data center, the traffic demand and growth rate is not only enormous, but also industry analysts have been unable to forecast it successfully. Cignal AI has needed to double their forecast in each of the past two years. Network bandwidth is growing at a similar rate, albeit at a much smaller capacity. One of the reasons for network forecast uncertainty is the introduction of scale across. This special class of data center traffic interconnects backend fabric of training compute clusters located in geographically distributed data centers, with extraordinarily high amounts of bandwidth required. Our internal projections, which are very preliminary, estimate that scale across bandwidth will account for one-third of total global network demand by 2029. But AI training isn't the only driver of new growth. The final chart illustrates how the energy consumed by inference workloads is forecasted by McKinsey to grow faster than training over the next few years. This indicates a significant shift from building out the foundational AI training engine towards using and monetizing the power of AI. Building on these positive growth signals, we believe we are only in the early stages of a substantial multi-year ramp in optical connectivity demand. While this chart is based solely on our internal assessment, it provides a view of directionality for the expansion in optical communications bandwidth in the network and breakdown in workloads over the network by the end of the decade. Starting with non-AI. The cloud is not going away. Cloud backlog figures continue to rise, and traffic associated with network cloud applications are growing at a steady pace. We expect that dynamic will continue going forward. By 2030, we estimate that the proportion of CapEx associated with non-AI workloads will still account for approximately 30% of the total. On top of this, we have seen a surge in inter-data center traffic relating to scale across and other high-capacity bandwidth flows associated with the distribution of AI training workloads. Models are getting bigger and will need to be trained over greater distances, and ongoing power, space, and cooling limitations will continue this trend. We estimate AI training workloads to consume approximately 30% of CapEx by 2030. Turning to AI inference. While the last three years have been primarily spent building out infrastructure to train large language models, the coming years will be spent building out the infrastructure both to distribute these models across the network and to monetize the investment. In so doing, inferencing will become the dominant driver of global bandwidth, and we are just at the very beginning of that inflection point. By 2030, we anticipate inference workloads to assume the dominant share of CapEx at approximately 40%, which brings the total spend on AI workloads to 70%. Inference has already started in the hyperscale domain, but it will quickly spread out to geographically distributed regions and metros. It is important to note that we are projecting all three drivers of connectivity demand to continue growing, albeit at different rates, well into the future. As an important example of this growth, the scale of AI data center infrastructure is evolving quickly. The growth of massive GPU clusters over multiple data centers is redefining the units we use to describe the scale cluster size, energy consumption, and bandwidth capacity. Over a few short years, the number of GPUs in an AI cluster has grown from being measured in thousands to millions, and corresponding cluster energy consumption has grown from megawatts to gigawatts. As a result, AI is redefining the fundamental units of scale in both the data center and the network. The expansion of cluster size is redefining the AI unit of compute, evolving from the single processor to a rack of GPUs to a data center comprising thousands of racks. This effectively multiplies the unit of compute by 100,000 times. It is also having a direct impact on the cross-sectional bandwidth carried between data centers and in the network. Historically, DCI connections would be required to carry hundreds of gigabits to a few terabits per second. This could be achieved using one or a few wavelengths within a single optical fiber. Today, fabric expansion between adjacent data centers is demanding petabits per second of cross-sectional capacity, and this is significantly more than a single fiber can handle. Moving from a single wavelength that can carry one terabit to a single fiber carrying 100 wavelengths, to a conduit that can carry 1,000 fibers. In redefining the new unit of capacity, this results in the ability to ultimately support hundreds of petabits per second per conduit. In the case of both compute and bandwidth capacity, all of these units will exist and coexist for a long time to come. Putting it all together, it is the sum total of these bandwidth and other AI-driven dynamics that give us such strong confidence in the durability of industry demand for optical connectivity and the criticality of network infrastructure. It is also serving to underwrite our views about the expanding market opportunity and specifically market growth. As we have said recently, we project the total addressable market, or TAM, for our business to effectively double over the next three years. More specifically, we see a three-year growth rate of 24%, where the market grows from approximately $27 billion today to approximately $52 billion in 2029. I now want to turn to breaking down this TAM a bit further. First, by going through some of the key architectural shifts within our key markets. We will then show you our view of the TAM by those markets, and then by looking through the lens of our product portfolio, after which we will show a view of the TAM broken down by product groups. With respect to all of these views on total addressable market, it is important to keep in mind that our growth projections are not just based on a rising tide of market growth, but rather the assumption that we will continue to take market share in this environment. The other presentations at our investor forum will explain exactly why. With the introduction of AI over the past three years, we have witnessed significant growth in high-bandwidth connectivity by the cloud providers, introducing new markets for high-performance optical connectivity. Within the traditional WAN markets, our business has provided bandwidth capacity that has enabled the circulatory system of the internet, first for long-distance networking, and then expanding to metro backbone and regional markets, and then access and aggregation at the network edge. Over the past decade, we have seen the build-out of data centers in support of globally distributed cloud applications. This drove new investments into front-end data center fabrics and the DCI services that connected those data centers together across the WAN. With the introduction of AI, a new back-end fabric has evolved to support AI's training function. Massive bandwidth growth in the fabric has stimulated rapid growth of markets for 400 gig, 800 gig, 1.6T, and soon 3.2T optical interconnects. It has also created new high-bandwidth connectivity between different data centers in the form of distributed training, or scale across. Now today, cloud's front-end fabric is expanding to accommodate increased cloud and AI adoption via inference, and it is also accelerating DCI connectivity to support the interconnection of many new data centers. We categorize these trends in the context of the evolution of optical connectivity inside the data center and in the context of a new market category we are introducing today that we are calling AI WAN. In this market, we include the different DCI applications associated with cloud networking, AI training, and AI inference. Turning to some of the key dynamics and architectural shifts within each of these markets. As we know, the traditional WAN markets have been historically over-indexed to the service providers. Now, with AI and the AI investment era, we can see how these markets and how we serve them are being significantly impacted. In the network backbone, as enterprises increase their use of AI and cloud providers look to build out their DCI connectivity reach around the world, bandwidth and fiber scale services are on the rise in the form of managed wholesale services such as MOFN, dark fiber, and other infrastructure services, creating new growth markets for the service providers. At the network edge, internet-class networks are not expected to meet the bandwidth and quality of service demands of AI, such as latency or determinism, which require a re-architect of the edge, opening a new market for inferencing across routing and switching interconnect infrastructure. In terms of network operations, extreme scale and complexity is impacting the service provider's ability to transform the network operations capabilities, which is creating new markets for the application of AI intelligence and AIOps to traditional OSS, management, and control domains. With respect to the market that we're calling the AI WAN, here, the AI-driven shifts in architecture are defining a new set of WAN-related applications. First, network DCI, or data center interconnect, continues to expand in support of both cloud and AI. Cloud providers' need for operational simplicity at high volume with lower price and power has challenged the traditional modems consumption model in metro and regional areas, introducing a new market for ZR and ZR+ class coherent pluggable transceivers. A new DCI mesh is evolving to interconnect new co-location data centers that are being built out in support of private enterprise hosting of AI models or neocloud data centers offering GPU as a service. With data center scale across, as we've noted, a limited supply of energy to data centers is forcing AI training workloads to expand across multiple data center locations, creating a new market for lossless, ultra-high capacity coherent optical connectivity over many fibers, forcing a rethink of traditional amplified fiber line systems. The data center markets, which are, of course, heavily over-indexed to the cloud providers, are also being significantly impacted by AI. Starting with a scale-out fabric, today's existing approach that combines packet switches and IMDD plugs is too large and power-hungry to meet next-generation specs. As a result, the market is now exploring new ultra-high capacity co-packaged and near-packaged optics solutions to overcome this challenge inside the data center. Between buildings across the data center campus, IMDD modulation is hitting bandwidth and distance limitations, making this the likely first deployment scenario for coherent light. With respect to data center scale-up, future generations of XPUs require higher bandwidth interconnectivity, but the distance that existing copper interconnects can transmit is reducing as data rates increase. This is opening markets for new copper cable solutions, but will ultimately lead to the further opticalization of the data center via ultra-dense optical input/outputs based on silicon photonic interposers and co-packaged optics. Finally, with respect to data center operations, the bandwidth and power required to manage millions of devices is straining traditional out-of-band management communication solutions, which are based on older technologies, and has created a new market for DCOM, a PON-based out-of-band network. Stepping back and looking at all of these architectural shifts, they are not only impacting the industry, but are also directly operating in favor of Ciena. These shifts are coming to us, and our strategy is purposely designed and aligned to take advantage of them. They are creating increased opportunities for our business, and Brodie and Dena will talk about how we are addressing them from a portfolio standpoint and why we win. They are at the core of what is driving the expanding addressable market. Now that we have further context of key dynamics within these three primary markets, we can return to our TAM chart and break it down by those markets. Within the $52 billion overall market in 2029, first, in the traditional WAN markets, we see very strong growth from $21 billion to $26 billion, representing an 8% growth rate. This is important for two reasons. First, that growth rate is much larger than the historical low to mid-single digits growth rate. Second, that growth is off of a very large base. Second, in the AI WAN market, we see very strong growth from a market going from $7 billion to $14 billion over the next three years, representing a 30% growth rate. Finally, on the inside the data center market, we see extreme growth, growing from a base of almost nothing today for Ciena to $12 billion in 2029. Turning now to Ciena and its products, we can see specifically how our strategic portfolio investments map across these markets. Across the top is an illustration of where our primary customer segments play in relation to these three markets and their sub-domains, with the service providers spending and operating in the WAN markets, as well as in the DCI portion of the AI WAN, for all of the reasons that I mentioned at the outset of the presentation, and with the cloud providers spending and operating all the way from the network backbone in the WAN, through the AI WAN, into the data center and data center operations. Perhaps more helpfully, our portfolio can be categorized into product groupings. Across the WAN and the AI WAN at the top of the page, you can see the optical systems portion of the portfolio. This consists of our WDM optical networking products, including modems and photonic line systems, including the new RLS Hyper-Rail product, supported by Navigator, our multi-layer domain controller for network operations. Below is a grouping of our interconnects portfolio, consisting of plugs, modules, and components. Starting with coherent modules and the ZR class of coherent pluggables for both WAN and AI WAN applications. It also includes Coherent-Lite for metro campus data center applications, co-packaged and near-packaged optics solutions, and disaggregated components for scale-up and scale-out applications, and DCOM for data center operations management. Separately and more easily illustrated on a standalone basis are the WAN routing and switching products, including the coherent routing and access and aggregation products, as well as our Blue Planet software OSS solutions, all of which is supported and underpinned by our global services portfolio of implement, maintain, and advanced services. An important takeaway is how, consistent with the evolution of network architectures and applications, several of our product solutions span across multiple domains and even the markets themselves. Now, against the backdrop of the product portfolio and those product groupings, we'll turn one more time to the TAM growth chart. Again, within the $52 billion overall market in 2029, there are a few key takeaways. First, we see a 2% growth rate combined for the routing and switching portion of the network in which Ciena plays, as well as network operations. These areas are heavily directed to the service providers, and the growth rate's indicative of the role being played by those solutions in networks and network applications going forward. Second, there is strong 15% growth in the services portfolio, which is highly correlated to the projected growth of the installed base of our hardware products. That's both optical systems and routing and switching. Third, there is 18% growth rate in the optical systems portion of the portfolio, which again, is significantly greater than the historical growth rate. Fourth and finally, we see almost 100% growth in interconnects, again, off of a small base, but still very significant given the role being played by plugs, modules, and components in both the data center and the AI WAN. This clearly represents the fastest-growing opportunity in the industry. Taken together, Ciena has a product portfolio that is designed to maximize the growth opportunities in the context of the industry dynamics and architectural shifts that we've discussed. You'll hear more about the product portfolio in other presentations at the investor forum. In summary, we have a tremendous opportunity ahead of us, with an almost generational set of AI-driven industry dynamics and architectural shifts. In that context, our vision and our strategy haven't changed and are incredibly well-aligned with those market opportunities. Our priority is on scaling all aspects of Ciena in order to turn that strategy into execution and take full advantage of the opportunities ahead of us. Turning to separate presentations from the investor forum. Brodie Gage, our EVP of Products and Technology, and Dino DiPerna, our EVP of Research and Development, will focus on how we're scaling our product portfolio, and specifically our optical systems and interconnect products and technologies to address the expanded market opportunity that I've discussed. They'll go through how we're addressing and will be addressing the industry dynamics and architectural shifts, and specifically why we will win in each of those cases. Grant Hoffman, our new EVP and Chief Supply Chain Officer, will discuss how we're framing supply chain as a strategic growth enabler, and specifically how we're scaling our supply across a number of dimensions to service the accelerated customer demand we're seeing today and what we're projecting over the next several years. Mark Graff, our Executive Vice President and Chief Financial Officer, will discuss how we are leveraging our industry leadership to scale the value of and from our business, with the goals of delivering strong revenue and profit growth, strong cash generation, and meaningful long-term returns for our shareholders. These presentations contain a great deal of very relevant content. I strongly encourage everyone to listen to them all. Thank you very much for listening. We look forward to continue engaging with the financial community on Ciena's journey going forward. Hello, my name is Brodie Gage, Senior Vice President and Chief Product and Technology Officer. Together with Dino DiPerna, who is our Senior Vice President and Chief Research and Development Officer, we are going to talk about the technology leadership Ciena has across our core markets. Ciena's addressable market covers a vast array of network applications. On the left-hand side of the chart, you can see our traditional network markets, typically service provider WAN networks from network edge through network backbone and network operations. In the middle, what we call the AI WAN, which is inclusive of network DCI and scale across fabrics. Finally, on the right-hand side of the chart, inside the data center, from scale up to scale out, as well as our DCOM solution for data center operations. Not only does Ciena play in each one of those network applications, it is the only supplier that brings leading technology across the full interconnect and optical systems continuum. We provide optical solutions inside the data center for scale up and scale out across the AI WAN and traditional network markets from metro, long haul terrestrial, as well as submarine. We are the only ones that can deliver optical solutions all the way from scale up to submarine transport. What we are going to focus on in this portion of the presentation is the traditional network markets on the left and the AI WAN applications inclusive of network DCI and fabric scale across. There are two major sections. One, we will talk about the architectural shifts, which is expanding Ciena's addressable market. Two, why does Ciena win across technology and systems leadership, scale and deployment advantage, and our innovation engine? As I just mentioned, there are architectural shifts in the network that are expanding Ciena's addressable market. One of the lens that you can put on our business is a customer lens, and we really break it down into two major customer units, the cloud providers and then the service providers. On the cloud provider side of the equation, the unit of currency used to be wavelengths, and we used to talk about 100 gig, 400 gig, 800 gig, et cetera. With the amount of demand and capacity growth that is going on to those cloud provider networks, whether it is for scale across, data ingestion, content distribution, or inferencing, the unit of currency is shifting from wavelengths to fibers. The other thing I will state is that scale across is still in the early innings. We have spoken multiple times about synchronous and asynchronous scale across applications. What we're seeing is there are more and more deployments within each of the hyperscalers, and there are more and more hyperscalers that are starting to do training across distance. Inference is going to have an impact on both the cloud provider network and the service provider network. Inference drives distributed connectivity and expands connectivity beyond centralized AI infrastructure. It will impact the cloud provider networks, of course, but the service providers are going to have a larger and larger role to play as inference becomes more pervasive. On the service provider side, these applications are driving new service models. We've talked about MOFN, or managed optical fiber network. There's applications for dedicated fiber, and even on-demand connectivity and new services. The net of all this is that these architectural shifts are increasing the strategic value of optical systems. With the capacity growth and the unit of currency shifting from wavelengths to fibers, it drives a new generation of optical platforms. On the left-hand side of the chart, you can see how AI capacity would impact amplifier huts along a long-haul route using traditional technology. If you wanted to support 20 petabits across the network, you would need 22 huts at each ILA site. That is not sustainable, and that's what's driving new photonic architectures like Multi-Rail, where we can drastically increase the capacity and the density of our amplifier solution. It's not only having an impact on line systems, though. If you're going to light up full systems or full fibers, it doesn't make a lot of sense to plug in your transponders one at a time in a pluggable form factor. What makes more sense is to have very low-cost, low-power connectivity from your routing layer into your transport layer and light up the full spectrum one at a time, versus one channel or one wavelength at a time. That drives much lower cost, much better density, and lower power and cost per bit. These new architectures and the amount of traffic that's going on the network requires innovation in how capacity is packaged, deployed, and managed. It also goes back to our philosophy that the winning hand is to not have one type of solution, just pluggable optics or just high-performance modems. The right solution requires both. Scale-across is one of the first use cases that starts to put this new architecture in place, and in fact, is one of the applications that's moved the unit of currency from wavelengths to fibers. We've talked about one hyperscaler deploying our solution across about 80 km of distance and carrying over 20 petabits per second of bandwidth. For that span and that type of capacity, it requires over 400 fiber pairs, and just for that one application. I want to highlight that each hyperscaler project are going to be different, number of GPUs, the number of fibers, unprotected or protected. But for very illustrative purposes, again, each hyperscaler will be different, but for illustrative purposes, deployments, say, for a 500,000 GPU cluster, can reach to about $500 million. That's not Ciena revenue, that's the TAM of the opportunity, or about $1,000 per GPU in that example. That's without amplifiers. Once distances grow and reach 1,000 km span length, for example, you now need inline amplifiers. In that case, the project size can double. If it was the exact same design, it could be doubling to $1 billion. If you went to 2,000 km of reach, that project, if it was exactly the same, which it won't be, but if it was, can jump to a 4X multiplier or $2 billion, because you also need to change the endpoint transponders from pluggables to high-performance modems. Again, this is a generalized example, and the specifics, the number of fiber pairs, et cetera, will be different depending on that project. Because of constraints on power and the need to locate data centers across regions to access that power, we expect every hyperscaler will want to do scale-across, both synchronous and asynchronous, over extended distances. We co-developed both our next-generation RLS line system, our Hyper-Rail product, with that customer for terminal configurations, with redundancy and protection switching built in. We developed a set of 800 gig coherent optics that are deployed in that use case. Going forward, we expect that same customer to deploy in more regions, and it'll take time. We see other customers coming on board to deploy both synchronous and asynchronous scale-across applications. When they do that, they require more dense amplifiers. That's where Hyper-Rail comes in. In some cases, as they increase the distance, our high-performance platforms and modems are required along with the plugs. As I mentioned earlier, we see scale-across as entering the early innings of a multi-year investment cycle. There will be more deployments, and there'll be more deployments over distance. Our customers cannot get GPUs in one data center, and as they distribute those GPUs, the network becomes more and more critical. This AI infrastructure build-out will create opportunities for the service providers as well. If you start on the right-hand side of this chart, with distributed training, the service providers have a role to play as well, whether it provides conduits or fiber or managed optical fiber networking. There are certain regions where the cloud providers do not have the regulatory rights to build out a network. They have to rely on the service providers, and the service providers will provide that network infrastructure on behalf of the cloud provider. As we move more and more to inferencing, the service providers have the real estate assets and the fiber connectivity into major data centers, central offices, and even the end consumer and enterprise customers. They're going to have a role to play in inferencing fabrics and providing the connectivity for inferencing traffic. If you go to the middle of the chart, an AI cloud distribution, the reality is that for data sovereignty, there's going to be more and more distributed AI cloud infrastructure. Certain regions are going to want to have their data stay within their continent or within their country. That's going to drive more and more cloud or AI cloud distribution in the network, and the service providers will have a role to provide that connectivity. As we move more and more towards the edge of the AI infrastructure, there's a need to enable enterprises to connect to the AI cloud or interconnect the private enterprise AI infrastructure as well. Finally, there's all the discussion around building, operating, and connecting AI infrastructure at the edge of the network. These service providers have the data centers, the offices, the fiber infrastructure to provide that edge AI infrastructure. You see more and more partnerships. I believe you'll see more and more partnerships between the XPU providers as well as the hyperscalers or cloud providers together with the service providers. AI is expanding where and how service providers can participate in the connectivity value chain. Again, we believe there's a significant opportunity for the service providers going forward. Now that we've covered the key trends in the industry and what's driving the capacity, the next question is, why does Ciena win? It really comes down to three major items. Number one is technology and systems leadership. We've driven best-in-class performance over multiple generations, and we have systems-level understanding of how technology is used and the challenges it must solve. We've been in this market for decades, and we leverage that experience to build best-in-class products. Number two is scale and deployment. We output, from a supply chain perspective, more photonics than anybody else in this industry, and we have the breadth across coherent modems, photonics, as well as routing and switching. We also have end-to-end service and support capabilities that we'll talk to in more detail, both for the service providers and the cloud providers. Finally, we have a best-in-class innovation engine. We get faster to market than anybody else, and we drive better cost and power per bit than anybody else in this industry. The other thing that we do is we build reusable technology building blocks. That allows us to leverage these technologies across multiple use cases and multiple applications. We're going to cover these in more detail. On the technology and system side, I'm going to start with our coherent modems, which drives traffic down the fiber. We've been the first for every generation of coherent modem, whether that was 100 gig with WaveLogic 3 or now 1.6 terabits with WaveLogic 6 Extreme. We've led in coming out with the next generation of technology every single time, and that, of course, drives the lowest cost and power per bit. The other thing, though, that we've been doing over the last decade is we've been co-developing our products with key customers, especially with the hyperscalers, and we built those platforms to suit their needs. Whether it's a data model perspective, APIs, streaming telemetry, operational aspects, et cetera, we really co-developed. What that's allowed us to do is to sell our coherent modems, yes, in our Waveserver system products, yes, in our pluggable form factors, but also as coherent modems that can be integrated into their platforms as well. So we've repeatedly translated our core technology leadership into new platforms, new form factors, and deployment models as these cloud architectures have evolved. If you take our latest generation WaveLogic 6e 200 gigabaud 1.6T capable modem and you compare it to all of our system competitors' 140 gigabaud 1.2T capable modem, on a typical long-haul network across the United States, what type of advantages do our customers get by deploying WaveLogic 6e? Number one, they get more bits per wavelength. So within the same footprint, a 200 gigabaud modem reaches 1.6T, where 140 gigabaud typically falls back to 800 gig. So you get double the capacity per modem. For 800 gig services, or 2x400 gig type services, a 1.6T modem on that typical network supports 96% of the routes with no OEO regeneration. A 1.2T terabit modem requires, on half of the routes, a regeneration point, which adds cost, complexity, and power. On a capacity-per-capacity basis, a 200 gigabaud modem supports 25% fewer modems to carry the same set of typical traffic across that network. It's a massive advantage for Ciena and the customers that deploy this technology. Our two-year technology lead delivers meaningful network advantage and has driven many competitive wins. Shifting gears now to photonics. Ciena believes that the photonic layer is the foundation for AI-scale connectivity. Some of our routing competitors talk about the routed optical network and the elimination of photonics completely from the network. That is the wrong way to build an infrastructure. For AI scale, photonics are going to become more and more important. We've seen that in the metro network, we've seen it in the long haul and submarine network, and we're now actually starting to see it in the campus scale across and even inside the data center for multiple use cases, whether that's front-end switching or for XPU interconnect. Photonics are the lowest power, lowest latency, and lowest cost way to move bits across an infrastructure. Investing in the photonic layer, like Ciena has done, also drives system knowledge. Because you understand what's happening at the photonic layer, Ciena is able to customize its coherent modems and other solutions with that wide expertise and that breadth of knowledge across both the coherent modems and the photonics. The most efficient network is one that keeps traffic optical for as long as possible. Similar to our coherent modems, we're now on our sixth generation of photonic system innovation. We've been working and co-developing with multiple hyperscaler customers now for over eight years on our next-generation open line system, which we call RLS. We started with a C-band optimized CDC solution. Then we went to C&L, integrated C&L. Then we went to what we call colorless direct attach and passive point-to-point type use cases, sub-C use cases at step three. Realizing that our customers have multiple deployment models and multiple huts and solutions they need to build into, we developed 300-millimeter depth configurations. We developed high-density scale across use cases and platforms. Now, our high-density Hyper-Rail solution is that sixth generation. We've built these systems and co-developed these solutions again for over eight years. We've built the APIs, NBI, declarative configuration, streaming telemetry, auto-calibration, ASE noise loading, layer zero control plane with spectrum sharing, optical protection, direct attach, security solutions, together with these hyperscaler customers. Our competitors must overcome years of accumulated photonic systems expertise, automation, and deployment know-how. Hyper-Rail is not a new start. It's the evolution of what we've been doing now for over a decade. Ciena has led in photonic technology development, and we've co-developed with multiple hyperscaler and service provider customers. What has that done from a share perspective? If you take the Dell'Oro Group Optical Transport Report up to Q2 2026, you can see that Ciena has recognized 4.5x more photonics revenue than our nearest systems competitor. We've driven 5x more photonic revenue than all others combined, and we've driven 25 points of market share gain over the last four years. We now have 70% market share in open line systems. Critical component of our strategy going forward to continue to lead in photonic line systems and continue that generational advantage, our next step is what we call RLS Hyper-Rail. As I mentioned earlier, as scale across goes across larger distance and inference starts to impact the network in a larger way, you can't rely on the existing solution. The number of huts required for the bandwidth is just too much. So there's a massive densification of amplifiers. We're going to target a 32X density improvement in our first release coming out the end of this year, but that's not where it'll end. We're targeting even further density improvements and more vertical integration as part of our solution. We will leverage all the co-development that we've already done with all the hyperscalers. We'll leverage all the same software that we've developed over the last eight years. Because we have incumbency and integration into operational systems, we have a massive advantage as we move to this next generation of Hyper-Rail technology. On the final chart of our photonic line systems, I just want to kind of differentiate what Ciena has done from a strategy perspective versus our nearest system competitor in the middle, and the merchant or component suppliers that are starting to get into the game as well. When we first started our photonic systems way back when, prior to 2010 even, we relied on component suppliers to build our full line cards and our full solutions. We developed the software, but the hardware was developed by merchant suppliers, the gain blocks, et cetera. From 2018, when we started the co-development with the hyperscalers, we decided we were going to take back control and have the design authority on both the hardware and the software. So yes, we buy components from merchant suppliers, but we build our own line cards and we build our own software on top of that. What that allowed us to do is customize form factors, customize density, customize where we put Raman and where we put EDFAs, et cetera, et cetera. It allowed us to control the software architecture and the software solution on top of it and differentiate on multiple levers. We're going to continue to do that and into Hyper-Rail and beyond, and leverage all the co-development and work that we've already done. Our nearest competitor on the system side continues to have full line cards built by component suppliers. Their ability to customize for use cases and for different customers is much less than what Ciena's been able to achieve. Now, they're just getting going, we believe, with Hyper-Rail and trying to develop some of their own solutions, but they got at least eight years to catch up in all the work that we've done. From a component supplier perspective, component suppliers really don't have that system knowledge or that software around the APIs, the data models, the operational integration into our hyperscaler or service provider customers. I think they're really lagging versus what we've been able to achieve over the last few years. The second major reason why we win is because our scale and deployment advantage. If you look at our services, of course, we offer all the traditional network services that many of our system competitors offer as well, and we do that on a global basis. Whether it is to manage, maintain, support, build, et cetera, we have always offered those services. I think really where we have differentiated is we have really gone after some hyperscaler-oriented services as well. I talked about co-development of technology, but we also co-develop and help them build large-scale optical networks around the world. We support flexible operating models from a system perspective, a plug perspective, or even from a component perspective. We help them accelerate their deployments. We have over 80 engineers, for example, that are helping one hyperscaler customer deploy their optical network at scale. We maintain and support their network around the world, and we even offer proactive optimization, migration, and evolution services to our customers as well. Really what that has led to is a doubling of our hyperscaler services revenue from 2024 to 2026. Our largest services customer is now a hyperscaler, and we really differentiate versus all of our competitors in having over 1,000 deployment optical domain experts. We have done over 5,000 projects or 20,000 systems per year, and we have over 400 global sparing depots around the world. So along with our technology differentiation, our services differentiation complements that, but also provides a large moat that our nearest competitors are going to have to cross. The third major reason why we win in this market is a best-in-class innovation engine. We own and become vertically integrated where we can differentiate. We design for supply flexibility. I spoke about photonics and how we have taken over the design authority on all our line cards. That allows us to choose multiple suppliers and be able to output more than anybody else. We optimize across material systems. We are fabless, but we leverage CMOS, BiCMOS, gallium arsenide, silicon photonics, and indium phosphide materials. Of course, we invest, and we have increased our investment to continue to lead in this market, the AI WAN and the traditional WAN networks. Here at Ciena, we are definitely proud of our science and advanced technology, but at the end of the day, the key metrics that drive our technology programs are power per bit, space per bit, and cost per bit. In fact, Ciena were the pioneers of coherent DSP-assisted optical communications technology, and that dates back some 25 years ago. In the late 1990s, while the industry was busy deploying 10-gig wavelengths, we were pushing the technology towards 80 gig and even 160 gig. The physics and tech associated with that work made it very clear that the traditional on-off light modulation schemes of the past would simply not scale. This drove us to adapt, of course, and advance the coherent techniques long used in mobile radio, but not yet applied in the domain of optical communications. Looking back over the past 15 years and five generations of our WaveLogic coherent modems, we've consistently been first to market in delivering that next step in capacity, connectivity, and reach. We've increased individual wavelength channel capacities by a factor of 16, from 100 gigabits on a wave to an industry-leading 1.6 terabits on a wave today. These adaptive WaveLogic modems also greatly simplify the deployment engineering required to turn up and operate these global massive-scale networks, and all the while, we've reduced the power and space by over 95% on a per-bit basis. A coherent modem or transceiver is much more than a single chip. The large CMOS DSP, or digital signal processing engine, is definitely one of the critical components, but the coherent modem also depends on high-speed analog ICs. These include drivers and transimpedance amplifiers, or TIAs. along with the high-performance electro-optic devices such as the transmit modulator, the receive coherent mixer, photodetectors, and of course, a tunable laser. Even the DSP device is not simply a homogeneous digital CMOS piece. It includes a digital core that performs the advanced signal processing used to encode and recover the signal, as well as specialized mixed signal blocks such as a high-speed digital analog and analog-to-digital converters, and the client-side SERDES. These high-speed converters can ultimately limit the transceiver performance regardless of the strength and ingenuity of the digital signal processing core. This applies to modems that are high-performance sub-assemblies, like the one on the bottom right, as well as to the coherent pluggable modules in 400 gig, 800 gig, and emerging 1600 gig. Indeed, Ciena owns the critical intellectual property and design control over all of those blocks within a coherent modem. Our expertise across device types, material systems, and expertise in optical propagation allows us to make the right trade-offs across the functions, ultimately improving both performance and cost. This is far more advantageous than simply assembling a collection of separately sourced components. Each of those components comes with its own margining and nuances, and they have to be treated as standalone black boxes. For example, we can use DSP functions in our math processor to compensate for analog device implementation artifacts. We can make it a little easier on the analog parts. We can also tune the design to specific characteristics of the different electro-optic materials that I'll speak about in a little bit. We can also shape the individual specs to deliver the best propagation performance per watt, which of course is the ultimate objective. Even companies that claim to build their own DSPs may still need to buy critical mixed signal blocks from third parties. That not only adds cost, but more importantly, it can often gate them in terms of time to market and the ultimate performance that they can achieve from the modem. Exploring the theme of vertical integration and various material systems a bit further, the Ciena strategy is vertically integrated by design while remaining fabless. Some might argue that owning the fab itself can offer an advantage, and this is certainly true in specific cases, for example, for cost. But that's only true if the fab runs at near full capacity, and more importantly, only if that particular material system remains the right choice over time and over different applications. The reality, of course, is far more nuanced. Indeed, optimized solutions depend on multiple technologies and material systems, each selected for the needs of that application and adapted as those requirements may evolve. Returning to the key modem building blocks that I showed in the previous charts, the electro-optic functions already rely on a various range of material systems. Indium phosphide, of course, is essential for generating the light that we use in optical communications, and it can also support functions such as modulation. But material systems such as silicon photonics can offer better manufacturability and cost in some of those applications. Looking to the future, there are newer emerging high-speed generations that will be deploying technologies such as thin film lithium niobate and even newer materials and crystals such as barium titanate for certain critical functions. The high-speed analog devices are typically implemented in silicon germanium bipolar or BiCMOS technologies. Lastly, the DSP itself is built on state-of-the-art CMOS. Ciena's WaveLogic 5 generation, for example, was the very first to use 7-nanometer CMOS. WaveLogic 6, again, was the very first to exploit the 3-nanometer CMOS technology. Of course, with an eye to the future, we have a new suite of 2-nanometer devices in development for our next generation of WaveLogic. The key point is that broad-based material system expertise combined with the flexibility to choose the right technology for each application is what's required to consistently deliver the best system-level solution. It's worth spending a moment on R&D leverage and reuse. At Ciena, we view these functions, electro-optics, the electronics, DSP algorithms, together with all the associated material systems, as a suite of reusable technologies. In fact, a toolbox. We evolve and optimize those building blocks to meet the requirements of each target network application. Examples range from data center interconnect, including scale across, through the metro, long haul, and transoceanic networks. We then instantiate those technologies across multiple form factors, including our Waveserver product line, the 6500 platform, suite of coherent modules, and a broad pluggables portfolio that includes QSFP, OSFP, and the emerging XPO formats. The Ciena advantage comes from combining these broad technology ownerships, deep system expertise, field expertise, and the efficient R&D leverage. To summarize this section, Brodie and I covered both our traditional WAN and the emerging AI WAN market segment opportunities, along with the reasons why Ciena continues to win and sustain its market leadership. Firstly, Ciena possesses a distinctive combination of deep multi-technology expertise, extensive system-level knowledge, and most definitely a culture of innovation that's always focused on redefining what comes next. Second, we have a proven ability to turn this innovation into market-leading products and real-world customer solutions. Third, multi-generational field deployment experience that converts this technology and products into operational advantages for our customers. Lastly, the financial scale and strength to continue investing in the innovation and technology development that's required for this fast-changing, demanding market. Of course, the results will speak for themselves. Ciena's estimated market share of 25% in service providers and 36% in the cloud market segment most certainly reinforces that leadership. Thank you. Dino and I are now going to talk about how Ciena is going to expand in and around the data center. In this section, as I mentioned, we are going to focus on inside the data center markets. We are going to have two major sections. Number one is why is AI changing connectivity requirements in and around the data center? Section two, we are going to focus on four use cases: Data Center Out-of-Band Management, or DCOM, scale-up, scale-out, scale-across, inclusive of Coherent-Lite technology. We will talk to the trends, what our solution is, and why we differentiate. What are the key challenges that we see with growth and scale inside the data center driven by this AI connectivity? Number one is just the data center infrastructure management. That has traditionally been built with copper cabling, Ethernet leaf and spine switches. It takes up rack space, it takes up power, and is really not set up to provide management at AI scale, and a new solution is required there. On scale-across, we have talked about the trends, but our hyperscaler customers cannot get GPU in the density and quantities they require in one data center. They require multiple data centers to house those GPUs, and now they need lossless, high-capacity connectivity over distance. On scale-out, there are new density requirements. There is insufficient front panel density for traditional transceivers. They are running into power and scaling issues, and as rate gets increased, the ability to support with traditional technologies goes away. For shorter distance scale-up, passive copper is getting shorter and shorter, again, as rates increase. We will talk to all four of these use cases and what our solution is. We will start with Data Center Out-of-Band Management, go to our Nitro linear redriver solution, talk to our Vesta CPX portfolio, and then finally, for scale-across and campus use cases, we will talk to Coherent-Lite. As data centers expand and the number of devices continues to grow within a data center, the out-of-band management solution needs to change. As I mentioned previously, the traditional solution consists of Ethernet switches and console servers that sit within the rack. Those are connected with multiple copper cables in a redundant fashion to an aggregation layer, and then that aggregation layer is connected up to a spine or router layer, again, through copper cabling. The challenge with that is that that copper cabling is very expensive and very complex, hard to track where those cables are going and why. The Ethernet switches and console servers take up rack space that could be utilized for TPUs or GPUs within the rack. They take up a lot of power, they take up a lot of space, and it is an extremely complex and static solution. This out-of-band management solution was ripe for new innovation. Recognizing that a new solution was required, Ciena decided to try to leverage its PON technology to provide a much lower cost, lower power, lower space solution, leveraging three major components. One is an 8112 converged router. That is a routing platform that we have built. Two is a PON OLT and ONU that you can plug in to that converged router or at the edge of the network. Three is a data center-optimized ONU that can sit now on top of the rack or underneath the rack, but not take up the rack space within it. It's passive optical networking. It requires zero rack units within the rack, and it's much higher density than existing solutions. We took that idea and that technology, we'd co-developed it with a leading hyperscaler, which was Meta, and we've been in deployment now since Q4 of 2024. I think we've shipped over 300,000 ONUs on top of racks now year to date, so we're scaling within that customer, but we also have multiple other customers that we're engaged with now as well. We're obviously not standing still with our DCOM solution. We're going to continue to lead and scale our existing solution, drive security improvements, speed of deployment, scale of deployment. We're going to go after new hyperscaler and neoscale demand and opportunities with new ONUs and OLT hardware. We've containerized our DCOM controller that allows us to deploy not only in our solutions or our routers, as well as third-party routing platforms that may already be inside the data center. We're leveraging our own software, but we're also working and contributing to have SONiC support for our solution as well. Finally, this market, of course, is applicable to multiple hyperscalers, but it's also applicable to the enterprise and small data centers as well. Each opportunity won't be as large as a hyperscaler, but they'll add up over time. So we're building general data center market applications, inclusive of having software, off-box software management and support capabilities. So we're reusing and repackaging the technology to address a broadening opportunity within the hyperscaler and neoscalers, as well as with the general market. Why is Ciena going to win in this DCOM market? Number one, we have great relationships and deep customer engagements with Meta, and we're going to continue to evolve this solution with that customer. We're engaged with multiple other hyperscaler and neoscale customers with this DCOM solution. We're focused on DCOM. We're not focused on the residential broadband access market. We're really focused on this DCOM market. We have a proven platform. We have outputted, as I mentioned, hundreds of thousands of units already. And we have flexible routes to market. I mentioned the containerized DCOM solution. That allows us to deploy again in our products, but also third-party routers and switches. What's not mentioned here, and is a key differentiator again, is our services. When you're deploying hundreds of thousands of units, the ability to commission, install, deploy, and support that type of scale globally is critical for our customers. Along with the technology and product, our services are a key differentiator here as well. As the interconnect rates continue to climb, the same physics that drove the coherent inflection point in the WAN is now forcing changes in data center interconnect architectures and technologies. This is showing up in two major technology shifts. The first, from copper to optical, and the second, from IMDD signaling optics to coherent and DSP-assisted optics. Inside the data center, interconnect rates are typically paced by the IO, or so-called lane speeds of the switch ASICs. The graphic on the left shows how these devices have evolved over time, with IO lane speeds shown on the y-axis. While most deployments today still use 100 gig lanes, 200 gig lane devices have emerged in 2026. Each switch ASIC lane rate generation maps to a typical interconnect rate that's shown on the y-axis of the graphic on the right. For example, eight by 100 gig lanes produce an 800 gig interconnect, with higher rates following the same progression as you move up the y-axis. The same chart on the right shows which technologies apply across different reaches, from meters to kilometers that are shown along the x-axis. It also roughly maps the familiar application categories, such as scale up, scale out, and scale across against those reaches. Although exact distances always vary, the technology transitions are clear. As interconnect rates increase, copper is useful over progressively shorter and shorter reaches, eventually shrinking to only a few meters. That space is being addressed by traditional IMDD or gray optics, and the emerging near-package and co-package optical approaches. Similarly, as rates continue to climb, traditional IMDD optics also become reach-limited, creating the need to transition to coherent-enabled optical solutions. Before turning to optical technologies, it's worth spending a few minutes on copper. Even the strong proponents of optics would agree that staying in the electrical domain as long as possible generally delivers the lowest cost and power. Not surprisingly, there are several copper-based solution options. The simplest is shown at the top, which is direct-attach copper, or DAC. It is by far the simplest and lowest power interconnect approach, and as the previous chart shows, it quickly reaches its performance limits as the bit rates go up. To extend the copper reach, the industry adopted active electrical cables, or AECs. These cable assemblies include a full transmit and receive electrical retiming at each end. While they improve the performance, they also add significant cost and power consumption. As these interconnect rates continue to rise, however, that performance boost becomes less meaningful, and this leads to a third option which has emerged, the active copper cable, or ACC. Based on a small, highly specialized Ciena linear device, an ACC can provide a meaningful performance improvement over a purely passive cable while using only a fraction of the cost and power of a full AEC. The table on the right compares ACC with AEC at 200 gig per lane. Most notable, ACC consumes barely one-sixth the power of AEC while giving up only a very small amount of reach and performance. The Ciena solution for active copper cables is a custom device known as the Nitro 2004. It is implemented in high-speed silicon germanium bipolar technology with four devices integrated into each ACC cable. It delivers stronger link performance than competing solutions while maintaining very low power consumption. It also preserves the operational simplicity of the passive DAC approach, plug-and-play operation, no configuration, no special power feeds required. We are currently qualifying the solution with multiple customers, including hyperscaler and ODMs. As discussed, copper-based solutions are attractive for their simplicity, cost, and power efficiency, but they eventually reach their performance limits, and at that point, the industry has to shift to optical IMDD solutions. For several generations now, the dominant model has been to connect the switch ASICs lanes directly to the front panel pluggable optical modules themselves. But as those interconnect rates increase, the signal propagation, even within the box, becomes more and more difficult, and that often drives these costly, power-hungry electrical retimers, similar to the ones I just described in the AEC on the previous page. The answer, of course, is to move the optics much closer to the switch ASIC, reducing the signal propagation channels and eliminating the need for expensive power-hungry retimers. This approach is known as co-packaged optics, or CPO. Some single-vendor implementations have emerged, but that ecosystem or that closed ecosystem model really undermines the supply chain advantages of today's system. Specifically, how do you maintain separate supply chains for the switch device and the optics while preserving that customer choice across multiple suppliers? The second, equally important, is decoupling the switch and optics innovation cycles. How do the data center operators upgrade each independently rather than being forced to change everything at the same time? The optimal approach preserves the advantages of both models, of course. Use modular pluggable optical modules that can be assembled directly beside the switch ASIC, much like the front panel pluggables, but connected to an internal interconnect assembly. Each module is individually pre-tested and confirmed as a known good part before being attached to that sub-assembly, and then ultimately support an open ecosystem through this Open CPX. This is an MSA that was launched over the past six months. Ciena is a founding member and holds both the chair and editor positions. The group now includes nearly 50 members. It should be noted that Open CPX and the Vesta 200 both support the CPO as well as the related variant NPO, or near-packaged optics. The Ciena solution is known as the Vesta 200 CPX. Each Vesta module provides a full bidirectional 6.4 terabit interface, incorporating 32 lanes of 200 gig optical IMDD. A set of 16 modules can deliver the full 100 terabit switch interconnect that's shown on the right-hand side of the slide. Ciena's individual modules combine highly specialized photonic and electronic ICs that deliver superior interconnect performance, very low power, and proven high reliability. Vesta also uses a 2D optical connector array for maximum density. This positions the solution very well for the next generation 200 terabit switches. They will require 32 of these modules. At present, we are sampling devices for multiple key customers, both hyperscaler as well as switching OEMs. Vesta technology certainly extends optical IMDD to its practical physical limits, but beyond that point, we have to move into the coherent optical domain. Data center use cases do vary, but generally fall into the three categories shown on the left. The first is just reach. As data centers have grown, servers and switches no longer fit within a single building, turning large facilities into multi-building campuses. Combined with the need to find suitable space and power, this has pushed interconnect distances from a few kilometers to as much as 10 to 20 km. The second use case is sheer fiber utilization. At shorter distances, operators have often dedicated a single fiber pair for each interface, but as both scale and distance increases, multiplexing multiple interfaces across different wavelengths becomes more and more important, and this will enable fiber utilization gains of 20 times or even more. The third case is the optical circuit switch, or OCS. As OCS is introduced, it inevitably adds optical impairments such as loss and/or noise, and therefore increasing the performance required from the interconnect technologies. Considering the 1.6T generation, these scenarios, individually or in combination, push beyond the practical reach of traditional optical IMDD signaling. 1600ZR technology could certainly address this requirement, but it is likely overkill for these applications and adds unnecessary cost and power. The optimal answer combines the best of both approaches, power and cost close to what an enhanced IMDD would require, with the performance needed to support the longer reaches and emerging use cases, and that solution is Coherent-Lite. Returning to our reusable building block approach, Ciena can select and optimize elements from our foundational coherent toolbox to target this Coherent-Lite application, covering the inside the data center, the campus, and selected scale-out environments. As a leader in coherent technology, Ciena can draw on a deep portfolio of algorithms, techniques, and system expertise to bring a Coherent-Lite solution to market. Our next-generation development includes an optimized Coherent-Lite engine, which is built around a dedicated 2-nanometer CMOS DSP. Combined with a new set of optimized electro-optics, this technology will support a 1.6 and 3.2 terabit Coherent-Lite plug in the OSFP form factor, as well as an ultra-high density 12.8 terabit plug for the next generation of liquid-cooled switches. Ciena also supports multiple routes to market, from complete pluggable assemblies down to individual components, including a standalone DSP. Building on that multiple routes to market theme, much of our interconnect technology portfolio is already available at multiple levels of the ecosystem. Our systems business certainly uses the full breadth of these interconnect products, from high-performance WaveLogic modules through the complete pluggable portfolio. The Vesta 200 is also likely to be incorporated into a full-spectrum transponder in the near future. Pluggable optics are sold both to end-user data centers, as well as network providers and to networking platform OEMs. At the module level, we also sell to platform partners, and this includes high-performance WaveLogic modems that have been designed into traditional network provider platforms, as well as hyperscaler homegrown boxes. Finally, our ASICs, including the coherent DSPs and the Nitro redriver, are sold as standalone components to module and cable manufacturers. To sum up, Ciena delivers differentiated technologies that solve the critical scaling challenges inside the data center. While the solutions differ by application, they all build on decades of optical communications and high-speed transmission expertise. Our close collaboration with data center operators and hyperscaler supports co-development. It aligns our solutions to the application needs and enables rapid high-volume deployment. Finally, our multiple routes to market reach customers across the ecosystem, maximizing both exposure and our opportunity. Thank you. Hello. My name is Grant Hoffman, and I manage the supply chain at Ciena. I have been with Ciena for just over two months, and I am excited to review progress and the future ahead. An advantage of being new is that you can quickly identify what is truly unique about a company. One thing that has become clear to me is that Ciena's supply chain is not just an operational organization. It is a strategic differentiator that has been built over decades. Today, I would like to walk you through five topics outlining why I believe this to be true. First, our global footprint and operating model. Second, how we invested ahead of growth. Third, how we are securing supply in an increasingly constrained environment. Fourth, how we are scaling through vertical integration and design for excellence, or DFX. Finally, how we are delivering products and services to customers at the scale and speed the market now demands. Before I talk about where we are going, I want to start with a broader perspective. I have spent much of my career in technology and manufacturing businesses that shipped tens or hundreds of millions of units annually. Throughout that experience, one thing became very clear. Supply chains aren't stressed by normal growth. They are stressed during inflection points. Many members of our leadership team have been with Ciena for two or three decades. When I asked them how often they have seen supply chain conditions like this one, the answer surprised me. Only three times. The first was the telecom expansion of the late 1990s. The second was during the pandemic, and now we are experiencing the third event. What is particularly interesting is that there were roughly 20 years between the first and the second events, but less than five years between the second and third. The causes are different, but the outcomes are remarkably similar. Supply shortages, extended lead times, capacity constraints, volatility, and the disruptions are becoming larger and more frequent. What separates winners from losers is whether they can learn from the last disruption before the next one arrives, and that is exactly what Ciena has done. Our partner engagement and geographic footprint is the result of years of strategic decisions. Our R&D and supply chain teams are co-located in Ottawa, accelerating product introduction and problem resolution. We operate through leading global manufacturing partners. We design our own test processes and assets. We maintain global distribution capabilities, and we operate a global reverse supply chain network. Perhaps more importantly, many of these decisions were made before they became obvious. Over a decade ago, we consciously focused manufacturing strategy based on resiliency. We invested heavily in USMCA-based ecosystems across Canada, Mexico, and the U.S. We built regional flexibility before regional flexibility became fashionable. Those decisions today are paying dividends. That is not luck. That is preparation. One of the most important lessons from prior disruptions is that capacity cannot be built overnight. By the time demand arrives, it is already too late, and that is why we have been investing aggressively ahead of growth. You can see that reflected here. We have increased capital investments roughly five times historic levels. We have more than doubled manufacturing capacity, and we have doubled distribution capacity. What is more important is that these aren't theoretical investments. They are translating directly into output capabilities across key product families, interconnects, utilizing our WaveLogic and RLS. Why are we doing this? Because demand is growing faster than traditional planning assumptions anticipated. The challenge for Ciena is not creating demand. The challenge is ensuring our supply chain can convert that demand into revenue and customer success. Our strategy is straightforward: invest ahead of the need, build capacity before it becomes constrained, and maintain the ability to scale alongside our customers. The next capability is one of our strongest differentiators, and that is vertical integration. Historically, network equipment companies relied heavily on third-party suppliers for critical technologies. That model works in stable environments, but becomes risky when demand accelerates rapidly. Our response has been to expand Ciena's development and ownership of critical technologies in our products. To support our growing demand, we leverage both third-party solutions and internally developed components in our WaveLogic products. Why does this matter? Because ownership of a large portion of the components creates control over roadmaps, control over supply assurance, and improved control over cost and execution, with the ultimate objective to reduce dependencies on factors outside of our control and increase our ability to deliver regardless of market conditions. The next three elements of supply assurance after vertical integration are long-term supply commitment, extended purchase agreements, and multi-vendor sourcing. Over the past several years, one lesson has become abundantly clear. When supply environments tighten, relationships matter, planning horizons matter, and balance sheets matter. With that understanding, we have intentionally moved beyond traditional industry practices. For many of our most critical technologies, we are extending purchase commitments and establishing multi-year agreements aligned to our growth outlook through the end of the decade. These commitments focus on some of the most constrained technologies in our industry, particularly across photonics, optical components, and modem-related technologies. In many cases, we have gone a step further by helping suppliers fund capacity expansion and technology investments. Why is that important? Because when demand outpaces supply, suppliers make choices. They prioritize customers who demonstrate long-term commitment, strategic alignment, and the financial strength to support growth. Ciena's balance sheet and disciplined approach have allowed us to become a preferred partner, creating access to future capacity and strengthening our ability to support customer demand. At the same time, we recognize that resiliency requires more than capacity. Every major industry disruption has reinforced a simple truth: single-source dependency creates risk. That is why we have deliberately expanded and qualified multiple suppliers across products and future platforms. We continue to broaden participation across strategically important components. What is particularly encouraging is that many of the leading industry technology providers want to partner with Ciena. They see the same long-term market opportunity that we do. They recognize our growth trajectory, and they understand the importance of our position within the optical networking ecosystem. The result is a supply network that becomes stronger, more resilient, and more scalable as we grow, providing a significant competitive advantage to helping ensure we can continue to deliver for customers through every phase of the market cycle. One very important item immediately stood out to me when I joined Ciena. There is a very close connection between the customer, R&D, and the supply chain. In many companies, products are designed first, and manufacturing figures out how to build them later. That doesn't scale. Instead, Ciena partners closely with our customers on new product specification and partners with supply chain on design for excellence principles. Those are design for manufacturability, design for cost, design for serviceability, design for reliability, and design for test. This means products are engineered with scale in mind from day one. Supply chain, engineering, operations, and manufacturing work together early in the development process, and the benefits are clear: faster product ramps, higher quality, improved cost, and fewer surprises when demand increases. As Dino discussed earlier, we are building on a tremendous engineering foundation. Our role is to ensure that innovation can be delivered at scale, and that is what DFX enables. As we scale the business, our ability to execute is increasingly determined by the strength of our talent and the breadth of our ecosystem. Over the past year, we have deliberately invested in both through targeted hiring, acquisitions, and strategic partnerships. By doing so, we are expanding the capabilities, expertise, and capacity required to support our next phase of growth. Together, through investments in strategic talent and the expanded ecosystem, we create a powerful multiplier effect. This strengthens our ability to retain customers, capture new opportunities, and scale efficiently, creating a durable foundation for sustained growth and long-term value creation. I am going to close with a supply chain capability that is core to customer retention, and that is service and delivery. At the end of the day, customers buy more than a product. They buy outcomes, products delivered on time, reliable deployments, strong operational support, and the ability to scale when their business scales. We combine traditional network expertise with hyperscaler-inspired agility to deliver a seamless customer experience across the entire lifecycle, from advisory and deployment through optimization and ongoing support. With one of the world's largest optical service networks supporting customers in more than 116 countries through extensive field, repair, and logistics capabilities, we help customers deploy faster, operate efficiently, and resolve issues quickly. This global scale and operational excellence strengthens customer loyalty, supports long-term growth, and reinforce our leadership position in the market. The goal is simple: make it easier for customers to grow with Ciena. The environment we are operating in today is one of the most supply-constrained periods of our industry's history, but it is also one of the greatest growth opportunities. The strategy you have heard today is built on four capabilities. We plan ahead of demand. We secure supply through strategic partnerships, long-term agreements, and vertical integration. We build for scale through close integration between engineering and operations, and we deliver through flexible global manufacturing, fulfillment, and service networks. These capabilities did not happen overnight. They are the result of decades of deliberate investment and lessons learned from prior industry disruptions. We diversified our footprint before resiliency became a requirement. We invested in strategic inventory and supplier relationships before capacity became constrained, and we built an operating model designed to scale alongside our customers. While we have done the work, we recognize in today's unprecedented hypergrowth environment, there is always more to do. We believe Ciena has one of the most resilient and scalable supply chains in the industry, positioning us to secure capacity, support our customers, and capture the growth opportunity ahead. We have planned for this moment, we have invested for this moment, and we are ready for the scale ahead. Thank you. Hello, everyone, and thank you for joining. This is Marc Graff, Ciena's Chief Financial Officer. Today, I will update you on the journey we have been on for the past few years, the amazing results we have achieved, and our plans and financial objectives through 2029. It was a little more than a year ago when Ciena laid out its near-term financial goals, namely structurally improving gross margins, driving towards world-class working capital, and having a focused capital allocation strategy. I am happy to report we have made excellent progress against all three of these objectives, as they have been the core on delivering the fundamentals for Ciena's business model. You have seen gross margins improve by over 200 basis points year-on-year through an aggressive cost reduction program, pricing actions, and accretive product mix. These dynamics are sustainable and have reset our gross margin base to solidly mid-40s. Our working capital management has been a focus not only for financial hygiene, but also to power the growth of the company. We have the resources to fund our growth over the long term and to ensure that our capital is deployed to meet growing customer needs while reinforcing efficiencies such as cash conversion cycles. Lastly, we have a disciplined capital allocation framework that focuses on investing in our current leadership portfolio as its top priority and maintaining an eye to inorganic opportunities as our second priority. Being the only pure-play optical systems and interconnect solution provider and a leader in the industry is the result of decades of deliberate investment decisions, trade-offs, and execution to meet this extraordinary market opportunity. Lastly, when capital is generated beyond these needs, we are committed to returning it to our owners, which over the last five years has averaged over 100% of free cash flow. The results speak for themselves. Since 2024, revenue has grown an average of 27% to our current $6.4 billion 2026 guides. Operating margins have expanded by 1,100 basis points over the same timeframe, and earnings per share have more than tripled. It is this earnings potential that we are unleashing. We have tripled our earnings per share with only a 27% revenue CAGR. We are able to deliver these results because we are the only pure play optical systems and interconnect provider on the planet, and that allows us massive leverage for both research and development and go-to-market activities. These impressive results have been gated only by industry supply. As we have talked previously, we are seeing an acceleration in the demand for optical systems and interconnects as customers realize the network is the unlock to monetizing AI investments. Over the past two years, we have seen orders, a proxy for unconstrained demand, double from 2024 to 2025, and increase at least another 50% into 2026. Yet revenue, a proxy for supply in a constrained environment, has increased only 27% on average over the past two years. As a result, over the same period, backlog has more than doubled from 2024 to 2025, and again from 2025 to 2026, to an expected $10 billion exiting 2026. Clearly, there is not a demand issue. What is clear is the need for the industry to accelerate its capacity requirements for a number of years to catch up to demand. While it is typically seen as a negative indicator, industry lead times need to compress in order to establish a healthier balance between supply and demand, effectively providing a cash unlock to a growing backlog. Based on our latest outlooks, we do not see a balanced supply-demand environment returning before 2028. As it applies to our directional outlook for 2027, we see a very similar dynamic occurring, namely demand outstripping supply. We continue to believe that we will exit the fiscal year with at least a 50% growth in orders to yield a minimum $10 billion in backlog, which covers our 2027 revenue outlook and starts to build the 2028 revenue base. As importantly, we have secured the supply to enable at least a 30% revenue growth going into 2027. Our supply chain team has pulled all the levers to ensure that we can deliver this floor and position Ciena for upside as additional supply may become available. In fact, we believe that we have secured supply to meet demand to extend the 30% trend through 2029, yielding a 2029 revenue target of approximately $14 billion. With customer commitments and long-term supply agreements, our revenue will exceed what we believe the top four hyperscalers' CapEx growth rate to be. Additionally, the competitiveness of our portfolio will enable continued share gains and exceed the TAM growth rate, driven primarily by our optical systems such as Hyper-Rail and WaveLogic 6e, and accelerated by the massive growth in our interconnects portfolio of plugs, DCOM modules, and the beginning of the CPO/NPO product ramps. We have requested supply commits for 2028 and 2029 to exceed these growth rates. The next three years will be an inflection point for the company as it cements its leadership position in the traditional and AI WAN and establishes its footprint inside the data center. While our ambitions are great inside the data center, we have taken a relatively modest approach to 2029 revenue from newer product lines like CPO, NPO, and Coherent-Lite. As Brodie and Dino reinforced, the move to optical within the data center is accelerating, and Ciena leads the industry in optical connectivity expertise. These products could provide significant upside to our baseline assumptions as the markets and applications develop for them. Together, there is $11 billion of 2029 TAM for these new markets in which we can participate. We are not relying only on growing revenue to drive profitability. We believe that we can achieve 50% gross margins by 2029, underwritten by our industry leadership. First, as I have talked about previously, we have made excellent progress on pricing and term discussions with our customers that extend over the next three years. We will see these actions start to impact 2027 and annualize for the remainder of the three-year period. Second, we will see improved unit costs as new products like WaveLogic 6n, our 800 gig plug, continue to ramp to mature volumes. Our engineering efforts to reduce product costs through both design and technology improvements continue to be an active investment. Last, and perhaps most exciting, is our evolving product portfolio delivering increased value, allowing structural improvements to our margins. Products like RLS Hyper-Rail and DCOM that deliver meaningful TCO advantages, and modules and components that have improved margin profiles underwrite structural improvements to our margins. With our ability to deliver more value to the market, we are also seeing the benefits of our business model take hold. On top of gross margin improvements that I've just discussed, we are able to get significant leverage from our R&D investments in optical technology across multiple product lines. This reusable engineering model proliferates leadership discoveries that can be productized for multiple use cases. This creates an enormous amount of operating leverage for the company. On top of that, we see expanding leverage from our go-to-market efforts as well. As our largest customers are multi-segment, multi-use case customers, we are able to deepen the co-creation and co-development efforts while growing revenue, creating additional leverage. Taken together, gross margin improvements and increased operating leverage, we see a path to an operating margin of 32%-35% by 2029. Our earnings strength is directly translating to improving what is an already strong balance sheet. With the completion of the 2026 convertible debt issuance, we have the strategic financial capability to continue to invest in the business while taking advantage of inorganic opportunities. With the scale that we are seeing in the business, we will continue to generate significant amounts of cash from operations, with an expectation of approximately quintupling or 5X-ing cash generated from operations. In addition, with the growth of earnings, I expect that we will take our already low net leverage ratio down to zero or even below. Which begs the question, what are we doing with all this capital? Ciena has a history of being a thoughtful and deliberate steward of our owners' capital, and that trend will continue. As I noted before, our first priority is to invest in the organic opportunities of the business, such as next generation of WaveLogic technologies, new material systems, and products that compete inside the data center. In fact, we are currently estimating that we will invest somewhere between $2.5 billion and $3 billion in R&D over the next three years. Second, we are constantly looking across the ecosystem for technologies and talent that will accelerate our market leadership. You've seen us take decisive action with our Nubis acquisition, which has positioned us well for expanding inside the data center. With our recently announced $200 million Ciena venture fund, we will take an even more active stand to seek out those opportunities. In total, our balance sheet has given us the flexibility of a BB+ rating and roughly $20 billion of M&A capacity to deploy. Lastly, once our organic and inorganic opportunities have been captured, we will return excess capital to our owners. Over the past five years, we've returned a cumulative 108% of our free cash flow to our owners. Moving forward, we are committed to returning a minimum average of 70% of free cash flow back to our shareholders, barring any major M&A activity. Overall, we are excited about the opportunities to generate leading returns for our owners. It's against this backdrop that we are realigning our reporting segments. It has been seven years since we adjusted our reporting structure, and the business has dramatically changed over that timeframe. First, we've seen the rise of the hyperscalers and AI as a major demand driver. Second, our product portfolio has become even more optically focused. Third, our opportunity inside the data center represents a new growth frontier with our interconnects portfolio. Last, but certainly not least, this new structure simplifies our ability to communicate with the investment community while increasing the transparency of our different segments. Specifically, we will report four segments that align to our $52 billion 2029 TAM. Optical systems with an estimated $21 billion TAM, growing at roughly an 18% CAGR through 2029, will include our line systems, WaveLogic, and Waveserver systems, as well as the Navigator Network Control Suite. Our interconnects portfolio will include the coherent plugs, DCOM, or Data Center Out-of-Band Management, our Nitro and Vesta products from the Nubis acquisition, as well as optical and electrical components. This segment is expected to grow at an average annual rate of 95% through 2029 to a TAM of $17 billion. Our global services segment will now include platform software services revenue and is expected to grow 15% with a 2029 TAM of $2 billion. Finally, we have combined our non-DCOM routing and switching business with Blue Planet into routing and other, with a 2029 TAM of $12 billion and a CAGR of 2%. We are excited about this new structure that will take effect with fiscal 2027 and be first reported with the Q1 2027 results. Pulling it all together, we are committed to a step function improvement in our financial model. Our three-year revenue target will be a 30% CAGR through 2029, achieving a $14 billion revenue level at roughly 50% gross margins. After 15%-18% operating investments, our business model's leverage will accelerate our operating margins to the mid-30s, with a free cash flow margin expected to be approximately 20%. Finally, we expect our adjusted tax rate to remain in the 20% range. As you can see, we are translating revenue growth into massive earnings power. To sum it all up, Ciena is in an excellent position to continue and, in fact, to increase our earnings power, leading to sustained returns for our owners' investments. We are committed to building on and extending our market share in these growing optical markets. We believe the inevitability of the data center opticalization plays to our strengths as both an optical systems expert and a trusted partner to global hyperscalers, neoscalers, and service provider customers. The foundation of our pure play optical systems and interconnects business model will drive significant earnings expansion as the world moves from copper and electrons to fiber and photons. We are absolutely committed to being good stewards of our owners' capital by investing first in the business and returning excess capital to our owners. Put simply, Ciena and its owners are positioned to enjoy increasing profits and returns. Thank you.

Data Source: Quartr Last updated: Sep 18, 2026

Earnings call transcripts are provided by Quartr and are published within hours after each earnings call.

Update frequency:

Quarterly

Data Sources Report a data error

• Slides